Forest Ecology

Forest Ecology
Forest Ecology FIFTH EDITION
Daniel M. Kashian
Wayne State University,
Detroit, USA
Donald R. Zak
University of Michigan,
Ann Arbor, USA
Burton V. Barnes (deceased)

University of Michigan,
Ann Arbor, USA
Stephen H. Spurr (deceased)

University of Texas at Austin,
Austin, USA
This fifth edition first published 2023
© 2023 John Wiley & Sons Ltd
Edition History
John Wiley & Sons Ltd. (4e, 1998); Ronald Press (3e, 1980; 2e, 1973; 1e, 1964)
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in
any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, except as permitted by
law. Advice on how to obtain permission to reuse material from this title is available at http://www.wiley.com/go/
permissions.
The right of Daniel M. Kashian, Donald R. Zak, Burton V. Barnes, and Stephen H. Spurr to be identified as the
authors of this work has been asserted in accordance with law.
Registered Offices
John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA
John Wiley & Sons Ltd., The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK
Editorial Office
9600 Garsington Road, Oxford, OX4 2DQ, UK
For details of our global editorial offices, customer services, and more information about Wiley products, visit us at
www.wiley.com.
Wiley also publishes its books in a variety of electronic formats and by print-on-demand. Some content that appears
in standard print versions of this book may not be available in other formats.
Limit of Liability/Disclaimer of Warranty

While the publisher and authors have used their best efforts in preparing this work, they make no representations
or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim
all warranties, including without limitation any implied warranties of merchantability or fitness for a particular
purpose. No warranty may be created or extended by sales representatives, written sales materials, or promotional
statements for this work. The fact that an organization, website, or product is referred to in this work as a
citation and/or potential source of further information does not mean that the publisher and authors endorse the
information or services the organization, website, or product may provide or recommendations it may make. This
work is sold with the understanding that the publisher is not engaged in rendering professional services. The advice
and strategies contained herein may not be suitable for your situation. You should consult with a specialist where
appropriate. Further, readers should be aware that websites listed in this work may have changed or disappeared
between when this work was written and when it is read. Neither the publisher nor authors shall be liable for any
loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or
other damages.
Library of Congress Cataloging-in-Publication Data

Names: Kashian, Daniel M., author. | Zak, Donald R., author. | Barnes,
Burton Verne, 1930-2014, author. | Spurr, Stephen H., 1918-1990, author.
Title: Forest ecology / Daniel M. Kashian, Wayne State University, Detroit,
USA, Donald R. Zak, University of Michigan, Ann Arbor, USA, Burton V.
Barnes (deceased), University of Michigan, Ann Arbor, USA, Stephen H.
Spurr (deceased), University of Texas at Austin, Austin, USA.
Description: Fifth edition. | Hoboken, NJ : Wiley, 2022. | Precedey by:
Forest ecology / Burton V. Barnes ... [et al.]. 4th ed. c1998. |
Includes bibliographical references and index.
Identifiers: LCCN 2022037810 (print) | LCCN 2022037811 (ebook) | ISBN
9781119476085 (paperback) | ISBN 9781119476054 (adobe pdf) | ISBN
9781119476146 (epub)
Subjects: LCSH: Forest ecology.
Classification: LCC QK938.F6 F635 2022 (print) | LCC QK938.F6 (ebook) |
DDC 577.3–dc23/eng/20220831
LC record available at https://lccn.loc.gov/2022037810
LC ebook record available at https://lccn.loc.gov/2022037811
Cover Design: Wiley

Cover Image: © U.S. Forest Service Photo
Set in 9.5/12pt STIXTwoText by Straive, Pondicherry, India

Dedication
We dedicate the 5th edition of Forest Ecology to the co-author of the 2nd and 3rd
editions and the lead author of the 4th edition, Burton V. Barnes (1930–2014). It would
be no easy task to find a more accomplished and humble leader in his field. He excelled
at his science, was a truly beloved teacher, and helped to shape the world view of
thousands of colleagues, friends, students, managers, and scientists alike, all with an
unmatched humor and a love of the natural world.
Burt Barnes was world-renowned as an expert in the ecology of North American aspens
and the ecological classification of forest ecosystems. His professional training was in
forest ecology, botany, and genetics, but he dabbled heavily in glacial geomorphology,
soil science, phytogeography, and woody plant physiology. Perhaps his greatest love,
however, was teaching, especially in the field, which drove his motivation for this
textbook. Generations of students have been touched by his love for the art and science
of teaching field ecology, which they will forever pass on to future generations. We, as
authors of this edition, have been personally and professionally shaped by him as a
mentor, colleague, and friend. His legacy is therefore unending. To him we say, in his
own words, “Thanks for everything you have done—and will do.”
Contents
Prefacexxiii
PA R T 1 Forest Ecology and Landscape Ecosystems
1 Concepts of Forest Ecology 3

Ecology, 4
Landscape Ecosystems, 4
Landscape Ecosystem and Community, 7
Ecosystem Structure and Function, 7
Examples of Landscape Ecosystems, 9
An Approach to The Study of Forest Ecology, 11
Applicability to Forest Management, 12
Suggested Readings, 13
2 Landscape Ecosystems at Multiple Scales 15

Overview of Spatial and Temporal Scales, 15
Spatial Scales of Hierarchical Landscape Ecosystems, 17
Climatic Classification, 18
Physiography, 20
Vegetation Types and Biomes, 20
Distinguishing and Mapping Landscape Ecosystems
at Multiple Spatial Scales, 25
Regional Landscape Ecosystems, 26
Regional Landscape Ecosystems of Michigan, 28
Local Landscape Ecosystems, 30
Local Landscape Ecosystems in Upper Michigan, 30
vii
viii C o n t e n ts
PA R T 2 The Forest Tree
3 Forest Tree Variation 35

Components of Phenotypic Variation, 35
Plasticity of the Phenotype, 36
Sources of Variation, 37
The Evolutionary Sequence, 38
Sexual and Asexual Systems, 39
Genetic Diversity of Woody Species, 39
Genecology, 40
Patterns of Genecological Differentiation, 41
Genecological Categories, 42
Factors Eliciting Genecological Differentiation, 42
Growth Cessation, 43
Growth Resumption, 46
Examples of Genecological Differentiation, 46
Eastern North American Species, 47
Scots Pine, 49
Wide-Ranging Western North American Conifers, 50
Ponderosa pine, 50
Douglas-fir, 52
Local Genecological Differentiation, 55
Factors Affecting Differentiation: Gene Flow and Selection Pressure, 56
Ecological Considerations at the Species Level, 57
Niche, 60
Hybridization, 60
Polyploidy, 63
The Fitness–Flexibility Compromise, 66
Epigenetics, 66
4 Regeneration Ecology 69
Regeneration, 69
Sexual Reproduction, 71
Maturation and the Ability to Flower, 72
Increasing Seed Production, 73
Reproductive Cycles, 74
Pollination, 75
Contents ix
Periodicity of Seed Crops, 76

Effects of Reproduction on Vegetative Growth, 78
Dispersal, 80
Seed Bank, Dormancy, and Germination, 82
Establishment Following Sexual Reproduction, 83
Post-Establishment Development, 90
Vegetative Reproduction, 90
5 Tree Structure and Growth 95

Tree Form, 95
Architectural Models, 97
Short and Long Shoots, 99
Patterns of Intermittent Growth, 100
Sylleptic and Proleptic Shoots, 102
Roots, 102
Kinds, Forms, and Occurrence, 103
Fine Root Relations, 105
Horizontal and Vertical Root Development, 107
Periodicity of Primary Root Growth, 108
Root Grafting, 108
Specialized Roots and Buttresses, 110
Stems, 110
Xylem Cells and Growth Rings, 111
Periodicity and Control of Secondary Growth, 112
Control of Earlywood and Latewood Formation, 114
Winter Freezing and Water Transport, 115
Water Deficits and Tree Growth, 116
PA R T 3 The Physical Environment

Forest Environment, 119
Site Factors, 120
Organization of Site Factors, 120
Interrelationships Among Site Factors, 121
Importance of Site in Forest Ecology, 122
0005430834.INDD 9 01-05-2023 17:43:22

x C o n t e n ts
6 Light 123
Distribution of Light Reaching the Ecosphere, 124
Plant Interception of Radiation, 125
Canopy Structure and Leaf Area, 125
Light Quality Beneath the Forest Canopy, 130
Sunflecks, 130
Light and Growth of Trees, 133
Light and Seedling Survival and Growth, 136
Light and Tree Morphology and Anatomy, 137
Light and Epicormic Sprouting, 139
Photocontrol of Plant Response, 139
Light and Ecosystem Change, 141
7 Temperature 143
Geographical Patterns of Temperature, 143
Temperatures at the Soil Surface, 145
Temperature within the Forest, 147
Temperature Variation with Local Topography, 148
Temperature and Plant Growth, 150
Cold Injury to Plants, 153
Dormancy, 155
Frost Hardiness and Cold Resistance, 156
Thermotropic Movements in Rhododendrons, 159
Winter Chilling and Growth Resumption, 161
Natural Plant Distributions and Cold Hardiness, 162
Deciduousness and Temperature, 164
8 Physiography 167
Concepts and Terms, 167
Characteristics of Physiography and their Significance, 168
Physiographic Setting, 169
Specific Landforms, 170
Elevation, 170
Form of Landforms, 170
Level Terrain, 170
Sloping Terrain, 171
Slope Characteristics, 171
Contents xi
Position on slope, 172

Aspect, 172
Slope inclination, 173
Parent Material in Relation to Landform, 173
Position of Landform in the Landscape, 174
Multiple Roles of Physiography, 174
Physiographic Diversity, Landscape Ecosystems, and Vegetation, 176
Mountainous Physiography, 176
Mountainous Terrain of California and the Pacific Northwest, 177
Physiography and Forests of the Central Appalachians, 180
Flatlands, 183
The Great Plains, 184
Pine Savannas of the Western Great Lakes Region, 185
Till Plains of the Midwest, 185
Southeastern and Southern Coastal Plain, 186
Floodplains, 186
Physiography and Firebreaks, 190
Microlandforms and Microtopography, 191
Tree Uprooting and Pit-and-Mound Microtopography, 191
Microtopography and Regeneration in Hardwood Swamps, 193
9 Soil 195
Parent Material, 195
Soil Formation, 197
Soil Profile Development, 197
Physical Properties of Soil, 200
Soil Texture, 200
Soil Structure, 201
Soil Color, 202
Soil Water, 202
Physical Properties of Water, 203
Soil Water Potential, 204
Chemical Properties of Soil, 206
Clay Mineralogy, 207
Cation Exchange and the Supply of Nutrients, 209
Soil Acidity, 210
Soil Organic Matter, 212
Soil Classification, 213
Landform, Soil, and Forest Vegetation: Landscape Relationships, 215
xii C o n t e n ts
10 Fire 219
Fire and the Forest Tree, 220
Causes, 220
Fire Regime, 220
Fire Types, Frequency, and Severity, 221
Fire Adaptations and Key Characteristics, 224
Strategies of Species Persistence, 227
Closed-Cone Pines, 228
Fire and the Forest Site, 230
Indirect Effects, 230
Direct Effects, 232
Organic Matter and Erosion, 234
Beneficial Effects of Fire, 236
11 Site Quality and Ecosystem Evaluation

and Classification237
Direct Measurement of Forest Productivity, 238
Tree Height as a Measure of Site, 239
Site-Index Curves, 240
Comparisons Between Species, 243
Advantages and Limitations, 243
Vegetation as an Indicator of Site Quality, 244
Species Groups of Ground Cover, 245
Indicator Plants of Coastal British Columbia, 246
Ecological Species Groups, 246
Plant Associations and Habitat Types in the Western United States, 247
Operational Site Classification Based on Vegetation, 247
Applications and Limitations of Vegetation, 250
Environmental Factors as a Measure of Site, 251
Climatic Factors, 251
Physiographic Land Classification, 252
Physiographic and Soil Factors: Soil-Site Studies, 252
Soil Surveys, 256
Multiple-Factor Methods of Site and Ecosystem Classification, 256
Ecosystem Classification and Mapping in Baden-Württemberg, 257
Applications of Multi-Factor Methods in the United States
and Canada, 258
Ecosystem Classification and Mapping in Michigan, 258
Ecosystem Classification in the Southeastern United States, 261
0005430834.INDD 12 11-23-2022 12:32:40

Contents xiii
Ecosystem Classification in the Southwestern United States, 262

National Classification Systems in the United States, 262
Ecological Land Classification in Canada, 263
Hills’ physiographic approach, 263
Other approaches used in Canada, 264
PA R T 4 Forest Communities
12 Animals in Forest Ecosystems 269

Plant Defense, 269
Investment in Plant Defense, 270
Plant Defense Against Insects, 272
Examples of Injury and Plant Defense, 273
Nutrition, 274
Plant Hybrid Zones as Reservoirs for Insect Diversity, 275
Plant Defense Against Mammals, 276
Roles of Animals in Plant Life History, 277
Pollination, 277
Seed Dispersal, 278
Fish and Reptiles, 278
Birds, 279
Mammals, 282
Germination and Establishment, 284
Decomposition, Mineral Cycling, and Soil Improvement, 285
Damage and Death, 286
Influence of Livestock on Forest Ecosystems, 288
13 Forest Communities 291

Community Concept, 291
Grounding Communities, 292
Florida Keys, 292
Interior Alaska, 294
Southern Illinois, 294
View from the Past: Community Concepts, 296
Schools and Terminology, 296
Concepts of Clements and Gleason, 297
0005430834.INDD 13 11-23-2022 12:32:40

xiv C o n t e n ts
Phytosociology in Europe, 298

Continuum Concept, 298
Community as a Landscape Ecosystem Property, 299
Examples of Spatial Variation in Forest Communities, 300
Discrete Forest Communities, 300
Coastal California: Giant and Pygmy Forests, 301
Forest–Grassland Ecotone, 302
Alpine Tree Lines, 302
Merging Forest Communities, 303
Eastern Deciduous Forest—Southern Appalachians, 303
New England, 304
Competition and Niche Differentiation, 305
Interactions Among Organisms, 307
Mutualisms in Forest Ecosystems, 307
Symbiotic Mutualisms—Mycorrhizae, 307
Nonsymbiotic Mutualisms, 307
Competition, 308
Forest Community Structure and Composition, 308
Vertical Structure, 310
Stand Density, 312
Competition and Overstory Composition, 313
Competition in the Understory, 314
Understory Tolerance, 315
Characteristics of Understory-Tolerant and -Intolerant Species, 316
Tolerance Ratings of Tree Species, 317
Examples of Understory Tolerance in Forest Ecosystems, 319
Nature of Understory Tolerance, 319
Environmental Factors Relating to Understory Tolerance, 322
Physiological Processes Relating to Tolerance, 323
14 Diversity in Forests 327

Concepts of Biological and Ecosystem Diversity, 327
The Value of Species Diversity, 328
Value of Biodiversity, 329
Common Threats to Diversity, 331
Measuring Diversity, 331
Levels of Diversity, 331
Measurement, 333
Inventory Diversity: Alpha Diversity, 333
Differentiation or Beta Diversity, 335
0005430834.INDD 14 11-23-2022 12:32:40

Contents xv
Diversity of Landscape Ecosystems, 336

Examples of Diversity, 337
Ground-Cover Species Diversity in Northern Lower Michigan, 337
Ecosystem Groups, 337
Ecosystem Types, 340
Ecosystem Diversity, 342
Causes of Species Diversity, 343
Diversity at Continental and Subcontinental Scales, 343
Paleogeography and Continental Relationships, 343
Glaciation, 344
Latitude and Elevation, 344
Diversity at Local Scales, 346
Physiography and Soil, 346
Community Composition and Structure, 349
Disturbance and Succession, 349
Focal Species in Conserving Diversity, 351
Foundation Species, 352
Keystone Species, 352
Endemics and Rare and Endangered Species, 353
Diversity and the Functioning of Ecosystems, 354
Biodiversity–Productivity Relationship, 354
The Role of Biodiversity in Ecosystem Stability, 356
Forest Management and Diversity, 357
Effects of Traditional Forest Management on Diversity, 357
Preserving Diversity in Managed Forests, 358
Ecological Forestry: Incorporating Biodiversity
into Forest Management, 359
Variable-Retention Harvest System, 360
Designing a Variable-Retention Harvest System, 361
How Well Does Variable Retention Conserve
Biodiversity?, 361
Epilog: Conserving Ecosystem and Biological Diversity, 363
PA R T 5 Forest Ecosystem Dynamics
15 Long-Term Forest Ecosystem and Vegetation Change 367

Change Before the Pleistocene Age, 367
Pleistocene Glaciations, 368
Ecosystem and Vegetational Change Since the Last Glacial Maximum, 370
0005430834.INDD 15 11-23-2022 12:32:40

xvi C o n t e n ts
Eastern North America, 370

Overall Migration Sequence and Patterns, 371
Ecosystem Change in the Southern Appalachians, 373
Western North America, 373
Patterns of Tree Genera and Species Migrations, 375
Migration Irregularities and Disturbance, 377
Migration From Glacial Microrefugia, 377
Independent Migration and Similarity of Communities Through Time, 379
16 Disturbance 381
Concepts of Disturbance, 382
Defining a Disturbance, 382
Disturbance as an Ecosystem Process, 384
Sources of Disturbance, 386
Major Disturbances in Forest Ecosystems, 387
Fire, 387
Roles of Fire in Forest Ecosystems, 387
Pines in New England and the Lake States, 389
Western Pines and Trembling Aspen, 390
Southern Pines, 393
Douglas-Fir in the Pacific Northwest, 394
Giant Sequoia, 394
Fire History and Behavior, 395
Northern Lake States, 395
Boreal Forest and Taiga, 397
Northern Rocky Mountains, 398
Fire Suppression and Exclusion, 399
Wind, 400
Widespread and Local Effects, 401
Principles of Wind Damage, 401
Broadscale Disturbance by Hurricanes, 403
Gulf and Southern Atlantic Coasts, 403
New England—1938 Hurricane, 404
Wave-Regenerated Fir Species, 404
Floodwater and Ice Storms, 404
Insects and Disease, 406
Catastrophic and Local Land Movements, 407
Logging, 407
Land Clearing, 409
Disturbance Interactions, 409
0005430834.INDD 16 11-23-2022 12:32:41

Contents xvii
Biotic Composition Changes, 411

Elimination of Species, 411
Addition of Species, 412
17 Forest Succession 413

Basic Concepts of Succession, 414
Primary and Secondary Succession, 414
Biological Legacies, 414
Successional Pathways, Mechanisms, and Models, 414
Autogenic and Allogenic Succession, 416
How is Succession Determined?, 416
Evolution of the Concept of Forest Succession, 417
Formal Ecological Theory, 417
How Does Succession Work?, 418
Clementsian Succession, 419
Stages of Succession, 420
Primary Succession, 420
Secondary Succession, 423
Successional Causes, Mechanisms, and Models, 427
Key Characteristics and Regeneration Strategies, 427
Availability and Arrival Sequence of Species, 428
Facilitation, Tolerance, and Inhibition, 428
Change in Ecosystems, 431
End Point of Succession?, 431
Succession as an Ecosystem Process, 432
Examples of Forest Succession, 434
Primary Succession on Recently Deglaciated Terrain, 434
Succession Following the Eruption of Mount St. Helens, 437
Secondary Succession Following Fire in Ponderosa Pine Forests
of Western Montana, 439
Secondary Succession and Gap Dynamics, 439
Gap Specialists: American Beech and Sugar Maple, 444
Fire and Oak Dominance—Oaks at Risk, 446
18 Carbon Balance of Trees and Ecosystems 449

Carbon Balance of Trees, 450
Photosynthesis, Dark Respiration, and Leaf C Gain, 450
Light and Leaf C Gain, 452
0005430834.INDD 17 11-23-2022 12:32:41

xviii C o n t e n ts
Temperature and Leaf C Gain, 454

Water and Leaf C Gain, 455
Soil Nitrogen Availability and Leaf C Gain, 456
Construction and Maintenance Respiration, 456
Allocation to Structure, Storage, and Defense, 459
Light and C Allocation, 461
Soil Nitrogen Availability and C Allocation, 461
Carbon Balance of Ecosystems, 465
Biomass and Productivity of Forest Ecosystems, 466
Measurement of Biomass and Productivity, 469
Climate and Productivity, 473
Soil Properties, Forest Biomass, and ANPP, 475
Biomass Accumulation During Ecosystem Development, 477
Soil N Availability and Belowground Net Primary Productivity, 482
19 Nutrient Cycling 487

Nutrient Additions to Forest Ecosystems, 488
Mineral Weathering, 488
Atmospheric Deposition, 490
Biological Fixation of Nitrogen, 493
Nutrient Cycling within Forest Ecosystems, 497
Nutrient Transport to Roots, 497
Nutrient Uptake and Assimilation by Roots, 498
Root Architecture, Mycorrhizae, and Nutrient Acquisition, 501
Root Architecture, 501
Mycorrhizae, 502
Plant Litter and the Return of Nutrients to Forest Floor and Soil, 504
Leaf and Root Litter Production, 505
Nutrient Retranslocation, 507
Nutrients in the Forest Floor, 509
Organic Matter Decomposition and Nutrient Mineralization, 512
Biochemical Constituents of Plant Litter, 513
Dynamics of Decomposition, 515
Nitrogen Immobilization and Mineralization, 519
Nitrogen Availability in Forest Ecosystems, 521
Nitrification, 522
Nutrient Loss From Forest Ecosystems, 523
Nutrient Leaching from Forest Ecosystems, 524
0005430834.INDD 18 11-23-2022 12:32:41

Contents xix
Denitrification, 525
The Cycling and Storage of Nutrients in Forest Ecosystems, 527
Nutrient Storage in Boreal, Temperate, and Tropical Forests, 527
The Nitrogen and Calcium Cycle of a Temperate Forest Ecosystem, 528
Ecosystem C Balance and the Retention and Loss of Nutrients, 530
Forest Harvesting and Nutrient Loss, 532
PA R T 6 Forests of the Future
20 Climate Change and Forest Ecosystems 537

Climate Change Concepts, 537
Effects on Temperature, 540
Effects on Precipitation, 543
Climate Change Effects on the Forest Tree, 546
Tree Growth and Mortality, 546
Phenology, 548
Regeneration, 551
Climate Change Effects on Tree Species Distributions, 554
Observed Range Shifts, 554
Projected Changes in Tree Species Distributions, 556
Projected Changes in Forest Type Distributions, 559
Climate Change Effects on Forest Disturbances, 561
Fires, 563
Insects and Pathogens, 568
Wind, 570
Climate Change Effects on Forest Carbon, 571
Climate Change Effects on Carbon Gain: Primary Productivity, 572
Climate Change Effects on Carbon Loss: Heterotrophic Respiration, 573
Feedbacks Among Disturbance, Climate Change, and Carbon
in Forests, 575
Fire, Carbon, and Climate Change in Forests of Yellowstone
National Park, 577
Adapting to Climate Change Effects on Forests, 578
Assisted Migration, 579
Refugia, 581
Forest Carbon Management, 583
0005430834.INDD 19 11-23-2022 12:32:41

xx C o n t e n ts
21 Invasive Species in Forest Ecosystems 587

Concepts of Invasive Species, 587
Definition of Invasive Species, 587
Characteristic Traits of Invasive Plant Species, 589
Non-Plant Invasive Species in Forests, 594
Impacts of Invasive Species on Forests, 597
Impacts of Invasive Plants on Forests, 597
Competition, 598
Altered Fire Regimes, 598
Carbon and Nutrient Cycling, 599
Impacts of Invasive Insects and Pathogens on Forests, 600
Chestnut Blight, Dutch Elm Disease, and Forest Succession, 601
Impacts of Invasive Animals on Forests, 604
A Primer of Invasive Species Management in Forests, 605
Early Intervention Strategies, 607
Management Approaches for Established Invasive Species, 607
Novel Ecosystems and Invasive Species, 608
22 Forest Landscape Ecology 611

Concepts of Landscape Ecology, 612
Forest Fragmentation and Connectivity, 615
Patches in Forest Ecology, 616
Forest Fragmentation, 617
Ecological Effects, 617
Connectivity, 620
Disturbances on Landscapes, 622
Effects of Heterogeneity on Disturbances, 622
Hurricanes in New England, 622
Landscape Pattern Effects on Disturbance Spread, 625
Effects of Disturbances on Heterogeneity, 625
Stand-Replacing Wildfires in Yellowstone National Park, 627
Historical Range of Variability, 630
Interactions of Landscape Patterns and Ecological Processes, 632
Leaf Area and Productivity, 633
Forest Carbon Dynamics, 635
Nutrient Dynamics, 636
0005430834.INDD 20 11-23-2022 12:32:41

Contents xxi
23 Sustainability of Forest Ecosystems 639

Concepts of Sustainability, 639
The Prevalence of Human Values in Forest Ecology, 639
Historical Perspective of Sustainability in Forests, 641
Ecosystem Services, 642
Toward a Definition of Sustainability, 644
Where Do We Go From Here?, 646
Epilog: Earth as a Metaphor for Life, 648
References651
Scientific Names of Trees and Shrubs 731
Index739
0005430834.INDD 21 11-23-2022 12:32:41

Preface
F orest Ecology deals with forest ecosystems—spatial and volumetric segments of the Earth—and
their climate, landforms, soils, and biota. It is designed as a textbook for people interested in
forest ecosystems—either in the context of courses in forest ecology and environmental science or
as an ecological reference for those in professional practice. This book is meant to provide basic
ecological concepts and principles for field ecologists, foresters, naturalists, botanists, and others
interested in the conservation and restoration of forest ecosystems.
Ecology, in general, has undergone several sea changes since the appearance of the first
edition in 1964, with enormous increases in public interest and scientific development of theory
and research. Ecology and the issues associated with it have become part of our modern lexicon.
The great number of advances in our ecological knowledge, as well as increased public interest,
presents forest ecologists with both opportunities and challenges to sustainably manage ecosystems
using our best understanding of ecosystem properties and processes. This book will hopefully be
useful in that process by providing an understanding of the ecological relationships of individual
trees and forest ecosystems.
The book has six major subdivisions. “Forest Ecology and Landscape Ecosystems” introduces
forests as whole ecosystems rather than tree communities, and at multiple scales. “The Forest
Tree” considers the genetic variations among individual trees, the causes of diversity within and
between species, regeneration ecology, and selected aspects of tree structure and function. “The
Physical Environment” treats the physical factors of forest ecosystems that form the forest site—the
influences of light, temperature, physiography, soil, and fire on the individual forest plant and
on plant communities. The concluding chapter in this part considers methods of evaluation and
classification of the forest site and ecosystems. In Part 4, “Forest Communities,” we consider the
forest community of trees and associated plants and animals that form a key structural component
of forest ecosystems—one part of the whole. We also consider the importance and measurement of
diversity of species and ecosystems. In “Forest Ecosystem Dynamics,” we examine the functional
relationships of the physical environment and the biota. We first examine changes in communities
and ecosystems over tens of thousands of years. We then consider the extent to which disturbance,
an ecosystem process, initiates change (termed succession) over shorter time scales of centuries.
Chapters on carbon balance and nutrient cycling present a detailed consideration of the pattern
in which carbon (i.e., energy) and plant nutrients flow within forest ecosystems and how natural
and human-induced disturbances alter these patterns. Finally, “Forests of the Future” explores the
role of humans in the sustainability of forests. Here we emphasize two of the most pressing issues
in forest ecology today, climate change and invasive species, and present a review of landscape
ecology which has humans at its center. We end the book with a treatment of sustainability itself.
In this edition, we have made great attempts to maintain the core organization and readability
of previous editions while adding those areas most relevant to forest ecology that have developed
over the last quarter-century. We have retained the important focus on landscape ecosystems (rather
than organisms and communities) that was developed in the fourth edition and have added critical
new ecological concepts and research that have developed in genetics, diversity, climate change,
invasive species, and sustainability. The ecological literature has only become more voluminous
over the past 25 years, and as in previous editions our use of the literature was selective, rather
than exhaustive. New references were most often chosen based on their accessibility to students
xxiii
xxiv P r e fac e
and practitioners as understandable examples of important ecological concepts. At the same time,
we have retained many older references that still provide excellent examples of fundamental eco-
logical concepts, many of which would otherwise be lost in obscurity.
As before, we have integrated woody plant physiology into multiple chapters, rather than
developing it in a single chapter, and in this edition have done the same with climate because of its
overriding influence on so many ecosystem processes. We have also further limited our treatment
of forest ecology with examples from temperate and boreal forests, with special emphasis on North
America and Europe. With a primary focus on the ecological principles of forests that form the
basis for management, we have largely avoided specific treatments of forest management tech-
niques and strategies throughout the book, although by necessity we have provided some examples
of forest management in the chapters on diversity and invasive species.
The revised edition would not have come to pass without the patience and dedication of
many colleagues who helped in its preparation. We are especially indebted to the following for
their reviews of one or more chapters: Dennis Albert, Brian Buma, Mark Dixon, Tom Dowling,
Jennifer Fraterrigo, Stephen Handler, Donna Kashian, Doug Pearsall, David Rothstein, Madelyn
Tucker, and Chris Webster. Other important contributions, either through provision of material,
enlightening discussions, or enthusiastic encouragement, were made by Jonathon Adams, Virginia
Laetz, and Dan Binkley. Victoria Meller was extremely supportive of D. M. Kashian and his time
spent away from campus in the completion of this revision, as were the graduate students in the
Kashian lab. Perhaps without knowing it, the graduate and undergraduate students in the 2018
and 2020 Terrestrial Ecology classes at Wayne State had an immense role in the thought processes
and revisions made in the 5th edition. D. R. Zak was partially supported by the National Science
Foundation and the Department of Energy during the preparation of this text.
Daniel M. Kashian
Detroit, Michigan, USA
Donald R. Zak
Ann Arbor, Michigan, USA
Forest Ecology and PA R T 1
Landscape Ecosystems
F orest ecologists work to understand the dynamics, structure, and function of

forest ecosystems. The first step in doing so is to approach ecosystems as geo-
graphic units or landscape ecosystems—real, three-dimensional, defined locations at the
Earth’s surface. Landscape ecosystems include organisms, species, and communities, but
also the air above and the soil below these living things such that organisms are but
one important part of a larger landscape unit. Our perception of forest ecosystems in
this book is that ecosystems are not simply extensions of forest communities, whereby
ecosystems are conceived as organisms and their environment. Instead, units of whole
ecosystems that integrate factors of climate, landforms, soils, co-occurring biota, and all
the interactions between them are the most appropriate units of study rather than their
individual parts that too often claim our immediate attention.
In understanding forest ecosystems, particularly in the context of an ecology course
or in solving ecological problems, ecosystems need to be conceived at several spatial and
temporal scales. It is very daunting to attempt to gain such understanding, and certain
well-known ecologists have suggested that the environmental complex is unknowable
and inexpressible. However, conceiving landscapes as ecosystems and proceeding “from
above,” that is, from the biggest to the smallest units to understand their spatial relation-
ships, makes synthesis possible and manageable. Although local sites and stands with
their familiar species appear a convenient starting point, we know that these are, in turn,
affected by geological and climatic factors of higher levels within which they are embed-
ded. Landscape ecosystems are nested within one another in a hierarchy of spatial sizes
such that it is best to consider the big picture first—the comprehensive view from the
outside—rather than the minutia of details as seen from inside of the ecosystem. Our
perspective in understanding ecosystems involves establishing a framework for studying
the components—a framework of the spatial and temporal, hierarchical pattern of eco-
systems of all sizes making up the ecosphere.
Therefore, in the first part of this book, before immersing in the detail of organ-
isms, sites, and their interrelationships, we wish to first examine landscape ecosystems,
and then place them within a perspective of spatial levels and their processes at different
Forest Ecology, Fifth Edition. Daniel M. Kashian, Donald R. Zak, Burton V. Barnes, and Stephen H. Spurr.
© 2023 John Wiley & Sons Ltd. Published 2023 by John Wiley & Sons Ltd.
1
2 PA RT 1 Forest Ecology and Landscape Ecosystems
temporal scales. In Chapter 1, we present the basic concepts of forest ecology within a
framework of landscape ecosystems. In Chapter 2, we present the concept of scale and
consider the hierarchy of these landscape ecosystems. These fundamental concepts pro-
vide the basis for studying the variation, life history, and ecology of individual organisms
that follow in Part 2.
Concepts of Forest Ecology CHAPTER 1
A forest is an ecological system dominated by trees and other woody vegetation. More than
simply a stand of trees or a community of woody and herbaceous plants, a forest is a complex
ecological system, or ecosystem, characterized by a layered structure of functional parts. Ecology
is the study of ecological systems and their interacting abiotic and biotic components. Forest
ecology, therefore, addresses the structure, composition, and function of forests. In forest ecology,
we study forest organisms and their responses to physical factors of the environment across for-
ested landscapes. Forests are widespread on land surfaces in humid climates outside of the polar
regions. It is with forests in general, and with the temperate North American forest in particular,
that this book is concerned.
There are many ways to study forest ecosystems. Most simply, a forest may be considered in
terms of the trees that give the forest its characteristic aboveground appearance or physiognomy.
Thus, we think of a beech–sugar maple forest, a ponderosa pine forest, or of other forest types, for
which the naming of the predominant trees alone serves to characterize the forest ecosystem.
Forest types are often considered to be composed of forest stands, which are trees in a local
setting possessing sufficient uniformity of species composition, age, spatial arrangement, or
condition to be distinguishable from adjacent stands (Ford-Robertson 1983).
A broader concept of a forest may take into account the interrelationships that exist between
forest trees and other organisms. Certain herbs and shrubs are commonly found in beech–sugar
maple forests, and these may differ from those found in ponderosa pine or loblolly pine forests.
Similar interrelationships may be demonstrated, for example, for birds, mammals, arthropods,
mosses, fungi, and bacteria. Thus, part of the forest ecosystem is the assemblage of plants and ani-
mals living together in a biotic community. The forest community, then, is an aggregation of
plants and animals living together and occupying a common area. It is thus a more organismally
complex unit than the forest type.
A third approach is to focus on geographic or landscape ecosystems. This approach is
centered conceptually and in practice on whole ecosystems and not just their parts. When our pri-
mary focus is real live chunks of Earth space, that is, landscapes and waterscapes (oceans, lakes,
rivers; hereafter included as parts of a landscape), we can effectively study their parts (e.g., organ-
isms, soils, and landforms) while recognizing that each is but one part of a functioning whole. We
emphasize this focus on ecosystems rather than on the individual organisms and species that are
parts of them.
In the past, the forest stand or the species has been the focus in natural resource fields such
as forestry and wildlife. However, we are really managing whole forest ecosystems, despite their
incredible complexity, because the diverse biota is inseparable from the physical environment that
supports it. A consideration of the field of ecology from this viewpoint provides an overall
perspective.
3
4 Chapter 1 Concepts of Forest Ecology
ECOLOGY
Broader fields of scientific inquiry are difficult to limit and define, and ecology is one of the most
indistinct. In 1866, Ernst Haeckel proposed the term oecology, from the Greek oikos meaning
home or place to live, as the fourth field of biology dealing with environmental relationships of
organisms. Thus, ecology literally means “the knowledge of home,” or “home wisdom.” Since its
introduction, the term has been applied at one time or another to almost every aspect of scientific
investigation involving the relationship of organisms to one another or to the environment
(Rowe 1989). Haeckel’s organismal focus of ecology has since been redefined and expanded to
include the physical aspects of the environment that provide life for those organisms (Hagen 1992;
Golley 1993). Thus, Rowe (1989, p. 230) suggests:
Ecology is, or should be, the study of ecological systems that are home to organisms at the surface of the
earth. From this larger-than-life perspective, ecology’s concerns are with volumes of earth space, each con-
sisting of an atmospheric layer lying on an earth/water layer with organisms sandwiched at the
solar-energized interfaces. These three-dimensional air/organisms/earth systems are real ecosystems—the
true subjects of ecology.
This approach to ecology emphasizes whole ecosystems as well as organisms, both volumet-
ric and having structure and function.
LANDSCAPE ECOSYSTEMS
The British botanist–ecologist Arthur Tansley (1935) introduced the term ecosystem, writing with
an emphasis on “the whole ‘system,’ including not only the organism complex but also the whole
complex of physical factors.” He also noted that from the point of view of the ecologist, ecosystems
“are the basic units of nature on the face of the Earth.” Tansley was a biologist and vegetation ecol-
ogist, and so his idea of ecosystem was centered on organisms (species or communities) rather
than geographic or landscape entities. With this bioecosystem approach, “ecosystem” derives its
meaning from particular plant or animal organisms of interest, and an “abiotic” environment
defined by the organisms as relevant or not is considered with lesser emphasis. In this approach,
every organism defines its own ecosystem, nearly infinite in number and difficult to study and use
as a basis for management and conservation.
On the other hand, others (e.g., Rowe 1961a and Troll 1968, 1971) view ecosystems centered
on geographic or landscape units (i.e., geoecosystems) of which organisms are but one important
structural component (Rowe and Barnes 1994). We term these units landscape ecosystems in
part to differentiate them from geology-based units of study (e.g., Huggett 1995). Landscape eco-
systems are geographic objects, with a defined place on the Earth. Landscape ecosystems have
three dimensions (volume) just as organisms do, including landforms and biota at the Earth’s sur-
face as well as the air above them and the soils below them (Figure 1.1). Other terms have been
introduced to express the same idea, but are less commonly used, such as the ecotope
(Troll 1963a, 1968) and the ecoterresa (Jenny 1980). This geographic/volumetric concept has been
discussed and adapted by professional and academic ecological societies (Christensen et al. 1996),
and is useful to field ecologists, naturalists, foresters and other land managers, and natural resource
professionals. The concept is described in detail in Chapters 2 and 11.
In addition to being geographic and volumetric, landscape ecosystems are hierarchical,
extending downward from the largest ecosystem we know, the ecosphere (Cole 1958), through
multiple levels of ecological organization (Figure 1.2). These levels include macrolevel units of
continents and seas, each of which contains mesolevel units of regional ecosystems (major phys-
iographic units and their included organisms), which in turn contain local ecosystems
(Hills 1952), the smallest level of homogeneous environment with organisms enveloped in it. We
therefore conceive the ecosphere and its landscapes as ecosystems, large and small, nested
Ecosystem boundaries
Macroclimate
Topoclimate
Biota
Landform
Surface water
Soils
Ground water
Bedrock
F I G U R E 1 . 1 The three-dimensional, volumetric nature of a landscape ecosystem. Ecosystems comprise the

atmosphere (macroclimate as well as the climate affected by surface relief), landforms and soils (underlain
by ground water and bedrock), and the biota that provide a physical connection between the air and the
Earth. Source: Bailey (2009) / Springer Nature.
Ecology
Organismal Biology
Cell
Organelle Cell Biology
olecu
Molecular Biology and
M
le
Biochemistry
F I G U R E 1 . 2 Objects of study from the most inclusive (ecosphere) to the least inclusive levels of organiza-
tion (cell and organelle and molecule below it). Note that each higher level envelops the lower ones as parts
of its whole. Some corresponding fields of study are also shown. Aggregates of organisms, such as popula-
tions and communities, are components of ecosystems at all scales. Like other components such as
atmosphere, landform, and soil, they do not appear in this diagram of first-order objects of study, but are
shown in Figures 1.3 and 1.4. Source: Modified from Rowe (1961a).
within one another in an ecological hierarchy, having processes at each level with their own
spatial and temporal scales (see Chapter 2).
Two landscape ecosystems in Figure 1.3 (Rowe 1984b) illustrate characteristic ecosystem
differences in hilly or mountainous terrain. The two ecosystems are distinguished by different geo-
morphologies (convex upper slope versus concave lower slope, and high versus low topographic
slope position) that mediate microclimate, soil water, and nutrient availability. The vertical dashed
line is placed at an ecologically significant boundary that spatially separates the two ecosystems.
Organisms in these ecosystems are sandwiched between the air and the Earth. Also shown in Figure 1.3
are the traditional fields of study in which individuals seek to understand each of the ecosystem com-
ponents, although forest ecologists aim to understand the integrated effects of all of these components.
In many parts of this book, we focus on organisms, species, and communities, but always
remembering that they are parts of volumetric, hierarchical ecosystems. For studies of organisms
in their immediate surroundings, a biological approach is often useful. Nevertheless, for
management of ecosystems, studies of forest productivity, and the conservation and restoration of
forest ecosystems, a landscape ecosystem approach is eminently practical and theoretically sound.
Forest ecologists not only study (i) organisms of these systems and their aggregates as communities
HILLSIDE SLOPE
Upper Lower
Fields of Study
Upper Air Meteorology

Layer Climatology
Lower Air Topoclimatology

Layer
Biota Ecology
Biology
Soil Pedology
Landform Geomorphology
& Surficial Geology
Mineral Strata Bedrock Geology
Water Table Hydrology
F I G U R E 1 . 3 Structural profile illustrating landscape ecosystems of upper and lower slopes. Air–earth
layers surround the organisms at the Earth’s energized surface. The vertical line is set at an ecologically
significant topographic break, dividing the upper and lower slope ecosystems. Source: Rowe (1984b) / United
States Department of Agriculture / Public Domain.
Landscape Ecosystems 7
and populations (see Chapters 3–5, 12, and 13), but also (ii) the functioning of local ecosystems
that involves complex interactions among organisms and their supporting environment
(Chapters 6–11, 13–14, and 16–19), and (iii) the spatial patterns of occurrence and interrelationships
of entire forest ecosystems (Chapters 2, 18, 19, and 22).
LANDSCAPE ECOSYSTEM AND COMMUNITY

The term ecosystem was introduced in 1935 by Tansley, but terms in various languages, such as taiga,
heath, bald, Auenwald (river floodplain forest), pampas, prairie, chaparral, maquis, hammock, mus-
keg, and bog, have long been used to depict interactions among air, organisms, and soil. Very often the
terms emphasize a distinctive plant community and may therefore imply that an ecosystem is simply
an extension of a community. Tansley’s definition of ecosystem as “organism + environment” leads to
this view, and general definitions such as “ecosystem = biotic community + environment” reinforce
this view. Although indispensable for forest ecologists and managers, vegetation may not always coin-
cide with climate–landform–soil-based ecosystems because of disturbance and/or unknown historical
factors. Thus, we emphasize the importance of geography and physiography within a regional climatic
setting as the basis of understanding not only vegetation but whole ecosystem structure and function.
This conceptual approach is not to say that communities are unimportant; they form the key
ecosystem component whose response is essential in affecting ecosystem change and indicating
the integrated effects of many site factors (Chapter 11). Communities and populations, however,
fundamentally differ from entities such as ecosystems, organisms, and cells because they are aggre-
gates of individuals but not functional systems. Entities such as ecosystem, organism, cell, and mol-
ecule (Figure 1.2) are “volumetric” levels of organization because they have structurally joined
parts that form a functioning unit (Rowe 1961a, 1992c). By contrast, communities and populations
are assemblages or aggregates of spatially separated trees and understory plants that have no
necessary, physical, structural connections.
ECOSYSTEM STRUCTURE AND FUNCTION

An ecosystem’s structure describes the spatial arrangement of its parts. In local forest ecosystems,
multiple strata of atmosphere and soil (Figure 1.3), as well as surface relief, influence species com-
position and its patterns of occurrence. Plants provide a physical connection between the soil strata
and the aboveground layers of air. The layered structure and related physical and chemical prop-
erties of soil are well studied and understood, but similar properties of the atmosphere are not
(Rowe 1961a; Woodward 1987), particularly as they affect forest organisms at the Earth’s surface.
Vegetation itself is also structured vertically, and the vertical structure and species composition of
communities as well as species interactions provide insights critical in understanding properties
and processes of ecosystems. In addition to species composition, vegetation structure is shaped by
tree form and stem density, canopy characteristics, shrub and herbaceous plant abundance, and the
amount and distribution of dead vegetation, among other characteristics. Thus, the physiognomy of
vegetation varies markedly in different regional and local ecosystem types (Chapters 5 and 6).
The horizontal spatial patterning of these structural components occurs at multiple spatial
scales, but is most often described across broad scales. Landscapes are structured horizontally into
mosaics of ecosystems that reflect differences in climate, geology, and physiography and their
relative effects on vegetation. The natural communities of these systems reflect limiting factors of
climate, soil water, nutrients, and disturbances, and in turn modify the physical factors. Different
ecosystem mosaics characterize mountains, plains, and river valleys due to fundamental differ-
ences in their physical factors and the vegetation adapted thereto (Chapter 8). The diversity of
landscape ecosystems can be assessed by understanding such mosaics (Lapin and Barnes 1995).
Such an understanding provides a spatial ecosystem framework for programs in biodiversity that
seek to conserve and manage the diversity of organisms and maintain and increase populations of
rare and endangered species (Chapters 14 and 22).
Local landscape ecosystems, besides having a structure of interconnected parts, are
functional units characterized by many processes that define their properties. Ecosystem-level
processes are part of the entire system and are not restricted only to physical or biotic parts. These
processes drive or mediate the flow of energy or matter and/or the cycling of materials in the sys-
tem. Organic matter decomposition; cycling of water, nutrients, and carbon; and biomass
accumulation are considered ecosystem-level processes. Other processes are often more associated
with plant species, populations, or communities, such as photosynthesis and respiration,
reproduction, regeneration, mortality, and succession. Despite their basis in organisms, these are
also ecosystem processes because they are mediated and regulated by characteristic ecosystem
factors of temperature, water, nutrients, and disturbances (such as fire, windstorm, and flooding).
Ecosystem function is often described with box-and-arrow diagrams, flow charts, and simulation
modeling as ways of disentangling very complex systems (Figure 1.4).
Landscape ecosystem structure and function are tightly coupled by the physical environ-
ment, the frequency and severity of disturbances that reset succession, and the life histories of the
plants and animals that comprise the biotic community. There is increasing evidence that many
aspects of an ecosystem’s function are linked to the diversity of its biota (Chapter 14). In turn, an
ABIOTIC ENVIROMMENT
SUBSTANCES OTHER RADIATION SPACE STRUCTURE
HUMANS
water, pressure, heat, media (soil, water,
as a super-
O2, CO2, fire, light air) and their motion,
organic
minerals pH, etc. area, substratum
factor
PRIMARY PRODUCERS
(other green plants
autotrophs*)
OTHER
ECO-
symbionts pollination,
SYSTEMS
distribution,
selection
by grazing
DECOMPOSERS DEAD
mineralizers organic
CONSUMERS OUTPUTS
substances
by respiration,
H scavengers herbivores H erosion, oozing
parasites away of water
bacterio- are omitted
predators H
phages soil
humus
‘limit’ of the ecosystem
dead living material or energy flow other influences
Humans as a partner *relations omitted

H
of the ecosystem
F I G U R E 1 . 4 A model of a landscape ecosystem detailing the flow of energy and matter among biota. The
model also describes the interactions between the physical environment and the biota. Source: Reprinted
from Ellenberg (1988) / Cambridge University Press.
Examples of Landscape Ecosystems 9
ecosystem’s biodiversity is strongly shaped by its physical factors such as climate, physiography,
and soil, which provide the context within which organisms survive, adapt, and evolve. In addition,
the functional aspects of an ecosystem are strongly affected by its geographical context and by its
spatial position relative to its surrounding neighbors on a landscape. Adjacent ecosystems, espe-
cially in mountainous or hilly terrain, affect one another by the lateral exchanges of materials and
energy. Water and snow, soil, organic matter, nutrients, and seeds are transported downhill; the
effects on other ecosystems depend on the size, shape, and composition of the systems. Therefore,
an understanding of the spatial pattern and configuration of landscape ecosystems can play an
important role in the management of ecosystems and their biota. An excellent overview of function
in terrestrial ecosystems is given by Chapin et al. (2012), its variation across landscapes considered
in Lovett et al. (2005), and the history of the ecosystem concept itself is reviewed by Golley (1993).
EXAMPLES OF LANDSCAPE ECOSYSTEMS

There are many examples of landscape ecosystems, some of which are easily discerned in the field,
whereas others must be carved out of geographic continua. For example, bogs in the glaciated ter-
rain of northern and boreal regions are easily discernible ecosystem types. They exhibit distinctive
physiography, soil, and vegetation, and they recur in the landscape. The distinct terraces of river
floodplains distinguish local ecosystem types that differ in microclimate, soil, drainage, vegetation,
and their dynamics. Mountains are characterized by gradients of elevation, aspect, and slope steep-
ness along which strikingly different ecosystems can be purposefully delimited and mapped,
although their boundaries may not be sharply demarcated. On old lake-bed terrain, dry, sandy
beach ridges are easily distinguished from adjacent depressions of swampy land. Differences in
water drainage and oxygen availability in these adjacent systems result in major differences in
their native tree-and ground-flora vegetation. Unseen but critical are the very different processes
that also distinguish these ecosystems.
Ecosystem components will change over time, giving rise to a sequence of different landscape
or waterscape ecosystems on a given place on the Earth’s surface. Striking changes in the fossil
record illustrate long-term changes in physiography, soil, and biota (Chapter 15), but short-term
changes also occur. Shorter-term changes are most obvious when natural or human-caused distur-
bances affect vegetation. A site-specific location may exhibit a range of different forest commu-
nities over 100 to 300 years, young and old, as disturbances or lack thereof affect the biota. Ecosys-
tem structure and function at this site change over time. The landforms and associated pattern of
parent material, soil, and climate also change over time, but more slowly than the suite of species
available to recolonize the disturbed site. Thus, what we term the ecosystem site type or simply
ecosystem type (land area supporting potentially equivalent ecosystems) can be distinguished
and mapped regardless of the forest community currently present.
One such example is illustrated in Figure 1.5 where deciduous forests occur on glaciated ter-
rain in southern Michigan. In this setting, three local ecosystem types are distinguished by differ-
ences in physiography (outwash and moraine landforms); soil; drainage; and overstory, under-
story, and ground-cover vegetation. The relatively fine-textured, silty soil on the outwash plain
(type 1) supports forest dominated by white oak, whereas the drier, coarse-textured, sandy out-
wash soil (type 2) supports a black oak community. The fine-textured, moist, clayey soil on
the rolling moraine landform (type 3) supports a community dominated by northern red oak. The
moraine formerly supported a beech–sugar maple forest. However, recurrent fires through the drier
white and black oak ecosystems killed the fire-sensitive mesic species of the adjacent moraine
ecosystem and led to dominance of northern red oak. An occasional beech is still found, and red
maple has invaded the shaded understory.
The western portion of ecosystem type 1 was clear-cut about 90 years ago, and a cover type
markedly different from the adjacent old-growth white oak forest has formed (Figure 1.5, type
(a) Ecosystem Type

1 2 3
American T'
T 1a-cut 1b-uncut Black Oak Northern Red Oak beech
White Oak
Fine Outwash
– Silt
and
sh – S
Outwa
Coarse
End Moraine
(Till)
(b) Forest Cover Type

1a 1b 2 3
cut uncut Black Oak Northern Red Oak
T T'
1 2 3
Ecosystem Type
Cutting Type Transect

Boundary Boundary T-T'
F I G U R E 1 . 5 Three local landscape ecosystem types in glacial terrain that are differentiated by landform,
parent material, soil, and vegetation: (a) Lateral transect from T to T′ showing vegetation and underlying
geological parent material. (b) Top view showing distribution of forest trees, cover types, and ecosystem
types. Following clear-cutting of part of type 1, two forest cover types (1a—early successional oak forest;
1b—old-growth white-oak forest) are distinguished. The diagram illustrates that different forest cover types
(1a and 1b) are not necessarily different ecosystem types. They represent two of many possible ecosystem
derivatives (disturbed in 1a versus relatively undisturbed in 1b) of a given ecosystem type. See text for
discussion.
An Approach to the Study of Forest Ecology 11
la). The overstory tree layer of the disturbed area is now dominated by white, black, and red oaks,
white ash, black maple, American elm, black cherry, and sassafras. These species either sprouted
from the base of cut trees, were already present in the white oak forest understory, or seeded in
from adjacent communities following cutting. Today, we can recognize two (types 1a and 1b) of
the many compositionally different forest communities that might occur at the site of ecosystem
type 1; these easily could be mapped as two different forest cover types. Because the cut-over area
(type 1a) has physiography and soil like type 1b, its vegetation may gradually become similar in
structure to those of ecosystem type 1b, providing that no major changes in climate, soil, or
ecosystem processes (including changed browsing pressure by herbivores) were caused by clear-
cutting. Thus, a given local ecosystem type, defined by relatively stable features of physiography
and soil, may have a suite of disturbance-induced cover types (Simpson et al. 1990). Therefore,
recognizing forest cover types alone is not necessarily likely to provide a useful estimate
of site potential for management or conservation. However, cover types are extremely use-
ful in management planning for wildlife, timber, water, and recreational use of existing forest
communities.
It is often the conspicuous forest cover type that receives our immediate attention. However,
the enormous complexity of geographic space and changes in its component ecosystems through
time require that major attention be directed to atmospheric, geologic, physiographic, and soil
properties of forest landscapes. In summary, for every landscape, a combination of factors should
be used to distinguish the pattern of local and regional landscape and waterscape ecosystems that
have similar ecological potential in the long run. Understanding ecological units at multiple spatial
scales (Chapters 2, 11, and 22) is needed to provide the basis for monitoring ecosystem change over
time and for ecosystem management.
AN APPROACH TO THE STUDY OF FOREST ECOLOGY

This book presents the scope of forest ecology as having at least two major parts. The first is
analysis: studying the details of ecosystem components (atmosphere, organisms, and Earth). The
second part is synthesis: understanding the broad ecological framework within which ecosystem
components are integrated. We therefore emphasize a “big picture” approach even as we discuss
much smaller details of the organic and inorganic parts of ecosystems, their interactions, and the
structure and function of entire ecosystems. The explicit separation and analysis of the biotic and
physical parts of ecosystems are not realistic or desirable from an ecological perspective, but may
serve as reasonable subdivisions to provide both analysis and synthesis. The sequence of presen-
tation is (i) understanding forest ecosystems at multiple spatial scales, (ii) the variation, life
history, and structure of the forest tree, (iii) the physical environment of forests, (iv) forest com-
munities, (v) forest ecosystem dynamics, and (vi) the outlook for and consideration of future forests.
In Part 1 of this book, landscape ecosystems are the primary objects of interest, but we
observe the importance of understanding spatial scale in examining them. Spatial scale influ-
ences the observations we make about ecosystems because ecological processes vary as spatial
scale changes. Advances in hierarchy theory in ecology have helped us to understand ecosys-
tems as volumetric segments of the ecosphere within which smaller volumetric units
are nested.
In Part 2, we consider that the forest tree largely owes its appearance, rate of growth, and size
to the environment in which it has grown throughout its life. In other words, the phenotype, the
individual as it appears in the forest, is the product of the environment acting on its genotype, its
individual genetic constitution. An understanding of trees as parts of ecosystems depends, in part,
upon the genetics of the trees themselves and the way their genetic heritage affects their response
to the environment they live in. In addition, life history features of regeneration as well as anatomical
and physiological aspects of forest trees in relation to environment are considered.
We examine the physical factors of forest ecosystems in Part 3. The site is the sum total of
environmental factors surrounding and available to the plant at a specific geographic place. These
factors are primarily the atmospheric, physiographic, and soil components of the physical environ-
ment, but also include important influences of growing vegetation which itself affects the micro-
climate and soil. Climate includes various atmospheric factors that vary across hours, days,
months, and years, such as solar radiation, air temperature, precipitation, humidity, wind, and
carbon dioxide content. We provide special consideration to sunlight (solar radiation) and tem-
perature, and we examine the relations and dynamics of water throughout the text in appropriate
contexts. Physiological processes related to photosynthesis, respiration, and growth are included in
Part 3 in chapters on light and temperature, but also in later chapters on carbon balance and nutri-
ent cycling.
Physiography—comprising landforms and soil parent material——plays a key role in
affecting climate as well as the amount and rates at which radiation and moisture are received and
distributed in forest ecosystems. Below the ground surface, the supply of soil moisture and nutri-
ents, the physical structure of the soil, microbial communities in the soil, and the nature and
decomposition pattern of organic matter, that is, factors pertaining to edaphic characteristics of
soil, all affect the growth and development of plants. We also treat fire as a site factor. The study of
the environmental factors and their effects on individual plants constitutes the field of autecol-
ogy, and we summarize those aspects of forest autecology most pertinent to an understanding of
forest ecosystems. Finally, we examine site quality and its evaluation, focusing on the degree to
which individual site factors and combinations of them are used to estimate the productivity of
forest ecosystems and determine management prescriptions for them.
In Part 4, we consider forest communities and their plants and animals as integral parts of
ecosystems. First, we examine the important roles of animals in affecting all phases of plant
development, as well as their effects on forest communities and ecosystem processes. Forest com-
munities are then treated, emphasizing their composition, occurrence, and the interactions (with
emphasis on competition and mutualism) of their constituent individuals. Finally, we examine
concepts of biological and ecological diversity, including their importance, measurement, and
conservation.
Forest ecosystems are ever changing, and we examine their dynamics in Part 5. We first con-
sider change in ecosystems over hundreds to millions of years. Primary bases for this change come
from geologic and physiographic studies of the land itself and determination of plant species and
their communities as deduced from the paleoecological record. Disturbance and succession as eco-
system processes are treated next, followed by chapters on whole ecosystem functioning, including
the carbon balance of trees and ecosystems and the dynamics of nutrients.
Part 6 examines future forest ecosystems in considering their sustainability. We first examine
the effects of climate change on forest ecosystems, including effects on individuals, populations,
communities, and whole ecosystems. We also explore the potential for mitigation of climate change
using forest management. We next consider the importance of invasive plants and animals for
forest ecosystem structure and function. Finally, we place humans in an appropriate context of
forest ecosystems in considering forest landscape ecology and the emerging concept of forest
sustainability.
APPLICABILITY TO FOREST MANAGEMENT

The management of forests is, as it should be, a continuously evolving effort as new science, tech-
nology, and consideration of the role of humans in forest ecosystems develop. Management of
forests may include decisions made at broad or local scales, but consideration is increasingly given
to whole systems, their complexity and heterogeneity, as well as their long-term sustainability and
integrity (Franklin et al. 2018) rather than single-focus use of their parts for recreation, wildlife, or
Applicability to Forest Management 13
timber. To ecologists, understanding ecosystems for their own sake has significant value in itself,
but the principles of forest ecology are essential for their guidance in conservation and management
practices. Spurr (1945) defined silviculture, or applied forest ecology, as the theory and practice
of controlling forest establishment, composition, and growth. Modern texts suggest that current
silviculture is most concerned with managing forests for future products and services in a manner
that has social stability (Ashton and Kelty 2018). Nevertheless, we explicitly emphasize that silvi-
culture is the theory and practice of controlling forest ecosystem composition, structure, function,
and heterogeneity, often at multiple spatial and temporal scales. Even local management opera-
tions may have significant and often long-term effects on surrounding landscapes far removed
from the specific site of human activity.
The management, conservation, and restoration of forested landscape ecosystems rely
heavily on understanding the structure and function of forest ecosystems and the autecology of
their organisms (i.e., forest ecology). More than ever, understanding how ecological systems
work is critical for making sound decisions regarding human intervention in forests. Under-
standing the structure and function of whole ecosystems will also equip us to best deal with the
ecological consequences of past human activities, such as deforestation and its wide range of
impacts, habitat loss, and fragmentation. Human activities associated with industrialization,
such as air pollution or the introduction of destructive new species, have resulted in local tree
mortality and, in certain ecosystems, widespread death and decline of forest trees and other
organisms. Increasing the carbon dioxide concentration in the atmosphere due to the burning
of fossil fuels and deforestation are creating climatic changes and concomitant changes in
forest ecosystems that are quickly increasing in their severity. Invasive plants and animals
introduced into forests are well known for disrupting ecosystem structure and function on
relatively short time scales. In all cases, one must understand the system in order to understand
the extent to which it is disrupted.
At the time of the last edition of this textbook over 20 years ago, a new paradigm in for-
estry had occurred that shifted the view of the forest from one of single commodities (timber,
wildlife, water, and recreation) to forestry understood as sustaining ecosystems and their
structure, diversity, heterogeneity, and function (Rowe 1994; Kohm and Franklin 1997). That
paradigm emphasized maintaining the integrity of ecosystems across landscapes by sustaining
their natural patterns and processes even when heavily manipulated by humans. This paradigm
has been further refined over the past two decades to include an emphasis on considering or
even incorporating natural disturbances, increasing ecosystem resistance and resilience,
embracing structural complexity, and maintaining a range of ecosystem conditions across
broad scales (Franklin et al. 2018). Whether intentionally or otherwise, this paradigm has the
landscape ecosystem approach at its core because it focuses on the maintenance of landscapes
as complete ecosystems. The only way to assure the sustained yields of forests, wildlife, and
water, now and in the future, is to maintain the integrity of their processes and keep their
biota in a healthy state. Such a practice demands an intimate familiarity with forest ecology as
we have described it.
SUGGESTED
R E A D I N G S
Bailey, R.G. (2009). Ecosystem Geography, 2e. (Chapters 1 Golley, F.B. (1993). A History of the Ecosystem Concept
and 11). New York: Springer-Verlag 251 pp. + 1 map. in Ecology. New Haven, CT: Yale Univ. Press 254 pp.
Chapin, F.S. III, Matson, P.A., Vitousek, P., and Chap- Franklin, J.F., Johnson, K.N., and Johnson, D.L. (2018).
in, M.C. (2012). Principles of Terrestrial Ecosystem Ecological Forest Management. (Chapter 1). Long
Ecology, 2e. New York: Springer 520 pp. Grove, IL: Waveland Press 646 pp.
Kohm, K.A. and Franklin, J.F. (ed.) (1997). Creating a Rowe, J.S. (1994). A new paradigm for forestry. For.
Forestry for the 21st Century. Washington, D.C: Island Chron. 70: 565–568.
Press 475 pp. Tansley, A.G. (1935). The use and abuse of vegetational
Rowe, J.S. (1961). The level-of-integration concept and concepts and terms. Ecology 16: 284–307.
ecology. Ecology 42: 420–427.
Rowe, J.S. (1992). The ecosystem approach to forestland
management. For. Chron. 68: 222–224.
Landscape Ecosystems at
CHAPTER 2
Multiple Scales
F orest ecologists often by default are also landscape ecologists—people whose interests and
practices necessarily encompass ecosystems of many different spatial scales. A landscape
is a heterogeneous land area composed of a group of intermeshed ecosystems, each with inter-
acting atmosphere, earth, and biota. Landscape ecology is the study of the spatial patterns of
ecosystems, the causes of the pattern, and how the pattern affects ecological processes. Many
of the core concepts of landscape ecosystems overlap with those of landscape ecology, but the
domain of interest of the two fields is not identical. In this chapter, we examine the spatial occur-
rence and complexity of landscape ecosystems, whereas the discipline of landscape ecology is
considered in Chapter 22.
OVERVIEW OF SPATIAL AND TEMPORAL SCALES

Forest ecologists ask questions, conduct field experiments, and resolve problems at different scales
of space and time, and these scales directly influence the problems posed, the questions asked, and
the techniques employed. Scale describes the dimensions of an object or process in space or time
(Turner and Gardner 2015). Scale is often confused with biological or ecological levels of organiza-
tion, such as cells, tissues, organs, organisms, and ecosystems (see Figure 1.2, Chapter 1), but the
concepts are quite distinct. For example, organisms are obviously an important level of organiza-
tion, but they may occur over very different spatial scales (e.g., 1–10 μm for bacteria to 84 m tall
giant sequoia) with very different life spans (e.g., 24 hours for mayflies to nearly 5000 years for
bristlecone pine). A given scale is defined by its grain (its finest spatial or temporal resolution) and
its extent (the size of a study area or the length of time considered), and there is some trade-off
between the two characteristics. For example, forest ecologists might increase grain size (e.g.,
sample plot size) in a larger study area (e.g., the size of a county) because of their interest in
capturing variation in species composition at a larger spatial scale. However, they might reduce
grain size (using smaller plots or quadrats) if they are interested in describing species variation
across a 1 ha woodlot. As a result, no single set of spatial and temporal scales is applicable to all the
questions addressed by forest ecologists (Turner and Gardner 2015).
Processes that occur at a given scale do so within a hierarchy. The process in question is
provided context (influenced and constrained) by the processes at the hierarchical level above it,
while the level below it provides its mechanisms. For example, it is vital in understanding the
growth of trees in forested wetlands to examine those processes that occur at broader and finer
spatial scales than those of an individual tree. Wetland tree growth is constrained by broader-scale
processes such as landscape position and soils, but the mechanisms that influence tree growth,
such as tree physiology, occur at finer scales. A classical hierarchical model shown in Figures 2.1
and 2.2 suggests that many forest managers, ecologists, and naturalists would have an interest at a
microscale (up to 100 ha and 500 years), which would encompass the size and life span of a typical
forest in North America (Delcourt and Delcourt 1988). Disturbance events such as clear-cuts, some
wildfires, and wind events, as well as biotic responses such as gap-phase replacement, competition,
15
16 Chapter 2 Landscape Ecosystems at Multiple Scales
SCALE PARADIGM F I G U R E 2.1 General spatial–temporal scales for

the hierarchical characterization of disturbance
109 regimes, ecosystem and biotic responses, and
megascale
vegetation units. Quaternary studies at the macro-
scale refer to investigations of species migrations and
extinctions in the time frame of 10 000–1 000 000 years
in the Quaternary Period (see Chapter 15) on a
TEMPORAL SCALE (yr)
continental scale. Source: After Delcourt and

106 Delcourt (1988) / Springer Nature.
macroscale
(QUATERNARY STUDIES)
mesoscale
103 (LANDSCAPE ECOLOGY)
microscale
100
100 104 108 1012
SPATIAL SCALE (m2)
and productivity might occur at this spatiotemporal scale (Figure 2.2). Many landscape ecologists
are concerned with the mesoscale (100–10 000 ha and 500–100 000 years), which encompasses
longer-term processes such as fire regimes, insect or pathogen outbreaks, and soil development, as
well as biotic processes such as secondary succession. Seasonal patterns of precipitation and
temperature and longer-term (decades to centuries) weather trends and climatic fluctuations affect
the distribution, growth, and interactions of plants and animals at microscale and mesoscale.
Macroscale and megascale (Figures 2.1 and 2.2) extend in space and time to include not only
events of ecosystem change and plant migration occurring at the level of physiographic regions
(macroscale), but also plate tectonics and evolution of biota on the spatial scale of continents and
the ecosphere (megascale).
The disturbance regimes, ecological responses, and vegetation units that are of concern to
forest ecologists may occur over a range of spatiotemporal scales rather than cleanly within one
shown in the hierarchical framework in Figure 2.2. Disturbance events such as wildfire or insect
and pathogen outbreaks may occur at a local site or over thousands of square kilometers. Wind
may uproot a single tree in an old-growth forest or may affect huge land areas for much longer
intervals, such as the 150 000 ha (370 000 acres) of forest blown down in the Boundary Waters
Canoe Area in northern Minnesota in 1999 (Mattson and Shriner 2001). Likewise, geomorphic
processes such as soil creep, debris avalanches, stream transport, and deposition of sediments
affect vegetation at levels from individual plants to large forest stands and occur on areas extending
from the size of sample plots (m2) and stream watersheds (ha) to mountain ranges (km2). Vegeta-
tion change or succession is another good example of an ecosystem process that may be examined
at different spatial and temporal scales. For example, vegetation change may occur in just a few
years in small tree-fall gaps, as widespread secondary succession over decades or centuries, or as
evolution over millennia on a continental spatial scale. Even the same disturbance type may occur
at different scales among ecosystems. For example, subalpine lodgepole pine forests of Yellowstone
National Park were historically burned by large fires (>5000 ha) every 100–300 years (Romme and
Despain 1989), which spans both the microscale and mesoscale in Figure 2.1. By contrast,
hemlock-northern hardwood forests of the Lake States were likely burned by very small fires
Spatial Scales of Hierarchical Landscape Ecosystems 17
ENVIRONMENTAL ECOSYSTEM AND

DISTURBANCE REGIMES BIOTIC RESPONSES VEGETATIONAL PATTERNS
100 104 108 1012 100 104 108 1012 100 104 108 1012
108 Global Terrestrial 108

Vegetation
Plate Evolution of
tectonics the Biota
Temporal Scale (years)
FM.
Zone
106 108
Glacial– Ecosystem Change
Interglacial Forma-
Biota Landforms tion
Climatic Cycles
Soil Development Speciation Soils
Macro-and-
Migration Microclimate
Climatic Fluctuations
103 Stand
Type 108
Human Activities Secondary Subtype
Succession
Tree
Gap-phase
Fire Regime Replacement
Pathogen Outbreak Competition Shrub
*Disturbance Events Productivity Herb
100 108
100 104 108 1012 100 104 108 1012 100 104 108 1012
*Examples: Wildfire, Spatial Scale (m2)
Wind Damage,
clear-cut, Flood,
Earthquake
F I G U R E 2 . 2 Disturbance regimes, ecosystem and biotic responses, and vegetational units viewed in the
context of four spatial–temporal scales shown in Figure 2.1. Source: Modified from Delcourt and Delcourt, 1988 /
with permission of Elsevier.
(<5 ha) over very long intervals (approximately 1000 years; Frelich and Lorimer 1991), which also
spans both the microscale and mesoscale.
SPATIAL SCALES OF HIERARCHICAL LANDSCAPE ECOSYSTEMS

The ecosphere is the largest and most all-inclusive ecosystem that includes within it volumet-
ric segments (i.e., landscape ecosystems). Three general levels of landscape ecosystems are
typically identified (Rowe and Sheard 1981; Rowe 1992b; Bailey 2009). Macroscale landscape
ecosystems (scale in this context refers to spatial scale) are seas and continents and their major
units, essentially encompassing the macro- and mega- scales shown in Figure 2.2. Continents
contain mesoscale landscape ecosystems, usually major physiographic features such as broad
plains, rolling hills, and mountain ranges. These major physiographic features contain microscale
landscape ecosystems that are local landforms supporting upland forests, swamps, and lakes.
This hierarchical framework proceeds from the more complex and heterogeneous ecosystem
units at broad spatial scales to less complex and relatively homogeneous at fine spatial scales.
Note the three levels of the hierarchy cited in the earlier text are arbitrary, as the complexity of a
landscape and the ecological questions at hand will determine the number of levels necessarily
distinguished.
Ecosystems at each level, whether at the broad regional scale (macroscale and mesoscale
landscape ecosystems) or at a local scale (microscale), all include climate, physiography, soil, and
organisms. It would be ideal to delineate and map landscape ecosystems by integrating all these
components at all levels, but this formidable task has been accomplished only at local scales (e.g.,
Barnes et al. 1982; Pregitzer and Barnes 1984; Spies and Barnes 1985a; Simpson et al. 1990). Despite
failing to integrate multiple factors, single-component classifications of ecosystems by climate,
vegetation, or physiography are useful to illustrate the spatial pattern of the focal component. In
the sections provided in the following text, we provide examples of these approaches as an avenue
to emphasize the complexity of landscape ecosystems at multiple spatial scales.
CLIMATIC CLASSIFICATION
Climate is the source of energy and water for all ecosystems, and thus it is the most important eco-
logical factor at the macroscale. Together with physiography, climate strongly controls genetic
differentiation of organisms and hence their distribution in space and time. Climate may be classi-
fied at the broadest spatial scale, such as the “ecoclimatic zones” of Bailey (1988; Figure 2.3), to
illustrate the pattern of climate across the ecosphere. In this way climate may be used to distin-
guish macroscale and mesoscale ecosystem units to which other ecosystem components can be
linked. Each of the 13 ecoclimatic zones in Figure 2.3, adapted from the climatic classification of
Köppen (1931) (modified by Trewartha 1968), is characterized by a particular climatic regime (the
characteristic seasons and daily ranges in temperature and moisture). Each climatic zone essen-
tially corresponds to a general zonal soil type (usually at the level of soil order) and a broad, late-
successional vegetation type (Table 2.1). These climatic zones span the continents and in many
cases include high mountains. Mountains provide vertical stacks of ecoclimatic zones or sequences
of altitudinal belts that are related to elevation, slope aspect, and associated climatic and soil char-
acteristics (Figure 2.4) within the context of the horizontal ecoclimatic zones. The sequence of
altitudinal belts may differ according to the geographic position of the zone in which it is located.
In this example, climate is the primary factor for classification of mountainous landscapes
into horizontal and vertical zones at the broadest spatial scales. At the scale of continents, gross
physiographic features and vegetative cover also play a major role in affecting climate both hori-
zontally and vertically. At the mesoscale, Bailey (2009, 1988) subdivides climatic zones using
physiography (specific landforms and their surficial form and parent material) as the key factor.
Physiographic features affect the distribution and composition of vegetation by influencing both
macroclimate and microclimate (Chapter 8).
0°
Ar Tropical Wet Cf Subtropical Humid

Aw Tropical Wet-Dry Do Temperate Oceanic
BSh Tropical/Subtropical Semi-arid Dc Temperate Continental
BWh Tropical/Subtropical Arid E Boreal
BSk Temperate Semi-arid Ft Polar Tundra
BWk Temperate Arid Fi Polar Ice Cap
Ca Subtropical Dry Summer H Highlands
FIGURE 2.3 Ecoclimatic zones of the world. Source: Bailey (2009) / Springer Nature.
Spatial Scales of Hierarchical Landscape Ecosystems 19
Table 2.1 Zonal relationships between climate, soil orders, and vegetation.a
Ecoclimatic zoneb Soil orderc Vegetation

Ar Oxisols (latosols) Evergreen tropical rainforest (selva)
Aw Oxisols (latosols) Tropical deciduous forest or savanna
BS Mollisols, aridisols (chestnut, brown soils, and Shortgrass
sierozems)
BW Aridisols (desert) Shrubs or sparse grasses
Cs Entisols, inceptisols Sclerophyllous woodlands
(Mediterranean brown earths)

Cf Ultisols Coniferous and mixed coniferous–
deciduous forest
(red and yellow podzolic)
Do Alfisols (brown forest and gray-brown podzolic) Coniferous forest
Dc Alfisols Deciduous and mixed coniferous–
deciduous forest
(gray-brown podzolic)
E Spodosols and associated histosols (podzolic) Boreal coniferous forest (taiga)
Ft Entisols, inceptisols, and associated histosols Tundra vegetation (treeless)
(tundra humus soils with solifluction)
a
After Walter (1984, p. 3).
b
Abbreviations of the zones are shown in Figure 2.3.
c
Names in parentheses are soil orders (USDA Natural Resources Soil Conservation Service, 1999).
Source: Bailey (1988) / United States Department of Agriculture / Public Domain.
N
6000 m
4000
2000
0
A B C D
70° 65° 55° 45° 35°

Edmonton Denver
1 2 3 4 5 6 7 8
F I G U R E 2 . 4 Vertical zonation (cover types 1–8) within boreal (B) and temperate semi-arid (D) ecoclimatic
zones (Figure 2.2) from 70° to 35° N latitude along the eastern slopes of the Rocky Mountains from approxi-
mately 102° to 115° W longitude. Climatic zones: A, transition; B, boreal; C, transition; D, temperate
semi-arid. Vertical zones: 1, ice region; 2, mountain vegetation above tree line; 3, boreal and subpolar open
coniferous woodland; 4, boreal evergreen coniferous forest; 5, boreal evergreen mountain coniferous forest;
6, coniferous dry forest; 7, shortgrass dry steppe; 8, boreal evergreen coniferous forest with cold-deciduous
broad-leaved trees. Source: Rowe (1988) / United States Department of Agriculture / Public Domain.
PHYSIOGRAPHY
Physiography is the Earth’s surface relief and its associated geologic substance or parent material
and is perhaps the single most useful component in distinguishing landscape ecosystems. The
Earth’s structures and forms mediate the fluxes of energy and materials at the Earth’s surface
where atmospheric and earth layers meet, and thus physiography often has a strong influence on
climate in a given place. Vegetation often corresponds to landforms and to the soil that develops
from their parent material. Therefore, the physiographic regions of North America (Figure 2.5)
provide a useful basis for distinguishing broadscale landscape ecosystems at the macroscale and
mesoscale.
Fenneman (1938) created a hierarchical physiographic classification in eastern North
America that provided the basis for Braun’s vegetation classification (1950) and Omernick’s ecore-
gion classification (1987). Fenneman’s physiographic classification provided the landscape frame-
work for Braun’s classification of deciduous forest regions and sections of the eastern United States
(Figure 2.6). The boundaries of these broad regional geologic units also matched well with major
climatic regions (Bryson and Hare 1974), such as those identified by the Holderidge (1947, 1967)
model of world bioclimatic formations or life zones. Holderidge’s classification identifies a series
of broad “life zones,” such as rain forest, wet forest, dry forest, tundra, and steppe, based on “bio-
temperature” (the sum of average monthly temperature above 0 °C divided by 12), precipitation,
and the ratio of potential evapotranspiration/precipitation (E/P). Lindsey and Sawyer (1970) and
Lindsey and Escobar (1976) used Holderidge’s classification to compare the macroclimatic condi-
tions among eight of Braun’s forest regions using data from 1100 National Weather Service stations.
The physiographically based forest regions are well separated, with exceptions where forest regions
had similar canopy species composition among the regions at low and moderate elevations. This
example illustrates that physiography is an effective way to delineate landscape ecosystems at
regional scales, particularly when coupled with the distribution of pre-European colonization or
near-natural vegetation.
VEGETATION TYPES AND BIOMES

The physiognomy (form) and species composition of vegetation may also serve to delineate the
extent and boundaries of ecosystems at broad scales (macroscale and mesoscale). In fact, vege-
tation has been the traditional criterion for distinguishing broad landscape units, at least in
part because it is easier to describe and map than physiography or climate. Moreover, vegeta-
tion is an effective integrator of other ecosystem factors. We emphasize, however, that any
single ecosystem component—vegetation, climate, soil, or landform—is an incomplete
indicator of the whole.
At very broad spatial scales, macroscale ecosystems roughly coincide with the formations,
plant community types, life zones, biotic provinces, or biomes that biogeographers traditionally
recognize to reflect macroclimatic regimes and major physiographic regions. The term formation
is applied to macroscale plant communities, but the term biome (biotic region) is more commonly
used because plants and animals are interrelated parts of these units. Broadscale climate patterns
such as air masses and movement patterns correlate well with the distribution of some biomes
(Tallis 1991; Figure 2.7). This relationship is particularly true for the boreal forest, whose southern
limit corresponds closely with the typical southern extent of Arctic air masses during the winter
and whose northern limit corresponds closely to the southern extent of Arctic air masses during
the summer (Bryson 1966; Barry 1967).
Thus, the location of major biotic regions (Figure 2.8) indicates the location and extent of the
Earth’s macroscale ecosystems, if we conceive these ecosystems as extensions of the biota
Vegetation Types and Biomes 21
St. Lawrence Lowlands Province Basin and Range Province

Central Lowlands Province Colorado Plateau Province
Great Plains Province Columbia–Snake Rivers Plateau Province
Blue Ridge Province Interior Mountains and Plateaus System
New England and Maritime Province Pacific Border System
Adirondack Province Rocky Mountains System
Piedmont Plateau Province Coastal Plain Province
Valley and Ridge Province Arctic Lowlands and Coastal Plain Province
Appalachian Plateau Province Greenland Shield Province
Ozark Plateau Province Hudson Bay Lowlands Province
Ouachita Mountains Province Canadian Shield System
Caledonian Province Interior Low Plateaus Province
Wyoming Basin Province Glacier or icecap
F I G U R E 2 . 5 Physiographic regions of North America and their regional landforms. Source: Brouillet and
Whetstone (1993) / Oxford University Press.
B
8
B
9c
9a
8b 9b 9d 9g
9a 9a 9f
8a
G 7
9e
3c
7 4c
3d 4d
1b 4
2c b
2a 1c
2d
3a 2e 1a
2b 5a
2f
3b 4a
6a
5b
G 5b
6
6
F I G U R E 2 . 6 Eastern North American map of deciduous forest regions and sections of the eastern and
midwestern United States. Formation, region, and section names: B = boreal or spruce–fir forest formation;
G = grassland or prairie formation; S = subtropical broad-leaved evergreen forest formation. Regions and
sections of the deciduous forest formation: 1 = mixed mesophytic forest region: a, Cumberland Mountains; b,
Allegheny Mountains; c, Cumberland and Allegheny Plateaus. 2 = western mesophytic forest region, sections:
a, bluegrass; b, Nashville Basin; c, area of Illinoian glaciation; d, hill; e, Mississippian Plateau; f, Mississippi
Embayment. 3 = oak–hickory forest region: southern division, sections: a, interior highlands; b, forest–prairie
transition area; northern division, sections: c, Mississippi Valley; d, Prairie Peninsula. 4 = oak–chestnut forest
region, sections: a, Southern Appalachians; b, Northern Blue Ridge; c, ridge and valley; d, Piedmont; e,
glaciated. 5 = oak–pine forest region, sections: a, Atlantic slope; b, Gulf slope. 6 = southeastern evergreen
forest region, sections: a, Mississippi Alluvial Plain. 7 = beech–maple forest region. 8 = maple–basswood
forest region, sections: a, driftless; b, big woods. 9 = hemlock–white pine–northern hardwoods region: Great
Lakes-St. Lawrence division, sections: a, Great Lake; b, Superior Upland; c, Minnesota; d, Laurentian;
Northern Appalachian highland division, sections: e, Allegheny; f, Adirondack; g, New England. Source:
Based on a map of physiographic provinces and sections of Fenneman (1938). Braun (1950) / McGraw-Hill.
(bioecosystems). Many textbooks with this biotic emphasis include similar diagrams and descrip-
tions of these major units. Detailed descriptions of the major vegetation types shown in Figure 2.8,
and subdivisions of these types are provided by several authors (Rowe 1972; Vankat 1979; Barbour
and Billings 1988; Barbour and Christensen 1993). In addition, books in the series “Ecosystems of
the World” (for example, Tropical Rain Forest Ecosystems (Golley 1983), Temperate Broad-Leaved
Vegetation Types and Biomes 23
Arctic
80
HL
60
W Polar
40 D ML
S W
Tropical
D
20
S
0 W LL Equatorial
S
20 D
Tropical
S D W
40 ML
W
Polar
Biomes of wet climates 60
Evergreen rainforest
Temperate forest
Biomes of climates with seasonal rain:
Tropical seasonal forest
80 Antarctic
Mediterranean scrub and woodland
Biomes of subhumid climates: HL = high-latitude climatic belt
Savanna and dry woodland ML = mid-latitude climatic belt
Steppe LL = low-latitude climatic belt
Boreal forest
Biomes of arid climates: D = dry
Desert
Tundra
F I G U R E 2 . 7 Distribution of biomes (left) and source regions and air masses (right) in relation to a
hypothetical land mass. Source regions are shown shaded; arrows show direction of prevailing winds.
HL = high-latitude climatic belt; ML = mid-latitude climatic belt; LL = low-latitude climatic belt; D = dry;
W = wet; S = seasonal precipitation. Source: Tallis (1991) /Springer Nature. Reproduced with permission
from Plant Community History by Chapman and Hall.
Evergreen Forests (Ovington 1983), and Temperate Deciduous Forests (Röhrig and Ulrich 1991))
provide comprehensive accounts of many forest communities and ecosystems.
Formations or biomes are broadscale groupings of ecosystems based on similar plant
community form and structure, or plant physiognomy. Plant physiognomy is a direct response
to physical factors, such as the temperate and tropical rain forests of warm and wet climates in
contrast to tundra and sclerophyllous scrub communities of cold and/or dry climates. Macro-
climate (general climate over a large geographical area) is often perceived as the primary envi-
ronmental factor controlling the form, vertical layering, and composition of these broad units.
However, physiographic features and vegetation also influence climate, especially where
abrupt discontinuities affect the exchanges of energy and materials at the surface (Hare and
Ritchie 1972). Figure 2.8 shows differences in the distribution of vegetation such as tundra,
coniferous forests, deciduous forests, grassland, and tropical rain forest that reflect, to a certain
degree, the “ecoclimatic zones” of Figure 2.3. There are, however, notable differences between
the two figures even at the broad scales of these major units.
One well-known classification system based on vegetation is the United States National Veg-
etation Classification (USNVC 2019). The USNVC is a hierarchical system with eight levels that
emphasizes physiognomy (in this case dominant growth forms) at its highest three levels
(Formation Class, Formation Subclass, and Formation) as it reflects environment at broad (global
to continental) spatial scales. For deciduous forests in the northern Lake States, for example, these
three highest levels would be forest and woodland (Formation Class), Temperate and Boreal Forest
L J
J
J
L J
B J
F L
J J
B
L
F
F KB B
B J
F
K
K K B
F
I B K L
K
I
I L
B B J L
K
L
I B
I J
B
B
D C
D D D
C C C D
D
C C
D C
D
C C
D C
A Temperate rain forest G Open scrub savanna
C
B Coniferous forest H Tree savanna C
Desert, desert scrub,
C Tropical rain forest I sagebrush, and saltbush
Tropical deciduous forest Tundra, Alpine fields, and
D J
and thorn forest other mountain vegetation
E Sclerophyll forest K Grassland
F Sclerophyll scrub L Temperate deciduous forest
J
J
J
J B J
J
J J
B
B
B
A
B
K B
B
I K
K
B
L
F B
I H
H C H
D D D C
D H
D
C
F I G U R E 2 . 8 The major biotic regions or biomes of Eurasia and North America. Source: Dasman (1984) /
John Wiley & Sons.
Distinguishing and Mapping Landscape Ecosystems at Multiple Spatial Scales 25
and Woodland (Formation Subclass), and Cool Temperate Forest and Woodland (Formation). At
increasingly lower levels of the hierarchy (Division, Macrogroup, and Group), biogeography and
floristics (similarities in species composition) are emphasized as well as physiognomy reflecting
environmental factors at spatial scales of continents to regions. The northern Lake States decidu-
ous forest would fall within the sugar maple–American beech–northern red oak Forest and Wood-
land Division, perhaps the sugar maple–yellow birch–eastern Hemlock Forest Macrogroup, and
the American beech–sugar maple–yellow birch Forest Group. The lowest levels of the hierarchy
(Alliance and Association) are based mainly on physiognomy and compositional similarity
reflecting local to regional environmental factors. The northern Lake States deciduous forest might
fall within the sugar maple–American basswood–white ash Forest Alliance and the sugar maple–
yellow birch–American basswood Forest (USNVC 2019). Although the classification interprets
physical factors (primarily climate but also geology, soils, and hydrology at lower levels) as the
ecological context for vegetation at each level, it does so as it is reflected by the vegetation, and thus
heavily emphasizes the biotic component of ecosystems rather than the ecosystems as a whole.
Vegetation classification is certainly useful for applications focusing on vegetation alone, but such
a classification should not be confused with one that classifies landscape ecosystems.
Classifications and maps of the individual ecosystem components of physiography, climate,
and vegetation are useful, but their boundaries do not typically correspond well. Landscape ecosys-
tems are volumetric units of many interrelated components, and thus mapping them demands the
simultaneous integration of physiographic features with those of climate, soil, and vegetation at
multiple spatial scales. Mapping necessarily depends on visible features that can be observed on
satellite and aerial imagery as well as on the ground, that is, landform and vegetation. However,
these individual components, though easy to observe and conceive, are not ecosystems themselves.
In particular, biomes are not ecosystems, but they are often treated as such when in fact they are
only reflections of the physical features of a given place. Likewise, physiographic features are not
ecosystems because they lack the biotic component that would make them ecological. Fortunately,
the ecological relationships of these components provide useful insights to ecosystem components
that are much harder to observe: the mobile animal community, soils, the hydrologic regime, and
the transparent and continuously shifting atmospheric conditions that comprise climate.
DISTINGUISHING AND MAPPING LANDSCAPE ECOSYSTEMS

AT MULTIPLE SPATIAL SCALES
A useful model to describe the progressive dividing of the ecosphere into a hierarchical series
of ecosystems would be Chinese boxes or Russian dolls, with a series of nested units fitting one
inside the other. The division of landscapes into smaller hierarchical units from the top-down is
termed regionalization. Regionalization provides the basis for a logical grouping or classification
of ecosystems that have practical significance for management or conservation. No one standard-
ized scheme encompasses the enormous diversity of ecosystem patterns around the world, as the
number and name of ecosystem levels will vary depending on the size and ecological configuration
of the landscape, the purposes for which the classification is developed, and personal preferences
for names of the levels. One example of such a hierarchical system of units was proposed by Klijn
and Udo de Haes (1994), using the prefix “eco” to indicate entire ecosystems in a holistic sense
(Table 2.2). Each level in the hierarchy can be matched with general levels of macro-ecosystem,
meso-ecosystem, and micro-ecosystem that correspond to the spatial scales (macro, meso, and
micro) shown in Figures 2.1 and 2.2. The corresponding ecological units of the system in wide use
today by the US Forest Service (Cleland et al. 1997) also correspond well to the hierarchical levels
of this example (Table 2.2).
A key feature of ecosystem geography is that ecosystems at different spatial scales are nested
within one another. Climate and gross physiography of a broadscale ecosystem (macroscale) affect
Table 2.2 Overview of hierarchical ecosystem classification at various spatial scales within the general
framework of macroecosystem, mesoecosystem, and microecosystem levels.
Ecosystem Size of basic US Forest Service

General ecosystem levela Mapping scaleb
unit nameb mapping unitb ecological unit namec
MACRO-ECOSYSTEM Ecozone 1: >50 000 000 >62 500 km2 Domain Division
Ecoprovince 1: 10 000 000–50 000 000 2500–62 500 km2 Province
MESO-ECOSYSTEM Ecoregion 1: 2 000 000–10 000,000 100–2500 km 2
Section
Ecodistrict 1: 500 000–2 000 000 625–10 000 ha Subsection
Ecosection 1: 100 000–500 000 25–625 ha Landtype
association
MICRO-ECOSYSTEM Ecoseries 1: 25 000–100 000 1.5–25 ha Landtype and
landtype phase
Ecotype 1: 5000–25 000 0.25–1.5 ha Landtype phase
Eco-element 1: <5000 <0.25 ha Landtype phase
a
Correspondence with spatial scales shown in Figures 2.1 and 2.2 (Delcourt and Delcourt, 1988): Macroecosystem level
approximates megascale and macroscale combined; mesoecosystem level approximates mesoscale; microecosystem
level approximates microscale.
b
After Klijn and Udo de Haes (1994).
c
US Forest Service National Hierarchical Framework of Ecological Units (Cleland et al., 1997).
Source: Modified from Klijn and Udo de Haes (1994).
ecological properties and processes of those at finer scales (mesoscale) and the fine-scale (micro-
scale) ecosystems they in turn contain. Climate and physiography of one broadscale ecosystem will
therefore affect its constituent ecosystems differently than the climate and physiography of an
adjacent broadscale ecosystem would affect its own finer-scale ecosystems. It therefore follows that
a given management practice in one broadscale ecosystem will not necessarily have the same
results in a different broadscale ecosystem. Recognizing ecosystem hierarchies therefore avoids the
mistake of extrapolating management policies and practices too broadly across the landscape.
Mapping ecosystem units at multiple scales discourages managing isolated bits and pieces of
ecosystems (species, stands, soils, etc.) and encourages a more integrated perspective of the link-
ages among ecosystems and their components. Robert Bailey’s (2009) book, Ecosystem Geography,
provides a detailed treatment of the theory and principles of ecosystem classification and mapping
at multiple scales.
REGIONAL LANDSCAPE ECOSYSTEMS

The process of delineating broadscale regional ecosystems has been described in detail
(Bailey 1983, 1988, 2009; Bailey et al. 1985; Bailey and Hogg 1986; Omernik 2004, 1995; Omernik
and Griffith 2014). Broadscale maps of ecoregions in the United States have been developed by
Omernik (1987) and Bailey (1995, 1994), the former used extensively by the US Environmental
Protection Agency and the latter by the US Forest Service. A detailed map and description of regional
landscape ecosystems of Michigan, Minnesota, and Wisconsin are also available (Albert 1995).
Ecological regionalization has a longer history in Canada, beginning with Halliday’s (1937)
macroscale forest classification, which was revised by Rowe (1972) as the Forest Regions of Canada.
Mesoscale classification then followed by Hills (1960) and Wickware and Rebec (1989) in Ontario,
Loucks (1962) in the Maritimes, and MacKinnon et al. (1992) in British Columbia. A comprehensive
and multipurpose classification of seven hierarchical levels from broad “ecozones” to site-specific
“ecoelements” provided a Canada-wide system (Wicken 1986).
Distinguishing and Mapping Landscape Ecosystems at Multiple Spatial Scales 27
P133
121 M122
M132 121
131
131
133 133
P212
M241
254
133
M211 211
253
241
M312
311 211 211
M311
M311
221
313 A314
251
M281 311
253 221
261 M311
232
P313
M262 M312 232

322
314
252 251
231
231
321
411
316
M316
M321
Highlands with Complex Zonation
M-mountains
M315
315 414 P-plateau
A413
421
414 A-altipiano
M412 Ecoregion Boundaries

M421
Domain
1000 Km Division
415 Province
F I G U R E 2 . 9 Ecoregion map for North America. Ecoregion code: 100-series, polar domain; 200-series,
humid temperate domain; 300-series, dry domain; 400-series, humid tropical domain. Key to letter symbols:
M = mountains; P = plateau; A = altiplano. Source: Bailey and Hogg (1986) / Environmental Conservation.
The highest hierarchical levels of the regionalization of North America (Bailey and
Hogg 1986) are termed domain, division, and province, and are reasonably equivalent to ecosys-
tems at the macroscale (Figure 2.9). Domains are distinguished by climate (polar, humid temperate,
dry, and humid tropical) and correspond broadly with “ecoclimatic zones” (Figure 2.3), physio-
graphic regions (Figure 2.6), and the biotic regions of North America (Figure 2.8). The four
domains differ among each other, but also themselves include great spatial heterogeneity in cli-
mate, physiography, soil, and vegetation. For example, the humid temperate domain (Figure 2.9)
includes temperate rain forests of the Pacific Northwest, sclerophyllous scrub of southern
California, grasslands of the Great Plains, deciduous forests of the eastern United States, and
c oniferous forests of the northeastern United States and adjacent Canada (Figure 2.8). As domains
are subdivided into divisions, provinces, and finer subdivisions below them, the heterogeneity of
ecosystem components decreases. Divisions are delineated based on finer distinctions in climate,
and provinces are distinguished according to potential natural vegetation (Küchler 1964). Vertical
zonation of landscape ecosystems in mountainous terrain based on elevational belts and associ-
ated mountain physiography of ridges, valleys, and slope–aspect orientation of hillsides is super-
imposed on the geographic framework we described in the earlier text (cross-hatched areas in
Figure 2.9; see also Figure 2.4 and related discussion).
Some argument has been made that watersheds or similar hydrologic units are suitable for
ecological classification (Lotspeich 1980; Montgomery et al. 1995), and in some cases are prefera-
ble to ecoregions delineated with climate and physiography because their definitions are much
clearer and more consistent than ecoregions (Omernik and Bailey 1997). However, using water-
sheds as the basis for ecosystem delineation includes three major pitfalls as described by Bailey
(2009). First, watersheds are not always clearly defined in many parts of the United States, and in
such cases, it is extremely difficult to elucidate the boundaries between them. Second, hydrological
units such as watersheds may be difficult to delineate on a map where surface-water flow does not
coincide well with the movement and flow of groundwater. Finally, it is quite common for the
stream system of a watershed to flow though very different climates and physiographic sys-
tems, such that some watersheds may include a great deal of heterogeneity in site factors that
cannot be easily reconciled even at very broad scales. Whereas broadscale areas with similar cli-
mate and physiography include ecologically relatable units with similar environments, watersheds
often contain ecosystems that are ecologically dissimilar and therefore are too heterogeneous to be
useful (Bailey 2009).
REGIONAL LANDSCAPE ECOSYSTEMS OF MICHIGAN

One of the early examples of delineating regional landscape ecosystems at the mesoscale
(within the province level of Bailey’s ecoregions; Figure 2.9) was completed for Michigan by Albert
et al. (1986). Using a multiscale, multifactor integrated approach, regional landscape ecosystems of
Michigan were determined, described, and mapped at three hierarchical levels termed region, dis-
trict, and subdistrict (Albert et al. 1986). The first two of these levels (region and district) were
determined by integrating climatic and physiographic/soil classifications (Figure 2.10). At this
spatial scale, detailed analytical methods were necessary to first develop a climatic classification of
Michigan using multivariate procedures and separate classifications for growing season tempera-
ture, winter temperature, and growing season precipitation variables (Denton and Barnes 1988).
These classifications were then combined with physiographic data to create a single classification.
The unique geographic location of Michigan consists of two large peninsulas oriented in a perpen-
dicular direction that project into the Great Lakes, which dramatically influences the climatic dif-
ferences of both peninsulas. The distinctiveness of warm, vegetationally diverse southern Lower
Michigan (Region I, Figure 2.10) and cold Upper Michigan (Region IV, Figure 2.10) is in marked
contrast to lake-moderated Regions II and III where district-level ecosystems parallel lakes Huron,
Michigan, and Superior. Within districts, major physiographic features and their associated soil
and vegetation were used to identify subdistricts. Macroclimate is increasingly homogeneous
within a given district and even more so within a subdistrict.
District 1 in southeastern Michigan (approximately 56 000 km2) provides a good example of
the spatial pattern of local ecosystems distinguished at the subdistrict level. Subdistrict 1.1 (shaded
area in Figure 2.10) is the heat island of the Detroit metropolitan area, which is superimposed on
an extensive glacial lake plain of Subdistrict 1.2. The lake plain extends south into Ohio and north
toward Lake Huron where it forms major parts of Districts 5, 6, and 7 (Figure 2.10). This relatively
flat plain of lacustrine clay and sandy soil is interspersed with sand beach ridges and adjacent wet
Regional Landscape Ecosystems of Michigan 29
20.3
Isle Royale
20.2 N
19.1 20.1
uperior
19.2 18.3 L ake S
18.2 16
18.1 14.2
15.4
17.2 13.2
14.1
17.1 15.3 13.1
IV 15.2 13.1
15.1 13.3
III
Univ. of
Mighigan
15.3
Biological
12.3
Station
12.3
n 12.2
a
10 12.2
L
hig
Mic
12.1
ak
11.2
eH
10 11.2
Lake
8.3
uron
11.1 8.2
II
9
10 7.1
7.2
MILES
0 10 20 30 40 50
5.1
0 20 40 60
KILOMETERS 6
3.2 4.2 4.2
SCALE 1:2,000,000 5.2
4.1
3.3 1.2
I
3.2
3.1 1.1
1.1 1.4
2.1
2.2
1.3 1.2
INDIANA OHIO
F I G U R E 2 . 1 0 Map of regional landscape ecosystems of Michigan. Three hierarchical levels are mapped:
Regions I–IV; districts 1 to 20, and subdistricts within the districts. Region I = Southern Lower Michigan,
districts: 1, Washtenaw; 2, Kalamazoo; 3, Allegan; 4, Ionia; 5, Huron; 6, Saginaw. Region II = Northern
Lower Michigan, districts: 7, Arenac; 8, High Plains; 9, Newaygo; 10, Manistee; 11, Leelanau; 12, Presque
Isle. Region III = Eastern Upper Michigan, districts: 13, Mackinac; 14, Luce. Region IV = Western Upper
Michigan, districts: 15, Dickinson; 16, Michigamme; 17, Iron; 18, Bergland; 19, Ontonagon; 20, Keweenaw.
Source: Albert et al. (1986) / United States Department of Agriculture / Public Domain.
epressions. In contrast to the flat lake-plain landform, Subdistrict 1.3 (just to the west) is charac-
d
terized by large, end-moraine ridges of rolling terrain and fine-to medium-textured soils. Adjacent
to it and further west is Subdistrict 1.4, characterized by hilly ice-contact terrain of kettle (wet or
dry depressions) and kame (steep hills) topography with distinctive microclimatic patterns very dif-
ferent from the other subdistricts. Within each of these subdistricts of uniform macroclimate, local
landscape ecosystems and their diverse forest communities occur in distinctive patterns reflecting
those of the diverse landforms.
LOCAL LANDSCAPE ECOSYSTEMS

Thus far we have described the delineation of regional ecosystem units at multiple spatial scales—
from domains to districts and subdistricts over six hierarchical levels. This regionalization process
was based upon progressive division of the landscape using gross climatic and physiographic fea-
tures and associated soil and natural vegetation. At some lower level in the hierarchy, we eventu-
ally reach a land area within which the regional macroclimate is relatively homogeneous. At this
level in the hierarchy, gross physiography and recurrent patterns of vegetation become most
important in distinguishing local landscape ecosystems (microscale). At this scale, major land-
forms and local landscape ecosystem types occur in a mosaic or complex network of interrelated
systems. Thus, delineation of local ecosystems cannot proceed exactly as it did at regional
levels, where boundaries are drawn around large areas distinguished by major discontinuities of
macroclimate, gross physiography, and vegetation (as in Figures 2.9 and 2.10). At a local level,
physiographic systems or major landforms (e.g., outwash plain, mountain slope or plateau, and
moraine) are recognized that recur within a climatically homogeneous area (subdistrict level).
Physiographic systems have a profound influence on topoclimate, soil development, the biotic
community, and successional trends (Chapter 8). Within these physiographic systems, local eco-
system types can then be grouped appropriately by these major physiographic or landform
complexes, the local landforms within them (such as floodplains on outwash physiographic sys-
tems or ground moraine on moraine physiographic systems), and with even more specific site
factors of microclimate, soil texture, and nutrient availability (Pregitzer and Barnes 1984; Spies
and Barnes 1985a; Archambault et al. 1990; Bailey 1995).
In any regionalization of a large land area, identifying the largest unit of relatively homoge-
neous macroclimate (i.e., the district level in Figure 2.10) is of the utmost importance. Homogeneous
macroclimate within such units has great ecological significance because of the intense genetic
adaptations of plants to climatic factors along latitudinal and elevational gradients (Chapter 3). In
an area of homogeneous macroclimate, individual species and groups of plant species (Chapter 11)
indicate different site–factor complexes of water, nutrients, and light (Spies and Barnes 1985b;
Archambault et al. 1989). The caveat is that these same species may not indicate the same relation-
ships in areas of different macroclimate (e.g., different districts in Michigan).
LOCAL LANDSCAPE ECOSYSTEMS IN UPPER MICHIGAN

Even within an area of homogeneous macroclimate and physiography, local ecosystem types occur
and recur in a complex spatial mosaic, particularly in glacial terrain (Figure 2.11). In the ice-
contact terrain of the Sylvania Wilderness Area in Upper Michigan, daily, monthly, and yearly
microclimate of local ecosystems, as well as their soil and vegetation, is markedly different. The
presence of many lakes, bogs, and coniferous swamps (types 16, 17, 18, and 19) and adjacent sandy
uplands (hemlock-northern hardwood types 5, 9, 10, and 11) illustrates the local complexity of ice-
block disintegration features that were formed as the last glacial ice sheet retreated. Many tiny
(<0.1 ha) depressions forming black-ash swamps (type 22) dot the entire area. A lake-and swamp-
edge ecosystem (type 14), characterized by sandy soil and fire-regenerated pines and white birches,
Local Landscape Ecosystems 31
KATHERINE
LAKE
CLARK
LAKE
LEGEND
Ecosystem
Type N
Intermittent Stream 0 1/4 1/2 MILE
Section Line 1°30´
0
Hiking Trails
Paved Road
0 .8 KM
F I G U R E 2 . 1 1 Example of the patterns of local landscape ecosystems that occur within the local level. The
map illustrates local landscape ecosystem types surrounding part of Clark Lake in the north central part of
the Sylvania Wilderness Area, Ottawa National Forest, Upper Michigan. Area shown mapped: 13.9 km2
(5.4 mi2); 1390 ha (3435 acres).
repeatedly occurs along the fire-prone margins of lakes and large swamps. In the unique shoreline
landscape position, offshore winds continuously dry the site and coniferous needles of the forest
floor, which ignite readily when lightning strikes the tall red and white pines. Geological processes
that formed the Sylvania Wilderness have therefore created a remarkable diversity of ecosystem
sizes, shapes, and spatial patterns of wetland, upland, and lake-shore types. Managing areas with
local ecosystems as small as 0.1 ha, especially when they occur in an enormously complex pattern
over thousands of square kilometers, is clearly a major challenge for wildlife ecologists, foresters,
and land managers. Insights on how to deal with such complexity are considered in Chap-
ters 11 and 22.
Multiscale, multifactor landscape ecosystem classification has proven to be incredibly useful
as a framework for many different aspects of ecological research, as reflected by its adoption across
the United States, Canada, and in many parts of the world. Since the last revision of this textbook,
for example, landscape ecosystem classifications have been completed for portions of North Caro-
lina (McNab et al. 1999), Alabama (Carter et al. 1999), the Gulf Coastal Plain in Georgia (Goebel
et al. 2001), northern Michigan (Kashian et al. 2003), South Carolina (Abella et al. 2003), Indiana
(Dolan and Parker 2005), Vermont (Ferree and Thompson 2008), the southern Appalachian Moun-
tains (Stottlemyer et al. 2009), and Nova Scotia (Cameron and Williams 2011), among many other
areas. In the late 1990s, The Nature Conservancy, a global environmental conservation organiza-
tion, developed a continental-scale ecoregional system based on a regional landscape ecosystem
framework and has used it to prioritize areas for conservation in the Western Hemisphere (The
Nature Conservancy 1996). Landscape ecosystem classifications resembling the system we
described in the earlier text have also been completed in Italy (Blasi et al. 2000), the Czech Republic
(Kusbach and Mikeska 2003), and Iran (Azizi Jalilian et al. 2020).
SUGGESTED
R E A D I N G S
Albert, D. A. (1995). Regional landscape ecosystems of Omernik, J.M. and Griffith, G.E. (2014). Ecoregions of the
Michigan, Minnesota, and Wisconsin: a working conterminous United States: evolution of a hierarchi-
map and classification. USDA For. Serv. Gen. Tech. cal spatial framework. Environ. Manag. 54: 1249–1266.
Report NC-178. North Central For. Exp. Sta., St. Paul, Rowe, J.S. (1992). The ecosystem approach to forestland
MN. 250 pp. + map. management. For. Chron. 68: 222–224.
Bailey, R.G. (2009). Ecosystem Geography, 2e. New York: Rowe, J.S. and Sheard, J.W. (1981). Ecological land
Springer-Verlag 251 pp. + 2 maps. classification: a survey approach. Environ. Manag.
Bailey, R.G. and Hogg, H.C. (1986). A world ecore- 5: 451–464.
gions map for resource reporting. Environ. Conserv. Spies, T.A. and Barnes, B.V. (1985). A multi-factor eco-
13: 195–202. logical classification of the northern hardwood and
Denton, S.R. and Barnes, B.V. (1988). An ecological conifer ecosystems of Sylvania Recreation Area,
climatic classification of Michigan: a quantitative Upper Peninsula of Michigan. Can. J. For. Res.
approach. For. Sci. 34: 119–138. 15: 949–960.
The Forest Tree PA R T 2
I n the three chapters of this part, our focus is on forest organisms and their populations
in relation to the forest ecosystems of which they are a part. In Chapter 3, the genetic
differentiation of populations is seen to be closely related to the specific climatic and
physiographic conditions in which plants grow. In addition, we examine the considerable
within-plant variation that is important in the survival and persistence of woody plant
populations. The processes and factors affecting variation within and among populations
and at the species level, such as hybridization and polyploidy, are also considered.
In Chapter 4, the life history of forest organisms is considered with emphasis on
the regeneration of woody plants. The important processes of reproduction, dispersal,
germination, and establishment are discussed. In Chapter 5, the focus is on how trees
grow in relation to site conditions. We examine the structural parts of woody plants, the
crown and its architecture, and the structure and growth of stems and roots. We also
emphasize the effects of water stress on the growth of plants.
33
Forest Tree Variation CHAPTER 3
G enetic adaptations of woody plants inseparably link these organisms to the environment.
Genetics are the governing biochemical control mechanism that can be passed on from gen-
eration to generation; without such a mechanism, a forest tree or any other organism could not
perpetuate itself. Both genetic and environmental factors together control the form and development
of an organism, and it is therefore a futile exercise to argue which controls the phenotype. Environ-
mental factors are the most obvious because they are generally visible and readily accessible, and
it is natural that ecologists were long preoccupied with their study. Genetic factors have since
become more fully appreciated and assessed in forest trees.
Organisms are able to live and reproduce in a given range of environments because of their
adaptedness (Dobzhansky 1968), an ability critical to forest ecology. In the present chapter, we
emphasize variation in organisms: sources of variation; the kinds and extent of variation within
and among individuals, populations, and species; and how physical and biotic factors elicit adaptive
changes in tree populations. Detailed treatments are available for the fields of population genetics
(Gillespie 2004; Hartl and Clark 2006; Hamilton 2009; Charlesworth and Charlesworth 2010; He-
drick 2011; Allendorf et al. 2012; Hartl 2020), geographic variation in trees (Morgenstern 1996),
and forest genetics and tree improvement (White et al. 2007).
COMPONENTS OF PHENOTYPIC VARIATION

Ecologists work with organisms both as individuals and as aggregations of individuals, such as
populations and communities. The genetic constitution of an individual is termed its genotype.
We do not physically witness a plant’s genotype because from the moment of fertilization the geno-
type is influenced by the environment—the plant’s internal environment of cells and biochemical
reactions and the external environment of light, temperature, and moisture. We can only see the
result—the phenotype—the observable properties of an organism produced by the genotype
together with the environment. This relationship can be expressed for the entire organism, or for
individual characters, by the simple formula P = G + E + GE. The phenotype (P) is the sum total
of the effects of these components: the genetic information coded in the chromosomes (G), the
environment (E), and the interaction of the genotype and the environment (GE). The genotype–
environment interaction is often disregarded, but in certain cases it is of major importance. The
genotype itself may be “observed” using biological assays, whereby the data obtained can be com-
pared against either a second individual’s gene sequence or a database of sequences. Genes control
physiological functions through complex pathways, and many of these influence the morphology
of the plant. The basic interrelationship of the genotype, environment, and plant processes of the
phenotype is illustrated in Figure 3.1.
Ecologists often examine the degree to which a phenotypic character is controlled by the geno-
type or by the environment. For example, consider two adjacent trees of the same species, where one
is a dominant tree exhibiting excellent growth in a forest plantation and the second is dying. The
dominant clearly has a superior phenotype, but to what degree is its genotype responsible, and to
35
36 Chapter 3 Forest Tree Variation
Sources of Phenotypic Types of Phenotypic

Differences Differences
GENOTYPE
Genes: DNA
RNA
PLANT PROCESSES PHENOTYPE

Chemical reactions Age
Photosynthesis Habit
Respiration Growth: rate
EXTERNAL ENVIRONMENT Water and mineral periodicity
Time uptake complexity
Space Transpiration
Light Translocation
Heat Cell division Resistance to:
Water pests
Soil moisture extremes
Nutrients light extremes
Chemicals heat extremes
Other organisms Symbiosis
F I G U R E 3 . 1 Relationship of an individual’s phenotype to its genotype and environment. Differences in

plant phenotypes have their origin in the genetic constitution (genotype) of an individual and the effects of
environmental factors on the genotype. Phenotypic differences of the individual become apparent as
physiological processes occur in the internal environment of plant cells, tissues, and organs.
Source: Courtesy of J. W. Hanover.
what degree does a favorable microenvironment or the random factor of better handling in the nurs-
ery account for the difference? Do environment and genotype contribute equally, or is one more
important than the other, and by how much? Despite its poor phenotype, the genotype of the dying
tree may or may not be superior to that of the dominant. Both the genetic and environmental compo-
nents are always involved, and no two phenotypes are exactly alike, but the relative degrees of genetic
and nongenetic influences on a given character are important factors.
To determine the relative effects of genotype and environment, forest geneticists compute
genetic (Vg) and environmental (Ve) variance in the process of quantifying the strength of genetic
control for a given character. The total phenotypic variance (Vp) is equal to the sum of these two
values (Vp = Vg + Ve). The strength of genetic control of a character, termed heritability (i.e.,
ability of a character to be passed on to successive generations), is then determined using the ratio
of genetic variance to the total phenotypic variance (heritability = Vg/Vp × 100). A high ratio of the
genetic to total variance (for example, 80) indicates a strong genetic control for the trait. In trees,
strong genetic control (high heritability) has been shown for stem straightness, the specific gravity
of stemwood, susceptibility to leaf rusts, and date of spring bud burst. Traits highly influenced by
the environment, such as height (strongly influenced by soil water and fertility) and diameter
(strongly influenced by stand density), are predictably under weak genetic control, and, accord-
ingly, have low heritabilities.
PLASTICITY OF THE PHENOTYPE

A given genotype may exhibit one phenotype in one environment and exhibit a different pheno-
type in another environment. This variability, termed phenotypic plasticity (Bradshaw 1965;
Schlichting 1986), is the degree to which a character of a given genotype can be modified (a non-
genetic change) by environmental conditions. Bradshaw (1965) cites the size of vegetative parts,
number of shoots, leaves, and flowers, and elongation rate of stems to be plastic traits; nonplastic
Sources of Variation 37
characters include leaf shape, leaf margin serration, and floral characteristics. Though deter-
mined for herbaceous species, there is good reason to believe these findings are consistent for
woody plants. In general, characters such as stem elongation are formed over long time periods
of meristematic activity, are more subject to environmental influences, and are therefore more
plastic. By contrast, characters such as reproductive structures are formed rapidly, or traits such
as leaf shape are impressed at an early stage of shoot development and are therefore less plastic.
Because trees are sessile and long-lived, plasticity has substantial adaptive value. An example
of plasticity with adaptive value is the rooting habit of many tree species, particularly Norway
spruce, white spruce, and balsam fir, that may develop either shallow or deep roots depending on
the soil environment (e.g., shallow in a poorly drained swamp versus deep in a sandy loam upland
soil). High plasticity for certain establishment and growth characters allows individuals of a
species to regenerate and maintain themselves on a variety of sites, as well as to endure decades or
centuries of fluctuating climate as adults.
The differences in phenotypes can be summarized using three hypothetical situations involv-
ing individuals of a given species:
Situation A Situation B Situation C

P1 = G1 + E1 P1 = G1 + E1 P1 = G1 + E1
P2 = G2 + E2 P2 = G1 + E2 P2 = G2 + E1
P3 = G3 + E3 P3 = G1 + E3 P3 = G3 + E1
The phenotypes in A illustrate the typical situation in the field, where all phenotypes have different
genotypes, but the environments are different enough to produce differences in the phenotypes.
Situations B and C illustrate experimental situations where we can either hold constant the geno-
type (in B) or test different genotypes in a given environment (in C).
Situation B illustrates plasticity, because different phenotypes of a single genotype (G1)
result from environmental differences; the environment has modified growth and development. In
nature, the degree of plasticity of a character is difficult or impossible to measure precisely because
each individual typically has a different genotype (as in A), and the extent of environmental mod-
ification can therefore only be inferred. For example, individual genotypes of an even-aged forest
in rolling terrain may occur from a dry ridgetop to a moist, fertile valley, and an increase in tree
height is observable as we progress from ridge top into the valley. If there lack major differences
among genotypes along the gradient, environment is likely to be the major factor controlling the
observed phenotypic differences in height. However, precise determination of plasticity for repre-
sentative genotypes would require experiments based on situation C.
In situation C, where the environment is constant for all individuals, phenotypic differences
result from differences among genotypes, and the amount of genetic variation can be estimated
directly from the phenotypes. In practice, the environment cannot be held constant, but the ideal
may be approached using growth chambers or relatively uniform field test plots and a replicated
experimental design. This “common garden” approach is widely applied in determining genetic dif-
ferences among selected individuals or populations (e.g., Pregitzer et al. 2013; Potts and Hunter 2021).
SOURCES OF VARIATION
As we have seen, phenotypes vary partly due to the genotype and partly due to the environment.
The major sources of genetic variation are mutation and gene recombination. Gene mutations
increase the number of alleles (the different forms of a gene) available for recombination at each
locus (position on a chromosome where a gene is located), thereby adding to the pool of genetic
variability.
Continuous variation is typical for most characters of plants. This is because most characters
are affected by many genes (termed polygenic) that simultaneously segregate and interact.
Continuous variation also arises from nongenetic causes. Relatively few traits are controlled by a
single gene with major effects, such as chlorophyll deficiency (albinism) in seedlings of species of
the pine family (Franklin 1970). Single-gene mutations may also result in differences in leaf mor-
phology of closely related taxa. For example, the rare Virginia round-leaf birch of the southern
Appalachian Mountains is similar to sweet birch except for its round leaves (Sharik et al. 1990).
Similarly, a single-gene mutation in a white ash population may have produced the single-leaf ash,
an ash with simple rather than compound adult leaves (Wagner et al. 1988). In addition to muta-
tion, additional sources of variation include:
Recombination, which spreads mutations and extracts maximum variability from them. Recombi-
nation is the major source of genetic variation of individuals that reproduce sexually. Natural
selection then acts upon the raw material of variation distributed by recombination.
Gene flow, which is the exchange of genes between different populations. Migrants, in the form of
pollen and seed, bring new genetic material to a population from other populations. The process
is called hybridization when the two populations involved are substantially different (such
as species).
Nongenetic variation, whose major sources are (i) the internal environment of the plant, and
(ii) the external or physical environment (climate, landform, soil, biotic factors). Factors of
the physical environment that modify plants and elicit genetic adaptations are considered
in Chapters 6–11.
Variation within an individual that is not directly related to factors of the external environ-
ment is far less appreciated. All cells of a tree have the same genetic information, but the plant’s
internal environment may affect its expression of genes and hence the traits we observe and mea-
sure. Striking physiological changes occur as a seedling develops into an adult tree. A series of
developmental phases is recognized, particularly the differences between the juvenile and adult
phases, apparently driven by changes that take place in apical meristems as they age. The most
universal feature is the inability of trees in the juvenile stage to flower (Chapter 4). Classic studies
by Schaffalitzky de Muckadell (1959, 1962) examined many other characters that differed between
juvenile and adult phases. In European beech and oak, as well as American beech and many oaks
in North America, the brown and withered leaves are retained over winter by trees in the juvenile
phase, but the leaves are deciduous in the adult phase. Entire portions of the lower trunk and
branches, even of very old trees, remain in the juvenile phase and exhibit the juvenile trait of leaf
retention. Reciprocal grafting experiments of juvenile and adult branches in European beech have
shown that the juvenile phase leafs out later than the adult phase. This juvenile trait may be of
adaptive value given that late spring frosts can pose a serious problem for regenerating beech.
THE EVOLUTIONARY SEQUENCE

The basic unit of evolutionary biology in sexual populations is the local interbreeding population,1
a group of potentially interbreeding genotypes that constitute the gene pool for that population.
The heritable variation of these genotypes allows natural selection to bring about evolutionary
changes. Natural selection is the differential and nonrandom reproduction of genotypes. Ulti-
mately it is this differential reproduction of genotypes, rather than just differential mortality of
1
A population is any group of individuals considered together because of a particular spatial, temporal, or other relationship
(Heslop-Harrison 1967).
Genetic Diversity of Woody Species 39
individuals, that brings about evolutionary change. Natural selection is therefore the mechanism
for evolution, the cumulative change in genetic makeup of populations over successive
generations—in simplest form, a change in gene frequencies.
Of the thousands or millions of zygotes of a species on a given site that might develop to
maturity and contribute offspring to the next generation, only a few survive. Sugar maple provides
an excellent example of the raw materials available to selection and its severity. Curtis (1959) re-
ported that 55.7% of 6 678 400 potentially viable sugar maple seeds germinated, leaving 3 673 100
seedlings per hectare. Only 198 740 seedlings remained by late summer of the same year, and only
35 380 seedlings remained alive 2 years later—less than 3% of those that germinated. He estimated
that the canopy gap resulting from the death of a mature sugar maple tree (about 6–7 m in diame-
ter) would initially support about 15 000 seedlings, which would be reduced to about 150 over the
first 3 years, and eventually only 1 or 2 trees would occupy the opening. Selection acts in every
phase of the plant’s life, but it is most effective on young seedlings eliminating in some species all
that are not well adapted to their immediate environment.
SEXUAL AND ASEXUAL SYSTEMS

The sexual breeding system characterizes most woody species and is dominated by cross-pollination.
Sexual systems allow evolution to proceed much faster because new genotypes are constantly gen-
erated. Most woody species also have some form of asexual reproduction or apomixis—any means
of reproduction, including vegetative propagation, which does not involve fertilization. In angio-
sperms, vegetative reproduction includes sprouting from the basal part of the stem (oaks, hickories,
ashes) and roots (aspens, sweetgum) following death or removal of the aboveground stem. Conifers
rarely form sprouts (except for redwood and juvenile individuals of loblolly pine), but the produc-
tion of adventitious roots from branches (termed layering) in larches, spruces, and firs is an impor-
tant mechanism of asexual reproduction in swamps or mesic uplands. Amazingly, some species
such as hawthorns may form seeds asexually. Whatever the mechanism, asexual reproduction pro-
motes uniformity and gives the plant immediate fitness in the prevailing environment.
Aspens of North America and Eurasia sucker from roots to form natural colonies termed
clones—an excellent example of asexual reproduction. A clone is the aggregate of stems produced
asexually from one sexually produced seedling. The clone is the most common growth habit of
aspens across the world (Barnes 1967). The tendency for aspens to reproduce asexually may be a
major factor in their ability to compete successfully in conifer-dominated landscapes, where fire is
common and stimulates sprouting. Aspens occur in large clones in parts of western North America,
some over 40 ha in size (Kemperman and Barnes 1976), and may live almost indefinitely by recur-
rent suckering so long as fire is not excluded from the landscape.
GENETIC DIVERSITY OF WOODY SPECIES

Genetic diversity refers to the variation or differences between organisms at the molecular level. It
reflects the presence of different alleles in a gene pool, and hence different genotypes within popu-
lations. As differing forms of genes are alleles, a gene with identical alleles across all individuals in
a population is considered monomorphic. If a gene has more than one allele in a population, it is
considered to be polymorphic. Increasing polymorphisms within a population increases genetic
diversity, which allows populations to better adapt to environmental variability. In a heteroge-
neous or unpredictable environment, small genetic differences among individuals increase the
likelihood that some individuals will exhibit genetically determined traits that allow them to sur-
vive and reproduce better than others, and thus undergo natural selection.
As described in the earlier text, the interactions of many genes with each other and the envi-
ronment typically make it difficult to infer amounts of genetic diversity using phenotype alone.
Studies of genetic variation are typically conducted at the molecular level, and have become more
predictive as modern advances in molecular biotechnology techniques provide an assortment of
very powerful methodologies to detect differences in DNA (Hartl 2020). Genetic variation was first
widely detected at the molecular level, using variations in allelic variants (allozymes) in proteins
at a single-gene locus. Determining gene frequencies for the allelic variants for many loci and for
many populations and individuals allowed the genetic diversity of species and their populations to
be estimated. Later, researchers were able to directly examine variation in DNA sequences with a
wide array of polymerase chain reaction (PCR)-based marker systems, allowing for more precise
detection of genetic variation and estimation of levels of diversity. A detailed treatment of the
measurement of genetic diversity is presented by Hartl (2020) and Cutter (2019).
Trees generally have more genetic diversity than herbaceous plants (Hamrick and Godt 1989;
Hamrick 2004; Nybom 2004). A series of early studies using allozymes (Hamrick and Godt 1989;
Hamrick et al. 1992) found that the genetic diversity of long-lived woody plants is 15% greater than
that of annual plants, 42% higher than herbaceous perennials, and 53% greater than short-lived
woody species. Geographic range is the best predictor of allozyme variation in long-lived woody
species. Genetic diversity is highest in widespread and regionally distributed species (Scots pine,
white spruce) and lowest in endemic woody species (red pine, Torrey pine, and balsam poplar).
Despite their high levels of genetic diversity within populations, trees have relatively little
diversity among populations; about 91% of the total genetic diversity of trees is within their popu-
lations (Hamrick 2004). This contrast is likely related to trees’ tall stature and low population den-
sities (creating the potential for longer dispersal distances and greater potential for gene flow;
Hamrick and Godt 1996), predominance of outcrossing (which reduces the probability of allele
loss), more chromosomes (which affects chromosomal recombination), more diversifying selec-
tion, and delayed maturity (which allows colonization of available space by a diverse group of
offspring; Austerlitz et al. 2000; Petit and Hampe 2006). These and other factors affecting gene flow
in trees have great implications for their genetic diversity. For example, a study of global wind pat-
terns and tree genetic diversity found that tree populations are more genetically similar when
linked by stronger winds, and downwind populations have higher genetic diversity (Kling and
Ackerly 2021).
There is evidence that trees are able to quickly adapt to new environmental conditions at
local scales (Petit et al. 2004) while simultaneously having low nucleotide substitution rates and
low speciation rates (Petit and Hampe 2006). The large effective population sizes of temperate
trees, which reduces the likelihood that rare alleles are lost to genetic drift, coupled with the occur-
rence of trees in heterogeneous microhabitats, are one explanation for their high levels of adaptive
genetic variation (Sork et al. 2013). Understanding the life history and ecological traits of plant
species and populations with high genetic diversity could be a powerful tool for prioritizing
conservation efforts (Chung et al. 2020).
In short, the genetic diversity present in trees is rather amazing relative to most other higher
plants. In the sections that follow, we again emphasize the inseparability of organism and environ-
ment in describing some of the patterns of this amazing diversity and the processes that
account for it.
GENECOLOGY
Genecology is the study of variation in plant species from an ecological viewpoint. Specifically,
genecology is concerned with the adaptive properties of any sexual population in relation to the
environment—species, subspecies, race, or local interbreeding population (Langlet 1971). Early
genecological studies (Turreson 1923) demonstrated that phenotypic variation among populations
that was correlated to ecological conditions often had a genetic basis and was not simply the result
of environmentally induced modification of individuals. Genecology has major practical implications
Genecology 41
for tree introduction attempts around the world, which have typically employed a trial-and-error
method too problematical and costly as a general practice. Understanding which environmental
and biotic factors elicit a genetic response, how finely populations are adapted to these factors, and
the patterns of adaptation along major environmental gradients would allow predictions of how a
given species population would perform in a new environment.
Comparative cultivation of seedling populations of forest trees, originating from environ-
mentally different sites, was first explored as early as the mid-1700s and refined throughout the
nineteenth century. Cieslar in Austria (1895, 1899) and Engler in Switzerland (1905, 1908) were
instrumental in showing that the genetic makeup of forest trees in the Alps left them adapted to
the climatic conditions of their respective environments (Langlet 1971). Cieslar published evi-
dence in 1895 of a continuous gradient of juvenile height growth for Norway spruce. This work
demonstrated genetic adaptation to growing-season conditions grading from low to high altitude
and confirmed observations of the previous century that seedlings of lowland areas proved worth-
less on mountain sites.
Turesson transplanted whole individuals from markedly different habitats into standard
conditions within a common garden (rather than cultivating seedlings from seed as Cieslar and
Engler did). The phenotypes Turesson observed in nature usually differed in growth habit (i.e.,
procumbent or erect) and in various morphological characters. If maintained in a common garden,
these differences indicated genetic differences among the populations studied. Natural popula-
tions were presumably exposed to the factors of their respective environments for generations. As
such, forces of natural selection guided the genetic differentiation of each population to adjust it to
the daily, seasonal, yearly, and even longer-term climatic and soil water fluctuations of its respec-
tive environment.
Because whole plants are preconditioned to the environment in which they grow (Rowe 1964)
and are difficult to transplant, forest scientists classically collected seeds from the desired popula-
tions, termed provenances, raised the seedlings in a common garden, and studied the differences
among the provenances. Such experiments are termed provenance or seed-source tests (e.g.,
George et al. 2017; Kashian and Barnes 2021). This type of testing reveals (i) significant genetic
differences among populations for specific characters, and (ii) the degree of genetic differentiation
among provenances under the environmental conditions of the common garden. Provenance
testing does not directly indicate which mechanisms cause the differences, although these may
be inferred.
PATTERNS OF GENECOLOGICAL DIFFERENTIATION

Genetically based ecological differentiation or divergence of populations is a recurrent feature of
plants, but whether it is discontinuous or whether it is continuous in nature has been met with
some controversy. A cline is a gradation in measurable characters which might be continuous or
discontinuous, stepped or smooth, or sloping in various ways (Huxley 1938, 1939). The term cline
does not mean or necessarily imply a genetically based gradation in a character and could refer to
any gradation of phenotypic characters observed along a natural gradient.
Genecological differentiation occurs as a multidimensional response of individuals of a
population to their environment and is unique for each population and species. In the following
text, we present the following generalizations of differentiation in forest species.
1. The pattern of differentiation is determined largely by the total natural range of a species,
the distribution pattern (continuous, discontinuous, mosaic) of a species within this range,
and the way in which the conditioning environmental factors vary. If a species that exhibits
continuous distribution over a wide range, particularly in latitude or elevation, is subjected
to more or less continuously varying climatic factors, genetic variation tends to be
continuous. By contrast, a discontinuous pattern may result if the species’ distribution or

the conditioning factors are discontinuous. The pattern of variation is often represented as a
series of contour lines whose spacing reflects the rate of change in the conditioning factors.
2. The spatial scale at which a given genecological study is conceived and conducted affects its
results (Heslop-Harrison 1964). Thus, broadscale studies are less likely to detect local
discontinuities (Chapter 22). For example, a wide-ranging investigation of a species along a
north–south gradient may expose a clinal variation pattern that may mask other clines
associated with elevation at a given latitude, or local discontinuities that may arise from
marked changes in microclimate or soil drainage.
3. Incomplete sampling and certain methods of data analysis may suggest a discontinuity even
in an extremely continuous cline (Langlet 1959).
4. A discontinuous pattern of genetically based variation is typically the exception rather than
the rule. This is because the major factors of climate that elicit genetic variation are continu-
ous. Clinal patterns have a fundamental basis in the genetic system of many forest trees
favoring high rates of recombination and outbreeding (cross-pollination) and associated
features of (i) long-lived individuals, (ii) high site stability, (iii) high and selective seedling
mortality, and (iv) physiological mechanisms in the adult to tolerate fluctuating environ-
mental conditions. By contrast, discontinuous variation is favored by inbreeding, low rates
of gene recombination, short life span, and fluctuating sites and communities, and is
therefore more typical in herbaceous than in tree species. These and other evolutionary
aspects of genecological differentiation are discussed by Petit and Hampe (2006), Hamrick
(2004, 1989), Hamrick et al. (1992), and Linhart (1989).
Genecological Categories The continuum of genetic differentiation occurring among species

populations along ecological or geographic gradients may be subdivided into classes. Species popula-
tions that differ significantly in one or more morphological and physiological characters in common
garden tests are generically called races. Races differ in the frequencies of one or more gene alleles
(Dobzhansky 1951) and may be further described as geographic or local races, and climatic, edaphic
(soil), or photoperiodic races. Races are basic elements of evolution, but if genetic differences are
especially great, the populations may be recognized as formal taxonomic categories or taxa such as
varieties, subspecies, or species. The term ecotype, introduced by Tureson (1922a,b), is also some-
times used to designate genetically different population units. He defined ecotype as arising from
the genotypic response of a population to a “definite habitat” or “particular habitat” (i.e., a local
ecological race). However, the term has been used in many different contexts, each having a different
genecological significance, and fails to provide a useful concept for a specialized local race. We there-
fore use “race” to distinguish units of genetic differentiation below the taxonomic rank of variety.
FACTORS ELICITING GENECOLOGICAL DIFFERENTIATION

Genetic differences in growth and other characters are often obvious when populations are grown
at latitudes or elevations substantially different from those of their native habitats. Factors affect-
ing the length and nature of the growing season in the native habitat, such as mean and extreme
temperatures, occurrence of early and late frosts, thermoperiod, photoperiod, and amount and
periodicity of rainfall, are major selective forces affecting survival, growth, and reproduction that
elicit genetic differences in plant populations.
Plant associates and selective pressures of insects, mammals, and birds also influence vari-
ous characters (see Chapter 12). Coevolutionary systems are evident for animals and reproductive
Genecology 43
traits of various woody species. For example, Janzen (1969) listed 31 traits of woody legumes in
Central America that eliminate or lower the destruction of seeds by bruchid beetles. Most defense
mechanisms against these predators are deterrents, such as biochemical repellents (alkaloids and
free amino acids) or increasing the number of seeds to the point of predator satiation, likely accom-
panied by a decrease in seed size.
Growth Cessation A plant’s efficient use of its growing season is closely related to the nature
and patterns of genetic differentiation. Temperate plants will be damaged or killed by early autumn
frosts if they grow too late in the season, but they may be eventually overtopped or suppressed by
competitors if they annually cease growth too soon. Species are able to anticipate seasonal fluctua-
tions by responding genetically to the more annually consistent factors of their native environment
(such as day length and heat sum) than to variable factors such as frost occurrence. Response to
a photoperiodic signal of shortening days late in the growing season in temperate regions sets in
motion a gradual and complex process of acclimation to dormancy (Chapter 7). This response of
plants to the timing of light and darkness (usually expressed as day length) is called photoperiod-
ism; it is a biological clock enabling plants to adjust their metabolism to certain seasonal fluctua-
tions. Unlike other environmental factors, day length changes everywhere (except at the equator)
in a consistent annual cycle, and plants are remarkably precise in monitoring these changes.
Photoperiod strongly influences when woody plants enter into dormancy, particularly in
species with northern ranges. These species are genetically adapted to a photoperiod that enables
them to become dormant before factors of their prevailing environment, such as freezing temper-
atures, become limiting. Early autumn frosts and cold winters are factors that significantly affect
survival and thus are highly selective for plants in northern climates or at high elevations. Hence,
photoperiod may be highly developed as a reliable mechanism in triggering the dormancy
sequence. For example, there exists a very strong relationship between growth cessation and lati-
tude for balsam poplar that reflects a close adaptation to day length (Figure 3.2). Balsam poplar
occurs in a broad range within the boreal forest of North America and especially Canada, extend-
ing over 26° of latitude from southern Ontario to northern Alaska. The relationship is very strong
because latitude, and its associated day-length regime, characterizes very well the differences in
climate that occur across this broad range. The relationship is remarkable because species that
range over mountainous terrain and encounter widely varying temperature and moisture condi-
tions within any given latitude are far less likely to exhibit a close relationship between growth
cessation and latitude.
Most genecological and provenance tests feature populations grown in day length regimes
different from their native habitats. For example, high-latitude provenances of black cotton-
wood (native to western North America) ceased height growth in June when planted at a low-
latitude site near Boston, Massachusetts (Figure 3.3; Pauley and Perry 1954; Pauley 1958).
Southern provenances moved north to the test site continued height growth until September
and October; some individuals continued to grow until their terminal shoots were killed by the
first severe frost. Generally, movement of a southern population northward into longer days
prolongs the active growth period and results in larger plant size of the southern populations
compared to the northern populations native to that site. However, if such a move is too far
north, plants become susceptible to early frosts, which may lead to injury, decreased growth, or
death. Conversely, movement of northern populations southward into shorter summer days
reduces the active growth period in comparison to native plants at or south of the test site. In
the black cottonwood example, individuals of high-latitude provenances grew only about
15–20 cm, whereas those from southern localities grew about 2 m (Pauley 1958). When trees
from the high latitude at the test site were exposed to longer days simulated by artificial light,
they grew over 1.3 m (Pauley and Perry 1954), demonstrating that growth is strongly regulated
by day length.
(a)
260
240 r = –0.977, P < 0.0001

Bud set (DOY)
220
200
180 21 June
160
r = –0.938, P < 0.0001
140
(b)
280
50% leaf yellowing (DOY)
r = –0.934, P < 0.0001

260
240
220
r = –0.818, P < 0.0001
45 50 55 60 65 70
Latitude (°N)
F I G U R E 3 . 2 Relationship between latitude of origin and date of growth cessation in place of origin for 35
balsam poplar provenances from Canada. Indicators of growth cessation include: (a) date of final bud set,
and (b) date of 50% canopy color change. Dark circles indicate experiment performed in Vancouver, British
Columbia, and open circles in Indian Head, Saskatchewan. Source: After Soolanayakanahally et al. (2013) /
John Wiley & Sons.
Although a genetically based clinal response was shown in relation to latitude for black cotton-
wood, the response is not simple and direct (compare with the balsam poplar example shown in
Figure 3.2) as evidenced by substantial variation among provenances from 44° to 48° (Figure 3.3).
These provenances included a variety of sources sampled from the Pacific Coast to western Montana
and over an elevational range from sea level to 1525 m. The difference in latitude is far less than the
variation in the length of growing season among these sources due to elevation, aspect, and microsite
conditions. Clinal genetic adaptation to growing season length was found within the narrow latitudi-
nal range of 45–47° (Figure 3.4), and elevation of the source probably explains much of the variation
not accounted for by latitude. Populations at low and high elevations at a given latitude have growing
seasons of different lengths much like populations among latitudes. Likewise, they become accord-
ingly adapted to different photoperiods, and in particular to a critical day length in autumn that reg-
ulates their entrance into dormancy. High-elevation populations necessarily cease growth earlier
than low-elevation populations due to earlier occurrence of killing frosts, and they adapt to longer
day lengths (occurring earlier in the year) than those at low elevations. Further, a high-elevation
Genecology 45
60 Black Cottonwood
Lat. of origin/date of ht.
growth cessation
Weston, Mass., 1950
55
50
Latitude
45
40
35
30
June July Aug. Sep. Oct. Nov.
Date Killing frost
F I G U R E 3 . 3 Relation between latitude and time of height growth cessation in black cottonwood.
Source: After Pauley and Perry (1954) / Arnold Arboretum of Harvard University.
Black Cottonwood
200 Av. growing season of origin/
date of ht. growth cessation
Average growing season, days
Clones from 45° – 47° N. lat.

Weston, Mass., 1950
150
100
July Aug. Sept. Oct.

Date
F I G U R E 3 . 4 Relation between length of growing season and time of height-growth cessation in black
cottonwood. Source: After Pauley and Perry (1954) / Arnold Arboretum of Harvard University.
population may have the same length of growing season as a low-elevation population that occurs
several degrees of latitude farther north, because of the compensation of latitude for elevation. Such
populations, if interchanged, would have a similar photoperiodic adaptation mechanism through
this equivalence in growing season length, and may show only negligible differences in growth rate.
The interrelationship between elevation and latitude often goes unrecognized in genecological
studies. Correlations of growth cessation or plant size and latitude of source are often confounded by
elevational differences. Wiersma (1962) modified a formula developed for Swedish conditions to
avoid this problem, using growing season (number of days ≥6 °C) to relate latitude to elevation. He
reported that a displacement of 1° of latitude north is equivalent to a displacement of 100 m upward
in altitude. Wiersma (1963) recomputed correlations of latitude of source and various characters
from published papers using this adjustment and found greatly improved relationships. Similarly,
Sharik and Barnes (1976) found that adjusting latitude for elevation substantially improved the cor-
relation between latitude of origin and height growth cessation for yellow birch and black birch
populations, as did Barnes (1977) for a provenance test of larch of European origin.
The functioning of this photoperiodic mechanism over gradients of latitude and elevation has
been demonstrated experimentally for Norway spruce, which is wide ranging in northern and
central Europe. The critical night length that stimulates bud set is about 6–7 hours in southern
populations, but only 2 hours in northern ones (Figure 3.5). A similar clinal pattern of variation is
also evident for elevation. Norway spruce populations at 700 m in Austria exhibited a critical night
length of 7 hours, and populations at 1400 m had a critical night length of 5 hours. A comparison of
populations on the west coast of Norway and those in northern Finland shows a strong growing
season length at a given latitude and elevation (four versus two dark hours) (Figure 3.5). In addition
to controlling growth cessation, photoperiod also affects the optimum growth of plants during the
growing season itself. Southern populations of both Scots and lodgepole pines grew vigorously in a
16-hour day, whereas northern populations grew optimally in an 18-hour day (Ekberg et al. 1979).
Growth Resumption While growth cessation for temperate trees is based largely on photoperiod,
the resumption of vegetative growth in the spring (flushing or bud burst) and flowering occurs
following a winter chilling requirement, and thus is controlled largely by temperature (Sarvas 1969;
Hunter and Lechowicz 1992; Chapter 7). Northern trees are genetically adapted to initiate flushing
or flowering once a certain number of “heat units” have accumulated above a threshold temper-
ature (Owens et al. 1977). This relationship has been reported for many conifers (Campbell and
Sugano 1979) and for 26 hardwood species of eastern North America (Hunter and Lechowicz 1992).
The mean temperature or heat sum2 of spring flowering for both northern and southern
populations of a given species is the same fraction of the whole year’s average heat sum for
their respective locality. That is, northern (or high-elevation) individuals of a species are
genetically adapted to flower and flush at a lower absolute heat sum than individuals of the
same species at a more southerly location where the total heat sum is greater. Therefore,
northern plants in nature may flower and flush relatively late in the spring if warming is de-
layed and slow. When their progeny are grown at warmer, lower latitudes or lower elevations,
northern plants tend to flush earlier than the native populations and vice versa when southern
populations are moved north. Notably, frost must be a significant selective force for this north–
south relationship to hold; species whose northern sources do not flush earlier (Nienstaedt 1974)
are species with more southerly ranges (black walnut, sweetgum, tulip tree, American syca-
more) or with a coastal, ocean-moderated distribution such as Sitka spruce (Burley 1966).
EXAMPLES OF GENECOLOGICAL DIFFERENTIATION

Much of the information regarding genetic differences in physiological and morphological characters
among populations comes from provenance or common garden tests established to find the most
suitable provenances for planting in one or more localities. Many of the early tests were therefore not
designed to answer genecological questions of how and why populations are adapted to their envi-
ronments. Today, many forest scientists instead sample tissues from natural populations, genotype
them with high-throughput sequencing methods (providing enormous amounts of information), and
2
A temperature or heat sum is the product of temperature above a certain base or threshold level (such as 0° or 5 °C) and
the time duration of that temperature. It may be expressed in degree-hours or degree-days.
Genecology 47
Number of
dark hours
0
9 3
1400
1000
700
altitude, m
F I G U R E 3 . 5 Clinal variation of the critical night length for bud set in Norway spruce populations of
different geographic origin in Fennoscandia and central Europe. Critical night length is based on the
number of dark hours bringing about bud set in 50% of the plants after pretreatment with continuous light.
Source: Ekberg et al. (1979) / John Wiley & Sons.
then perform analyses using a genome (an organism’s complete set of genetic instructions) sequence
for reference (Eckert et al. 2010). Compelling arguments have been made to retain common garden
tests in combination with genotyping in understanding local adaptations (de Villemereuil et al. 2016).
Provenance studies, particularly those monitored for several decades (Morgenstern and Mullin 1990),
have uncovered major adaptive responses, primarily along latitudinal and altitudinal gradients.
Eastern North American Species Common garden tests have revealed an enormous amount
of genetic variation in deciduous angiosperms. Photoperiodic races associated with latitude have
been reported for many species (Barnes 1991; Morgenstern 1996) and for a variety of charac-
ters, including survival, time of growth cessation and flushing, height and diameter growth, frost
resistance, winterkill and crown dieback, tree form, and foliage color. Similarly, there is much
evidence that clinal genecological differentiation is widespread in many eastern conifers, includ-
ing jack pine (Mátyás and Yeatman 1992; Morgenstern 1996), loblolly pine (Wells and Wake-
ley 1966; Wells et al. 1991), eastern white pine (Wright 1970; Morgenstern 1996), and black and
white spruce (Park and Fowler 1988; Morgenstern 1996). The rate of change in clinal variation
is well illustrated by a study of the genetic variation of slash pine seedlings (Figure 3.6; Squil-
lace 1966), which also revealed weakly defined or highly fluctuating gradients. Many of the 25
characters studied revealed reversals in the general cline, as illustrated in the variation pattern of
needle length (Figure 3.6). From a minimum of 16 cm in southernmost Florida, needle length of
the progenies increased to its maximum values, 19–20 cm, in south-central Florida and then pro-
gressively decreased northward.
There has been tremendous interest in the variation patterns of the resistance of loblolly
pine, a wide-ranging and commercially important species of the South, to fusiform rust (Wells and
Wakeley 1966; Wells et al. 1991). In general, sources west of the Mississippi River and in southeast-
ern Louisiana are resistant to the rust, but sources east of this area, except those from the northern
14
15
14
15
15.6
15.3 14.5
16
15.3 15.1
15.2 14.3 14.2

16.2
16 16
15.1
14.2 15
15 14.8 15.2
15
14.6 15.6
14.4 14.7 14.9 16.4
15.7 15.8
14.6 14.8
15.6
14.4 14.1 14.6 15.3 16.1
14.8 14.6 13.9 15.2 15.5
17
14
15.4 16.1
15 16 17 17.0 16.1 16.5
18
19
17.3
15.9
19.1
19
17.0 17.9
18
18.6
18.1
17.0
18.7 18.9
Legend
19
NORTHERN LIMIT OF TYPICAL SLASH PINE 18.8
10.9 STAND LOCATION AND TRAIT VALUE 20.2
20
18
20 17
18.0
19 16
18
17
16.0
16
F I G U R E 3 . 6 The pattern of variation of needle length (cm) in seedling progenies of slash pine.
Source: Squillace (1966) / with permission of Oxford University Press.
Genecology 49
Atlantic Coast, are susceptible. This pattern is probably explained by past climatic conditions. The
wetter conditions in the West and along the northern Atlantic Coast during the last glacial
maximum (18 000 years ago; Figure 3.7) were optimal for selection for fusiform rust resistance,
whereas drier conditions in Florida gave rise to susceptible populations that migrated into present
day Alabama, Georgia, and South Carolina. It is probable that the DeSoto Canyon (Figure 3.7), east
of the Mississippi River, acted to isolate eastern and western populations as selection for
resistance occurred.
Scots Pine Scots pine is the Earth’s most wide-ranging pine and has been investigated in many
studies at European (dating from 1745) and North American test sites. Genetic differences in
height growth, foliage color, stem form, rooting habit, resistance to insects, fruitfulness, and time
of bud set have been demonstrated (Langlet 1959; Troeger 1960; Wright et al. 1966, 1967). Many
of these characters exhibit a clinal pattern. Parts of a cline such as races, ecotypes, or varieties are
used as the basis for selecting seed-collection zones, particularly where major differences in tree
Wisconsin
glaclation
*
Spruce
macrofossils WET
“Lost” pines
*
*WET Pinus
taeda
yon
WET DRY
Can
Mis nyon
De Soto
Ca
siss
Pinus
lpp
taeda
i
CONTINENTAL Pinus caribaea-/-eltiottii

SHELF
F I G U R E 3 . 7 Hypothetical distribution of loblolly pine and slash pine at the Wisconsin glacial maximum,
18 000 years ago. Caribbean pine (Pinus caribaea) and slash pine (P. elliottii) may have been a single species at
that time and are shown as such. Source: After Wells et al. (1991). Reprinted by permission of the Bundesforsc-
hungsanstalt für Forst-und Holzwirtschaft, Grosshansdorf, Germany.
characteristics are important. American Christmas tree growers, for example, prefer Scots pine
varieties from Spain and southern Europe because they remain green in winter rather than those
of northern Europe that turn yellow (Wright et al. 1966).
Wide-Ranging Western North American Conifers Western North America has provided a
wealth of research examining the genetic response of species to heterogeneous environments. The
heterogeneity of physiography and climate across a wide range of latitude, longitude, and ele-
vations from Mexico to northern British Columbia and Alberta, Canada have elicited both very
localized differentiation and broad patterns of variation. Douglas-fir and lodgepole pine exhibit the
most variation because of the diversity of sites they occupy (Critchfield 1957, 1985; Wheeler and
Critchfield 1985; Rehfeldt 1988). Both species have coastal populations yet range eastward to high
elevations in the Rocky Mountains and arid interior lands. Ponderosa pine also exhibits broad,
subcontinental variability, substantial variation related to elevation in various parts of its range,
and significant local differentiation. We describe in the following text the different kinds of genetic
responses of ponderosa pine and Douglas-fir to their heterogeneous environments.
Ponderosa pine Ponderosa pine ranges in North America from central Mexico to southern British
Columbia, Canada. It is classified in the subsection Ponderosae of the genus Pinus and is subdi-
vided into two varieties: var. scopulorum in the Rocky Mountain portion of its range (with two
major races; Figure 3.8) and var. ponderosa in the western part of its range (with three races;
Figure 3.9; Conkle and Critchfield 1988).
The major races of ponderosa pine exhibit differences in morphology and biochemistry as
evidenced by range-wide provenance testing, likely related to water use and drought. The Rocky
Mountain variety was named for the compact, brushlike, bushy-tuft (scopulate) appearance of its
foliage (Figure 3.8) in contrast to the open, plumelike foliage of far-western populations (Figure 3.9).
The western variety is morphologically distinct because of its general lack of two-needle fascicles
compared to the Rocky Mountain variety (Figure 3.10). The number of needles per fascicle is influ-
enced by climate and site conditions, and there are fewer per fascicle in younger trees (Haller 1965).
Two-needle fascicles reduce water loss by having less surface area and fewer stomates, and they
require less energy to produce than three-needle fascicles. These features are of survival value in
dry, harsh Rocky Mountain conditions.
Studies of the monoterpene components of xylem resin by Smith (1977) also illustrate differ-
ences among geographic races (Figure 3.11). Beginning in the southern California race, a clock-
wise pattern of decrease is seen for α-pinene and limonene through the Pacific, North Plateau, and
Rocky Mountain races to the southwestern race. Conversely, 3-carene is negligible in southern
California, but present in significant amounts in the other races. Conkle and Critchfield (1988)
emphasize the correspondence of physiographic barriers and the distinct monoterpene races and
their sharp transition zones.
Despite clear morphological and biochemical differences among races of ponderosa pine,
genetic marker studies are yet unable to detect such differences in races at the molecular level and
will likely rely on future genome sequencing to do so. Nevertheless, physiological traits related to
drought tolerance have been demonstrated to have a direct genetic basis (though not yet specific to
ponderosa pine). Epigenetic studies of several pine species have demonstrated shifts in gene
expression of several physiological traits that control drought tolerance when seedlings of ponder-
osa pine are drought-stressed (see summary in Moran et al. 2017). Genes upregulated (activated)
during drought stress include those related to physiological traits such as water uptake or the pro-
duction of protective molecules and heat shock proteins; those downregulated are those involved
in growth, including cell division. Most research in this area thus far has been limited to pine seed-
lings rather than adults. Given that such shifts in gene expression may vary widely among trees of
different ages, and most gene expression studies do not examine difference among populations
Genecology 51
Pinus ponderosa
var. scopulorum
Transition
zone
Rocky
Mountain
Southwestern
F I G U R E 3 . 8 Eastern geographic races of ponderosa pine, excluding poorly understood populations in

central Mexico. Source: After Conkle and Critchfield (1988) / United States Department of Agriculture /
Public Domain.
(Hamanishi and Campbell 2011), several researchers have suggested that a combination of gene
expression/epigenetic studies and provenance testing is the new frontier of understanding differ-
ences in drought-tolerance mechanisms among races of coniferous species (Moran et al. 2017).
The basics of epigenetics are discussed near the end of this chapter.
Each of these lines of evidence illustrates the enormous variation that can be expected over
a diverse range of regional and local ecosystems that differ in macroclimate and microclimate, ele-
vation, and evolutionary history. Great genetic variation also occurs within these varieties and
races, and it is strongly related to length of the frost-free period along elevational gradients (Conkle
1973; Read 1980; Rehfeldt 1986a,b, 1990).
Pinus ponderosa
var. ponderosa
North
Plateau
Transition
zone
Pacific
Southern California
F I G U R E 3 . 9 Western geographic races of ponderosa pine. Source: After Conkle and Critchfield (1988) /
United States Department of Agriculture / Public Domain.
Douglas-fir The genetic differentiation of Douglas-fir has been investigated more than any other
North American tree species. Two major geographic varieties—(i) the coastal or green type, Pseu-
dotsuga menziesii var. menziesii and (ii) the interior, Rocky Mountain, or blue type, Pseudotsuga
menziesii var. glauca—span an enormous geographic range in western North America. The two
types are found on extremely variable sites in coastal and mountain areas of the Pacific Northwest
and in the Rocky Mountains from Mexico to Alberta and British Columbia, Canada. Broadscale
provenance testing of Douglas-fir has been conducted extensively in species-poor Europe; the most
extensive test included 182 native provenances with plantings across 30 countries. Overall, there
was a relatively clear separation between coastal and interior varieties in many traits, including
frost sensitivity, phenology, morphology, and growth traits associated with latitude and elevation
(Kleinschmit and Bastien 1992). Many populations of the coastal variety that perform best in Eu-
rope and the Pacific Northwest come from areas of high genetic diversity (Silen 1978; Kleinschmit
and Bastien 1992).
Genecology 53
< 1–3
60–85
< 1–2
20–50
< 5–15
F I G U R E 3.1 0 Average frequencies of two-needle fascicles on (mature/young) native ponderosa pines

(Haller 1965). Similar results were reported by Wiedman (1939) and Read (1980) for native and plantation
grown pines. Source: After Conkle and Critchfield (1988) / United States Department of Agriculture /
Public Domain.
A recent study of coastal Douglas-fir in Washington and Oregon examined growth and pheno-
logical traits and their relationships to environmental variation in seedlings of 1338 parents from
1048 locations (St. Clair et al. 2005). The objective of the study was to identify genetic variation and
races within the coastal variety of Douglas-fir, which appeared to form two races located on the east
and west sides of the Cascade Mountains in Washington as a result of variation in minimum winter
temperatures and late spring and early fall frost dates. The dates of bud set, emergence, and growth
were strongly related to elevation and cool-season temperatures, and variation in bud burst and
growth partitioning to stem diameter versus height was related to latitude and summer drought.
Seedlings from the east side of the Washington Cascades were smaller, set their buds later, and burst
their buds earlier than populations from the west side when planted in a common garden (St. Clair
et al. 2005). Douglas-fir not only exhibits high genetic variation throughout its range but displays
remarkable variation in morphological and physiological characters within physiographic regions of
both coastal (Campbell and Sorensen 1978; Campbell 1986, 1991) and Rocky Mountain races
(Rehfeldt 1989).
Pacific
40 North Plateau and Rocky Mt.
60
30 β-pinene
50
Percent
Limonene
20 40 Myrcene
Percent
α-pinene
30
10 Δ3-carene
20
0
10
Monoterpene
0
Monoterpene
Transition
zone
S. California 60 Southwestern
50
50
40
40
Percent
Percent
30
30
20
20
10 10
0 0
Monoterpene Monoterpene
F I G U R E 3 .1 1 Average amounts of the major monoterpene components in xylem resin of native ponderosa
pines (Smith 1977). Source: After Conkle and Critchfield (1988) / United States Department of Agriculture /
Public Domain.
Genecology 55
Local Genecological Differentiation As described in the earlier text, ecologically based

genetic differentiation occurs at the macro or range-wide scale for virtually all tree species and
mirrors the variation of the environment. The question of local adaptation thus becomes
relevant—the degree to which populations are adapted to local segments of geographic and
ecological gradients and fine-scale differences in soil water, elevation, aspect, and soil type.
Early evidence for herbaceous species demonstrated extensive local adaptations to soil and
microclimatic conditions (Antonovics 1971; Bradshaw 1971; Snaydon and Davies 1972, 1976).
More recent evidence suggests that trees are also able to adapt to local environmental con-
ditions relatively quickly (Petit et al. 2004). Chemical and physical properties of soils elicit
sharp discontinuities in plant distribution, and strong edaphic preferences are quite common
for woody species, but examples of intraspecific genetic adaptation to local soil type are not
well documented.
Local differentiation has been shown for populations on adjacent sites representing the
extremes of soil water conditions. For example, Wright et al. (1992) demonstrated genetic dif-
ferences between populations of jack pine growing on adjacent wet and dry sites. Those origi-
nating from the wet site had significantly more tertiary roots than those from dry sites, which
probably enhances nutrient uptake in the spring when the soil is likely to be saturated and
anaerobic. Likewise, genetic differences were reported between a few populations of northern
white-cedar from wet and dry sites (Habeck 1958; Musselman et al. 1975), but failed to be dem-
onstrated for upland and lowland populations of black spruce (Fowler and Mullin 1977; Boyle
et al. 1990).
Local races are often formed by the contrasting ecological conditions created by slope aspect.
Differentiation caused by aspect has been demonstrated for Douglas-fir seedlings originating from
north and south slopes in southern and central Oregon (Hermann and Lavender 1968; Camp-
bell 1979), and for Sitka spruce in southeastern Alaska (Campbell et al. 1989). Elevation and aspect
are often associated with local differentiation, presumably due to the heterogeneity of microcli-
mate created by these physiographic factors. For example, microevolution of the coastal race of
Douglas-fir was demonstrated by Campbell (1979) in a single 6100 ha (61 km2) watershed in central
Oregon with elevations ranging from 500 to 1600 m. Seedlings of 193 trees located throughout the
watershed were grown in a common garden. Substantial genetic differentiation was noted in 14 of
16 traits, including seed germination rate, height, bud burst, bud set, and dry weight. Variation was
mainly associated with elevation for traits related to vegetative growth. Estimates of maladaptation
showed that none of the seedlings of a subpopulation would be adapted to locations 670 m higher
or lower on the same slope.
It was once assumed that vast stands of wind-pollinated trees are genetically homogeneous,
but significant genetic heterogeneity may be identified among local populations. Studies of local
populations of lodgepole pine (Knowles and Grant 1985), trembling aspen (Mitton and Grant 1980),
ponderosa pine (Linhart et al. 1981), and Engelmann spruce and subalpine fir (Shea 1990) demon-
strate significant genetic heterogeneity over short distances, for example, within 2 ha for ponder-
osa pine (Linhart 1989). All trees occupying an area of approximately 2 ha in a Colorado ponderosa
pine stand were mapped and their genetic constitution determined for seven polymorphic loci.
Most trees fell into one of six clusters that differed from one another at one or more loci (Lin-
hart 1989), a population structure promoted by the patchy nature of disturbance and regeneration
in ponderosa pine forests. However, this fine-grained genetic heterogeneity is not necessarily
related to adaptive differences among sites, but localized ecological races are possible wherever
selection pressures are strong enough and gene exchange somewhat limited. For example, Shea
(1990) found significant differences in allele frequencies between wet and dry sites dominated by
Engelmann spruce and subalpine fir.
FACTORS AFFECTING DIFFERENTIATION: GENE FLOW AND SELECTION PRESSURE

Gene flow in plants among populations via pollen, seeds, and other propagules is a cohesive force
acting to keep populations from diverging, and thus influences the extent of differentiation. By
contrast, gene flow is impeded by isolating factors such as spatial distance between populations, or
ecological isolation (south versus north slopes, wetlands versus uplands). Selection is also impor-
tant, because populations will remain similar if subjected to similar selection forces, but will dif-
ferentiate if they are not. The tall stature of trees, among other characteristics of the life form,
creates the potential for high gene flow (Hamrick and Godt 1996), particularly over short distances,
and local differentiation is therefore unlikely. Gene flow over short distances occurs mainly by
high selection pressure.
Gene flow is restricted by several factors besides limited dispersal of pollen and seeds, includ-
ing: (i) a limited number of sexually mature trees, (ii) individual variation in flowering times, and
(iii) biological and ecological factors that control zygote viability and seedling establishment, such
as genetic incompatibility, frost, drought, shade, and herbivory. The age of seed-bearing and peri-
odicity of seed crops vary widely among species, and trees also vary in their reproductive capacity;
some are highly fruitful, some moderately so, and others are completely barren year after year.
Only a few breeding trees in a population contribute substantially to subsequent generations of the
many trees that could potentially exchange genes. For example, Schmidt (1970) found in an
89-year-old Scots pine stand that only 30% of the trees were responsible for 71% of the female stro-
bili and 64% of the male strobili.
The phenology of flowering (time of pollen release and female receptivity) is vital in pol-
lination and is closely related to air temperature and relative humidity. The phenology of trees
within a given population is more likely to be synchronized with that of trees close together
than with those progressively farther away. Although the timing of flowering generally favors
local gene exchange, gene flow from distant sources is a definite possibility. Despite its poten-
tial to be transported many miles, viable pollen may reach the distant population too early or
too late to compete effectively with local pollen. This effect becomes even more important in
Scots pine and probably other pines and conifers, given that the capacity of the pollen cham-
ber is limited (Sarvas 1962), and pollen grains have an unequal chance to fertilize the eggs.
Only two of the many pollen grains reaching the micropyle of the ovule have an opportunity
to fertilize the eggs of each ovule, one of which eventually develops into the embryo. There-
fore, pollen grains of neighboring trees have a higher probability of achieving fertilization
than those of distant trees because of the greater probability of their being first to reach the
micropyle.
The range of seed dispersal in forest trees is limited, except perhaps for very small-seeded
species such as birches, hemlocks, poplars, and willows. Bird dispersal in oaks and beeches (John-
son and Webb 1989) and pines (Vander Wall and Balda 1977) may extend dispersal to 20 km or
more (Chapter 12). By contrast, pollen may be carried great distances by wind (Lanner 1966). An-
dersson (1963) reported that pollen was blown from Germany to southern Sweden, a distance of
72 km, and that in a high pollen year in Sweden, clouds of pollen were heavy enough to be mis-
taken for forest fires. Genetic markers have provided more precise estimates of effective dispersal
distance of pollen in trees. For example, a study of Scots pine in Spain by Robledo-Arnuncio and
Gil (2005) found that over 4% of matings resulted from pollen dispersed greater than 30 km away.
A study of European ash in Scotland documented more than half of the individuals successfully
pollinated occurred from populations greater than 10 km away (Bacles et al. 2006). And finally,
about 27% of individuals of black cottonwood in the Pacific Northwest were pollinated by popula-
tions 3–10 km away (DiFazio et al. 2004).
Widespread pollen and seed dispersal might lead to the conclusion that populations over
large areas are prevented from diverging because of widespread gene exchange. Extensive gene
Ecological Considerations at the Species Level 57
flow of trees suggests that adaptation would occur at broad scales and that local adaptation is
unlikely in the absence of very strong selective forces (Boshier et al. 2015). Several recent studies
have shown explicitly with genetic markers that adaptation in trees for various traits may occur at
fine spatial scales, including those of maritime pine (Budde et al. 2014; Vizcaíno-Palomar
et al. 2014), European beech (Csilléry et al. 2014), sugar pine (Eckert et al. 2015), whitebark pine
(Lind et al. 2017), Pacific silver fir (Roschanski et al. 2016), and the model organism black cotton-
wood (Holliday et al. 2016). The combination of high local differentiation for adaptive traits in the
face of extensive gene flow is rather paradoxical (Petit and Hampe 2006).
The mechanisms of local adaptation in trees are long studied. Most pollen and seeds of trees
are dispersed close to the source, and their frequency declines rapidly with distance (Ell-
strand 1992). Individuals of most wind-pollinated species in the north temperate zone in natural
stands are pollinated and fertilized by their neighbors less than about 100 m away (Koski 1970;
Shea 1990; Adams 1992). Monitoring flow and destination of pollen from individual trees and from
different stands remains extremely difficult, and the extent of gene exchange is difficult to estimate
but is probably often underestimated (Savolainen et al. 2007). Gene flow may be sometimes
restricted to nearest neighbors, favoring inbreeding, whereas at other times gene exchange may
occur over considerable distances, thus significantly affecting the gene pool of receptor populations.
Natural selection guides the genetic makeup of populations in a given ecosystem against
the prevailing level of gene flow. Environmental selection pressures play a significant role over
distances of only 2–4 m in herbaceous species (Aston and Bradshaw 1966). A classic study by
Barber and Jackson (1957) demonstrated the effect of intense selection in glaucous and nonglau-
cous (green) phenotypes of urn tree (eucalyptus) in Tasmania. Green phenotypes are found at
low elevations and more sheltered sites, while glaucous individuals (having leaves covered with
a whitish wax effective in freeze avoidance) are more frequent in exposed and colder environ-
ments along a gradient from low to high elevation. Clines of glaucousness are correlated with
frost occurrence, with more glaucous populations occurring in the frostier localities (Bar-
ber 1955). The change from green to glaucous types occurred completely over a vertical distance
of 122–152 m (0.8–1.6 km ground distance) in the adult populations. Gene flow is likely via insect
and bird pollinators, as indicated by the production of glaucous seedlings from nonglaucous
mother trees and vice versa. Nevertheless, intense selection eliminates glaucous seedlings as
they mature in low elevation forests, and at the higher elevations green seedlings are eliminated,
building high genetic diversity over a short distance even in the face of considerable gene flow.
More generally, the annual variation in selective forces experienced by woody plants, such as
temperature, precipitation, or frosts, may have a stabilizing rather than a selective effect that
may contribute to local adaptation (Savolainen et al. 2007; Boshier et al. 2015).
ECOLOGICAL CONSIDERATIONS AT THE SPECIES LEVEL

Populations diversify in response to genetic differentiation as they change through time and radi-
ate into new areas. Mechanisms that act to separate species are those that favor reproductive isola-
tion (Table 3.1). Many mechanisms, called pre-zygotic barriers, act singly or in concert to pre-
vent gene flow among plant populations, and this process is an important route of speciation.
Geographic separation is the most obvious cause of reproductive isolation, resulting when popula-
tions are separated by physical barriers such as mountain ranges, waterways, or wide canyons, and
the gene pools thereby occur in different regions. Such isolation may result in allopatric specia-
tion, whereby each population undergoes unique development based on the demands of the envi-
ronment of their given region or differences in their specific gene pool. Plant populations may also
experience mechanisms that favor reproductive isolation even when the populations co-occur in
the same region. Sympatric speciation occurs in such instances, whereby resulting species are
able to maintain their distinctness even when individuals of each are within effective pollinating
Table 3.1 Pre-and post-zygotic barriers that contribute to reproductive isolation in plants. Except
for geographic separation, barriers are listed in the order they occur.
Timing Type of barrier Effect

Pre-zygotic barriers (fertilization and/or zygote formation prevented)
Pre-pollination Geographic Populations are physically separated into
different regions
Pre-pollination Ecological/niche Seeds, seedlings, and/or adults of different
differentiation populations in the same region have the highest
fitness in differing habitats
Pre-pollination Phenological Populations in the same region differ in phenology
(flowering times)
Pre-pollination Ethological Adults of populations in the same region have
specialized pollinators specific to one population
and not the other
Post-pollination Pollen interactions Populations in the same region may pollinate
each other, but incompatible gametes reduce or
eliminate fertilization
Post-zygotic barriers (fertilization occurs, but zygote is inviable, weak, or sterile)
Post-pollination Hybrid seed formation Hybrid zygotes are less likely to mature into seeds
Post-pollination Hybrid inviability Hybrid seeds are less likely to germinate and
or weakness establish than parental seeds or establish as
weak hybrids
Post-pollination Hybrid sterility Hybrid plants are more likely to be sterile than
parental plants
Post-pollination Hybrid offspring First-generation hybrids are normal, fertile, and
inviability and sterility vigorous, but hybrid offspring are inviable
or sterile
Source: Adapted from Baack et al. (2015).
range of one another and hybridization would be possible. Sympatric species may hybridize and
form hybrid individuals, such as in loblolly pine and shortleaf pine or yellow birch and paper
birch, but the parent species remain distinct.
Many pre-zygotic barriers that occur prior to pollination prevent gene flow among popula-
tions or species occurring in the same region (Baack et al. 2015). Ecological separation or niche
differentiation (see niche discussion in the following text) promotes reproductive isolation when
populations co-occur in the same region but occupy different habitats and become reproductively
isolated. Ecological separation is effective when trees experience reduced fitness in some habitats
compared to others at some point in their lifestyle. For example, black ash and white ash co-occur
in the eastern United States, but black ash is restricted to poorly drained soils where white ash is
unable to persist, and white ash to uplands where black ash is unable to compete. Phenological
separation reproductively isolates populations and species when phenology—especially flowering
periods—is not synchronized. Ethological separation occurs when species or populations have spe-
cialized pollinators, a barrier especially common in the tropics. Finally, gametic incompatibility is
a pre-zygotic barrier that occurs after pollination, when different populations or species may polli-
nate each other, but incompatible gametes prevent or reduce fertilization (Table 3.1; Stebbins 1971;
Baack et al. 2015).
Following various degrees of population differentiation, populations exhibiting strong ge-
netic differences in morphology and physiology are classified as formal taxa (species, subspecies,
or varieties). Stebbins (1971) has described this process as occurring in four stages. First, a single
population that exists in a homogeneous environment may eventually differentiate into races and
subspecies as individuals of the population migrate into new environments. Next, some of these
races and subspecies may become geographically isolated as differentiation and migration con-
tinues. Third, the geographically isolated races or subspecies further differentiate at a genetic level
and may eventually become reproductively isolated. Finally, changes in the environment may
allow populations once geographically isolated to co-occur in the same area (sympatric), remain-
ing distinct because they have differentiated to the point of reproductive isolation (Stebbins 1971).
Not all species have the same degree of divergence, as noted in their degree of morphological
difference and in their reproductive isolation. “Good” species, from the standpoint of reproductive
isolation, are generally sympatric.
There exist several additional isolating mechanisms, called post-zygotic barriers, that
operate even when pollination and fertilization are able to occur between co-occurring species and
hybrids are formed (Stebbins 1971). First and foremost, outcrossed zygotes that form when popula-
tions or species pollinated each other are often less likely to mature into seeds than parental
zygotes. Similarly, outcrossed zygotes may be unable to germinate or become established (hybrid
inviability) or if established may be unable to form reproductive organs (hybrid weakness). An
example of a weak hybrid is the hybrid that occurs between bigtooth and trembling aspen (Populus
× smithii) in the northeastern United States and southeastern Canada. Alternatively, hybrid steril-
ity may occur because reproductive organs develop abnormally, or chromosome incompatibility
leads to meiosis failure. Notably, the hybrids created from the outcrossing of sympatric species may
be vigorous and fertile, but their offspring (F2 generation) are largely inviable or sterile (Steb-
bins 1971; Baack et al. 2015).
In some cases, hybridization may dissolve or “swamp out” species distinctions in a flood of
intermediates. In this situation, complete reproductive isolation has not yet occurred, and the pop-
ulations may be more appropriately regarded as subspecies. This situation has been documented in
Canada, where intermingling occurs between the morphologically similar white spruce and En-
gelmann spruce, and balsam poplar and black cottonwood. Subspecies have been recognized in
both cases (Taylor 1959; Brayshaw 1965; Viereck and Foote 1970).
Biologists have long recognized that no single definition of a species is entirely applicable to
classify the enormous diversity of organisms nor to serve the various purposes desired by different
scientists. This problem of defining speciation has been considered concisely by Stebbins (1971),
Heslop-Harrison (1967), and Solbrig (1970), and in detail by Stebbins (1950 and 1970) and Grant
(1963, 1971, 1977). Ecological and genetic studies have identified the ability for populations to
diverge even when there is considerable gene exchange (Arnold 2015), and the emerging ability to
study entire plant genomes (genomics) has largely reshaped our understanding of species and spe-
ciation. Genome sequencing has provided many examples of gene exchange in nature between
good species, and additional examples should be expected as additional species are sequenced.
One modern definition of a species is a group of genotypes that are more similar to each other than
to other groups across the genome as a whole (Mallet 2006). Such a definition focuses on distinct
genetic backgrounds among species in light of the fact that some genes or genomic regions may be
similar among species (Hartl 2020).
As the field of genomics continues to develop, its methodologies will undoubtedly be critical
to understanding how speciation occurs at the molecular level. Genomics may be used to (i) iden-
tify the specific loci that control species differences or cause reproductive isolation; (ii) quantify the
number of loci subject to selection that are needed for the evolution of new species and how they
are distributed through the genome; and (iii) identify and quantify gene flow between populations
(Jiggins 2019). In general, genomic approaches utilize many types of genetic markers to compare
species, populations, and individuals, and the genetic processes that occur among them. For
example, Zheng et al. (2021) sequenced 55 trees from three hornbeam species (genus Carpinus) in
China to examine the gene flow and selection that occurred during the speciation process. One of
the species, Carpinus tibetana, was only recently identified and named, and exhibited very low
genetic diversity attributed to its extremely small population size. The other two species exhibited
evidence of gene flow between them; divergent and selected genomic regions included genes asso-
ciated with temperature, stress response, and plant development that were likely adaptations to
local environments (Zheng et al. 2021). The use of genomics to study speciation is a rapidly devel-
oping field, but the basics have been reviewed by Jiggins (2019), Trense and Tietze (2018), and
Campbell et al. (2018), among others.
NICHE
The genetic differentiation among species that has resulted from ecological interactions has
served to place each species into a different niche. The term niche expresses where, when, and
how a species is genetically adapted to persist with other species it interacts with on a site. The
niche of a species is the result of the multidimensional specialization of that species in space and
time. By definition, no two species can occupy the same niche without one being outcompeted by
the other (Chapter 13). Three components are recognized in examining how woody species differ-
entiate their niches: a spatial component (the physical site conditions to which the species is
adapted), a temporal component (whether a species dominates early or late in succession), and a
functional component—physiologically based genetic adaptations, sometimes termed natural
history or life history traits, such as number of seeds produced; dispersal time and mechanism;
growth rate; and tolerance of shade, drought, fire, and flooding. A species’ functional component
is the set of genetic adaptations that enable it to (i) occupy a geographic range and the characteristic
local sites it includes, and (ii) dominate at a characteristic time during succession on a given site.
These three niche components identify where (spatial), when (temporal), and how (functional) a
species competes and persists in regional and local landscape ecosystems. We use the term niche
as the most concise formulation of this genetic specialization.
The spatial component of the niche is illustrated by silver maple, which occupies river
floodplains, whereas the related sugar maple thrives on upland sites. Their niche differentiation
results from the different physiological adaptations of these species, which make them relatively
more competitive on these respective sites. The temporal component is illustrated by paper birch
and eastern hemlock on an upland site in the hemlock–northern hardwood forest. Paper birch is a
pioneer species that dominates a site following fire, early in the course of succession. Hemlock
seedlings establish simultaneously on the same site, but grow slowly under the birch canopy, and
then dominate the site a century or more later as the birches decline and die. In this case, the
species are niche-differentiated by the physiological adaptations that enable them to dominate the
site at different times rather than by site conditions. The functional component is illustrated by
the physiological adaptations of paper birch and hemlock. Birches colonize a burned site quickly
and their seedlings grow rapidly with leaves of high photosynthetic efficiency on sites with high
light levels. By contrast, hemlock seedlings are photosynthetically efficient at low light levels and,
using other physiological adaptations to obtain soil water and nutrients, they are able to develop
under the birches and replace them in 100–200 years.
HYBRIDIZATION
Hybridization, the crossing between individuals of populations having different adaptive gene
complexes (races, subspecies, species), is frequent in natural populations of many woody plant
groups. Many tree and shrub hybrids were discovered and reported during the twentieth century. For
example, widespread disturbances by humans have enabled the European white poplar to initiate
naturally occurring hybrids with a native poplar on three continents: the gray poplar (Populus ×
canescens) in Europe, Rouleau’s poplar (Populus × rouleauiana) in eastern North America (where the
white poplar was introduced as an ornamental tree during European colonization), and the hairy
poplar (Paulownia × tomentosa) in China. Similarly, the introduction of eastern cottonwood into gar-
dens of France and England enabled it to hybridize with the native European cottonwood to produce
the highly successful black poplar hybrid (FAO 1980). Although most natural hybrids demonstrate
hybrid weakness, two of the best examples of so-called hybrid vigor in trees are clones of the European
black poplar hybrid and Rouleau’s poplar (Little et al. 1957; Spies and Barnes 1981).
Many hybrids were formerly treated as normal divergent species and given binomial names,
for example, Populus × acuminata (hybrid between narrowleaf cottonwood and Fremont cotton-
wood (Crawford 1974; Eckenwalder 1977, 1996)). Hybrids are now considered nothospecies
(hybrid species) and may be designated by a taxonomic formula, as in the case of the hybrid bet-
ween shingle oak and northern red oak (Quercus imbricaria × Quercus rubra), or as a binomial
with the multiplication sign (×, signifying hybrid) placed directly before the epithet, for example,
Quercus × runcinata (Wagner 1983). These and most hybrids were originally identified as such
using morphological characteristics, such as leaf shape, intermediate between the parent species,
but molecular markers are more typically used in recent decades.
Hybridization is of major evolutionary significance because it acts as an evolutionary cata-
lyst (Stebbins 1969, 1970; Arnold 2015). An estimated 70% of all flowering plants originated from
past natural hybridization between different species or genera (Whitham et al. 1991; Arnold 1994).
Many hybrids of woody species are of major ecological and practical significance, with importance
to management and horticulture due to their rapid growth, good form, disease resistance, or frost
hardiness (Duffield and Snyder 1958; Nikles 1970; Wright 1976; Zobel and Talbert 1984).
Importantly, hybridization may also act as a mechanism for reversing the divergence of species
(White et al. 2007). A recent study in China used genetic markers to identify loss of genetic
diversity and an associated higher extinction risk of an endangered oak species due to introgres-
sion (An et al. 2017).
Natural hybridization is probably more common today than it was in pre-European coloni-
zation forests due to massive disturbances associated with humans that affected habitats and plant
populations in the late nineteenth century and most of the twentieth century. Such disturbances
have greatly increased the likelihood of hybridization and gene flow by creating open, disturbed
sites and thereby reducing competition and favoring establishment and survival of hybrid plants.
Natural hybrids often occur on disturbed sites and in zones of contact between species (Brayton
and Mooney 1966; Remington 1968). Disturbed areas often act as intermediate or hybrid habitats
(Anderson 1948) where neither parent is well adapted. Hybrids between the narrowleaf cotton-
wood and Fremont cottonwood in northern Utah (Whitham 1989) illustrate the occurrence of
hybrids in a contact zone between parent species. The hybrids occupy a zone of overlap between
the morphologically distinct parents that occur in upper and lower elevations, respectively, along
the Weber River. Leaf-gall-producing aphids are concentrated on leaves of hybrids in a 13 km zone
where the host’s parents interbreed, yet pure Fremont cottonwood is totally resistant to the aphid,
and aphids rarely colonize leaves of pure narrowleaf cottonwood. Susceptibility to this parasite
illustrates one kind of weakness that is often found in hybrids. Despite the boom of new hybrid
reports in the twentieth century, relatively few detailed ecological studies have compared the
establishment of hybrids to their parents or the presumed differences between the so-called hybrid
habitat and that of the parents.
Hybridization may enrich the gene pool of species by the process of introgressive hybridiza-
tion or introgression: the infiltration of genes (or small portions of a genome) from one species
into another due to hybridization and repeated backcrossing. Introgression is simply gene flow
between species and occurs in three phases: (i) initial formation of F1 hybrids, (ii) their backcross-
ing to one or other of the parental species, and (iii) natural selection of certain favorable
recombinant types (Davis and Heywood 1963; Figure 3.12). When the third of these phases is
achieved—that is, the recipient species benefits by natural selection from the transfer of genetic
Species A Species B
Initial cross:
A = 50%; B = 50%
1st Backcross:
Generations
A = 25%; B = 75%
2nd Backcross:
A = 12.5%; B = 87.5%
3rd Backcross:
A = 6.25%; B = 93.75%
4th Backcross:
A = 3.12%; B = 96.88%
F I G U R E 3 .1 2 Hypothetical example of introgression—the interbreeding of species followed by backcross-

ing of some genes from one parent into at least some members of the population of the other.
material—the process is often called adaptive introgression (Suarez-Gonzalez et al. 2018).

Hybridization between two closely related species is likely to result in higher gene flow than when
the species have well-integrated but different gene pools. Genes from one population will be incor-
porated into the gene pool of the other (regardless of taxonomic level—species, subspecies, etc.) if
they improve the well-integrated harmony of the foreign gene pool (Whitney et al. 2006), but their
frequency reduced if they tend to disrupt the harmony (Barton 2001; Shaw and Mullen 2011).
Roughly 25% of plant species exchange genes with relatives, twice the rate estimated for animals
(Mallet et al. 2016).
Oaks are notorious for hybridization (Little 1979; Miller and Lamb 1985; White et al. 2007).
The Quercus genus currently includes more than 450 species, many of which hybridize (Rush-
ton 1993). In California alone, more than 20 hybrids have been recognized, 11 of which are named
(Pavlik et al. 1991), and introgression is therefore likely among closely related oak species.
Hybridization and possible backcrossing among oaks were graphically illustrated by Wagner and
Schoen (1976) between shingle oak, having entire, nonlobed leaf blades, and associated northern
red and black oaks that have multiple leaf lobes (Figure 3.13). Some plants are possible back-
crosses to northern red oak (Figure 3.13b) or shingle oak (Figure 3.13e), and introgression is
suspected.
Woody plant populations exhibit enormous variation in morphological characters, such that
most modern studies of parent and hybrid variation employ molecular genetic techniques to
estimate the extent of hybridization and gene flow. Prior to the availability of genomic and genetic
markers, divergence caused by intense selection was often misinterpreted as introgression (Barber
and Jackson 1957). Thus, in determining the amount of gene flow between species, it is important
to establish through more than observations of morphological characters of phenotypes that (i)
hybridization and backcrossing have actually taken place, and (ii) increased variation of the parent
species occurs outside the area of hybridization and is due to hybridization rather than intense
selection along an environment gradient.
The use of chloroplast DNA (cpDNA) markers has brought about a better understanding of
the contributions of hybridization to adaptation and speciation. Chloroplasts contain a separate
(a) (d)
(b) (e)
(c) (f)
F I G U R E 3 .1 3 Silhouettes of leaves from (a) northern red oak, (b)–(e) hybrid of northern red oak and
shingle oak, and (f) shingle oak. Source: Wagner and Schoen (1976) / FAO.
genome from that of the nucleus, and the cpDNA in most cases is inherited from only the maternal
parent. Detecting similar cpDNA haplotypes—sets of polymorphisms in the chloroplasts that are
inherited together from a single parent—is useful in inferring evolutionary relationships among
species, as well as evidence of hybridization and/or introgression. For example, Saeki et al. (2011)
identified 38 cpDNA haplotypes in red maple and 7 in silver maple in eastern North America. The
greater haplotype diversity of red maple probably resulted from its greater ecological range and its
ability to be a habitat generalist compared to silver maple, which is restricted to river floodplains.
Moreover, five of the seven silver maple cpDNA haplotypes were shared with red maple in areas of
geographic overlap (the Lower Mississippi Valley and a northern area centered on Vermont and
northern Michigan), suggesting that introgression was likely (Figure 3.14). The shared haplotypes
in the Lower Mississippi Valley were likely derived from silver maple, whose preferred habitat is
centered on this region, and the shared haplotype in the post-glacial northern landscapes likely
originated from red maple (Saeki et al. 2011). A similar study of birches in eastern North America
also documented high haplotype diversity and evidence of introgression (Thomson et al. 2015).
Genetic markers have uncovered evidence of introgression in several additional tree genera,
including poplars (Lexer et al. 2005) and willows (Hardig et al. 2000), as well as oaks (Belahbib
et al. 2001).
POLYPLOIDY
Variation at the species level is not only influenced by the combination of two different chromo-
somal sets of genes, as in hybridization, but also by the number of similar sets of chromosomes
that individuals of a species possess. Polyploids are organisms with three or more sets of
chromosomes (compared to haploid organisms with one set, or diploid organisms with two). The
ploidy level of a species (triploid = 3 sets, tetraploid = 4 sets, etc.) is measured in relation to the
(a) (b)
A. rubrum A. saccharinum
Canada Canada
Group 1 Group 1
Group 4
Group 3
Group 3
Group 2
Group 2 Atlantic Atlantic
Ocean Ocean
Group 1
Group 2
Gulf of Mexico Gulf of Mexico
F I G U R E 3 . 1 4 Maps of genetic boundaries for (a) red maple and (b) silver maple in eastern North
America denote two regions of geographic overlap and likely areas of introgression. Source: From Saeki
et al. (2011) / John Wiley & Sons.
base or x chromosome number established for the genus or family, usually the lowest haploid
(gametic) number for the group. For example, the base number for the genus Pinus is 12 (x = 12),
and all pine species have 24 chromosomes (2x = 24, because they are diploid). By contrast, the
birches have a base number of 14 chromosomes and exhibit many polyploid levels among
species. These include diploids (black birch and river birch; 2x = 28 chromosomes), tetraploids
(bog birch, paper birch, and European white birch; 4x = 56), and a hexaploid (yellow
birch; 6x = 84).
Polyploids arise when related species hybridize and the chromosomes of the hybrids then
double, a process that produces new species. Recent research has also denoted this process whole
genome duplication. Polyploids are of considerable evolutionary and ecological significance in
part because they exhibit a wide geographical distribution, including alpine, arctic, and tropical
environments. In addition, they occur on a wide range of sites from mesic to more extreme sites
characterized by harsh conditions (Stebbins 1985).
Polyploids are successful largely because of their greater ability to colonize new or disturbed
habitats relative to that of their diploid ancestors (Stebbins 1985). These sites may be associated
with zones of contact and hybridization between genetically differentiated diploid populations.
The resultant hybrid polyploids often contain gene combinations for aggressive colonization that
are buffered and maintained by the polyploid condition. Many polyploids are notable for their
ability to colonize and persist on cold, harsh sites, particularly those that were formed during and
following Pleistocene glaciation.
The success of some polyploids under severe site conditions relative to related diploids is
explained in part by the uniformity-promoting mechanism of polyploidy. Polyploidy is analogous
to a sponge, absorbing mutations but rarely expressing them. By contrast, diploids more easily ex-
press mutations or new recombinations because of the low chromosome number and fewer of
each kind of chromosome. Given that polyploids may have four, six, or more chromosomes of each
kind, new gene combinations are less likely to produce a major change in the phenotype. Thus,
polyploids represent an efficient buffering system, resisting the effects of natural selection on
particular genes and promoting and preserving phenotypic uniformity (Mosquin 1966). The
narrow adaptational limits of high latitude and weedy polyploids are an adaptive feature
corresponding to the narrow and relatively uniform environments of boreal, arctic, and disturbed
or weedy habitats. The abundance of many polyploids, such as high-latitude birches and willows,
results from the widespread availability of their preferred habitats.
Polyploidy may affect how an individual interacts with the environment (Maherali et al. 2009;
Segraves and Anneberg 2016), and not all polyploids are found in the harshest environments.
A study of polyploids in paper birch (Li et al. 1996) found several traits that reduce water loss, such
as reduced photosynthesis and stomatal conductance during water stress, as well as smaller sto-
mata, higher stomatal density, and higher leaf pubescence. Notably, polyploid paper birches tend
to be found in warmer and drier ecosystems compared to diploids (Li et al. 1996). Trembling aspen,
which is either diploid or triploid, also provides an excellent example of the potential adaptive
importance of polyploidy (Kemperman and Barnes 1976; Mock et al. 2012). Triploid aspens have
been found to have higher stomatal conductance and water use efficiency, higher carbon uptake,
bigger leaves, and faster growth rates (Greer et al. 2018), and are most frequently found at lower
latitudes in warmer and drier climates (Greer et al. 2016). By contrast, aspen in the cooler and
wetter portion of its range is nearly completely diploid (Mock et al. 2012; Blonder et al. 2020). Trip-
loids have also been characterized as having less resilience to climate-induced stress (Greer
et al. 2018).
Certain tropical floras also exhibit high levels of polyploidy, some of which are of very
ancient origin. As in temperate zones during recent centuries, newly opened habitats were avail-
able during the period when angiosperms diversified, and increasing polyploidy probably facili-
tated the establishment and spread of new groups of angiosperms during the early stages of their
history (Stebbins 1970). Thus, the ability of polyploids to colonize newly opened environments is
the common denominator of their success in diverse regions of the world, whether the time of
their origin was ancient or modern. Temperate species such as American basswood and tulip tree
are examples of ancient polyploids that have outlived their ancestors. By contrast, a newly evolved
polyploid birch, an octoploid, was discovered on a disturbed lake margin in southern Michigan
(Barnes and Dancik 1985). Overall, polyploidy illustrates a significant and intimate ongoing pro-
cess of site–plant relationships throughout evolutionary time.
Polyploidy is rare in gymnosperms; only about 5% of gymnosperms exhibit polyploidy
(Khoshoo 1959; Ahuja 2005). For many decades, very few conifers were known to be poly-
ploids (Khoshoo 1959), the most notable being coastal redwood, a hexaploid with 66
chromosomes. All known coniferous polyploids reside in the cypress family (Cupressaceae)
and none in the pine family (Pinaceae). A recent screen of the ploidy level of over 96% of the
genus Juniperus found a much higher frequency of polyploidy than previously thought, with
15 Juniperus taxa being tetraploid and one being hexaploid (Farhat et al. 2019). It is likely that
polyploids are more common in Juniperus because of the overlap in their geographical ranges,
providing ample opportunity for hybridization between species. By contrast, about 50–80% of
all angiosperm species are known to be polyploids. Amazingly, about 70% of all angiosperms
are estimated to have experienced one or more episodes of polyploidy in their ancestry (Mas-
terson 1994), and 2–4% of all speciation events in angiosperms involved polyploids (Otto and
Whitton 2000). Early estimates reported that 70% of monocots were polyploids (Goldb-
latt 1980), and between 80 and 90% were reported for grasses (Stebbins 1985). Some woody
angiosperm genera have no polyploid species (Juglans, Aesculus, Cercis), whereas many others
each have species of various ploidy levels (Prunus, Salix, Betula, Alnus, Magnolia, Acer, and
many others) (Wood et al. 2009). Modern genomic techniques have shifted the focus from the
proportion of polyploid angiosperms to the number of polyploid episodes per lineage. Recent
discussions of the evolutionary aspects of polyploidy are provided by Jiao et al. (2011), Soltis
(2005), Soltis et al. (2009, 2014, 2015), De Bodt et al. (2005), Van de Peer et al. (2017), and Rice
et al. (2019).
THE FITNESS–FLEXIBILITY COMPROMISE

A species must compromise between having flexibility—the potential to change—and having
adaptedness to its present environment if it is to survive and persist over time. Flexibility that will
permit further adaptive change is favored by variability-promoting mechanisms such as cross-
pollination and a high rate of recombination (Mosquin 1966). Inbreeding, apomixis, polyploidy,
and a low rate of recombination are uniformity-promoting devices that favor fitness in a given
environment. Most woody species exhibit mixtures of uniformity- and variability-promoting
devices, shaped by selection, that serve such a compromise well. The proportion of each type of
device exhibited by a given species is closely related to the limiting ecological factors of the envi-
ronment that supports it. For example, trembling aspen is widely distributed in northern, glaci-
ated, and disturbed sites, and eastern cottonwood is also wide-ranging but primarily found at
lower latitudes in river floodplains. Both species are primarily dioecious (male and female flowers
borne on different individuals) and thus are effective cross-pollinators. A great amount of varia-
tion is generated in these species to be widely circulated through abundant seed production and
widespread wind dispersal. Aspen is highly proficient at clonal growth via vegetative propagation
by root suckers, which assures genetic uniformity; this adaptation has been widely favored by fire.
By contrast, cottonwood is found in a virtually fire-free environment and rarely produces root
suckers in nature, but its branches and young shoots root easily in soil. This trait is selectively
advantageous and of immediate fitness in river bottoms subject to flooding, whereby branches are
broken off by ice and debris and deposited in new soil. Both species have strongly developed
mechanisms, closely linked to their respective environments, that provide both fitness and
flexibility.
Pines are typically cross-pollinating species, but maintain a certain amount of self-pollinating
ability in their breeding systems. Fire plays a major role in establishment of pines around the world
such that the colonizing ability of one or a few survivors following fire, if necessary, by self-
fertilization, is of major significance. Post-fire colonization is likely to be accompanied at first by
an increase in the degree of inbreeding, but outbreeding is restored as stand density increases
(Bannister 1965). Fire-imposed selection pressure has been instrumental in promoting a very high
degree of self-fertility in red pine (Fowler 1965a), and unlike most other pines, seedlings resulting
from self-pollination are as vigorous as those from cross-pollination (Fowler 1965b). A high degree
of self-fertility has been achieved at the cost of variability, however, because red pine is one of the
most uniform of all woody-plant species and has a relatively restricted range compared to its asso-
ciates, jack pine and eastern white pine.
EPIGENETICS
A developing area of research in understanding phenotypic plasticity is epigenetics, which is the
study of heritable changes in gene expression and function unexplainable by changes in DNA
sequences (Richards 2006; Bird 2007). These changes occur due to a set of molecular processes that
alter gene expression by activating, deactivating, or reducing the activity of particular genes. In
turn, these changes in gene expression can produce phenotypic variation in plants that is either
perpetuated during development and/or is heritable across generations. Because epigenetic
changes may affect plant phenotype and ultimately fitness, be inherited across multiple genera-
tions, and vary across populations and individuals, they are likely to contribute to the ability of
plants to colonize, adapt, or evolve in variable environments (Amaral et al. 2020). For example,
epigenetic phenomena have been shown to be responsible for a “memory” analog in plants to
stressors such as pathogens, herbivory, or drought. Upon experiencing a stressful environmental
event, epigenetic changes allow plants to store information about the stress such that the individual
may respond differently if the stress is repeated (Lämke and Bäurle 2017). This epigenetically
c ontrolled memory may last from several days to years, and may even be passed on to offspring,
helping trees to quickly respond to recurring biotic or abiotic stressors (Hilker and Schmül-
ling 2019). Examples of studies that document epigenetic responses in trees to environmental
stressors include those examining salt stress (Liu et al. 2019), heavy metal exposure (Cicatelli
et al. 2013), drought (Sow et al. 2021), temperature extremes (Dewan et al. 2018), and pathogenic
infection (Sollars and Buggs 2018).
Perhaps the most well-documented example of epigenetic memory is found in Norway
spruce, a species native to areas of mild summers and cold winters in Europe. The temperatures
that occur during embryo development and seed maturation have been shown to create important
and long-lasting changes in the phenotype of seedlings that develop from those seeds, such as
changes in the timing of bud set or bud break (Johnsen et al. 2005; Besnard et al. 2008; Yakovlev
et al. 2010; Carneros et al. 2017). The ability of a seedling to “remember” the environmental con-
ditions present when seeds were produced has important implications for their ability to rapidly
adapt to the environment at hand, which has evolutionary significance if the environment is novel
or changing. A study by Kvaalen and Johnsen (2008) showed that this epigenetic memory system
in Norway spruce is able to produce differences in bud set or bud break resembling those seen
among populations separated by as much as six degrees of latitude. In addition to evolutionary
consequences, this epigenetic phenomenon will likely be important for tree breeding, conservation
of genetic resources, and forest management (Amaral et al. 2020).
Much of this chapter has focused on the importance of genetic variation in species, for which
there are decades of research with regard to forest ecosystems. As it is a rapidly emerging field, our
focus on species’ epigenetic diversity and its potential for phenotypic variation and adaptation is
purposely reserved. However, knowledge of epigenetics will likely revolutionize our understanding
of individual tree variation within the next decade. Detailed reviews of epigenetics in forest trees are
provided by Amaral et al. (2020), Carbó et al. (2019), Sow et al. (2018), Pascual et al. (2014), and
Yakovlev et al. (2012). Implications of epigenetics for ecology in general are described by Herrel
et al. (2020), Rey et al. (2020), Kilvitis et al. (2014), Richards et al. (2010), and Bossdorf et al. (2008).
SUGGESTED
R E A D I N G S
Adams, W.T., Strauss, S.H., Copes, D.L., and Griffin, Saeki, I., Dick, C.W., Barnes, B.V., and Murakami, N.
A.R. (ed.) (1992). Population Genetics of Forest Trees. (2011). Comparative phylogeography of red maple
Boston, MA: Kluwer 420 pp. (Acer rubrum L.) and silver maple (Acer saccharinum
Hamrick, J.L. and Godt, M.J.W. (1996). Effects of life L.): impacts of habitat specialization, hybridization
history traits on genetic diversity in plant species. and glacial history. J. Biogeogr. 38: 992–1005.
Philos. Trans. R. Soc. B: Biol. Sci. 351: 1291–1298. Stebbins, G.L. (1985). Polyploidy, hybridization, and inva-
Morgenstsern, E.K. (1996). Geographic Variation in sion of new habitats. Ann. Mo. Bot. Gard. 72: 824–832.
Forest Trees. Vancouver, BC: Univ. British Columbia White, T.L., Adams, W.T., and Neale, D.B. (2007). Forest
Press 209 pp. Genetics. Cambridge, MA: CABI Publishing 708 pp.
Petit, R. and Hampe, A. (2006). Some evolutionary con- Wood, T.E., Takebayashi, N., Barker, M.S. et al. (2009).
sequences of being a tree. Annu. Rev. Ecol. Evol. Syst. The frequency of polyploid speciation in vascular
37: 187–214. plants. Proc. Natl. Acad. Sci. 106: 13875–13879.
Regeneration Ecology CHAPTER 4
I n Chapters 4 and 5, we consider the life history of forest trees within the context of landscape
ecosystems. In this chapter, we examine woody plant regeneration as a critical part of the forest
tree’s life cycle, emphasizing the environmental factors influencing this process from reproduction
to establishment. In Chapter 5, we discuss aspects of the structure and function of shoots, crowns,
and roots as related to site conditions and growth of forest trees. As elsewhere, we do not attempt a
full treatment of tree physiology; excellent treatments of the physiology of woody plants are avail-
able in the books of Kozlowski (1971a,b), Zimmermann and Brown (1971), Fitter and Hay (1987),
Raghavendra (1991), Kozlowski and Pallardy (1997a,b), and Pallardy (2008). The life history of
plants has been examined from the perspective of plant population biology by Grime (1979, 1988)
and Silvertown and Lovett-Doust (1993).
The growth and form of a plant and its associated populations are affected by the physiological
processes that occur in cells and organs. In turn, physical factors of the ecosystems in which plants
reproduce, establish, and grow direct these physiological processes. Individuals of a population are
recruited and established from a “bank” of seeds stored on or in the forest floor (Phases I and II;
Figure 4.1). Once established, plants grow in height and mass, which requires space, light, nutrients,
and moisture that may be insufficient to allow vigorous growth of all individuals (Phase III). During
Phase III, some plants die (unbranched stems, T), others thrive (shown by branched systems), and
others may persist for many years as part of the ground cover (“stored” as seedlings or sprouts) until
their recruitment proceeds into the understory and overstory. The individuals are eventually restrained
by limited abiotic resources and biotic limits (such as herbivory) of the ecosystem, indicated by the
vertical bars on either side of the population (Figure 4.1). The environment of the understory and
forest floor changes as the plant canopy develops (shown by the dashed arrow) that, in turn, affects the
recruitment of new individuals. Seeds are dispersed to the forest floor in Phase IV, where reproduction
occurs. In this chapter, we emphasize regeneration as a complex of ecosystem processes involving
sexual and vegetative reproduction, dispersal, and establishment in relation to environmental factors.
REGENERATION
The process of regeneration, including seed production and maturation prior to dispersal, allows
plants to maintain and/or expand their populations over time. In sexual reproduction, seed
production is followed by dispersal of fruits and seeds, germination, and finally the
establishment of seedlings on the forest floor, as follows:
Reproduction Fruit and seed Active or dormant Seed Plant

disspersal seed bank germination establishment
Regeneration may also occur via vegetative reproduction, whereby stems of existing plants
develop to maintain and expand the forest community.
69
70 C h a p t e r 4 Regeneration Ecology
IV
III
II
I
F I G U R E 4 . 1 The plant life cycle with trees shown as a series of repeating modular units of shoots.
Phases shown:
I The bank of seeds on and in the forest floor.
II The establishment and recruitment of seedlings.
III Growth in height, mass, and number of modular units; vertical bars represent environmental constraints
on growth; dashed line indicates the influence of the overstory on establishment and recruitment.
IV Seed production and dispersal.
Source: After Harper (1977) / Elsevier.
Whether by sexual or asexual reproduction, regeneration is an ecological process ensuring

that successive generations of plants develop as part of a given landscape ecosystem. The established
plants that do not die grow upward into successive forest strata (e.g., from the ground cover layer
to the understory) via the process of recruitment. Finally, even fewer individuals in the under-
story are recruited into the overstory or canopy layer of the forest. At each step in the regeneration
process, physical site factors such as light, temperature, moisture, nutrients, wind, and disturbance
regimes are critical in directing the outcome, as are biotic factors such as herbivory, disease, com-
petition, and mutualisms with plants and animals.
The functional characteristics and site factors of many widespread models of regeneration
in forest ecosystems illustrate the interconnectedness of environment and biota and under which
conditions each “strategy” is successful (Table 4.1). Episodic regeneration and recruitment of a
new generation as a single cohort are common with the strategies of vegetative reproduction, fire-
induced cone opening, or wind-dispersed seeds, but continuous or sporadic regeneration and
recruitment over a period of years are more common with persistent juveniles (Table 4.1). Seeds
in the active seed bank have no dormancy requirement (such as aspens, cottonwoods, and many
pines) and germinate readily once seedbed conditions are favorable. In temperate forests, seeds of
many trees and shrubs lie dormant over winter and germinate the following spring.
Sexual reproduction is the basic mode by which plants maintain their populations and per-
sist in space and time. Sexual reproduction allows plants to adapt to changing environments
because the process forms a zygote that is genetically different from either parent and from other
offspring. By contrast, vegetative or asexual reproduction (i.e., clone formation) in plants creates
Sexual Reproduction 71
Table 4.1 Regenerative strategies of widespread occurrence in woody plants.
Conditions under which strategy

Regeneration method Functional characteristics
appears to enjoy a selective advantage
Vegetative: clone V New vegetative shoots remaining Productive or unproductive sites
maintenance attached to parent plant until well subject to low or high intensities of
or expansion established disturbance
Active seed bank Ba Viable seeds that have no dormancy Sites with favorable weather or
requirement; reside in seed bank less seasonally predictable disturbance
than a year by climatic or biotic factors
Dormant and Bd Viable but dormant seeds present Sites subjected to temporally
persistent seed bank throughout the year; some persisting unpredictable disturbance
more than a year
Fire-induced open- F Heat of fire opens cones where seeds Sites prone to relatively frequent,
ing of cones are stored; seeds germinate immediately intense fires
following favorable moisture and site
conditions
Widely wind- W Propagules numerous and exceedingly Sites subjected to spatially
dispersed seeds buoyant in air; widely dispersed and unpredictable disturbance
often of limited viability
Locally dis- L Propagules few and heavy; dispersed by Sites predictable in vicinity of
persed seeds gravity and animals; seed buried parent plant
Persistent juveniles Bj Offspring derived from an independent Sites subjected to low intensities of
propagule; seedling capable of disturbance
long-term persistence in a juvenile state
Source: Modified from Grime (1988).
genetically identical stems or ramets from a sexually produced parent plant (the genet) to form a
spontaneous clone. Almost all woody plants are capable of some form of cloning. The multi-
stemmed clone best illustrates the concept of an individual tree as a population of modular units,
the shoots, where each shoot becomes a ramet of the original plant (Figure 4.1). By fragmenting its
genotype, the plant gains growing space, water, and nutrients and eventually increases its capacity
for sexual reproduction. Vigorous asexual reproduction by a plant is not known to diminish its
sexual ability to produce flowers and seeds.
SEXUAL REPRODUCTION
The sexual reproduction and regeneration process is rather complex and involves several key
factors (Figure 4.2). This process includes not only the production of fruits and seeds from flowers
but also seed dispersal, formation or addition to a seed bank, germination and establishment,
recruitment, and finally sexual maturation and flower production to begin the cycle anew. We
begin the following text with maturation as the initiating factor.
MATURATION AND THE ABILITY TO FLOWER

Forest trees progress from a juvenile phase, characterized by no flowering,1 to the adult phase when
flowering occurs. Maturation refers to the gradual set of complex physiological changes in woody
plant meristems that occur with increasing chronological age (Wareing 1987). Many characteristics
1
Angiosperms produce flowers, whereas conifers do not bear flowers but produce cones (strobili) that bear naked ovules.
For simplicity, we will use the term flowering to denote the reproductive process of both groups.
Seed reproduction cycle
Vegetative buds
– Climate
Sexually mature
– Plant condition
woody plant
– Resource availability
Bud loss
Reproductive buds Flowering pollination
– Climate
– Predation Fruit loss
– Climate
Fertilization
Established plant – Predation
Maximum – Resource
fruit size availability
– Resource availability
– Predation Mature seed
– Climate
Fruit loss
– Competition
– Climate
– Predation
Seeding bank
Seed dispersal
Germination Primary dispersal
Timing Secondary dispersal

– Climate Seed bank Timing
– Substrate – Climate
– Species Annual seed rain – Species
‘Stored’ seeds – Dispersal agent
– Seed longevity Distance
– Predation – Climate
– Dormancy – Species
– Stand structure
F I G U R E 4 . 2 The sexual regeneration process of woody plants. Source: Zasada et al. (1992) / Cambridge
University Press.
differ between the juvenile and the adult phases, including ability to flower, leaf morphology and
retention, disease resistance, rooting ability, and growth rate (Bonga and Aderkas 1993; Green-
wood and Hutchison 1993). We emphasize here that maturation is not the same as tree age; seed-
lings and young trees only have meristems in the juvenile phase, but older trees do not necessarily
contain all meristems in the adult phase. In older trees, meristems in the juvenile phase occur at
branch positions close to the trunk and near the base of the crown, apparently where meristems
remain dormant for long periods (Bonga and Aderkas 1993). The adult phase appears first at the
top of the crown as a tree ages and increases in size, apparently where the highest number of divi-
sions have occurred in the meristems. Overwinter retention of dead leaves in oaks and beeches is
indicative of the juvenile phase in a tree’s crown.
The duration of the juvenile period varies markedly among species, from 1 to over 40 years
(Owens 1991). Some conifer species may remain in the juvenile stage for 45 years or even their
entire lives, to the point that species having very different juvenile versus adult morphologies
have been misclassified into the incorrect genus (Pallardy 2008). Flowering appears to occur
sooner in fast-growing, shade-intolerant species such as paper birch and Virginia pine compared
to slow-growing, shade-tolerant trees such as American beech and eastern hemlock. For example,
the juvenile period is estimated to last 5–10 years in Scots pine, whereas in European beech it is
30–40 years (Wareing 1959). The length of the juvenile phase may also be greatly influenced by
the site conditions at different geographic locations (Kozlowski and Pallardy 1997b; Ross
et al. 1983). Overall, plants with a reduced juvenile phase are short-lived, and those with a long
juvenile phase are long-lived with a long reproductive life (Harper and White 1974; Bender
et al. 2000).
There is inconsistent evidence that trade-offs exist between reproduction and growth (see
discussion in the following text; Obeso 2002; Knops et al. 2007), but onset of the adult phase is more
closely related to tree size than to age for many species (although height and diameter of young
plants are often correlated). Small, suppressed trees in the forest understory may never flower, even
at ages of 50–100 or more years. For many species, the attainment of a certain minimum size, rather
than age, is the critical factor in attaining the adult or flowering phase. For example, European larch
normally remains in the juvenile phase for 10–15 years, but was “forced” to flower in just 4 years
when seedlings were quickly grown to the size threshold in a greenhouse (Wareing and
Robinson 1963).
We emphasize here that flowering itself is not directly triggered by the attainment of a
critical size, but through the expression of flowering-time genes that accumulate with growth.
For example, a critical level of FT2 expression is required to initiate flowering (ending the
juvenile phase) in eastern cottonwood (Hsu et al. 2006). Moreover, trees in natural stands vary
widely in their genetic disposition to flower, particularly in number of reproductive buds pro-
duced and the ratio of female to male buds. Even some adult trees situated favorably in the
overstory canopy do not flower or rarely flower, whereas adjacent trees are highly fruitful.
Although trees attaining the adult phase are able to flower annually or seasonally thereafter,
they may not do so every year due to a variety of environmental and genetic conditions
(Pallardy 2008).
Variation in flowering among individuals of the same species may also be explained by
sexual differentiation in some species. Species of several genera (Acer, Ailanthus, Diospyros,
Fraxinus, Gymnocladus, Maclura, Populus, Sassafras, etc.) bear unisexual male and female
flowers on different individuals (dioecious condition). The dioecious trait ensures outcrossing
between genetically different individuals and precludes self-pollination (selfing). Selfing and
inbreeding generally lead to growth depression in trees (red pine is a notable exception; see
Chapter 3). Angiosperms rarely produce viable self-pollinated seeds. Conifers (especially pines)
are more likely to produce viable self-pollinated seeds, although the resulting seedlings are often
outcompeted and eliminated at an early stage in natural populations. Some tree species, includ-
ing most conifers, bear both male and female unisexual flowers on the same tree (monoecious
condition). Notably, some individual monoecious trees may be predominantly male, whereas
others are predominantly female.
Increasing Seed Production Assuming there is a trade-off between reproduction and growth,
a long juvenile phase is desirable for growth, but undesirable for fruit or seed production. Reducing
the length of the juvenile phase is most effective when seedlings are grown so that they attain a
large size as rapidly as possible. This increased growth rate may be accomplished with appropriate
light conditions, by applying fertilizers, or using other cultural methods. Gibberellic acid applica-
tion and other cultural treatments cause early flowering in most conifer species (Ross et al. 1983),
but the use of asexually reproduced clones rather than sexually reproduced seedlings is often used
for tree crops in high-yield plantations (Ahuja and Libby 1993). Trees of many species in the adult
phase may be stimulated to increase flower and seed production by a variety of cultural methods
(Matthews 1963; Kozlowski 1971a).
REPRODUCTIVE CYCLES
The reproductive cycle of trees and shrubs is closely adapted to and influenced by the physical site
factors where they grow. For example, many woody species that grow on river floodplains and
other wetland sites (most willows, cottonwoods, elms, silver maple, river birch, red maple, etc.)
flower in the early spring and disseminate seeds four to six weeks thereafter. The seeds are dis-
persed onto moist seedbeds (e.g., recently flooded river sites and wetlands) where they germinate
readily (active seed bank, denoted by Ba, and widely dispersed seeds, denoted by W, in Table 4.1).
For these species, only a brief period of a few weeks is needed to proceed from pollination to seed
dispersal, and the small seeds develop quickly. Millions of seeds are produced, some of which inev-
itably find a favorable seedbed for establishment so that a large food reserve in the seed is unnec-
essary. By contrast, most other trees in the north temperate zone have fruits and seeds that require
two to four months to develop throughout the growing season and are disseminated in the fall or
winter. In these cases, the trees produce medium-size and large seeds that lie dormant over the
winter and germinate the following spring (dormant and persistent seed bank, denoted by Bd, in
Table 4.1). The seedlings of such species (such as upland oaks and hickories) are soon subjected to
drying soils and soil-water stress, requiring that a root system develops to cope with these droughty
conditions. The large amount of stored food in the seeds of oaks and hickories provides the
resources for this early root development (see discussion later in this chapter). A description of the
reproductive cycle of many North American forest trees is available from the United States
Department of Agriculture (USDA, 1990).
Virtually all tree species initiate reproductive buds or primordia (earliest bud stage) during
the growing season of the year before anthesis, or the opening of the flowers. Flower buds are
often visible along the current year’s shoots of conifers and in the axils of leaves of hardwoods in
the fall of the year prior to flowering. The reproductive cycle of Douglas-fir includes the initiation
of lateral bud primordia in April along the vegetative shoot that develops inside the terminal veg-
etative bud (denoted by letter A in Figure 4.3). Some or many of these primordia may become
pollen or seed cone buds depending on the favorability of the internal environment. When the
vegetative bud bursts, needles flush (denoted by letter B in Figure 4.3), the shoot elongates (de-
noted by letter C in Figure 4.3), and the lateral buds become visible (by late July or August) along
the young shoot (denoted by letters D and E in Figure 4.3). The new lateral buds will enlarge and
flush the following spring to produce new vegetative shoots or male or female cones. Lateral buds
at the base of the young shoot tend to become pollen cones, whereas buds toward the tip of the
shoot become either seed cones or vegetative shoots; lateral bud fate is obvious by October.
Lateral bud primordia may follow any one of multiple development pathways. Some bud
primordia abort early, degenerate, and leave no trace. Others, called latent buds, form bud scales
and then cease to develop. If the terminal bud is removed, for example, by herbivory, the latent
buds may develop into vegetative buds. Latent buds play an important role in oaks, for example,
whose foliage is killed by spring frost or defoliated by insects, because they are able to develop an
entirely new set of leaves. Lateral bud primordia not aborted or developed into latent buds develop
into pollen cone buds, seed cone buds, or vegetative buds. The fate of lateral bud primordia is
largely determined by the internal nutrition and hormonal relations in the shoot and tree. The
same number of primordia may be initiated in two consecutive years, but the ratio of vegetative to
reproductive buds may differ significantly. Thus, the inter-annual variation in seed cone produc-
tion in Douglas-fir, and probably many other forest trees, results from the proportion of primordia
that develop into cones, not the number of primordia initiated.
In Douglas-fir and most conifers, pollination and fertilization take place in the same year;
cones mature during the summer, and seeds are released in the fall. This cycle is similar in many
hardwoods. By contrast, fertilization occurs 12 months after pollination and the reproductive cycle
is a full year longer in the pines and the red oak group.
NOV • FEB
T MA
RC
OC F H
AP
PT Meiosis and RI
SE
G
pollen development L
E F
G
M
H
J
AU
AY
E
Leaf, bract and G
microsporophyll Cone buds burst “flowering”
I
Pollination I
Y
initiation
JU
H
JUL
Seed cones Fertilization

D
NE
Pollen engulfed
Lateral bud primordia become pendant
Pollen growth
become determined
JUNE
JULY
J
Shoot elongation Seed cones enlarge
K
Lateral bud primordia enlarge
Bud scale initiation
B
B
C
K
MAY
Vegetative bud burst flushing
AUG
Embryo and seed development
D
A L
Cones mature
Onset of vegetative
and open
RIL
bud growth
SEP
L
AP
Lateral bud initiation
T
A
Seed shed
F I G U R E 4 . 3 Reproductive cycle of Douglas-fir. The entire cycle extends over 17 months. Lateral buds are
initiated in April and differentiate into vegetative, pollen, or seed cone buds during the ensuing 10 weeks.
Pollination of the seed cones occurs the following April and the mature seeds are shed in September of the
second year. The various stages are identified by letters A–L and are briefly described. The approximate
length of each stage is shown by the arrows. Source: Allen and Owens (1972) / Canadian Forest Service.
POLLINATION
The timing of pollen release and female receptivity in deciduous species is closely related to
whether pollination occurs by wind or animals. Wind-pollinated species such as aspens, birches,
elms, and red maple flower in the early spring before leaf flush. Insect-pollinated species, such as
basswood, tulip tree, and black cherry, typically flower during leaf flush. Conifers are exclusively
wind-pollinated, which is promoted by the overproduction of pollen and the positioning of the
female cones at the ends of shoots in the upper third of the crown. Wind-pollinated trees produce
far more pollen than insect-pollinated species; a single anther of wind-pollinated birch may pro-
duce as many as 10 000 pollen grains compared to an insect-pollinated maple, which produces an
order of magnitude fewer pollen grains (Faegri and Iverson 1989). Wind pollination is much more
common in temperate and boreal forests than in tropical forests and in early-successional forests
compared to late-successional forests (Pallardy 2008).
Weather conditions, especially temperature, affect the shedding of pollen and the receptivity
of female flowers. Although anthesis of male and female is synchronized, it is not precisely
synchronous on the same tree, thus reducing the likelihood of self-pollination. Flowering is rapid
in wind-pollinated species. Most of the pollen for individual stands of Scots pine in Finland is shed
within three to seven days, and over an even shorter period for some individual trees (Sarvas 1962).
Pollen discharge was highly correlated with high temperature and low humidity, and the day of
maximum shed usually coincided with the warmest day of the flowering season. Moreover, pollen
discharge was delayed in unfavorable springs until the occurrence of one or several favorable days
adequate for pollen shed and spreading. In 14 years of monitoring pollination of Scots pine and
European white birch, not once was a major part of the pollen crop destroyed by unfavorable
weather. These data highlight the distinct advantage of rapidity for wind pollination (Sarvas 1962).
By contrast, severe weather conditions may cause failure of fertilization (as opposed to pollination)
and the process of ovule development. For example, regeneration failure of European linden in
northwestern England is reported due to temperatures too low to permit fertilization (Pigott and
Huntley 1981).
PERIODICITY OF SEED CROPS

Abundant flowering and seed production occur irregularly and erratically from year to year in
natural populations of forest trees. This cyclic nature of seed production, highlighted by a year of
exceptionally high production known as a mast, greatly promotes the establishment of tree seed-
lings and may be one of the most important life history traits in woody plants. Periodicity of fruiting
and seed production is a pattern of repeated cycles of mast years followed by one or more years of
lesser production. This pattern is particularly pronounced in beeches, oaks, and chestnuts, and is
thought to be an antipredator adaptation. Many early-successional species—typically fast-growing
and shade-intolerant species such as aspens, birches, cherries, cottonwoods, junipers, and many
pines—exhibit mast years more frequently with fewer years between them than late-successional,
slower-growing, more shade-tolerant species such as beeches, firs, sugar maple, and northern red
and white oaks. In species prone to seed predation, masting was initially hypothesized to satiate
seed predators because more seeds are produced than can be consumed (Silvertown 1980; Nilsson
and Wästljung 1987; Kelly and Sullivan 1997). Mast flowering may have also evolved because of
wind pollination, because many wind-pollinated species are pollen-limited due to an economy of
scale (i.e., wind pollination is far more efficient when pollen density is high; Norton and Kelly 1988;
Smith 1990). Both predator satiation and wind pollination may lead to masting, as in a study of
yellow birch in New Hampshire (Kelly et al. 2001). Finally, some authors have suggested that mast-
ing is advantageous because it enhances seed dispersal during mast years (Crone and Rapp 2014;
Pearse et al. 2016).
Trees tend to exhibit a regular cycle of seed production when studied over a long period of
time, although some baseline number of seeds may be produced each year by the dominant trees.
For example, Scots pine in central Finland produced 100 seeds m−2 every 5 years, and 150 seeds m−2
at 8-year intervals over a 50-year period (Koski and Tallqvist 1978). Similarly, abundant cone crops
of western white pine occur about every three to four years, each followed by a marked drop in
cone production the following year (Figure 4.4; Rehfeldt et al. 1971; Eis 1976). Notably, individual
trees tend to exhibit similar periodicity even when there is considerable variation in inherent pro-
ductive ability among trees (Figure 4.4). Good cone and seed crops typically occur at three- to
seven-year intervals in pines, but they may occur nearly every year in some species (such as
Virginia, Monterey, jack, and lodgepole pines). A three-year cycle of good cone crops has been re-
ported for several northwestern fir species (Franklin 1968), and a five-year cycle for Douglas-fir.
The periodicity of flowering and the minimum seed-bearing age for most North American trees
and shrubs are summarized by the US Department of Agriculture (1974, 1990).
Factors that influence the initiation of flower primordia in turn affect seed production. Very
often primordia initiation has been related to climate variables—particularly temperature and
moisture—but their influence varies depending when they occur during the growing season. Wet
and cool growing seasons prior to the initiation of flower primordia have been correlated with high
seed production in temperate forests for spruces (Buechling et al. 2016), oaks (Fearer et al. 2008),
240
MATURE CONES
180
120
Tree 19
Tree 58
60 Tree 22
Tree 17
0
1950 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67
F I G U R E 4 . 4 Periodicity of cone production of four western white pines in northern Idaho over 18 years.
Source: After Rehfeldt et al. (1971) / with permission of Oxford University Press.
beeches (Piovesan and Adams 2001), and Douglas-fir (Lowry 1966). At the same time, warm
summer temperatures during the early phase of floral primordia development have also been
related to high seed production in these same species. The relationship is complex because two,
and sometimes three, reproductive cycles are proceeding simultaneously (in pines and the red oak
group), making it difficult to determine which portion of the cycle is being affected by temperature
and moisture conditions. In addition to climate, temporal changes in resource uptake and utiliza-
tion (particularly nitrogen) have been linked to mast events (Miyazaki 2013; Han and Kabeya 2017),
but this relationship remains relatively unexplored. The maturing seed crop is a large sink for pho-
tosynthate and associated plant nutrients, and thus strongly influences the number of new flower
buds initiated and the development of those already formed.
In addition to flower primordia initiation, seed production is also governed by factors that
may preclude fertilization and ovule development following the initiation of flower primordia or
cause the loss of maturing flower buds, fruits, or cones (Figure 4.2; Sork et al. 1993). Newly formed
reproductive buds and immature fruits and cones may be killed or injured by spring frosts. Wind
pollination and dispersal are negatively affected by high precipitation at anthesis (Allen and
Platt 1990; Pearse et al. 2016), and low temperatures limit fertilization and ovule development.
High temperatures or severe drought during fruit or seed maturation may cause immature fruits
and cones to abort or be greatly reduced in size, although several studies have shown that moderate
increases in temperature during seed maturation increase seed production (e.g., Lowry 1966; Sork
et al. 1993; Fearer et al. 2008; Buechling et al. 2016). The seed crop is also subject to mechanical
damage by strong winds, hail, and insect damage. In general, no single factor appears to drive
masting in all species; one hypothesis is that mast seeding may be the result of an ancient ecologi-
cal relationship in ancestral populations, probably related to both climate and resource availability,
which was amplified though selection in the tree species we see today (Allen et al. 2017).
The abundance and periodicity of flowering are strongly influenced by physical site factors
of light, temperature, and moisture as well as the genetic and physiological mechanisms of the
plant. In a closed-canopy forest, most seed production results from the large dominant trees that
have crowns well exposed to sunlight; smaller trees with narrow and suppressed crowns yield few
if any seeds. Large, open-grown trees, free from aboveground competition for light and below-
ground competition for water, typically flower more frequently and abundantly than equally large
individuals of the same species in a closed-canopy forest. Such open-grown trees may produce
large quantities of seed each year if they are well pollinated. Isolated, open-grown conifers may
produce many seeds by self-pollination, decreasing their viability and reducing the growth of seed-
lings they produce. In addition, more favorable site conditions are likely to result in a greater flower
and seed crops. A classic study by Sarvas (1962) found that on poor sites in southern Finland
(height of dominant trees, 16 m at 100 years) fewer than 30 seeds were produced per square meter.
On sites of medium fertility (height of dominants, 23 m) and high fertility (height of dominants
27 m), 60 seeds m−2 and 90 seeds m−2 were produced, respectively. Similarly, pollen yields were
much higher on fertile sites (35 kg ha−1) than on infertile sites (9 kg ha−1).
In a study of mast fruiting in black, northern red, and white oaks in east-central Missouri,
Sork et al. (1993) considered both weather variables and the effects of prior seed production on the
current year’s acorn crop. They found that the size of an acorn crop was determined by both abun-
dance of flowers produced and their survival to mature fruit production. Furthermore, they found
that the length of reproductive cycles differed among the species (two, three, and four years for
black, white, and northern red oak, respectively), although evidence exists for widespread
synchronous flowering of oak species in mast years (Downs and McQuilken 1944; Beck 1977).
They concluded that there was an inherent mast-year periodicity for each species that additionally
may be affected by weather at critical times. Favorable spring temperatures during the season of
fruit maturation appeared to be critical for good acorn crops of all species, whereas summer
drought had a negative effect.
EFFECTS OF REPRODUCTION ON VEGETATIVE GROWTH

Flowering and the production of fruit and seed crops reduce vegetative growth. The common hy-
pothesis to explain this relationship is that photosynthate is a limiting resource that must be allo-
cated among reproduction and growth. Reproductive structures (flower, fruits, and cones) are
major sinks for photosynthate that will be unavailable to support growth, representing an impor-
tant trade-off (Harper 1977; Obeso 2002). Reproduction also affects terminal and lateral buds,
which receive less photosynthate and nutrients so that immature shoots telescoped within them
(next year’s shoots) are also reduced in size. These patterns have been long documented for fruit
trees (Pallardy 2008). Many correlational studies have also inferred this trade-off in forest trees,
noting that heavy seed crops decrease both height and radial growth. For example, Holmsgaard
and Bang (1989) reported that the stem volume increment of Norway spruce was reduced 12–25%
the year of an unusually large or “bumper” crop and from 4 to 9% in the two following years. For
the evergreen Mediterranean tree holm oak, acorn production was negatively correlated with
annual and late summer–autumn stem growth during mast years, but these trade-offs were greater
in smaller trees on poor sites (Martín et al. 2015). The negative relationship between cone produc-
tion and radial growth has been documented over a large geographic area for firs, larches, spruces,
pines, Douglas-fir, cedars, and hemlocks in North America (Koenig and Knops 1998). The strength
of this trade-off appears to be species-specific (Berdanier and Clark 2016) and increases with tree
age (Genet et al. 2010).
The evidence for a trade-off between reproduction and stem growth in trees is not clear, how-
ever (Obeso 2002). A study of five oak species in California found that radial growth and acorn pro-
duction were indeed negatively correlated, but only in species that produce acorns in the same year
that pollination occurs; species requiring two years for seed maturation showed no such relation-
ship (Knops et al. 2007). Importantly, increased rainfall increases radial growth, but decreases seed
production because it interferes with effective wind pollination. As a result, growth and reproduction
are independent of each other, but vary in opposite directions due to the same environmental con-
ditions, and therefore their negative correlation does not imply a causal mechanism (Figure 4.5).
Likewise, a study of European beech found that radial growth was affected by reproductive effort
and summer drought when they occurred simultaneously, but unaffected when one factor was
(a) (b)
*** Q. Lobata Q. Lobata ***
0.2
(1-year) (1-year)
0.3
0.1
0.0
0.0
–0.3 –0.1
*** –0.2 **
0.6 *** Q. douglasii Q. douglasii ***
(1-year) 0.4 (1-year)
0.3
0.2
0.0
0.0
Mean correlation between growth and reproduction

–0.3 –0.2
*** ***
Mean correlation with rainfall
0.6 *** Q. agrifolia Q. agrifolia ***

(1-year) 0.4 (1-year)
0.3
0.2
0.0 0.0
–0.3 –0.2
*** **
0.6 *** Q. chrysolepis 0.4 *** Q. chrysolepis
(2-year) (2-year)
0.2
0.3
0.0
0.0
–0.2
–0.3 –0.4
*** –0.6 **
*** n.a
0.9 Q. kelloggii Q. kelloggii
(2-year) 0.2 (2-year)
0.6
0.0
0.3
–0.2
0.0
–0.4
–0.3
–0.6 ** –0.6 **
Radial Acorn Growth yr1, Growth yr1,
growth crop Acorns yr1 Acorns yr2
F I G U R E 4 . 5 Correlation of (a) rainfall with growth and reproduction, and (b) growth with reproduction
for five California oak species. Stem growth of all species increases with increasing rainfall, but reproduction
decreases. Species whose seeds mature in one year exhibit a negative correlation between growth and
reproduction in the mast year; this relationship occurs in the second year for species requiring two years for
seed maturation. Source: Knops et al. (2007). Copyright (2007) National Academy of Sciences, U.S.A.
absent (Hacket-Pain et al. 2017). These studies suggest that if a trade-off exists between growth and
reproduction, environmental factors play an important role in the strength of the trade-off.
Crown dieback as well as reduced radial growth has been documented in the event of a
heavy seed crop. For example, an enormous yellow birch seed crop of 1967 in Ontario was eight
times greater than the good seed year that occurred the previous year (Gross and Harnden 1968;
Gross 1972). Resource allocation to flowering and fruiting suppressed early leaf expansion, and
leaves in heavily fruiting crown areas were only 75% of their normal size. These small leaves led to
poor shoot growth and small buds or lack of buds on the shoots, marked crown dieback occurred
in 1968, and new crowns developed below the dead branches. Radial increment in 1967 was only
47% of the annual average increment, and diameter growth of current shoots was less than 5% of
that in the previous year. Negligible fruit crops were produced the two following years.
DISPERSAL
Seeds are dispersed by gravity, wind, water, animals, and by combinations of these agents to micro-
sites where they may germinate and become established seedlings. The method by which seeds of
a species are dispersed enables it to compete in a particular spatial and temporal niche. In general,
most tree species disperse seeds locally, with a high proportion falling within perhaps 40–50 m of
the parent. Seeds of certain species are widely distributed by water and wind, particularly early-
successional trees such as willows, sycamores, cottonwoods, birches, and aspens (widely dispersed
seeds in Table 4.1). Most seeds of even these light-seeded, shade-intolerant, short-lived species are
dispersed locally, but some are blown by wind or carried by water far from the parent to recently
disturbed sites where establishment is more probable. Widespread seed dispersal was a successful
means of sexually reproducing early-successional species in pre-European colonization forests,
where disturbances were common.
Most trees and shrubs of northern latitudes disperse their seeds locally, with a seed rain
decreasing exponentially with distance from the tree (Figure 4.6). In a study of Engelmann spruce
4000
A
3000
Sound seed deposited, m–2
B
2000
1000
C
–20 0 50 100 150

Distance from forest edge, m
F I G U R E 4 . 6 The distribution of seeds of Engelmann spruce in good seed years from the edges of forest
stands into adjacent openings. (A) Togwotee Pass (1964), Teton National Forest, Wyoming; (B) Falls Creek
(1952), Flathead National Forest, Montana; (C) Griffin Top (1964), Dixie National Forest, Utah; (D) Fisher
Creek (1964), Payette National Forest, Idaho. Source: After Harper (1977) / Elsevier.
seed dispersal from four stands into adjacent openings, Dobbs (1976) found that most seeds fall
near the parent, and approximately 70% of the seeds fell within 50 m of the edge of the standing
timber. For white spruce in British Columbia, at 50, 100, and 200 m from the stand, seed rains of
19%, 12%, and 4%, respectively, were reported (Dobbs 1976). Only 9%, 5%, and 1% were reported for
the same species and distances in Alaska (Zasada 1985). Seed rain in the interior of a forest stand
is considerably shorter than into an adjacent opening because many seeds are intercepted by tree
crowns of the overstory and understory.
Heavy-seeded species such as walnuts and oaks have seeds dispersed close to the parents,
primarily by animals (locally dispersed seeds in Table 4.1). Dispersal relatively close (10–30 m) to
the parent usually occurs onto soil-site conditions similar to those where the parent tree became
established and is considered to be a selective advantage. Some species in this category have
evolved additional adaptations to ensure that they are able to establish and compete with the plants
and animals of the parent sites. Black walnut, for example, is highly sensitive to site conditions,
and is not competitive where the soil is too wet, too dry, or too shallow. Its heavy seeds are an
adaptation for this site specificity, ensuring that they reach the ground under the trees. Squirrels
and other animals disperse the seeds to microsites away from the parent tree, but similar in soil-site
conditions.
Most pines have moderately heavy, winged seeds that are wind-disseminated mostly within
50 m of the parent tree. In a study of longleaf pine, about 80% of the seeds caught in traps placed
at increasing distances from the parent fell within 40 m of the tree. Similarly, for shortleaf pine in
Arkansas, Yocom (1968) reported that one-half of the seeds trapped in a forest opening fell 10 m
from the edge of the stand, and 85% fell within 50 m of the stand edge. Because pines are typically
regenerated by fire and fire is likely within the lifetime of the parent tree, regeneration is likely to
occur in the vicinity of the parent. Conifers having wingless and/or heavy seeds such as junipers
and the stone pines (limber, whitebark, and Swiss stone, among others) and pinyon pines (wings
remain attached to the cone scales as they open) are dispersed primarily by birds. Bird-and-mam-
mal dispersal is extremely important for these and many other forest species and is discussed in
detail in Chapter 12.
If seeds are stored in a seed bank on the tree, dispersal may occur immediately after seed
maturation or periodically, depending on the establishment adaptations of the seedlings to con-
ditions of different ecosystems (Zasada et al. 1992). Seeds of elms, aspens, and silver maple, all
spring germinators, are dispersed immediately after ripening over a relatively short period in the
spring, germinating and establishing on moist or wet sites. Some pine species (jack, lodgepole,
Virginia, Monterey, etc.) have seeds that mature within cones that may remain closed (termed
serotinous cones) for short or long periods of time (fire-induced opening of cones, denoted by
F in Table 4.1; Chapter 10). These species are adapted to dry, fire-prone sites. Cones of some
serotinous species may open periodically when threshold temperatures are achieved. Other
cones remain closed until opened by fire-generated heat, dispersing the seeds onto the fire-
prepared seedbed for germination following summer rains. Most other pines have a four- to
eight-week seed dispersal period from cones that open shortly after ripening. This duration may
be affected by weather conditions (Figure 4.7). For eastern white pine in Maine in 1965, 75% of
the filled seeds were shed over a three-week period under warm and dry conditions. By contrast,
cool, moist conditions of 1968 delayed and lengthened the dispersal period, requiring seven
weeks to shed the same amount of seeds. Seeds of hemlocks, spruces, and firs are typically dis-
persed over many months in moist, cool climates. Dispersal begins in September or October for
firs and continues over the winter as the cones disintegrate and release the seeds. In black
spruce, cones are retained on the tree for several years in a “semi-serotinous” state, opening in
warm and dry conditions (or when a fire occurs) and closing in cool and moist conditions. Semi-
serotiny is advantageous because it ensures a continuous seed supply and that the seeds are not
all susceptible to wildfire on the forest floor.
50
40 Empty
Percent of total seed fall
Filled
30
1965 1968
20
10
0
6 13 20 27 4 11 18 25 1 8 15 22 29 30 Nov.–
Sept. Sept. Sept. Sept. Oct. Oct. Oct. Oct. Nov. Nov. Nov. Nov. Nov. 8 Apr.
F I G U R E 4 . 7 The time of seed fall of eastern white pine in 1965 and 1968, southwestern Maine.
Source: After Graber (1970) / United States Department of Agriculture / Public Domain.
Seed dispersal
n Seedlings
tio
ina
m
r
Ge
Active
Predation seed
Dormant uli bank
Stim
seed
bank
Decay and
senescence
Death
F I G U R E 4 . 8 Conceptual model of the seed bank and the dynamics of the population of seeds.
Source: Modified from Harper (1977), © 1977 by J. L. Harper. Reproduced by permission of Academic
Press, Inc.
SEED BANK, DORMANCY, AND GERMINATION

Seeds disseminated onto the forest floor are stored in the seed bank until they germinate (Figure 4.8).
Seeds of many forest trees arrive at the forest floor in a dormant state and remain in the dormant seed
bank until internal and external conditions are favorable for germination. The residence period of
viable tree seeds in the dormant seed bank (Table 4.1) is relatively short; being relatively large, they
are eaten by animals, and they decompose easily. Seeds of other species, particularly those dissemi-
nated in the spring (e.g., eastern cottonwood and silver maple), are capable of germinating just a few
days or weeks following dispersal, entering directly into the active seed bank (Figure 4.8 and Ta-
ble 4.1). Non-dormant seeds are typically adapted to certain favorable germination conditions pre-
sent immediately following dispersal. This adaptation applies to seeds of spring-disseminated, river
floodplain species. Also, seeds of many species of the hard-pine group (including jack, lodgepole,
Establishment Following Sexual Reproduction 83
longleaf, pitch, ponderosa, and Virginia) lack a dormancy requirement. Being strongly adapted to
fire, their seeds are ready to germinate following rains on the fire-prepared seedbed. Other pines,
including all pines of the soft-pine group (white and stone pines, among others), exhibit a pro-
nounced period of seed dormancy, although they also depend on fire for establishment.
To germinate, viable seeds must (i) imbibe water, (ii) activate metabolic processes, and (iii)
initiate growth of the embryo. Seeds are considered to be in a state of dormancy if they are unable
to satisfy any of these requirements (i.e., blockage of any of these processes). Morphological and
physiological changes must occur before dormancy is broken and the seed is capable of germinat-
ing. These changes take place slowly for seeds that overwinter on the forest floor, and the seeds
typically germinate the following spring (dormant seed bank in Table 4.1). The moist, cool condi-
tions of the forest floor over a period of weeks or months decrease germination inhibitors, increase
germination-promoting hormones, and create favorable internal conditions for germination. Seeds
of other species, including basswood, white ash, and black cherry, may require much longer
periods (one to three years after dispersal) before their dormancy requirements are satisfied. Like-
wise, seeds of many shrub species (Ribes, Rhus, Ceanothus, and many chaparral species) may be
stored for several years in the soil and germinate vigorously after fire.
ESTABLISHMENT FOLLOWING SEXUAL REPRODUCTION

Sexually produced seeds of woody plants produce roots, stems, and leaves in forming a seedling
upon germination that must cope with the environment. Many seedlings perish soon after germi-
nation, but those that survive and grow vigorously are regarded as established. Establishment may
take one to five years or more depending on the species and site conditions and is the most critical
stage in the life history of a tree. As discussed in the earlier text, species have evolved many adap-
tations to produce and disperse seeds in such a way that maximizes the chance that many seedlings
will encounter favorable conditions and survive. The regeneration ecology of each species is closely
tied to the physical and biotic factors of the sites where it typically establishes. Notably, seedlings
must survive on their seedbed, but they must also compete both with other seedlings of their own
species as well as plants already established on the site. Assuming natural processes and distur-
bance regimes remain, these interactions bring about an apparent order and some degree of orga-
nization that enable us to associate particular forest communities with specific sites.
Establishment strategies may be categorized into epigeous (above the soil surface) and
hypogeous (below the soil surface) germination and early seedling development. In the epigeous
condition (nearly all conifers and most angiosperms), the cotyledons are elevated above the sur-
face by the elongating hypocotyl, often still with the pericarp (fruit wall) attached (Figure 4.9a,b).
In hypogeous species (mostly nut-producing species such as oaks, hickories, walnuts, and buck-
eyes), the cotyledons remain below the surface, attached to the seedling for weeks or months while
the epicotyl grows upward and develops true leaves (Figure 4.9c,d). Four stages in cotyledon
development are recognized (Marshall and Kozlowski 1976):
1. Storage: Cotyledon cells are packed with stored foods such as fats, carbohydrates, proteins,
and mineral nutrients, which are utilized in the first days of seedling growth.
2. Transition: Following exposure to light, chloroplasts develop, and chlorophyll synthesis
begins, stomates develop, epidermal cells expand, and large intercellular spaces form in the
mesophyll.
3. Photosynthesis: The process of photosynthesis begins to further the development of the shoot
and tap and lateral roots. Appreciable photosynthesis begins four to six days after radicle
emergence in black locust, red maple, and American elm (Marshall and Kozlowski 1976). Peak
photosynthesis occurs 8–15 days after radicles emerge and continues for about four weeks.
4. Senescence: Dry weight decreases and some mineral nutrients are translocated back into
the seedling as the cotyledon function declines.
(A) (B)
f
d
d
1 cm
0
d b
b
c a
i c
(C) (D)
e 1 cm
g 0
d
a
i
c h
F I G U R E 4 . 9 Epigeous and hypogeous seed germination and development. (A) Epigeous germination of
pin cherry seedlings at 1 and 10 days; (B) epigeous germination of red pine at 1, 2, 6, and 10 days; (C)
hypogeous germination of Allegheny plum seedlings at 1 and 9 days; (D) hypogeous germination of bur oak
seedlings at 1, 5, and 12 days. (a) Seed; (b) hypocotyl; (c) radicle; (d) cotyledons; (e) epicotyl; (f) leaves; (g)
plumule; (h) pericarp; (i) primary root. The plumule consists of the epicotyl and the emerging leaves. Source:
A, C, and D after USDA (1974) / United States Department of Agriculture / Public Domain; B by W. H.
Wagner, Jr.
Epigeous species store relatively little food in the endosperm or cotyledons and rely strongly
on the cotyledons for photosynthesis to stimulate early root development. Thus, the cotyledons are
extremely important to the development of seedlings during the first few weeks, and any damage
such that might be caused by animals or frost inhibits seedling growth. By contrast, hypogeous
seedlings have large, fleshy cotyledons that remain below ground during seedling development
and are enclosed in the pericarp (Figure 4.9c,d). The large amount of stored food favors extensive
root development prior to the development of a transpiring shoot and leaf system. The cotyledons
also store considerable water and enough food to reestablish the epicotyl should it become dam-
aged. Furthermore, the cotyledons are protected from browsing animals because they are under-
ground and inside the pericarp.
The hypogeous system is typical of large-seeded woody species whose seeds are typically
buried by small animals and thereby dispersed relatively close to the parent tree (locally dispersed
seeds, denoted by L in Table 4.1). Many oaks and hickories grow on dry sites characterized by
summer drought, and their seedlings typically establish themselves in the shade of trees whose
roots compete for soil water and nutrients. In this context, production of fewer well-provisioned
seeds appears to be an adaptive advantage. Mesic sites with available soil water throughout the
growing season often support mixed hardwood forests with both hypogeous (bitternut hickory and
northern red oak) and epigeous (sugar maple, white ash, and beech) systems.
Young seedlings pass through a succulent stage during the first several weeks, when tissues
are soft and highly susceptible to fungal infections, damage by insect larvae and other animals,
smothering, and desiccation. Species whose seedlings have shallow young roots are especially sub-
ject to mortality compared to those with deeply penetrating “tap” roots. There follows a juvenile
period when the seedling becomes increasingly hardy but still subject to mortality once tissues
begin to harden. A classic study of seedling survival of Douglas-fir (Lawrence and Rediske 1962)
documented low cumulative germination and survival resulting from high seed destruction by
fungi and animals prior to germination (46%) and high first-year germination failure (27%), as well
as other factors (Figure 4.10).
The kinds of sites and seedbeds where seedlings of different species become established are
important for understanding site–species relationships and for predicting seedling survival. Most
tree species depend on disturbances (e.g., fire, uprooting of trees by wind, or flooding) to provide
suitable seedbeds for their establishment. Fast-growing, light-demanding species (pines, willows,
cottonwoods, aspens, birches, etc.) are favored by catastrophic disturbances that provide open and
extreme sites. The adaptations of fire-dependent species are presented in Chapter 10.
The establishment of many tree species is favored or not severely limited by partial shade of
an overstory, but heavy shade typically prevents establishment and/or persistence of most species.
Similarly, most seedlings can establish on continuously moist forest soils, but only few are able to
establish and persist as soils dry out and soil-water stress becomes severe. Each species therefore
has a unique set of adaptations that facilitates establishment under certain physical and biotic con-
ditions. In the following text, we divide (and simplify) the broad continuum of regeneration sys-
tems into three groups of species that differ in establishment pattern in relation to site conditions
and the type and severity of disturbance.
1. Early-successional species (often called pioneer species) seed into open areas following
major disturbances such as fires, flooding, windstorms, and landslides. Relatively little tree
competition is present, and the environment is often harsh (hot, cold, dry, wet, and/or
windy). Germination and growth on these sites are rapid, and seedling roots penetrate
deeply enough into the soil (as in the pines) to make the seedlings that eventually become
established resistant to yearly summer drought. Seedlings of floodplain species tolerate
variable water levels and are able to generate adventitious roots from their stems if covered
with water or silt. Fire-dependent pioneer species, such as the southeastern pines of North
Weeks after seeding

2 4 6 8 10 12 14 16 18 20 22 52
Feb. March April May June July
25
Accumulative germination
and seedling survival
20
Germination
15
% total seed
Survival
10
0
Frost
Seed movement
Seed loss to wildlife
Emergence of seedings
Post emergence – damping off
Duration of some factors Clipping injuries

influencing germination and
survival of seedings Desiccation
Feb. March April May June July
2 4 6 8 10 12 14 16 18 20 22 52
Weeks after seeding
F I G U R E 4 .1 0 The cumulative germination and survival of Douglas-fir seeds during the first growing
season. 440 Scandium46-tagged seeds were seeded by hand in February and their fate carefully followed.
Source: Lawrence and Rediske (1962) / with permission of Oxford University Press.
America (longleaf, shortleaf, loblolly, and slash), typically establish in dense, even-aged
stands on recently burned areas, and the seedlings develop rapidly during the first year
(Figure 4.11; Wakeley 1954). Tap root development is rapid because considerable soil water
is needed to meet transpiration demands in the needles and shoots which develop rapidly in
the ensuing juvenile period.
DAYS SINCE SOWING

0 40 80 120 160 200 240 280 320
20 51 cm
Seeds Primary Secondary Terminal Winter buds
germi- needles needles buds appear;
nate appear appear appear some burst
TOP HEIGHT (INCHES)
A F
LONGLE
10 25 cm
S H
Branches
SLA
appear LY
Second terminal LOL
bud appears LOB
TLEAF
SHOR
Upper branches
appear
Stems crook Basal branches appear GROUND LINE
0 0 cm
10 25 cm
ROOT DEPTH (INCHES)
LO
B LO
L LY
SL
AS
20 SH H 51 cm
O
RT
L EA
F
Tap root increases
in diameter
Lateral Mycor- Corky LONGL

Roots stop Growing E AF
roots rhizae layer
growing in length points appear
appear appear appears Growing points multiply
30 76 cm
0 40 80 120 160 200 240 280 320
DAYS SINCE SOWING
MARCH APRIL MAY JUNE JULY AUG. SEPT. OCT. NOV. DEC. JAN.
F I G U R E 4 .1 1 The course of first-year shoot and root development of four species of pines under favorable
conditions for establishment (in a forest nursery) in Louisiana. Source: Wakeley (1954) / United States
Department of Agriculture / Public Domain.
2. Gap-phase species establish under the existing forest canopy and are shade-tolerant
enough to persist until a small disturbance enables them to penetrate a gap in the canopy
(Chapter 16). Common gap species include white ash, black cherry, white and northern red
oak, red maple, yellow birch, basswood, black walnut, slippery elm, white spruce, and to a
lesser extent Douglas-fir and white pines. Notably, a seedling of any species may reach the
overstory canopy via a gap, if positioned in the right place at the right time. Gap-phase
species are able to tolerate the deeply shaded forest understory, but can also grow quickly
enough to colonize the gap before the gap is filled by nearby crowns or by competing
species. Gap size is important for this process, as light-requiring pioneer species are better
able to regenerate in large gaps than gap-phase species or the slow-growing, very shade-
tolerant species that can reach the canopy in small gaps. The phenomenon of gap or patch
dynamics is discussed in detail in Chapter 16.
3. Shade/understory-tolerant species have seedlings able to establish in the shaded
understory on moist or mesic sites (i.e., where soil water is available throughout the
growing season) and persist for years or decades (persistent juveniles, denoted by Bj in
Table 4.1). Seedlings of these species (e.g., sugar maple, American beech, hemlocks, true
firs, and western redcedar) slowly penetrate the canopy as overstory trees die or windthrow
provides openings for them.
Seedlings of both gap-phase and shade-tolerant species (groups 2 and 3) establish under a
forest canopy and must persist in the understory until conditions are favorable for their growth into
the overstory. Species in these groups have evolved many different kinds of adaptations that allow
their seedlings to persist in the understory and respond to release depending on (i) the physical and
biotic conditions of the understory, (ii) the species’ intrinsic growth rate, and (iii) the type of distur-
bance that provides canopy openings. Both groups (but especially group 3) accumulate a few to hun-
dreds or millions of seedlings per hectare in the understory. Both groups also have different functional
adaptations that enable them to tolerate and survive the physical rigor (low light and soil-water
stress) and biotic hazards (herbivores and diseases) of the understory environment. Gap-phase
species better tolerate soil-water stress than do shade-tolerant species, but the reverse is true for low-
light conditions. In addition, deciduous gap-phase species are able to sprout after fire or animal
damage and continue to persist in the understory. Most shade-tolerant species have slow-growing
seedlings which require more soil water, typically dominate on mesic sites, and may accumulate
large seedling populations that experience high mortality. Seedlings of shade-tolerant species may
endure for 20–30 years or more to shade out any seedlings that might try to establish under them;
understory hemlocks may even be 100 or more years old. Both gap-phase and shade-tolerant trees
that eventually reach the overstory were recruited from a bank of established and persisting seedlings.
Two associated species of the northern hardwood forest, yellow birch (gap-phase) and sugar
maple (shade-tolerant), have contrasting establishment strategies. These species grow side by side
in the northern forest, but with very different physiological adaptations and realized niches. Seed
dispersal for sugar maple occurs with leaf drop (before snowfall) in the early fall. Yellow birch
seeds (more than 50 times smaller than sugar maple seeds) also mature in the early fall, but they
are dispersed gradually throughout the winter and may be blown for great distances on top of the
snow (Figure 4.12a; Tubbs 1965). Sugar maple seeds germinate under the snow in the early spring
within a layer of leaves where temperatures are only about 1 °C (Figure 4.12b,c), while yellow birch
seeds remain on top of the snow (Figure 4.12a). Radicles of maple seedlings penetrate the wet leaf
mat and establish by the millions in a good seed year, often forming a carpet of first-year seedlings
on the forest floor. By contrast, birch seeds come to rest on the forest floor after snowmelt and ger-
minate in late spring at higher temperatures (about 10 °C). Birch seed radicles are tiny and unable
to penetrate the thick leaf mat, which tends to dry out rapidly. Most yellow birch seedlings desic-
cate and die, but some birch seeds come to rest by chance on rotting logs, moss-covered rocks, or
mineral soil of mounds formed by uprooted trees where they may establish if sufficient light is
available and the substrate does not dry out.
Sugar maple seedlings are highly shade-tolerant and are able to survive the forest floor envi-
ronment better than most other tree species in its geographic range. Yellow birch seedlings cannot
endure heavy shade, but grow much faster than sugar maple at higher light levels. Thus, an
F I G U R E 4.1 2 Sites of seed dispersal and germination of sugar maple and yellow birch in the northern
forest. (a) Yellow birch bracts and seeds litter the surface of melting snow on April 30 in a northern hard-
wood stand on the Upper Peninsula Experimental Forest, Michigan. The overstory is composed primarily of
sugar maple of seeding age, but seldom are maple seeds observed on top of or within the snow cover even
after a bumper crop. Right arrow indicates a yellow birch seed; left arrow points to a bract. (b) Removing the
snow from the exact spot shown in (a) to the top of the previous fall’s leaf layer reveals no seeds of any
species. The leaves are compressed into a soggy mat that is often partially frozen. Spring ephemerals have
pushed through the mat (arrow). (c) When the top leaf layer shown in (b) is removed, the ability of sugar
maple to germinate underneath a snow cover in early spring is revealed. Arrows point to germinated seeds.
In those areas sampled on the Experimental Forest, the bulk of sugar maple seed was found under a layer of
leaves, whereas yellow birch seeds were distributed on the top of the snow as illustrated in (a). Source: After
Tubbs (1965); U. S. Forest Service photos.
a ppropriately positioned yellow birch is able to outcompete the slower-growing maples for a place
in the overstory canopy upon the formation of a large gap. Sugar maple is able to dominate both
understory and overstory in many ecosystems of the upland hemlock–northern hardwood forest
because of its effective seed production, dispersal, germination, and establishment adaptations.
POST-E STABLISHMENT DEVELOPMENT

Tree species exhibit rapid growth in the juvenile phase when environmental conditions are favor-
able. Growth in trees slows greatly with increasing age and begins to decline in old age. The high
rate of stem mortality that occurs during the establishment period decreases with time. Trees con-
tinue to compete for light, water, and nutrients as they increase in size, and their density (number
of stems per unit area) decreases (Chapter 13). Relationships of shoot, crown, and root structure to
regeneration ecology and tree growth are considered in detail in Chapter 5.
VEGETATIVE REPRODUCTION
Reproduction by vegetative means rather than sexually with flowers and fruits may be critical for
the survival of woody plant populations in some situations, such as following disturbances. Vege-
tative reproduction is achievable (and is a major fitness trait) for all woody angiosperms once
established (Chapter 3), but is less common in conifers. This difference between conifers and
angiosperms may be one of the major reasons for angiosperm dominance since the Cretaceous
Period. Vegetative reproduction is advantageous because it allows the plant to survive and reestab-
lish itself in place and sometimes to expand its spatial coverage of an area. It also allows plants to
quickly colonize disturbed areas, as in pioneer species such as aspens, willows, sumacs, dogwoods,
and many shrub species. Vegetative reproduction produces ramets genetically identical to the par-
ent plant which grow rapidly from an already-established and functioning root system—quite
unlike seedling establishment. Ramets may persist for long periods of time once established. In
fire-prone ecosystems, woody species that form clones resprout vigorously after fire, preventing
erosion and rapidly recoupling nutrient cycles.
Vegetative regeneration contributes meristems to the bud bank from the crown, basal stem
and root collar, and underground roots and rhizomes (Figure 4.13). Vegetative buds are released by
temperature, fire, wind breakage, herbivores, or other stimuli, forming new shoots and initiating
clonal regeneration and expansion. Most woody plants regenerate vegetatively through sprouting
(development of new shoots from adventitious buds) and rooting of stems. Sprouting occurs in
different parts of the plant:
Basal stem: New shoots may develop from dormant or adventitious buds at the basal part of the
stem (i.e., at the root collar where stem joins root) of an established plant. Sprouts from basal
stems may form multiple-stemmed shrubs or trees such as oaks, hickories, basswoods, ashes,
walnuts, birches, alders, prickly-ash (Reinartz and Popp 1987), and species of many other genera.
Root: New shoots may arise from adventitious buds on roots. Root sprouting is common in clone-
forming trees and shrubs such as aspens, sassafras, sumacs, and sweetgum, as well as in
American beech.
Rhizome: New shoots may develop from underground horizontal stems. The distal end of a rhi-
zome becomes erect and bears leaves and flowers. Sprouting from rhizomes is typical of many
shrubs such as salmonberry (Tappeiner et al. 1991) and other Rubus species, and trees such as
flowering dogwood, striped maple, Gambel oak (Tiedemann et al. 1987), and dwarf chestnut oak.
Lignotuber: New shoots may arise from a buried mass of stem tissue termed the lignotuber, as in
Gambel oak and eucalypts.
Vegetative Reproduction 91
Clone Expansion
Clone Regeneration
Growth and development
Factors stimulating release:

Bud release fire, wind, temperature
herbivory, cutting
Bud Bank
Mortality
Individual, Ecosystem
Due to:
drought, fire
herbivory, flooding
accidental breakage
disease
Crown Basal Stem Roots and

and Root Collar Rhizomes
Axillary buds
Adventitious buds Adventitious buds
Adventitious buds
Origin of Meristems
F I G U R E 4.1 3 Conceptual model of vegetative regeneration and clonal expansion of woody plants.
Source: Zasada et al. (1992). Reprinted with permission of Cambridge University Press.
Vegetative reproduction that occurs by rooting of aerial stems and subsequent shoot
development includes:
Stolons or runners: Stolons are the arching branches of shrubs, such as red-osier dogwood and
Rubus species, which take root when they come in contact with the soil surface and form new
plants. Runners (procumbent stems) of plants such as creeping strawberry-bush and Virginia
creeper take root at various points along their length as they encounter the soil surface.
Fragmentation: Branches that are broken off by ice, flood debris, or wind or senesced from
willow, cottonwood, and other stream-side species may take root and become established after
being buried in soil of the stream bank or the active floodplain.
Layering: Lower branches of boreal and northern conifers such as spruces, firs, and larches are
often pressed into the soil by the weight of snow or by woody debris. These branches take root
and form clones around the perimeter of the parent tree.
Tipping: Northern white-cedars in swamps are often uprooted just enough by wind so that
they very slowly tip over and eventually lie procumbent on the peaty surface of the swamp
(Curtis 1959). The lateral tree branches turn upright and develop into independent trees as moss
and organic matter cover the old tree, and roots form at the base of each branch.
In addition to maintaining individual genotypes following disturbance, vegetative

reproduction expands the area occupied by many shrubs and trees such as blueberries, dogwoods,
sumacs, sassafras, beeches, aspens, sweetgum, and sycamores. Trembling aspen clones, consisting
F I G U R E 4 .1 4 Large clone of trembling aspen, 43 ha in extent, with the boundary outlined; south-central
Utah. Note the smaller clones around the large clone; the differences in tone indicate differences in fall
coloration of clone foliage. Source: After Kemperman and Barnes (1976). Reprinted with permission of
National Research Council Canada.
of many genetically identical ramets, are very common following wildfires. In the Intermountain
West of North America, trembling aspen clones may encompass very large areas; one such clone
in south-central Utah commonly known as the “Pando Clone” is 43 ha in area (Figure 4.14;
Barnes 1975; Kemperman and Barnes 1976; DeWoody et al. 2008; Mock et al. 2008) and is reported
to be one of the world’s largest organisms (Grant et al. 1992; Grant 1993). Vegetative reproduction
is much more common in maintaining and expanding some clonal species than is the production
of new seedlings, including trembling aspen (Romme et al. 2005) and the northern subshrub win-
tergreen (Moola and Vasseur 2008).
Rather than spreading expansively by sprouting from roots, many forest shrubs spread by the
repeated rooting of aerial stems that root and produce new ramets when pressed into the forest
floor. For example, clones of the shade-tolerant vine maple develop in the understory of Douglas-
fir forests in the Pacific Northwest when stems are pinned to the forest floor by falling trees and
branches and then take root (O’Dea et al. 1995; Figure 4.15). Vine maple establishes by seed follow-
ing fire or windstorms and then develops as multiple-stemmed groups. The maple branches are
vinelike and flexible and droop to the forest floor over time. Almost 99% of the stems that repro-
duce vegetatively by layering are pinned to the forest floor by fallen trees or branches, and repeated
layering events by the genet and its ramets extends the clone over a large area (Figure 4.15). This
clonal process allows vine maple to persist in the understories of Douglas-fir forests, perpetuated
in part by the mutualistic relationship with the dominant trees themselves.
Most vegetative reproduction by conifers is by layering in the northern, boreal, and alpine
areas, but even in these examples, conifers reproduce primarily by sexual means. Only a few coni-
fer species are able to sprout vigorously and regenerate themselves after fire, grazing, or cutting.
Examples include the redwood, which sprouts vigorously from root collar, stump, and stem, and
Vegetative Reproduction 93
Decumbent vine
maple stem
Fallen tree
Point of layering
18 yr
83 yr 55 yr
68 yr
125 yr 23 yr
45 yr
10 m
64 yr
45 yr 33 yr
83 yr
60 yr
80 yr
F I G U R E 4.1 5 Top view of a vine maple clone in a 130-year-old Douglas-fir forest showing decumbent
stems and points of layering. Aerial stems are not shown. Irregular enclosed areas indicate vine maple
crown coverage. Note the large range in age among ramets of the clone. Source: O’Dea et al. (1995) /
John Wiley & Sons.
various hard pines that sprout from dormant buds along the stem after fire or cutting, including
pitch pine, Virginia pine, and pond pine (Kozlowski 1971a). Near tree line in the southern Rocky
Mountains, Marr (1977) reported “tree islands” of subalpine fir and Engelmann spruce that moved
along the ground by repeated layering and growth to leeward. Movement of 5 m in 11 years was
common, and some clonal islands apparently moved at least 15 m. In this example, seedlings
apparently became established in sheltered microsites, but the clones expanded and colonized
adjacent microsites inhospitable to establishment.
In addition to the adaptive benefits described in the earlier text, vegetative reproduction
ensures the survival of a single genet either by resprouting from its roots or protected stem follow-
ing destruction of the aboveground parts or by spreading mortality risk of the genet among exist-
ing ramets that are capable of independent survival (Cook 1979). For example, persistence of green
ash in the presence of emerald ash borer in eastern North America is likely to be maintained by
basal sprouting from top-killed trees after attack by the invasive beetle (Kashian 2016). In a five-
year study of pure green ash stands in southeastern Michigan, basal sprouting was the dominant
mode of regeneration where 58% of the trees had been killed by the beetle. Sprouts grew far faster
than either established ash seedlings or surviving overstory ash trees, and 27% of the sprouts pro-
duced significant seed during a mast year that occurred during the study. Thus, the presence of
vegetative reproduction from top-killed ash facilitated continued sexual reproduction that will
likely lead to the species’ persistence, although at much lower abundance with far smaller and
shorter-lived trees (Kashian 2016). Sprouting in this example, which notably does not occur from
every killed tree or in every population of green ash, is therefore critical to ash persistence.
Vegetative reproduction probably contributes more to short-term persistence and spatial domi-
nance of woody species, but it is clearly sexual reproduction that provides the genetic variation for
persistence in the face of longer-term environmental change.
SUGGESTED
READINGS
Grubb, P.J. (1977). The maintenance of species-richness Owens, J.N. (1991). Flowering and seed set. In: Phys-
in plant communities: the importance of the regen- iology of Trees (ed. A.S. Raghavendra), 247–271.
eration niche. Biol. Rev. 52: 107–145. New York: Wiley.
Knops, J.M., Koenig, W.D., and Carmen, W.J. (2007). Pallardy, S.G. (2009). Physiology of Woody Plants, 3e. San
Negative correlation does not imply a tradeoff bet- Diego, CA: Academic Press 454 pp.
ween growth and reproduction in California oaks. Silvertown, J.W. (1980). The evolutionary ecology of
Proc. Natl. Acad. Sci. 104: 16982–16985. mast seeding in trees. Biol. J. Linn. Soc. 14: 235–250.
Kozlowski, T.T. (1971). Growth and Development of Zasada, J.C., Shank, T.L., and Nygren, M. (1992). The
Trees, Vol. I. Seed Germination, Ontogeny, and Shoot reproductive process in boreal forest trees. In: A Sys-
Growth. New York: Academic Press 443 pp. tems Analysis of the Global Boreal Forest (ed. H.H.
Obeso, J.R. (2002). The costs of reproduction in plants. Shugart, R. Leemans and G.B. Bonan), 211–233.
New Phytol. 155: 321–348. Cambridge: Cambridge Univ. Press.
Tree Structure and Growth CHAPTER 5
T wo kinds of growth occur when a seedling develops into a large tree: primary growth,
or the extension of each growing point forming the shoots of the crown and the roots, and
secondary growth, involving the expansion of stem and root diameter. Each species has a
characteristic aboveground and belowground structure and form that represent a combination
of these growth processes, following an inherent architectural model and influenced by environ-
mental factors. In this chapter, we emphasize the tree crown, stem/shoots, and selected aspects of
root structure and growth in relation to environmental factors, especially soil water. We consider
growth and development of shoots and roots in detail, but a treatment of other important aspects
of tree and shrub physiology is beyond our scope. For detailed treatments of tree physiology, the
reader is directed to books by Kramer and Boyer (1995), Kozlowski and Pallardy (1997a), Larcher
(2003), Pallardy (2008), and Lambers and Oliveira (2019). For considerations of tree structure and
function, classic works by Büsgen and Münch (1929) and Zimmermann and Brown (1971) are
notable as well as by Oldeman (1990).
TREE FORM
Tree form, habit, or architecture is as incredible as it is fascinating. Diverse regional and local eco-
systems favor many species, vertical vegetation strata, and long periods of rhythmic or continuing
growth, especially in tropical forests. Tree architecture is determined by many factors such as
apical control of growth, shoot types, branching orientation, timing of meristematic activity, her-
bivory, and shedding of shoots. Tree architecture is also shaped by physical factors of light, temper-
ature, soil water, and nutrient availability. Readers are urged to examine more detailed treatments
of tree architecture by Wilson (1984), Barthelemy et al. (1991), Bell (1991), Turnbull (2005), and
Hollender and Dardick (2015).
Generally, tree form refers to the shape, size, and/or habit of the crown of a tree. The form of
a woody plant is determined by its genetically pre-determined architectural plan, and by the envi-
ronmental influences that affect its development. Each species has obtained a precisely deter-
mined growth plan or architectural model selected by and adapted to the physical and biotic factors
of the ecosystems where it lives. In northern and alpine environments, where snow, ice, and wind
are strong selective factors, trees (primarily conifers) are adapted to grow in a conical form
(Figure 5.1a). It is clear that this trait is under strong genetic control when species from high alti-
tudes (spruces, larches, and firs) maintain their conical form in parks far from their native sites.
The conical form results when the terminal shoot or leader grows systematically faster than the
lateral branches below it, creating a prominent, central stem. This tree form is called excurrent
and is typical of northern conifers and a few deciduous trees, such as sweetgum and tulip tree.
The excurrent form results from strong apical control, whereby the terminal leader
inhibits the growth of the lateral branches below it, and thus grows faster than them (Cline 1997;
Sterck 2005). Apical control is a process that shapes the growth and development of the entire
tree, including the length and angle of branches and stem diameter (Wilson 2000; Sterck 2005).
95
96 C h a p t e r 5 Tree Structure and Growth
(a)
Balsam fir Engelmann spruce

(b)
Black oak American elm
F I G U R E 5 . 1 Examples of (a) excurrent and (b) decurrent or deliquescent forms of trees. Source: After
Hosie (1969). Reproduced by permission from Native Trees of Canada, 7th ed. by R. C. Hosie; published by
the Canadian Forest Service and Fitzhenry and Whiteside Ltd., 1969.
Apical dominance is related to apical control, but specifically refers only to the effect of the
terminal shoot on the current year of growth rather than on the tree as a whole (Brown
et al. 1967). Together, apical dominance and control are strong determinants of tree architecture.
The subtle difference between these processes is important because apical control is not simply
a summation of annual apical dominance, and the molecular mechanisms driving apical domi-
nance are far better understood compared to those driving apical control (Hollender and
Dardick 2015).
Climate has a strong influence on tree form such that not even all conifers are excurrent. For
example, pines of the warm and humid southeastern United States, such as loblolly and longleaf,
exhibit a strong central stem, but a more rounded crown than spruces and firs of the far north. In
the arid Southwest, pinyon pines and junipers tend to have short, compact, rounded, and bushy
crowns. This form is related to the high temperature and soil-water stress of their environment as
well as the sparse nature of the forests in the region. These conditions favor extensive root
development and compact crowns rather than tall excurrent growth that would severely expose the
crown to prolonged, drying winds.
Tree Form 97
In many deciduous species, the lateral branches grow nearly as fast or faster than the terminal
leader. Weak apical dominance and control may allow the main stem to fork repeatedly, giving rise
to a broad, spreading form (Figure 5.1b). This decurrent or deliquescent form is typical of elms,
oaks, maples, and many other species that grow in less harsh environments compared to conifers.
In these species, a multiple-terminal leader system is often developed, most pronounced in
American elm and many oaks. The characteristic “vase shape” of many elms is related to the
branching and re-branching of lateral shoots that results in an effective disappearance of a main
stem (Pallardy 2008). Such broad, spreading crowns may have an adaptive function in spacing out
the relatively large deciduous leaves to allow access to light compared to the densely packed
needles of many excurrent conifers.
Within the group of decurrent angiosperms, some species that grow along rivers and streams
or along a forest edge, besides having spreading crowns, may exhibit markedly leaning trunks com-
pared to upland or forest-interior species. For example, Loehle (1986) reported that black willow,
sycamore, and river birch had a lean from the vertical of 21°–22°, whereas the average lean of three
upland oaks was only 7°. The growth response toward light, phototropism, markedly affects tree
form of certain species and enables these shade-intolerant species to obtain light from directly over
the river. However, other bottomland species such as American elm, eastern cottonwood, and red
or green ash rarely display such a pronounced lean, but instead exhibit different branching archi-
tectures to optimize light capture by the crown for photosynthesis.
ARCHITECTURAL MODELS
The total diversity of tree forms were generalized into 23 architectural models (which also apply to
shrubs and herbs) by Hallé and colleagues (1978; Figure 5.2). Most of these models describe
tropical species, but several are directly applicable to temperate conifers and deciduous trees. A
given species may share features of more than one model, as for eastern hemlock (Hibbs 1981).
The models were developed based on whether shoot growth is rhythmic (seasonal) or continuous,
whether the tree is unbranched (as in some palms) or branched, whether or not branching is or-
thotropic (erect) or plagiotropic (horizontal), and whether flowers are located on terminal or
lateral shoots. Hallé et al. (1978) identified several of these models appropriate for temperate plants
and named them after a well-known botanist (Figure 5.2): Leeuwenberg (sumacs, red-osier
dogwood, and rosebay rhododendron), Rauh (most species of pines, red and live oak, and probably
most oaks), Massart (redwood and true fir species such as balsam fir and European silver fir), Tom-
linson (most multiple-stem palms and many sedges and grasses), Troll (American beech and
American elm), and McClure (many bamboo species). Note that excurrent forms of pines are
grouped with spreading forms of oaks, indicating the broad nature of these architectural plans. A
set of non-standard tree forms, as summarized by Hollender and Dardick (2015), have also been
described as applicable to crop trees and those used for landscape design. The forms include
columnar/pillar, weeping, dwarf/semi-dwarf, and compact/highly branched, and most have been
at least partially described at the molecular level.
Architectural models for trees are most easily observed when trees are in the seedling or sap-
ling stage because their form has been least subject to the environmental stressors or damage that
occur as they mature. Therefore, examining tree architecture clarifies how woody species respond
to stresses and crown damage over time. When new shoots arise from a damaged tree, they arise
according to the species’ original architectural growth plan. This process, termed reiteration (Bar-
thelemy et al. 1991), also occurs in old crowns and in spontaneous vegetative reproduction from
roots or stems. This replication of architecture without an obvious stimulus has been proposed as
a way for some tree species to maintain their crown and thus extend their longevity (Ishii
et al. 2007). Thus, the crown shape of woody plants results from a genetically determined architec-
tural plan, modified by the environment, and reiterated as the plant matures.
Leeuwenberg’s model Rauh’s model Massart’s model
Tomlinson’s model Troll’s model McClure’s model
F I G U R E 5 . 2 Generalized architectural models of trees, selected from the 23 models of Hallé et al. (1978).
Besides tropical species, the following temperate groups or species are characterized: Leeuwenberg’s
model—sumacs, rosebay rhododendron; Rauh’s model—most pines, red and live oaks; Massart’s model—
redwood and true firs; Tomlinson’s model—multiple-stem palms; Troll’s model—American beech, American
elm; McClure’s model—many bamboo species. Source: Tomlinson (1983) / with permission of Sigma Xi, The
Scientific Research Honor Society.
The canopy architecture of individual species is combined in a given temperate or tropical

ecosystem to produce a particular canopy structure, usually of several layers. Multiple tree and
shrub layers characterize moist tropical and temperate forests (for example, see Figure 6.11 in
Chapter 6), whereas dry tropical or temperate forests may exhibit a totally different vertical canopy
structure. Canopy structure directly influences the amount of leaf surface area present to capture
light for photosynthesis. Trees with narrow columnar forms are associated with dry or high-eleva-
tion sites, whereas broader crowns are associated with mesic or humid sites (Landsberg 1995).
Tropical studies have suggested that the tallest trees have the widest and flattest crowns with
Tree Form 99
upwardly angled branching, the second layer trees are as deep as they are wide, and in the third
layer, trees crowns are conical with horizontal branching (Pallardy 2008). The interrelated factors
of tree form, vertical canopy structure, and leaf area, as they influence interception of light, photo-
synthesis, and overall tree productivity, are discussed in Chapter 6.
Short and Long Shoots The kinds of shoots and the patterns of shoot growth influence tree
form and competitive ability in different ecosystems, as well as reproductive and regeneration
potential. Plants have an open system of growth whereby the shoot (the stem with its collection of
leaves and buds) is the modular unit of construction. In trees, the modular unit implies that buds
open and expand, the shoot elongates, and then develops a new bud for the process to repeat. Mul-
tiple kinds of shoots make up the modular units and also the patterns of shoot growth that affect
tree structure and function.
All woody plants bear long shoots, and some species have both long shoots and short shoots.
Long shoots contain nodes, the areas where leaves and buds are found on the stem, and inter-
nodes, the areas between successive nodes. Long shoots lengthen mostly from elongation of the
internodes. The annual growth of long shoots of yellow birch may range from a few centimeters to
several meters. Annual long-shoot growth of over 3 m is not uncommon for young eastern cotton-
wood trees on alluvial sites.
Short shoots, also called dwarf shoots in conifers or spur shoots in deciduous trees, exhibit
very little or no internode elongation (a few millimeters or less) and typically mark annual shoot
growth only by groups of adjacent ringlike scars left by the bud scales. For example, we collected a
28-year-old short shoot of yellow birch that measured only 7 cm long (compare to the length of long
shoots above). Short shoots are characteristic of beeches, birches, maples, and fruit trees such as
cherries and apples. In the gymnosperms, short shoots are found in the deciduous genera including
larches and ginkgo. In the pines, needle-bearing fascicles are modified short shoots. They are deter-
minate and persist more than one season, but do not extend into long shoots. Temperate trees tend
to feature lateral short shoots, but short shoots are terminal in most tropical trees. Long shoots may
be either terminal or lateral in both temperate and tropical forests and often occur in the upper
crown where the main canopy growth occurs. In European white birch, for example, short shoots
make up over 90% of all shoots on a tree, but 80% of the shoots at the top of the tree are long shoots.
Leaves on short shoots are produced at the same place every year and their relatively fixed position
allows them to serve as primary photosynthetic units (Bell 1991, p. 254). For example, the broad-
paired, short-shoot leaves of hobblebush, a shrub that grows in shaded understories of conifer for-
ests of the Appalachian Mountains, are positioned to maximize light capture.
Short shoots bear only “early” leaves whose primordia are pre-formed in the dormant or rest-
ing bud; they also may bear flowers, cones, or thorns. Short shoots are characterized by determi-
nate (fixed) growth, meaning that shoots are pre-formed in the dormant bud, expanding in the
spring, and then ceasing growth before setting buds late in the growing season. In contrast to short
shoots, long shoots may be either determinate or indeterminate. Indeterminate (free) growth
features shoots where the basal portion is pre-formed in the bud and elongates in the spring, but
the shoot continues to grow throughout the growing season if light and other site conditions are
favorable. Indeterminate long shoots are advantageous in that they extend the framework of the
plant into new space. Indeterminate shoots are especially well developed in well-lighted crown
positions of fast growing, pioneer species of pines, birches, cherries, cottonwoods, balsam poplars,
aspens, sassafras, and sycamores (Marks 1975). Determinate long shoots bear only early leaves,
whereas indeterminate long shoots bear early leaves along their basal portion and “late” leaves,
those formed during the current growing season (rather than pre-formed in the bud), along the
upper portion. Late leaves produced after the early leaves develop may bear little resemblance to
early leaves (Figure 5.3), but represent the internal environment at the time and place of primordia
formation and development rather than the external environment.
F I G U R E 5.3 Silhouettes of leaves from a leader

shoot of trembling aspen. Leaves 2–7 are typical
early leaves; 8 is a transition form to late leaves; and
leaves 12–18 are typical late leaves. Source: Barnes
2 3 4 5
(1969) / with permission of JD Sauerlander’s Verlag.
6 7 8
12 13 14 15
16 17 18
Patterns of Intermittent Growth Most woody plants exhibit intermittent (periodic, episodic,
and rhythmic) rather than continuous growth. Some trees exhibit continuous growth by producing
a pair of leaves every few months, such as in palms and mangroves, but even these do not con-
stantly extend their shoots (Borchert 1991). Meristems or buds typically are induced into a resting
or dormant state by cold temperature or soil-water stress; buds in a resting state resume growth
within one to two weeks once the favorable conditions resume, but dormant buds do not. Intermit-
tent growth in temperate areas (evidenced by leaf flushing and shoot extension) is controlled by
temperature; in the tropics it is controlled by soil water. Intermittent growth in some trees that is
unrelated to environmental changes is called rhythmic growth (Borchert 1991).
Long shoots are responsible for height growth and crown extension and therefore are an
advantage to use the growing season effectively. Zimmermann and Brown (1971) illustrated vari-
ous ways temperate trees make use of favorable growing conditions by designating four types of
long-shoot extension:
1. A single flush of terminal growth followed by formation of a resting bud (seasonally
determinate long shoots).
2. Recurrent flushes of terminal growth with terminal bud formation at the end of each flush.
3. A flush of growth followed by shoot-tip abortion.
4. A sustained production of leaves including early-and late-leaves up to the time of terminal
bud formation (seasonally indeterminate long shoots).
Type 1: As described in the earlier text for determinate growth, many northern species make a
burst of growth in spring or early summer and then form terminal or end buds (Figures 5.4
and 5.5) before mid-summer soil-water stress intensifies. Ecological conditions largely determine
when the burst of growth begins. For example, red and eastern white pines exhibit different tim-
ing of shoot growth in North Carolina as compared to that in New Hampshire (Figure 5.4). Once
Tree Form 101
80
White pine Red pine

70 North Caroline New Hampshire
Season’s growth in each month, %
60 Red pine
North Caroline
50
40
White pine
New Hampshire
30
Mean of three
southern pines
20 North Carolina
frost-free season
10
New Hampshire
N. Carolina
Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec.
F I G U R E 5 . 4 Variations in seasonal height-growth patterns of red pine and eastern white pine in North
Carolina and in New Hampshire, and of three southern pines (loblolly, shortleaf, and slash) in North
Carolina. The northern pines have pre-formed shoots and usually have one annual growth flush, whereas the
southern pines grow in recurrent flushes. Source: Kramer (1943) / with permission of Oxford University Press.
the terminal bud forms, shoot and leaf primordia develop within the bud scales and expand the
following year. The current year’s shoot growth is therefore primarily dependent on environmen-
tal conditions of the previous year when leaf primordia were formed. Other example species of
this type of fixed growth include some oaks, hickories, and buckeyes.
Type 2: Trees in regions with periodically favorable growth conditions, such as the pines of the
southeastern United States and Central America and Monterey pine in California, typically exhibit
recurrent flushes of growth during the growing season. This type of growth is also the most typical
one among conifer and deciduous trees in the subtropical and tropical regions. The environmental
conditions of the site, especially soil water, determine the number of flushes of growth.
Type 3: In 16 tree genera (Zimmermann and Brown 1971), the shoot tip aborts either following the
extension of the pre-formed leaves (beeches) or after many late leaves have been produced later
in the growing season (black willow, birches, and mimosa). In many cases, shoot-tip abortion is
associated with the shortening days of summer. Shoot extension occurs from the last fully formed
lateral bud the following spring. The ecological significance of shoot-tip abortion is unclear, but
it may occur as a way to continue shoot extension as long as favorable conditions exist.
Type 4: In a few southern species such as tulip tree, sweetgum, and eastern cottonwood, some
shoots grow continuously, including both early and late-leaf development, before setting a
terminal bud. The potential of severe frost damage is low for these species within their range, and
selection pressures for shoot-tip abortion may be weak. These species also occupy moist sites so
that early cessation of growth for drought avoidance is unimportant. Species exhibiting this type
of shoot extension and true terminal bud formation tend to have symmetrical crowns and a more
pronounced main stem than in trees exhibiting shoot-tip abortion.
150 Willow
140
130
120
Yellow–poplar
110
100 Ash
Shoot growth, cm
90
80
70
60
50
40 Hickory
30 Buckeye
20
10
10 30 50 70 90 110 130 150 170 190

Days
F I G U R E 5 . 5 Rate and duration of shoot growth among several woody species in the Georgia Piedmont.
Measurements of shoot elongation were made biweekly on 9 trees ranging from 8 to 15 years in age: painted
buckeye, mockernut hickory, red ash, yellow-poplar (tulip tree), and black willow. Source: Zimmermann and
Brown (1971) / Springer Nature.
Sylleptic and Proleptic Shoots It is common that recently formed buds open and extend
late during the same growing season in which they are formed. This growth occurs laterally and
produces long shoots. Development of a lateral shoot from a lateral bud before the bud is fully
formed is referred to as sylleptic growth, which forms a sylleptic shoot. This pattern is far less
common in temperate regions than in the tropics, although sassafras, alternate-leaf dogwood,
sweetgum, and alders may regularly form sylleptic shoots (Wilson 1984). In prolepsis, lateral buds
at the base of a terminal bud develop lateral shoots after a rest period. Proleptic growth typically
takes place following the winter-dormant period in temperate areas, but lateral meristems of some
temperate species may resume growth briefly after setting buds if late season weather is warm and
moist. These shoots may not adequately harden before winter and may be damaged by freezing
temperatures. Proleptic shoots are common in fast-growing pines (Rudolph 1964). Lammas
growth occurs when the terminal shoot resumes growth following bud set, usually in temperate
areas. When lateral meristems develop late in the season (prolepsis) but the terminal shoot does
not, several laterals may dominate the top of the tree in the next several growing seasons and even-
tually replace the terminal shoot. A heavily forked form is the result as often seen in young, open-
grown white pines (Wilson 1984). The term for lammas growth was coined when late-season
shoots apparently occurred about the time of European holidays, Saint’s Day (June 24), and in
England, Lammas Day (August 1).
ROOTS
Two of the many important functions of tree roots are the firm anchoring of the tree in the soil
and the absorption of water and nutrients. The entire root system serves to anchor the plant,
whereas absorption is accomplished mainly by fine (<0.5 mm), short-lived, non-woody roots.
Larger roots also store carbohydrates and other materials, synthesize organic compounds, trans-
Roots 103
Table 5.1 Comparative root development of seedlings of deciduous and coniferous species.
Species Age (months) Number of roots Total root length (meters) Growing conditions
Black locust 4 7124 325.5 Greenhouse
Loblolly pine 4 419 1.6 Greenhouse
Flowering dogwood 6 2657 51.4 Greenhouse
Loblolly pine 6 767 3.9 Greenhouse
White oak 12 196 2.3 Forest
Loblolly pine 12 148 1.0 Forest
Source: Kozlowski and Scholtes (1948) / with permission of Oxford University Press.
port water and nutrients to the crown, secrete chemicals, and, for some species, generate vegeta-
tive shoots. The growth and functioning of roots have long received attention from many authors,
and recent major reviews include Baluška et al. (1995), Pregitzer and Friend (1996), and
Gašpariková et al. (2001).
Root structure and function differ between angiosperms and conifers. A classic study of root
development demonstrated that deciduous seedlings developed more extensive root systems than
loblolly pine in greenhouse and forest environments (Table 5.1; Kozlowski and Scholtes 1948).
Decomposition of leaves from deciduous trees is rapid (Ellenberg 1988, p. 59), such that the exten-
sive and finely divided root systems of these species may have evolved to efficiently intercept and
absorb the nutrient ions released from decomposing leaves (Kozlowski et al. 1991). Fine roots of
deciduous species are generally smaller in diameter than those of conifers (Voigt 1968) to maxi-
mize surface area and associated nutrient absorption (Pregitzer et al. 2002). By contrast, fine roots
of the more primitive conifers probably have a greater ability to extract nutrient ions from the soil,
especially important where soil nutrient status is poor (Chapter 19).
KINDS, FORMS, AND OCCURRENCE

Tree root systems consist of a framework of relatively large, woody, long-lived roots supporting a
mass of small, short-lived, non-woody absorbing roots, many of which are associated with fungi
(these root–fungal relationships are discussed in Chapter 19). Seedlings of many species develop
tap roots that provide early stability and survival on dry and fire-prone sites. Tap roots are common
in pines and in deciduous species that have large seeds with heavy food reserves (oaks, hickories,
chestnuts, walnuts, etc.). Roots of the main framework grow quickly (up to 1 m per year) for
distances up to 20 m or more (Wilson 1984). Vertical sinker roots may develop from the lateral
system as it develops, reaching into lower soil horizons and helping to anchor the root system.
The form and structure of the root system are largely genetically controlled, but site condi-
tions also influence root depth and extent. Root systems vary widely among species; trees able to
produce deep roots or highly branching roots are able to extract water and nutrients from a larger
soil area than trees with more restricted root systems (Kozlowski 1971b). Thus, species adapted to
droughty sites such as pines and upland oaks and hickories often have deep-penetrating tap roots.
By contrast, root systems are more often shallow where the water table is near the surface as in
swamps or peat bogs, or where deep root growth is restricted by bedrock or dense or poorly aerated
soil. Shallow rooting on such sites often leads to the common uprooting of trees by wind (i.e.,
windthrow). Lateral root extent also may vary with soil, extending further with drier, better aer-
ated, sandy soils compared to clay soils (Rogers and Booth 1960). A sample of the great variety of
rooting forms that develop in relation to different soil and water table conditions is presented in
Figure 5.6.
(a) (b) (c) (d)
(e) (f) (h)
(g)
(i) (j) (k)
F I G U R E 5 . 6 Modification of root systems of forest trees by site. (a) and (b) Taproots with reduced upper
laterals: patterns found in coarse sandy soils underlain by fine-textured substrata. (c) Taproot with long
tassels, a structure induced by extended capillary fringe. (d) Superficial laterals and deep network of fibrous
roots outlining an interlayer of porous materials. (e) Flattened heart-root formed in lacustrine clay over a
sand bed. (f) Plate-shaped root developed in a soil with a reasonably deep groundwater table. (g) Plate-
shaped root formed in organic soils with shallow groundwater table. (h) Bimorphic system of platelike
crown and taproot, found in leached soils with a surface rich in organic matter. (i) Flat root of angiosperms
in strongly leached soils with raw humus at surface and hardpan below. (j) Two parallel plate roots
connected by vertical sinkers in a hardpan spodosol. (k) Pneumatophores of mangrove trees in tidal lands.
Source: Wilde (1958) / John Wiley & Sons.
Aerial roots such as strangling roots and stilt and prop roots are formed by some tropical
trees (Bell 1991). Yellow birch and eastern hemlock approximate this root form in eastern North
America by germinating on decaying logs and stumps and producing roots that grow downward
and into the soil. When the woody debris decays away, the roots are exposed and stilted (Figure 5.7).
Older yellow birch trees are able to generate aerial roots from the stem cambium in humid envi-
ronments such as the southern Appalachian Mountains. These roots develop downward, either on
the outside of the trunk or through the decaying heartwood of the standing tree, and into the soil
(Spurr and Barnes 1980, p. 91).
Roots 105
Photo by Dan Kashian

F I G U R E 5 . 7 Stilted roots of an eastern hemlock in northern Lower Michigan. A seed of this eastern
hemlock germinated on a decaying log. The roots of the growing seedling grew through and around the log
and into the ground. The log has since decayed away, leaving space beneath some of the roots.
FINE ROOT RELATIONS

Lateral roots are formed along the main roots that may become long roots or remain short, such
that the root system consists of several orders of woody, perennial long roots and one or more
orders of small or fine non-woody roots (Figure 5.8). Fine roots (<1 mm) make up very little of the
biomass (weight) of a tree’s root system but most of its length and surface area (Jackson et al. 1997;
Pallardy 2008). In a study of northern red oak, for example, Lyford (1980) found that fine roots
branch and re-branch to form many root orders down to a minimum size of about 0.07 mm
(Figure 5.8). A mature northern red oak (40–70 years old) in Massachusetts would therefore have
90 million fine-root tips in its system, and an average of 1000 fine roots per cubic centimeter of
forest floor! These fine roots were estimated to occupy only 3% of the cube, but have the surface
area of more than 6 cm2, not counting fungal hyphae or root hairs.
In most soils, fine roots are concentrated in the organic and upper mineral soil horizons
(Kozlowski 1971b) where it is well aerated, soil water is most readily available following precip-
itation events, and nutrients are at their highest concentration. A classic study of two pine and
two oak forests of the Piedmont of North Carolina found that 94–97% of the roots <2.5 mm in
diameter were located in the upper 13 cm of soil (Coile 1937). In northern red oak, many roots
arising from the lateral root system grow upward and into the organic layers of the forest floor,
where they proliferate into many small-diameter, non-woody roots (Figure 5.8; Lyford 1980).
This upward growth has also been described for some temperate conifers (Coutts and
Nicoll 1991).
The number of fine roots per unit volume varies greatly during the growing season and by
soil depth. Variation during the growing season occurs in part because fine-root turnover is
extremely rapid. Over 60 years ago, Kalela (1957) reported a continual flux of death and replacement
0 1 2 3 4
METERS
F I G U R E 5 . 8 Diagrams of a horizontal woody third-order lateral root of red oak emphasizing roots that
ascend to the surface and elaborate into many small-diameter, non-woody roots in the forest floor. Top view
of roots (top panel) illustrates proliferation of small roots in forest floor. Side view (bottom panel) illustrates
sinker roots as well as those conforming to the surface topography and ascending into forest floor. Squares
are 1 m on a side. Source: Wilde (1958) / FAO.
of fine roots for Scots pine stands in sandy soils of Finland. Forty-five years later, Hobbie et al.
(2002) found that about 70% of fine-root carbon was new each year in a study of Douglas-fir in the
Pacific Northwest. As much as 33% of annual photosynthate may be allocated to fine roots (Jack-
son et al. 1997). Kalela had concluded that fine roots were sensitive to local ecological conditions,
with roots dying en masse when conditions were unfavorable (drought, cold, lack of photosyn-
thate, etc.), but were rapidly formed again once conditions became favorable. Using this reasoning,
one might predict that fine-root mortality may peak during the winter due to low temperatures
and lack of photosynthate. The degree of fine-root turnover described by Kalela has been con-
firmed in studies of carbon balance and nutrient cycling (Chapters 18 and 19), but the timing of
the turnover has varied in studies over the last half-century. For example, fine roots of ponderosa-
pine seedlings lived the longest when produced in the fall, but longevity was shorter when they
were produced at higher temperatures in the spring and summer (Johnson et al. 2000). Hendrick
and Pregitzer (1992) observed that fine-root mortality in a northern hardwood forest was continu-
ous throughout the growing season, with a peak in autumn. In a beech forest in northern Ger-
many, peak concentration occurred in the spring and in organic matter layers where soil water,
aeration, and mineralization of nutrients were most favorable.
A final point about fine roots is that their branching patterns vary widely among species. A
study of nine North American tree species (Pregitzer et al. 2002) revealed that the majority of fine
roots for most species were <0.5 mm in diameter. Sweetgum roots were thick and unbranched
compared to those of sugar maple or white oak, which were much thinner and more intricately
branched (Figure 5.9). In general, fine roots of deciduous species tend to be smaller and more
finely branched than coniferous species, although all have fine roots dominated by short, thin
individual roots. Besides representing an analog to differences in canopy form and architecture,
Roots 107
Acer Liriodendron Populus

saccharum tulipifera balsamifera
1 cm 1 cm 1 cm
Quercus Juniperus Picea
alba monosperma glauca
1 cm 1 cm 1 cm
Pinus Pinus Pinus
edulis elliottii resinosa
1 cm 1 cm 1 cm
F I G U R E 5 . 9 Digitized representative samples of fine-root-branching patterns of nine North American tree

species. All species include short, thin branches, but deciduous species (except for sweetgum, Liriodendron
tulipifera) tend to have thinner and more finely branched roots compared to conifers. Source: Pregitzer et al.
(2002) / John Wiley & Sons.
these differences in fine-root architectures among species suggest that assuming roots as large as
2 mm function as fine roots across all species may be inaccurate. In many species, roots as small as
1 mm may still primarily serve transport or storage rather than absorption functions; the position
of a root on an individual branch is probably more important for its function than its diameter
(Pregitzer et al. 2002).
HORIZONTAL AND VERTICAL ROOT DEVELOPMENT

Perennial long roots have the potential to horizontally develop a significant distance from the main
stem of the tree. Lateral roots of pines, birches, and black locust may extend 10 or 20 m as they
continually occupy new volumes of soil (Lyr and Hoffmann 1967). In a meta-analysis of rooting
extent, Stone and Kalisz (1991) found that the maximum distance of lateral rooting for most tree
species was less than 30 m; elms, walnuts, oaks, and aspens were able to exceed 30 m, and willows
achieved 40 m. Lateral spread is related to the nature of the rooting medium and is more extreme
in sandy soils than in clay. For example, in a sandy Wisconsin soil, roots of 20-year-old red pine
trees extended about seven times the average height of the trees (Kozlowski 1971b). In a classic
study of the rooting of fruit trees, lateral roots were found to spread about three times as far as the
crown on sandy soils, but only twice as far on loamy soils, and one-and-a-half times as far on clay
soils (Rogers and Booth 1960). The importance of distant roots for soil water and nutrient
absorption appears to be relatively low (Stone and Kalisz 1991).
Rooting depth is particularly important for accessing soil water in deeper soil layers, and
may have survival value for some species in times of drought. Deep-rooting trees found on
prairies, such as certain species of oaks, locusts, and walnuts, are well-documented examples of
the ability of trees to survive drought by accessing deep soil water. Reviewing the literature for 113
temperate tree species, Stone and Kalisz (1991) reported that 70% had rooting depths shallower
than 5 m and 33% shallower than 3 m. Although rooting depths of 2–3 m are not uncommon for
many forest trees, deep-rooting species (rooting depths of 3–6 m or more) include eastern redce-
dar, black and honey locust, bur oak, and osage-orange (Kozlowski 1971b). Extraordinary rooting
depths of 61 m have been reported for one-seed juniper by Cannon (1960), who located them in
uranium mines on the Colorado Plateau. Roots of the desert shrub mesquite were found 53 m
below the surface in southern Arizona (Phillips 1963). Stone and Kalisz (1991) cited various
studies reporting rooting as deep as 8 m for locust, 10 m for apple, and 24 m for oak and pon-
derosa pine.
PERIODICITY OF PRIMARY ROOT GROWTH

Much like shoots, extension growth of roots is intermittent. For many temperate tree species, root
growth initiates earlier in the spring and proceeds later into the fall than growth of shoots. Spe-
cifically, root growth tends to peak when soil temperature and moisture conditions are most
favorable in the spring and fall. Root growth often decreases during summer drought periods and
slows or stops with the onset of winter and low soil temperatures. In a study of root and shoot
growth of red and sugar maple in Virginia, root growth began in early March, about one month
before budbreak (Harris and Fanelli 1999). Root growth rate decreased sharply at budbreak, but
resumed and continued as shoots grew. Root growth continued even after leaf senescence
occurred, until soil temperatures deceased below 7 °C (45 °F) in the fall (Figure 5.10). Many
exceptions to these trends have been observed, however. For example, short-shoot needles expand
before root growth begins in larch species (Lyr and Hoffmann 1967), and a series of horticultural
trees in New York State began root growth only after shoot growth was initiated (Harris
et al. 1995).
Roots tend to cease growth during the winter, but do not enter a dormant phase as shoots
do within winter buds. Rather, roots of many species may grow during the winter if tempera-
ture, soil water, and aeration are favorable (Lyford and Wilson 1966; Lyr and Hoffmann 1967).
Winter root growth has been reported from regions where winters are mild and soils are frost-
free, such as in the southern United States, coastal British Columbia, Crimea, and in parts of
Europe. Woods (1957) reported considerable growth of longleaf pine roots in the surface layer of
sandy soils of the southeastern coastal plain of the United States during the winter, and Kramer
(1969) reported root growth during every month of the year for other pine species in the
same region.
ROOT GRAFTING
Natural root grafting among individuals of a species is relatively common in woody plants
(Graham and Bormann 1966), especially in pure stands (commonly pines, spruces, and other
conifers) that exhibit a high concentration of roots near the soil surface. Grafting occurs when
the cambium layers of two closely associated roots fuse together, thus connecting their phloem
and xylem tissues, allowing the transport of carbohydrates, growth substances, and pathogens
between the two trees. For example, root grafts among shade-intolerant lodgepole pines have
been shown to support shaded trees that would have otherwise been outcompeted via the trans-
mission of carbohydrate (Fraser et al. 2006). Water transport between two trees is less common
and requires a more complete fusion of root tissue. However, a study in red pine plantations
observed the survival of girdled trees for 18 years due to continued water transport supported by
root grafting between girdled and ungirdled stems (Stone 1974). Grafting has proposed to have
Roots 109
Acer rubrum
4 Shoots
Roots
2
Budbreak
% Increase per day
4 Acer saccharum
Shoots
Roots
Substrate temp.
3
2
Budbreak
% Increase per day
0 0
5
Temperature (C)
1
10
2 15
20
3
25
4 30
1 1 1
Ja 1
M M 1
Ju
1
Se
1
N Ja 1 M 1 M 1
Ju
n ar ay ly pt ov n ar ay ly
95 96
F I G U R E 5.1 0 Comparison of weekly extension for shoots and growing roots of red and sugar maple over a
14-month period. Root and shoot growth follow distinct but different seasonal patterns. The growth of roots
for both species initiates before shoots and continues longer into the growing season. Source: Harris and
Fanelli (1999) / with permission of Horticultural Research Institute.
some adaptive significance, such as increased mechanical support or the formation of beneficial
communal root systems, but relatively little empirical data exist to suggest that it occurs in
response to more than incidental conditions (Loehle and Jones 1990; Lev-Yadun and
Sprugel 2011).
Root grafting plays a critical role in the transmission of systemic diseases such as Dutch elm
disease and oak wilt (caused by Ophiostoma ulmi or Ophiostoma novo-ulmi and Bretziella
fagacearum, respectively) and in root rots caused by Armillaria mellea, Heterobasidion annosum,
and Phellinus weirii. In American elm, the roots of nearly all trees within 2 m of one another and
about 30% of those within 8 m are grafted (Verrall and Graham 1935; Himelick and Neely 1962).
This interconnection permits rapid transmission of the fungus from infected to uninfected trees
through the vessels of the roots. In the root rots, the fungus spreads in part by growing and decay-
ing its way between adjacent trees via root grafts (as in H. annosum).
SPECIALIZED ROOTS AND BUTTRESSES

Specialized roots may develop in relation to site conditions. Perhaps the best example of such
structures are the pneumatophores (roots that function as a respiratory organ; Figure 5.6k) pro-
duced by various species in periodically flooded river valleys or subtropical and tropical swamps
(Kozlowski 1971b). In bald cypress, “knee roots” are formed at the air–water interface in flooded
sites, but do not form when bald cypress trees are grown in upland, non-flooded sites unless they
are injured. The aerial portion of the root exhibits high cambial activity compared to the sub-
merged portion of the root. These knee roots are not important in root aeration in bald cypress, but
may strengthen the roots for firmer anchorage (Gill 1970). Pneumatophores in other species may
contain lenticels and facilitate the exchange of gasses with the roots.
Buttresses—flattened, vertical plates that develop at the base of the stem—serve to hold the
tree upright when site conditions limit the development of suitable vertical root growth. They are
most common in tropical species, but occur in temperate trees such as American elm and Lom-
bardy poplar. In bald cypress and water tupelo of the southern United States, buttresses develop as
a response to the level of flooding (Varnell 1998). On flooded sites, buttresses form because (i)
ethylene accumulates at the waterline, and (ii) growth regulators are unable to translocate beyond
the waterline. As a result, the number or size of individual cells increases dramatically at this loca-
tion on the stem (a process called hypertrophy) where the trunk is neither completely aerated nor
completed saturated. Stem diameter at any height therefore represents the frequency of occurrence
of flooding at that height.
Buttressing may also develop from mechanical stresses. Based on studies of 191 trees of
African whitewood in Ghana, Johnson (1972) reported that small buttresses were found on well-
drained upland soils, whereas large buttresses occurred on poorly drained mid- or lower-slope
sites. When soil conditions permit deep rooting, the vertical root system is able to absorb stresses
that would otherwise concentrate at the lateral root–stem juncture. By contrast, limited vertical
root development due to impermeable layers or a high water table concentrate stresses at the stem
base, thus inducing marked buttress growth. Wind and associated swaying may also result in but-
tressing. For example, a study of Sitka spruce by Nicoll and Ray (1996) documented considerable
buttressing on the leeward side of spruce stems compared to the windward sides to support the
stem against prevailing winds. Notably, entire spruce root systems also had more structural root
mass on the leeward side than the windward side of the tree.
STEMS
Trees and shrubs are called “woody” because they form consecutive layers of structural tissues to
the primary stem. These layers are termed secondary growth, and they strengthen the stem,
support the crown, and increase the transport of food, water, nutrients, and hormones between the
shoots and the roots. A meristematic sheath of cells called the vascular cambium surrounds the
stem, shoots, and roots (Figure 5.11). It creates the successive layers of secondary growth and con-
sists of a zone of actively dividing cells. The vascular cambium builds tissues in two directions:
toward the center of the tree and toward the outside of the tree. Toward the center of the tree, the
cambium gives rise to xylem, which conducts water and the nutrients dissolved in it, typically
Stems 111
Photo courtesy of Champion International Paper Company

Outer bark
Inner bark
Cambium
Sapwood
Heartwood
F I G U R E 5.1 1 Generalized structure of a tree trunk (i.e., a stem) showing the position of major tissues.
Tissues include the outer bark (dead and compressed phloem cells), the inner bark (living phloem-food-
conducting cells), cambium (sheath of living cells that give rise to phloem and xylem cells), sapwood (outer
band of xylem-water-conducting tissue), and heartwood (inner core of xylem-water-conducting cells).
from the roots toward the crown. The xylem cells become lignified and form the dead, woody axis
of the tree. It is remarkable that the indispensable function of conducting water rapidly over long
distances necessarily requires dead cells (Zimmermann 1971). Dead xylem also provides the struc-
ture for height growth that distinguishes forests from prairie and tundra. There are two types of
xylem—the central core, or heartwood, and the outer portion, or sapwood (Figure 5.11). Most
water is conducted through sapwood, but heartwood is purely structural and readily decayed in old
trees. Heartwood is often much denser than sapwood (Woodcock and Shier 2002). Huge, com-
pletely hollow trees (for example, basswood, American beech, sycamore, etc.) may live for decades
supported only by a thin shell of phloem, cambium, and sapwood.
Toward the outside of the tree, the cambium generates the phloem or inner bark, which
conducts photosynthate created in the leaves to the rest of the plant. Phloem cells are not lignified
and function for only a few years before collapsing and becoming part of the outer bark (Fig-
ure 5.11). Phloem cells are therefore as regularly constructed as xylem cells, though they do not
accumulate. The anatomies of the xylem, phloem, and bark are summarized by Esau (1977) and
Bell (1991), and the physiology by Zimmermann and Brown (1971) and Pallardy (2008).
XYLEM CELLS AND GROWTH RINGS

Diameter growth of trees results from the annual formation of layers of xylem cells by the cambium.
In conifers, the primary water-conducting cells are the vertically oriented (perpendicular to the
F I G U R E 5. 12 Transverse section of wood of ponderosa pine

showing a transition in size and shape of earlywood (E) and
latewood (L) in successive growth rings. Source: Zimmermann
and Brown, 1971; reproduced with permission from Springer
Nature.; Photo by Claud L. Brown.
ground in the stem), overlapping tracheids (shown in cross section in Figure 5.12). These cells are
thick-walled, tapered, 3–5 mm long and up to 100 times longer than wide (Kozlowski 1971b), and
arranged in uniform radial rows in conifers. Parenchyma cells and ray tracheids are oriented hori-
zontally (parallel to the ground in the stem) and transport water, food, and other substances later-
ally. Perforations in the walls of the tracheids allow water and other substances to transfer among
adjacent cells. Conifer growth rings are distinguished by differences in cell diameter and the cell-
wall thickness of the tracheids. Thus, an annual growth ring consists of both a light and a dark
portion, with the dark portion produced most recently (Figure 5.12).
In angiosperms, xylem cells are both more specialized and more diverse, including vessels,
tracheids, fibers, and parenchyma. The majority of the xylem cells in angiosperms are fibers, but
cells called vessels are those that conduct most of the water through the stem. Vessels are single
cells whose end walls have many perforations, such that many cells lined up end to end act as tubes
that may be up to several meters long. This system of vessels allows rapid transport of water and
nutrients to the shoots. In ring-porous angiosperms (including ashes, hickories, oaks, elms, and
black locust), vessels formed early in the growing season have much larger diameters (up to 100
times as great) than those formed late in the growing season (Figure 5.13). Growth rings of
ring-porous species are easily distinguished because of the large vessels in the earlywood.
Diffuse-porous species (such as aspens, willows, birches, basswood, and maples) have smaller
vessels that are more uniformly distributed throughout the growth ring (Figure 5.14). Growth
rings of diffuse-porous trees are not as clearly distinguished as those of ring-porous trees.
PERIODICITY AND CONTROL OF SECONDARY GROWTH

In north temperate species, annual diameter growth typically results from the growth of a single
annual ring. The widths of annual rings vary among species, tree ages, and site conditions. Where
tree growth is affected more by weather than site conditions, secondary growth is sharply reflective
Stems 113
F I G U R E 5 .1 3 Transverse section of wood of a ring-porous

black oak showing very large vessels in the first formed
earlywood (E) and the preponderance of fibers in the latewood
(L). Source: with permission from Springer Nature.; Photo by
Claud L. Brown.
of temperature and/or precipitation during the growing season when the ring was generated.
Where the growing season is very short, such as in subarctic regions and high elevations, temper-
ature is often more limiting than precipitation (Kozlowski et al. 1991). By contrast, growth may be
nearly continuous throughout the year in tropical and subtropical areas, and many tropical trees
lack annual growth rings.
Most secondary growth occurs in temperate areas during the spring and early summer
when soil conditions are favorable, but the duration of the growth period varies widely among
species. For 21 tree species on one site in the Georgia Piedmont, the length of the most rapid
period of growth varied from 70 to 209 days (Jackson 1952). In a study of European species,
Norway spruce, larch, and ash formed growth rings rapidly, with little wood laid down after July
(Ladefoged 1952). Species of birch, beech, alder, maple, and oak laid down wood from May until
early September, with up to one-third of the ring being formed in the late summer. For seven
boreal tree species in northern Quebec, radial growth was similar among species but did not
appreciably begin until early June, explained mainly by increasing soil and air temperature (Tar-
dif et al. 2001).
The initiation of cambial growth in the spring occurs when growth-regulating hormones,
auxin, and other substances (promoters and inhibitors) from expanding buds and leaves trigger
activity. In conifers, cambial activity begins in the crown at the base of actively developing buds
and leafy shoots prior to extension and moves rapidly downward throughout the shoots, branches,
and the main stem. The process is similar but far slower in diffuse-porous hardwoods. Cambial
activation is almost simultaneous throughout the entire stem in ring-porous species, which is an
important adaptation for forming new vessels to replace those becoming nonfunctional from the
previous year. The newly formed vessels are vital in ring-porous species to supply water to develop-
ing foliage.
F I G U R E 5. 14 Transverse section of a diffuse-porous

hardwood, tulip tree, showing fairly uniform distribu-
tion of vessels throughout the growth ring and radial
flattening of the last formed latewood cells. Source:
Zimmermann and Brown, 1971; reproduced with
permission from Springer Nature.; Photo by
Claud L. Brown.
Control of Earlywood and Latewood Formation Much research has focused upon the
formation of earlywood and latewood and their properties and physiology, particularly because of
their importance to pulpwood and timber quality. Earlywood is formed by rapid cell division early
in the growing season; cell enlargement is prolonged, and the duration of cell wall thickening is
short, creating larger tracheids with walls only thick enough to avoid cell implosion (Hacke
et al. 2001; Pratt et al. 2007). These earlywood cells supply most of the demand of the crown for
water. Latewood is formed when the rate of cell division decreases later in the growing season, the
cell enlargement duration period is shortened, and cell walls thicken for a longer period (Cuny
et al. 2014). The narrow diameter of these cells limits their utility for water transport (Sperry
et al. 2006). As a result, latewood tends to be denser than earlywood because it contains thicker cell
walls per unit volume. The specific gravity of latewood in conifers is about two to three times that
of earlywood (Kozlowski 1971b), but the density of both earlywood and latewood varies among
years. The density of earlywood in a given year is related to temperature in the previous growing
season, specifically as it relates to photosynthetic activity and its impact on cell size. By contrast,
latewood density in a given year is related to early spring temperatures in the same year, specifi-
cally as it relates to the activity of the cambium and thickness of the cell walls (Björklund
et al. 2017). The type of cells produced (earlywood or latewood) is also likely to be influenced by
the balance of growth hormones, especially as they interact with environmental factors, but their
production is unlikely to be controlled by hormonal signaling alone (Buttò et al. 2020).
Stems 115
Dendrochronology, or the study of tree rings, has long been used to assess global climate
change because tree rings are able to record the natural variation of the past climatic record. For
example, both ring width and maximum wood density of the rings have been used to reconstruct
changes in average summer temperatures over the last 6000 years in northern Sweden, Finland,
and Siberia (Pearce 1996). At these high latitudes (inside the Arctic Circle and near the northern
limit of Scots pine), small changes in average summer temperature greatly affect the number of
days warmer than the threshold for radial growth (5 °C/41 °F). The width and density of tree rings,
and particularly of latewood cells within the rings, dramatically reflect these changes in tempera-
ture. Thus, understanding the structure of wood is at the forefront of distinguishing long- and
short-term patterns of climate change. Accurately understanding and interpreting patterns of tree-
ring width and density also permits crossdating ring samples from a tree, whereby each individual
tree ring is assigned its exact year of formation by matching patterns of wide and narrow rings
between samples from the same tree and between trees from different locations. The body of
research utilizing dendrochronology to study forests in North America and elsewhere is enormous
and well beyond our scope. The science of dendrochronology is treated extensively by Trouet
(2020), Amoroso et al. (2017), Speer (2010), Cook and Kairiukstis (1990), Schweingruber (1989),
and Fritts (1976).
WINTER FREEZING AND WATER TRANSPORT

Winter freezing of water in xylem cells is problematic for water conduction in trees of northern
latitudes, particularly for ring-porous species with very large earlywood vessels. The presence of
ice is less of an issue than dissolved gasses in liquid water, which form bubbles when frozen. When
the water thaws, the bubbles create gaps or embolisms in the water columns, breaking their conti-
nuity and forming an air lock against transpiration. Zimmermann and Brown (1971) proposed
three coping mechanisms of trees in this situation:
1. Small-diameter xylem cells: Very small conducting cells in some species may form
bubbles small enough to redissolve when the water thaws, rejoining the vertical water
column before transpiration begins. Conifers (firs, spruces, etc.) and certain diffuse-porous
trees with very small vessels (alders, willows, trembling aspen, paper birch, etc.) are thus
easily able to survive the extreme freezing and thawing regime of the boreal or alpine forest.
Notably, these species are common in boreal and alpine regions. No ring-porous species are
boreal, and most (such as oaks, hickories, and walnuts) are most abundant in the central or
southern portions of North America.
2. Root pressure: Some diffuse-porous species such as birches may develop enough root
pressure to fill the air gaps formed in small vessels.
3. Spring formation of large vessels: Ring-porous species that have large earlywood vessels
conduct much of their water within the most recent growth ring. These same species are
able to very quickly activate their cambium and produce new vessels prior to leaf flush in
the spring.
Spring vessel formation is advantageous because water transport through large vessels is
very rapid upon leaf flush as it is required by the developing and transpiring leaves. However,
many species within this group, such as oaks, hickories, walnuts, and ashes, must delay their flush-
ing in the spring until the xylem system is ready to supply the water required by the foliage. Delayed
flushing has the disadvantage that the leaves must develop very rapidly in the warm temperatures
of late spring when transpiration is high, and that a delay in flushing ultimately results in loss of
growth relative to those species that flush earlier. Spring vessel formation is also advantageous
because late-flushing species have a greater probability of escaping late frost than early flushing
species. An additional disadvantage, however, is the dependence on the most recent ring for water
conduction, because injury or disease to the ring would cause significant damage to the tree.
American chestnut and American elm are examples of species that have been dramatically affected
by diseases that block their water vessels (see Chapter 21).
WATER DEFICITS AND TREE GROWTH

Tree growth responds more to water stress than any other perennial factor of the forest site, and
thus forest-site productivity depends heavily on soil-water availability in many parts of the world.
Environmental soil-water stress affects height and diameter growth, flowering and fruiting, seed-
ling germination and establishment (Chapter 4), and virtually every aspect of plant morphology
and physiology (see also Chapters 9, 11, 18, and 19). A definitive review of soil-water deficit is by
Zahner (1968), with more recent reviews by Kramer and Boyer (1995) and Pallardy (2008).
In temperate climates, soil-water deficits tend to occur during the middle of the growing sea-
son when air temperatures are high, soil water replenished by winter snowmelt and spring rains is
no longer present, and fully flushed and growing forests have removed soil water via transpiration.
The effect of soil-water stress on height growth is seen clearly in the low heights that trees attain in
dry climates or on dry sites compared with those growing on moist sites, as well as reduced shoot
growth of seedlings and saplings under soil-water stress.
Effects of soil-water stress vary among species, and they depend partly on the species’
seasonal pattern of shoot elongation. Species that maintain shoot elongation throughout the
growing season such as birches, tulip tree, and loblolly pine are affected by late-season droughts.
Species that complete height growth and set buds by midsummer, including red pine, eastern and
western white pine, white ash, sugar maple, American beech, and northern red and white oaks,
are unaffected by late-summer droughts of the current year. However, a late-summer water deficit
in the current year will affect height growth during the following growing season. This pattern
occurs because photosynthate accumulates following bud set and is stored for use in the next
growing season, but the amount of photosynthate produced and stored will be reduced under
drought conditions. For example, 72% of the variation in annual shoot growth in young red pine
was explained by water deficits of both the previous and current growing seasons together (Zahner
and Stage 1966). However, height growth of tree species that exhibit shoot elongation throughout
the growing season is unaffected by soil-water conditions in the previous year (Zahner 1968).
Diameter growth is also affected by soil-water stress. Periodic changes (daily and annual) in
the radial growth of conifers and angiosperms are directly related to soil-water availability
(Figure 5.15). For example, the period of highest radial growth for seven boreal species in northern
Quebec was explained by precipitation and subsequent soil-water availability (Tardif et al. 2001).
In a study of deciduous forests in Indiana, 13 of 20 species exhibited decreased radial growth in
response to soil-water stress; the 7 remaining species unaffected by drought were classified as xero-
phytic rather than mesophytic (Brzostek et al. 2014). Water absorption by roots may lag behind
transpiration later in the growing season in temperate regions, dehydrating crown tissues, slowing
photosynthesis, and limiting growth below the potential for that time of year (Zahner 1968).
Normal radial growth usually resumes with late-season rains if the mid-season water deficit has
not been severe.
Plants resist drought by avoiding and tolerating drought stress (Levitt 1980b). Woody plants
mainly avoid drought stress using morphological and anatomical adaptations that maintain
relatively high tissue-water content in environments where water is limited by minimizing water
loss and maximizing water uptake. Major avoidance mechanisms include: (i) the ability of roots to
extract large amounts of water from the soil; (ii) high root-to-shoot ratio; (iii) the ability to reduce
leaf surface area (including rolling, folding, and shedding of leaves, and maintaining a dense
Water Deficits and Tree Growth 117
0.002
1955
Basal area
1952
Basal Area Cumulative m2
0.001
0
15 14 14 13 13 12 11 11 11
Mar. Apr. May June July Aug. Sept. Oct. Nov.
20
Centimeters of Available Water
15
10
1955
Soil Water
5
1952
0
15 14 14 13 13 12 11 11 11
Mar. Apr. May June July Aug. Sept. Oct. Nov.
F I G U R E 5.1 5 Basal area growth per tree for shortleaf pine and trends of available moisture for relatively
wet (1955) and dry (1952) growing seasons. Note that the growth rate slowed in mid-June of 1952 but did not
slow until mid-August of 1955, at the time in each year when available soil water had been depleted to about
5 cm or below. Source: Zahner (1968) / Elsevier.
pyramidal crown); (iv) the ability to reduce transpiration via stomatal control—closing stomata
most of the day and/or the ability to close stomata very rapidly in response to stress; (v) thick cuti-
cle to reduce cuticular transpiration, and/or (vi) a high proportion of water-conducting tissue to
nonconducting tissue. The ability of trees to absorb large quantities of water may be more impor-
tant for drought avoidance than reducing water loss through transpiration. For example, drought
avoidance by stomatal closure has the disadvantage of reduced photosynthetic rates, but species
efficient at absorbing water can maintain high tissue-water content even when stomata open for
photosynthesis and transpiration occurs. Drought tolerance is the ability of plants to endure low
tissue-water content through adaptive traits that maintain cell turgor using osmotic adjustment
and cellular elasticity. Drought tolerance is less common than drought avoidance in woody plants,
but many species withstand considerable dehydration before stomatal closure. Woody plants are
heavily influenced by many other site factors in addition to drought, and these are discussed
in Part 3.
SUGGESTED
READINGS
Barthelemy, D., Edelin, C., and Halle, F. (1991). Canopy Pallardy, S.G. (2008). Physiology of Woody Plants, 3e. San
architecture. In: Physiology of Trees (ed. A.S. Raghav- Diego, CA: Academic Press 454 pp.
endra), 1–20. New York: John Wiley. Pregitzer, K.S., DeForest, J.L., Burton, A.J. et al. (2002).
Hallé, F., Oldeman, R.R.A.A., and Tomlinson, P.B. Fine root architecture of nine North American trees.
(1978). Tropical Trees and Forests. Berlin: Springer- Ecol. Monogr. 72: 293–309.
Verlag 441 pp. Tomlinson, P.B. (1983). Tree architecture. Am. Scientist.
Kozlowski, T.T. (1982). Water supply and tree growth. 71: 141–149.
Part I. Water deficits. For. Abstr. 43: 57–95. Zimmermann, M.H. and Brown, C.L. (1971). Trees:
Kramer, P.J. and Boyer, J.S. (1995). Water Relations of Structure and Function. New York: Springer-
Plants and Soils. New York: Academic Press 495 pp. Verlag 336 pp.
Lyford, W. H. (1980). Development of the root system of
northern red oak (Quercus rubra L.) Harvard Forest
Paper No. 21:1–30.
The Physical PA R T 3
Environment
FOREST ENVIRONMENT
We emphasized in Part 1 that ecosystems integrate factors of climate, landforms, soils,
biota, and all their interrelationships. This integration defines ecosystems as more than
simply organisms (biotic) and their environments (abiotic). Although integrating these
ecosystem components is our goal, it is useful to examine the physical factors of eco-
systems as they affect plant distribution and processes. The sum total of the physical
factors of an ecosystem determines the forest site. Importantly, plants and animals that
occupy a site modify the local physical environment. For example, decomposing organic
matter forms soil horizons as it interacts with mineral soil particles and water. The fact
that abiotic and biotic factors are so heavily integrated emphasizes the difficulty of treat-
ing factors separately, despite the usefulness and simplicity of such an approach. The
chapters in this section fall within the subdivision of ecology known as autecology, the
study of organismal relationships to environment—in this case those of forest trees or
other organisms.
We examine in this section the physical factors that influence plants where they
live. The forest site is the place where plants live—defined by a specific geographic loca-
tion and by the suite of physiographic, edaphic, hydrologic, and climatic factors that sus-
tain the plants. To varying degrees, these abiotic factors are modified by the plants and
animals of the locality. A given physical site factor may directly affect plants and other
forest organisms, but more often its action is affected, intensified, or diminished by inter-
actions with biotic factors and other physical factors.
The term habitat is widely used in the literature, conventionally when the major focus
is on organisms. For example, “the white-tailed deer’s habitat is the northern white-cedar
swamp” (as if the swamp belonged to the deer), where typically the vegetation that pro-
vides shelter and food defines the habitat. The white-cedar’s habitat, however, is defined by
physical site factors including air, soil, and drainage. The term habitat may therefore take
on several meanings. Because our focus is on ecosystems rather than organisms, we use the
preferred term site (or habitat supporting plants) whenever possible to refer to the physical
environment (as modified by biotic effects) of landscape ecosystems.
119
0005430838.INDD 119 01-05-2023 17:33:06

120 PA RT 3 The Physical Environment
SITE FACTORS
It may be fairly easy to enumerate the physical factors that support the existence and
growth of forest organisms. However, it is exceedingly difficult to understand and eval-
uate the sum total of the interactions among the physical factors and their interactions
with biotic factors that make up the complex we term site.
The site supplies a set of very basic factors. The tree grows with its crown in the air,
gathering light, warmth, carbon dioxide, and oxygen. The tree also grows with its roots
in the soil, gathering the mineral nutrients and water necessary for photosynthesis and
other processes. The availability of each of these factors to the tree, however, depends
upon an endless system of changing climate, day length, and soil development—changes
that are in part related to the developing vegetation and associated animals.
ORGANIZATION OF SITE FACTORS

Biotic and physical site factors may be organized into broad groups that may be con-
sidered separately. The following diagram shows that site factors interact to yield the
light, heat, water, etc. that are directly available and used by the plant. Physical and biotic
factors have multiple effects on the plant.
PHYSICAL SITE FACTORS FACTORS DIRECTLY

AVAILABLE TO THE PLANT
CLIMATE
radiation, air temperature, LIGHT
rainfall, relative humidity, fog, energy for photosynthesis
wind, lightning, etc.
HEAT
energy for metabolic
PHYSIOGRAPHY
processes
landform, parent material,
aspect, slope position
WATER
maintains cell function PLANT
SOIL transports nutrients
texture, structure, nutrients,
organic matter, drainage
CHEMICAL FACTORS
pH, pollutants, nutrients
BIOTIC FACTORS
associated plants, animals and DISTURBANCE FACTORS
microorganisms above - and damage through wind, fire,
belowground glaze, animals, human activities
Relationships between site factors directly and indirectly responsible for plant distribution and growth.
Source: Modified from Ellenberg 1968. Reprinted from Wege der Geobotanik zum Verstdändnis der
Pflanzendecke © by Springer-Verlag Berlin, Heidelberg 1968. Reprinted with permission of Springer-Verlag
New York, Inc.
Climate is an overriding driver of ecosystems on the Earth. The physical factors

of climate related to the atmosphere affect the aboveground growth and persistence of
plants, such as solar radiation, air temperature, air humidity, wind, lightning, and the
0005430838.INDD 120 01-05-2023 17:33:07

Site Factors 121
CO2 content of the air. Climate also determines belowground temperature, moisture (via
precipitation), CO2, and the weathering of nutrients from rock substrates. Climate factors
are repeatedly discussed and integrated into nearly every chapter of this book. The recep-
tion of light and the effects of light on plant growth are given special consideration in
Chapter 6, as is the influence of temperature in Chapter 7.
Physiographic and soil factors (Chapters 8 and 9) include the topographic and
structural features of landscapes together with all the physical, chemical, and biological
properties of the soil. The relief and form of the land, the nature of the parent material,
physical and chemical properties of soil, and soil biota are particularly important in plant
growth and in determining site quality.
Biotic activities, both visible and microscopic, change climate, physiography, and
soil—and therefore affect site quality. Large organisms (such as trees, grazing animals,
and humans) create the most obvious changes in the microclimate and the soil. Small
organisms occurring in great numbers (including fungi, bacteria, earthworms, rodents,
and many others) can also substantially change the site. Because of the influence of
biota on site, a discussion on various biotic factors is integrated into most of the chap-
ters in this part. In addition, the roles of animals on site quality are discussed later in
Chapter 12 (Part 4).
Fire is also a critical factor affecting forest species and site quality, and it is consid-
ered as a physical force in Chapter 10. For example, the burning of soil organic matter
and the heating of the surface horizons of mineral soil result in changes in the physical
and chemical properties of the soil and its soil biota.
Once the individual factors affecting forest site have been considered, it is necessary
to integrate them into a whole if we are truly interested in accurately characterizing the
forest site. Approaches to the determination of forest site quality and the evaluation of
ecosystems are considered in Chapter 11.
INTERRELATIONSHIPS AMONG SITE FACTORS

It is both conventional and convenient to discuss the various factors of the forest site one
at a time, but we approach such a discussion with caution. The consideration of factors
individually does not imply that any is an independent force affecting forest biota. It is
easy to study ecology in terms of simple direct relationships, making such statements
as “at a temperature of 55 °C, the plant suffers direct heat injury” or “precipitation
limits the distribution of this species.” Such statements ignore the reality that plants
live in the total complex of the environment, where a change in any one factor may
well bring about a changed requirement or tolerance of the plant for other factors. The
individual environmental factors are not isolated, independent forces operating on the
plant but are interdependent and interrelated influences that in the end must be con-
sidered together.
0005430838.INDD 121 01-05-2023 17:33:07

122 PA RT 3 The Physical Environment
The law of the minimum holds that the rate of growth of an organism is controlled
by that factor available in the smallest amount—often referred to as the “limiting factor.”
Ecologists have restated the law of the minimum to account for the interactions existing
between the various single factors that operate at different scales. For instance, upon
observing that seedlings die under dense forest canopies, an ecologist may conclude that
low light intensity is responsible for the mortality. In fact, many environmental factors
are affected by dense forest cover. Light intensity is certainly low, but so also is the supply
of water in a soil permeated by the roots of the many trees present. The temperature
regime and wind speed under a dense canopy are greatly changed from those in the open
or even under an open forest. The organic matter and the soil microorganisms, too, will
be greatly affected. In short, the whole environmental complex is related to forest can-
opy cover, and it is misleading, if not downright incorrect, to attribute changes in plant
response to any single factor (Chapin et al. 1987).
IMPORTANCE OF SITE IN FOREST ECOLOGY

In this book, we consider site before biota for good reason. In a forest ecosystem, the
site is more concrete, more stable, and more easily defined than the biota that occupy it.
Therefore, perhaps counterintuitively, the site constitutes a better basis for the descrip-
tion of the forest ecosystem than do the trees, other plants, and animals. The emphasis
on earth science as the basis for ecology is seen in the books on geoecology by Matthews
(1992) and Huggett (1995).
The forest site occupies a fixed geographical area and is capable of fairly precise def-
inition, especially using physiographic features or landforms. In contrast, various types of
forest communities may develop on a particular site type, depending upon chance, past
history, or changing environmental conditions (Chapter 1). Moreover, several quite dif-
ferent forest communities may be characterized by a given species or even a given group
of species. Ideally, it is the combination of site and biota, the landscape ecosystem itself,
that is necessary in describing forest landscapes and in understanding how forest ecosys-
tems function.
0005430838.INDD 122 01-05-2023 17:33:07

Light CHAPTER 6
E arth’s ecological systems are driven by the energy provided by light. Light energy is used
in photosynthesis, and its signals are used in the photoregulation of plant growth and
development. Photosynthesis is carried out primarily in the chloroplasts of higher plants as indi-
cated in this simplified equation:
CO2 H 2O light energy

(CH 2O)n O2
Carbon Water Glucose Oxygen
This process features photons, packets of light energy, that excite electrons and move them
from a lower to higher energy state. Green plant chloroplasts have a reaction center containing two
molecules of chlorophyll, the green pigment that captures light energy. When the chloroplast
absorbs a photon of light, an electron is transferred out of the reaction center, leaving the system
with enough energy to split water in the cell and release its oxygen:
2H 2O O2 H 4e
The hydrogen ions and the electrons released in the cell bind to carbon dioxide to form
sugar molecules, the source of energy used in respiration. Respiration is the process by which
tissue biosynthesis and maintenance occur (Chapter 18). In the simplest terms, this process is
how light energy is converted to and stored as chemical energy in photosynthetic products. Pho-
tosynthesis is the initial, basic step in the transfer of energy from light to the food chains of the
ecosphere.
In addition to using light energy, plants respond and adapt to changes in environmental
factors using the signals provided by light. Photosensory systems of plants as well as animals,
fungi, and microorganisms acquire and process information about light direction, duration,
intensity, and spectral quality. In higher plants, photoregulation occurs at all phases of the plant
life cycle, from seed formation and germination to the genetic adaptation of populations to their
site conditions (Chapters 3 and 4). Dormancy is initiated by seasonal changes in light quality,
and daily changes in light conditions elicit leaf orientation and chloroplast distribution within
leaf cells.
In this chapter, we examine the distribution of light at the Earth’s surface, the interception
of light by plant canopies and its use in photosynthesis, and the effects of light on tree growth and
leaf morphology. In Chapter 18, we focus on the ecological aspects of carbon fixation and the allo-
cation of photosynthate to growth, maintenance, storage, and defense. These processes have
important implications for the cycling and storage of carbon in forest ecosystems.
123
124 Chapter 6 Light
DISTRIBUTION OF LIGHT REACHING THE ECOSPHERE

Solar radiation may be reflected or absorbed by nearly anything in the atmosphere, such as clouds,
smoke, ozone, smog, or greenhouse gases. Radiation reaching the Earth’s surface on overcast days
may be one-quarter to one-half of that for cloudless days under the same conditions. Most radia-
tion under overcast conditions is scattered by the atmosphere and other interceptors, and then
reaches the ground indirectly. The total amount of radiation reaching the surface is even further
diminished by absorption and scattering when the sun is lower in the sky.
Total solar radiation per unit surface area is not the same thing as radiation in the visible
range, or what we perceive as light. Energy is measured in Joules (J), or when integrated over time,
watts (1 W = 1 J s−1). Solar radiation is divided into the ultraviolet, the visible, and the infrared spec-
tra. Wavelengths at the ultraviolet end of the spectrum (100–400 nm) are shorter and nearly com-
pletely absorbed by the atmosphere. Long wavelengths fall into the infrared spectrum (700–1000 nm),
which contains about 50% of total radiation. Most of the radiation arriving at the Earth’s surface is
shortwave radiation with wavelengths shorter than 200 nm. Radiation emitted from the Earth into
space—including the reflection of about 30% of incoming radiation known as albedo—is mostly
longwave radiation, or radiation in the infrared region.
Radiation with wavelengths between 400 and 700 nm—a very narrow portion of total solar
radiation—is visible to the human eye. Light in many historical studies in ecology was based on
measurements that represent the amount of light detectable by the human eye—called illumi-
nance. Illuminance is commonly expressed in the English system as foot-candles and in the metric
system as luxes. Photosynthetically active radiation (PAR) of visible light is a measure of the
total number of photons within the visible spectrum striking a surface and describes the effects of
light on plants. Plants selectively absorb radiation for photosynthesis, and leaves appear green to
the human eye because plants reflect most green light (wavelengths 500–580 nm).
The amount of solar radiation, or insolation, reaching a horizontal surface on the Earth ranges
from 800 to 1200 W m−2. These values change based not only on atmospheric content, as described in the
earlier text, but also because of the slope and aspect of the ground. When radiation strikes the ground at
angles greater or less than 90°, the radiation is distributed over a larger surface area and the amount of
energy received (the photon flux, in moles or μE m−2 s−1) is less (Figure 6.1). In the Northern Hemisphere,
south-facing slopes receive more radiation per unit area than north-facing slopes because they face the
sun at its highest angle during the day. Moreover, eastern slopes receive their highest irradiance in the
morning, whereas western slopes receive their highest irradiance in the afternoon (Figure 6.2).
A A A
B α
A
α
F I G U R E 6 . 1 Distribution of insolation at different aspects and slopes. Photon flux changes according to
the cosine of the angle of incidence (∝). Angle of incidence varies with the orientation of the receiving
surface as well as with the position of the light source. Energy received at B = photon flux (moles m−2 s−1) of
Acos∝. Source: Hart (1988) / Unwin Hyman.
Distribution of Light Reaching the Ecosphere 125
F I G U R E 6.2 Relative irradiance received throughout

the day on a 100% (45°) slope in the southern
Appalachians (north latitude 35° 30′) on June 21. (A) C
South; (B) north; (C) east. The west slope curve is a A
mirror image of that for the east slope; the maximum
Relative irradiance
irradiance occurs in the afternoon. Source: Bryam and
Jemison (1943) / United States Department of
Agriculture / Public Domain.
B
6 A.M. Noon 6 P.M.

Time of day
PLANT INTERCEPTION OF RADIATION

Plants utilize solar radiation efficiently by the way their leaves absorb, transmit, and reflect it. Leaf
absorption is the fraction of incident light absorbed by the leaf, and transmittance is the amount
of light that passes through the leaf and comes out the other side. Absorption is logarithmically
related to transmittance, with absorption = 0 corresponding to 100% transmittance, and
absorption = 1 corresponding to 10% transmittance. Reflectance is the amount of incident light
neither absorbed nor transmitted by leaves. Only about 4–5% of total solar energy is absorbed and
converted to chemical energy, but only 1–2% of total solar energy is used in photosynthesis, a
remarkably small amount considering the amount of plant biomass produced globally on an
annual basis (see Chapter 18). Plant leaves are most efficient at absorbing ultraviolet and visible
light, but inefficient at absorbing near-infrared light (the 0.7–1.5-μm range, high in energy—see
Figure 6.3). Leaves have high absorptivity of far-infrared light where solar irradiance is very low
(Gates 1968). By contrast, reflectance and transmittance are highest for near-infrared light and
lowest for visible and ultraviolet light. These properties are similar among many species, even
when comparing plants from many different environments and with varying leaf anatomy. Thicker
leaves seem to decrease transmittance and increase reflectance in many species (Knapp and
Carter 1998). The widest difference among species for absorption and transmittance occurs in the
portion of visible light that appears green (wavelengths 500–600 nm; Knapp and Carter 1998;
Hagemeier and Leuschner 2019).
CANOPY STRUCTURE AND LEAF AREA

Interception of solar radiation depends very strongly on the structure of the forest canopy, deter-
mined largely by the surface area, shape, density, height, and depth of many adjacent tree crowns.
Our understanding of canopy structure effects on light is based heavily on wet tropical forests.
These forests have high tree and shrub diversity that incorporates wide ranges of shade tolerance
and crown architecture, so they are able to produce a multi-layered canopy structure (Figure 6.4a).
As a result, virtually all vertical space may be occupied in many forests of the wet tropics, extend-
ing from the tallest overstory trees down to herbaceous plants occupying the forest floor, and light
intensity at the forest floor is extremely low. By contrast, canopy structure is often much less com-
plex in temperate forests. In many trembling aspen or paper birch forests, for instance, overstory
trees form a single-layered canopy with relatively sparse crowns that allow ample light to reach the
126 Chapter 6 Light
100 F I G U R E 6.3 Mean reflectance, transmittance,

and absorptance (thick lines) of the leaves of 26
80 species. The thin lines around the mean
represent +1 standard deviation of the mean.
Reflectance (%)
Source: Knapp and Carter (1998) / with

60
permission of John Wiley & Sons.
40
20
0
100
80
Transmittance (%)
60
40
20
0
100
80
Absorptance (%)
60
40
20
0
400 500 600 700 800
Wavelength (nm)
well-developed shrub and ground flora (Figure 6.4c). Coniferous forests of the Pacific Northwest
may also have a single canopy layer, but their vertical canopy depths extend deeply toward the
forest floor (Figure 6.4b), utilizing available sunlight to a much greater extent than aspen or birch
forests. Canopy depth varies substantially among shade-intolerant, mid-tolerant, and tolerant
species, with higher shade tolerance permitting leaf maintenance lower in the canopy. In the
Pacific Northwest, for example, the canopy of grand fir, a very shade-tolerant conifer, is much
deeper than that of mid-tolerant Douglas-fir and intolerant ponderosa pine (Figure 6.5).
Structural differences in the canopy determine the amount of leaf surface area present to
capture light energy, quantified by forest ecologists using leaf area index (LAI). For deciduous
canopies, LAI is defined as the one-sided green area of leaves (m2) per unit ground surface area
(m2), also called the projected leaf area. Because their leaves have no anatomical top or bottom sur-
face, LAI for conifers may be defined as the projected LAI, half the total needle surface per unit
Canopy Leaf Area Light

(a) Structure Distribution Intensity
30
Canopy Height
(meters)
15
0
0 50 0 50 100
(b)
30
Canopy Height
(meters)
15
0
0 50 0 50 100
(c)
30
Canopy Height
(meters)
15
0
0 50 0 50 100
Percent of Percent of
Total Full
Leaf Area Sunlight
F I G U R E 6 . 4 Canopy structure, leaf area distribution, and light intensity in (a) a wet tropical, (b) western
conifer, and (c) a trembling aspen ecosystem. Notice the relationship between canopy structure and the
distribution of leaf area within each forest. Differences in the vertical distribution of leaf area result in
dramatically different light profiles within these ecosystems.
ground area, or total needle surface per ground area, all with units m2 m−2. LAI varies widely
among ecosystems, ranging from 1 m2 m−2 in sparse juniper woodlands in southwest Wisconsin
(Schuler and Smith 1988) to 8.4 m2 m−2 for a sugar maple-dominated forest in northern Wisconsin
(Fassnacht and Gower 1997) to 13 m2 m−2 in coastal old-growth Douglas-fir–western hemlock eco-
systems of the Pacific Northwest (Thomas and Winner 2000). Moreover, LAI may vary significantly
with forest age and tree density even within the same forest type (Kashian et al. 2005a). At least
some inaccurately high estimates of LAI in old forests (as high as 22 m2 m−2) result from using
methodology for estimating LAI developed in younger forests.
128 Chapter 6 Light
100 F I G U R E 6.5 Differences in the vertical distribution

of leaf area in grand fir, Douglas-fir, and ponderosa
pine. The crown of grand fir, a very shade-tolerant
species, has a deeper vertical distribution of leaves,
compared to the shade-intolerant ponderosa pine.
Crown Height %
grand fir
50 Douglas-fir
ponderosa pine
0
0 20 40
Leaf Area %
15 F I G U R E 6.6 The relationship between leaf

Wet
coastal w. hemlock area index and ecosystem productivity among
selected forest ecosystems of the Pacific
Northwest. Productivity is expressed in
megagrams of carbon per hectare per year.
Ecosystem Productivity (Mg ha–1 y–1)
mountain
hemlock
10
Douglasfir
western
juniper
5
ponderosa pine
sagebrush
Dry
0
0 5 10 15 20
Leaf Area Index (m2 m–2)
LAI and productivity are closely related in many forest ecosystems because leaf area determines
the amount of photosynthetically active surface area available to convert light energy into plant bio-
mass. The amount of leaf area also determines the amount of surface area from which water can be
lost via transpiration, however, such that climate and site, leaf area, and productivity are all interre-
lated. In Rocky Mountain juniper woodlands found east of the Cascade Mountains, for example, low
amounts of precipitation in the semi-arid climate constrain leaf area, which in turn reduces ecosystem
productivity (Figure 6.6). By contrast, high amounts of precipitation on the coast of the Pacific
Northwest support western hemlock, which attains a much greater leaf area, is less affected by water
loss in a humid climate, and ultimately results in greater rates of ecosystem productivity. The relation-
ships between water availability, leaf area, and productivity are discussed in detail in Chapter 18.
Tree species vary widely in their ability to intercept solar radiation, and this variation is obvious
even for broad vegetation types. Relative illumination beneath a deciduous forest in the winter may
approach 50–80% of full sunlight, 10–15% under open, even-aged pine stands, <1–5% under temperate
hardwoods in summer, and as low as 0.1–2% beneath tropical rain forests (Huber 1978). As we sug-
gested in the earlier text, we observe differences in light interception among canopy tree species
because they differ in shade tolerance and crown architecture. Shade-tolerant species are able to pro-
duce dense or deep (extending closer to the ground) crowns whose inner or lower foliage is able to
survive in lower light conditions. In a study of light interception by canopy trees in southern New
England, canopies dominated by eastern hemlock or American beech transmitted <1% of PAR to the
understory (Canham et al. 1994). Both species are very shade-tolerant and exhibit dense canopies;
eastern hemlock, notably, exhibits a canopy along 95% of its total tree height. By contrast, canopies
dominated by mid-tolerant species with sparser crowns, such as northern red oak or white ash, trans-
mitted >6% PAR to the understory. Compared to eastern hemlock, canopy depth of white ash and
northern red oak represented only 39 and 47% of total tree height, respectively. Because transmission
of PAR through canopies of northern red oak is nearly 6 times that through eastern hemlock
canopies, the presence of a single hemlock tree could potentially create a zone of deep shade in a
northern red oak forest (Canham et al. 1994).
Because tree species differ in the amount and quality of light they transmit, it follows that a mix-
ture of species within a forest canopy is likely to be more efficient at intercepting incoming radiation.
Forest canopies of mixed species have been shown to accomplish canopy packing, whereby species
that have complementary canopy architecture and physiology are able to more efficiently use available
growing space (Pretzsch 2014; Jucker et al. 2015), thereby increasing light interception and absor-
bance and reducing transmittance to the understory. Often the effect of a species-diverse tree canopy
has a greater impact on light interception than the identity of the species themselves (Sercu et al. 2017).
For example, a study conducted across Europe showed that mixed canopies of European beech and
Scots pine had a 14% higher absorption of PAR than pure canopies of either species (Forrester
et al. 2018). Although sometimes the presence of different canopy shapes from different species alone
increases the interception of light by the forest canopy, other times the presence of multiple species
changes the crown shape of the species present to increase canopy packing even further (Figure 6.7).
F I G U R E 6 . 7 Potential effects of mixed (a) (b) (c)

species on canopy packing and thus light
interception and absorption. When two
species have (a) and (b) complementary
shapes, their presence has (c) an additive
effect of space filling. Canopy packing
may increase even further (and thus light
interception increases) if mixing allows
(d) wider crown expansion, (e) higher
stand density, or (f) both. Source:
Pretzsch (2014) / with permission
(d) (e) (f)
of Elsevier.
130 Chapter 6 Light
Technological advances in forest canopy measurement and characterization over the past
two decades have allowed a greater appreciation for the complexity and importance of canopy
structure. Canopy structural complexity (CSC) is a concept that describes the vertical and
horizontal arrangement and complexity of both photosynthetic and non-photosynthetic tissue
within forest canopies. The concept of CSC encompasses LAI, which estimates photosynthetic sur-
face area in a single dimension, but is multidimensional, also including variables relating to can-
opy height, arrangement, openness, and variability (Atkins et al. 2018a). In essence, CSC is able to
quantify the spatial pattern of leaf area, both vertically and horizontally, in a forest canopy.
Increases in some CSC parameters such as canopy variability and height are correlated with higher
levels of light absorption, whereas increases in other parameters such as canopy openness and
clumping are correlated with lower levels of light absorption (Atkins et al. 2018b). Measures of
CSC are novel and are still developing, but they are likely to better predict how forest canopies uti-
lize incoming solar radiation than LAI alone (Atkins et al. 2018b). Likewise, CSC is emerging as an
improved predictor of ecosystem productivity compared to LAI (Fotis et al. 2018; Gough et al. 2019).
It appears that among all measures of CSC, maximum canopy height may be a primary driver of
CSC because it allows the development of more complex arrangements of vegetation that improve
the efficiency of resource use (including light) and thus increases productivity (Gough et al. 2020).
LIGHT QUALITY BENEATH THE FOREST CANOPY

As we suggested in the earlier text, differences in the amount of available light during the growing
season are the main drivers of plant response in the understory of forests. Here, we describe the
quality of radiation reaching the understory. Beneath forest canopies, PAR consists of (i) small
amounts of light transmitted through leaves, (ii) unscattered radiation reflected downward by
foliage, known as beam enrichment, and (iii) light that passes unabated through openings in the
canopy. The primary source of understory PAR is through canopy openings, but transmitted light
and beam enrichment are particularly important beneath closed or very dense canopies (Hutchin-
son and Matt 1976). Diffuse transmission represented only 0.5% and 0.7% of total understory PAR
in a Douglas-fir–western hemlock forest in the Pacific Northwest and a northern hardwood forest
in southern Ohio, respectively, but >6% understory PAR in red spruce–balsam fir forests of western
Tennessee (Canham et al. 1990).
Much of the light transmitted or reflected by the forest canopy to the understory is in the far-
red spectrum (700–750 nm), and light quality is very often reported as a red (600–700 nm) to far-red
(R/FR) ratio. Most of the UV-B, blue, and red wavelengths are absorbed by the canopy, and the re-
maining light transmitted to the understory is enriched in green and far-red light spectra. In general,
therefore, shady understory light conditions are typically characterized by low PAR and a low R/FR
ratio (Messier and Bellefleur 1988; Fiorucci and Fankhauser 2017). However, a classic study by Vé-
zina and Boulter (1966) documented differences in light quality among forest ecosystem types.
Light beneath a sugar maple canopy was dominated by light in the far-red spectrum with lesser
amounts of green, blue, and red light and ultraviolet radiation (Figure 6.8). Note that red light
increases sharply in Figure 6.8 in the morning and early afternoon, when direct sunlight reached
the understory through small canopy openings (sunflecks; see the following text). By contrast, no
significant differences were found among various wavelengths in the visible spectrum beneath the
red pine canopy, given its lack of selectivity in transmission. The proportion of radiation transmitted
by forest canopies is also higher on cloudy or overcast days compared to clear days, and the sugar
maple canopy in particular was less selective than on clear days (Vézina and Boulter 1966).
SUNFLECKS
Much of the forest floor in many forest ecosystems is in heavy shade for the majority of the growing
season. Through small openings in the canopy, direct-beam radiation reaches the understory to
Light Quality Beneath the Forest Canopy 131
In the open
200
100
90
80
70
60
50
40
Absolute energy, 10–5 ly min–1 nm–1
30
U.V.
Blue
20 Beneath the canopy Green
Red
Far–red
10
9
8
7
6
5
1
09.00 10.00 11.00 12.00 13.00 14.00
Time, E.S.T.
FIGURE 6.8 Light quality in the open and beneath the canopy of a forest dominated by sugar maple. Spikes in
the understory light regime represent the presence of sunflecks. Source: Vézina and Boulter (1966) / Canadian
Science Publishing.
create sunflecks. These short-lived patches of sunlight sweep over the ground over the course of
the day, momentarily bathing small areas with unfiltered radiation (Figure 6.9), though with
reduced PAR compared to full sunlight. A key feature of sunflecks is that they are not present at
any given location of the understory for more than 10% of the time, yet they disproportionately
contribute to the amount of PAR available for understory plants (Chazdon 1988). Sunflecks are
often clustered in time, separated by long periods of few or no sunflecks (Vierling and Wess-
man 2000). Canham et al. (1990) reported that the proportion of total growing season PAR received
as sunflecks beneath closed forest canopies at several temperate sites and one tropical rain forest
132 Chapter 6 Light

F I G U R E 6 . 9 Sunflecks are short-lived patches of direct solar radiation reaching the understory through
small holes in the canopy. These patches sweep light over the ground during the course of a day, creating an
important source of radiation for understory vegetation. Sunflecks in this photo provide enough light
irradiance to support a thick seedling layer of red ash beneath a closed canopy.
site ranged from 47% to 68% of the total amount of light penetrating through canopy openings.
Notably, sunfleck duration averaged only 5.7–7.1 minutes. Most sunflecks are even shorter in dura-
tion, however, with 95% lasting less than 2 minutes (Pearcy et al. 1994).
The ability of an understory plant to utilize sunflecks depends on a complex balance of rapid
changes in chloroplasts, electron transport carriers, and the potential accumulation of pools of
photosynthetic metabolites. Plants may require several minutes to adjust their photosynthetic
mechanism to adapt to this new high radiant energy flux (Fitter and Hay 1987). In addition, stoma-
tal opening on the illuminated leaves must occur quickly enough to allow carbon dioxide uptake
to occur for photosynthesis while the sunfleck is still in place. Some tree seedlings, such as sugar
maple and white ash that open stomata rapidly and exhibit slow stomatal closure with increasing
light, are probably able to make use of sunflecks (Davies and Kozlowski 1974). In general, how-
ever, understory plants benefit little from individual short-duration sunflecks because stomata take
at least 5–10 minutes to open (Pallardy 2008). However, the induction time of photosynthesis of a
leaf is reduced each time it is exposed to relatively high light levels. This means that sunflecks,
even if short in duration, will prepare the leaf for better utilization of the next subsequent sunfleck.
Thus, the sunflecks most beneficial to understory plants are those of long duration or which occur
in rapid succession. A full description of the physiological responses of plants to sunflecks is
provided by Way and Pearcy (2012).
Notably, the effects of sunflecks are not always beneficial. Leaves of understory plants are
typically well adapted to shade, and brief exposure to the bright light of sunflecks has the potential
to damage leaves due to photoinhibition, overheating, or water stress. A study by Ustin et al. (1984)
compared red-fir seedlings on north and south slopes in upper montane forests of the central
Sierra Nevada Mountains of California (Ustin et al. 1984). On the south slopes, plots with few seed-
lings had mean daily irradiance twice as high as plots with many seedlings, primarily due to sun-
flecks that occurred 3.5 times as often at irradiance levels above 1025 μE m−2 s−1. Photon flux
densities >500 μE m−2 s−1 exceeded the light saturation point for net photosynthesis in red fir, and
Light and Growth of Trees 133
thus substantially increased the energy load (heat) without corresponding increases in carbon
gain. Sunfleck tolerance strategies are many and may include physical restructuring of chloro-
plasts, changes in leaf angles in some species, utilization of leaf pigments, and the emission of
isoprene, the latter for the protection of photosynthesis against heat stress (Way and Pearcy 2012).
LIGHT AND GROWTH OF TREES

Growth and development of forest trees are tightly linked to photosynthesis, which of course is
dependent on light. Absorption spectra are described in the earlier text as the wavelengths of
visible light absorbed by plants; here we discuss action spectra, which are the specific wavelengths
most effective for photosynthesis. We described in the earlier text that green foliage reflects visible
light in the green wavelengths (Figure 6.3), and thus absorption is highest in the blue-violet and
yellow-red wavelengths. We might predict that plants absorb only the wavelengths that they will use
to drive photosynthesis, and thus the action spectra for most plants also include blue-violet and
yellow-red wavelengths. Several studies have shown that not all green wavelengths are reflected and
often >70% are absorbed (Figure 6.3). Action spectra for most green plants are highest in red wave-
lengths (600–700 nm), followed by blue wavelengths (400–500 nm), and then by green wavelengths
(500–600 nm; McCree 1972). Several classic studies of woody plants identified action spectra even
before the concept was articulated. For example, Burns (1942) showed that orange-red wavelengths
were most effective in stimulating photosynthesis in Norway spruce, eastern white pine, and red
pine, whereas the shorter violet-blue waves were relatively unimportant. Similarly, light absorption
in Scots pine seedlings was highest in the red and the blue regions, although the photosynthetic
efficiency of blue light was lower (Linder 1971). An important point is that absorption in both the
blue and red regions is needed for maximum photosynthetic efficiency.
Many studies have examined how variation in the amount of light affects the rate of photo-
synthesis. Respiration is the opposite reaction to photosynthesis—the process that draws energy
from the carbon that is fixed into biomass via photosynthesis and releases CO2 to the atmosphere
(Chapter 18). Light respiration occurs mainly in leaves, results in a relatively high loss of fixed
carbon, and is genetically controlled. Dark respiration occurs in all living cells, acts to build and
maintain living tissues, and is more closely coupled to environmental factors. Net photosynthesis
(photosynthesis minus dark respiration) must be positive—meaning more CO2 is taken up than
given off—for a leaf to fix carbon into its tissues. When light levels are low, the rate of photosynthe-
sis is also low, and the leaf (and ultimately the plant) loses more CO2 than it gains. The light
compensation point is the light intensity at which carbon gain from photosynthesis and carbon
loss from respiration are equal, and net CO2 is neither lost nor gained. In dense forests with very
dark understories, light intensity in the understory is generally too low for photosynthesis to exceed
respiration without the presence of sunflecks and adaptions by understory plants to trap diffuse
light (Huber 1978).
Light compensation points vary widely among species, and particularly among species with
differing shade tolerances (Figure 6.10). A study of 115 species of woody plant seedlings found that
compensation points were lower for shade-tolerant species than for shade-intolerant (“sun-
loving”) and mid-tolerant species because of lower dark respiration rates in shade-tolerant species
(Craine and Reich 2005). For example, the light compensation points of ponderosa pine and Scots
pine (shade intolerant) are nearly three times as high as for eastern white pine (mid-tolerant), but
those for eastern hemlock, American beech, and sugar maple (shade tolerant) are extremely low
(Burns 1923). Shade plants generally require 0.5% of full sunlight at the compensation point, and
sun plants require 1.5% (Perry et al. 1969). Given that understory light intensity may be very near
the compensation point for as much as 90% of the growing season (Ellsworth and Reich 1993), low
compensation points and dark respiration rates for shade-tolerant species are critical for them to
achieve positive net photosynthesis (Figure 6.10).
134 Chapter 6 Light
Light saturation points

Shade-intolerant
plants
Net photosynthesis
Shade-tolerant
plants
Light compensation points
0
Light intensity (PAR)
Dark respiration rates
F I G U R E 6 .1 0 Hypothetical photosynthetic light saturation curves for plants of differing shade tolerance.
Carbon gained through photosynthesis is greater than that lost through dark respiration when net photosyn-
thesis is positive. Net photosynthesis increases asymptotically with increasing light intensity to a saturation
point for all plants, light is limiting at intensities less than the saturation point, and CO2 is limiting once the
saturation point is reached. Light saturation is lower, dark respiration rates are lower (less negative), and
light compensation points are less for shade-tolerant plants.
Woody plants continue to increase their photosynthetic rate as light intensity increases
above the compensation point, eventually reaching an asymptote at 25–50% full sunlight. The
point at which photosynthesis no longer increases with increasing light intensity (and carbon fix-
ation no longer increases) is called the light saturation point. At this point, the reaction is not
light-limited, but is instead affected by the supply of CO2 , PO34 , and enzymatic processes. The
light saturation point is higher for shade-intolerant woody plants, such that they are able to attain
higher levels of net photosynthesis (Figure 6.10). Classic studies by Kramer and Decker (1944)
and Kozlowski (1949) identified a much higher light saturation point for loblolly pine seedlings
compared to associated hardwood species, which reached their maximum photosynthetic rate at
30% or less full sunlight. Such a finding partly explains better relative competitive ability of hard-
woods under open pine stands where the relative illumination is typically near 30% full sunlight
and is common for many hardwood species that succeed pine (Figure 6.11). Higher potential rates
of photosynthesis generally trade off with the higher physiological cost of synthesizing and main-
taining higher proportions of photosynthetic machinery per unit of biomass.
The relationships between net photosynthesis and light intensity are similar for sun leaves
and shade leaves in a single canopy as they are for shade-intolerant and shade-tolerant plants.
These relationships affect the branching patterns of species depending on their shade tolerance
and growth rate. Fast-growing shade-intolerant species tend to have branches more widely spaced
than those of shade-tolerant species whose shade leaves acclimate to the light conditions of the
inner crown by anatomical features (see discussion in the following text). Many floodplain species
(willows, cottonwoods, silver maple, American elm), for example, have broad and often wide-
spreading crowns that facilitate light capture by foliage of the external and internal crown.
Many other factors affect the competitive ability of different species under given light
conditions, including the color, shape, presence of waxes and hair, and the arrangement of
Light and Growth of Trees 135
F I G U R E 6 .1 1 A conspicuous hardwood understory in an open 80-year-old loblolly pine stand, North

Carolina. Source: Roth (1990). Reprinted with permission of Gebrüder Borntraeger Verlagsbuchhandlung,
Stuttgart, Germany.
leaves. These factors regulate the amount of light that reaches the chlorophyll and drives pho-
tosynthesis. For example, pines have rounded needles arranged in dense clusters, often inclined
upward rather in a plane perpendicular to incident sunlight, which results in scattered light
and shading of one needle by one or more others (Kramer and Clark 1947; Carter and
Smith 1985). Many hardwoods associated with pines in the eastern United States, such as
dogwood, have broad, thin leaves, arranged to minimize leaf shading and oriented perpendicu-
lar to the direction of incident radiation. In this comparison, chlorophyll in a pine needle
receives a much smaller fraction of the light incident to it than does chlorophyll in a dogwood
leaf. As such, pine is able to maintain a high net photosynthetic rate in high light conditions of
the overstory, and dogwood is able to efficiently photosynthesize in low light conditions of the
understory.
Conifers and evergreens in general have a relatively low photosynthetic capacity per unit
area of foliage (Waring 1991; Reich et al. 1995). From the tropics to the boreal forest, the maximum
photosynthetic capacities of evergreen species are usually half to less than a quarter of that of asso-
ciated deciduous species. Moreover, many deciduous species have equal photosynthetic efficiencies
at low and high light intensities, whereas many evergreen species are more efficient when light
intensity is high (Pallardy 2008). Such a pattern does not always imply that conifers and evergreen
136 Chapter 6 Light
species are less competitive than deciduous species, however, in part because conifers are able to
tolerate lower resources on poorer sites. Moreover, in environments where deciduous and conifer-
ous trees occur together, conifers often develop a canopy dense enough to capture most of the
incident radiation in the visible range (Jarvis and Laverenz 1983). Once a dense, closed-canopy
stand of conifers dominates an area, as after fire, all but the most shade-tolerant co-occurring
deciduous species are at a disadvantage. Finally, the ability of evergreen species to achieve some
carbon fixation via photosynthesis while associated deciduous plants are dormant also improves
their competitive ability over the long term.
Both the angle and azimuth of the leaf blade affect its interception of solar radiation. In
many deciduous species, leaf petioles are able to orient the leaves to either increase absorption of
radiation or avoid it. Positioning the leaf perpendicular to incoming radiation maximizes the
amount of light absorbed and heat received. Normally this orientation is achieved over several to
many days during the normal growth process, but many species are able to re-orient leaves at the
temporal scale of minutes or hours. For example, woody leguminous species such as redbud,
locust, or Kentucky coffeetree contain a small group of cells at the base of the leaf petiole that
facilitates the movement of individual leaves into acute angles in response to radiation and, most
often, its associated temperature. The horizontally flattened petioles of aspen species around the
world (trembling aspen, big tooth aspen, European and Asian aspens) enable the leaf blade to
move even in slight breezes, which permits light to bathe leaves more effectively than if they were
stationary. Aspen leaf movement also reduces excess heat on the leaf surface when irradi-
ance is high.
LIGHT AND SEEDLING SURVIVAL AND GROWTH

In general, at least 20% of full sunlight is required for survival of woody plants over a period of
years, the actual amount varying with the species and with growing conditions. Measuring the
height or weight (dry matter) growth of plants over a period of years under different degrees of
irradiance has shown that several pine species grow best in full sunlight, although diameter
growth is more affected than height growth at 50% sunlight (Pearson 1936, 1940; Shirley 1945b;
Logan 1966a). Similarly, height growth of other coniferous species is virtually unaffected by
partial light, such as white spruce, white pine, and Douglas-fir (Gustafson 1943; Shirley 1945b;
Logan 1966a; Lassoie 1982). Several studies suggest that height growth of many hardwood
species actually improves in partial sunlight. Height growth data for six hardwood species in
Missouri (McDermott 1954) showed that five of the six species grew tallest in either 33% or 50%
relative illumination (Table 6.1). The maximum seedling height of white birch, yellow birch,
silver maple, and American elm seedlings in Ontario was at 45% relative illumination, but it was
at 25% for highly shade-tolerant basswood and sugar maple (Logan 1965, 1966b). At least some
recent studies support the idea that shade-tolerant tree seedlings have higher growth and
survival in low light than shade-intolerant species, but shade-intolerant species grow faster at
higher light intensity (e.g., Walters and Reich 1996, 2000a). Notably, fewer species achieve
greater dry weight under partial light compared to full light because of increased allocation to
roots (Chapter 18).
Root development of seedlings is also limited by low light conditions, such that seedlings of
most forest trees have relatively shallow and poorly developed root systems in shaded conditions.
In Table 6.1, American elm may achieve its greatest height growth at 33% sunlight, but above-
ground biomass is 3.2 times that of belowground biomass under these conditions. Elms developing
in full sunlight are shorter, but the root system is far better developed under these conditions (top/
root ratio = 1.7). The reduction of root growth in low light is critical for the competitive ability and
survival of woody plant seedlings in a forest understory, where root competition substantially
reduces the amount of available soil water and nutrients. Although there are many causes of
Light and Tree Morphology and Anatomy 137
Table 6.1 Seedling height (HT) and top/root weight ratio (T/R) of newly germinated seedlings as influenced
by amount of sunlight (relative illumination, RI).
20% RI 33% RI 50% RI 100% RI

Species
HT (cm) T/R HT (cm) T/R HT (cm) T/R HT (cm) T/R
American elm (Ulmus 73 3.1 81 3.2 69 3.0 37 1.7
americana) 13 weeks
Winged elm (Ulmus 59 5.2 71 3.9 76 4.4 28 1.9
alata) 13 weeks
Sycamore (Platanus 43 4.6 41 3.5 40 4.1 33 2.0
occidentalis) 15 weeks
River birch (Betula 26 13.4 33 9.3 52 8.3 41 3.6
nigra) 10 weeks
Red maple (Acer 26 4.1 27 3.7 29 3.5 26 2.0
rubrum) 13 weeks
Alder (Alnus 21 3.4 20 3.5 38 4.7 30 2.8
rugosa) 14 weeks
Source: Based on McDermott (1954).
seedling mortality, growth of seedlings in low light often occurs too slowly to develop a root system
capable of taking up soil water and nutrients necessary for long-term survival of the plant (Coomes
and Grubb 2000).
These patterns of biomass allocation to aboveground versus belowground structures are
not likely to be a direct result of light availability alone. Instead, they have been proposed as a
trade-off between maximizing photosynthetic efficiency in low light by allocating more biomass
to aboveground structures for light capture at the expense of roots and maximizing the efficiency
of water and nutrient capture in high light conditions where transpiration rates are higher and
belowground resources may be less available (Smith and Huston 1989; Pallardy 2008). As a
result, some species may tolerate deeper shade on soils with higher water and nutrient avail-
ability, and less shade where soils are poorer (Kobe et al. 1995; Walters and Reich 2000a). A
detailed discussion of aboveground and belowground carbon allocation is presented in
Chapter 18.
LIGHT AND TREE MORPHOLOGY AND ANATOMY

Morphological differences between sun-grown and shade-grown plants of the same species may
result from their respective exposure to varying light conditions. These changes are of ecologi-
cal importance in understanding a species’ plasticity—its capacity to adjust to shaded
conditions—or its response to sudden release following removal of the overstory by windstorm
or by cutting.
Perhaps most studied for their responses to changes in radiation are the leaves of trees and
shrubs, especially because the leaf is the plants’ principal photosynthetic organ. In most cases, a
species is able to adjust the physiology and structure of its leaves developed in the shade compared
to those developed in full or near-full sunlight. Similar to how the rates of photosynthesis vary
among shade and sun leaves, the altered structure and physiology of leaves on a plant grown
entirely in the shade resemble those of leaves found at the base or interior of a tree canopy. Typi-
cally, shade leaves have a larger surface area per unit weight, and are thinner and less deeply lobed,
with a thinner epidermis, less conducting tissue, fewer stomata, more intercellular space and
spongy mesophyll, more chlorophyll per unit dry weight, and lower respiration (Pallardy 2008;
138 Chapter 6 Light
EPIDERMIS
PALISADE MESOPHYLL
SPONGY MESOPHYLL
RESIN DUCT
IN SUN
EPIDERMIS
PALISADE MESOPHYLL
STELE
SPONGY MESOPHYLL
RESIN DUCT
IN SHADE
F I G U R E 6 .1 2 Structural variability in western hemlock needles grown in full sun (top) and in the
understory beneath a dense Douglas-fir canopy. Note the thinner shape, larger surface area, thinner
epidermis, and reduced palisade in the shade-grown needle. Source: Tucker and Emmington (1977) / with
permission of Oxford University Press.
Larcher 2003) than comparable sun leaves of the same tree (Figure 6.12). The higher surface area
of shade leaves increases their ability to gather light per unit of respiring surface area, and the thin-
ner epidermis transmits more light energy compared to sun leaves (Daubenmire 1974). A study by
Eschrich et al. (1989) also found that shade leaves of European beech had one layer of palisade
mesophyll, but two layers when grown in the sun.
Although shade leaves develop in response to reduced light, temperature is also an impor-
tant factor resulting in differing structure and anatomy. Temperature generally increases with irra-
diance, and water stress generally increases with temperature. Sun leaves therefore tend to experi-
ence higher water stress compared to shade leaves, but leaves within the crown and in its lower
portions on clear days are under markedly less water stress compared to the sun leaves at the top
of a tree. Therefore, many characteristics of sun leaves are as likely to be important adaptations to
reduce water stress with increasing temperature as they are for differences in light. For example,
the deep lobing of sun leaves apparently aids in the convective cooling of surfaces. Likewise, thick-
er leaves with a thicker epidermis act to reduce water loss in sun leaves, and their thicker cuticles
reflect incoming radiation to reduce temperature.
The plasticity in leaf anatomy described in the earlier text varies among species, even
among those in the same genus. The leaves of black oak, a xeric, shade-intolerant species,
exhibit far more anatomical plasticity compared to those of northern red oak, a mesic, mid-
tolerant species (Ashton and Berlyn 1994). Some species exhibit little plasticity in leaf anatomy
and thus little difference between their sun leaves and their shade leaves. The anatomy of such
plants is suited for a limited set of environmental conditions, and thus they are usually found
only in the overstory or only in the understory. A study by Jackson (1967) found that leaves of
shade-tolerant trees such as red maple, American beech, and flowering dogwood had one layer
of palisade mesophyll regardless of light intensity, but those of shade-intolerant species such as
tulip tree or black cherry were more plastic in the layers of palisade they exhibited across
light regimes.
Photocontrol of Plant Response 139
LIGHT AND EPICORMIC SPROUTING

Light may have important effects upon bark thickness and the stimulation of adventitious buds.
Adventitious buds in the bark may be activated to suddenly form epicormic branches when the
bark is exposed to direct solar radiation, such as by the thinning of forest stands via management
or disturbance. Under sudden changes from shade to light, epicormic branches may develop in
most forest trees, both conifers and hardwoods, although they are much more common in decidu-
ous trees (Pallardy 2008). The occurrence of epicormic branches after thinning is particularly
serious with many of the oaks, maples, birches, and other hardwoods where stem quality is of par-
amount economic importance. Although light has often been assumed to be the factor that triggers
hormonal responses within a tree that lead to epicormic sprouting, the direct influence of light on
sprouting is not clear (Meier et al. 2012). Some recent research suggests that shading the stem
limits the initiation of epicormic branching (e.g., Gordon et al. 2006), but just as many studies
report weaker correlative results (e.g., Wignall et al. 1987). A study by Nicolini et al. (2001) found
heavy epicormic sprouting in natural stands of European beech without significant canopy gaps or
sudden changes in light.
Many other factors in addition to light have been attributed to the initiation of epicormic
sprouting, such as various stressors, low tree vigor, weakened crowns, or tree pruning (Meier
et al. 2012). Many field observations have shown that epicormic branches form on the exposed
boles of many trees with weak, poorly developed crowns. European foresters have been able to
reduce the amount of epicormic branching on oaks by maintaining vigorous crowns and keeping
the bole clothed with understory trees such as beech and hornbeam. It is clear that carefully con-
trolled experimentation is required to determine which factor or combination of factors actually
initiates the change of physiological circumstances that lead to the development of epicor-
mic branches.
PHOTOCONTROL OF PLANT RESPONSE

Most critical plant processes are subject to the same photocontrol system, including stem elonga-
tion, root development, dormancy, germination, flowering, and fruit development. This system,
whereby light is absorbed by a reversible pigment system in the plant, is the physiological basis of
photoperiodism and affects seasonal rhythm, timing, and amount of tree growth. Photoperiod-
ism was discussed in Chapter 3 as the response of the plant to the relative length of the day and
night and the changes in this relationship throughout the year.
Garner and Allard (1920) and Garner (1923) proposed the term photoperiodism, and showed
that both the initiation of reproductive processes in plants as well as vegetative development could
be greatly affected by varying the relative length of day and night. Specifically, it is the length of an
interrupted dark period that is the main factor in photoperiodism that drives plant responses,
though we typically speak of “day length” rather than “night length.” Many plants were not found
to be especially sensitive to day length in these early studies. In other plants, known as short-day
plants, flowering could be induced when the plant was exposed to days shorter than a certain criti-
cal length. Likewise, some plants could be termed long-day plants because they responded to days
longer than a given critical length or even to continuous illumination (Galston 1994). The photo-
periodic effect on the flowering of woody plants is not well studied, in part because it has been
difficult to clearly assign long-day and short-day flowering types to trees and shrubs as has been
done in herbaceous species. Most woody plants are therefore considered to be day-neutral, meaning
that day length is unimportant for flowering.
Photocontrol is important not just for how it affects the growth and development of plants,
but also that plant responses to differing day lengths are clearly different for different species and
even individuals within a species. As described in Chapter 3, photoperiodic requirements vary with
the latitude and altitude of the source population with a species. Many tree species are therefore
140 Chapter 6 Light
able to differentiate on a genecological basis because they have been naturally selected for a
particular photoperiod associated with the limiting factors of the growing season.
Plant response to light occurs because light is absorbed by the phytochrome system of the
plant (Taiz and Zeiger 1991; Galston 1994). Phytochrome is a protein that acts as one of the more
important photoreceptors in plants. Phytochrome is expressed across many tissues and develop-
mental stages in plants (Li et al. 2011) and directs plant physiological changes in response to
changes in red and far-red light conditions. Phytochrome exists in two reversible forms:
red light 0.66 m

darkness
Pr Pfr Pr
far-red light 0.73 m
One form, Pr, has an absorption peak in the red region of the spectrum (650–670 nm) and is
the physiologically inactive form, appearing turquoise in color to the human eye. The other form,
Pfr, is physiologically active, has an absorption peak in the far-red region (705–740 nm), and
appears more green in color. Absorption of light at the appropriate wavelength converts one
phytochrome form to the other; when the red-absorbing form receives red light, it is changed into
the far-red-absorbing form (Pfr). When the far-red-absorbing form receives far-red light, it reverses
form back to Pr. Besides reversibility, the system is characterized by a slow drift in darkness from
the far-red absorbing form to the red-absorbing form. The detection of phytochrome, its properties,
and effects in morphogenesis of plants are reviewed by Vince-Prue (1975), Larcher (1995), Hart
(1988), and Pallardy (2008).
The phytochrome system drives growth cessation in short-day conditions and continuous
growth over long days. Red light dominates the spectrum near the end of the daylight period, with
a red light to far-red light ratio (R/FR) of around 1.15 (Smith 1982) and more than 70% of the
pigment is in the Pfr form at this time. The pigment reverts to the Pr form once the sun sets and
far-red light is dominant (R/FR 0.05–0.7; Smith 1982), and less than 10% remains in the Pfr form
depending on the length of the dark period (Downs 1962). As days shorten, less red light is avail-
able and thus less of the pigment will revert to the physiologically active Pfr form; this process is
exacerbated by lengthening nights. Woody plants eventually cease growth and enter dormancy as
the physiologically inactive Pr form accumulates. Conversely, under long days or continuous light,
growth is promoted because the pigment is in the Pfr form. This physiological system is responsible
for the development of a continuous sequence of photoperiodic races of trees in the north temperate
zone across species’ ranges. Unlike growth cessation, spring leaf flushing in woody plants is
controlled mainly by temperature rather than photoperiod once the plant’s chilling requirement
has been met (Chapter 3).
In addition to sensing seasonal changes in light, phytochrome allows plants to sense the
presence of neighboring plants. Light reflected from tree leaves or transmitted through tree can-
opies or neighboring vegetation is enriched in far-red light (lower R/FR), reducing Pfr and increas-
ing Pr. The accumulation of Pr beyond a specific threshold will initiate a series of signal transduc-
tion pathways that will in turn induce a set of shade-avoidance responses in the plant (Casal 2012).
Shade-avoidance responses include rapid elongation of stems and petioles, thinner leaves, reduced
branching, and reduced chlorophyll. These responses theoretically represent an effort to quickly
elevate the plant toward higher-quality light in crowded populations. Shade avoidance is much
more common in herbaceous plants compared to tree seedlings and is typically measured in trees
as the ratio of vertical to horizontal shoot growth. A study of 12 deciduous trees in Ontario re-
vealed that shade-tolerant understory tree species such as musclewood and hop-hornbeam had
low shade avoidance in low light compared to canopy tree species such as sugar maple (Henry and
Aarssen 2001). Thus, shade avoidance in shade-tolerant trees may be most important for those
species that rarely leave the understory.
The signaling action of phytochrome is sensitive to temperature and acts as a light-independent
thermal sensor similar to the way it is able to sense light. Recent work on herbaceous plants has
Light and Ecosystem Change 141
shown that Pfr reverts to Pr through a process called thermal reversion, which depends strongly on
temperature and is hastened as temperature increases. Thermal reversion is important in controlling
the amount of time Pfr is able to signal physiological changes in the early hours of darkness (Jung
et al. 2016). Higher temperatures are also effective at reducing Pfr in low light such as that present
in a shaded understory (Legris et al. 2016). Understanding that Pfr is high at the end of the daylight
period, stem elongation such as that occurring as shade avoidance will be suppressed in the early
hours of darkness (Van Buskirk et al. 2014). The fact that phytochrome responds to temperature as
well as light therefore suggests that it controls plant elongation at night as well as during the day.
The enhancement effect of cooler temperatures on Pfr (Jung et al. 2016) may have implications for
a plant’s cold tolerance, although this relationship has yet to be explored.
LIGHT AND ECOSYSTEM CHANGE

We cannot close a chapter about light without briefly discussing its importance for vegetation
change in ecosystems. Vegetation is one of the most visible and measurable expressions of the
change experienced by landscape ecosystems over time. Vegetation change, often termed
succession (Chapter 17), occurs within the context of site factors (landform, soil properties, tem-
perature, drainage, etc.), often following disturbances. Light is a critical part of the process of
succession, but it is fundamentally different from the other site factors. We rarely can find adjacent
sites that are permanently dark and light, but we can easily locate adjacent sites with vastly
different soil moisture, soil fertility, and temperature regimes.
Though strongly constrained by site factors, forest vegetation itself modifies the site factors
and disturbance regimes over time and thus plays a key role in ecosystem change. Vegetation
does little to affect landforms, at least from a geological sense, but it has a stronger effect on soil
development at the temporal scale of centuries to millenia. The dominant trees in an ecosystem,
however, may affect an enormous change in light availability over mere years or decades—
independent of the other site factors. The occurrence and species composition of the forest
below the dominant overstory (subdominant overstory, understory, shrub, and herbaceous
layers) are therefore heavily dependent upon light levels mediated by the density and crown
architecture of species in the overstory canopy (Figure 6.4). These light levels in turn affect or
determine many of the concepts discussed in this chapter, such as the complexity of the canopy,
shade tolerance, shade avoidance, root development of the understory, and the frequency and
duration of sunflecks at the forest floor.
Once the overstory fully develops in an ecosystem, trees having greater shade tolerance than
those than compose the overstory tend to establish and grow up in the shaded understory. Many site
factors in addition to light strongly influence species composition and density in any ecosystem. How-
ever, the plants themselves probably have the biggest impact on light availability compared to all site
factors, and light is a key factor in the dynamics of vegetation (see Chapter 17). In the next chapter we
focus on temperature, which is closely associated with light availability in forest ecosystems.
SUGGESTED
R E A D I N G S
Craine, J.M. and Reich, P.B. (2005). Leaf-level light Hart, J.W. (1988). Light and Plant Growth, 204. Boston,
compensation points in shade-tolerant woody seed- MA: Unwin Hyman.
lings. New Phytol. 166: 710–713. Kozlowski, T.T. and Pallardy, S.G. (1997). Growth Con-
Gough, C.M., Atkins, J.W., Fahey, R.T. et al. (2020). trol in Woody Plants. San Diego, CA: Academic
Community and structural constraints on the com- Press 641 pp.
plexity of eastern North American forests. Glob. Ecol. Larcher, W. (2003). Physiological Plant Ecology, 4e, 534.
Biogeogr. 29: 2107–2118. New York: Springer-Verlag.
142 Chapter 6 Light
Pallardy, S.G. (2008). Physiology of Woody Plants, 3e. San Way, D.A. and Pearcy, R.W. (2012). Sunflecks in trees
Diego, CA: Academic Press 454 pp. and forests: from photosynthetic physiology to global
Pretzsch, H. (2014). Canopy space filling and tree crown change biology. Tree Physiol. 32: 1066–1081.
morphology in mixed-species stands compared with
monocultures. For. Ecol. Manage. 327: 251–264.
Temperature CHAPTER 7
S imilar to light, the temperature regime at the Earth’s surface is determined by the energy
provided by incoming solar radiation. Incoming radiation interacts with secondary heat trans-
fers resulting from terrestrial radiation and air movement to produce the mean annual temperature
at a given location. The surface layers of air near the Earth are heated during the day to an extent
that depends on the amount of infrared radiation received. Thus, the greatest amount of heating
occurs at tropical latitudes, at high elevations, and where the air is free from water vapor, clouds,
and particulates. By contrast, nighttime temperatures are most determined by the rate and amount
of heat radiation from terrestrial objects and the atmosphere that were heated during the day.
In this chapter, we first consider broadscale patterns of temperature as determined by climate
without considering the interrelated effects of precipitation or water vapor. We then examine tem-
peratures at the soil surface, where the critical phases of plant germination and establishment take
place, within the forest, and in relation to topography and surficial features. Our main focus in this
chapter is temperature and plant growth, particularly how plants resist freezing, the process of
dormancy, and the distribution of plants in relation to cold hardiness. Detailed discussion of the
physiology of plants and plant growth as they relate to temperature is presented by Fitter and Hay
(1987), Sakai and Larcher (1987), and Pallardy (2008).
GEOGRAPHICAL PATTERNS OF TEMPERATURE

Broad patterns of temperature at the surface of the Earth are controlled largely by the rotation
of the Earth around its axis (which controls daily variation in temperature of a specific place)
and around the sun (which, combined with the Earth’s tilted axis, controls seasonal tempera-
ture). These patterns make latitude an important determinant of broadscale temperature;
generally, temperatures are warmest near the equator because incoming solar radiation
occurs most directly at the center of the globe. Tropical areas (between about 23.5°N and
23.5°S latitude) have an average temperature of about 79 °F (26 °C). Geometrically, the center
of the globe is also the area least affected by the Earth’s tilted axis, such that the variation bet-
ween winter and summer temperature is small. In fact, the temperature difference between
day and night in tropical areas is likely larger than that between summer and winter. Average
temperatures increase with latitude either north or south of the equator because the angle of
incidence of solar radiation is less direct, and seasonal variability increases because the
Earth’s tilt places these areas markedly closer to the sun in summer and further in winter. For
comparison, the average temperature at International Falls, Minnesota (a continental region
at about 49°N latitude) is 38 °F (3 °C), with an average of 56 °F (13 °C) in the summer and
19 °F (−7 °C) in the winter.
Temperatures are not strictly determined by latitude, as the proximity of an area to large
bodies of water such as oceans may alter temperature at a specific latitude. The most common
example of the influence of water bodies is in comparing maritime to continental locations. The
143
144 C h a p t e r 7 Temperature
German-Russian climatologist Wladimir Köppen first documented the difference between these
two types of geographical regions in the late 1880s. Maritime climates tend to be cool with relatively
little daily or seasonal variation in temperature. This is because the oceans adjacent to maritime
regions moderate the temperature as the warmer water heats the air in the winter and land heats
the air in the summer via conduction, and these heat transfers create convective currents within
the atmosphere. Land has a lower heat capacity compared to water, and thus continental climates
lose and gain heat much quicker than maritime climates. As a result, continental climates are cold,
lack the moderating influence of water bodies, and tend to have wider variation in seasonal tem-
peratures. For example, Vancouver, British Columbia lies at the same approximate latitude as
International Falls but in a maritime climate, and its average temperature is higher (52 °F/11 °C).
Moreover, Vancouver’s average temperature varies much less, from 68 °F (20 °C) in the summer to
43 °F (6 °C) in the winter. There are many other meteorological phenomena related to atmospheric
and oceanic currents and precipitation that contribute to the difference between continental and
maritime climates in addition to temperature.
Freshwater lakes, if large enough, can have a similar moderating influence on otherwise
continental climates, though their influence is less than that of oceans. Perhaps the best example
of the moderating influence of large lakes occurs in the Great Lakes region. Climate on the eastern
shore of Lake Michigan has a higher average annual temperature, higher average temperature dur-
ing the summer, and less extreme minimum temperatures in the winter compared to only 10 miles
(16 km) further inland. This region is renowned for its production of cherry crops that are other-
wise produced in the maritime climate of the Pacific Northwest. Even a small inland lake
(<8 mi2/21 km2) in the center of Michigan’s Lower Peninsula has been shown to slightly moderate
the climate on its eastern shore (Albert et al. 1986).
In addition to latitude and proximity to large water bodies, altitude has a strong effect on
temperature at broad scales. Temperature generally decreases with increasing elevation in much
the same way that it decreases with increasing latitude; a long-used conversion factor is that 100 m
of elevation is equivalent to 1° of latitude in terms of temperature (Wiersma 1963). As such, the
form of woody plants near the tree line in mountainous terrain begins to resemble those in northern
boreal and subarctic regions. Latitude and altitude do not act independently on vegetation, and the
variation in tree line across the globe is a good indicator of how altitude and latitude work in
concert. In general, tree line occurs at lower elevations at high latitudes and at higher elevations
near the equator, and this trend is less extreme for mountains in maritime climates compared to
continental climates (Figure 7.1). Although Grace (1989) hypothesized that trees cannot grow
where the combination of latitude and altitude results in an average temperature of the warmest
month <50 °F (10 °C), mean growing season temperature worldwide at tree line occurs between 42
and 45.5 °F (5.5 and 7.5 °C; Körner and Paulsen 2004).
One important factor determined by temperature at a particular area is growing season
length. The growing season length for vegetation in a particular area is the number of days when
aboveground plant growth is able to take place during the year. Quantitatively, growing season
length is the number of days when soil temperature at a depth of 50 cm (20 in.) is greater than 5 °C
(41 °F). However, it is often more simply estimated as the number of days between the average date
of the last killing frost recorded in the spring and the first killing frost recorded in the fall, where a
killing frost is considered to be −2 °C (28 °F) or lower. As one might expect, growing season length
is longest where freezing temperatures are least likely (the tropics, maritime areas, or at lower ele-
vations) and shortest where they are most likely (closest to the poles, continental areas, or at higher
elevations). Notably, some tree processes occur well outside of their growing season, such as respi-
ration and even photosynthesis in conifers.
Another important aspect of temperature for vegetation is the heat sum, which is a mea-
sure of the amount of heat to which plants are exposed. A heat sum for a given species in a
given place is typically measured in degree days. Degree days are calculated by summing over
Temperatures at the Soil Surface 145
6000 snow line

tree line
5000
4000
Altitude (m)
3000
2000
1000
0
70° 60° 50° 40° 30° 20° 10° 0° 10° 20° 30° 40° 50° 60°
North Latitude South
F I G U R E 7 . 1 Tree-line (and associated snow line) position around the globe by latitude. With some
variation, tree line occurs at lower elevation nearest the poles and higher elevation nearest the equator.
Source: Körner (1998) / with permission of Springer Nature.
the course of a growing season the product of (1) the number of degrees over a base tempera-
ture in degrees Celsius and (2) the number of daylight hours of that temperature. The most
common base temperatures used for calculating a heat sum are 5 °C (41 °F) and 10 °C (50 °F),
with the assumption that significant plant growth will not occur below the base temperature.
The importance of the heat sum for woody plants is that an important event will occur once a
certain heat sum has been achieved, such as flowering, bud burst in the spring (Worrall 1983;
Hari and Häkkinen 1991), or other important phenological events. Therefore, populations of a
species in a location are able to adapt to and are able to monitor the temperature regime of that
location. For example, red maple is one of the earliest flowering trees in the spring and has been
documented to flower at 1–27 degree days, but white ash in the same location flowers at 30–50
degree days (Burns and Honkala 1990). Importantly, however, northern and southern popula-
tions of widely distributed species of woody plants are able to adapt their respective heat sums
to the temperature regime at hand, which eventually plays an important role in genecological
differentiation (Chapter 3).
TEMPERATURES AT THE SOIL SURFACE

When we speak of variation in soil temperature, we are speaking chiefly about warming of the
ground exposed to the sun during daylight and its subsequent cooling at night. Solar radiation is
absorbed during the day both by the soil surface and the air immediately above it, and these
processes largely determine daytime temperatures. At night, however, temperatures depend on
the amount of heat absorbed by terrestrial objects including soil and how quickly the heat is lost
to the atmosphere (typically as infrared radiation) once solar radiation is no longer present. These
factors are in turn determined by the amount of vegetation and litter cover. Soil temperature is
therefore influenced both by the physical properties of the soil itself as well as the climate in
which the soil occurs. From a vegetation perspective, soil temperature is important because it
affects the biological activities of plant roots and the soil organisms that are mutualistic with
them. Most chemical and biological processes increase with temperature, such that soil tempera-
ture has a strong effect on the ability of tree roots to grow or uptake water and nutrients, and on
microbial and other activities that affect important belowground processes such as mineralization
and decomposition (Davidson et al. 1998; Heinze et al. 2017). Soil temperature would also be
expected to have an important influence on the establishment and survival of tree seedlings.
The amount of water contained in a soil strongly controls its temperature regime. The heat
capacity of soils, or the amount of heat necessary to raise its temperature, has many implications
for the effects of soil moisture on soil temperature. As discussed in the earlier text, water has a high
heat capacity relative to air, meaning that almost five times as much heat is required to raise the
temperature of a volume of water compared to the same volume of air, and thus heating and
cooling occur more slowly in wet soil compared to dry soil. In addition, heat moves from warm
areas to cold areas, and the rate of this movement is known as thermal conductivity. Organic
soils have lower thermal conductivity compared to mineral soils, and soils with smaller particles
(clays) have lower thermal conductivity compared to sandy soils (Weil and Brady 2017). As a result,
tree roots found on drier, sandier soils are more likely to suffer freezing injuries compared to wetter
or more clayey soils because the latter retain their heat longer into the fall. Finally, soils lose an
enormous amount of heat when water evaporates from the soil surface during the day. A study by
Maguire (1955) showed that watering bare soil in California, for example, resulted in its cooling by
9–23 °C (48–73 °F) over the next three days.
Soil temperature changes with depth, with surface soil much more prone to variability in
temperature compared to subsurface soil. Subsequently, deeper soils take much longer to reach
maximum and minimum soil temperatures compared to the surface, if they ever achieve those
extremes at all. Temperatures measured deep in the soil frequently remain constant throughout
the year and approximate the mean annual temperature of the area (Poulson and White 1969).
Soil color also affects soil temperature. Dark soils absorb more radiant energy than light
soils, which have a higher albedo (Oke 1987). The importance of soil color for temperature was
classically demonstrated by Isaac (1938) in southern Washington. He found that soils blackened by
recent fire reached a temperature of 73 °C (163 °F) when the air temperature was 38 °C (100 °F), but
comparable gray mineral soil heated up only to 64 °C (147 °F) and yellow mineral soil to 62 °C
(144 °F). These trends logically extend to wetter soils, with a slightly dampened effect (so to speak).
A higher moisture content of soil darkens their color, such that wetter soils absorb more solar radi-
ation than drier soils, but the high heat capacity of the soil water results in a slowed warming of
the wetter soils.
Of course, the soil surface is often not barren and is in fact covered with layers of organic
matter. Organic layers on the forest floor tend to heat very rapidly during the day without transmit-
ting much of this heat to the soil below it. At night, heat loss from the mineral soil is reduced by an
organic layer, which itself cools quickly without incoming radiant energy. This tendency for high
variability in temperature is one reason that organic matter is not an ideal place for seedling estab-
lishment, as either its high daytime temperatures in the summer or the freezing nighttime temper-
atures in the fall are potentially lethal to young seedlings. In eastern Connecticut, a surface of
eastern white pine needles litter in a small clearing was recorded to be 68 °C (154 °F) on a day when
the air temperature reached only 24 °C (75 °F), but the surface temperature of bare mineral soil was
46 °C (115 °F) and that of a moss surface layer reached only 39 °C (102 °F) (Smith 1951). Similarly,
soil at 10 cm depth beneath bare mineral soil in high-elevation spruce–fir forests of British Colum-
bia was found to be 8 °C (46 °F) warmer than sites covered by forest floor and vegetation (Balisky
and Burton 1995). Snow is also an excellent insulator and reduces heat loss from the soil during the
winter months (Hardy et al. 2001).
A striking example of the importance of soil temperature is found for forest ecosystems of
the Chena River floodplain in interior Alaska (Viereck 1970). The annual soil temperature regime
varies markedly for a sequence of ecosystems dominated by willow, balsam poplar, white spruce,
and black spruce at successive distances from the river. Near the river’s edge the gravelly soil is cold
in winter and warm in summer. However, this wide annual fluctuation is reduced inland until in
soil under black spruce, soils are cold throughout the summer and waterlogged when not frozen.
Temperature within the Forest 147
Many other examples could be given of the importance of soil temperature, illustrating not only
differences among landscape ecosystems but also its influence in bringing about vegetation
changes over time at a specific site.
TEMPERATURE WITHIN THE FOREST

Broad-level patterns in temperature are but one aspect of climate that sets the template for smaller,
local variations in temperature experienced by the woody plants in a forest. When considering the
effects of water as well as temperature, these local variations in climate are termed microclimate.
Microclimate influences the distribution of local ecosystems as well as the plants that occur within
and among them. The most obvious differences in microclimate are those that occur with slope
and aspect in topographically diverse terrain, which we discuss as part of physiography in
Chapter 8. Here and in the next section, we discuss the importance of the forest itself in affecting
the temperature component of microclimate. In the following section, we discuss the influence of
local topography on microclimate.
Within a forest, the canopy and tree stems shelter the region near the ground from solar radia-
tion and reduce the mixing of air, such that the temperature and its variation are significantly differ-
ent from areas in the open (Geiger et al. 2009). Shading by the forest canopy results in reduced
warming of air temperature during the day, but infrared heat re-radiated from plants and soil at night
tends to be captured beneath the canopy. As a result, air temperature warms less during the day and
cools off less during the night compared to open areas, and the forest canopy therefore has a moder-
ating effect on the air and soil temperature beneath it. Relative humidity is closely related to air tem-
perature and also increases at night beneath a forest canopy. Data collected within and adjacent to an
eastern white pine plantation by Spurr (1957) gave a summer air temperature range of 15.9 °C
(60.2 °F) within the forest compared to 21.6 °C (70.9 °F) in the open (Table 7.1). Throughout the year,
at a height of 1 m from the ground, the maxima and means were lower and the minima higher within
the forest as contrasted to the open station. The mobility of air, even beneath a forest canopy, allows
some degree of mixing with nearby areas such that the shading provided by the canopy has a larger
effect on soil temperature than air temperature (Morecroft et al. 1998; Porte et al. 2004). Many impor-
tant forest processes respond to the moderating effect of forests on air temperature, soil temperature,
and relative humidity, including the survival and growth of tree seedlings.
Table 7.1 Mean weekly maxima, minima, and mean temperatures (°C) in the open and under a dense
20-year-old eastern white pine plantation.a
Winter Spring Summer Fall

Open
Maximum 5.1 22.8 29.7 14.2
Minimum −18.4 −2.2 8.1 −7.3
Range 23.5 25.0 21.6 21.5
Mean −6.7 10.4 18.9 3.4
Under forest
Maximum 2.7 19.9 25.6 11.0
Minimum −16.7 −2.8 9.7 −5.7
Range 19.4 20.2 15.9 16.7
Mean −7.1 9.8 17.7 2.7
a
Petersham, MA; by 13-week quarters, 1943–1944.
Forest structure and species composition have important effects on the strength of the mod-
erating effect on air temperature, in large part because their differences in leaf area influence the
degree of absorption of incoming solar radiation. In a study of deciduous, spruce–fir, and pine
forests, von Arx et al. (2012) found that forest cover in general decreased daily maximum air tem-
perature as much as 5.1 °C (41.2 °F). Deciduous and non-pine conifer forests moderated daytime
maximum and average air temperature about twice as much as pine forests, and pine forests cooled
significantly less at night. The moderating effect of the canopy varies with its density regardless of
species composition, with higher-leaf area canopies having a greater moderating effect (von Arx
et al. 2013). Forests having old-growth characteristics (tall canopies, high biomass, complex vertical
structure) had reduced maximum temperature in the warmest month of the growing season,
reduced mean monthly maximum temperature, and reduced variability in mean weekly tempera-
ture (Frey et al. 2016).
Temperature also varies vertically within a forest. Many researchers have documented
changes in temperature with increasing height above the ground beneath both deciduous and
coniferous forest canopies (Harley et al. 1996; Zweifel et al. 2002), but it is important to recognize
the effects of the canopy itself on temperature variation. The classic study of variation in tempera-
ture along a vertical gradient was performed in an American beech-dominated forest in Ohio by
Christy (1952). Summer and fall temperatures were similar at the soil surface and beneath the
canopy, but the widest range of temperatures occurred in the canopy itself. In the winter and
spring, temperature variation was similar at all vertical strata because of the absence of the canopy
that allowed for greater mixing of the air within the forest. The effects of both changing tempera-
ture and varying temperature at increasing distance from the forest floor are significant for their
potential effects on photosynthesis and carbon capture within the canopy (Ellsworth and
Reich 1993; Bauerle et al. 2009).
TEMPERATURE VARIATION WITH LOCAL TOPOGRAPHY

The effects of local topography upon local temperature have long been studied (most recently
Kelsey et al. 2019; Arduini et al. 2020; and Rupp et al. 2020). Common among these studies is the
pattern that low, often concave landforms accumulate cold air on still, cold nights. Cold air is
denser and flows downward from adjacent high land. Radiant cooling on high ground forms a
shallow layer of cold air near the ground, and minimum temperatures at night remain fairly high.
When cold air pools in narrow valleys or small, closed depressions, these processes may result in
air temperatures near the feature’s bottom as much as 8 °C (46 °F) lower than that of the “thermal
belt” further up the slopes. The resulting condition is known as an inversion in mountainous ter-
rain, in that the temperature increases with elevation in contrast to the usual decrease in tempera-
ture with elevation. Smaller concave areas are known as frost pockets because the temperatures
commonly occurring near the ground result in late spring frosts, early fall frosts, and consequent
short growing seasons. These areas of cold air pooling need not be deep. Spurr (1957) found that
minimum temperatures in a small depression only 1 m deep were comparable to those in a nearby
deep valley 60 m below the general land level. Interestingly, the same low landforms tend to accu-
mulate radiant energy during the day and reach high maximum temperatures near the ground,
while mounds, ridges, and other higher convex surfaces tend to remain cooler.
An excellent example of the effects of cold-air drainage on forests can be found in the jack
pine forests of northern Lower Michigan. Jack pine is tolerant of low-nutrient, dry, cold sites.
Northern pin oak is a secondary tree species that also tolerates dry, infertile sites, but is far more
sensitive to frost. Ice-block depressions on this landscape are sites where cold air frequently pools
at night. Even shallow depressions exhibit distinctive effects of this phenomenon, where regener-
ation of pine is sparse near the bottom of the depressions and oak is limited by freezing tempera-
tures (Kashian et al. 2003). Where the depressions are deep, the slopes of the depressions are often
Temperature Variation with Local Topography 149
steep enough to exhibit the effects of aspect. South-facing slopes of these deep depressions are
often very dry and devoid of trees, and oaks are limited to the higher terrain (Figure 7.2).
Despite the importance of local topographic influence on temperature variation for the dis-
tribution of plants and tree growth in frost pockets, soil-water drainage is likely to be an important
interacting factor. The same characteristics of a landform that ensure cold-air drainage, such as the
existence of convex surfaces, will make a soil very well drained. Thus, a very well-drained site is apt
Photos by Dan Kashian
F I G U R E 7 . 2 Effects of microclimate on vegetation as mediated by topography. The top image shows a

shallow ice-block depression (approximately 6 m/20 ft deep) after a recent wildfire in a cold, continental region
of Lower Michigan. The center of the depression lacks pine regeneration due to frequent frosts during the
growing season. Note the bushy, resprouting oaks that are frost-sensitive and limited to the upper rim of the
depression. The lower image is a large ice-block depression (approximately 26 m/85 ft deep); the sides of the
depression are steep enough to lack significant tree growth on its south-facing slope (foreground). Again, note
the oaks (seen at the horizon) are limited to the very top of the depression on the north-facing slope.
to have a warmer microclimate suitable for plants of a more southern distribution. Likewise, a very
poorly drained soil is apt to result from a concave land surface where both cold air and water will
pool. Very poorly drained sites are apt to be characterized by temperature extremes and a short
growing season and might prove suitable for plants of a more northern distribution. Such exam-
ples are common in northern Ohio and southern Michigan where boreal and northern forest
species (black spruce, tamarack, northern white-cedar, black ash, yellow birch, leatherleaf, among
others), relics from their late-glacial northward retreat, persist in wet depressions that also act as
frost pockets.
As a final note, cold-air drainage may be influenced by the vegetation itself as well as by local
topography. Given that a forest moderates the area beneath its canopy, as discussed in the earlier
text, the canopy tends to warm more during the day than the soil surface, and cold-air drainage
occurs in the subcanopy and understory while the warmer air above the canopy moves upslope
(Froelich and Schmid 2006). As the canopy cools faster than the ground at night, this cold-air
drainage is likely to weaken (Mahrt et al. 2000). Similarly, a phenomenon resembling topographic
frost pockets may result from the creation of small clearings in a forest by windthrow, insects or
diseases, or cutting. The tree crowns of the surrounding forest act to channel cold air into the clear-
ings on still, clear nights; larger clearings are more prone to winds that will help to drain the cold
air away and are thus less likely to pool the cold air.
TEMPERATURE AND PLANT GROWTH

We often consider plants as entire organisms and how their growth might be affected by external
factors such as temperature. In reality, different parts of the organism experience temperature
differently, especially when the plant is large (such as an overstory tree). In that the canopy of a
forest moderates the temperature beneath it, we would expect that the parts of a tree that occur
below the canopy would experience temperature differently than the canopy itself. The temper-
ature of tree stems, for example, approximates the air temperature in the shade, but may be
quickly heated by sunlight. High tree-stem temperature of sunlit trees may be modified by
upward water movement through the xylem, and bark—depending on its thickness and color—
may keep stems warmer than the ambient air temperature in the winter and cooler in the summer
(Nicolai 1986).
Tree roots experience temperature of the soil environment; as discussed in the earlier text,
soil temperature is typically cooler than the air in the summer and warmer in the winter. Roots are
more tolerant to a wider range of temperatures than are shoots, although optimum temperatures
for root growth vary by species (Gray and Brady 2016), even for species growing on the same site.
Thus, roots are able to grow well into the fall after shoot growth has ceased, and can begin growth
in the spring when soil temperatures remain cold but air temperatures have stimulated shoot
growth. In a study of six European tree species, Alvarez-Uria and Körner (2007) found that most
root growth occurred at temperatures warmer than 9 °C (48 °F) and was restricted below 6 °C
(43 °F). Root growth of ponderosa pine is more dependent on soil temperature and shoot growth
more dependent upon air temperature (Larson 1967). Optimum root growth occurred in 15 °C
(59 °F) air and 23 °C (73 °F) soil. Optimum root growth for northern red oak, basswood, and white
ash also occurs at relatively high soil temperatures (Larson 1970). Temperature of the crown also
affects root growth, because increases in respiration and transpiration that occur with increasing
temperature utilize carbohydrates and water, making them less available for use by the roots. For
example, root growth of silver maple seedlings decreased markedly as temperature of the crown
increased from 5 to 30 °C (41–86 °F; Richardson 1956).
In contrast to stems and roots, the leaves of plants are able to regulate their temperature by
dissipating some of the energy they absorb as a way of avoiding injurious or lethal high tempera-
tures. Re-radiation alone dissipates approximately 50% of absorbed heat energy (Gates 1980).
Temperature and Plant Growth 151
Additional heat loss is achieved through transpiration, which cools the leaf as a large amount of
heat energy is consumed in changing water to water vapor. A third process, convection, transfers
heat from the leaf to the surrounding cooler air as it moves across the leaf’s boundary layer (the
thin layer of still air found on all surfaces). Cooler nighttime air temperatures relative to the leaf
surface reverse this process, and heat is transferred from the air to the leaf by both convection and
conduction (Nobel 2009). We will discuss a series of plant adaptations that interact with these
processes to stabilize leaf temperature in a given environment, maintaining heat in cooler air and
facilitating heat loss in warmer air.
As temperature increases, plant activities increase up to an optimum temperature and then
decrease until lethal high temperatures are reached. Temperate tree species generally increase
their rate of photosynthesis from freezing to an optimum at 15–25 °C (59–77 °F; Pallardy 2008).
Notably, respiration often peaks at temperatures higher than photosynthesis, such that net photo-
synthesis is inhibited. Plant processes generally function as long as living cells and their proteins
are stable and their enzymes active, usually 0–40 °C (32–104 °F) depending on species, age, season,
and population source. Physiologically, temperature most strongly influences (i) the activity of
enzymes that catalyze photosynthesis and respiration, (ii) the solubility of carbon dioxide and oxy-
gen in plant cells, (iii) transpiration, (iv) the ability of roots to absorb water and minerals from the
soil, and (v) membrane permeability. A detailed discussion of effects of temperature on photosyn-
thesis is presented in Chapter 18.
Plant growth and metabolism depend on water in liquid form, and these processes slow or
stop when water is frozen into ice. Photosynthesis can occur when the ambient air temperature is
below freezing, down to about −8 °C (18 °F), because solar and terrestrial radiation often warms
tissues to near or above freezing. Low temperatures inhibit the rate of photosynthesis, but appre-
ciable photosynthesis may occur in winter in conifers. For example, although the optimum tem-
perature range for photosynthesis in Douglas-fir is between 10 and 25 °C (50 and 77 ºF), net pho-
tosynthesis at 0 °C is still 70% of the amount occurring at 10 °C (Lassoie 1982). Winter photosynthesis
in many conifers, however, is limited because low soil temperatures tend to reduce metabolic
activity and reduce membrane permeability so that uptake of water and nutrients needed for pho-
tosynthesis is low.
Photosynthesis decreases when temperature increases beyond the optimum, respiration
eventually exceeds photosynthesis, and cell death occurs when temperatures approach 55 °C
(131 °F). These temperatures are those of the plant tissues themselves, not air temperature, and
thus air temperatures must be even higher to warm up the plant tissues to the lethal point. In this
manner, prolonged exposure to high but sublethal air temperatures may eventually kill plants that
might survive short exposure to very high air temperatures. Exposure time is critical for survival of
seedlings because extremely high temperatures may be reduced by even the lightest shade, as is the
case when cotyledons of the seedlings shade the basal portion of the seedling stem. A good
treatment of high temperature stress and heat injury of plants is presented by Levitt (1980a) and
Pallardy (2008).
As discussed in Chapter 6, leaf arrangement and orientation change in response to light
irradiance and in the process may reduce the amount of solar energy absorbed and hence prevent
overheating of the leaf. Red maple seedlings are able to deflect their leaf blades almost vertically
downward in high light and return to the horizontal position when shaded (Grime 1966). This
process may partially explain the versatility of red maple in colonizing both dry, exposed sites
and shaded habitats. Leaf morphology, coloration, leaf pubescence, and maintenance of protein
integrity also enable woody plants to function effectively at high temperatures (Levitt 1980a).
Near-lethal temperatures for most woody plants are largely confined to the exposed ground-
air boundary in forests. As a result, direct heat injury in forest trees occurs most significantly in
small seedlings, which have relatively unprotected live tissues in this critical zone. Heat damage in
seedlings occurs when the soil surface reaches about 52 °C (126 °F; Helgerson 1989). High temperatures
may indirectly cause damage to trees via their effects on water loss, particularly from leaves. Many
mature hardwoods and conifers suffer leaf damage due to water deficiency of cells along the leaf
margin and the tips of conifer needles. For example, many woody species suffered leaf damage in
northern California when temperatures suddenly rose above 38 °C (100 °F) in areas that had had
an exceptionally cool spring (Treshow 1970).
Various phases of the temperature regime, such as day temperature, night temperature, heat
sums, and the difference between day and night temperature, all affect growth, each in a species-
specific or even a population-specific way. A series of experiments in the 1950s and 1960s demon-
strated that the response of tree seedlings to the temperature regime varied widely by species
(Kramer 1957; Hellmers 1962, 1966a,b; Hellmers et al. 1970). Engelmann spruce seedling growth
responded most to night temperature (Figure 7.3), but redwood responded to day temperature,
reaching its maximum growth in the moderate range of 15–19 °C (59–66 °F). Both species grew
best at the same day temperature (19 °C/59 °F), but spruce grew better when nights were warm and
also tolerated warmer days better than redwood. Redwood grew best when day and night temper-
atures differed only slightly, consistent with the lower diurnal variation in its native maritime cli-
mate. By contrast, Engelmann spruce was apparently adapted to the greater diurnal changes of
continental, high-elevation environments.
Several species (Jeffrey pine, erectcone pine, eastern hemlock) show a marked growth response
to the heat sum during a day, regardless of the time of application. For example, Jeffrey pine exhibit
the most growth when 300–400 degree-hours are accumulated, regardless of the specific different day
and night temperatures. Still other species show a primary growth response when nights are consid-
erably cooler than days. This difference between day and night temperatures is termed thermope-
riod. Low night temperatures coupled with moderate day temperatures are important in the
flowering and fruit set, flavor, and quality of various crop plants and fruit trees (Treshow 1970).
Strong responses to thermoperiod are common in seedlings of many forest trees. Maximum
shoot growth of loblolly pine seedlings occur with thermoperiods of 12 °C (54 °F), with night
F I G U R E 7.3 Average height of 36

Engelmann spruce seedlings from each of
30 combinations of day and night tempera-
tures where the plants were grown for
24 weeks. Source: Hellmers et al. (1970) /
with permission of Oxford University Press.
200
150 23
Height, mm
19
, °C
100
ure
15
rat
pe
11
tem
50
ht
Nig
0 3
35 27 23 19 15
Day temperature, °C
(a) (b) (c)

Pacific coast Southwest Rocky Mountains
80 Placerville, Cal. Safford, Ariz. Colorado, Spgs.
elev. 550 m elev. 2775 m elev. 2165 m
(1800 ft) (9100 ft) (7100 ft)
60
Height, mm
40
17 17 17
20
23y 23 23
y
Da
Da
Da
0 30 30 30
22 14 (10) 7 22 14 (10) 7 22 14 (10) 7
°C °C °C
Night Night Night
F I G U R E 7 . 4 (a)–(c) Mean height growth of ponderosa pine progenies from three provenances in three regions.
Treatment included 16-h days and nine combinations of three day and three night temperatures. The vertical bars
show the range of the mean ± standard errors of the mean. Source: Callaham (1962) / John Wiley & Sons.
t emperature colder than day temperature (Kramer 1957). A similar response has been consistently
reported for Douglas-fir, although the optimal temperature differential has varied among the
studies conducted and the provenances tested (Hellmers and Sundahl 1959; Lavender and
Overton 1972). Atlantic white-cedar and Fraser fir both maximize total seedling growth at a ther-
moperiod of 8 °C (46 °F) with warm days and cool nights (Hinesley 1981; Jull et al. 1999). Red fir
exhibit two effects of thermoperiod (Hellmers 1966a). First, for maximum height growth to occur,
a warm day must be followed by a cool night, whereas a cool day must have a cold night. Second,
maximum height growth is obtained when the thermoperiod is 13 °C (55 °F). Although maximum
growth under a 17 °C (63 °F) day and a 23 °C (73 °F) day is nearly equal, this growth occurs only
when the cooler day was followed by a 4 °C (39 °F) night and the warmer day with a 10 °C (50 °F)
night (a 13 °C thermoperiod in both cases).
In contrast to the three species previously described in the earlier text requiring cold nights
and warm days, seedlings of ponderosa pine grow best when nights are warmer than days (Calla-
ham 1962; Larson 1967). Furthermore, pines from diverse parts of the range show significantly
different responses to the temperature regime (Figure 7.4). Seedlings from east of the Rocky Moun-
tains (Figure 7.4c) require a high night temperature for optimum growth, but seedlings from the
Southwest grow remarkably fast under cold days and hot nights. Meanwhile, seedlings from the
west slope of the Sierra Nevada Mountains in California maximize growth with lower night tem-
perature. Similarly, red maple height growth varies in response to day and night temperatures
based on where it was located across its range (Perry 1962).
COLD INJURY TO PLANTS

Lethal cold temperatures occur periodically throughout the entire zone of tree growth in the
temperate and boreal regions of the Earth. Thus, they affect to a greater or lesser degree the distri-
bution and growth of trees in these zones. In addition to growing season temperature and water
availability, cold temperatures may be important for the distribution of temperate trees (Chapter 2).
For example, the northern and western extents of loblolly pine are strongly correlated with a
combination of cold temperatures occurring simultaneously with low soil water (Hocker 1956).
Plant tissues, particularly those sensitive or actively growing, may be damaged or killed
when freezing occurs quickly and ice crystals form within the protoplasm. Even slow freezing may
kill plants at temperatures of −15 °C (5 °F) to −45 °C (−49 °F) when cooling occurs at rates com-
monly found in nature (Weiser 1970). During this process, intercellular water freezes, dehydrating
cell contents until a point is reached when only “bound” (unfreezable) water remains in the proto-
plasm (Wolfe et al. 2002). Further decreases in temperature pull the bound water away from the
protoplasm, initiating protein denaturation and ultimately killing the cell. In tropical plants, death
may occur at above-freezing temperatures ranging from 0 to 10 °C (32–50 °F). Recent work has sug-
gested that the primary sites of freezing damage in plants are cell membranes. Frost injury is indi-
cated first by an increase in the leakage of electrolytes across the plasma membrane and out of
cells. Lowering temperatures also affect the phase and configuration of the lipid bilayer of the cell
membrane (Uemura et al. 2006; Ambroise et al. 2020).
Freezing temperatures have strong effects on whole plant organs as well as cells and tis-
sues, and their extent of damage depends strongly on the timing of their occurrence. If plants
are flowering, freezing temperatures may kill flowers and thus limit fruit production. For
example, a study of Gambel oak in Utah found that low-elevation oaks suffered freezing of
their catkins and a subsequent failure of the acorn crop, while oaks at high elevations did not
(Neilson and Wullstein 1980). Freezing can also kill or damage portions of the cambium in
tree stems. Outer portions of the stem freeze and thus contract more quickly than the inner
portions, which can create cracking of the stem. If freezing occurs during the growing season
before xylem cells in the stem have thickened and lignified, ice formation between cells may
crush the outermost zone of weaker cells and leave a scar within the stem that is evident in a
stem cross section. These “frost rings” have been used to identify significant past cooling
events during the life of a tree, such as the eruptions of major volcanoes (LaMarche and
Hirschboeck 1984).
Aboveground portions of the plant are most commonly the focus of studies of cold injury
to plants, but roots are similarly or even more affected. Root damage can occur for a variety of
physical reasons related to soil moisture in cold temperatures, but like shoots, cellular damage
is more likely the reason for cold injury than mechanical damage. Fine roots are most suscepti-
ble to freezing damage, which are responsible for water and mineral uptake and whose regrowth
uses resources at the expense of the crown (Gaul et al. 2008). An important difference between
roots and shoots is that shoots tend to increase their frost hardiness as the tree ages, whereas
the roots do so much less. For example, some oak species exhibit an increase of 9–10 °C (48–
50 °F) of frost hardiness of the shoot cambium over a seedling’s first three years, but roots
increase only 1–2 °C (34–36 °F; Sakai and Larcher 1987). This difference appears to be related to
a higher water content in roots compared to shoots and the added insulation provided by the
soil. A detailed discussion of the frost hardiness and cold tolerance of roots is provided by Am-
broise et al. (2020).
Most trees in the temperate and boreal zones become increasingly inactive as the day length
shortens and temperatures decrease near the end of the growing season. Protoplasm water content
is reduced as the concentration of other cytoplasmic materials increases during the onset of dor-
mancy. Water moves out of cells and freezes in extracellular spaces, leaving the cells undamaged,
and thus many species are able to survive and even avoid injury by subfreezing temperatures. This
ability of individuals to initiate and continue the reduction of water in their cells to correspond
with the progressively decreasing temperatures of their native site in autumn is critical to the cold-
acclimation process. Cold hardiness occurs slowly, and modern plants have been selected naturally
to survive the gradual lowering of temperatures in autumn and winter and their gradual rise
in spring.
DORMANCY
We describe plants that have transitioned to a hardy or dormant condition from a tender or
sensitive condition as acclimated or hardened. Three stages of dormancy are normally recog-
nized (Vegis 1964; Perry 1971; Sakai and Larcher 1987): pre-dormancy (early rest), true dor-
mancy (winter rest), and post-dormancy (after rest) (Figures 7.5 and 7.6). Short days trigger
growth cessation and initiate metabolic changes characteristic of pre-dormancy at temperatures
of 10–20 °C (50–68 °F; Figures 7.5 and 7.6). During this stage, growth is reduced, and starch
accumulates in the roots (Dumont et al. 2011). Pre-dormancy may be reversed by warmer tem-
peratures and increased photoperiod. Otherwise, these changes facilitate further metabolic
changes in the second stage of acclimation, sometimes called full dormancy, triggered by tem-
peratures below about 5 °C (41 °F) and especially subfreezing temperatures including the first
frost of autumn (Figure 7.6). Full dormancy involves the production of proteins, membrane
lipids, and metabolites involved in frost hardiness (Beck et al. 2004). Plants have increased cold
resistance in full dormancy, which can be reversed after a chilling requirement. The third stage
of acclimation is the attainment of true dormancy, which is exclusive to woody plants. This
stage is induced by low temperatures of −30 °C (−22 °F) to −50 °C (−58 °F). During this stage,
water remaining within the cell vitrifies or turns to a glassy material lacking the crystalline
structure of ice that is lethal to cells (Strimbeck et al. 2015). Truly dormant buds and seeds
cannot be immediately induced to normal growth. Some species, such as birches, European
Shoot
development Days
shorter
Drought
Temp.
tochrom
Phy e falls
Bud primordia
Shoot growth
Growt
Genes
d o minate
opening
stopped
formed
Bud
h p ro m .
Enzymes
p re
p re
rs
om
ito
d
in a ib
te In h
Temp.
rises
Postdormancy Predormancy
True
Days dormancy
longer
Cold weather
F I G U R E 7 . 5 External influences (shaded U) and internal interactions affecting the seasonal alternation of
vegetative and floral activity and dormancy in woody plants. Note that the growth inhibitors that influence
bud formation in the fall and the growth promotors that initiate bud opening in the spring are activated by
the phytochrome system. Source: Sakai and Larcher (1987) / Springer Nature.
Beginning first stage of acclimation F I G U R E 7.6 A typical seasonal pattern of

0 Beginning second cold resistance in the living bark of red-osier
stage of acclimation dogwood (Cornus sericea) stems in Minne-
sota. In nature, acclimation in this hardy
First autumn
frost shrub and in a number of other woody
–25
species proceeds in two distinct stages, as
Lowest survival temperature, °C
shown. The beginning of the second stage of

acclimation characteristically coincides with
–50 the first autumn frost. Source: Weiser (1970) /
with permission of American Association for
the Advancement of Science.
–75
–100
Sept. Oct. Nov.

Date
beech, and some oaks, do not enter true dormancy and will pass readily from pre-dormancy to
post-dormancy. Plants in their native ecosystems are not at risk from damage when truly dor-
mant. However, plants are susceptible to severe injury or death from abrupt departures from the
general temperature rise or decline. Alternate descriptions of the phases of dormancy are
common in the literature (e.g., Lang et al. 1987).
FROST HARDINESS AND COLD RESISTANCE

The timing of the cold acclimation process is a major factor in the development of frost hardiness,
and geographical races within a species may develop as a result. The more rapid changes in day
length that occur in northern latitudes lead to earlier onset of the acclimation process in northern
races, which are then more resistant to freezing compared to southern races. Examples of eastern
tree species that have developed racial differences in forest hardiness include eastern cottonwood
(Mohn and Pauley 1969), red maple (Perry and Wang 1960), sugar maple (Kriebel 1957), sweetgum
(Williams and McMillan 1971), and white ash (Alexander et al. 1984), among others. Flint (1972)
studied the hardening of northern red oak populations grown from seed and found a strong rela-
tionship between the temperatures at which stems are killed by frost and the latitude, annual
minimum temperature, and estimated extreme minimum temperature of their original site. Oaks
whose sources experienced colder average minimum temperatures required colder temperatures
to kill the stems (Figure 7.7). When examining the progression of hardening and de-hardening of
twigs from four populations at the extremes of the range, northern sources hardened earlier
(Figure 7.8).
Geographic races are genetically adapted to minimum temperatures in midwinter as well as
to the continuously changing gradient of temperature (Chapter 3; Figure 7.8). Overall, species are
able to resist freezing injury well below the minimum temperatures they typically confront in their
native sites. Woody plants of the boreal and northern forest (paper birch, spruces, pines, tamarack,
trembling aspen, willows) survive prolonged subzero temperatures. Amazingly, if the freezing rate
to −30 °C (−22 °F) is gradual and plants are fully acclimated, these and numerous hardy species
are known to withstand experimental freezing to −196 °C (−321 °F), the temperature of liquid
50
45
December 12. 1968
ý = 0.548x + 29.0
F = 31.9***
40
35
Temperature
November 14.1968
(°C)
ý = 0.339x + 18.0
F = 30.4***
30
October 8.1968
ý = .457x + 14.0
F = 34.8***
25
20
15
15 20 25 30 35
Estimated Average Annual Minimum Temperature
at Origin (–°C)
F I G U R E 7 . 7 Regression of killing temperature on estimated average annual minimum temperature at

seed-source origin of northern red oak at three sampling dates: 9 October (○), 14 November (●), and 12
December (□) 1969. Source: Flint (1972) / John Wiley & Sons.
nitrogen (Weiser 1970; Sakai and Weiser 1973). In contrast to boreal and alpine species, many trees
from regions characterized by mild fall and winter temperatures such as those of the Pacific Coast
region and the Lower Mississippi Valley and southern Coastal Plain are injured at temperatures
colder than −20 °C (−4 °F) to −30 °C (−22 °F).
Plants cope with freezing temperatures either by tolerating them (tolerating the presence of
ice crystals and related dehydration in extracellular spaces) or by avoiding them (preventing the
formation of ice crystals in the first place). These two mechanisms are the basis for freezing resis-
tance and may be found simultaneously occurring in the same plant (Levitt 1980a). In terms of
tolerance, freezing within the protoplasm is the only source of direct injury by freezing to plants,
and there is no tolerance to intracellular freezing at temperatures occurring in nature (i.e., it is
always fatal). However, ice formation between cells—even the formation of masses several times
larger than the cells themselves—is rather common. Thus, the main way plants achieve frost resis-
tance (“hardiness”) is by tolerating extracellular freezing, which involves both the tolerance and
avoidance of dehydration caused by cold temperatures (Figure 7.9). These mechanisms are dis-
cussed in detail by Levitt (1980a) and Sakai and Larcher (1987).
Freeze avoidance by plants occurs by either avoiding freezing temperatures altogether or by
avoiding ice formation (adaptations 2, 3, and 4 in Figure 7.9). Most obviously, plants may live
where frost is uncommon, absent, or infrequent, but they may have also insulating mechanisms
–10
–20
Temperature
(°C)
–30
Pope, Arkansas
Union, Georgia
–40 Penobscot. Maine

St. Louis, Minnesota
–50
Oct Nov Dec Jan Feb Mar Apr May Jun July Aug Sep Oct
1968 1969 1969
F I G U R E 7 . 8 Progress of hardening and de-hardening of twigs from four geographic sources representing
the corners of the natural range of northern red oak. Curves show the respective temperatures that cause
freezing death to oak twigs. Source: Flint (1972) / John Wiley & Sons.
Freezing resistance
Freezing avoidance Freezing tolerance
1) Avoidance Avoidance of (Tolerance of Tolerance of

of freezing ice formation intracellular extracellular
temperatures freezing) freezing
Avoidance 4) Avoidance of 5) Avoidance 6) Tolerance

of any intracellular of freeze of freeze
freezing freezing dehydration dehydration
2) Absence of 3) Supercooling
freezable
water
FIGURE 7.9 Six mechanisms of freezing resistance. Source: Levitt (1980a) / with permission of Elsevier.
such as bud scales and bark that provide at least initial protection. If the plant cannot avoid
freezing, it may avoid ice formation and thus freezing injury via three adaptations: antifreeze,
dehydration, and supercooling (undercooling) (Levitt 1980a). Reducing water in plant tissues, and
thus decreasing the probability of ice formation, may be achieved in part by increasing water loss
via transpiration and decreasing water uptake by roots via increased suberization. Increasing both
the proportion of unfreezable (bound) water and the concentration of solutes (especially sugars
and carbohydrates) also reduces the likelihood of ice formation. Sugar accumulation in the vacu-
oles can either prevent or reduce ice formation and thus also favor the avoidance of freeze dehy-
dration (5 in Figure 7.9).
Supercooling is an important mechanism of freeze avoidance that reduces the freezing
point of plant water below that of pure water (Levitt 1980a). The supercooling point is the lowest
subfreezing temperature attained before ice will form. Cytoplasm freezes between −1 and − 3 °C
(27–30 °F); this level of supercooling is the mode of freezing resistance in the meristematic tissues
of flower buds of many woody plants, such as azalea, blueberries, cherries, and plums (Weiser 1970;
George and Burke 1977). Supercooling represents a temporary, unstable state that may provide
protection against brief frosts (Sakai and Larcher 1987). In very small cells that are gradually hard-
ened, more persistent, deep supercooling may occur to very low temperatures (George et al. 1974;
Quamme 1985; Sakai and Larcher 1987). Deep supercooling of the ray parenchyma cells in the
xylem of many woody species prevents freezing at temperatures as low as −37 to −47 °C (−35 to
−53 °F). The distributions of 49 North American tree species were found to be related to the level
of deep supercooling (George et al. 1974; George and Burke 1976). Plants protected by this resis-
tance mechanism are limited to an area where minimum winter temperatures rarely are less than
−45 °C (−49 °F).
Freeze avoidance can also be accomplished by plants through the synthesis of specific sol-
utes within their fluids. Given that cytoplasm will freeze at lower temperatures as solute
concentration increases, these solutes are effective at lowering the water freezing point (Zacharias-
sen and Kristiansen 2000). Some extremely hardy boreal deciduous genera such as Salix, Betula,
and Populus have no supercooling mechanism to protect them. Instead, plants of these genera
permit freeze dehydration of their cells via very thin and elastic cell walls (Sakai and Larcher 1987).
Most freezing tolerance depends on the capacity of living cytoplasm to tolerate freeze-induced
dehydration even though supercooling is also an important resistance mechanism for some tissues
and species.
De-acclimation and de-hardening need to occur to move plants out of dormancy.
De-hardening occurs mainly due to increasing temperatures (Beck et al. 2004) and the length of
the dormant period (Kalberer et al. 2006). Rapid thawing of tissues and cells is harmful to cell
membranes, but de-hardening tends to occur more quickly than hardening, and the rate of
de-hardening depends on the rate of increase in temperature. Moreover, it takes relatively little
time to lose frost hardiness as the temperature rises. For example, a species of blueberry exposed to
5 °C (41 °F) during dormancy in midwinter took only seven days to lose more than 30 °C (86 °F) in
frost hardiness (Taulavuori et al. 2002). Once the de-hardening process is complete, plants may be
unable to re-harden again before an additional season of growth (Kalberer et al. 2006).
Thermotropic Movements in Rhododendrons Leaf movements and orientation are

important adaptations in response to cold temperatures and freeze–thaw cycles in the genus
Rhododendron (Nilsen 1992). This genus contains over 600 species of trees and shrubs distributed
widely across the northern hemisphere; this distribution is related to such adaptations.
Temperature-sensitive or thermotropic movements, leaf drooping, and curling into a pencil-
shaped coil (Figure 7.10) are behaviors widely reported by scientists and rhododendron
enthusiasts. Harshberger (1899) may have been the earliest to write about thermotropic
F I G U R E 7 . 1 0 Drooping and curled leaves of rosebay rhododendron in winter. Source: Nicholas A. Tonelli/
Flickr/CC BY 2.0.
ovments, observing that the petiole bends sharply downward through an angle of about 70°.
m
Overall, rhododendron species that exhibit the most intense thermotropic movements are also
the hardiest.
Leaf curling and drooping are distinct responses to climatic factors and have different eco-
logical significances (Nilsen 1992). Leaf curling is a direct response to leaf temperature regardless
of light or water availability. Leaf drooping is a response to leaf water potential, which in turn is
influenced by leaf temperature, light intensity, and soil-water stress. Thus, lower temperature or
higher light conditions result in increasing pendent behavior of the leaf. On winter days in the
southern Appalachian Mountains when air temperature is constant and slightly below 0 °C, leaf
angle of rhododendrons is vertical and leaf curling is 100% during the daylight hours. If light inten-
sity increases and raises leaf temperature, there is some uncurling. Air temperature increasing
from well below 0 °C to well above zero during the morning (diurnal changes of 24 °C/75 °F have
been documented in the southern Appalachians) causes rapid leaf uncurling and a movement of
leaves horizontally.
The ecological significance of these leaf movements is hypothesized to be the avoidance
of cold injury due to freezing and thawing. The leaves of rosebay rhododendron in the southern
Appalachians often freeze in winter at −8 °C (18 °F), but leaf temperature routinely decreases to
−15 °C (5 °F; Nilsen 1985, 1987). At the same time, daytime winter temperatures may reach
15 °C (59 °F) and initiate rapid thawing. Therefore, rhododendrons in cold northern, alpine, or
Arctic sites must have leaves adapted to tolerate both freezing and thawing, in repetition. Ther-
motropic movements may therefore be adaptations for reducing the rate of thaw. If frozen leaves
were positioned horizontally, irradiance hitting a leaf may warm them too rapidly to avoid cell
damage. It takes only seconds for leaf temperature to increase by 16 °C (61 °F) when a sunfleck
hits a rhododendron leaf (Bao and Nilsen 1988). However, thawing in response to a sunfleck or
by irradiance through a leafless canopy occurs much more slowly if the leaf is pendent
and curled.
A second advantage of thermotropic movement relates to the interaction of cold tempera-
ture and high light on photosynthesis. Rhododendron leaves are evergreen, but do not photosyn-
thesize in winter. High irradiance may cripple the plant’s photosynthetic machinery when leaves
are cold by damaging membranes of chloroplasts. This damage may impair photosynthesis by as much as
50% the following summer (Bao and Nilsen 1988). Notably, rhododendrons growing under a leaf-
less tree canopy experience the highest radiation of the year when temperatures are the coldest
(Nilsen 1985). The drooping of rhododendron leaves may therefore serve to protect cell mem-
branes from damage in high light. Moreover, leaf curling reduces leaf surface area available to
intercept light, and injury to the photosynthetic mechanism is reduced or avoided. Rosebay and
many other rhododendron species grow in dim light of forest understories, and their evergreen
leaves function for several years. Fully functional leaves are therefore critical for their survival,
which would be threatened by winter injury to their photosynthetic mechanism and a subsequent
reduction to their net carbon gain.
WINTER CHILLING AND GROWTH RESUMPTION

Buds and seeds of most temperate woody species require a period of winter chilling before
growth is resumed in the spring, typically at temperatures near 3–4 °C (37–39 °F; Harrington
et al. 2010). Once minimum temperature requirements are met, physiological changes take
place that enable growth to resume in the post-dormancy period. Plants are most sensitive to
freezing temperatures just after de-hardening and during leaf flushing (Kalberer et al. 2006),
and thus a chilling requirement prevents freeze-related injury or death by premature resump-
tion of growth triggered by a midwinter warm spell. The sensitivity of newly flushed leaves to
cold temperatures was evident in the springs of 2007 and 2010 in the eastern United States. In
those years, forests suffered major freeze damage due to uncharacteristically early-spring warm
temperatures that were followed by late-spring frosts (Gu et al. 2008; Augspurger 2009; Hufkens
et al. 2012).
The degree of dormancy attained by an individual tree and its chilling requirement vary
with the climatic adaptation of a local race of species. For example, red maples from New York
State attain true dormancy and have a chilling requirement of approximately 30 days or more
(Perry and Wang 1960). By contrast, red maples in south Florida have no chilling requirement
yet still drop their leaves in late autumn and resume growth in early spring. The southern
maples are unable to withstand severe freezing temperature, but such events are rare in
southern Florida.
When the chilling requirement is met and dormancy is broken, processes such as bud burst
(leaf flushing), bud-scale initiation, and shoot axis elongation are regulated by temperature, a pro-
cess known as forcing. In vegetative buds, dormancy breaks when mitotic activity begins, often
well before the buds flush (several weeks earlier in conifers; Owens et al. 1977). After dormancy is
broken, heat sums are very effective at predicting when a stage of development, such as flowering
or flushing, will occur or how rapidly a shoot will elongate. Notably, winter chilling and forcing
work in tandem in many woody plants: if chilling over winter is decreased, a higher heat sum is
necessary for bud burst, and longer or colder chilling periods reduce the heat sum needed for bud
burst (Cannell and Smith 1986). This pattern prevents very early bud burst when winters are warm
and reduces the possibility of freeze damage. The pattern also suggests that warming winters may
not be equally disruptive for all species. For example, in a study of chilling and forcing in tree
species of the Pacific Northwest, western hemlock required less forcing for bud burst than Douglas-
fir at low levels of chilling, but more forcing for bud burst than Douglas-fir at high levels of chilling
(Harrington and Gould 2015). The processes involved in the seasonal growth–dormancy cycle for
trees, including dormancy, chilling, and growth resumption (Figure 7.11), are also under strong
genetic control, as discussed in Chapter 3.
Seeds of many temperate forest species also require a chilling period, typically accompanied
by moist conditions, and thus overwinter on the forest floor before germinating the following
spring. This period is termed stratification, and it alters the hormonal balance between growth
Non-growing season
Dormancy
Chilling
maintained Dor
Gro man requirement
cy wt cy
an ed he r fulfilled
ot
pro orm
el bled
m
ea
na
III
se
d
II IV
Temperature
Grow
enab
enable y
c
Dorman
permissive
d
led
for growth
th
I Gap
Bud set Bud flush
Growth cessation
Critical daylength
for growth
cessation
Active growth
Growing growth
TRENDS in Plant Science
F I G U R E 7 . 1 1 Transitions in seasonal growth–dormancy cycling in trees. The growing season is the length
of time between bud flush and bud set, and the dormant season occurs opposite to the growing season.
Growth cessation is achieved in response to photoperiod in the fall, and growth resumption by a chilling
requirement in some species and the accumulation of a heat sum. Stage I approximates pre-dormancy; stage
II, full dormancy; stage III, true dormancy; stage IV, post-dormancy; stage V, active growth. Growth is
possible but does not occur during the “gap” between stages IV and V because of environmental constraints.
Source: Rohde and Bhalerao (2007) / with permission of Elsevier.
inhibitors and promoters to favor germination when conditions are ideal for growth. Seeds of other
species, including spring-fruiting trees (elms, cottonwoods, aspens, willows) and many fire-
dependent species (including most hard pines), lack a cold requirement and may germinate after
dispersal as soon as sufficient soil water is available.
NATURAL PLANT DISTRIBUTIONS AND COLD HARDINESS

Temperature and the ability of plants to resist freezing are critical factors limiting the natural dis-
tribution of plants. Winter cold hardiness appears to be related to the northern limit of species’
distributions (Kreyling et al. 2014), but the frost hardiness and hardening capabilities of adult indi-
viduals as discussed in the earlier text suggest that minimum winter temperatures may not be a
critical constraint. Instead, the timing of bud burst in relation to spring temperatures appears to be
the main control of the northern limit of tree distributions (Kollas et al. 2014; Vitra et al. 2017).
Freezing temperatures that occur in the spring just prior to and during bud burst are a main
Natural Plant Distributions and Cold Hardiness 163
selective agent, because this is the period when freezing resistance is weakest and plants are unable
to re-harden. Tree species are not limited by their relative abilities to withstand minimum winter
temperatures at the northern extent of their range. Rather, they are limited by their relative abilities
to withstand spring frosts that become more common further northward.
In a study of the frost hardening of four temperate deciduous genera, Vitra et al. (2017)
found that hornbeams and cherries, which had a lower forcing requirement and thus flushed ear-
lier in the spring, were able to withstand lower temperatures compared to beeches and oaks, which
had a higher forcing requirement and flushed later in the spring. The greater freeze resistance of
the early-flushing species thus seems to have developed from their vulnerability to late winter
frosts, because their bud burst coincides with a period when both warm and freezing temperatures
are possible. The relationship of the timing of bud burst to freeze resistance suggests that spring
rather than winter temperatures are pivotal in shaping northern distribution of temperate trees.
Notably, genera such as oaks that flush later do so at the cost of a reduced growing season, whereas
early-flushing species with a shorter duration of dormancy benefit from a longer growing season
at the cost of higher resource investment into frost resistance (Basler and Körner 2012; Muffler
et al. 2016).
An additional factor limiting the distribution of tree species is the effect of cold temperatures
on reproductive structures and flower buds, which tend to be more sensitive to frost than vegeta-
tive buds (Larcher and Bauer 1981). The failure to reproduce sexually can markedly affect migra-
tion and range extension, thus shaping species distribution. Perhaps more importantly, the ability
of species to regenerate and thus establish at the limits of their range depends on the survival of
seedlings in cold temperatures. Figure 7.12 illustrates frost resistance of several tissues of holm
oak, a relatively hardy, evergreen Mediterranean oak, whose seeds germinate in autumn and win-
ter when water availability is favorable (Larcher and Bauer 1981). Germinating seedlings and
surface-feeding roots are especially prone to freezing compared to other anatomical parts and
developmental stages. Adult trees are able to thrive where temperatures of −12 °C (10 °F) occur
each winter, but natural regeneration of the species is completely suppressed if winter tempera-
tures regularly drop below −8 °C (18 °F).
–12 –15
–9 –5
–9 –12 –13
–14 K –18 K –22
–15 –15 –20

–18 –22
–10
–20 –24
–20 –24
–6
–5 –4
–6 –3
–4 –5 –1
Damage below –8 –15 –20 –25 °C
F I G U R E 7 . 1 2 Freezing resistance of the various strata and age groups in a stand of holm oak (Quercus
ilex). Numbers are degrees centigrade. Source: Sakai and Larcher (1987) / Springer Nature.
Similarly, in North America, many species with southerly distributions can survive, reach
large size, and even produce fruit in climates far north of their natural range where their establish-
ment has been provided by humans. For example, sweetgum is used for landscaping in many
northern states far from the northernmost limits of its natural range. Northern catalpa has been
planted up to 650 km north of the northern extent of its natural range from northeastern Missouri
to southeastern Illinois. Osage-orange, whose northernmost natural range is in southern Oklaho-
ma and Arkansas (34° N), has been widely planted and successful as far north as southern Michi-
gan (42° 30′ N) or about 1000 km north of its northernmost natural range. At more northerly sites,
these species fail to regenerate and compete in natural communities.
DECIDUOUSNESS AND TEMPERATURE

Compared to maintaining thick evergreen leaves throughout the year, trees conserve energy by
shedding thin deciduous leaves in winter and protecting meristems from water loss (Breckle 2002).
This is considered an important mechanism to avoid freezing during the winter and to reduce the
loss of carbon via respiration when the potential to gain carbon through photosynthesis is low
(Givnish 2002). Losing the ability to photosynthesize without leaves (excluding those thin-barked
trees with stems that are able to photosynthesize) means that deciduous species must be able to accu-
mulate enough photosynthate during the growing season to support a full year’s maintenance,
growth, and reproduction. In part, deciduous trees compensate for shorter growing seasons by
producing thinner leaves with higher photosynthetic rates per unit mass compared to conifers,
whose long-lived leaves are thicker and have lower rates of photosynthesis (Reich et al. 1999;
Wright et al. 2004).
Deciduous woody species also address this trade-off by dominating forests in areas where
growing seasons are long, warm, and humid with readily available soil water and nutrients (approx-
imately from 35 to 45°N latitude and from 70 to 95°W longitude in North America). To the west,
deciduous forests are replaced by grasslands (Figures 2.6 and 2.8, Chapter 2) and dry-site conifers
(junipers, pinyon pines, and ponderosa pine) at low elevations where humidity is low, soils are dry,
and fires are frequent. On the southern Coastal Plain, growing seasons are long, summers are
warm, and winters are not extreme, but fire is frequent, and soils are nutrient-poor. These latter
conditions also favor conifers and other evergreen species over deciduous species.
The relationship between deciduousness and temperature is most obvious at higher latitudes
and elevations, where growing conditions become harsh and extreme, and conifers tend to replace
deciduous forests. With increasing latitude and elevation, the growing season shortens, summers
are cooler, soils are poor, and prolonged drying winds are common. Cold conditions inhibit
biological activity including decomposition, reducing nutrient availability compared with
comparable soils at lower latitudes and elevations. Moreover, cold soils result in unfavorable con-
ditions for root growth and for absorption and conduction of water and nutrients. Therefore,
deciduous and coniferous trees as well as forest shrubs and herbs that are better adapted to extreme
conditions outcompete those adapted to more favorable conditions. If we consider leaf longevity as
a proxy for the evergreen (high longevity) versus deciduous (low longevity) condition, then decid-
uousness (or lack thereof) is closely tied to the length of the growing season, which in turn is best
predicted by temperature (van Ommen Kloeke et al. 2012). Temperature and other site factors such
as soil water that significantly affect plant distribution are very strongly affected by physiographic
features, which are considered in the next chapter.
Natural Plant Distributions and Cold Hardiness 165
SUGGESTED
READINGS
Kozlowski, T.T., Kramer, P.J., and Pallardy, S.G. (1991). Levitt, J. (1980). Responses of Plants to Environmental
The Physiological Ecology of Woody Plants. New York: Stresses, vol. 1. New York: Academic Press Chapter 5
Academic Press 454 pp. (The Freezing Process), Chapter 6 (Freezing
Larcher, W. and Bauer, H. (1981). Ecological significance Injury), Chapter 7 (Freezing Resistance—types,
of resistance to low temperature. In: Physiological measurement, and changes).
Plant Ecology I, Responses to the Physical Environ- Sakai, A. and Larcher, W. (1987). Frost Survival of Plants.
ment, New Series, vol. 12A (ed. O.L. Lange, P.S. New York: Springer-Verlag 321 pp.
Nobel, C.B. Osmond and H. Ziegler), 403–437. Weiser, C.J. (1970). Cold resistance and injury in woody
New York: Springer-Verlag. plants. Science 169: 1269–1278.
Physiography CHAPTER 8
M any of our most vivid and unforgettable mental images are of the landscape where we grew
up or where we live. Moreover, the shape of the land and the silhouettes it creates provide
one with a sense of place, and sometimes a sense of purpose. The author Edward Abbey was fond
of relating the landscapes of the American Southwest to a sense of place and purpose, writing in
his book Desert Solitaire (Abbey 1968):
May your mountains rise into and above the clouds. May your rivers flow without end, meandering through
pastoral valleys tinkling with bells, past temples and castles and poets towers into a dark primeval forest
where tigers belch and monkeys howl, through miasmal and mysterious swamps and down into a desert of
red rock, blue mesas, domes and pinnacles and grottos of endless stone, and down again into a deep vast
ancient unknown chasm where bars of sunlight blaze on profiled cliffs, where deer walk across the white
sand beaches, where storms come and go as lightning clangs upon the high crags, where something strange
and more beautiful and more full of wonder than your deepest dreams waits for you – beyond that next
turning of the canyon walls.
Indeed, mountains inspire a sense of natural grandeur and power with their distinctive shapes and
striking seasonal changes. Prairie landscapes with their endless sweeps of flat or gently rolling
land have an irresistible and memorable quality whose beauty and harshness are portrayed in
many books and films. Lakes, rivers, and streams and their associated lake shores and stream
banks provide unforgettable mental maps to which we can link diverse biotic and human patterns
of occurrence and activities in space and time. Landscapes of rocky coastlines, sand hills, wetland
depressions, tropical lowlands and highlands, tidal flats, arid plateaus, and mountain slopes all
reveal close interrelationships between the form of the land or water with the climate above, the
parent material and soil below, and the organisms sandwiched between. At different spatial scales,
from subcontinental plains and mountains to pit-and-mound microsites in a hemlock–northern
hardwood forest, physiographic features or landforms provide the best path for understanding the
structure and dynamics of landscape ecosystems and the organisms they contain. We consider
foundations of ecological geomorphology in the following sections by examining the concepts and
ecological attributes of physiographic features and examples of their pervasive importance for the
forest ecologist.
CONCEPTS AND TERMS

The term physiography is an abbreviation for physical geography—the surface features of an
area. Surface features such as mountains or outwash plains have form at their surface that repre-
sents relief from the horizontal. These features also have parent material beneath their surface that
represents the underlying geological material from which soil develops (Bailey 2009). Thus, in
forest ecology, we define physiography as the form and substance of the surface features of a given
167
168 C h a p t e r 8 Physiography
regional or local area. Physiography differs from geomorphology, which is the science dealing
with the nature and origin of topographic features of the Earth, in that it integrates both the shape
of the surface as well as the geological surface materials (i.e., parent material) associated with the
features. The specific physiographic features themselves are landforms (e.g., mountain slopes,
plateaus, outwash plains, and river floodplains), created by erosion, sedimentation, or movement.
We use the term landform in the context of specific physiographic features of the Earth’s surface
and the term physiography in the broad context of a major ecosystem component. Physiography is
a major ecosystem component that not only provides spatial form and structure to a landscape and
its ecosystems but significantly affects ecosystem function. For instance, land in a high position
influences adjacent low-lying lands and their biota through its effect on climate and hydrology.
Land adjacent to a river affects its course, rate of flow, and water quality; and the river affects the
land as flood waters erode and deposit sediments, forming levees and floodplain bottoms.
We discussed in Chapter 2 that the major ecosystem components of climate, physiography,
soil, and vegetation may be used to distinguish and map landscape ecosystems at regional and local
levels, and also to understand ecosystem structure and function. Physiography is exceedingly
important in this regard because it is the component least affected by short-and long-term natural
and human disturbances, forming a relatively permanent framework of landscape ecosystems
(Rowe 1984b). Physiography influences ecosystem function because it controls the climatic regime
at and above the surface, and it controls soil development and acts to regulate soil processes below
the surface. Landforms and their parent materials modify the fluxes of radiation, soil water, and
nutrients thereby regulating plant establishment, distribution, growth, and productivity.
Let us consider a newly exposed glacial landform with different local climates on its slopes:
one slope relatively warm and dry, and the other relatively cool and moist (Figure 8.1). Different
plant species will initially colonize these different slopes based on their tolerances and require-
ments for temperature and soil water, and in turn animals will respond to the vegetation and
microclimate. Across a landscape, the size, shape, and relation of one landform to another shape
plant and animal communities, as well as genetic races that develop (Chapter 3). Soils develop
from the parent material of the new landform as it interacts with local climate, biota, and time.
Over time, similar ecosystems develop on similar landforms, and repetitive patterns of landforms
give rise to repetitive patterns of vegetation. Thus, a landform map is a proxy for a soils map, as well
as a map of local climate. In this regard, a landform map acts as a best approximation of a map of
regional and local vegetation (Rowe 1969). In addition, a landform map reflects the history and
patterns of disturbances such as wildfire and flooding, because the shape and parent material of
specific landforms may influence the severity and frequency of these disturbances.
CHARACTERISTICS OF PHYSIOGRAPHY AND THEIR SIGNIFICANCE

We list and briefly describe the specific characteristics of physiography that are useful in
understanding ecosystem structure and function in the hierarchical framework that follows. Many
physiographic characteristics are significant at many different spatial scales, from continents to a
local wetland depression less than 0.1 ha in area.
Physiographic Characteristics:
Physiographic setting (e.g., glaciated or unglaciated terrain, coastal or interior landscapes)
Specific landform (e.g., mountain, outwash plain, river terrace)
Elevation
Size
Form (shape or configuration)
Level landforms
Sloping landforms
Characteristics of Physiography and their Significance 169
Glacier
Newly Exposed Local Climates

Landform
Warm
Dry
Cool
Moist
Biotic Communities
Grasses
Sparrows Shrubs
Warblers T
I
M
E
Soils
Brown
Soil Dark Brown
Soil
F I G U R E 8 . 1 Landscape development since deglaciation. A hill of glacial till emerges from the melting ice
and develops its local climates. These and the parent-material substrate develop the ecological basis for
arriving plants that in turn provide habitat for adapted animals. Soils develop through interactions of
landform, local climate, and biota. Arrows indicate interactions and feedback so that, for example, the
properties of developing soils and communities influence the evolution (by erosion and deposition) of the
landform. Source: Rowe (1984b) / United States Department of Agriculture / Public Domain.
Slope shape or configuration (straight or planar, concave or convex)

Slope position (upper slope, mid slope, lower slope)
Slope aspect (direction a slope faces)
Slope inclination (slope percent or degree of slope)
Parent material of the feature or landform (kind of substrate, including rock type, particle
size, mineralogy)
Position of landform in the landscape (high-level, low-level)
Position of landform in relation to other landforms
PHYSIOGRAPHIC SETTING
Physiographic setting represents the broadest spatial scale of physiography that provides the
context for specific landforms. An overriding consideration is whether the landscape has been gla-
ciated by a continental ice sheet or not, because many landforms are specific to glaciated terrain.
The characteristics of features such as rivers, lake beds, mountains, and hills differ depending on
whether the feature occurs in glaciated or unglaciated terrain. Similarly, coastal regions represent
a very different physiographic setting than interior regions, and specific landforms that occur in
both regions are quite different in form and substance between the two settings.
SPECIFIC LANDFORMS
Specific kinds of landforms and their sizes determine to a large degree the physical factors of eco-
systems and the productivity and patterns of occurrence of the organisms with them. Specific land-
forms occur at spatial scales from continental to local (<1–100 ha). Mountain ranges, plains, and
large river basins are examples of specific landforms that occur at the spatial scale of continents.
Their properties and effects are appropriately considered in the classic treatments of physiography
by Fenneman (1931, 1938) and have since been discussed in detail in most geography texts. We
describe local landforms in the following sections.
ELEVATION
Elevation above sea level affects local climate, and for a given physiographic region or specific
landform, elevation is especially important when the climate associated with it is limiting to plant
establishment and growth. Two factors related to elevation that explain its effects on local climate
are the impacts of elevation on temperature and precipitation. First, air temperature decreases as
it moves upward through the Earth’s atmosphere at a rate of change known as the lapse rate, such
that temperatures experienced by plants are generally cooler at higher elevations than at lower
elevations. Moist air that is lifted through the atmosphere cools as it rises, losing its ability to hold
water and releasing the water as rain or snow called orographic precipitation. Much of this pre-
cipitation falls on the windward side of a mountain or mountain range, such that the leeward side
tends to be drier (often called a rain shadow). In general, sites at high elevations are often cooler
and wetter than sites at low elevations.
Elevation is therefore only a rough proxy for climate, because climatic factors also vary at
a given elevation with latitude and other physiographic elements (aspect, slope percent, etc.). At a
known latitude, however, elevation of a site may predict plant performance fairly well. For example,
the vegetation of riparian ecosystems differs at low elevations compared to high elevations whether
one is in the Cascade Mountains in Washington State or the Smoky Mountains in Tennessee.
It follows that the vegetation of low-elevation ecosystems at high latitudes may resemble the vege-
tation of high-elevation ecosystems at lower latitudes (see Chapter 7). For example, communities
dominated by red spruce at sea level in eastern Canada have many similar species to those found
at elevations over 1500 m in the southern Appalachians.
FORM OF LANDFORMS
As we discussed in the earlier text, the two key elements of physiography are the shape and parent
material of specific landforms. The shape of landforms, in particular, influences how radiation,
soil water, and nutrients are received and distributed and in turn the vegetation patterns of an area.
Level Terrain Whether land is level or sloping is one of the initial distinctions made
r egarding landscapes. Landforms with characteristically level or nearly level terrain include
plains, river terraces, old lake beds, prairies, and plateaus, all of which are either deposited or
influenced by water. Some exceptions to this association exist, such as the exceedingly flat till
plains of parts of Indiana and Ohio that were deposited by glacial ice as it rapidly retreated at a
relatively uniform rate. The Tipton Till Plain in central Indiana is so flat and monotonous that
one observer noted: “. . . the traveler may ride upon the railroad train for hours without seeing
a greater elevation than a haystack or pile of sawdust” (Petty and Jackson 1966).
The lateral movement of water on level landforms is limited, and thus precipitation may
more quickly move vertically through the soil depending on the nature of the parent material.
Level terrain with coarse-textured parent materials has high vertical water movement and tends to
be dry. In warm, dry seasons, water drainage is rapid and water nearest the soil surface is quickly
absorbed by plant roots or is evaporated. If the soil has layers or bands of less-permeable material
or bedrock at or near the surface, vertical water movement will be slowed or restricted. Even in
very dry soils, such bands may subtly or markedly affect plant growth and community composition
(Kashian et al. 2003). Level terrain is typically the location of ecosystems where water accumu-
lates, such as lakes, marshes, swamps, and bogs. Most of the world’s wetlands are found on
relatively flat terrain, including coastal plains, old lake beds and areas surrounding lakes, and areas
associated with river floodplains where groundwater is found at the surface.
Flat terrain also features strong effects of temperature on microclimate. Level terrain limits
cold-air drainage, such that small differences in the terrain in the form of small depressions or flat
land surrounded by even slightly higher land create natural sinks for accumulation of cold air at
night (Chapter 7). Low nighttime temperatures in such frost pockets further affect the composition
and growth of vegetation and animal populations.
Sloping Terrain Sloping terrain is ecologically different from flat terrain because water may
move laterally, either internally or overland along a slope so that poorly drained soils may seldom
occur. Sloping terrain generally features three shapes: straight or planar (surface with zero
curvature), convex (curving outward, such as a hill), or concave (curving inward, such as a depres-
sion). These types are combined in a variety of ways in different landforms. On convex surfaces,
water, organic matter, and dissolved nutrients tend to move out of the feature. Concave surfaces
tend to accumulate soil water, nutrients, and organic matter, often making them more favorable for
plant growth. Plants lower on convex surfaces often attain larger sizes than individuals of the same
species upslope (Chapter 11).
Other important characteristics of sloping terrain are aspect, or the direction a slope faces,
and slope inclination, or the steepness of the land from the horizontal. Slope length (in m or
km), for example, such as a short, steep slope versus a long, gradual slope, may affect plant distri-
bution in certain landscapes because a longer slope evokes steady rather than abrupt changes in
environmental conditions. In a classic study of deciduous swamps in northern Ohio, Sampson
(1930) reported that an abrupt elevation change of 18–25 cm above the wet swamp forest of
American elm, black ash, and silver and red maple gave sufficient local aeration for American
beech and sugar maple or even oak and hickory trees to occur. By contrast, areas with long, gentle
slopes rising 3 or 4 ft (about 1 m) over a mile (6.4 km), although their elevation may be several feet
higher than the abrupt rises, were still too wet to support American beech or sugar maple.
Glaciated landscapes often feature high and moderately-to-steeply sloping end and lateral
moraines, as well as flat, gently sloping, or rolling ground moraines. Ice-contact terrain exhibits
steep, hilly topography of kames and eskers with relatively abrupt and short slopes interspersed
with ice-block depressions of variable depths termed kettles (often containing lakes and wetlands).
Ocean and lake beach ridges, terraces, and dunes exhibit locally distributed low, rolling features.
Similar to unglaciated terrain, the recurrence of landforms and their parent materials on the
landscape provides predictable patterns of ecosystem types and plant communities.
Slope Characteristics Slope position, slope aspect, and slope inclination influence micro-
climate, soil depth, soil profile development, and the texture and structure of the surface soil.
Because these microclimate and soil factors in turn influence vegetation, slope characteristics have
a heavy bearing on the composition, development, and productivity of the ecosystem.
Ridge top
Upper slope
Mid slope
Lower slope
Cracks Wedge-Shaped Aggregates
Cobbles Subangular Blocky Structure
Prismatic Structure Basalt
F I G U R E 8 . 2 Cross-sectional diagram of changes in soil depth and morphology with slope position for a
vernal pool in Riverside County, California. The diagram depicts only a 3-m rise in elevation over a 30-m
horizontal distance. Source: Hawkins and Graham (2017).
Position on slope The tops of high features of the landscape, such as ridge tops or upper convex
slope surfaces, tend to be drier than is the average for the region. These features are differentially
exposed to intense solar radiation and high winds and are subject to higher rates of soil-water
drainage and erosion. By contrast, lower slopes with concave surfaces tend to be moister than
average for the region, are sheltered from high winds, accumulate water, organic matter, and soil,
and are subject to cold-air drainage. Midslopes are generally intermediate in their characteristics
(Figure 8.2). Soil development differs in response to these differences in moisture and temperature.
Soil differences are most pronounced when a slope represents a large change in elevation, but even
small elevational changes are subject to these processes (Hawkins and Graham 2017).
Slope position, used as a proxy for relative elevation (i.e., ridge top; upper, mid, lower slope),
has a strong influence on site productivity (Chapter 11) and is one of the most useful criteria for
classifying and mapping forest ecosystems. Many researchers have shown slope position to be the
single most useful factor in evaluation of the growth potential of forest trees (Ralston 1964; Car-
mean 1975, 1977). Slope position is also convenient in that it is able to be remotely sensed, histor-
ically using aerial photography and more recently with digital elevation models (DEMs) and
LiDAR (Light Detection and Ranging) technology.
Aspect The aspect of a slope, that is, its orientation with regard to the sun’s position, has a direct
influence on microclimate. In northern temperate regions, the sun moves across the southern
horizon and is therefore to the south during the warmest part of the day. As a result, south-facing
slopes receive the highest intensity of sunlight and thus its sites are generally hot and dry. In a
detailed study of eight environmental factors of four southerly aspects in the desert foothills of
Arizona, Haase (1970) found the sequence of warmest and driest aspects to be south,
south-southwest, southwest, and south-southeast. On combining aspect with slope inclination
(see the following text), the hottest and driest sites are those with slopes perpendicular to the sun’s
angle during the middle of the summer day, such that both steeper and more gentle slopes receive
less radiation. The amount of sunlight received on a site governs air and soil temperature, precipi-
tation, and soil water, all of which affect plant establishment and growth.
North-facing slopes receive less sunlight and are invariably cooler and moister in the
Northern Hemisphere. East and west slopes show similar but less extreme variation. East-facing
slopes are exposed to direct sunlight in the cool of the morning and are somewhat cooler and
moister than west-facing slopes, which are exposed to the sun for an equal amount of time as
east-facing slopes but receive the radiation in the hotter part of the day. As described in the earlier
text, these differences in microclimate have important implications for site productivity. In mixed
upland oak forests of the Appalachian Mountains, northeast aspects are the most productive,
being approximately 15% more productive than south and west aspects, which are the least pro-
ductive (Chapter 11). In a study of succession in subalpine landscapes of Colorado, succession
proceeded more quickly on north-facing slopes than south-facing slopes, primarily driven by
differences in soil moisture and air temperature (Donnegan and Rebertus 1999).
Slope inclination Slope gradient is measured in terms of percentage or in degrees, and steeper
slopes provide greater surface per unit area measured horizontally. For this reason, good forest
sites of moderate slope often contain more trees and produce greater timber yields per hectare
measured horizontally than do comparable level sites. Movement of water, snow, and soil is more
rapid on steep slopes compared to gentle slopes, and the danger of erosion, avalanche, and mass
soil movement is much greater. Notably, uprooting is one of the most common effects of steep
slopes on trees, and this process itself contributes to mass movement of soil (Norman et al. 1995).
These processes favor species dependent on disturbances that expose mineral soil and cause can-
opy openings. For example, a thick forest canopy is rarely formed in southern Wisconsin on steep
hillsides, so that light penetration occurs from the side and species such as black and white oak
persist. Succession away from black and white oak is more rapid on level terrain where light avail-
able to plants of the forest floor is controlled by the tree canopy rather than disturbance events
more common on steep slopes (Curtis 1959). On rocky landscapes, rockfalls on steep slopes may
damage trees at a frequency that inhibits tree growth (Moos et al. 2021).
PARENT MATERIAL IN RELATION TO LANDFORM

The substance component of landform is important because it allows us to approximate broad classes
of soils from what we understand about parent material. Parent materials are typically specific to
landforms regardless of whether the materials developed in place or were deposited in association
with glaciation (Chapter 9 and Figure 9.1). Soils that develop from residual limestone or sandstone
rock support markedly different plant communities. Parent material develops by weathering of
specific rock types, stream action, and geomorphic processes such as avalanche and debris flow.
Wind initiates landforms, most commonly sand dunes (from sand parent material) and loess deposits
(from silt), which reflect the particle size of the parent material. The landforms of river floodplains,
including the front, levee, first bottom, and backswamp, reflect the nature of their parent material
and flooding patterns. On glaciated landscapes, flat (old lake beds, glacial outwash plains) and hilly
(moraines, ice-contact terrain) landforms have characteristic parent materials that distinguish them
and, together with the shape of the feature, help determine forest composition, productivity, and suc-
cessional pathways. Despite a perceived lack of major topographic differentiation on glaciated land-
scapes, landforms constrain the distribution of parent material in such a way that they may be the
most important driver of ecosystem structure. For example, forest composition resulting from natural
reforestation of a central Massachusetts landscape following clearing for agriculture 150 years ago
was found to be more influenced by landform than agricultural history or natural disturbances
including fire, hurricanes, or logging (Gerhardt and Foster 2002). The mineralogy and chemistry of
parent material are discussed in more detail in Chapter 9.
POSITION OF LANDFORM IN THE LANDSCAPE

The position of a specific landform (e.g., outwash plain, ravine, ridge, etc.) on the landscape may
influence plants and animals differently. Broadly, landforms in the rain shadow of a mountain
range exhibit different climates, soils, and vegetation than similar landforms on the windward side
where precipitation is markedly higher for a given elevation. For example, some landforms on the
western (windward) side of the Cascade Range in Washington State may have up to 4 m of annual
precipitation compared to less than 50 cm for similar landforms on the eastern side. A high,
exposed ridge may support different plants and site potential compared to one with similar geology
and parent material in a lower and protected position. In non-mountainous glacial terrain, an
elevated, dry outwash terrace is more favorable for tree growth than an equally dry terrace below
it in a cold, frost-prone basin. Thus, the position of a given landform on the landscape may provide
different site conditions for plant establishment and growth.
A related phenomenon to landscape position is landform adjacency. The ecological effect of
a forested wetland is much larger on land adjacent to it than land at a greater distance. At a broad
scale, landforms adjacent to large water bodies (oceans or lakes) are more strongly affected than
lands further inland or at further distance from a lake. At a local scale, a first-bottom ecosystem
adjacent to a river is sharply different in soil and vegetation than an elevated, second-bottom eco-
system further from the river due to its greater likelihood and frequency of flooding.
In summary, many different physiographic characteristics act together to influence vegeta-
tion and structure and function of landscape ecosystems and have been used as the stable basis for
the management of forested ecosystems. For example, Angus Hills (1952) pioneered the use of
physiography on glacial landscapes (integrating slope position, aspect, slope gradient, and parent
material) to determine forest composition and productivity in Ontario, Canada (Chapter 11). Hills
demonstrated how topographic-mediated differences in microclimate (warm versus cold) and soil
water can bring about marked differences in late-successional communities (Figure 8.3).
MULTIPLE ROLES OF PHYSIOGRAPHY

Physiography results from geologic and geomorphic processes, and its characteristics already
described influence ecosystem structure and function. This interrelated suite of characters influ-
ences ecosystem processes at multiple spatial scales and influences soil development, climate, dis-
turbance regimes, and the distribution of vegetation. The roles of landforms in affecting ecosystem
processes have been summarized into four classes, all of which are interrelated (Swanson
et al. 1988):
1. Quantity of materials available (Figure 8.4a): As described in the earlier text, the kind of
landform and its elevation, slope aspect, inclination, and position influence the quantity of
solar radiation, air, water, nutrients, and other materials received at the site. In addition to
the effects described in the earlier text, these factors may be important determinants of
other drivers of ecosystem structure and function such as soil moisture and disturbance
(Méndez-Toribio et al. 2016).
2. Flux and transport of materials (Figure 8.4b): The same factors mediate the flow of air,
water, materials (including organic and inorganic particulate matter, dissolved materials),
organisms, propagules, and energy within and across ecosystem boundaries by: (i) affecting
gravitational gradients, (ii) guiding flows of wind and water, and (iii) forming barriers to
movement. River floodplains and corridors are perhaps the best example of landforms that
broadly direct the flow of many materials.
3. Disturbance (Figure 8.4c): Landforms and their physiographic features influence the
spatial pattern, frequency, and severity of disturbances caused by fire, wind, water,
Multiple Roles of Physiography 175
Late-successional
Microclimate and soil plant community
Normal microclimate Maple–beech

over moist soil
Normal microclimate Oak–ash

over wet soil
Normal microclimate Oak–history

over dry soil
Warmer microclimate Tulip–walnut

over moist soil
Warmer microclimate Sycamore–tulip

over wet soil
Warmer microclimate Oak–chestnut

over dry soil
Colder microclimate Elm–ash–oak

over moist soil
Colder microclimate White spruce–

over wet soil balsam fir
Colder microclimate Hemlock–

over dry soil yellow birch
F I G U R E 8 . 3 Physiographic diagram illustrating the effect of local topography (slope position and aspect) in
determining late-successional species composition of landscape ecosystems in the temperate continental
zone of southern Ontario, Canada. Source: Modified from Bailey (1988) and Hills (1952).
avalanche, and soil movement, as well as human-caused disturbances of erosion, pollutants,

and barriers to natural fluxes of water and nutrients. Disturbances by wind are particularly
affected by landforms, with wind speeds highest on exposed ridge tops if it moves perpen-
dicular to a ridge, at mid-slope if the wind direction is at an acute angle, and at the valley
bottom if the wind moves parallel to the ridge (Frelich 2002). Hurricanes tend to be most
damaging to forests on aspects that face the wind direction (Boose et al. 1994).
4. Spatial position of landforms (Figure 8.4d): The juxtaposition of one landform in relation
to another (landform adjacency) may affect regional and local climate and hydrology,
movement of materials, and disturbance regimes. The arrangement of landforms across a
landscape may be important in determining patterns of major disturbances such as large
fires, windthrow areas, or landslide events, because all or portions of some landforms are
more or less susceptible to a given disturbance.
(a) (b)
(c) (d)
F I G U R E 8 . 4 Summary of the roles of physiography as represented by four classes of landform effects:

(a) influences on materials received (e.g., effects of aspect, slope position, and slope inclination on solar
radiation and precipitation); (b) influences on fluxes and transport (e.g., effects of landscape position on
water input to lakes); (c) influences on disturbances (e.g., effects of aspect and slope position on wind
events); (d) spatial position of landforms (e.g., effects of adjacent landforms on the susceptibility to land-
slides). Source: Swanson et al. (1988) / with permission of Oxford University Press.
This simple four-class summary suggests that physiography directly or indirectly affects all
aspects of plant life: regeneration, geographic distribution and abundance, growth and produc-
tivity, and death. In the following sections, we examine the influence of physiography at multiple
spatial scales on landscape ecosystems and their biota, from the subcontinental to the microtopo-
graphic level.
PHYSIOGRAPHIC DIVERSITY, LANDSCAPE ECOSYSTEMS, AND VEGETATION

An interconnectedness of physiography and vegetation is clear throughout the world. Distinctive
patterns of ecosystems and their vegetation are strongly related to geomorphology and geomorphic
processes, and vegetation in turn influences physiographic processes.
MOUNTAINOUS PHYSIOGRAPHY
The influence of physiography on ecosystems is perhaps most obvious in mountainous terrain.
Several examples serve to illustrate the intricacy and critical importance of physiography and
mountain landforms.
Physiographic Diversity, Landscape Ecosystems, and Vegetation 177
10 20 30 40 50 60 70 80 90 100
White Dark Pine

13,000 Distance, miles MOUNT DANA
Western White Pine

Mountain Hemlock
Mountain Hemlock
Western White Pine
12,000
Lodgepole Pine
11,000
Giant Sequoia
Red Fir
10,000
Incense Cedar
9,000
Transition Brush
Altitude, feet
8,000
White Fir
Sugar Pine
Lyall Fork
Western Yellow Pine

Juniper
7,000 Leevining Pinon
Incense Cedar
Creek Sage
6,000 MONO
LAKE
5,000 Vernal Falls

Digger Pine
Live Oak
4,000 Sub-alpine
Chaparral
Fir
50´´ Jeffrey Pine 50
3,000
Rainfall, inches
Woodland
40
Blue Oak
Oak South Fork 35´´

2,000 Woodland 35´´
Brush
30´´ 30
Grass 25´´
20´´ 20
1,000 San Joaquin River Chaparral Yellow Mixed Conifer
Marginal Willows and Woodland Pine and Sugar Pine Fir Sub-alpine Alpine 10
10´´ Grass
F I G U R E 8 . 5 Physiographic diagram of the Sierra Nevada Mountains illustrating the distribution of

dominant plants in relation to elevation and precipitation. Variations in topography along this generalized
gradient are shown by the dotted lines. They define large stream valleys and ridges in this rugged terrain.
The amount of precipitation along the transect is represented by vertical bars along the bottom of the
diagram. Source: Harlow et al. (1996) / McGraw-Hill.
Mountainous Terrain of California and the Pacific Northwest A figurative transect

across the eastern half of the San Joaquin Valley and over the Sierra Nevada Mountains in
California illustrates how the physiographic position of valley and mountain features influence
precipitation that, together with temperature, determines the distribution of vegetation from near
sea level to over 3650 m (12 000 ft) (Figure 8.5). Conifer-dominated ecosystems appear on western
slopes above about 610 m (2000 ft), just above the dry valley and oak woodland. Further inland and
higher in elevation is a red fir-dominated forest, with annual precipitation at 125 cm compared to
50 cm in the oak woodland. Conifers maintain their dominance to nearly 3350 m (11 000 ft), as pre-
cipitation declines and where low temperatures and thin soils prevail, and the growing season is
very short. Precipitation drops sharply on the opposite (eastern) slopes of the Sierra Nevada range
within a significant rain shadow, where only the most drought-adapted species can survive.
The Pacific Northwest states of Washington and Oregon provide an excellent example of
physiographic diversity, where highly varied landscapes result in an enormous diversity of regional
and local ecosystems (Franklin and Dyrness 1980). Broad vegetation types are associated closely
with similarly broad physiographic provinces, which in effect describe regional landscape
ecosystems (Figure 8.6). Physiographic provinces delineate major landscapes characterized by
differences in climate and geology, with substrates contrasting in physical and chemical properties.
Variations in precipitation, temperature, and vegetation are associated with gradients in elevation,
latitude, aspect, and distance from the ocean (Swanson et al. 1990). Coastal marine terraces and
broad inland valleys (Puget Sound area and Willamette Valley) exhibit the lowest relief and support
unique vegetation types (Figure 8.6b). At the other extreme are the mountains of the northern
Cascade Range and Olympic Mountains of Washington (Figure 8.6a) where relief from ridge crest
to adjacent valley floor may be 1000–1400 m, again with the vegetation reflecting these major
differences in physiography and climate (Figure 8.6b). The arrangement of landforms and land-
form adjacency are important in provinces such as the Willamette Valley between the Coast and
(a)
NORTHERN OKANAGAN HIGHLANDS

CASCADES
OLYMPIC
PENINSULA
PUGET
TROUGH
COLUMBIA BASIN
SOUTHERN
WASHINGTON
CASCADES
ES
RANG
EY
ALL
COAST
EV
ETT
BLUE MOUNTAINS
LAM
WIL
CAS TERN
ES
CAD
WES
HIGH LAVA PLAINS

OWYHEE
UPLAND
KLAMATH BASIN AND RANGE

MOUNTAINS
F I G U R E 8 . 6 Comparison of physiographic provinces and generalized vegetation areas of Oregon and

Washington. (a) Physiographic–geologic provinces. (b) Major vegetation areas. Source: Franklin and Dyrness
(1980) / Oregon State University Press.
(b)
124° 123° 122° 121° 120° 119° 118° 117°
48° 48°
PSA
47° 47°
46° 46°
45° 45°
WV
44° 44°
PR
43° 43°
42°
42°
124° 123° 122° 121° 120° 119° 118° 117°
Legend
INTERIOR VALLEYS OF
FORESTED REGIONS WESTERN OREGON
Picea sitchensis Zone WV Willamette valley
Tsuga heterophylla Zone Umpqua and Rogue valleys
PSA Puget Sound area STEPPE REGIONS
Mixed Conifer and Mixed Evergreen Zones STEPPE (without Artemisia tridentata)
Pinus ponderosa Zone (board sense) SHRUB-STEPPE (with Artemisia
PR Pumice region tridentata)
Abies grandis and Pseudotsuga DESERT SHRUB
menziesii Zones
Juniperus occidentails Zone
Subalpine forests (including Abies amabilis,
TIMBERLINE AND
A. lasiocarpa, A. magnifica shastensis,
APLINE REGIONS
and Tsuga mertensiana Zones)
FIGURE 8.6 (Continued)

F I G U R E 8.7 Interfingering of
vegetation zones in mountainous
topography. Streams are shown running
downslope in valleys. A given vegetation
ZONE C zone is distributed higher in elevation on
Contour line
ridges than in valleys. Source: Franklin
and Dyrness (1980) / Oregon State
ZONE B University Press.
ZONE A
Cascade ranges or the Columbia Basin in the rain shadow of the Cascade Mountains. In addition,
the parent material of Pacific Northwest landforms is important in determining the great diversity
of soils (Franklin and Dyrness 1980).
The broad zones of natural vegetation correspond closely to broad physiographic provinces
in the Pacific Northwest, but the effect of landform on vegetation is also obvious at local scales.
Vegetation zones may occur as sequential belts along an elevational gradient on mountain slopes,
but more often their lower elevation limit occurs lower in valleys than on adjacent slopes
(Figure 8.7). This “interfingering” occurs because the site conditions in valleys differ in tempera-
ture and soil water compared to adjacent slopes of similar elevation. Vegetation zones are relatively
broad so that a given species, Douglas-fir for example, may occur in moist sites in the lower pon-
derosa pine zone and on relatively dry ridges higher up in the montane grand fir zone.
Franklin and Dyrness (1980) showed that forest vegetation zones in the Pacific Northwest
may be differentiated using precipitation (drought stress) and temperature (temperature growth
index), both of which are strongly influenced by physiography (Figure 8.8). Oak–juniper wood-
lands and ponderosa pine zones form a group of forests limited by moisture that occurs in the
rain shadows of eastern Washington and Oregon and in southwestern Oregon. The mixed conifer
and coastal temperate groups form extensive temperate forests. Within this group are the unique
temperate conifer forests of the world (zones: 1, Tsuga heterophylla (western hemlock); 11, Abies
grandis (grand fir); 12, mixed evergreen; including Sitka spruce, Douglas-fir, and others) (Frank-
lin et al. 1981). These unique forests grow under highly favorable temperature and moisture
conditions reaching extremely large size and old age. Finally, subalpine and boreal groups are
severely constrained by temperatures, reflected in short growing seasons and harsh mountain
environments.
Physiography and Forests of the Central Appalachians Forest species and commu-
nities are closely associated with local geologic and topographic conditions throughout the differ-
ent sections of the Appalachian Mountains (southern, central, northern, Blue Ridge, Ridge and
Valley, etc.; Hack and Goodlett 1960; Leak 1982). Hack and Goodlett (1960) studied a 142 km2
(55 mi2) densely forested area in the drainage of the Little River of northern Virginia, located in the
Ridge and Valley section of the central Appalachians. Local distribution of species and forest com-
munities between about 600 and 1300 m in elevation coincides with well-defined differences in
landform and disposition of water. Dry pine forest, dry–mesic oak forest, and mesic “northern
hardwood” forests are found in characteristic positions in first-order valleys (Figure 8.9). Pine and
pine–oak forests, dominated by pitch pine, table mountain pine, and several oak species, are found
on noses, ridges, and other slopes that are convex away from the mountain such as those shown in
MOISTURE MAJOR ENVIRONMENTAL CONSTRAINT

120 TEMPERATURE MAJOR ENVIRONMENTAL CONSTRAINT
NEITHER MOISTURE NOR TEMPERATURE

MAJOR ENVIRONMENTAL CONSTRAINTS
5
100
3 OAK–
JUNIPER
(NUMBER OF OPTIMUM GROWTH DAYS)
POND- WOOD-
TEMPERATURE GROWTH INDEX
12 EROSA LANDS
80 9 PINE
1
MIXED
CONIFER 6
60 10 4
COASTAL
TEMPERATE
1 -Tsuga heterophylla
2 11 2 - Abies amabilis
40 3 - SW Oregon Pinus ponderosa
4 - Interior Pinus ponderosa
7 5 - Quercus woodland
8 6 - Juniperus occidentalis
SUBALPINE 7 - Tsuga mertensiana
20 AND 8 - Abies lasiocarpa
BOREAL 9 - SW Oregon mixed conifer
10 - Pseudotsuga menziesii
11 - Abies grandis
12 - Mixed evergreen
0
0 5 10 15 20 25 30 35
DROUGHT STRESS INDEX (ATMOSPHERES)
F I G U R E 8 . 8 Distribution of some major forest zones of Oregon and Washington within an environmental
field based on moisture (maximum plant moisture stress during the dry season) and temperature (optimum
growth days). Source: Franklin and Dyrness (1980) / Oregon State University Press.
Figure 8.9a. Dry–mesic oak communities—variable in composition but dominated by oaks—are

found on the straight slopes between nose and valley or in drier hollows. Mesic northern hardwood
communities of sugar maple, basswood, yellow and black birches, and northern red oak are
restricted to moist hollows and other concave slopes that concentrate water. These mesic commu-
nities occupy the floodplains of larger valleys, extend up the floors of smaller valleys, and in the
first-order valleys extend into the hollows at the valley heads. In Valley 3 (Figure 8.9b), dry pine
forests are largely found near the top of the valley on its west-facing slope, where temperatures are
higher during the growing season and soil moisture is reduced. Mesic northern hardwood forest,
by contrast, is found on the east-facing slope and in the hollow of the valley. Trees of the northern
hardwood type in the hollow of north-facing Valley 3 are the tallest, largest, and have the greatest
volume compared with those of the other two forest types.
The association between the local distribution of species and communities and their position
in a valley is strongly determined by physiographic characteristics. Important characteristics
include: (i) valley size, because a large area is essential for the northern hardwood community to
persist in a hollow; (ii) orientation of the valleys and side slopes, shown by north and east slopes
being commonly forested with northern hardwoods, whereas south and west slopes are typically
(a)
t of ridge
N Approx cres 3000´
Nose - Contours
Convex Outward
Side Slope -
Contours Straight
Hollow - Contours
2900´ Concave Outward
Road
2800´ Channelway - Contours
Sharply Concave Outward
2600´ 2700´
Footslope -
Transitional Area Between
Side Slope and Channelway
200 0 200 400 Feet
0 61 122 Meters
(b)
COVER TYPE:
00
30
00
Pitch pine forest type

31
00
32
00
33
00
34
Northern hardwood forest type
3500
Oak forest type
3600
3700
3800
N 3900
500 0 500 1000 1500 FEET

CONTOUR INTERVAL 100 FEET
DATUM IS MEAN SEA LEVEL
F I G U R E 8 . 9 Contour maps of typical first-order valleys, Little River Basin, central Appalachian
ountains, northwestern Virginia. (a) Map of landforms of upper part of Valley 1, west side of Crawford
M
Mountain. Legend gives landform types of first-order valleys; vegetation distribution is not shown. (b) Map
of forest type distribution for Valley 3, northeast of Reddish Knob. Source: Hack and Goodlett (1960) / United
States Government Printing Office/ Public Domian.
oak; and (iii) nature of the bedrock, because formations favoring soil-water accumulation and
retention are more favorable for the more mesophytic species. Northern hardwoods occupy the
hollow of north-facing Valley 3 (Figure 8.9b), but oak forest is found in the hollow as well as
the side slopes in Valley 1 because it faces southwest (Figure 8.9a), leading to drier soils. Trees of
the oak type in the hollow of Valley 1 are larger and taller than those of the same species on the
side slope due to better soil-water and nutrient conditions, and ground-cover species differ between
hollow and side slopes. These patterns suggest three different landscape ecosystem types (nose;
side slope; and hollow, footslope, and channel) and two forest cover types (oak forest and northern
hardwood forest) in Valley 1 (see Chapter 1).
Detailed studies of vegetation in five additional first-order valleys showed that physiography
as controlled by geomorphic processes has predictable effects on vegetation, although species com-
position varied among valleys and among hollows (Hack and Goodlett 1960). Three valleys sup-
ported northern hardwood forest where topographic features provided a relatively moist
environment. The three valleys faced favorable aspects (either northeast or northwest), were
relatively large and deeply cut into the mountain, and had parent material favoring water reten-
tion. The topographic features of these valleys are controlled by geomorphic processes, specifically
debris avalanches and floods, caused by runoff from cloudbursts, which produce slopes that are
concave-upward, thus favoring mesic northern hardwood species. By contrast, the two valleys sup-
porting oak forest were physiographically similar to Valley 1 (Figure 8.9a), providing a drier envi-
ronment. The topography of the drier valleys resulted from soil creep, which produces
convex-upward slopes that typically support drier species. Overall, the present landscape is the
product of a long period of interaction between vegetative and geomorphic processes.
The central and southern Appalachians are some of the regions where the relationship of
forest composition and physiography is best studied. Technological advances over the last two
decades have allowed ecologists to quantitatively integrate the multiple components of physiog-
raphy to interpret their ecological effects on forest composition and productivity. For example,
Iverson et al. (1997) used DEMs and soil series maps to predict forest composition and productivity
(site index; Chapter 11) in southeastern Ohio. Landscape features derived from DEMs and
soil series maps included slope steepness and aspect, curvature of the landscape, soil-water-holding
capacity, and downslope water flow. These factors were used to build a single integrated mois-
ture index (Figure 8.10) that could be easily related to various ecological processes across the
landscape. The resulting index was used to successfully predict, for example, that oak was most
productive in the valley bottoms and least productive on upper south-facing slopes and ridgetops.
This index derived from physiographic data has also been shown to be closely correlated to under-
story vegetation, species richness, litter depth, soil pH, and bird species distribution (Iverson
et al. 1997). Similarly, Bolstad et al. (1998) predicted forest composition using elevation and terrain
shape (cove, ridge, or side slope) extracted from DEMs for the southern Appalachians. Species
characteristic of cove ecosystems were positively related to concave locations on the landscape,
while dry oak and pine species were negatively related. Northern hardwood species were more
closely related to elevation than to terrain shape. Methods such as these demonstrate the value of
incorporating physiography into our understanding of forest vegetation composition and distribu-
tion and have since been embraced by many ecological modelers and geographers.
FLATLANDS
Flat and gently sloping terrain is clearly less topographically diverse compared to mountainous
terrain. Nevertheless, subtle differences in topography and variable parent material result in dis-
tinctive effects on the distribution and composition of vegetation. Such subtle variations in physi-
ography may affect the spread of fire, the rate of nutrient cycling, and forest composition, among
other processes.
(a) (b) (c)
Slope steepness/aspect Downslope flow Landscape curvature
(d) (e)
Soil water holding capacity
Score
0-10 51-60
11-20 61-70
21-30 71-80
31-40 81-90 Integrated Moisture Index
41-50 91-100
F I G U R E 8 . 1 0 Maps depicting standardized scores (0–100) for (a) slope steepness and aspect, (b) downslope
water flow, (c) landscape curvature, and (d) soil-water-holding capacity for the Vinton Furnace Experimental
Forest in southeastern Ohio. These four physiographic factors were combined to create (e) an integrated soil
moisture index for the landscape. Source: Maps (a)–(d) after Iverson et al. (1997); map (e) courtesy of
Louis Iverson.
The Great Plains Grasslands dominated the flat or rolling plains and gentle slopes of the
North American Great Plains at the time of European colonization (Wells 1970), and forests are
not often considered as having broad ecological importance in the region. It is the relative lack of
forests across this landscape, however, that is instructive about the dominant process driving the
structure and function of the ecosystems of the Great Plains: fire. Vast, frequent, wind-driven
prairie fires would readily destroy tree seedlings and saplings, particularly those of conifers that
lack the ability to sprout, from areas where fires were too frequent. Forests of the plains therefore
existed where they were sheltered from fires, such as along rivers, in rough, dissected, boulder-
strewn escarpments of the uplands, and on relatively level sites on the leeward sides of lakes and
rivers. As is true with more complex distributions of species on forested landscapes, physiography
facilitates this treeless landscape through its influence on climate and fire (Wells 1970). Across the
entire width of the Central Plains, rougher and more dissected topography favors the occurrence
and spread of forest at the expense of grassland (Wells 1965).
Regional physiography drives the dry climate of the Great Plains (continental climate and
their location east of a major mountain range), but other physiographic features also make fire
common in this region. The landscape is vast and unbroken in its flat or rolling smoothness,
and these features, as well as dry soils, support flammable, grassy vegetation. It is easy to ima-
gine that grass fires swept by wind spreading across a flat or rolling plain, whether ignited by
lightning or humans, would continue indefinitely until quenched by rain or stopped by an

abrupt break in topography. Trees therefore persisted in escarpments with sparse grasses, per-
haps temporarily encroaching on grassland during short fire-free periods. The mosaic of eco-
systems of the Great Plains is therefore due to the regional flatness and continuity of the
physiography as much as it is to climate. The interrelationships of climate, physiography, and
the production of fuels from grassy vegetation create the “perfect storm” for frequent fire on
this landscape.
Pine Savannas of the Western Great Lakes Region Fire is also favored by the flat and
gently rolling physiography of the western Great Lakes region and adjacent Canada. In combination
with hot spring–summer periods and droughty, sand soils, physiography favors frequent fires
(though less frequent than in the Great Plains) and perpetuates fire-dependent species and com-
munities. In parts of Minnesota, northern Wisconsin and Michigan, and southern Ontario, flat
outwash or rolling plains, dry sand soils, and flammable vegetation led to frequent fires that main-
tained communities of nearly pure jack pine or jack pine and oak, often interspersed with large
openings or areas of sparse and scattered trees (Curtis 1959; Vogl 1964, 1970; McAndrews 1966).
Similar to the Great Plains without major topographic breaks, lightning or human-ignited fires are
likely to have swept uninterrupted across the landscape until reaching a major water body or
burning onto physiography supporting less flammable vegetation.
These “jack pine barrens” are unique to the upper Great Lakes region (Heinselman 1981)
and occupied over 400 000 ha in Michigan (Voss 1972, p. 62) and over a million hectares in
Wisconsin (Curtis 1959). The account of a reporter accompanying an expedition by botanists
William J. Beal and Liberty Hyde Bailey across the northern Lower Peninsula of Michigan in
June 1888 carried the title: A BARREN WASTE. MOSQUITOES THE LARGEST AND MOST
PROMISING PRODUCT YET FOUND (Voss and Crow 1976). The landscape is vividly
described:
These plains are clothed with a scant vegetation, the most conspicuous and common characteristic plant
being the jack, or scrub pine . . .. Fires often sweep over the plains, destroying all vegetation. Young pines
soon spring up in these burnt areas . . .. These groves of young trees are often miles in extent, and are so
dense that the traveler is completely hidden from view at the distance of a few paces . . .. The sward-like
glades are found to be clothed with stunted huckleberries or blueberries, miserable growths of sweet fern,
and a few other pinched and starved plants which can endure the heat and dryness of the sands . . .. On the
whole the plains are exceedingly uninviting in aspect.
Physiography is particularly important in this landscape because subtle topographic differences

are related to local soil differences, and both have large effects on vegetation composition and
growth. For example, physiographic differences perceived to be minimal have strong effects on
jack pine height growth, which in turn affect the duration of habitat suitability for the threatened
(formerly endangered) Kirtland’s warbler, which breeds nearly exclusively on this unique landscape
(Kashian et al. 2003; Walker et al. 2003).
Till Plains of the Midwest The extensive till plains of Indiana and Ohio are relatively flat or
gently rolling and crossed only by a number of low moraines, ridges, and shallow drainages. They
form the central core of the classic beech–maple forest region described by Braun (1950; Figure 2.6,
Chapter 2). The topography is extremely flat, but the parent material is heavily textured with soils
of moderate to poor drainage. The soils favor the dominance of American beech and sugar maple
in this region to the virtual exclusion of other tree species and the absence of fire. Leaf litter is
moist and rapidly decomposing, leaving insufficient fuel to carry a fire. Thus, the till plains are an
example of physiography where the substance (parent material) is more important than the form
in its direct and indirect control of dominant species and exclusion of fire.
Southeastern and Southern Coastal Plain A relatively flat plain supporting a mosaic
of landforms stretches along the southeastern and southern coast of the United States, from New
Jersey to Texas. Landforms include flat plains and terraces, embayed rivers, and sandy ridges and
swales near the Atlantic Coast giving way to rolling hills adjacent to the Piedmont Plateau. Over
short distances and very low topographic relief on the Coastal Plain, vegetation may vary from
grassland and savanna to shrubland, to needle-and broad-leaved sclerophyllous woodland, and to
rich mesophytic forest (Christensen 1988a; Myers and Ewel 1990). The region also supports the
most diverse assemblage of freshwater wetlands in North America (Ewel 1990). These stark
vegetative differences result in part from the effects of physiography on water table, drainage
patterns, and hydroperiod (length of time soils are saturated during a year); soil water and nutrient
cycling; and fire frequency and severity.
Fire is an important disturbance throughout much of the Coastal Plain (Chapter 16), and the
occurrence of rare local ecosystems there results from favorable physiographic and soil conditions
together with their relative protection from fire based on their position in the landscape. For example,
the positions of four distinctive ecosystem types in western Louisiana were reconstructed using
the General Land Office Surveys of 1821 (Delcourt and Delcourt 1974; Figure 8.11). The mesic
magnolia–holly–beech upland hardwood community occurred on thick loess deposits of the upland,
the magnolia–beech–holly bottomland hardwoods in ravine and river lowlands, and the tupelo-gum–
cypress community in swamp land adjacent to the Mississippi River. A mixed oak–pine–beech
ecosystem was found where the loess cap thinned out toward the northeast (Figure 8.11). The domi-
nance of southern magnolia, American beech, and their mesophytic associates in the Coastal Plain
was largely due to their protection by firebreaks from natural and human-set fires.
Floodplains Large rivers all across the world exhibit similar fluvial (river) processes and
atterns of physiography. Accordingly, specific fluvial landforms are identifiable, and complex pat-
p
terns of ecosystems and microsites occur within them. Rivers and their landforms are linear
R5W R4W R3W R2W R1W
T1S
MAGNOLIA–HOLLY–BEECH
UPLAND HDWDS.
MAGNOLIA–BEECH–HOLLY T2S
BOTTOMLAND HDWDS.
WHITE OAK–PINE–BEECH
FOREST
TUPELOGUM–CYPRESS SWAMP
T3S
31°W
LOUISIANA
MISSISSIPPI
31°N
T4S
F I G U R E 8 . 1 1 Pre-European colonization forest communities of West Feliciana Parish, Louisiana, as

reconstructed from 1821 General Land Office survey records. The dashed line represents the position of the
Mississippi River in 1821. Source: Delcourt and Delcourt (1974) / John Wiley & Sons.
f eatures so that a principle of similar effects is the result, in that similar ecosystems with
characteristic physiography, soils, and riverine vegetation may extend for tens or hundreds of
kilometers along a major river (at least before river processes were altered by dams, channels, and
human-made levees). Processes of flooding, sediment transport and deposition, and erosion and
abrasion by ice and water movement result in a degree of physiographic uniformity on the
landscape, whether a lowland river or a mountain stream. Nanson and Croke (1992) identified
three basic types of floodplains, including high-energy floodplains with non-cohesive sediments
(steep headwaters whose physiography is shaped episodically by extreme flow events); medium-
energy floodplains with non-cohesive sediments (featuring regular flow events in broad valleys);
and low-energy floodplains with fine-textured, cohesive sediments. Our emphasis in this section is
on the third of these floodplain types. Notably, stream floodplains or riparian zones are zones of
direct interaction (i.e., in flooding, bank cutting, and sedimentation) between terrestrial and
aquatic ecosystems (Gregory et al. 1991). Riparian characteristics and processes have received
detailed attention (e.g., Swanson et al. 1982b; Brinson 1990; Gregory et al. 1991; Malanson 1993;
Ward et al. 2002).
The pattern of zonation of landforms, ecosystems, and their vegetation is similar for large,
low-gradient, lowland rivers where landforms develop largely from lateral channel migration. In
this floodplain type, constant river meandering in the floodplain cuts into outer banks while
depositing sediments and forming point bars and new land downstream on inner banks
(Figure 8.12). Sedimentation during floods gradually but constantly alters the topography of the
floodplain. Coarser sediments are deposited adjacent to the river channel by floodwater, forming
OPEN WATER
BACKSWAMP
MISS
CLAY PLUG
ISSI
POINT BAR
PPI
cutoff
LEV
MEANDER SCROLLS EE
RIV
ER
OX RIDGES & SWALES

BOW
BA LE LAK
C V E
KS EE
W
AM
P
F I G U R E 8 . 1 2 Typical section of Mississippi River floodplain near False River, Louisiana. Physiographic
features include: (1) natural levees adjacent to the channel; (2) ridge-and-swale topography (meander scrolls)
associated with lands on and between former point bar/levee deposits as the channel migrated laterally and
downslope; (3) oxbow lake where former channel has been cut off; (4) backswamps of lower topography
behind the levees, and (5) point bars on inside curve of channel where deposition is rapid. Source: Brinson
(1990) / with permission of Elsevier.
well-drained ridges or natural levees. “Gallery” or riverside forests are typical of this landform
where high light availability along the river enables trees to lean over the bank and develop large
spreading crowns (Loehle 1986). However, such riverside species must tolerate periodic inunda-
tion during periods of flooding as well as drought when water levels recede, and the coarse sedi-
ments drain rapidly. Silts and clays are deposited beyond the levee in a poorly drained zone or
backswamp of the first bottom. A series of bottoms or terraces typically lie beyond the first bottom
and are progressively drier and less frequently flooded. Rivers further alter the physiography of
their floodplains as entire channels may be abandoned, forming oxbow lakes that eventually fill in
with sediment and vegetation. These depressions and lakes or sloughs support plant species best
adapted to inundation and anaerobic soils. Many floodplains exhibit ridge-and-swale topography
where channels, point bars, levees, and backswamps were cut off and abandoned by the mean-
dering stream (Figure 8.12), creating an enormous number of landforms inhabited by plant species
with different tolerances to hydroperiods and substrates.
The pattern of landform zonation and associated species occurrence is found across river
bottomlands of the southeastern United States (Figure 8.13). Floodplains are variable in their
number and kinds of landforms and the names associated with these features, but many landforms
of riverine ecosystems are quite typical (Wharton et al. 1982): (i) the river channel margin where
deposition builds a raised point bar that becomes new land and then the “front,” (ii) the ridge or
natural levee where coarse sediments are deposited, and (iii) a series of bottoms or terraces behind
the levee. Local relief of ridges, swales (low areas between ridges), and oxbows or sloughs are
found across the first bottom, although ridges and swales may differ in elevation by as little as 2 m.
Finer sediments are deposited across the first bottom, forming low, broad areas of poorly drained,
Upland forest (White Oak, Biackgum, White Ash, Hickories,

Winged Elm, Loblolly Pine)
Bald Cypress–Tupelo
Sycamore–Sweetgum–American Elm
Willow Oak– Upland
Sugarberry– Sweetgum–
Overcup Oak or Water Oak– Forest
American Elm or Willow Oak
Water Hickory Diamondleaf Oak-
Red Ash
Black Willow Swamp Chestnut Oak–
Cottonwood Sweetgum– Cherrybark Oak
Willow Oak
A B C D E F G H G I J
River Front First bottom (Terrace) Second Bottom Upland

Channel
F I G U R E 8 . 1 3 Correspondence between alluvial floodplain landforms and forest types for rivers of the
southeastern US. A, river channel; B, natural levee, C, backswamp or first terrace flat; D, low first-terrace
ridge; E, high first-terrace ridge; F, oxbow; G, second-terrace flats; H, low second-terrace ridge; I, high
second-terrace ridge; J, upland. Vertical scale is exaggerated. Source: Wharton et al. (1982) / U.S. Department
of the Interior / Public Domain.
slack-water clay or silty clay soils. Higher land of the second bottom is flooded less frequently and
for a shorter time, and thus supports a different set of species (Wharton et al. 1982).
Variation in the types of landforms on floodplains may arise from geographic location, size,
and velocity of the stream, kinds of materials transported, and the physiography and parent
material of adjacent terrain (Brinson 1990). In glaciated landscapes, for example, different fluvial
landforms may result when floodplains are located on different physiographic systems (outwash
plain, moraine, etc.). In a study of landscape ecosystems in riverine systems of northern Lower
Michigan (Baker and Barnes 1998), floodplains found on sandy outwash plains exhibited straighter
and shallower river channel segments and broad, continuous first bottoms (Figure 8.14). Where
outwash soils were deposited over moraines, floodplains had deeper and more sharply curving
river segments, narrow and discontinuous first bottoms, larger second bottoms, and generally a
more diverse set of fluvial landforms. These differences were attributed to differences in parent
material related to the physiographic system: sandier material characteristic of outwash plains is
more easily erodible and allows more lateral migration of the channel, but finer-textured glacial till
(a) (b)
Outwash Outwash/moraine
river valley river valley
(c) Outwash Plain
Transition from
outwash/moraine
to outwash plain
First Bottoms
Second Bottoms
Terraces
Rivers and Oxbows
Outwash/Moraine
F I G U R E 8 . 1 4 Typical sections of floodplain in outwash plains and outwash over moraine physiography
along the Pere Marquette River, northern Lower Michigan. (a) River valleys in outwash plains have broad,
continuous first-bottom floodplains with shallow, straighter river channels. (b) Valleys in outwash over
moraine physiography have narrow and broken first bottoms and larger second bottoms with deeper,
winding river channels. (c) These differences are evident at transitions across the landscape between the two
physiographic systems. Source: Baker and Barnes (1998) / with permission of Canadian Science Publishing.
parent material of the outwash over moraines causes deepening rather than lateral movement of
channels. Thus, broad physiography (outwash plain, moraine) affects the development of fluvial
landforms, which in turn affects the distribution and diversity of landscape ecosystems, which in
turn affects species occurrence and community composition (Baker and Barnes 1998).
The physiography of floodplains changes rapidly (temporal scales of years or decades)
relative to uplands where the geomorphic processes of sediment erosion, transport, and deposition
are far slower (temporal scales of centuries or millenia). The fact that fluvial landforms are vulner-
able to changes over short time periods suggests that vegetation may have an important influence
on the distribution and shape of fluvial landforms in floodplains. Indeed, a series of recent studies
of European rivers suggests that vegetation interacts with water and sediment to affect the river
channel and the landforms on its floodplain (Corenblit et al. 2007, 2009, 2016; Gurnell
et al. 2012, 2016; Wohl 2013). In particular, aquatic vegetation, riparian trees, and dead wood in the
river channel are effective trappers and stabilizers of sediment, leading to channel narrowing and
bending (Gurnell and Grabowski 2016) and the formation and alteration of fluvial landforms near-
est the channel, such as fronts, levees, and potentially first bottoms and backswamps. Thus, the
dynamic nature of physiography within floodplains provides for a positive feedback between flu-
vial landforms and floodplain vegetation that likely varies among landscape ecosystems.
As described in the earlier text, species adapted to floodplains must tolerate a range of hydro-
logical conditions depending on the landform where they grow. The physiological tolerances of
species to flooding related to topographic positions and effects of flooding have been described in
detail (Gill 1970; Hook 1984). Notably, diverse swamp ecosystems are associated with floodplains
throughout eastern North America, especially on the Coastal Plain (Christensen 1988a; Ewel 1990).
A great many studies emphasize floodplain landforms and vegetation for rivers from all regions of
the United States including northern Alaska (Bliss and Cantlon 1957), Oregon (Hawk and
Zobel 1974), North Dakota (Johnson et al. 1976), Texas (Chambless and Nixon 1975), New Jersey
(Wistendahl 1958; Frye and Quinn 1979), Virginia (Osterkamp and Hupp 1984; Hupp and Oster-
kamp 1985), and Indiana (Lindsey et al. 1961). An excellent review of North American rivers is
presented by Brinson (1990), and those of other continents are described in a book on forested
wetlands of the world (Goodall et al. 1990). Several reviews have examined the association bet-
ween vegetation and fluvial landforms (Ward et al. 2002; Lytle and Poff 2004).
PHYSIOGRAPHY AND FIREBREAKS

Firebreaks are places on a landscape that may interrupt the spread of a fire, such as rivers and
streams, lakes, or rough topography. Whereas fire spreads rapidly over flat land, significant surface
relief may disrupt the continuity of fuels, alter wind movements, retard air flow, and generally
slow and stop advancing fires (Grimm 1984; Taylor and Skinner 2003). Topography serves as a
natural firebreak most influentially in complex compared to gentle terrain (Heyerdahl et al. 2001;
Povak et al. 2018) where physiographic variation is most complex. In the Rocky Mountains of
North America, abrupt shifts between different aspects—such as valley bottoms and ridge tops or
between north and south aspects—may serve as effective natural firebreaks, as do cooler and
moister northern-facing slopes in general (Holsinger et al. 2016; Meigs et al. 2020). Several recent
studies have shown that low places on the landscape, such as depressions and valley bottoms, also
act as natural firebreaks because they generally have higher soil and fuel moisture than the sur-
rounding landscape (Ouarmim et al. 2014; Krawchuk et al. 2016; Wilkin et al. 2016; Povak
et al. 2018; Meigs et al. 2020).
There are many specific examples of how physiography affects fire spread and influences
vegetation distribution and composition. As described in the earlier text, escarpments and rivers in
the Great Plains acted as natural firebreaks and enabled trees to survive in an otherwise flat
landscape prone to sweeping fires. Tall, dense, deciduous forests on prairie soils in southcentral
Microlandforms and Microtopography 191
Minnesota, known as the Big Woods, existed more because of firebreaks than any other factor
(Daubenmire 1936; Grimm 1984). Remarkably, these forested ecosystems were dominated by
fire-sensitive species such as American elm, basswood, sugar maple, and hop-hornbeam. The loca-
tion, size, shape, and orientation of physiographic features in the Boundary Waters Canoe Area of
northern Minnesota (lakes, streams and wetlands, bedrock ridges, valleys and troughs) are related
to historic fire patterns and distribution of vegetation because of how they either favored or inhib-
ited the spread of fires (Heinselman 1973).
The highly shade-intolerant and fire-dependent longleaf pine dominated pre-European col-
onization forests throughout much of the Coastal Plain (Christensen 1988a; Myers 1990). How-
ever, in central Florida, late-successional communities lacking longleaf pine and dominated by
evergreen and deciduous angiosperms occurred on fire-protected islands and peninsulas. Because
of the pervasive presence of fire on uplands, mesophytic species such as American beech and
southern magnolia were relegated to the sites most protected from fire, such as the high bluffs east
of the Apalachicola River in the Florida Panhandle. A variety of mesophytic species occurred here
because of its protected position and its rich, moist soils, including two rare species that are
endemic to this bluff region—Florida yew and Florida torreya.
MICROLANDFORMS AND MICROTOPOGRAPHY

Microtopographic features, at scales less than 10 m2 or even 1 m2, also have important effects on
ecosystems and specifically influence regeneration of plants, succession, and local soil development.
Examples of microtopographic features include microsites created by tree-uprooting events, fallen
trees on the forest floor, and the elevated bases of living and dead trees in wetlands. The form and
substance of these biotic “landforms” create soil-site heterogeneity and provide favorable micro-
sites for plant establishment and growth over very large, forested landscapes.
TREE UPROOTING AND PIT-A ND-M OUND MICROTOPOGRAPHY

Probably the most widespread example of microphysiography and microsites on these landforms
is the pit-and-mound microtopography created by the uprooting of trees. Trees are consistently
felled by wind in forests, uprooted annually by strong storms, and over longer temporal scales by
hurricanes or tornadoes. An uprooted tree falls with much of its large roots intact, lifting the soils
among the roots and creating a depression or pit at the former position of the main mass of struc-
tural roots (Figure 8.15). The upthrown root and soil mass typically forms a mound adjacent to the
pit as it decays, forming another kind of microtopographic feature on the forest floor that is itself
colonized by woody plants. Pits and mounds usually occur in pairs, and their sizes vary because of
soil-site conditions, tree size and rooting habit, and the quantity of soil that slumps or erodes back
into the pit (Schaetzl et al. 1989a). Collectively over time, pairs of pits and mounds can form
significant microtopography over large areas. For example, in a landscape of sandy, glacial drift in
New Brunswick, 36% of the area was occupied by 1255 mounds per hectare and 12% by 1455 pits
per hectare (Figure 8.16; Lyford and MacLean 1966). Approximately 10–50% of the forest floor in
temperate forests may be covered by pit-and-mound microtopography (Schaetzl et al. 1989a).
Microtopographic features usually persist for fewer than 500 years, although Schaetzl and Follmer
(1990) documented a range of ages of these features from 10 to 2420 years using radiocarbon
dating. A few mounds in hardwood forests in the Upper Peninsula of Michigan have been
radiocarbon-dated to be as old as 4000–6000 years (Šamonil et al. 2013).
Tree uprooting and pit-and-mound topography do not occur randomly across a landscape;
they are associated with ecosystems subject to severe windstorms, particularly those having parent
material and soil that favor shallow rooting. Ecosystems with wet mineral or organic soils (high
water tables), rocky soils, or soils developing root-restricting horizons (e.g., clay-heavy soils) more
(a)
Root Plate
Mound
Pit
Treethrow Pit
STAGE 1: STAGE 2: STAGE 3:
Tree on Steep Slope Uprooted Tree Pit/Mound Pair
(Mound is Large & Downslope from the Pit)
(b)
Root Plate
Mound
Pit
Treethrow Pit
STAGE 1: STAGE 2: STAGE 3:
Tree on Flat Terrain Uprooted Tree Small Pit/Mound Pair
F I G U R E 8 . 1 5 The uprooting process, showing the formation of a pit-and-mound pair by soil slump off
the root plate, and the resulting mixed horizons within the mound. (a) Steep slope. (b) Flat terrain.
Source: Schaetzl and Follmer (1990) / with permission of Elsevier.
commonly feature tree uprooting. On glaciated upland landscapes, tree uprooting is much more
common on clay-rich parent material (till) than loamy or sandy outwash parent material. In slop-
ing or mountainous terrain, the abundance of pit-and-mound microlandforms is closely related to
landforms positioned low in the landscape (i.e., low, wetland areas with high water tables) and
those high in the landscape having steep slopes or exposed to strong winds. Pit-and-mound topog-
raphy also tends to be more pronounced in old-growth (larger trees) compared to second-growth
forests because the size and persistence of pits and mounds are closely related to the size of the
uprooted trees (Sobhani et al. 2014; Plotkin et al. 2017).
The ecological importance of pit-and-mound topography is that it creates heterogeneous soil
and microclimatic conditions where it occurs. Uprooting of mature trees may disturb an average
of 12–16 m2 of soil to a depth of 1 m or more (Lutz 1940; Peterson et al. 1990), bringing lower soil
horizons to the surface and creating significant soil mixing. As sediments slump and erode from
the root mass, mineral soil is exposed and often produces irregular or discontinuous horizons
within the soil mound (Schaetzl et al. 1990). Soil development is often accelerated beneath pits
more than on mounds (Stone 1975; Šamonil et al. 2015). Notably, the falling of trees downslope on
mountain slopes results in net downslope transfer of sediment and can trigger mass movement
and debris flow (Swanson et al. 1982a; Schaetzl et al. 1990).
Microlandforms and Microtopography 193
PLOT 2
N
(a)
0 20 40 60 80 100 FEET
PLOT 1
(b)
A A´
1-2˝ 2-3˝ 3-4˝ 4-5˝ 5-6˝ 6-7˝ 7-8˝ 8-9˝ 9-10˝ 10-11˝ 11-12˝
TREE DIAMETER
F I G U R E 8 . 1 6 Pit-and-mound microtopography. (a) Map of the spatial distribution of pits and mounds in a
30 × 100-ft plot in red spruce–balsam fir forest near Fredericton, New Brunswick, Canada. Dark areas are
pits; stippled areas are mounds. (b) Location and diameter size class of live trees (black circles) in relation to
pits (areas with dark vertical lines) and mounds (lightly stippled areas). Source: Lyford and MacLean (1966).
Microsite heterogeneity in pits and on mounds provides a variety of environments for

c olonization by woody species, and establishment of trees is almost exclusively on the lower slopes
of mounds where soil water and nutrients are most favorable (Figure 8.16). Pit centers may be
colonized soon after tree fall (Peterson, et al. 1990), but they are a poor location for tree establish-
ment and development because of wet conditions and deep accumulations of leaf litter (Schaetzl
et al. 1989b). Mounds tend to be more exposed to wind and are cleared of litter, lack competing
roots, and provide mineral-soil seedbeds ideal for species with tiny, wind-dispersed seeds
(e.g., hemlock and birch). In addition to favorable soil conditions, the canopy gap created by the
uprooting process provides sufficient light for establishment and growth of even shade-intolerant
species, if only temporarily. Thus, the uprooting process and pit-and-mound microlandforms not
only influence community composition of different ecosystems but also affect microscale eco-
system changes in light, topography, and soil conditions that, in turn, mediate vegetation change
(Clinton and Baker 2000; see Chapter 17). Reviews of the extensive literature on uprooting and
microtopography are provided by Schaetzl et al. (1989a,b, 1990) and Schaetzl and Follmer (1990).
MICROTOPOGRAPHY AND REGENERATION IN HARDWOOD SWAMPS

Microtopography and its effects in forest ecosystems are also associated with fallen and rotting tree
trunks and other coarse woody debris on the forest floor. These organic sources, often called nurse
logs, provide moist and competition-free establishment sites for seedlings of certain species. Nurse
logs in swamps of the southeastern Coastal Plain often bear distinctive vegetation (Schlesinger 1978;
Huenneke and Sharitz 1986). Swamps are waterlogged, periodically flooded systems that represent
a harsh growing environment for plants, and many other kinds of microtopography provide
favorable sites for regeneration. Titus (1990) identified 19 types of microsites in 4 different cate-
gories in a riverine swamp in Florida: (i) frequently submerged soil sites such as swamp bottoms,
(ii) raised-soil sites (elevated soil, soil near a stump, cypress knee, or shrub base, etc.), (iii) dead-
wood sites (fallen log or branch), and (iv) live-wood sites (trunk, root, cypress knee, palm tree
base). Distribution patterns of the seedlings of 25 woody species corresponded closely to flooding
duration and microsite substrate type, with the highest seedling densities on raised-soil sites. The
role of microtopography is enormous in wetland sites around the world where standing water and
low oxygen adversely affect seed germination, seedling establishment, and growth of woody plants
(Diamond et al. 2021).
SUGGESTED
READINGS
Gerhardt, F. and Foster, D.R. (2002). Physiographical H.H. Shugart and D.B. Botkin), 374–405. New York:
and historical effects on forest vegetation in Central Springer-Verlag.
New England, USA. J. Biogeogr. 29: 1421–1437. Huggett, R.J. and Cheesman, J. (2002). Topography and
Grimm, R.C. (1984). Fire and other factors controlling the Environment. Pearson Education 274 pp.
the Big Woods vegetation of Minnesota in the mid- Nanson, G.C. and Croke, J.C. (1992). A genetic classification
nineteenth century. Ecol. Monogr. 54: 291–311. of floodplains. Geomorphology 4: 459–486.
Hack, J. T. and Goodlett, J. C. (1960). Geomorphol- Rowe, J.S. (1988). Landscape ecology: the study of ter-
ogy and forest ecology of a mountain region in the rain ecosystems. In: Symp. Proc. Landscape Ecology
central Appalachians. Geol. Survey Prof. Paper 347. and Management. (ed. M.R. Moss). Montreal: Poly-
66 pp + map. sci. Publ. Inc.
Heinselman, M.L. (1981). Fire and succession in the Swanson, F.J., Kratz, T.K., Caine, N., and Woodmansee,
conifer forests of northern North America. In: Forest R.G. (1988). Landform effects on ecological processes
Succession, Concepts and Application (ed. E.C. West, and features. Bioscience 38: 92–98.
Soil CHAPTER 9
F orest trees, like all other terrestrial plants, require five primary resources for growth and
development: solar radiation, carbon dioxide (CO2), water (H2O), nutrients, and a porous
medium for physical support. Although plants obtain energy from the sun and CO2 from the
atmosphere, the remaining resources are provided by soil. Consequently, soil forms the “foundation”
of forest ecosystems in more ways than one. As you will see, soil is critical to the cycling of nutri-
ents (Chapter 19), a process that influences the growth of individual trees and the functioning of
entire ecosystems. In this chapter, we provide an overview of the physical, chemical, and biological
properties of soil that regulate the availability of soil resources to plants, particularly forest trees.
We begin with a discussion of the soil-forming process and explain how climate, geology, and biota
influence the distribution of forest soils.
There are many definitions of soil, as it pertains to the growth of terrestrial plants. For our
purpose, we define soil as a porous medium consisting of minerals, organic matter, water, and
gases. The combined influences of climate, topography, biota, and time differentiate geologic mate-
rials into soil. As such, soils are as diverse as the climates in which they occur, the landforms on
which they develop, and the plants that grow upon and within them. One would expect soils sup-
porting tropical rain forests to differ markedly in their physical, chemical, and biological properties
from those beneath forests in temperate or boreal climates for the reasons we have briefly outlined
in the text above. In the pages that follow, we review the processes that give rise to such differences
and focus on how they influence plant growth and ecosystem function.
PARENT MATERIAL
The Earth’s surface is blanketed by a wide array of geologic materials, differing in their chemical
composition and degree of consolidation. The relatively unweathered geologic material from
which a particular soil has developed is called parent material. It constitutes the basic substrate
for soil formation and exerts a substantial influence on the physical, chemical, as well as biological
characteristics of soil. Parent materials, as you will read, are typically associated with characteristic
kinds of landforms. Weathering is an important component of soil formation, because physical
abrasion and chemical dissolution differentiate freshly exposed geologic material (i.e., parent
material) into soil. Living organisms play an integral role in this process too, wherein organic acids
produced by plant roots and soil microorganisms solubilize minerals, allowing their elemental
constituents to be leached and deposited at depth. Additionally, the hydrolysis of CO2 resulting
from root and microbial respiration produces acidity, which further contributes to the dissolution
of minerals and the weathering process.
Parent materials are broadly classified as consolidated and unconsolidated (Figure 9.1).
Consolidated parent materials include igneous, sedimentary, and metamorphic rocks. A descrip-
tion of them can be found in most introductory geology texts, and Fairbridge (1972) provides a
particularly detailed discussion of their formation and chemical composition. Soil developing in
195
196 Chapter 9 Soil
Unconsolidated
Consolidated
Deposited in lakes
Lacustrine
Deposited by steams
Alluvium
Rocks ed
and ort Deposited by water
sp Marine
an
minerals
te r tr Deposited by ice
Wa Glacial till
Formed Deposited by water Outwash
Ice transported
in Ice contact
place Wi Lacustrine
nd
tra ed Alluvium
ns ort
sp
G
Residual po n Marine
ra
r te tra
vit
parents d nd
Wi
y
tra
material
ns
po
Deposited by wind Dunes (sand)

rte
Loess (silt)
d
Deposited by gravity
Colluvium
F I G U R E 9 . 1 Diagram illustrating how various kinds of parent material are formed, transported, and
deposited. Source: Modified from Brady (1990). Reprinted with permission of the Macmillan Publishing Co.,
Inc. as modified from The Nature and Properties of Soils, 11th ed., by Nyle C. Brady. Copyright ©1996 by
Macmillan Publishing Co., Inc.
consolidated geologic substrate is said to be formed in residual parent material (Figure 9.1). Forest
soils derived from residual parent materials occur throughout North America, primarily in areas
that have not been influenced by glaciation, moving water, or oceanic uplift. Forests of the Pied-
mont Plateau and Appalachian Uplands in the eastern United States occur on these materials, as
do forests of the Sierra Nevada, Cascade, and Rocky Mountains in the western United States.
Rates of soil formation on residual parent material composed of hard minerals can be quite
slow, and, in some situations, deep soils may never develop because erosion rates exceed those of
soil formation. This situation commonly occurs in mountainous regions in which steeply sloping
topography and exposed rock combine to form relatively thin soils. Nevertheless, neither all
consolidated parent materials weather at slow rates nor give rise to thin soils. The relative resistance
of rock to physical and chemical weathering is as follows (Birkeland 1974):
quartzite, chert > granite, basalt > sandstone,siltstone > dolomitee, limestone.
In general, rocks composed of insoluble minerals (i.e., quartzite, SiO2) are relatively more
resistant to weathering than those containing soluble minerals (i.e., calcite, CaCO3 or dolomite,
CaCO3•MgCO3), which rapidly weather in warm, humid climates. Residual parent materials also
yield very different chemical constituents as they are abraded and dissolved during the weathering
process. For example, the average SiO2 and Al2O3 contents of igneous and sedimentary rocks are
similar; however, their calcium (Ca) and sulfur (S) contents can dramatically differ (Table 9.1).
There can be considerable differences in the chemical constituents among soils derived from
residual parent material, with implications for tree growth. In the Ozark Mountains of Arkansas,
for example, different forest ecosystems develop on two kinds of residual parent material: black
and white oaks are overstory dominants on soil derived from chert (SiO2), whereas eastern red
cedar and northern red oak are dominant on limestone-derived (CaCO3) soil (Read 1952). Although
both of these parent materials are derived from sedimentary rock, differences in their chemical
composition and weathering rate markedly influence the distribution of forest trees.
Soil Formation 197
Table 9.1 Generalized chemical composition of igneous, metamorphic, and sedimentary rocks.
SiO2 Al2O3 Fe2O3 FeO MgO CaO K 2O P2O5 S

%
Igneous 69.9 15.2 2.0 2.0 2.2 4.0 3.3 0.20 0.04
Metamorphic 58.2 15.5 2.9 4.8 3.8 6.0 2.6 0.30 0.03
Sedimentary 49.9 13.0 3.0 2.8 3.1 11.7 2.0 0.16 0.18
Source: Modified from Ronov and Yaroshevsky (1972).
Unconsolidated parent materials are mineral particles that have been transported by water,
ice, wind, or gravity (Figure 9.1). They are chemically similar to the rock from which they origi-
nate, but are distinguished from residual parent material by being moved from their point of
geologic origin. The agent of transport has a substantial influence on the physical, chemical, and
biological properties of soils formed in unconsolidated parent materials—differences that often
influence forest composition and ecosystem productivity. In general, sediments deposited by water
and wind have a narrow particle-size distribution, whereas those deposited by ice contain frag-
ments that range in size from microscopic clay particles to boulders several meters in diameter. A
complete discussion of transported parent materials and the soils that they give rise to can be
found in Buol et al. (1980).
One example of how unconsolidated parent materials influence forest composition and eco-
system productivity comes from the glaciated portions of eastern North America. In this region,
the Wisconsin Glaciation (14 000 years before present) left behind a landscape consisting almost
exclusively of unconsolidated parent materials. Stratified materials were deposited by glacial melt-
waters (outwash), semi-stratified materials were deposited near the margins of stagnant ice (ice
contact), and unstratified materials were deposited directly by glacial ice (till). These parent mate-
rials differ markedly in their particle-size distribution and their ability to supply both water and
nutrients for plant growth. In the northern portion of Michigan’s Lower Peninsula, for example,
dry-oak-dominated (northern pin, black, and white oaks) ecosystems consistently occur on sandy
glacial outwash (72% coarse and medium sand). Mesic northern hardwood ecosystems occur on
till-derived soils with lower sand contents (55% coarse and medium sand; Host et al. 1988).
In addition to differences in species composition, aboveground productivity varies by a factor of
three between the dry oak ecosystems (1.3 Mg ha−1 y−1) and the mesic northern hardwoods
(3.4 Mg ha−1 y−1). Similar relationships also have been observed among geology, soil, forest compo-
sition, and ecosystem productivity in the glaciated portions of Wisconsin (Pastor et al. 1984).
Clearly, the type of geologic material in which soils develop exerts a profound influence on the
composition and function of forest trees and the ecosystem they compose.
SOIL FORMATION
As parent material weathers, and is colonized by plants and animals, it differentiates into more or
less distinct horizontal zones, giving rise to a soil profile. The type of soil profile that develops
depends upon the interaction of (i) climate, (ii) parent material, (iii) plants and animals occupying
the soil, (iv) slope and aspect of the land, and (v) the amount of time that has elapsed. Soil formation
is, in part, a chemical process resulting from the weathering of geologic material exposed to air and
water, and in part a biological process resulting from the activities of organisms growing on and in
soil, both roots and soil microorganisms.
SOIL PROFILE DEVELOPMENT

Forests comprise the natural vegetation in many of the moister parts of the Earth in which precip-
itation provides more water than can be evaporated or transpired by plants. Under such conditions
198 Chapter 9 Soil
two factors dominate the soil-forming process: (i) precipitation in excess of evaporation and
transpiration moves downward through the soil dissolving and transporting soluble minerals,
and(ii) tree roots remove both water and nutrients from the soil, transpiring most of the former
and eventually returning most of the latter to the soil surface as leaves, twigs, fruits, cones,
seeds, and fine roots. In temperate regions, the typical forest soil can be differentiated into five
zones or horizons, so identified by the soil-forming process occurring within them (Figure 9.2). A
soil horizon is differentiated from the overlying and underlying portions of soil by attributes that
can be easily identified by field observation.
The accumulation of organic matter on the mineral soil surface, or O horizon, is an attribute
unique to forest soils. This horizon consists of leaves, twigs, flowers, fruits, cones, and seeds that
have been deposited on the soil surface; the fine roots (<0.5 mm in diameter) also permeate this
soil horizon in some situations. The O horizon lies above the mineral soil and is distinguished from
OI Organic, Tree Leaves, slightly decomposed

Oe Organic, moderately decomposed
Oa Organic, highly decomposed
A Horizon of mineral soil mixed with

organic matter
E Horizon of maximum eluviation of

silicate clays, Fe and Al oxides, etc.
Bhs Horizon with accumulation of

Humus, Fe and Al oxides, and in
solis clay. Most clearly expressed
portion of the B Horizon
Bs Horizon of accumulation of Fe
and Al oxides
C Horizon of parent material.

Presence of Ca and Mg
carbonates in some soils
R Bedrock
F I G U R E 9 . 2 A theoretical mineral soil profile showing the major horizons that may be represented; realize
that not all soils contain all of these horizons. Their development is a function of climate, parent material,
topography, biota, and time. Source: Brady and Weil (1996) / Pearson. Reprinted with permission of the
Macmillan Publishing Co., Inc. from The Nature and Properties of Soils, 11th ed., by Nyle C. Brady. Copyright
©1996 by Macmillan Publishing Co., Inc.
Soil Formation 199
it by a high organic-matter content (>20% organic matter if soil has no clay; >30% if soil is more
than 50% clay). It can be divided into three subordinate horizons, each reflecting different stages of
decomposition: (i) the Oi horizon contains relatively “fresh” organic matter whose origin is easily
recognized (e.g., a white pine needle, or a sugar maple leaf), (ii) the Oe horizon is composed of
partially decomposed plant parts which are recognizable regarding their origin, and (iii) the Oa
horizon reflects the latter stages of decomposition and consists of well-decomposed organic matter
of unrecognizable origin (i.e., humus).
The A horizon marks the surface of the mineral soil and is characterized by: (i) the leaching
or eluviation of soluble minerals that migrate in the downward flow of water, and (ii) the
accumulation of organic matter originating from the overlying O horizon. In many forest soils,
the A horizon is dark in color (e.g., black) and is well structured, owing to its relatively high
organic-matter content (4–12%). The majority of aggregates contained within this horizon are of
crumb (<1–5 mm diameter) or granular (<1–10 mm diameter) size (see soil structure that follows).
The A horizon also is characterized by a large number of fine roots which actively forage for nutri-
ents released during organic-matter decomposition. Because of the shallow distribution of fine
roots and their decomposition products, most forest soils are characterized by a thin A horizon. By
contrast, A horizons of grassland ecosystems support deeply rooted grasses that incorporate
organic matter to depths of 50 cm.
In humid climates, the downward flux of water can move humus, silicate clays, Fe oxides,
and Al oxides from the surface soil, leaving behind light-colored, weathering-resistant minerals,
especially quartz (SiO2). The layer so formed is an E horizon, and it is distinguished from the
overlying A horizon by its light color (e.g., light gray or white). In cool, humid climates, E horizons
often develop under coniferous forest growing on course-textured (sand to sandy loam) parent
material. The acidity of this horizon is typically higher than the overlying or underlying horizons
due to the loss of base cations and presence of inorganic acids. Root densities within the E horizon
are low because few plant nutrients reside within this highly leached horizon.
Materials leached from either the A or E horizon migrate downward and are deposited at
depth to form the B horizon. The B horizon results from the process of illuviation
(i.e., accumulation) and is distinguished from other soil horizons by this important soil-forming
process. Materials that accumulate in the B horizon include silicate clays, humus, and Fe and Al
oxides, and can have a great impact on the physical, chemical, and biological properties of the B
horizon. In arid and semi-arid regions, CaCO3, CaSO4, and other salts can accumulate in the
B horizon.
The C horizon is the unconsolidated parent material underlying the A, E, and B horizons
and is outside the influence of the processes giving rise to the horizons above it. The A, E, and B
horizons can be derived from the same material contained within the C horizon. However, in areas
where geologic activity has deposited a relatively thin layer of mineral material over a previously
existing deposit, the A, E, or B horizon may be derived from a different parent material. This
situation is common along streams and rivers, where flooding and the subsequent deposition of
sediment can bury preexisting soil profiles. In regions where soil forms in residual parent material,
the C horizon is replaced by an R, the horizon that denotes the presence of the underlying
consolidated rock.
Note that the A, B, and C horizons refer to zones that have been leached, enriched, and unaf-
fected by soil-forming processes, respectively. It does not necessarily follow that the upper mineral
horizon is always the A horizon, or that all horizons are present in every soil. Following sheet
erosion, the B or C horizon may be exposed on the surface. Likewise, very “young” or unweathered
soils, like those forming on sand dunes or on any other recent geologic deposit, may lack a B
horizon and consist only of an A horizon and a C horizon. In these landscape positions, clay,
humus, or Al and Fe oxides often have not accumulated to any extent, thereby causing the absence
of a B horizon.
200 Chapter 9 Soil
PHYSICAL PROPERTIES OF SOIL

As plant roots grow within the soil, they anchor the aboveground portion of the plant and supply
it with water and nutrients. Plant roots also require oxygen for respiration, the supply of which is
controlled by the rate at which oxygen diffuses through water and other gases in soil. The physical
properties of the solid, liquid, and gaseous phases of soil have a substantial influence on the supply
of water, nutrients, and oxygen for plant and microbial metabolism, in addition to the availability
of physical space to anchor aboveground plant structures. Providing physical support for above-
ground tissues is of particular importance, because plants must properly orient leaves to capture
solar energy for use in photosynthesis. In shallow and poorly drained soils, windthrow is common
because belowground physical space is limited by the occurrence of shallow bedrock in the former
and anoxic (without O2) conditions in the latter. In the pages that follow, we discuss the physical
properties of soil that are of particular importance for the growth of forest plants and the func-
tioning of forest ecosystems.
SOIL TEXTURE
Soils are composed of mineral particles with a wide array of sizes and shapes. These particles, or
soil separates, are grouped into three size classes: sand, silt, and clay. Sand particles range in size
from 2.00 to 0.02 mm in diameter, silt ranges from 0.02 to 0.002 mm in diameter, and the clay
fraction is less than 0.002 mm. By convention, soil is composed of mineral particles that are less
than 2.0 mm in diameter; larger particles, like gravel and cobbles, are termed the coarse fraction,
because they do not contribute to water and nutrient supply for plant growth. Soil texture refers
to relative proportion of sand-, silt-, and clay-sized particles contained in a particular soil. This
physical property plays an integral role in regulating the availability of water and nutrients for
plant uptake, as well as the rate at which gases (O2 and CO2) are exchanged between soil and the
overlying atmosphere.
In addition to grouping mineral particles by size, particles in different size classes can be
distinguished by their physical and chemical properties. For example, sand-and silt-sized particles
are chemically identical to the rock from which they originate and are called primary minerals.
They are round or irregular in shape, composed primarily of quartz (SiO2) or other silicate min-
erals such as orthoclase (KAlSi3O8) or plagioclase ([Ca, Na][Al, Si] AlSi2O8). As a consequence of
their size and shape, sand particles have relatively low surface areas (1–2 m2 g−1) with large pores
between individual particles (1 g of soil would form a mound in the palm of your hand equal to the
size of a United States quarter coin). These attributes provide sandy soils with good aeration, but
limit the amount of water they can retain for plant use. Because silt-sized particles have higher
surface areas (45 m2 g−1), they hold relatively larger quantities of water compared to a soil consist-
ing primarily of sand.
The clay fraction of soil found in the B horizon comprises secondary minerals, which
result from the physical and chemical weathering of primary minerals. In temperate soils, clays
consist primarily of aluminosilicate minerals that differ markedly from sand and silt in their shape
and mineralogy. These minerals form plate-like structures or micelles that are referred to as
phyllosilicate (leaf-like silicate) clays. Due to their shape, they have a high surface area
(80–800 m2 g_1), enabling them to hold relatively large quantities of water for plant use. Phyllosili-
cate minerals differ widely in their physical and chemical properties from Al and Fe oxides, which
compose the clay fraction of soils in the humid tropics. Differences in clay minerals found in
temperate and tropical soils will be discussed in more detail later in this chapter (see Section 9.4,
Chemical Properties of Soil). The physical weathering of mineral material, such as through the
grinding action of glacial ice moving across the landscape, also can produce clay-sized particles
that can be found throughout a soil profile. These retain the chemical properties of the minerals
Physical Properties of Soil 201
100 0
90 10
80 20
70 30
Clay
60 40
y
Pe
cla
rc
nt
en
50 50
rce
Silty
ts
Pe
ilt
Sandy clay
40 clay 60
Clay loam Silty clay
30 Sandy clay loam 70
loam
20 Loam 80
Sandy loam Silt loam
10 90
Loamy Silt
Sand sand
0 100
100 90 80 70 60 50 40 30 20 10 0
Percent sand
F I G U R E 9 . 3 The soil textural triangle indicating the relative proportions of sand, silt, and clay composing
each textural class. Note that some textural classes encompass a range of particle-size distributions (i.e., the
clay class ranges from 60 to 100% clay), whereas other textural classes are narrowly defined (i.e., sand class
ranges from 90 to 100% sand). The particular soil (70% sand, 20% silt, and 10% clay) represented by an
asterisk in the sandy loam class can be used to illustrate the use of this diagram. Begin by finding the lines of
equal value for 70% sand and 20% silt. The intersection of those lines with the 10% clay lines locates the
position of the asterisk, and the textural class of the soil. Importantly, the heart center of this diagram (clay
loam) is the particle-size distribution that provides the maximum supply of water and nutrients to plants.
from which they are derived, which markedly differ from phyllosilicates clays that are formed by
weathering and deposition into the B horizon.
Soils are grouped into textural classes based on their sand, silt, and clay content. For example,
a soil consisting of equal proportions of sand, silt, and clay is classified as a loam (Figure 9.3). For
a soil to be classified as clay, 60% of its particles must be less than 0.002 mm in diameter. By con-
trast, sands are soils in which more than 90% of the mineral particles range from 0.02 to 2.00 mm
in diameter. The textural classes illustrated in Figure 9.3 have been delineated with specific refer-
ence to plant growth. As you will read later in this chapter, water availability, nutrient supply, and
aeration all are substantially influenced by this all-encompassing soil property.
SOIL STRUCTURE
Primary soil particles (i.e., sand, silt, and clay) are arranged into secondary structures called
aggregates or peds, which result from the combined activities of plants and soil microorganisms.
Plant roots enmesh and compress mineral particles and bind primary particles to one another.
Further pressure can be exerted during wetting and drying cycles, because clays expand as they
202 Chapter 9 Soil
hydrate and contract as they dry. Organic compounds (polysaccharides) excreted by plant roots
and produced by microbial activity during decay of plant litter function as cementing agents
binding to the surface of one or more soil particles. In combination, these processes make soil
aggregates highly stable structures that often remain intact even when immersed in water.
Aggregates range from single-grain structure, in which soil particles are totally unattached,
to massive structure, in which all soil particles adhere to one another in large clods. Beach sand is
an example of the former, whereas the latter often occurs in poorly drained soils with a high clay
content. The degree of aggregation most conducive to plant growth lies somewhere between these
extremes. Within this mid-range, aggregates vary in size from granules several millimeters in
diameter to blocks, prisms, or columns several centimeters in size. Granular aggregates commonly
occur in the A horizons of many forest soils, whereas blocky aggregates often occur in B horizons.
Plant roots generally occupy the spaces between aggregates, rather than growing through or within
them. These spaces, or macropores, have a substantial influence on the rate at which water and
gases move into and through the rooting zone. As a result, a well-aggregated soil will typically hold
more water and will have better aeration than an unstructured soil of the same texture.
SOIL COLOR
Color provides insight into many physical and chemical properties of soil, particularly organic-
matter content and drainage (or aeration). The surface horizons of mineral soils are generally dark
in color, reflecting organic-matter additions from leaf and root litter. A deep-colored surface soil,
dark in color (e.g., black or dark brown), usually contains relatively greater amounts of organic
matter than a thin, light-colored surface soil. The subsoil of sandy forest soils in northerly climates
can contain accumulations of organic matter and iron that have been leached from the surface soil.
This subsurface accumulation is easily identified by its dark chocolate-brown color (i.e., Bh or
Bhs horizon).
In addition to providing insight into organic-matter content, soil color can provide qualitative
information regarding soil drainage. Most soils contain large amounts of iron (Fe), which, in an
oxidized state (Fe3+), is bright orange or red. This condition occurs when soils are well aerated and
O2 rapidly diffuses into the soil profile. However, during prolonged periods of water saturation, O2
in soil can be depleted if the demand by plant roots and soil microorganisms exceeds the diffusion
rate of O2 in water. In these situations, Fe is reduced (Fe2+) producing compounds blue-gray to gray
in color. Zones of mottling, a patchwork of yellow-orange and blue-gray colors, indicate the
presence of both oxidized and reduced conditions. This situation occurs where the level of
the water table fluctuates within the soil profile, and color can be used to determine if it lies
within the rooting zone of plants. High water tables in poorly drained landscape positions greatly
restrict the rooting depth of trees, making them prone to windthrow. Using soil color, one can
easily identify soils in the field with poor drainage that often restrict the growth of many forest trees.
SOIL WATER
Water availability controls the global, regional, and local distribution of plants on Earth. For
example, forests occur in regions in which the annual amount of water supplied by precipitation
exceeds that which is lost through evaporation and transpiration. Although broadscale patterns of
precipitation control the total amount of water entering ecosystems, it is the interaction of water
molecules with soil particles that largely influences the amount of water that actually can be used
by an individual plant for growth.
Water flows along a continuum extending from the atmosphere, through the plant, and into
soil. The force driving water movement along this continuum is transpiration; water is actually
“pulled” from soil and through plants into the atmosphere by this process. Transpiration at the leaf
surface creates tensions that are translated down water columns extending through plants and into
soil. Inasmuch as soils can be “dry” to plants, but can still contain substantial amounts of water.
This relationship may occur if the force holding water in soil exceeds the force (i.e., tension) created
by the transpiration of water at the leaf surface. Understanding the dynamics of water along the
atmosphere–plant–soil continuum has clear relevance for the study of plant growth and ecosystem
function, because these dynamics directly control the amount of water available for plant use.
Physical Properties of Water The physical properties of water greatly influence its availabil-
ity to plants. Water (H—O—H) molecules have a net positive charge on one side of the molecule
and a net negative charge to the other. Mineral particles in soil also have charged surfaces that
attract water molecules, due to the dipole nature of the water molecule. The attraction of positively
and negatively charged bodies is termed adhesion, a force in soil that greatly influences the
amount of water available for plant use. Adsorped water molecules, those attracted to charged sur-
faces in soil, are linked to other water molecules through hydrogen bonding, a chemical bond
linking the oxygen atom (−) of one water molecule to a hydrogen atom (+) of another. Hydrogen
bonding gives rise to the cohesive force, or cohesion, joining water molecules into chains or poly-
mers that extend away from the surface of mineral particles.
Due to the strong attraction of water molecules to charged surfaces in soil, adsorped water mol-
ecules are closely packed and exist in an energy state less than that of pure water. Although water in
direct contact with mineral particles is strongly held to their surface by adhesion, that force dimin-
ishes as the distance from the solid surface increases, much like the attraction of a magnet for iron
diminishes as the distance between them increases. When soils are saturated, some water molecules
are only weakly attracted to the surface of mineral particles, because they lie at relatively large
distances from any charged surface (i.e., weak force of attraction due to adhesion). Water draining
from saturated soil does so because the Earth’s gravitational pull exceeds the adhesive and cohesive
forces holding a portion of soil water. As soil continues to dry, either through plant uptake or evapo-
ration, the forces holding the remaining water molecules steadily increase as the layer of water sur-
rounding particle surfaces diminishes. Adhesive and cohesive forces holding water in soil are tensions
that must be overcome if plants are to extract water from soil. Because the forces holding water in soil
can exceed those imposed by plants, only a proportion of the water in soil is available for plant use.
That is why soils can contain sufficient quantities of water, but can “feel dry” to plants.
In addition to being attracted to charged surfaces, water molecules also are attracted to ions
with net positive (cations) or negative (anions) charges. Salts, such as NaCl, dissolve in water
because the attraction of water molecules for cations (Na+) and anions (Cl−) is much greater than
the attraction between cations and ions. The strong attraction of water for positively or negatively
charged ions causes water molecules to lose energy as they hydrate either type of ion. Because
water molecules lose energy as they associate with cations and anions, water in soil has a lower
energy status than pure water.
The semi-permeable membrane surrounding plant and microbial cells in soil (i.e., plasma-
lemma) functions as a “molecular sieve,” allowing water to transverse while excluding larger,
hydrated ions. The movement of water molecules across any semi-permeable membrane in
response to differences in ion concentration (i.e., the energy status of water inside versus outside
the cell) is osmosis. Water molecules moving across a semi-permeable membrane exert a force
known as osmotic pressure. Osmosis, and the energy it produces, is of particular relevance to plant
growth, because it influences the movement of water into and out of plant and microbial cells. For
example, plants under salt stress suffer from a lack of available water, because dissolved ions in soil
water lower its energy status to a point where it is less than that in the plant cell. Water flows from
the plant into soil solution by osmosis in this particular situation. This can cause the water content
of plant cells to decrease to such a low level that physiological processes are impaired, and the
plant is no longer able to maintain turgor.
204 Chapter 9 Soil
Soil Water Potential The movement of water in soil, its uptake by plant roots, and its loss to
the atmosphere from the leaf surface are all energy-related phenomenon. In soil, adhesion,
cohesion, the presence of dissolved ions, and the Earth’s gravitational pull are the primary forces
influencing the energy status of soil water, and hence the movement of soil water and the
proportion of it are available for plant use. Forces in soil acting on water molecules can be pres-
sures (gravity) or tensions (adhesion and cohesion), both of which are measured in megapascals
(MPa; values less than 0 MPa are tensions and those greater than 0 MPa are pressures). In
combination, these forces control how much of soil water is actually available for plant use.
Soil-water potential is the energy status of soil water; it also can be thought of as the effec-
tive concentration of water in soil. By definition, the potential of pure, liquid water at 20 °C and at
standard atmospheric pressure is 0 MPa. Pure water is used as a standard reference point from
which we measure the influence of adhesion, cohesion, dissolved ions, and gravity on the energy
status of soil water.
Adhesion and cohesion, the forces holding water molecules to charged surfaces and to one
another, give rise to the matric potential of soil water. Because adhesion and cohesion lower the
energy status of soil water (i.e., relative to pure water), matric potentials are less than 0 MPa. As
such, matric potentials are tensions holding water molecules to one another and to the surfaces of
charged particles in soil. The presence of dissolved ions, which also lower the energy status of soil
water relative to pure water, gives rise to the osmotic potential of soil water. Gravitational
potential results from the downward force of gravity and its ability to extract water from soil.
Gravitational potentials are greater than zero (i.e., pressures), because the downward pull of gravity
extracts water from soil. In combination, matric, osmotic, and gravitational potentials give rise to
the total soil-water potential, which represents the summed energy status of soil water. In most
well-drained soils, matric potential is the most important factor regulating the supply of water to
the root surface.
Because adsorption and dissolved ions lower the free energy status of soil water, plants must
expend energy to remove water from soil. As such, the water potential (or energy status) of plants
must be lower than that of soil, if water is to flow from soil into plant roots, up the stem, and out
of leaves. Transpiration at the leaf surface drives the flow of water along the energy-related path
from soil to the atmosphere. The atmospheric water potential, which greatly influences the
transpiration rate of plants, is largely determined by relative humidity and air temperature. The
concentration of water in the atmosphere is much less than the concentration of water in either
plants or soils, and as a consequence, atmospheric water potentials are more negative (i.e., at a
lower energy status) than those of plants or soil. It is not unusual for atmospheric water potentials
to attain values of −100 MPa, values 10–100 times more negative than those in plants or soil
(Bidwell 1974). Because water flows from a region of high potential (i.e., high-energy state) to one
of low potential (i.e., low-energy state), water moves from soil into plant roots, through the vascular
system of the plant, and into the atmosphere via stomates in leaves. Large negative atmospheric
water potentials drive the process of transpiration and the flow of water from soil, through plants,
and into the atmosphere.
Although large negative water potentials at the leaf surface are translated downward to the
root surface, plants are generally unable to extract soil water held by potentials less than −1.5 MPa
(i.e., a tension of 1.5 MPa; Figure 9.4). Under these conditions plants wilt and are unable to regain
turgor even following the addition of water. At a potential of −1.5 MPa, soil water has attained the
permanent wilting point, which defines the lower limit of plant available water (Figure 9.4).
Field capacity represents the upper limit of plant available water and is the amount remaining in
soil after it has freely drained due to the downward pull of gravity (a potential of +0.01 MPa). The
quantity of water bounded by field capacity and the permanent wilting point represents the
Soil-water Hygroscopic Wilting Field

interface water point capacity
Soil 100 3.1 0.01 MPa

1.5 MPa
solid MPa MPa
Flows
with
gravity
Capillary water
100 MPa
Tension
0.01 MPa
Very thin Very thick

film Increasing moisture film
F I G U R E 9 . 4 Diagrams showing the relationship between the thickness of water films and the tension with
which water is held by soil particles. Tensions are presented in megapascals in the upper illustration. The
thickness of water films in relationship to matric potential is presented in the lower figure. Source: Brady
(1974) / Springer Nature. Reprinted with permission of the Macmillan Publishing Co., Inc. from The Nature
and Properties of Soils, 8th ed., by Nyle C. Brady. Copyright ©1974 by Macmillan Publishing Co., Inc.
vailable water content (ml cm−3) of soil. It differs substantially from the saturation water
a
content, which is the total amount of water that can be stored in all soil pores (i.e., no air-filled
pores; Figure 9.4). The reader is referred to Kramer (1983) for a complete treatment of plant–water
relations and soil-water dynamics.
Soil texture substantially influences the available water content of soil. Figure 9.5 illustrates
the relationship between water content (mL of water per cm3 of soil) and soil-water potential for a
clay, loam, and sand. At field capacity (i.e., water held by tension more negative than −0.01 MPa),
the clay soil holds approximately 0.6 ml of water per cm3 of soil, almost 2.5 times more water than
the sand (calculated from Figure 9.5). The adhesive properties of water, in combination with the
large surface area of clay-sized particles, allow the clay to hold more water at any given potential
than the sand. Notice that silt loam in Figure 9.5 contains the greatest quantities of plant available
water. It does so because of the favorable distribution of macropore and micropore spaces. The
available water content of sand, calculated from Figure 9.5, is 0.15 ml cm−3, approximately 50% of
that held by the silt loam (0.36 ml cm−3) and clay (0.33 ml cm−3).
206 Chapter 9 Soil
0.40
0.32
Soil H2O Content mL cm–3
Field
capacity
0.24 Available water
0.16
Wilting
0.32 coefficient
Unavailable water
0
Sand Sandy Loam Silt Clay Clay
loam loam loam
Fineness of texture
F I G U R E 9 . 5 The relationship between soil texture and the available water content of soil. Note that field
capacity increases from the sand to the silt loam and then levels off as the proportion of clay increases.
Because the permanent wilting point increases linearly as a function of soil texture, the largest amount of
plant available water occurs in soil with a silt loam texture. Source: Brady and Weil (1996) / Pearson.
Reprinted with permission of the Macmillan Publishing Co., Inc. from The Nature and Properties of Soils,
11th ed., by Nyle C. Brady. Copyright ©1996 by Macmillan Publishing Co., Inc.
CHEMICAL PROPERTIES OF SOIL

Plant life is constructed from a surprisingly small suite of elements, whether we consider the
majestic redwoods of northern California or the single-cell alga growing on the soil surface. In
Table 9.2, we summarize the chemical-building blocks from which plant life is constructed and
their source within terrestrial ecosystems. Each is required for plant growth and development,
albeit in different quantities. Macronutrients are those elements required in relatively large
Table 9.2 Macronutrient elements required by plants and their source

within terrestrial ecosystems.
Element Symbol Source

Carbon C Atmosphere
Hydrogen H Water
Oxygen O Atmosphere, water
Nitrogen N Organic matter, atmosphere
Phosphorus P Mineral soil, organic matter
Potassium K Mineral soil, organic matter
Sulfur S Mineral soil, organic matter, atmosphere
Magnesium Mg Mineral soil
Calcium Ca Mineral soil
Source: Brady (1990) / Springer Nature.

Chemical Properties of Soil 207
Table 9.3 The biochemical function of plant macronutrients, their form of uptake, and typical leaf
concentrations in plants.
Form Leaf
Element Biochemical function(s)
assimilated concentration
Carbon (C) Form the basic building blocks of all CO2, H2O 90–98%
Hydrogen (H) biologically active compounds
Oxygen (O)
Nitrogen (N) Nucleic acids, amino acids, proteins, NH4+, NO3− 1–4%
chlorophyll, anthocyanins, alkaloids
Phosphorus (P) Nucleic acids, nucleotides, sugar H2PO4− 0.1–0.4%
phosphates, phospholipids
Potassium (K) Enzyme co-factor, osmotic regulation, K+ 1%
cell ion balance
Calcium (Ca) Pectin synthesis and cell-wall formation, Ca2+ 0.8%
metabolism/formation of nucleus and
mitochondria, enzyme activator
Sulfur (S) Amino acids, proteins, sulfolipids SO42− 0.2%
Magnesium (Mg) Chlorophyll, enzyme co-factor Mg2+ 0.2%
Source: Salisbury and Ross (1992) / Wadsworth Publishing.
amounts. They are commonly found as constituents of nucleic acids, proteins, carbohydrates,
lipids, and chlorophyll (Table 9.3). Micronutrients (Fe, Mn, Bo, Mo, Cu, Zn, Cl, and Co), as their
name implies, are required in relatively small amounts and occur as co-factors in enzymatic
reactions. Although micronutrients are required in small amounts, they are nonetheless important
in the biochemical functioning of plants and entire ecosystems; all are supplied to plants by
chemical processes in soil. Further discussion regarding the biochemical and physiological
functions of plant nutrients can be found in Salisbury and Ross (1992).
Although plants assimilate carbon dioxide and oxygen from the atmosphere, the majority of
macronutrients and micronutrients are supplied by ion-exchange reactions, mineral weathering,
or organic-matter decomposition—processes all occurring within soil. The supply of nutrients
often limits the growth of individual plants and entire ecosystems. Nitrogen (N) availability, for
example, is known to limit the growth of many boreal and temperate forests (Flanagan and Van
Cleve 1983; Pastor et al. 1984), whereas phosphorus (P) has been observed to constrain forest
growth in the humid tropics (Vitousek 1984; Vitousek and Sanford 1986). In the following
discussion, we explore some of the chemical processes in soil that regulate the supply of nutrients
for plant growth in both temperate and tropical forest soils. These processes, in combination with
plant uptake and litter decomposition, control the cycling of nutrients within forest ecosystems
(Chapter 19).
CLAY MINERALOGY
In studying Table 9.3, note that many macronutrients exist as cations, the ionic form assimilated
by plant roots. The mineralogy of clay particles and the ion-exchange reactions mediated by their
negatively charged surfaces substantially influence the supply of cations for plant growth.
Ion-exchange reactions in soil also are an important mechanism influencing the retention and loss
of nutrients from forest ecosystems, especially following disturbances such as harvesting, large-
scale windthrow, or fire.
208 Chapter 9 Soil
Aluminum or Oxygen or
Silicon Oxygen magnesium hydroxyl
F I G U R E 9 . 6 The structure of the silica tetrahedra and the alumina octrahedra that form the basic building
blocks of phyllosilicate minerals.
Table 9.4 Physical and chemical properties of some common phyllosilicate minerals.
Property Fine-grained mica Montmorillonite Kaolinite Chlorite

Ratio of octahedral to 2:1 2:1 1:1 2:1:1
tetrahedral sheets
Size (μm) 0.1–5.0 0.01–1.0 0.5–5.0 0.1–2.0
Shape Plates Flakes Hexagonal Variable
Flakes Crystals
External surface (m2 g−1) 50–100 70–120 10–30 70–100
Net negative charge 100–180 80–120 2–5 15–40
(cmol kg−1)a
a
Centimoles of negative charge per kilogram of dry soil.
Source: Brady (1990) / Springer Nature.
Clay minerals form during the weathering process and, to some extent, chemically reflect
the primary minerals from which they originate. Phyllosilicate minerals dominate the clay fraction
of temperate soil and originate from a wide array of primary minerals including feldspar, ortho-
clase, and hornblende. Although phyllosilicate clays may differ in mineralogy, they are all formed
from the same chemical-building blocks. The primary structures of these minerals are the silica
tetrahedra (SiO4) and the alumina-magnesia octahedra (AlO6 or MgO6; Figure 9.6). By sharing O
atoms at their corners, these subunits can link to form tetrahedral ([AlO6]n) or octahedral sheets
([MgO6]n), secondary structures that confer many unique properties to phyllosilicate minerals.
Fine-grained mica, vermiculite, chlorite, montmorillonite, and kaolinite are common phyl-
losilicate minerals in temperate soil, and some of their physical and chemical properties are sum-
marized in Table 9.4. The weathering process, which gives rise to clay minerals, substantially
influences on the extent to which phyllosilicate clays attract and bind cation nutrients. As a
consequence, unweathered soils and highly weathered soils differ markedly in clay mineralogy
and consequently their ability to retain and release cation nutrients for plant growth.
Isomorphic substitution conveys a net negative charge on phyllosilicate minerals enabling
them to attract cations and supply them for plant growth. This process occurs as one type of clay
mineral (i.e., montmorillonite) weathers and gives rise to another (i.e., kaolinite). During the
formation of phyllosilicate minerals, atoms of the same size, but of lower charge, replace either Si
in tetrahedral sheets or Al in the octahedral sheets. The extent to which a particular clay is
substituted is influenced by the chemical constituents of the parent material and the degree to
which it has weathered. This property is a permanent attribute of phyllosilicate minerals and is
relatively unaffected by changes in soil acidity. Because of a low degree of isomorphic substitution,
kaolinite, a highly weathered clay mineral, has a low net negative charge that is balanced by a
small number of cations adsorbed into its crystalline surface (Table 9.4). Montmorillonite, a clay
mineral less weathered than kaolinite, is highly substituted in its octahedral sheets and has a
relatively large net negative charge, therefore attracting a greater number of cations to its surface.
The aforementioned examples illustrate how weathering directly controls the ability of phyllosili-
cate clays to adsorb and exchange cations with soil solution, a factor that has a profound effect on
the supply of nutrients for plant growth and the ability of ecosystems to retain nutrients
against leaching.
Following long periods of intense weathering (i.e., 100 000–1 000 000 years), Si is lost from
the structure of phyllosilicate minerals leaving behind oxides of Fe and Al, which constitute the
clay fraction of some highly weathered soils in the humid tropics. These oxides differ markedly in
their chemical characteristics from the phyllosilicate minerals from which they originate. In con-
trast to the ordered crystalline structure of phyllosilicate minerals, oxides of Fe and Al are amor-
phous (without form) and exist in a less-ordered semi-crystalline state (Schwertmann and
Taylor 1989; Sposito 1989). More importantly, these minerals exhibit a pH-dependent variable
charge, in contrast to the stable, net negative charge of phyllosilicate minerals. This property
results from the large number of hydroxyl groups (−OH) contained within these minerals; the for-
mulae for goethite [FeO3(OH)3] and gibbsite [Al(OH)6] illustrate this point.
Soil acidity in highly weathered tropical soils plays an important role in the functioning of
clay minerals and their ability to supply plants with nutrients. In acidic soils, the hydroxyl groups
of the Al and Fe oxides are protonated, conveying a net positive charge and the ability to adsorb
and exchange anions. However, these minerals lose protons (H+) in relatively alkaline soil (e.g.,
AlO(OH)5− + H+), thereby producing a net negative charge and the ability to adsorb and exchange
cations. As a consequence, soils dominated by Al and Fe oxides have a pH-dependent charge and
also a point of zero charge at which neither cations nor anions are adsorbed; both cations and
anions are susceptible to loss through leaching. Land management practices which alter soil pH
clearly have the potential to alter the ability of tropical soils to adsorb and retain plant nutrients.
For further elaboration on the dynamics of variable-charge soils and nutrient mobility within
them, we refer readers to Sollins et al. (1988) and Uehara and Gillman (1981).
CATION EXCHANGE AND THE SUPPLY OF NUTRIENTS

The cation-exchange capacity of soil is a general measure of plant nutrient availability and rep-
resents the total amount of cations (centimoles of positive charge) that can be adsorped by a kilo-
gram of soil. Cations in soil solution exist in an equilibrium with those adsorbed to cation-exchange
sites on clay micelles. When cations in soil solution are assimilated by plants, those adsorbed to
clay particles are released and a new equilibrium is established. As such, adsorbed cations can be
thought of as a “reservoir” of plant nutrient in soil. Because cations vary in size and charge, they
are adsorped and exchanged in a predictable manner reflecting their affinity for exchange sites on
clay micelles. In general, cations with a small, hydrated radius and a large positive charge are most
strongly held. We have listed the cations commonly found in soil solution in order of
decreasing affinity:
+
Al3+ > H+ > Ca 2+ > Mg2+ > K + = NH 4 > Na + .
Weakly adsorbed cations such as K+ and NH4+ are more available for plant growth; they also
are more susceptible to leaching than others higher in the order.
210 Chapter 9 Soil
Table 9.5 The relationship among soil texture, clay mineralogy, and the cation-exchange capacity
in surface forest soils of the eastern United States.
Clay Cation-exchange Organic

Vegetation/physiography Texture Mineralogy
(%) capacity (cmol kg−1) matter (%)
I. Northeastern US
Jack pine/outwash plain Sand 2 Mixed 13 6.4
Northern pin oak–black Sand 4 Mixed 11 6.0
oak–white pine/outwash
plain
Sugar maple–bass- Loamy 6 Mixed 18 12.0
wood/moraine sand
II. Southeastern US
Loblolly pine/coastal plain Silt loam 7 Montmorillonitic 11 3.9
Loblolly pine/coastal plain Silt loam 20 Montmorillonitic 20 3.7
Loblolly pine/coastal plain Silty clay 37 Montmorillonitic 28 2.3
Water oak–willow–oak/ Silty clay 39 Montmorillonitic 30 7.0
central prairie region
Loblolly pine/upper Loamy sand 5 Kaolinitic 4 0.9
coastal plain
Loblolly pine/upper Sandy loam 11 Kaolinitic 7 1.3
coastal plain
Loblolly pine/upper Clay loam 28 Kaolinitic 26 2.3
coastal plain
We summarize the cation-exchange capacity for forest soils of different texture and miner-
alogy to illustrate the combined influence of weathering and parent material on supply of cations
for plant growth (Table 9.5). Note that cation-exchange capacity increases as the percentage of clay
also increases. Also note that soils dominated by kaolinite, a weakly substituted and highly weath-
ered phyllosilicate mineral, generally have lower cation-exchange capacities than those soils con-
taining montmorillonite. The soils of mixed mineralogy contain a mixture of phyllosilicate clays
and have relatively high cation-exchange capacities, even though clay contents are relatively low.
As you will read later in this section, the high organic-matter contents of these soils greatly con-
tribute to their ability to supply plants with cation nutrients.
The proportion of cation-exchange sites occupied by Ca2+, Mg2+, K+, and Na+ is the percent
base saturation of soil. These cations are not technically bases, because they do not directly neu-
tralize H+ in soil solution. Rather, they reduce soil acidity when they are adsorbed in place of H+, a
topic that we will further elaborate in the following section. Percent base saturation is a chemical
property of particular relevance to plant growth, because it is both a general measure of cation
nutrient availability and soil-buffering capacity; that is, the ability of soil to resist a change in pH
following the addition of H+. In general, soils that have a high proportion of exchangeable bases
have a high capacity to supply plants both with base cations and buffer acidic inputs.
SOIL ACIDITY
Soil pH is commonly used to quantify acidity and, by definition, is the negative log of the hydrogen
ion concentration in soil solution (pH = −log [H+]). In combination with adsorption and exchange
reactions, soil acidity substantially influences the supply of nutrients for plant growth. It does so
F I G U R E 9 . 7 The relationship between soil pH

and the availability of plant macronutrients and Fungi
micronutrients is shown. The influence of soil
pH on microbial activity within soil also is Bacteria and Actinobacteria
illustrated.
Ca, Mg
Fe, Mn, Zn, Cu, Co
Mo
4 5 6 7 8 9
by controlling the solubility of minerals composing soil. Figure 9.7 illustrates the availability of
plant nutrients along a pH gradient ranging from very acidic soils to those with high pH. Note that
the availability of most nutrients is greatest at neutral pH values (Figure 9.7). By influencing min-
eral solubility, soil pH also affects the weathering rate of parent material, the formation of clay
minerals, and soil development. Thus, the weathering process and soil acidification often go hand
in hand. The activity of soil microorganisms also is influenced by soil acidity as we see in
Figure 9.7.
There are several sources of H+ that contribute to the lowering of soil pH, the weathering of
parent material, and the removal of base cations. Perhaps the most important source of H+ in soil
results from the respiration of plant roots and soil microorganisms. Carbon dioxide produced dur-
ing respiration dissolves in soil solution and forms carbonic acid (H2CO3). The prolonged exposure
of soil minerals to this relatively weak acid results in their solubilization and the eventual removal
of base cations over sustained periods of weathering. Much stronger organic acids (e.g., fulvic and
humic acids) are produced as byproducts of the microbial decomposition of plant tissues. Plant
roots also exude organic acids that similarly act on soil minerals, solubilizing them over time. Over
extended time periods (i.e., 10 000–1 000 000 years), these sources of acidity facilitate the weathering
of soil minerals and lower pH.
The industrial activities of humans also have influenced soil chemistry. Oxides of nitrogen
(NOx) and sulfur (SO2), released during the burning of fossil fuels, can further oxidize in the
atmosphere to produce nitric (HNO3) and sulfuric (H2SO4) acids. The addition of these acids in
precipitation has raised concerns in eastern North America and Central Europe. Because the con-
stituents of “acidic deposition” are relatively strong acids, they have the potential to act on soil
212 Chapter 9 Soil
minerals in the same manner as the acids produced by the metabolism of plants and soil microor-
ganisms. Some soils in eastern North America, particularly those with coarse textures where gra-
nitic bedrock is shallow, are sensitive to acidic deposition because of their low cation-exchange
capacity, and low base saturation.
It is important to consider soil acidity, mineral weathering, and the removal of base cations
simultaneously, because they are co-occurring processes in soil. As mentioned earlier, base satura-
tion represents the ability of soil to buffer the input of H+ from chemical and biological sources.
Base cations buffer the soil reaction when they weather from soil minerals and replace H+ adsorbed
to exchange sites. The H+ so released initially enters soil solution; however, it is easily leached from
the soil resulting in a decrease in acidity. Clearly, over long time periods, the ability of soil minerals
to relinquish base cations to weathering can be exhausted. In such a situation, pH declines to the
point where the soil reaction is dominated by Al3+ and Al hydroxides. At a very low soil pH (<4.0),
Al exists as Al3+ and, along with H+, occupies the majority of cation-exchange sites. In soil solution,
Al3+ can react with water in the following manner:
Adsorbed Al3 Al3 in soil solution Al3 + H 2 O AlOH 2 + H .
In studying this equation, note that the reaction of H+ with bases in soil solution, or its leach-
ing from the soil profile, will shift the equilibrium to the right, producing a “new” H+ during
formation of AlOH2+. This mechanism constitutes a buffering system that maintains acidic soils at
a low pH. Furthermore, Al3+ released in the reaction is toxic to plants and can greatly restrict root
growth (Runge and Rode 1991). High concentrations of Al3+ in soil solution are known to reduce
root elongation, kill root meristems, and interrupt the functioning of the plasmalemma in above-
ground tissues (Foy et al. 1978).
Most forest soils range from extremely acid (pH 4.0) to slightly acid (pH 6.5). Where a
particular forest soil lies along this gradient is substantially influenced by organic-matter additions
(e.g., leaves, roots, twigs, reproductive structures, fine roots) from overstory trees and the acids pro-
duced during microbial decomposition. The general trend is for conifers such as pines, spruces,
hemlock, and Douglas-fir to increase surface-soil acidity (i.e., decrease pH) to a greater extent than
hardwoods or northern white-cedar.
An example of how individual trees influence soil acidity is provided by tulip tree and east-
ern hemlock in eastern Kentucky (Boettcher and Kalisz 1990). Although these trees co-occur on
the same soil parent material, the soil pH under tulip tree (pH 4.7) is consistently greater than that
beneath eastern hemlock (pH 4.0). In addition, quantities of the exchangeable bases Ca2+ and K+
also are lower beneath eastern hemlock. In eastern Washington, organic-matter additions from
western hemlock (pH 4.0) lower surface soil pH to a much greater extent than western red cedar
(pH 5.9) when both species occur on the same soil parent material. These examples illustrate that
differences in the litter biochemistry of forest trees can have a substantial influence on the chemical
properties of soils.
SOIL ORGANIC MATTER

Although organic matter composes a relatively small fraction of most forest soils (e.g., <1–15%), it
has a profound effect on a wide array of physical, chemical, and biological properties. As noted
earlier, soil organic matter (i.e., aboveground and belowground plant litter) contributes to aggregate
formation, which in turn influences the amount of soil water available to plants. It also functions
as a “storehouse” of plant nutrients, supplying most of the nitrogen used in the annual growth of
forest ecosystems as it is metabolized by soil microorganisms. Importantly, soil organic matter is
the substrate for the growth and maintenance of heterotrophic microbial populations in soil. It is
through the metabolic activities of these organisms that nitrogen and other plant nutrients bound
Soil Classification 213
in plant litter and soil organic matter are released into soil solution for plant use. Because the nutri-
ents so released via microbial metabolism can be re-assimilated by plants, soil organic matter rep-
resents an important “weigh station” in the cycling and storage of nutrients within forest
ecosystems (see Chapter 19).
The organic matter entering the soil originates from aboveground and belowground sources
of plant litter. Aboveground sources consist of leaves, reproductive structures, twigs, and tree
stems, whereas roots (fine and coarse) are the primary belowground source of litter. In most forests,
belowground litter from fine roots equals or exceeds aboveground litter production (i.e., leaves,
seeds, flowers). In general, plant litter contains approximately 15–60% cellulose, 10–30% hemicel-
lulose, 5–30% lignin, and 2–5% protein (Paul and Clark 1996). In soil, these compounds are metab-
olized by microorganisms, producing energy, CO2, H2O, and humus as end products. The
biochemical constituents of deal fungal and bacterial cells also are a component of humus, com-
prising approximately 25% of this material. Humus, which composes the Oa horizon, is a complex
and chemically resistant material that gives surface soils their dark color and unique chemical
properties. Due to its advanced state of decay, humus does not physically or chemically resemble
the plant material from which it originated. Humus also is chemically resistant to microbial deg-
radation and can remain in soil for periods of 100–3000 years (Paul and Clark 1996). The surface of
humus can have a net negative charge, resulting from the dissociation of H+ from hydroxyl (−OH),
carboxylic (−COOH), or phenolic (C6H11–OH) groups. At high pH values, the cation-exchange
capacity of humus (150–300 cmol kg−1) can exceed that of many silicate clays. As such, cation-or
anion-exchange reactions mediated by humus represent an important mechanism influencing
nutrient availability in soil. In some soils, approximately 50% of the total cation-exchange capacity
of soil can be attributed to humus. The relatively high cation-exchange capacities of the sandy soils
of mixed mineralogy in Table 9.4 directly result from their high organic-matter contents.
The organic-matter content of soil reflects a balance between the addition of organic matter
from plant production and its loss during microbial decomposition. Because forest harvesting can
alter both rates of litter input and loss through decomposition, it also has the potential to alter the
quantity of organic matter and associated plant nutrients stored in soil. In Chapter 19, we further
consider the impact of forest harvesting on soil organic-matter dynamics and the cycling of plant
nutrients.
The organic-matter content of soil exerts an important influence on the available water
content of soil. Soil organic matter holds relatively large quantities of water at field capacity, but its
permanent wilting point also is proportionally high, providing only small quantities of plant avail-
able water. However, organic-matter content is the primary factor influencing soil-aggregate
formation. In turn, soil aggregation influences the proportion of micropore and macropore spaces,
which directly controls the water-holding characteristics of soil. Consequently, soil organic matter
exerts its main influence on the water-holding characteristics of soil through its influence on soil
structure, rather than by how much water it can directly hold for plant use. Well-aggregated, fine-
textured soils with ample organic-matter contents (5–10%) generally hold large quantities of avail-
able water, making them good substrates for plant growth.
SOIL CLASSIFICATION
A taxonomic system of soil classification, referred to as the soil taxonomy, is widely used throughout
North America. It was developed by the soil survey staff of the United States Department of Agri-
culture to classify soils in regard to their potential for agricultural management (Soil Survey
Staff 1975). This system is based on measurable morphological characteristics present within a
particular soil profile. The primary advantage of this approach is that the soil profile itself, rather
than the soil-forming process, is classified. Soil taxonomy is modeled after the plant taxonomic
system, with categories ranging from order (broad grouping) to series (narrowest category).
214 Chapter 9 Soil
The main soil orders found beneath North American forests are entisols, inceptisols, spodo-
sols, alfisols, histosols, andisols, and ultisols; there are other soil orders, but they generally occur
beneath desert, grassland, or agricultural soils.
Entisols (recent soils) beneath forests are mineral soil without, or with only, the beginnings,
of horizon development. They often occur on talus slopes, floodplains, sand dunes, and where bed-
rock lies close to the land surface. Inceptisols (from Latin inceptum) are more weathered than
entisols and contain a weakly developed B horizon. These soils have a wide geographic distribu-
tion, and, with the exception of arid climates, can be found in most regions in North America.
Inceptisols are common forest soils in the Pacific Northwest, Rocky Mountains, and the eastern
United States. They often occur in well-drained, upland landscape positions, but also can be found
along river corridors.
Alfisols (from the chemical abbreviation for aluminum and iron) typically form in cool to
hot humid areas and are common under deciduous forests in the eastern United States. They are
characterized by shallow dark-surface horizons, medium to high base saturation, and the
accumulation of silicate clay in the B horizon (i.e., Bt horizon; t denotes the accumulation of ped-
ogenic clay). Alfisols are more weathered than inceptisols, but are less weathered than spodosols.
In cold and temperate climates, the process of leaching can give rise to the formation of
spodosols (from Greek spodos, wood ash). These soils are characterized by the presence of a
strongly developed E horizon and the accumulation of humus and oxides of Fe and Al in the B
horizon (Bh or Bhs horizons). Spodosols often form beneath boreal forests occurring on sandy par-
ent materials. In general, these soils are best developed beneath spruce and fir, species whose litter
generally acidify the surface soil, and are common in the northeastern United States, northern
Lake States region, and the Pacific Northwest. It should be noted that spodosols also can form
beneath coniferous forest in warm climates. In Florida, spodosols are often encountered in
landscape positions in which coniferous forests are seasonally flooded.
Andisols are derived from volcanic ejecta with the soil profile containing at least 60%
volcanic glass. These soils form any place on Earth where volcanic activity has delivered a blanket
of volcanic ash or related materials as surface deposits. Most often, these soils are poorly devel-
oped, but can be highly productive for forest growth as well as agriculture. In North America,
andisols are located in the Pacific Northwest beneath highly productive forest often dominated by
Douglas-fir.
Histosols form in poorly drained landscape positions and are characterized by a high
organic-matter content (≥20%). These soils can form anywhere the land surface is continually sat-
urated with water, and thus occur in all climates and have a global distribution. Forest vegetation
occurring on these organically derived soils includes: black spruce bogs in the northern Lake
States, black ash–red maple swamps in the Northeastern United States, and the pocosin and
cypress swamps of the Southeastern United States. Large expanses of forested histosols also can be
found in Scandinavia, Siberia, and Canada.
Ultisols (form the Latin word ultimus) are a common forest soil on old land surfaces in
warm, humid climates, such as those of the southeastern United States. These soils are character-
ized by an accumulation of silicate clay in the B horizon. However, ultisols are distinguished from
alfisols by a low base saturation—the result of more intense weathering. These soils are widely
distributed in the eastern US, extending southward from Maryland to Florida and westward from
the East Coast to the Mississippi River Valley. They also occur in portions of the Pacific Northwest
and eastern California and can be found on old land surfaces in Australia, Africa, India, southern
China, and southern Brazil.
Oxisols (from the French oxide and the Latin word for soil, solum) support forest vegetation
in the tropical and sub-tropical regions of Central and South America, Southeast Asia, and Africa.
The subsoil of these highly weathered soils contains an accumulation of kaolinite and oxides of Al
and Fe. The old land surfaces on which these soils occur, in combination with the intense
weathering of humid tropical climates, can give rise to profiles exceeding 15 m in depth.
Landform, Soil, and Forest Vegetation: Landscape Relationships 215
Forest and agricultural ecosystems differ in ways that limit the use of the soil taxonomy to
classify forested landscapes. In the western United States, for example, a wide range of forest hab-
itat types can be found on the same taxonomic unit of soil (Neiman 1988), making it difficult to use
soil classification to predict the occurrence of forest vegetation as well as its productivity. The pri-
mary reason for such a disparity is that forest vegetation reflects a myriad of interacting physical
and ecological factors such as physiography, harvesting frequency and intensity, and prior land
uses that are not considered by soil taxonomy. Nevertheless, soil factors that reflect moisture and
nutrient regimes, such as texture, aggregation, and coarse-fragment content, are often related to
the occurrence of some forest types (Neiman 1988). Because the relationship between forest com-
munities and the soil developing beneath them is multifactorial and dynamic, it is likely that any
single-factor classification, such as the soil taxonomy, will be of limited use in predicting the distri-
bution and growth of forest ecosystems. In Chapter 13, we further discuss the limitation of single-
factor systems for classifying forested landscapes.
LANDFORM, SOIL, AND FOREST VEGETATION: LANDSCAPE RELATIONSHIPS

Changes in topography, vegetation, and parent material across a landscape give rise to marked
differences in soil formation, even within a relatively small geographic area. In Michigan’s Upper
Peninsula, for example, forest ecosystems dominated by white pine, sugar maple, and red oak can
occur in well-drained, upland landscape positions where a thin blanket of glacial drift overlies bed-
rock (Figure 9.8). Soils forming in these landscape positions are inceptisols, characterized by a
shallow profile and the minimal development of a B horizon. The deeper deposits of glacial drift
White Pine
Red Oak White Pine
E. Hemlock
N
Red Oak
80
Sugar
70 Maple
Sugar Maple
Relative Elevation (m)
60 Oe
cm Basswood
E
50 Bh
20
40 cm Oe
40 E N. White-cedar
30 20 Bh White spruce
60
2R Oe
20 40 cm A
80
60 E
10 2C 20
100
Bs cm Oe
0 80 40 A
120
20 cm Oe
100 60 A
140 cm
3R 40 2Bs 20
120 80 Oa
2C Bs 20
Inceptisol 140 60 40
100 Water
2C 40 tabl e
80 60
120 60
Inceptisol
100 80
140 4C 2C 80
120 100
100 2C
Spodosol
120
120
140
140
— Inceptisol—
Histosol
F I G U R E 9 . 8 A physiographic cross section in the Upper Peninsula of Michigan illustrating the influence of
topography and forest vegetation on soil profile development. Source: Pregitzer et al. (1983)/ John Wiley &
Sons. Reprinted with the permission of the Soil Science Society of America, Segoe, Wisconsin, USA.
216 Chapter 9 Soil
downslope give rise to spodosols, which occur beneath a canopy of sugar maple and basswood;
these soils have well developed E and Bh horizons. In lowerslope positions also dominated by
sugar maple, inceptisols form in relatively recent deposits of more recent colluvium. The formation
of an inceptisol in this landscape position is related to the relatively short duration over which the
parent material has weathered. Swamp forests dominated by black ash and northern white-cedar
occur in the poorly drained bottomslope positions where the accumulation of organic matter has
led to the formation of histosols (Figure 9.8). The patterns of soil formation described in the earlier
text are repeatable features in the landscape, occurring in other locations with a similar set of soil-
forming factors (i.e., parent material, vegetation, time).
In the southeastern United States, topography exerts a similar influence on the soil devel-
oping beneath the flatwood vegetation of the lower coastal plain (Figure 9.9). In this region,
relatively small elevational differences differentiate well-drained from poorly drained landscape
positions. This land surface is relatively old (100 000 years before present) compared to the relatively
recent (approximately 8000 years before present) deposition of glacial materials in Michigan’s
Upper Peninsula. As a consequence, the parent material giving rise to this soil has been affected by
climate, biota, and topography for a longer duration.
In uplands dominated by longleaf pine, relatively dry conditions give rise to inceptisols—
poorly developed soils with minimal horizon formation. The warm, humid climate of this region,
in combination with sandy parent material and coniferous vegetation, gives rise to spodosols in
somewhat poorly drained landscape positions with a fluctuating water table. Histosols further
form in very poorly drained landscape position in the Southeast United States, similar to the
landscape distribution of these soils in other regions.
Loblolly Pine
Gallberry Slash pine
Slash Pine Palmetto Hardwoods
Cypress
Wiregrass Slash
Turkey Longleaf
oak Pine pine
Clay
accumulation Organic matter accumulation
Well to moderately well drained Somewhat poorly drained Poorly to very poorly drained
0
15 Yellow-
A
ish Leached
brown E zone
Light
75 Organic gray
matter Bh s
Gray
to
pale Finer
yellow
105
Sands Clay Clay

Inceptisol Alfisol Spodosols Alfisol Histosols
F I G U R E 9 . 9 The relationship among topography, vegetation, and soil development in the coastal plain
flatwoods of the southeastern US. Organic-matter accumulation in the surface horizon dramatically
increases as one moves from the well-drained upslope positions to the poorly drained bottomslope positions.
In this region of the United States, relatively small changes in elevation elicit large changes in soil drainage,
profile development, and overstory composition. Source: Adapted from Pritchett and Smith 1970. Reprinted
with permission from Tree Growth and Forest Soils, © 1970 by Oregon State University Press.
Landform, Soil, and Forest Vegetation: Landscape Relationships 217
SUGGESTED
READINGS
Binkley, D. (1995). The influence of tree species on forest Paul, E.A. and Clark, F.E. (1996). Soil Microbiology and
soils: processes and patterns. In: Proceedings of the Biochemistry, 2e. New York: Academic Press 340 pp.
Trees and Soils Workshop, Agronomy Society of New Sanchez, P.A. (1976). Properties and Management of
Zealand Special Publication No. 10 (ed. D.J. Mead Tropical Soils. New York: Wiley 618 pp.
and I.S. Cornforth), 1–33. Canterbury, New Zealand:
Weil, R.R. and N.C. Brady 1996. Nature and Properties of
Lincoln University Press.
Soil, 11 Prentice-Hall, New Jersey. 740 pp.
Fire CHAPTER 10
I n North America, fire has affected virtually all of the upland forests in the South, the Lake
States and adjacent Canada, the West, and most of those in the Northeast, Appalachian
Mountains, and central states have been burned more or less frequently. In the boreal forests
of Alaska and Canada, fire has been a powerful natural factor affecting vegetation and wildlife.
Even wetlands such as swamps (Cypert 1973; Ewel 1990), bogs, and marshes have burned,
although less frequently, and their vegetation affected. Fire was extensively used by indigenous
people prior to European colonization. Books by Kozlowski and Ahlgren (1974), Wein and
MacLean (1983), Johnson (1992), Agee (1993), Whelan (1995), Bond and van Wilgen (1996),
Johnson and Miyanishi (2001), and Pyne (1982, 2019) provide an entry into the voluminous
literature on fire.
Fire has always been a critical process shaping the evolution of species and the functioning
of ecosystems in which they reside. Prehistoric fire and the remarkable interactions of fire with
humans and wildlife are considered by Schüle (1990). Fire regimes, which are defined by the
frequency, intensity, severity, extent, and time of occurrence of fires, are characteristic of different
regional and local landscape ecosystems. Fire is a principal influence on plant traits and life cycles
as well as on diversity and ecosystem processes such as carbon, nutrient, and water cycling;
biomass accumulation; and succession. Fire plays many major roles in landscape ecosystems
around the world (Wright and Heinselman 1973; Swanson 1981). It influences:
Geomorphic and hydrologic processes of hillslopes and stream channels.

Physical and chemical properties of soils.
Nutrient loss.
Biomass accumulation.
Genetic adaptations of plants.
Plant composition and diversity, mortality, regeneration, growth, and succession.
Wildlife habitat and wildlife population dynamics.
Presence and abundance of forest insects, parasites, and fungi.
In this chapter, we consider fire as a physical site factor, examining its effects on forest species
and forest site quality. The role of fire as a disturbance factor in forest ecosystems and their com-
munities is considered in Chapter 16, and its interactions with climate change are discussed in
Chapter 20.
219
220 Chapter 10 Fire
FIRE AND THE FOREST TREE

CAUSES
Fossil charcoal, which represents evidence of past natural fires, has been identified in the
Carboniferous coal deposits of 400 million years ago and in the Tertiary deposits of brown coal
(Harris 1958; Komarek 1973). Meteorites were a rare ignition source of fires, and falling igneous
rock from volcanic eruptions undoubtedly caused local fires then as it does today. Primarily, how-
ever, fires were caused by lightning, which today causes about 50 000 wildland fires worldwide
each year (Taylor 1974). Notably, this estimate is less than 1% of the estimated 182 million cloud-
to-ground discharges that occur in the forests and grasslands of the world each year. Between 70
and 100 lightning flashes are estimated to occur every second worldwide (Pearce 1997), but not all
strike the ground. In the United States, about 10 000 lightning-caused fires occur each year, and
about 80% of these are in the Rocky Mountains and Pacific Coast states. When fuel and climatic
conditions are conducive to ignition by lightning, a single lightning storm may start many small
fires in these regions.
Humans have been the most significant cause of fires throughout much of the modern world.
Prior to European colonization, indigenous people set many fires, and Europeans later followed
suit. The high intensity and frequency of such fires, which are often associated with logging and
land clearing, have in many cases altered the character of forests and affected site quality.
FIRE REGIME
Studies of fire history have helped to define the kind(s) of fire and the prominent immediate effects
of fire that characterize an area. A fire regime is typically characterized by type, frequency, inten-
sity, severity, size, and seasonality, with fire type, frequency, and severity the most important. Fire
type includes ground, surface, and crown fires. Fire frequency refers to the recurrence of fire
in a given area over time and may be expressed in a number of ways (Agee 1993). Return interval,
or the average number of years between successive fires, may be expressed for a given point (for
example, a single fire-scarred tree or a small group of trees) or for an area. Frequency is also ex-
pressed as fire rotation or fire cycle, which is the length of time required to burn over an area equal
to that under consideration. These approaches and their computation are described and contrasted
in detail by Agee (1993).
Fire intensity refers to the amount of heat released by the fire, often estimated by the length
of the flames, whereas severity expresses the effect of fire on soil or vegetation (seed bank,
mortality of plants). Specifically, severity is usually expressed as a function of tree mortality or soil
damage. For example, an intense spring fire, when the soil is moist, may be low in severity, whereas
a low-intensity summer fire during drought conditions may be severe in its effects on soil prop-
erties. Intensity is typically estimated in kilowatts per meter length of fire front (kW m−1) and in
flame length (m) along the fire front (Johnson 1992; Agee 1993). The seasonality of fire may have
differential effects on vegetation and soil. For example, in the southeastern coastal plain of the
United States, the proper proportion of winter fires to fire-free years is critical to the perpetuation
of pure longleaf pine forest because pine regeneration is prevented by annual burning and summer
fires (Marks and Harcombe 1981).
Three primary fire regimes are recognized (typically based on severity, or whether fires are
nonlethal to the dominant aboveground vegetation or cause stand replacement): nonlethal under-
story fires (including frequent or infrequent surface fires), stand-replacing fires (short- or long-
frequency crown fires), and mixed-severity fires. The latter regime applies to combinations of
understory and stand-replacing fires that may occur in two ways: (i) a variable fire regime of fre-
quent, low-intensity surface fires typically followed by an infrequent, stand-replacing fire, or (ii) a
mixed fire regime of individual fires alternating between nonlethal understory burning and
Fire and the Forest Tree 221
stand-replacing fires, creating a fine-scale pattern of young and older trees. The second type of
mixed-severity regime is more often discussed in the literature.
It is important to understand that nearly all fires result in a mosaic of severities, ranging
from unburned or hardly burned to completely scorched, mass mortality events within the same
fire perimeter. As a result, most burned areas are a mosaic from unburned survivors, to low-severity
areas with few trees killed, to high-severity areas with many or all trees killed. This variation is
often ignored for simplicity in describing how fires affect forests, but this relative proportion of
severities is often what implicitly defines a fire regime.
Fire Types, Frequency, and Severity Van Wagner (1983) identifies five main types of
fire in the northern forest based on their physical behavior, including two kinds of surface and
crown fires:
Ground fires that smolder in deep organic layers (less than 10 kW m−1).
Surface backfires that burn against the wind (100–800 kW m−1).
Surface headfires that burn with the wind (200–15 000 kW m−1).
Crown fires advancing as a single front (8000–40 000 kW m−1).
Crown fires that include high-intensity spotting fires (up to 150 000 kW m−1).
Surface fire is the most common type of fire, which burns over the forest floor, consuming litter,
killing aboveground parts of herbaceous plants, shrubs, and small trees, and typically scorching
the bases and crowns of larger trees. Mortality of shrubs and trees increases with greater fuel
accumulation on the surface. Surface fires are very sensitive to wind speed; surface headfires may
attain quite high intensities in brush in leafless hardwood stands and in open forests where trees
are sparse, or crowns are high above the ground. In addition to fire intensity, the amount of tree
mortality depends on the species, the age of the tree, and rooting habit. Young pines may succumb
to a surface fire, whereas older individuals of the same species survive because their thicker bark
protects the cambium from heat damage and their crown height above the flames protects the can-
opy. A shallow rooting habit, whether due to the inherent nature of the species or site conditions
(wetland or bedrock), increases susceptibility to fire injury compared to that of the deep rooting
habit typical of upland species such as oaks and hickories.
Surface fires tend to kill young trees of all species (though often just the aboveground
portion) and most trees of less fire-resistant species of all sizes. However, pole-size to mature trees
of fire-resistant species survive light surface fires in varying proportions. Survival in a surface fire
for most fire-resistant tree species is not typically dictated by damage to the stem cambium, but by
their susceptibility to root injury and to crown scorch by hot gases rising above the flames. For
example, Van Wagner (1970) reported that a light surface fire will leave mature crowns of red pine
undamaged, whereas a hot surface fire will kill a red pine stand just as surely as a crown fire
10 times as intense. Observations of fire-damaged red and white pines suggest that the trees will
die if more than 75% of the crown is killed.
Fires sweeping the forest floor may generate ground fires, which burn in thick accumula-
tions of organic matter, such as peat, that overlies mineral soil. They are flameless when they burn
below the surface, and they kill most plants with roots growing in organic matter. Ground fires
burn slowly and generate very high temperatures. In moist organic matter, heat from the fire
dries out material adjacent to the burning zone and perpetuates a zone of combustible fuel. Ground
fires tend to be persistent and serve as reignition sources for surface fires. Recent examples from
near-arctic regions have documented summer ground fires burning in peatlands that smolder
beneath the snow throughout the winter! Such fires, sometimes called “zombie fires” by wildland
firefighters, may reignite aboveground vegetation early in the spring.
222 Chapter 10 Fire
Surface fires are often closely associated with crown fires. Surface fires, fueled by
a ccumulations of organic matter and whipped by winds, may scorch and ignite crowns of trees.
Crown fires travel from crown to crown, often in dense even-aged stands, killing most trees in their
path. Notably, large, intense crown fires may leave unburned strips (sometimes called “stringers”;
Kashian et al. 2012) due to powerful, downward air currents (Simard et al. 1983) or other mecha-
nisms. Where the forest is patchy or consists of small groups of trees such as in much of the
dry-climate, ponderosa pine forests of western North America, some individuals or patches of
trees may torch or briefly carry a crown fire, but an extensive crown fire is highly unlikely. Conifers
are more susceptible to crown fires because of the high flammability of their foliage and the greater
likelihood of their occurrence in pure stands compared to broad-leaved species. Sparks and burning
embers may be carried ahead of the main fire to start spot fires, or new surface fires often far away
from the site of the main crown fire. Notably, especially in mountainous areas, surface fires are
common, and their frequency is related to climatic factors of temperature and precipitation associ-
ated with elevation. However, these surface fires can lead to severe crown fires given the right
climatic and fuel conditions.
Fire frequency, intensity, severity, and burning pattern are primarily controlled by climate,
fuel accumulation and flammability, soil water, and especially topography. Fire frequency was
probably greatest in grasslands prior to European colonization, where burns every 2–3 years were
common (Wells 1970). In dry ponderosa pine forests of western North America, the average inter-
val between fires varied from 2 to 18 years (Weaver 1974; Dieterich and Swetnam 1984; Savage and
Swetnam 1990). Average intervals vary because of the method of determination (Agee 1993) and
ecosystem type; pine forests of drier low-elevation ecosystems burned more frequently than those
of moist slopes and higher elevations. For example, Arno (1976) reported fire frequency over an
elevational gradient (1150–2600 m) for three watersheds in the Bitterroot Mountains of western
Montana for the period 1735–1900. Low-elevation forests of ponderosa pine and Douglas-fir
burned at about 9-year intervals (range 2–20 years). Intervals between burns increased with
increasing elevation, approaching 35 years at high elevations (range 2–78 years).
Intense surface fires may damage the cambium and scar the tree. When an individual tree is
scarred multiple times, the fire scar records the actual fire frequency of a single place (Spurr 1954).
One ponderosa pine in Arno’s study was scarred by 21 fires from 1659 to 1915, an average interval
of 13 years (Figure 10.1). Nearby trees were scarred by additional fires, demonstrating that fires do
not always burn hot enough to cause injury, particularly when they occur so frequently that only
light accumulations of fuel have occurred. Similarly, Dieterich and Swetnam (1984) studied a
small, suppressed ponderosa pine in Arizona containing 42 fire scars over the 178-year period from
1722 to 1900. A detailed historical fire chronology of the 97-ha area showed that whereas the mean
fire interval for the tree was 4 years, the interval for the area was 2 years, that is, not every fire was
recorded by this tree. Thus, records of fire frequency may reliably indicate the frequency of severe
fires, but may underestimate the frequency of all fires. Light to moderate surface fires, although
often undetected, are nevertheless important in regulating seedling composition and distribution
and reducing the concentration of fuels on the forest floor.
In contrast to frequent fires in dry, low-elevation, submontane ponderosa pine and mixed
conifer forests, fires occur infrequently but regularly in high-elevation subalpine ecosystems of the
Pacific Northwest and northern Rocky Mountains (Agee 1993). Dominant trees are firs (often sub-
alpine fir), spruces (often Engelmann spruce), mountain hemlock, and whitebark pine. Fire fre-
quency is spatially heterogeneous because of local variation in ecosystem conditions and fuel
accumulation. For example, fire-return intervals varied from over 1500 years for ecosystems domi-
nated by mountain hemlock to differing intervals for whitebark pine ecosystems: from 50 to
300 years (Arno 1980) to a low of 29 years (Morgan and Bunting 1990).
In the Sierra Nevada Mountains of California, a frequency of 9 years between fires was
found at one locality in the giant sequoia–mixed conifer forest during the period 1705–1873
F I G U R E 1 0 . 1 Multiple fire-scar cross section of a ponderosa pine
Photo by Peter M. Brown, Rocky Mountain Tree-Ring Research

showing 11 fire scars from 1623 through 1996.
(Kilgore 1973), comparable to the overall 8-year frequency of fire reported for ponderosa pine
forests (Weaver 1951). As is typical of many other western forests, moister east and north slopes
do not burn as readily as drier south and west slopes, but they may burn more intensely than
those that burn more frequently. A similar relationship develops along elevational gradients from
warm and dry low elevations to the cool and moist slopes of high altitudes. Studies of redwood in
California also illustrate that fire-return intervals are site-specific and not species-specific.
Fire-return intervals for moist coastal redwood sites at the northern part of the range may be as
long as 500–600 years (Viers 1980), but intervals of 33–50 years (Viers 1980) and 20–29 years are
reported (Jacobs et al. 1985; Finney and Martin 1989) on drier, interior and more southerly sites.
Similar variation in return interval by specific ecosystem conditions is also reported for
subalpine forests.
224 Chapter 10 Fire
In conifer forests of the Boundary Waters Canoe Area of northern Minnesota, detectable
fires were relatively frequent (Heinselman 1973, 1981), burning at approximately 4-year intervals
during the period 1727–1886. European colonization activities increased fire frequency, and the
interval dropped to 2 years. Major fires that burned over large areas occurred at longer intervals of
21–28 years.
The emerging pattern from fire history studies is one of cyclic occurrence that is determined
by the multiple and interrelated physical and biotic features of site-specific ecosystems. Fire
regimes of regional and local ecosystems are affected by climate, physiography, vegetation type
(affecting fuel accumulation and its flammability), and human activities. Variations in these factors
have resulted in fires of different intensities and severities—from frequent, light surface fires,
which kill few trees but reduce fuel accumulations and create patterns of seedling and sapling
distribution, to infrequent, high-severity crown fires that kill most trees and regenerate entire
stands. This ecosystem-dependent range of fire frequency, intensity, and severity was instrumental
in the evolution of the life span and species characteristics and, more generally, the kind of vege-
tation present in each region (Chapter 16). There exist obvious spatial patterns of regional forest
composition throughout North America that have resulted from variations in fire frequency and
severity (Figure 10.2).
FIRE ADAPTATIONS AND KEY CHARACTERISTICS

Certain key characteristics and adaptive traits of forest trees and shrubs appear to have developed
in response to fire. Many individual characteristics are typically cited, such as thick insulative bark
and sprouting ability, but probably few of these are genetic adaptations that are exclusively
fire-determined. Seasonally hot, dry sites favor fire and support species that not only are likely to
have fire adaptations but also are physiologically adapted to live in such severe environments.
Selective forces of the total environment affect the plant’s evolutionary response. This point is con-
firmed by the discovery that many species of tropical rain forests persist in human-modified eco-
systems now characterized by frequent fire. Such “fortuitous adaptations” allow species to survive
fires or quickly establish following fire regardless of evolutionary derivation (Kauffman and
Uhl 1990). Such adaptations include thick bark, vegetative sprouting, and seeds stored in the soil.
Therefore, our emphasis in this section is on those traits of woody species that are suited to survive
and reproduce in ecosystems characterized by fire, that is, fire-dependent ecosystems.
The characteristics probably most directly elicited by fire are closed cones (cone serotiny,
i.e., delayed or heat-opening of cones); thick fire-resistant bark; buried or protected buds and
subsequent sprouting; early flowering, fruiting, and dispersal; and the “grass stage” in longleaf and
other pines. In addition, the relative life span of many forest species has been shaped by the fre-
quency of stand-replacing fires (Gill et al. 1981). As a result, moderately long-lived, fire-dependent
species such as Douglas-fir and eastern white pine are able to compete with long-lived, late-
successional associates such as western hemlock and western redcedar, or sugar maple and
American beech on mesic sites that burn infrequently but within their life span.
The key characteristics of forest species enabling their persistence in fire-prone ecosystems
may be grouped into four general categories to emphasize major life history features related to fire
and site: (i) avoid fire damage, (ii) recover following fire damage, (iii) quickly colonize sites after
fire, and (iv) promote or facilitate fire. Species especially fire-resistant or susceptible typically
exhibit the extreme expression of these characteristics. Examples of each category are cited in the
following text:
(i) Avoiding Fire Damage:

Thick, insulated bark (many pines and oaks, western larch, giant sequoia, and redwood).
Buried buds protected by soil (aspens, sweetgum, sumacs, and many shrubs).
Coniferous Forest
and Woodland
Tundra • Infrequent, severe
• Very infrequent surface and crown
Marine Coniferous surface • Jack pine frequent,
Forest severe surface
• Infrequent or crown
crown
• More frequent Mixed Forest
severe surface • White and red pine
frequent surface
Mountain Vegetation • Jack pine frequent,
• see below severe surface
• Northern hardwoods
severe surface
Cool Desert
• Moderately frequent
Deciduous Forest
surface
• Probably infrequent
Coniferous Forest crown or severe
Woodlands, Shrubs,
• Frequent surface surface
and Grasslands
• Woodlands probably
frequent surface
• Chaparral frequent Warm Desert
severe surface • Very infrequent
• Grassland frequent Grasslands and Savannas
surface
surface • Frequent surface
Mountain Vegetation
Alpine Very infrequent surface
High-elevation forests Moderately frequent surface and/or

infrequent crown
Midelevation forests Moderately frequent surface and/or more

frequent crown or severe surface
Low-elevation forests Frequent to moderately frequent surface:

few infrequent crown
F I G U R E 1 0 . 2 Pre-European colonization fire regimes of broad vegetation types. Based on broad ecoregions
of Bailey (1995). Source: Vale (1982) / with permission of American Association of Geographers.
“Grass stage”—longleaf pine of the United States (see Chapter 16), and several pines of
Mexico and the Caribbean region (Mirov 1967, p. 417), and chir and Merkus pines of Asia.
Deep rooting—taproot in young plants (upland oaks and hickories).
Rapid juvenile growth—crown grows above the surface fire zone and heat-resistant bark
is formed (pines).
Basal crook—dormant buds on the lowermost stem are protected from fire by a crook of the
stem that brings the buds in contact with mineral soil (pitch, shortleaf, pond, and other
hard pines).
Branch habit and self-pruning ability—rapid self-pruning of branches decreases the
likelihood of crown fire (larches, pines, and Douglas-fir in closed stand conditions), whereas
226 Chapter 10 Fire
low or drooping branching habit and poor self-pruning ability increase the likelihood of
crown fire (true firs, hemlocks, and spruces).
Stand habit—open-grown stands decrease the probability of crown fire and also afford less
fuel (western larch, ponderosa pine, and longleaf pine).
Fire-resistant live foliage—hardwoods are much less flammable than conifers; among
conifers, larches have less flammable foliage than pines, Douglas-fir, spruces, and true firs.
Rapid foliage decomposition—slows fuel accumulation and reduces the opportunity for
fire ignition and spread (ashes, elms, sugar maple, and basswood).
The differential resistance of tree species of the northern Rocky Mountains to fire damage
and mortality was shown in a classic study by Flint (1930). He determined species’ relative
resistance using characteristics such as bark thickness of old trees, rooting habit, resin in old bark,
branch and stand habit, relative flammability of foliage, and abundance of lichens on stems.
Western larch, ponderosa pine, and Douglas-fir are very resistant, whereas western hemlock and
subalpine fir are low to very low in resistance. Although species vary greatly in resistance, severe
fires largely erase differences in resistance (Wellner 1970).
(ii) Recovering from Fire Damage:

Sprouting—new shoots arise from living stems or roots after fire damage:
Stems:
Rhizomes (many shrubs and herbaceous plants).
Root collar or crown (oaks, paper birch, black cherry, redwood, and chaparral species)—
basal sprouting is rare in conifers, but in hard pines (shortleaf, pitch, Monterey, etc.) sprouts
may occasionally arise from dormant buds formed in the axils of primary needles at the base
of the stem (Stone and Stone 1954).
Lignotubers (burls)—relatively large stem swellings at or below ground incorporate many
clusters of dormant buds, common in redwood and nearly all eucalypt species and
Mediterranean shrubs (Gill et al. 1981; James 1984).
Bole—dormant buds along the bole initiate new shoots after the crown is scorched or killed
(pitch pine, redwood, big-cone Douglas-fir, and eucalypt species).
Roots:
Important in some trees and shrubs (aspens, rock elm, sweet gum, sassafras, sumacs, and
Acacia species).
Deep rooting—taproot or sinker roots provide food reserves for rapid regeneration of new
shoots (upland oaks and hickories).
(iii) Colonizing Burned-Over Areas:

Early flowering and seed production—enable a species to reproduce itself sexually on a
site where fires regularly or occasionally occur at short to moderate intervals (jack pine,
lodgepole pine, pitch pine, and Xanthorrhoea species of Australia (Gill et al. 1981).
Light, wind-borne seeds—facilitate widespread dissemination of seeds, some of which
may reach burned sites (birches, aspens, willows, hemlocks, and larches).
On-plant seed storage and fire-triggered dehiscence of fruits and cones—closed

cones that contain viable seeds persisting on branches are typical of many species of the pine
family (including jack, lodgepole, pitch, pond, sand, table mountain, knobcone, and
Monterey pines) and cypress species of arid sites of southern California (Vogl et al. 1988).
Black spruce exhibits semi-serotinous cones; the cone scales open upon drying and close
when wet. Seed dispersal is therefore periodic, sometimes occurring over a 2-year period.
Shrubs of the Australian genus Banksia retain follicles on their branches that remain closed
unless they have been in contact with flames (Gill et al. 1981). Many other taxa of the
Australian flora also have this wood-fruit habit. Also, some eucalypt species of southeastern
Australia retain their fruit 4 or more years with seed release following fire.
No dormancy—certain species of fire-prone ecosystems have seeds that do not enter true
dormancy. They germinate readily at any favorable time following post-fire rains (jack,
lodgepole, longleaf, and pitch pines).
Heat-induced germination—hard-coated seeds of certain species of Arctostaphylos, Cea-
nothus, and Rhus tend to lie dormant in the soil (in redstem ceanothus up to 150 years
(Mutch 1976)). Germination is favored by fire, which cracks the seed coat and generates the
heat needed to stimulate germination.
(iv) Promoting Fire Occurrence:

Traits that increase the likelihood of fire and thereby favor regeneration of the species
over others:
Flammable foliage and bark—needles of pines, many other conifers, and the foliage of
some angiosperms are highly flammable, decompose slowly, and form a ready fuel source for
surface fires (Mount 1964; Mutch 1970; Gill 1977). The bark of certain species, such as paper
birch and some eucalyptus species, is decay-resistant yet highly flammable, creating an
excellent and persistent fuel source on the forest floor.
Retention of foliage near the ground—promotes crown fires (understory conifers, espe-
cially firs, northern white-cedar and western red cedar; juvenile oaks).
Short stature—brings foliage close to the ground where a surface fire may spread to the
crown (young or slow growing trees with flammable foliage, jack pine).
Retention of dead, lower branches—serves as “ladder” fuel promoting fire spread to the
crown (narrowleaf cottonwood, jack pine, and some oaks).
STRATEGIES OF SPECIES PERSISTENCE

Species persist in fire-prone ecosystems using one or more of these and other adaptations, regen-
erating themselves either by seed or by vegetative means. Rowe (1983) distinguished five kinds of
“strategies,” based on an understanding of fire adaptations, life cycle, and mode of persistence of
plants in ecosystems of northern and boreal forests. Plants reproducing primarily by seeds are
termed invaders, evaders, and avoiders. Plants reproducing primarily by horizontal and vertical
extension of vegetative parts are the resisters and endurers. These groups were conceived for
boreal conditions, but have wide application elsewhere, as Agee (1993) has shown for the species
of western North America.
Invaders are the early arrivers that are successful because they produce large amounts of
short-lived, wind-disseminated seeds. Once established, they flower and fruit profusely or spread
vegetatively. They are typically the shade-intolerant, pioneering, “fireweed” plants that establish
228 Chapter 10 Fire
vigorously after fire regardless of its frequency or intensity. Many herbs exemplify this group;
woody invaders include paper birch, aspens, and willows.
Evaders store seeds in the canopy, humus, or mineral soil, placing seeds to evade high tem-
peratures, followed by rapid seed germination and establishment. The plants themselves may be
killed by fire, but their seeds are protected. They include short-lived ephemerals as well as intoler-
ant to shade-tolerant species that persist into later successional stages (Rowe 1983). Woody evaders
of the boreal forest include jack and lodgepole pines and black spruce.
Avoiders arrive late in succession and prosper where fire cycles are long. They essentially
lack direct adaptations to fire and are often said to occupy unburned areas or ecosystems relatively
undisturbed by fire. They are typically mesophytic and shade-tolerant. Included in the group are
many herbs and shade-requiring mosses and lichens; balsam fir and white spruce are woody exam-
ples in northern and boreal forests. Sugar maple, basswood, beech, and southern magnolia are
examples of deciduous avoiders of eastern and southern forests.
Resisters are the few intolerant species whose adult stages can survive low-severity fires. They
continue growing vegetatively in spite of fire (whereas evaders are likely to be killed by fire). Boreal
examples include jack and lodgepole pines and cottongrass, a circumpolar sedge, whose dense
tussock form resists fire. Giant sequoia; ponderosa, red, shortleaf, loblolly, and longleaf pines;
western larch; and many thick-barked oaks are also resisters. This group is not mutually exclusive
with evaders.
Endurers are composed of the large and diverse group of species, both shade-tolerant and
shade-intolerant, that re-sprout following fire. They regenerate from roots, root collars, rhizomes,
and other belowground organs whose vertical position in the insulating humus and mineral soil
strongly determines these species’ persistence and abundance. Aspens, birches, and many shrubs,
including ericaceous bog species, are woody representatives of the group.
These groups of species are closely related to the site-specific ecosystem where they occur
and its particular fire cycle and fire behavior. Ecosystems characterized by high-severity fires favor
invaders, evaders, and endurers; those with low-severity surface fires favor resisters as well as
evaders and endurers. Species may belong to more than one group. For example, jack and lodgepole
pines may be invaders, evaders, and resisters because of their multiple adaptations in fire-prone
ecosystems.
Closed-Cone Pines Many species’ life-history traits are strongly related to fire frequency
and severity in fire-prone ecosystems. Such is the case with the varied group of pines bearing serot-
inous cones: jack and lodgepole pines of northern and western North America; pitch pine and
pond pine of the Coastal Plain of the eastern United States (Lutz 1934; Ledig and Fryer 1972;
Givnish 1981; Ewel 1990); table mountain pine of the southern Appalachians (Zobel 1969); and
knobcone, Bishop, and Monterey pines of California (Vogl 1973). Closed cones may persist on the
tree for many decades and still bear viable seeds. In lodgepole pine, millions of seeds per hectare
may be stored in serotinous cones. In such cones, the cone scales are bonded by resin that melts
above 45 °C (113 °F), typically due to heat generated from fire. The scales open as they dry, and the
seeds are disseminated. Amazingly, seeds of jack and lodgepole pines may remain viable even
when cones in the tree crowns are exposed to temperatures of 900 °C (1652 °F) for 30 seconds.
Both jack and lodgepole pines may bear various proportions of serotinous cones and those
that open readily (nonserotinous) under normal climatic conditions. These proportions vary both
within an individual tree and among trees in a population. The range of closed cones per tree in
jack pine is from 0 to 100%; the average for the species is about 78% (Schoenike 1976). Open-cone
trees are found in the southern portion of the range, whereas trees with serotinous cones predom-
inate in the boreal and western range (Figure 10.3). In the south, jack pine occurs in mixed stands
with oaks and red pine, and fires are more often light surface fires than in the boreal forest. In the
boreal forest, jack pine is regenerated by crown fires that open the cones and prepare the seedbed,
Legend
0–50
51–75
76–95
96–100
FIGURE 10.3 Variation in the percentage of closed cones per tree in jack pine. Source: Based on Schoe-
nike (1976).
thereby perpetuating the predominance of the serotinous habit. Similarly, in lodgepole pine, stands
originating after a severe burn typically produce trees with a high percentage of closed cones (Muir
and Lotan 1985).
More recent studies of geographic variability in serotiny have examined the proportion of
trees in a population that bears serotinous versus open cones. For example, the percentage of
serotinous trees for jack pine in northwest Wisconsin was 83% for populations where forests were
characterized by high-severity crown fires, but only 9% where forests burned more frequently by
low-severity surface fires (Radeloff et al. 2004). In Yellowstone National Park in northwestern
Wyoming, Tinker et al. (1994) found a higher percentage of trees with serotinous cones
in lodgepole pine stands on xeric sites compared to mesic sites, and a significant negative correla-
tion of closed cones with elevation. Expanding this work, Schoennagel et al. (2003) found that the
proportion of serotinous trees was higher but varied with stand age at low elevations where
stand-replacing fires burned every 135–185 years. At high elevations where fires burned every
280–310 years, the proportion of serotinous trees was low and did not vary with stand age. More-
over, younger stands at low elevations were least likely to have a high proportion of serotinous
trees, but this proportion increased with stand age. Elsewhere, populations of a given species with
a very high proportion of cone serotiny also typically occur in the most frequently burned ecosys-
tems (Givnish 1981). For example, serotiny is common in pitch pine populations of the Coastal
Plain, but rare in the Appalachian Mountains where it occurs with serotinous populations of ta-
ble mountain pine that grow on more fire-prone sites. Taken together, these data suggest a very
close relationship between fire regimes (especially frequency and severity) and expression of the
serotinous trait across large landscapes.
230 Chapter 10 Fire
Givnish (1981) emphasized the importance of natural selection by fire in maintaining the
serotinous habit in pitch pine in a local portion of its distribution—the Pine Barrens—on the
Coastal Plain of New Jersey. In the Pine Plains area of the barrens that topographically favor
frequent fire every 6–8 years (Lutz 1934), 99% of the individuals are serotinous. Cone serotiny
decreases with distance in all directions from the Pine Plains; infrequently burned lowland sites
(fire frequency 16–28 years) averaged 84% serotinous individuals. Givnish and Lutz stressed that
the flatness and physiographic location of the Pine Plains affected the incidence of fires and
promoted high cone serotiny, in that fires from many directions and distances could carry fire
uninterrupted into the Pine Plains.
FIRE AND THE FOREST SITE

Many writers once considered and described fire to be a destructive agent with few or no beneficial
aspects. The success of prescribed fire in improving silviculture, reducing fuels, and managing
wildfire risk has led to the realization that the effect of fire on forest ecosystems is complex and
may often be entirely beneficial. The many important functions of fire and its beneficial effects in
forest ecosystems are considered in Chapter 16. In this section, we examine fire effects on site con-
ditions that affect the regeneration and growth of plants. We distinguish between the indirect
effects of fire on site quality, through its effects on vegetation, organic matter, and soil organisms,
and its direct effects on soil properties and microclimate.
INDIRECT EFFECTS
The indirect effects of fire on site quality depend upon changes in the vegetation, as discussed in
detail in Chapters 16 and 17. A high-severity fire will kill most or all of the plants above the soil
surface, and even a surface fire will kill smaller vegetation and affect dead organic matter. Soil’s
physical, chemical, and biological properties are heavily influenced by live vegetation and dead
organic matter (Chapter 9), and thus fire may invoke major changes upon the soil via its effects on
these two factors. When vegetation is killed, loss of shade affects soil temperature, ceased transpi-
ration affects soil moisture, and the death of roots affects soil stability. In these examples, fire
effects are on the organic components of the ecosystem which in turn affect the soil, and thus those
effects are indirect.
Perhaps the most obvious way fire indirectly affects the site is by its role in the type of vege-
tation growing there. Vegetation re-establishing after fire tends to be made up of light-seeded
species that colonize from outside the burned area, species with perennial root systems capable of
vegetative reproduction, and species with dormant seeds stimulated by heat. Many legumes and
chaparral species (Ceanothus spp.; Hanes 1988) fall in these categories, and the abundance of these
and other nitrogen-fixing plants often increases after fire. In such a case, although previously accu-
mulated nitrogen may be volatilized by the fire, a rapid increase in available nitrogen often occurs
and the overall site quality may be temporarily improved. By contrast, in many parts of the world,
recurrent fires favor the development of shrubby vegetation composed of sprouting species with
characteristically tough foliage low in nutritional value and slow to decompose. Heather in
northern Europe, blueberry and other ericaceous species and bracken in many countries, scrub
oaks around the Northern Hemisphere, and many chaparral types (the broad-sclerophyll scrub or
brush-land vegetation) of California (Hanes 1988) and the Southwest, all are plants that become
dominant after heavy and repeated fires.
Fire that removes forest tree cover and forest floor organic matter may substantially change
the temperature regime of soils. In general, soil temperatures generally increase following fires
because of the removal of vegetation and organic matter and the blackening of the surface. Average
maximum soil surface temperatures on burned sites may be from 3 to 16 °C higher than on
Fire and the Forest Site 231
comparable unburned areas (C. Ahlgren 1974a). Soils of openings in southern forests are colder
than those under the canopy, whereas in northern forests the reverse is true (Komarek 1971). In
the boreal forest, tree canopies promote the deepest frost beneath them, and in permafrost areas
(ground that is permanently frozen) surface soil thaw is least beneath a canopy (Rowe and Scot-
ter 1973). A large part of the Alaskan boreal forest is in the zone of discontinuous permafrost (Van
Cleve et al. 1986). Forest canopies in winter intercept snow that would otherwise provide insula-
tion from the cold on the ground, and organic matter built up under forest cover in summer pro-
vides insulation from heat. Fire therefore contributes both to increased heat flow into the soil in
summer (via removal of tree cover, creating a blackened surface, and reducing surface organic
cover) and decreased heat outflow in winter (increased snow cover). These factors cause perma-
frost to melt at increasing depths below the surface following fire, although the effects are gener-
ally temporary (Certini 2005). Viereck (1982) reported that thaw depth increased steadily for
9 years following fire. Such belowground changes significantly affect surface hydrology, vegetation
composition, and site productivity.
Removal of live vegetation by fire affects soil moisture as well as temperature, although the
actual effect is variable depending on fire severity. An immediate effect may be that the loss of veg-
etation that once intercepted precipitation before it reached the soil and transpired away soil water
would increase soil moisture. At the same time, however, the loss of organic matter on the forest
floor may increase evaporation from the soil surface. High-severity fires that consume soil organic
matter may greatly reduce its porosity and infiltration (see in the following text), as well as its
water-holding capacity, leaving it with significantly less soil moisture. For example, a modeling
study of post-fire soil-water balance in New Mexico found higher soil moisture following light-or
moderate-severity fires due to reductions in evapotranspiration when the vegetation was killed
(Atchley et al. 2018). However, high-severity burns were more strongly characterized by larger
increases in surface runoff of water compared to increased soil moisture, resulting in drier soils,
suggesting that there is a threshold of fire intensity at which soil moisture is affected.
The removal of organic matter is probably the most obvious effect of fire on soils. Fire regu-
lates dry-matter accumulation on the forest floor and in coarse woody debris, thereby affecting fuel
loadings and ultimately controlling the severity of burning. As previously discussed, fire severity
affects the density and composition of forest vegetation, which in turn influences site quality.
Abnormally long intervals between fires, such as those caused by prolonged fire exclusion, often
lead to large fuel accumulations and high concentrations of organic matter. For example, a study
conducted in the oak-dominated forests of the Missouri Ozarks showed that in a regional ecosys-
tem characterized by surface fires that burn every 8–15 years, 50% of the litter re-accumulated
within 2 years after fire, 75% within 4 years, and 99% after about 12 years (Stambaugh et al. 2006),
well within the period between fires. Under fire suppression in this and other forest types, unchar-
acteristically intense fires that eventually occur may preclude or delay the reestablishment of
normal vegetation on the site or change the type of vegetation present, although indirect effects of
change or lack of vegetation are much less than the direct effects of severe burning.
Soil microbes and fungi significantly influence site quality by decomposing organic matter,
fixing nitrogen, and providing aeration (Chapter 19). The effects of burning on soil organisms are
highly variable depending on fire intensity, depth into the soil profile, time elapsed following
burning, the nature of the soil, and the vegetation of the site (I. Ahlgren 1974b; Rundel 1981;
Dooley and Treseder 2012). Fire immediately reduces the biomass of microorganisms, mostly in
the upper 5 cm of soil; fatal temperatures for many microbes may be less than 100 °C (DeBano
et al. 1998), which is below the temperature achieved during many fires. Light surface burns may
therefore have almost no effect on microbial biomass, but extreme high-severity fires have been
demonstrated to completely sterilize the upper soil horizons (Certini 2005). In a meta-analysis of
42 studies of microbial and fungal responses to fire, Dooley and Treseder (2012) found that fires
reduced microbial biomass on average by about 33%. In addition to lethal heat, post-fire declines in
232 Chapter 10 Fire
microbes have been attributed to declines in soil carbon following fires (Waldrop and Harden 2008).
General trends for fungi following fires are less clear, although they are more heat-sensitive and
thus may be more highly affected (although some fungi have been reported to be fire-adapted;
Wicklow 1975). Fungal biomass decreases more so than microbial biomass after fire, estimated by
Dooley and Treseder (2012) to be by about 48%.
Microbial abundance typically increases after fire, often multifold, apparently due to the
sudden availability of organic substrate available in the soil, increases in soil pH, increases in soil
temperature and moisture following removal of tree canopies by fire, and other soil chemical
changes associated with burning. Deposition of ash increases soil pH and the availability of
nitrogen (Wan et al. 2001), the latter of which may increase due to higher mineralization rates and
stimulate further microbial growth. Fungi, which are tolerant of acidic soils, may decline after fire
in favor of bacteria when ash deposits increase soil pH. Microbial and fungal growth may also
benefit from reduced evapotranspiration that increases soil moisture and reduced shading that
increases soil temperature when fire kills aboveground vegetation, although temperature increases
are most beneficial in colder climates (Smith et al. 2008; Treseder et al. 2004). Where re-growth of
fire-adapted vegetation is rapid, one would expect that such changes in soil temperature and mois-
ture are relatively short-lived.
In their meta-analysis, Dooley and Treseder (2012) found that the negative effects of fire on
microbial biomass lasted about 15 years, whereas aboveground net primary productivity (ANPP;
Chapter 18) may recover as quickly as 4 years after a fire. Organic matter in the soil accumulates
much more slowly than ANPP, suggesting that the post-fire recovery of microbial biomass depends
more strongly on re-accumulation of soil organic matter than on re-vegetation. For example, Litton
et al. (2003) found that microbial biomass was still lower in burned stands compared to unburned
stands even 13 years after a high-severity fire in subalpine lodgepole pine forests, probably because
of a lack of organic layer re-accumulation relative to rapid forest re-vegetation.
Other soil fauna such as beetles, spiders, mites, collembola, centipedes, and millipedes are
typically reduced by burning but increase thereafter (I. Ahlgren 1974b). Invertebrates in general
are more mobile than microbes and fungi, which allows them the potential to move deeper into the
soil to escape the heat of a fire. Ants are less affected than other fauna because their behavior
enables them to survive in lower soil layers and they are adapted to xeric conditions of post-fire
topsoil. The social organization and rapid colonizing ability of ants also enable them to reestablish
populations rapidly after fire. Notably, reduction of litter on the forest floor by fire has a negative
effect on the diversity and abundance of soil invertebrates, such that re-accumulation of the litter
layer is likely to be a limiting factor for them much as it is for microbial biomass (Certini 2005).
DIRECT EFFECTS
Fire directly affects site quality via (i) the burning of organic matter above and on the mineral soil,
and (ii) the heating of the surface layers of the soil. The burning of organic matter releases carbon
dioxide, nitrogenous gases, and ash to the atmosphere and deposits minerals as ash. Ash from
wood and litter is more soluble than the organic matter from which it was formed. Thus, in creat-
ing ash, fire increases available minerals, at least temporarily, lessens soil acidity and increases
base saturation, and decreases total soil nitrogen.
Two classic studies are indicative of many studies on fire effects. In the ponderosa pine
region of Arizona, burning increased soluble nutrients when the surface layer of unincorporated
organic matter was ashed (Fuller et al. 1955). This increase in turn raised pH, available phospho-
rus, exchangeable bases, and total soluble salts; caused a decrease in organic matter and nitrogen
to a depth of 20–30 cm; and increased microbial activity. Unfortunately, the soil surface was com-
pacted by rains when litter was removed in this study, reducing the rate of water infiltration. In the
Douglas-fir region of eastern Washington, Tarrant (1956b) investigated the effects of slash burning
on physical soil properties. Light burning increased the percolation rate of water within the upper
8 cm of soil, but severe burning seriously impeded water drainage by about 70%.
Minerals in ash are readily available to plants because of its small particle size and high sol-
ubility, but these same characteristics also make the ash susceptible to leaching by precipitation. If
the ash is washed down into the soil so that the roots can absorb the nutrients dissolved from it, site
quality is usually temporarily improved. Such a process is typical on level, coarse soils of sandy to
loamy texture. However, site quality is lowered if the ash is leached down below the tree roots (or is
washed off the surface by runoff), which is apt to happen on very coarse sands (leaching) or clayey
soils with considerable slope (runoff).
Total nitrogen may be lost through volatilization, which is closely related to fire intensity
(Rundel 1981). Maximum ground temperatures during forest fires typically range between 200 and
300 °C (Rundel 1983), but may reach 500–700 °C in heavy fuels such as slash (Dunn and DeBa-
no 1977). Nitrogen volatilization begins when temperatures reach 200–400 °C (Neary et al. 1999).
Working in the coastal Douglas-fir region, Knight (1966) found no nitrogen loss in soils heated to
200 °C, a 25% loss at 300 °C, and a 64% loss at 700 °C. Nitrogen loss is also proportional to the
amount of fuel consumed, and considerable nitrogen may be lost when fire is intense. For example,
a high-severity fire in a second-growth, mixed-conifer stand in northcentral Washington resulted
in a loss of about 97% nitrogen originally in the forest floor and a loss of 66% of the nitrogen in the
A horizon (Grier 1975). In this case, nitrogen re-accumulation is expected to occur relatively rap-
idly due to N fixation by symbionts associated with snowbrush (Ceanothus velutinus; see
Chapter 19). Although major volatilization of nitrogen occurs with fire, available forms of nitrogen
are commonly higher on burned than unburned sites (Rundel 1981).
Much of the nitrogen lost through burning of litter and vegetation is organic nitrogen and in
a form unavailable to plants. The organic matter layer in coniferous forests in boreal regions, for
example, contains a large amount of unavailable nitrogen (Viro 1974). Organic nitrogen that is not
volatilized must be mineralized and converted into ammonia for vegetation to have access to it
(Chapter 19). The degree to which fire affects site quality thus depends largely on the ability of the
succeeding vegetation and soil bacteria to replace the available nitrogen lost in burning. Higher
post-fire pH due to ash deposition can improve the soil environment for free-living, N-fixing
bacteria and thus may initiate an increase in available nitrogen, albeit slowly. Nitrogen loss is often
considered to be a deleterious effect of fire, but fire at the same time increases the abundance of
ammonium (Neary et al. 1999). Ammonium is eventually transformed to nitrate and is lost to
leaching if not taken up by plants, such that nitrogen availability lowers to pre-fire levels in only a
few years (Certini 2005).
Importantly, heating and burning of organic matter are of much more consequence than
heating of the mineral soil, as the mineral soil horizons rarely experience the same heat as the
organic horizons above. Nevertheless, even low-severity fires may heat the surface soil to 50 and
100 °C (122 and 212 °F) at 5 cm depth (Agee 1973). Where organic horizons are very thick and are
burned with hot fires, mineral soil at the surface may reach 275 °C (527 °F) (Sackett and Haase 1992).
Temperatures of surface soil in very hot fires may approach 700 °C (1292 °F) beneath slash piles
and in such cases may exceed 100 °C as deep as 22 cm below ground. Removal of the entire organic
horizon by fire may expose mineral soil to direct irradiation by the sun, elevating soil temperatures
for months or even years (Neary et al. 1999). Soil aggregates may be broken down in this heated
zone, first by the heat and later by the direct striking action of raindrops, resulting in loss of soil
structure and lowered infiltration capacity of the surface soil, increasing surface runoff. In many
cases, burning heavy accumulations of slash in piles causes intensely hot fires and may alter the
physical structure of the underlying soil. For example, lower tree densities and markedly slower
conifer growth were documented on burned slashpile sites in northwestern Montana compared to
adjacent unburned areas (Vogl and Ryder 1969). The reduction in growth was attributed to
impaired water infiltration and other physical soil properties.
234 Chapter 10 Fire
ORGANIC MATTER AND EROSION

Fire effects on the site are greatest where the soil is composed almost entirely of organic matter,
and where the removal of organic matter by fire exposes highly erodible soils to heavy rain. Burning
peat bogs after drainage or a series of dry seasons results in complete destruction of the soil and
reproduces swamp conditions. The lost organic soils would take thousands of years of peat
accumulation to replace them, and such areas can be virtually eliminated as potential sites for
forest growth. In the United States, many bogs on the Atlantic Coastal Plain and in the recently
glaciated parts of the Northeast and the Lake States have been destroyed as forest sites by fire.
Similarly, burning organic matter located directly atop rock will eliminate the soil. In glaciated
portions of Canada and the northeastern United States and in many mountainous regions, thin
accumulation of humus provides the only nutrition for forest trees, and fire may completely
remove this thin layer (Kelsall et al. 1977).
Removal of organic matter by fire that exposes highly erodible soil on steep slopes may result
in catastrophic erosion or landslides (Rundel 1981; Swanson 1981). Post-fire landslides require
brief (15 minutes or more) heavy rainstorms to trigger debris flows. Such events are natural and
have occurred regularly in steep areas that burn; rainstorms that initiated 77% of debris flows in
the southwestern United States recur approximately every 2 years (Staley et al. 2020). The classic
example of this process is in southern California, where chaparral species and the organic matter
from them protect the granitic or clayey soils lying at approximately the angle of repose in steep
mountains. Swanson (1981) estimates that over 70% of the long-term sediment yield from steep
chaparral slopes is the result of fire effects. Chaparral is highly susceptible to burning, especially
when it reaches about 30 years of age, because of the dense, highly flammable fuel accumulations,
shrubby growth habit, and seasonal dry, windy periods (Biswell 1974; Keeley and Keeley 1988).
Following wildfire, Biswell (1974) reports that debris movement may increase dramatically,
especially on south-facing slopes (Figure 10.4). In this instance, most of the movement is during
the dry period. If the characteristic heavy rainstorms of this semiarid region strike before
re-vegetation of the burn, whole slopes may wash downhill.
Fire-caused erosion and landslides occur in many different physiographic contexts. Fires can
also cause serious erosion in the northern Rocky Mountains where erosive soils occur on steep
Photo courtesy of the Pacific Southwest Research Station,
USDA Forest Service
F I G U R E 1 0 . 4 Dry-creep erosion may be severe immediately after intense wildfire in old-growth chaparral
on slopes above the angle of repose.
slopes, the fires tend to be of high severity, and multiple burns frequently occur. In general, the
amount of sediment transported a year after fires ranges from very low in flat terrain and with dry
post-fire weather to extreme in steep terrain with heavy thunderstorms. Soil erosion typically
increases with fire severity. Notably, erosion is a temporary process and decreases as vegetation
recovers following fires, but it can lead to significant losses of nutrients from ecosystems (Neary
et al. 1999, 2005).
The burning of litter and organic matter in the soil may be significant in causing reduced
infiltration, increased surface runoff, and erosion in many areas of the western United States
where water-repellent soils have been reported (DeBano et al. 1967; Meeuwig 1971; Dyrness 1976).
Soils can become resistant to wetting in multiple ways. Removal of vegetation and litter layers by
fire allows the force of raindrops to be directly absorbed by the soil rather than by the organic
matter, which degrades soil structure at the surface and greatly reduces infiltration. In such cases,
droplets do not readily penetrate and infiltrate, but “ball up” and remain on the soil surface for
variable periods of time. This phenomenon is widespread in sandy soils throughout much of the
wildland areas of western North America supporting chaparral or coniferous vegetation, as well as
many other parts of the world (Foggin and DeBano 1971). By increasing surface runoff, this pro-
cess reduces soil moisture, allows nutrient-rich ash to be eroded away, and effectively dries the soil,
reducing microbial activity, nutrient availability, and plant growth (Neary et al. 1999).
Fire can actually promote the development of water-repellent soils. In unburned areas,
non-wettable or hydrophobic organic molecules produced by litter decomposition coat surface soil
particles, creating a weak water-repellent layer between the litter layer and the mineral soil
(Figure 10.5; DeBano 1969). High-intensity fires consume litter and volatilize the hydrophobic
substances, which diffuse downward and attach to cooler soil particles lower in the soil profile
(Figure 10.5). This deeper water-repellent layer forms as a result of fire when soil temperatures rise
above 176 °C (349 °F), but is destroyed when temperatures exceed 288 °C (550 °F) (DeBano
et al. 1976). As a result, water repellency is high after summer and fall wildfires such that rain
(a) Unburned (b) Fire (c) Burned
Litter Non-wettable zone Wettable layer
Water repellent layer Decreasing Non-wettable layer

temperature
Wettable soil Wettable soil Wettable soil
F I G U R E 1 0 . 5 Soil non-wettability before, during, and after fire. (a) Before fire, the non-wettable substances
accumulate in the litter layer and mineral soil immediately beneath it. (b) Fire burns vegetation and
litter layer, causing non-wettable substances to move downward along temperature gradients. (c) After a
fire, non-wettable substances are located below and parallel to the soil surface on the burned area.
Source: DeBano et al. (1967) / United States Department of Agriculture / Public Domain.
236 Chapter 10 Fire
falling on the soil surface infiltrates readily until impeded by the shallow non-wettable layer, facil-
itating surface runoff. In regions where the terrain is flat, this process contributes to drying of soils,
but in steep regions can cause serious erosion (Neary et al. 1999).
BENEFICIAL EFFECTS OF FIRE

Thus far we have emphasized mainly the deleterious effects of fire on soils and site quality (other
than the release of ammonium), but controlled burning has shown to benefit site quality in many
forest ecosystems. On some sites, such as in the boreal forest where thick layers of humus tend to
develop under coniferous vegetation, burning improves site quality because it enhances the post-
fire thermal regime. Increased soil temperatures in burned areas hasten spring development of
roots and shoots. Extremely high post-fire soil temperatures may cause seedling mortality and
delay forest regeneration, although fire-dependent species on fire-prone sites are typically adapted
to tolerate such extreme conditions.
As long demonstrated by many classic studies, the effect of fire on site quality is often
relatively subtle. Metz et al. (1961) showed that repeated burning in the sand plains of the Southern
Pine region at worst had no major detrimental effect on site quality and, at best, was locally bene-
ficial to the soil. Likewise, nine successive annual surface burns under even-aged shortleaf and
loblolly pines had little effect on the nutrient content or structure of the topsoil on loess soils on
flat terrain in Arkansas (Moehring et al. 1966). Studies of soil biota after 20 years of annual pre-
scribed burning on a very fine sandy loam on the Coastal Plain indicated little evidence that fire
adversely affected metabolic processes (Jorgensen and Hodges 1970). Burning had no effect on
total concentration of fungi, although it did reduce their abundance by decreasing the organic
horizon. The number of bacteria in the organic layer was reduced by annual burning but not in
mineral soil. Stark (1977) found that controlled burning in a Douglas-fir–larch forest on sandy
loam soils in northwestern Montana generally reduced accumulated fuels without nutrient loss,
runoff, or erosion, and estimated that burning could be conducted on this soil for a very long time
with no problem of soil fertility. Notably, ecosystems such as the sand plains of the Lake States,
where organic matter is the major source of nutrients for plants, may not benefit from burning,
especially for species with relatively high nutrient requirements (Stoeckeler 1948, 1960).
SUGGESTED
READINGS
Agee, J.K. (1993). Fire Ecology of Pacific Northwest Forests. Environment, New Series, vol. 12A (ed. O.L. Lange,
Washington, D.C.: Island Press 505 pp. Chapters 1, P.S. Nobel, C.B. Osmond and H. Ziegler). New York:
4, 5, and 6. Springer–Verlag.
Arno, S.F. (1980). Forest fire history of the northern Wein, R.W. and MacLean, D.A. (1983). The Role of Fire
Rockies. J. For. 78: 460–465. in Northern Circumpolar Ecosystems. New York:
Baker, W.L. (2009). Fire Ecology in Rocky Mountain Wiley 322 pp.
Landscapes. Washington, D.C.: Island Press 632 pp. Whelan, R.J. (1995). The Ecology of Fire. Cambridge:
Johnson, E.A. (1992). Fire and Vegetation Dynamics. Cambridge University Press 346 pp.
Cambridge University Press 129 pp. Chapters 4 and 6. Wright, H.E. Jr. and Heinselman, M.L. (1973). The eco-
Johnson, E.A. and Miyanishi, K. (ed.) (2001). Forest logical role of fire in natural conifer forests of west-
Fires: Behavior and Ecological Effects. New York: ern and northern North America: introduction.
Academic Press 594 pp. Quat. Res. 3: 319–328.
Rundel, P.W. (1981). Fire as an ecological factor. In:

Physiological Plant Ecology I., Responses to the Physical
Site Quality and Ecosystem
CHAPTER 11
Evaluation and Classification
T hus far we have considered separately the forest tree and the physical factors that affect it
within a landscape ecosystem framework. In this chapter, we begin an important focus on
“site” of the forest ecosystem—the often-complex combination of factors that not only influences
tree and forest productivity but affects the entire range of management and conservation appli-
cations. Forest ecosystems consist of (i) geographic position in space and its associated physical
factors (site), and (ii) the biota (forest trees and associated plants and animals) that occupy these
places and that are supported by site factors. The physical factors of climate, physiography, soil,
and disturbances determine the site conditions that affect the occurrence and diversity of organ-
isms, their productivity, and how they change in space and time. The biota in turn modify the
direct effects of the physical factors. Obvious examples of how biota affect physical factors include
how forest canopy trees affect understory light, temperature, and soil development, or how uproot-
ed or fallen trees in uplands and wetlands (Chapter 8) create microtopography that affects forest
composition and dynamics. Therefore, the forest site is the sum total of all the factors affecting the
distribution and growth of forests or other vegetation (see site diagram, p. 120).
For centuries humans have evaluated the site quality of forests for a variety of purposes. Both
simple and complex methods have been attempted with varying degrees of success. We conclude
this section of this book, covering the forest environment, with a discussion of approaches to site-
quality evaluation. Forest managers were historically concerned mostly with trees and typically
with the potential land quality for fiber production, and single-factor methods of evaluation tend
to reflect these objectives. Emphasis has largely shifted from the stand (bioecosystem concept) to
the land or ecosystem (geoecosystem concept), such that we no longer consider site-quality evalu-
ation only in terms of tree or stand production. Instead, forest managers seek an understanding of
whole ecosystems as an attempt to assure their sustainability (Rowe 1992b; Franklin et al. 2018;
Palik et al. 2021).
Forest and land managers evaluate sites and ecosystems for recreation, aesthetic values, wa-
ter and wildlife conservation and management, ecosystem and biological diversity, and mainte-
nance of ecosystem processes in addition to, or to the exclusion of, fiber production. Approaches
involving multiple factors and mapping of entire landscape ecosystems have increased in popular-
ity because they provide both the basis for biomass production and the ecological framework for
ecosystem conservation, management, and restoration at multiple scales. We also include in this
chapter a review of single-factor site evaluation and conventional use of tree height to estimate site
quality, as this process has produced a wealth of insights concerning plant growth and occurrence
for over a century.
A forest scientist must integrate many physical site factors to effectively estimate forest site
quality in terms of forest productivity. Site factors are interdependent, but are also dependent in
part upon the forest, which is itself a major site-forming factor. Because of these interactions, esti-
mating site quality from an evaluation of a few important site factors, important as it is in practical
forest ecology, is only approximate. Nevertheless, estimating forest productivity is of the utmost
237
238 Chapter 11 Site Quality and Ecosystem Evaluation and Classification
importance in forest ecology and ecosystem management. Productivity, or actual site quality, may
be measured directly for a few forests where accurate long-term records of stand development and
growth have been maintained. Most often, it can only be estimated indirectly by one or more of the
alternatives that are considered in this chapter:
Vegetation of the forest:
Tree height (site index method).
Ground vegetation (indicator species and species groups).
Overstory and ground-cover vegetation in combination.
Factors of the physical environment:
Climate.
Physiography.
Soil survey and soil-site methods.
Multiple-factor and multiple-scale approaches (using some or all the abovementioned factors,
disturbance regime, and forest land-use history).
Multiple-factor methods have broader applicability than productivity estimation alone, and
have reflected the shifting emphasis toward multiple forest values. Significant reviews of tradi-
tional forest site-quality estimation have been published by Coile (1952), Rennie (1962), Ralston
(1964), Jones (1969), and Carmean (1970a, 1977, 1996), and comprehensive reviews of site-quality
evaluation, its history, methods, and application, by Carmean (1975) and Hagglund (1981), also are
available.
DIRECT MEASUREMENT OF FOREST PRODUCTIVITY

Actual forest productivity, at least in terms of fiber production, is typically measured in terms of
the gross volume of bole wood per unit area per year over the normal rotation. This gross mean
annual increment (m.a.i.) may be computed from long-term permanent sample plot data. Classic
studies showed Douglas-fir on a pumice soil site on the North Island of New Zealand to yield a
gross m.a.i. of 31 m3 ha−1 yr−1, and yields of 36 m3 ha−1 yr−1 may be expected from Monterey pine on
similar sites (Spurr 1963). In the United States, gross m.a.i. may reach 15 m3 ha−1 yr−1 on high-
quality sites, but average productivity in temperate forests of North America and Europe is approx-
imately 5 m3 ha−1 yr−1. Note that growth in these values is presented as gross—the total amount of
wood accumulated by all trees within a given unit of time without accounting for natural mortality,
removal by humans, or decrease in wood volumes due to rot. Gross values allow a manager to
obtain comparable increment measurements of aboveground production. In reality, actual gross
productivity data are rare, and actual yield is conditioned by genetic factors, age or rotation, stand
history, and stand density as well as site factors. Nevertheless, actual growth is the proven produc-
tivity of a site and therefore represents the closest available approximation of potential productivity.
Theoretically, on a given site, a stand of a given species and age will produce the same
amount of wood per year at various densities as long as the site is fully occupied. Trees in un-
thinned stands will fully occupy a site as long as they retain good crown development and vigor.
Likewise, trees in thinned stands or even in open stands will normally fully occupy the site, even
when excess crown space is available in the stand, if they are able to fully utilize available soil water
and nutrients. Thus, thinned and unthinned stands should provide the same total increment if
mortality within a growth period, whether from cutting activity by humans or from natural causes,
is included.
Tree Height as a Measure of Site 239
Modern forest ecologists no longer attempt to estimate forest productivity using the growth
of the boles alone, but in terms of all components of the forest ecosystem. Ecosystem ecologists
now measure not only the stems but also the branches, leaves, and roots; the organic matter in the
forest floor; and even the animals inhabiting the forest, to provide a more exact appraisal of
the entire forest ecosystem. A detailed discussion of forest biomass and carbon productivity and its
measurement is provided in Chapter 18.
TREE HEIGHT AS A MEASURE OF SITE

The height of free-grown trees is well known to be the best indicator of a site’s capacity to produce
wood of that species. Free-grown trees are defined as individuals of a given species and age that
have grown without suppression by an overstory canopy since their establishment. The height of
free-grown trees is the tree dimension least influenced by stand density and may thus be used as
an index of site quality in even-aged stands that vary in density and silvicultural history. Tree height
is especially appropriate for shade-intolerant and midtolerant species compared to tolerant species
whose individuals may be suppressed in the understory for variable periods of time.
Tree height of the majority of a forest stand at a specified age is commonly termed site
index. In applying site index, a set of height/age curves are developed for a geographic area, such
as the Coastal Plain and Piedmont physiographic regions of the southeastern United States
(Figure 11.1). Using the mean total height and age of a representative sample of dominant “site
trees,” this set of curves can be used to determine the site index for any given loblolly pine stand in
the area. In Figure 11.1, the site index is 50 for a site supporting dominant pines that are 55 ft tall
at 60 years, and is 95 for a site supporting pines that are 85 ft at 40 years. In the United States, site
index is defined as the average height that the dominant portion of an even-aged stand will have at
a standard age. This standard age is generally 50 years in the eastern United States, and 100 years
140
120
50 years, coastal and Piedmont areas, southeastern U.S.
Total
Height in feet of average dominant trees, by site index at
height 110
(ft)
120
100
90
100
80
70
80
60
60 50
40
40
20
20 30 40 50 60 70 80 90 100
Total age, years
F I G U R E 1 1 .1 Site-index curves for second-growth loblolly pine in the Coastal Plain and Piedmont areas of
the southeastern United States. Source: Hampf (1965) / United States Department of Agriculture /
Public Domain.
for the longer-lived species of the West. Occasionally, other standard ages are specified for a
particular species or region, for example, 25 years for pulpwood rotations in the South. Impor-
tantly, site index is simply a number that is used to indicate the relative productivity of a stand of
a given species at one place compared to stands of the same species on other sites. While useful for
such simple comparisons, site index is unable to indicate why a given site may be poor, medium, or
good for growth and yield of the species currently growing there.
Determining and measuring dominant or codominant trees is often problematic, and thus
the trees that should be measured for site determination are carefully selected. Often the height of
an objectively determined sample of the larger trees in the stand is measured. In Britain and some
other countries, the concept of top height is widely used, which refers to the arithmetic mean
height of the 250 largest-diameter trees per hectare. It is often not feasible to measure so many
trees per hectare for height, and heavy thinning practices commonly reduce the stand to fewer
than 250 trees per hectare. In such cases, a mean height based on the largest 100 trees per hectare
(sometimes defined as mean predominant height) is used. Top heights may also be based upon
even fewer of the largest diameter trees when many are not available, such as the 60 largest-
diameter trees per hectare, the 8–12 largest trees, or even upon the 1 largest tree per hectare.
Tree height is the single best measure related to the site productivity of a given species, but it is
not necessarily unrelated to other factors, nor is there always a perfect correlation between stand top
height and site productivity. Stand density, particularly extremes of stand density, may influence
height growth, and under such circumstances site-index curves are developed separately for different
stand density classes, as has been done for lodgepole pine in the Rocky Mountains (Alexander 1966;
Alexander et al. 1967). Moreover, it is not uncommon for the same site index to be determined for a
given species on two different sites because of differences in stand density rather than similarities in
productivity (Assmann 1970; Curtis 1972). For example, for Douglas-fir in northern Idaho and north-
western Montana, the density of trees is greater in even-aged pure stands than in uneven-aged mixed-
species stands of the same dominant height (Sterba and Monserud 1993). Although the site index for
Douglas-fir for these two sites is the same, the productivity is different.
Within regional ecosystems, genetic factors likely are strong drivers of height growth, espe-
cially across an area heterogeneous in climate and landform (Monserud and Rehfeldt 1990). Ge-
netic control of tree height growth within a species tends to be weaker at local scales, and site
factors such as soil water and nutrients are more important. Naturally occurring aspen clones on
sandy soils in northern Lower Michigan have been found to differ greatly in height on the same
site, with some clones more than twice as tall as adjacent clones of the same age (Zahner and
Crawford 1965). Such variation is much more likely in species that develop multi-stemmed clones
than in species where each stem is a different genotype. Competition among clonal trees is primar-
ily between stems of the same genotype, such that individual trees in slow-growing clones are
more likely to survive than in stands where competing individuals are of different genotypes. In
such a case, sampling the largest dominant trees within the same clone may overestimate site
quality if the clone is genetically superior, and the mean total height from at least five different
clones rather than five “site trees” is probably more appropriate.
A final point is that the condition of the site at the time of stand establishment, as well as
competition from other vegetation in early years, may affect height growth. Both naturally reseed-
ing and planted pines will usually show different growth trends on old fields, where soil nutri-
ents, microorganisms, structure, and drainage have been altered, compared to cutover sites.
SITE-I NDEX CURVES

Determining site index on the basis of tree height typically utilizes a height-over-age growth
curve to estimate height at a standard age. In the past, such curves for North America were devel-
oped from a series of regression curves, based upon a single guiding curve and harmonized to
have the same form and trend (Figure 11.1). This approach has many weaknesses that have been
well described (Spurr 1952b). Most obviously, the technique is sound only if the average site
quality does not change among age classes. For example, it is often the case that younger stands
are found on generally better sites (perhaps because of early logging on these sites), while re-
maining old growth stands are concentrated on poorer sites. In such a situation, the average curve
will be warped upward at younger ages and downward at older ages. The reverse situation can
also occur.
The assumption that the shape of the height-growth curve is unchanging among sites is a
second major weakness of the conventional technique. This generalization may provide poor
results when differences in climate, landform, or soil conditions are present. For instance, if the
depth of soil is limited by physical reasons, tree-height growth may be normal up to the point
where the depth of the soil becomes a limiting factor (curve B; Figure 11.2). In another ecosystem
and soil, the same species may grow slowly until roots reach an underlying enriched horizon or a
deep-lying water supply, after which growth will be accelerated (curve C). The shape of these two
growth curves may differ from the normal growth curve on a normal soil (curve A).
Similarly, harmonizing site-index curves as part of the standard technique assumes that site
differences are apparent at early ages. Harmonizing site-index curves assumes that a site produc-
ing a taller tree at age 50 or 100 will also produce a taller tree at all other ages. However, many
plantations and even-aged natural stands on marginal sites grow normally at young ages and
exhibit sharply decreased growth in middle life. For example, planted black walnut trees on seven
contrasting sites in southern Illinois (Figure 11.3) showed rapid early growth, even on the poorer
sites, but slowed abruptly after 10 years (Carmean 1970b). These polymorphic patterns are closely
related to soil conditions. In Figure 11.3, trees on plot 1 are growing on a deep, well-drained alluvial
silt loam, whereas those on plots 4–7 are growing on a bottomland silt loam soil underlain at 1 m
or less by a gravelly subsoil. Curves harmonized with a standard average curve would be unable to
show such plant–soil relationships.
Height-growth patterns vary not only across the geographic range of a species but also among
different local ecosystem types (local areas of contrasting soil and topography). Height-growth
F I G U R E 11 .2 Theoretical effect of soil Height

on height growth. A: normal height
growth in homogeneous soil. B: height
growth on good but shallow soil. C: C
height growth on soil poor at the surface
but with a fertile horizon beneath. A
Age
70
1
60 S.I. 80
2
50
3
S.I. 60
Total height, ft
40
S.I. 40
4
5
30
6
20
7
10
0
0 10 20 30
Total age, years
F I G U R E 1 1 .3 Polymorphic height-growth curves of dominant black walnut trees from plantations in

southern Illinois. The height–age curves are averages from 36 sectioned trees growing in plots on 7 different
sites. Plot 1: deep, well-drained alluvial silt loam; plot 3: similar to 1 but with restricted internal drainage;
plots 4–6: bottomland silt loam underlain at 0.5–1 m by gravelly subsoil; plot 7: bottomland silt loam
underlain at 0.3 m by a gravelly, cherty subsoil. Source: Carmean (1970b) / with permission of Oregon State
University. Reprinted with permission from Tree Growth and Forest Soils, © 1970 by Oregon State University
Press. Also shown are the regional harmonized site-index curves for black walnut (Kellogg 1939).
atterns of oak, for example, not only vary between different soil texture groups but also vary with
p
aspect and slope within soil groups. Likewise, height-growth patterns of jack pine similarly vary
across even minor differences in soil texture as well across landforms (Kashian and Barnes 2000).
Indeed, polymorphic site-index curves have been repeatedly demonstrated to better characterize
the variable height-growth patterns of forest trees compared to simple monomorphic pattern por-
trayed by regional harmonized curves (Carmean 1970b, 1975).
It is not uncommon for relatively small differences in curve shape to exist for a broad area.
For example, in northern Ontario, jack pine exhibited only slight differences in average height-
growth patterns among several different glacial landforms and soils, and only one set of curves
were developed for this broad range (Carmean et al. 2001). However, pines on the poorest sites
(8 m at 50 years) exhibited a relatively linear growth pattern, whereas those on the best sites (22–
24 m at 50 years) were more curvilinear (rapid growth rate for the first 50 years and a gradually
decreasing rate to 100 years). For black spruce, the most linear pattern was on the poorest sites
where spruce grew on nutrient-poor, organic soils, and the pattern became increasingly curvilin-
ear as site quality improved (Carmean et al. 2006). Polymorphic site-index curves have been
prepared for many species (Carmean 1975, 1996), typically for broad regions. Polymorphic
curves for specifically mapped ecosystem types were developed for Norway spruce and other
species for intensive forest management in southwestern Germany (Barnes 1984), as well as for
jack pine in northern Lower Michigan.
Weaknesses of standard site-index curve techniques suggest that height-growth curves

should be based upon actual measured growth of trees for specific sites rather than averaging
together height and age values from plots across a species’ range. Doing so requires measuring
actual tree growth among a series of soil-site conditions such that separate growth curves may be
developed if necessary. These data may be collected from long-term permanent sample plots.
Where permanent plot data are absent, stem analysis may be used to reconstruct past growth over
time by sectioning trees from top to bottom. The pitfalls of this technique are more easily overcome
than those of temporary plot techniques. The appropriate regression methods for estimating site
index based on stem analyses are discussed by Curtis et al. (1974a,b).
A third method to reconstruct height growth with age is to measure the distance between
nodes on the boles of trees that produce distinct annual whorls, such as Douglas-fir and many
pines. Such measurements can be used to create accurate height-growth curves (Beck 1971; Car-
mean 1996). Height growth is fairly constant in the middle portion of the height-over-age growth
curve, such that the average annual distance between whorls may be used to predict height
growth over a future short period. The use of mean annual height increment over the middle
period of height growth, the growth-intercept method, is well established as a measure of site
quality. The 5-year intercept above breast height has proved better correlated with site quality
than total height for loblolly, shortleaf, and slash pines (Wakeley and Marrero 1958). The tech-
nique has also been adapted to red pine, Douglas-fir, southern pines (Carmean 1975), black
spruce, jack pine, and balsam fir (Mailly and Gaudreault 2005). Importantly, growth-intercept
models have resulted in different site-index curves for jack pine growth in natural stands com-
pared to plantations (Guo and Wang 2006).
COMPARISONS BETWEEN SPECIES

Different species growing on the same site will show variation in height response to a given forest
site, and to a much greater degree than trees of differing genetic character within a species. Site-index
curves may be quite different for two different species on the same site, because individual species
differ in their maximum growth rates (Aber and Melillo 1991) and they respond differently to envi-
ronmental conditions. Nevertheless, site index as predicted from height measurements of one species
may sometimes be used to estimate the site index of the other. Carmean (1975, 1996) lists such
studies, including examples from the southern pines, eastern hardwoods, northern and boreal trees,
and western conifers; examples for northern Idaho conifers are given by Steele and Cooper (1986).
ADVANTAGES AND LIMITATIONS

Site index is the method for determining site quality that is most often employed in North
American forests because it provides a convenient and reliable estimate of forest productivity.
Site index is also useful in selecting the most productive tree species for specific sites. Developing
site index curves for specific ecosystem types or groups of ecosystems with similar site condi-
tions and genetic potential is generally useful, although it should be applied with great caution
due to the various problems cited previously. Site index is not appropriate, however, for deter-
mining silvicultural practices geared toward regeneration and care of stands. Instead, site-
quality studies and ecosystem classification using vegetation and physical site factors are more
appropriate and have attracted universal attention. Such studies provide the ecological basis for
determining silvicultural practices for various management goals as well as a basis for determin-
ing productivity. In the following text, we discuss how these other approaches used to estimate
site quality are useful not only for understanding wood production potential, but also in under-
standing the broader ecological significance of site quality in managing ecosystems for wildlife,
water, recreation, and wilderness.
VEGETATION AS AN INDICATOR OF SITE QUALITY

Tree species presence, relative abundance (also called “coverage”), and relative size reflect the
nature of the local forest ecosystem of which they are a part and therefore may be useful indicators
of site quality. Because of happenstance and elements unrelated to site quality, the correlation bet-
ween site quality and these factors may not always be apparent. Factors such as plant competition
and mutualisms, herbivory, low light conditions in the understory, or past events in the history of
the vegetation such as drought, insect outbreaks, or human disturbance may disrupt the relation-
ship. In many instances, however, vegetation sufficiently reflects site characteristics to make it a
successful index of site quality. Plants themselves may be used as phytometers, or the measure of
site, integrating the effects of many climatic and soil factors that are difficult to measure directly.
They are not precise phytometers, however, and their indicator value changes with changes in
regional macroclimate and physiography.
Tree species are long-lived, relatively unaffected by stand density, and easily identified in all
seasons of the year, and as such are useful indicators. The occurrence of species with an extremely
narrow ecological amplitude is indicative of a particular site. Demanding hardwoods such as black
walnut, white ash, and tulip tree develop best only on moist, well-drained, protected sites rich in soil
nutrients and characterized by well-developed organic soil horizons. Most trees, however, have a
wider ecological amplitude and may occur and prosper on a wide variety of sites such that their
presence is of little indicator value. For such species, relative abundance and size may be a better
indicator than their presence. For example, in similar eastern hardwood forests to those described
in the earlier text, the greater the proportion of northern red oak to black and white oak in the forest,
the better will be the soil-water conditions and the general site quality. Therefore, the sizes of free-
grown dominant oaks of these three species in even-aged stands of uniform density may be used as
an index of site quality.
Plant species of the shrub, herb, and moss–lichen layers (termed groundflora or ground
cover) of the forest understory are often more restricted in their ecological tolerances than tree
species and may therefore be more useful as plant indicators. This relationship is particularly
strong in the boreal forest, where the dominant tree species are spruces, firs, pines, and birches
that are few in number and widespread in their distribution on various sites, and thus have
relatively poor indicator value. Site classification methods based upon ground-cover indicator
species have been successful under such circumstances. They are most easily applied where
regional variation in altitude, precipitation, and soil conditions is high and where humans have
not markedly altered the original vegetation. These methods have also been developed and used
where such variation is more subtle and past human disturbances are nontrivial. Notably,
ground-cover species are more apt to be influenced by stand density, past history, and the com-
position of the forest than are tree species. Plant indicators have been adapted successfully to
many temperate forests of North America. Overstory dominants and physical ecosystem factors
are increasingly used with ground-cover vegetation to estimate site quality and classify and map
forest ecosystems.
The classic example of plant indicator species is Cajander’s system of site types for Finland
(1926), designed to distinguish site-quality classes in pure and mixed stands of spruce, pine, and
birch. Pine height growth, for example, differed on four such site types (Figure 11.4); this system
remains useful decades later for the site conditions where it was developed (Lahti 1995). The Cladi-
na type is characterized by a lichen understory and occurs on dry, sandy, poor-quality heaths.
Mosses and lichens are also present on the Calluna type, but heather (Calluna vulgaris) is typically
the predominant species. The Vaccinium type is more dominated by low shrubs than mosses or
lichens, particularly lingonberry (Vaccinium vitis-idaea), and occurs on moderately dry, sandy soils
and on glacial ridges. Finally, the Myrtillus type, named for the predominant Vaccinium myrtillus,
generally lacks lichens and is found on richer soils supporting late-successional spruce forest, but
frequently converted to pine by fires, felling, and silvicultural control.
Vegetation as an Indicator of Site Quality 245
Height of stand
(ft) (m)
100 30
type
illus
Myr t
type
in ium
Vacc
e
n a typ
Callu
66 20
ype
ina t
Clad
33 10
0
20 40 60 80 100 120 140
Age of stand, years
F I G U R E 11 .4 Height-over-age curves of Scots pine on four site types in Finland characterized by different
indicator species.
It should be remembered that sites typically vary along a gradient rather than forming
istinct and mutually exclusive site classes. For example, a sharp soil and microclimatic con-
d
trast may occur between an upland and adjacent wetland swamp or between a sandstone-
derived residual soil and a limestone-derived residual soil, but otherwise site changes tend to
be gradual. Segmenting this ecological gradient into classes or types is more useful in ecologi-
cal comparisons and in land management. For example, Rowe (1956) identified five classes
along a soil-water gradient from dry to wet in the mixed boreal forest in Manitoba and Sas-
katchewan. Understory vegetation was assigned to each class depending on their soil-water
requirements, thereby forming groups of species that indicate distinctive positions along the
soil-water continuum.
SPECIES GROUPS OF GROUND COVER

A site may be characterized in terms of very few species, but these key indicator plants may or may
not be present at a given locality because of chance, past forest history, or present competitive condi-
tions. For example, it is awkward to classify a site or ecosystem type as an Oxalis–Cornus type if both
species are missing, yet other characteristics such as the sum total of the vegetation and the growth
characteristics of the ecosystem may otherwise fit the classification well. These pitfalls may be over-
come by using many ground-cover species grouped together using their similar environmental
requirements as determined by intensive field study. Ecological species groups have been deter-
mined by a number of different approaches (Daubenmire 1952; Rowe 1956; Sebald 1964; Mueller-
Dombois and Ellenberg 1974; Barnes et al. 1982; Ellenberg 1988; Klinka et al. 1989) depending on
the nature of site and vegetation conditions of the study area and the purpose for which the method
is to be used. The concept of ecological species groups is attributed to Duvigneaud (1946), but plant
ecologists have long recognized the occurrence of plants with similar ecological distributions.
Ecological species groups are developed using one of two different general approaches. In
the first, each species is assigned an indicator value for individual site factors (e.g., low nitrogen,
high nitrogen). For example, Ellenberg (1974) listed the indicator value (in classes 1–9) of approx-
imately 200 vascular plants for each of the gradients of light, temperature, continentality, soil
water, pH, and soil nitrogen in the western part of central Europe. In the second approach,
groups of species that reflect a similar combination of site factors are constituted. In both
approaches, the “importance” or abundance of the groundflora species present on a site is quan-
titatively rated with percent coverage, using the aerial crown or foliage coverage of the species.
Coverage for a given species is defined as the proportion of a ground area that is covered by a
vertical projection of the foliage of all individuals of the species. This process is tedious, time-
consuming, and subjective, such that ecologists typically estimate coverage on a scale from <1 to
100% using a series of cover classes whose number varies depending on user purpose, site condi-
tions, and number and rarity of the species present (Mueller-Dombois and Ellenberg 1974; Spies
and Barnes 1985b; Klinka et al. 1989).
Indicator Plants of Coastal British Columbia Over decades V. J. Krajina and his students
characterized the plant–environment relationships of over 3000 species in British Columbia. As
part of this effort, Klinka et al. (1989) described the indicator value of 416 species of coastal British
Columbia using 4 site attributes: climate, soil water, soil nitrogen, and ground-surface material.
The gradient of each attribute was segmented into classes (6 for climate, 6 for soil water, 3 for soil
nitrogen, and 5 for ground-surface material), and each species was placed into a class for each attri-
bute, forming 20 indicator species groups. For example, indicator values were assigned to 337
species in the 6 classes for soil water: (i) excessively dry to very dry (17 species), (ii) very dry to
moderately dry (50 species), (iii) moderately dry to fresh (74 species), (iv) fresh to very moist (107
species), (v) very moist to wet (59 species), and (vi) wet to very wet (39 species). A given site or eco-
system is then characterized by the presence–absence and coverage (abundance) of different
indicator species in groups of each of the four attributes.
Ecological Species Groups Ecological species groups were first used for forest ecosystem
classification, mapping, and site-quality evaluation in the southwestern German state of Baden-
Württemberg (Sebald 1964; Dieterich 1970; Barnes 1984). Plant species are grouped together that
repeatedly co-occur in areas with similar combinations of soil water, nutrients, light, and other
factors, as they are perceived to have similar ecological requirements or tolerances. Species groups
consist of herbs, shrubs, and (less commonly) mosses and lichens of the groundflora, and each
group is named for the most characteristic species. In North America, the approach has been
applied in distinguishing and mapping landscape ecosystem types in old-growth forests in Michi-
gan (Pregitzer and Barnes 1984; Spies and Barnes 1985b; Simpson et al. 1990) in highly disturbed
oak–hardwood forests of southern Michigan (Archambault et al. 1990) and Wisconsin (Hix 1988),
in jack pine forests of northern Michigan (Kashian et al. 2003), in forests of the southern
Appalachian Mountains (Abella and Shelburne 2004), and in ponderosa pine forests of Arizona
(Abella and Covington 2006a), among other systems. Methods used to develop and evaluate the
groups are given by Spies (1983), Spies and Barnes (1985b), and Archambault et al. (1990). An
example of 2 contrasting groups of the 16 groups used to map forest ecosystems in Upper Michigan
(Spies and Barnes 1985b) illustrates their indicator value:
Clintonia borealis group (6 species): Most common on moist, very infertile soils supporting coni-
fers. Dry to wet. Very infertile to infertile. Shade tolerant.
Caulophyllum thalictroides group (5 species): Characteristic of very fertile, moist to very moist
soils. Moist to very moist. Fertile to very fertile. Shade tolerant.
Co-occurrence of plants in the field is due to many interacting physical and biotic factors,
including mutualisms and competition with other plants, and the groups constituted represent the
integrated effects of multiple-factor gradients. The absolute indicator value of a plant for any
individual factor may therefore be difficult to determine. Once constituted, the groups may be used
to help distinguish landscape ecosystems in the field by their presence or absence and by the
relative coverage of plants in each group (see spectral format in Table 11.1). Importantly, ecological
species groups are never used alone, but always with attributes of physiography, soil, microclimate,
and the composition and vigor of overstory trees.
How widely can the perceived indicator value of an individual species or species group be
extrapolated beyond its local use? Applications of this approach in Germany and Michigan show
that the indicator value of a given species changes from one regional landscape ecosystem to
another. As macroclimate, landform type and pattern, species genetics, and species competition
and mutualisms change from one region to another, the relative indicator value of a species also
changes. Among regions, the usefulness of certain species increases (especially as new species
assert dominance) and that of other species declines. As such, a given species does not always indi-
cate the same relative level of a given site factor (e.g., high soil fertility, low soil water) in different
regions. For example, four different sets of ecological species groups were required to map
landscape ecosystems in four macroclimatically different areas of northern Lower Michigan and
Upper Michigan, suggesting that a regional landscape classification (e.g., Albert et al. 1986;
Figure 2.10 in Chapter 2) is a useful framework to determine how widely to extrapolate a set of
ecological species groups. In any case, it is clear that indicator species and groups should be used
with great care.
PLANT ASSOCIATIONS AND HABITAT TYPES IN THE WESTERN

UNITED STATES
Widespread deforestation in many parts of the world has made it difficult to evaluate or classify
sites using native or pre-European colonization forest trees and groundflora. An exception to this
trend occurs in the northern Rocky Mountains, where relatively pristine old-growth forests have
facilitated the development of an indicator system employing both trees and understory plants.
Rexford Daubenmire (1952; R. and J. Daubenmire 1968) used groups of understory species, termed
subordinate unions, in combination with late-successional overstory species, termed dominant
unions, to distinguish forest or plant associations. A habitat type (literally, the type of late-
successional vegetation on a particular site or habitat) is the collective area of a given plant
association. Different areas with the same habitat type may have similar environmental condi-
tions, and thus the concept has value for land management.
Operational Site Classification Based on Vegetation Land classification using the

habitat type method (Pfister et al. 1977; Pfister and Arno 1980) has been applied to millions of
hectares of federal forests in the Rocky Mountains, the Intermountain Region, parts of the
Southwest, California (Wellner 1989), and in the Pacific Northwest (Hall 1989). Habitat types are
typically applied to parts of states (e.g., Montana, Pfister et al. 1977; northern Idaho, Cooper
et al. 1991), to individual National Forests or parts thereof (Alexander et al. 1986), or a community
type (trembling aspen, Mueggler and Campbell 1986). Habitat types are typically developed
independent of a regional ecological framework.
The habitat type approach is based on the concept of the natural potential climax (polycli-
max theory of Tansley 1935; R. and J. Daubenmire 1968; Pfister and Arno 1980), which assumes
that overstory and understory vegetation integrate and express the environmental complex for a
specified geographic area. A habitat type represents all parts of the landscape that support or
may support the same primary climax vegetation (Alexander 1986). A climax community is
defined as the final stage of forest succession for a given habitat; without natural or other disturbance,
Table 11.1 Representative ecological species groups and ecosystem types (site units) along moisture gradients, upper Neckar growth district,
Baden-Württemberg, southwestern Germanya.
Ecosystem type (site unit)

Ecological species group Moderately fresh → Moderately dry Fresh → wet
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Vaccinium myrtillus group ● ● ● ● • ● •
Leucobryum glaucum group • • •
Bazzania trilobata group • • •
Deschampsia flexuosa group ● • ● ● • • • ● • •
Pirola secunda group • • • • • • • •
Milium effusum group • • • • ● ● • • • • • ● ● ● ●
Elymus europaeus group • • • •
Aruncus sylvester group • ●
Ajuga reptans group • • • • • • ● • •
Stachys silvatica group • • • • ● ●
Chrysanthemum corymb group • • • •
Carex glauca group • ● ● •
Molinia caerulea group • •
Key: ●, species of the group abundant; •, species of the group moderately abundant; •, species of the group rare.
a
Thirteen of Sebald’s 24 groups are shown. Major differences among the groups are indicated by space between the sets of groups. Two sets of site units (ecosystem types) are ordered along
gradients from moderately fresh to moderately dry (units 1–10) and from fresh to wet and somewhat poorly drained (units 11–15).
Source: After Sebald (1964).
0005430852.INDD 248 11-14-2022 15:26:57

the climax community develops and perpetuates itself (Cooper et al. 1991, p. 134). Habitat types
identify areas of similar climax vegetation (biota), not necessarily areas of the same ecosystem
type. The climax vegetation upon which the classification is based is called a “plant association.”
Classically in the Rocky Mountains, the habitat type system determines late-successional
overstory dominants (termed series) that occur along an elevational gradient from grassland to
alpine tundra (Figure 11.5). Understory species (shrubs and herbs) in each series are used to
identify many different plant associations, and the type is named for the potential climax
community type or plant association (e.g., the Pseudotsuga menziesii/Calamagrostis rubescens
habitat type). The series level is typically denoted by the most shade-tolerant, late-successional
tree species adapted to the site and a dominant or indicator understory species of the plant
association. Series order habitat types along an elevational gradient that approximates climate.
For example, ponderosa pine occupies areas that are warmer and drier than areas where Douglas-
fir or Engelmann spruce and subalpine fir are dominant (Figure 11.5). A finer level, termed
phase, is used to designate major within-type variation in understory vegetation associated
with geographic, topographic, or edaphic features (Youngblood and Mauk 1985). The ubiquitous
Pseudotsuga menziesii/Calmagraostis rubescens habitat type in Montana has four phases which
range geographically from northwestern to southwestern Montana and are found at elevations
from 823 to 2377 m (Pfister et al. 1977).
Habitat types and other classification systems based on vegetation are widely used in forest
management on public lands for timber, wildlife, range, and watershed management (Lay-
ser 1974; Ferguson et al. 1989). Specifically, such systems have been used for growth and yield
evaluation (Monserud 1984; Stage 1989), recreational use studies (Helgath 1975), forest protec-
tion (Arno 1976), fire effects (Fisher 1989), fire management (Arno and Fischer 1989), and
Larix lyallii
Pinus albicaulis
Tsuga mertensiana
Alpine tundra
Timberline
h.t.s.
Abies lasiocarpa
Tsuga beterophylla
Pinus monticola
Thuja plicata
lasiocarpa
Tsuga
series
Abies
mertensiana
Pseudotsuga menziesii
Abies grandis
Larix occidentalis
climax
Pinus contorta
Picea
Pinus ponderosa
Tsuga beteropbhylla series
Tbuja plicata series
Abies grandis series

Picea series
Pseudotsuga menziesii series
Pinus ponderosa series
Grassland
F I G U R E 11 .5 Distribution of forest trees along an elevational gradient in the Rocky Mountains of

northwestern Montana. Arrows show the relative elevational range of each species; the solid portion of the
arrow indicates where a species is the late-successional (climax) dominant, and the dashed portion shows
where it is early successional (seral). Source: Pfister et al. (1977) / United States Department of Agriculture /
Public Domain.
natural area preservation (Schmidt and Dufour 1975; Wellner 1989). Recent research has advo-
cated the use of habitat types in forest conservation planning, particularly in Europe (Kovač
et al. 2016, 2020; Culmsee et al. 2014). The use of habitat types and communities has provided
an important framework for management and has contributed to inventorying, describing,
and classifying forested lands. At least 127 classifications were developed primarily on US
Forest Service lands from 1952 to 1987 (Wellner 1989), and 909 habitat types, community types,
and plant communities were described for the Rocky Mountains alone by the mid-1980s
(Alexander 1985).
The habitat type approach arose from a very old tradition of vegetation classification
(Daubenmire 1989). Early foresters focused their attention narrowly upon forest cover types and
their delineation, with an emphasis on stands rather than land such that a vegetation taxonomy
was very useful. Habitat types are based on the phytosociological classification approach devel-
oped by Cajander (1926) and transplanted by his students to Canada (Pfister and Arno 1980).
Cajander’s approach is taxonomic, developing a classification of vegetation from plot samples and
then applying the classification to other sites. Habitat types remain useful especially where forest
land management problems concern particular cover types that are to be maintained, logged, and
regenerated with the same species (Rowe 1984b). Using a taxonomic key to habitat types, and
supplemented by selected topographic and soil properties, each forest stand can be classified and
assigned an appropriate prescription for management.
Moreover, habitat types provide a framework for extensive management of large areas,
although increasing the area in question naturally increases the number of taxonomic units. For
example, a reworking of approximately 4.9 million hectares of northern Idaho, including the area
of the Daubenmires’ (1968) original study, yielded a five-fold increase in taxonomic units. It is
notable that the large extent of a habitat type—including all land areas potentially capable of pro-
ducing a similar plant association at climax—incorporates considerable variation in physiography
and soils. Hanks et al. (1983), working in northern Arizona, report:
Two land areas with obvious differences in measurable environmental factors may fall within the same
habitat type, if they are equivalent with respect to plant requirements. For example, a habitat type can
occur on two sites with different soils and climatic regimes when greater moisture holding capacity of soils
on one site compensates for a drier climate. Because of compensating factors a vegetation type may occur
on different physical environments.
Daubenmire (1976) also cites an example from the northern Rocky Mountains: “. . .the
Pseudotsuga/Physocarpus forest occurs on steep north-facing slopes at its lowest altitudinal limits,
moves onto zonal soils at intermediate elevations, then onto the shallow soils of steep south-facing
slopes at its highest limits” (Daubenmire and Daubenmire 1968). The important point is that with
its basis in vegetation alone, a given habitat type may not necessarily identify a landscape ecosys-
tem type that is homogeneous in physical environment. Indeed, soil classifications have been
found to poorly align with habitat types (Daubenmire and Daubenmire 1968; Neiman 1988), in
part because separate disciplines have different goals, assumptions, and methods. As Daubenmire
(1970) recognized, the soil-site properties that play key roles in patterning vegetation are not
among those emphasized in soil classification. Neiman (1988) concluded in his broad study of
northern Idaho that further delineation of habitat types, based on soil variation, would permit
greater accuracy in predicting site capabilities and response to disturbance.
APPLICATIONS AND LIMITATIONS OF VEGETATION

In Site evaluation and classification, vegetation integrates the effects of many interacting
factors, and key species may indicate specific site conditions. Like any single ecosystem compo-
nent, however, the use of vegetation has limitations, as discussed by Rowe (1984b). First,
Environmental Factors as a Measure of Site 251
v egetation is strongly controlled by macroclimate and microclimate, and thus its use should
occur within regional ecosystem hierarchies that utilize climate (Albert et al. 1986; Bailey 2009).
Second, vegetation is highly sensitive to disturbance and thus having information about his-
toric disturbance regimes is important. Existing vegetation and potential natural vegetation
need to be understood for land-management applications (Eshelman et al. 1989). Third, vegeta-
tion is floristically complex and may require the user to identify the entire complement of
vascular plants and even mosses and lichens. Fourth, vegetation changes over time, and succes-
sional patterns should be understood, especially where original vegetation has been heavily
disturbed by humans (Neiman 1988). Fifth, vegetation varies widely in occurrence, abundance,
coverage, and biomass. Plants that are lacking may be due to chance or historic causes, such
that reliable sampling methods are necessary in assessing plant indicator value. Finally, vegeta-
tion varies greatly in its vertical layering, which is of major significance in animal ecology and
wildlife management.
ENVIRONMENTAL FACTORS AS A MEASURE OF SITE

Often trees or understory plants are less useful than site factors in assessing forest productivity.
Such situations might include those where vegetation is lacking or heavily disturbed. Examples are
non-forested or agricultural lands slated to be reforested; areas recently subjected to wildfire, log-
ging, heavy grazing, or other severe disturbances; and areas to be converted between forest types.
In these cases, site productivity must be estimated from an analysis of the physical environment
rather than from the vegetation.
Physical factors, such as climate, physiography, and soil as discussed in previous chapters,
may be used singly or in combination. The usefulness of any factor depends on its consistent, sim-
ple, and inexpensive measurement, as well as its high correlation to forest productivity. Many site
factors are not useful as site indices because few data are available for the factors or because it is
difficult or expensive to obtain these data. Other factors are not used because they are not sensitive
enough as a measure of site quality. The most useful environmental factors are those that are “lim-
iting,” or in short supply regarding the demand by forest trees, so that small changes in the supply
of the factor will result in measurable changes in tree growth. Thus, different environmental
factors are useful in different regional and local forest ecosystems.
Climate factors often provide a rough index of productivity among adjacent forest regions or
among altitudinal zones within a geographic region. Climate is a much stronger driver of genetic
differences over a species’ geographic and altitudinal range than is soil. Temperature and precipi-
tation data are most often used to compare forest growth in various geographic regions or altitudi-
nal zones, assuming similar soil conditions or at least assuming soil conditions closely related to
climate. Tree growth will vary widely within a given climatic region depending on physiographic
and soil conditions; soil factors are therefore particularly useful in local studies of site quality.
Increasing the number of factors used in creating an index of forest productivity will improve
its assessment, because the site portion of a forest ecosystem is defined as the sum total of many
factors. Unfortunately, increasing the number of factors typically increases the cost and the time
needed to complete the analysis. Where site classifications are developed for application to a broad
area, multiple factors should be used at both regional and local scales.
CLIMATIC FACTORS
The crowns and boles of trees live in the air and are affected by it, and thus climate has a strong
effect on tree growth. Macroclimatic factors have long been used to distinguish differences
among major forest regions (Schlenker 1960; Rowe 1972; Findlay 1976; Ecoregions Working
Group 1989). However, long-term climatic data from forest sites are often very difficult to obtain.
In addition, data interpretation for individual interrelated variables is problematic, and vegeta-
tion itself is often a more easily measured and useful integrator of the complex of climatic factors.
Most site-quality evaluations for management decisions take place at regional scales (Chapter 2)
where the average climate may not vary widely, although local climate may vary significantly
from place to place within the management unit. Local climate has seldom been used in site
evaluation, in part because other variables such as land-use and forest management history have
their own effects on forest growth. In addition, the control of local climate on local topography
and soil suggests that an ecosystem classification based upon topography and soil implicitly
assumes a local climate classification as well. For example, the same factors that make a soil very
well drained are apt to insure good cold-air drainage, and a very poorly drained soil is apt to
result from a topography that inhibits the drainage of cold air as well as of soil water. Thus, very
poorly drained sites are also characterized by climatic factors such as temperature extremes and
a short growing season.
PHYSIOGRAPHIC LAND CLASSIFICATION

Physiography directly or indirectly affects many other key ecosystem factors, and thus it provides
a framework for managing ecosystems at both regional and local scales. As emphasized in
Chapter 8, the shape, parent material, and location of specific landforms influence the reception
of solar energy and the runoff and retention of precipitation (Rowe 1984b; Swanson et al. 1988). In
the mid-south of the United States, Smalley (1991) developed and applied a landform-based taxo-
nomic system for about 12 million hectares of the Cumberland Plateau and Highland Rim physio-
graphic provinces of Virginia, Tennessee, Kentucky, and Georgia. Following the system of land
stratification of Wertz and Arnold (1972, 1975), Smalley classified sites on these landscapes using
five hierarchical levels (province, region, subregion, landtype association, landtype). Landtypes,
the most detailed level (spatial scale 1:10 000 to 1:60 000), are described by nine factors of geogra-
phy, geology, soil, and the most commonly occurring trees and shrubs. Notably, slope position and
aspect were particularly important in distinguishing landtypes characteristic of highly dissected
terrain in the Cumberland Mountains of Kentucky, Tennessee, and Virginia (Figure 11.6;
Smalley 1984).
Land classification based on physiography is particularly appealing because of a general lack
of soil survey information and a long history of human disturbances that have drastically changed
how natural vegetation relates to site. As a result of human disturbances, existing forests vary widely
in their stand conditions, productivity is reduced far below its potential, and few stands are appro-
priate for a direct measure of site productivity. In terrain with high topographic relief, landforms
may correlate better to tree growth than soil series or the existing tree vegetation. In such a situation,
landtypes become the basic unit of land management rather than existing forest stands whose
boundaries may be artifacts of past land use. If necessary, this broad landform base facilitates sub-
division into finer ecosystem types based on local topography, soil, and ground-cover vegetation.
PHYSIOGRAPHIC AND SOIL FACTORS: SOIL-S ITE STUDIES

Many researchers have examined the potential of using physiography and soil to predict forest site
quality. For areas where physiography (specific landforms, topographic factors; see Chapter 8),
soil, and stand conditions are highly variable, such an approach would be very advantageous. Like-
wise, using physical factors rather than biota would be necessary where trees are lacking, the area
is stocked with unwanted species, or the area is dominated by trees unsuited for site-index mea-
surements. Carmean (1975, 1977, 1996) cites over 180 soil-site studies for the United States and
Canada and discusses their applications and limitations; Burger (1972) cites many additional
studies in Canada.
N 27
26 27
23
23 22 28
22
28
22
23 25 22
2 24
24
26
23
24
25
25
LE,
24 SHA
A N IAN TONE,
LV S
NSY AND
19 PEN TONE, S
25 I LT S
S
AL.
& CO
ALLUVIUM
COLLUVIUM
F I G U R E 1 1 .6 Land types characteristic of dissected forest terrain of the Middlesboro syncline and
Wartburg Basin–Jellico Mountains of Kentucky, Tennessee, and Virginia. Legend: 2, shallow soils and
sandstone outcrops; 19, mountain footslopes, fans, terraces, and stream bottoms; 22, upper mountain
slopes–north aspect; 23, upper mountain slopes–south aspect; 24, colluvial slopes, benches, and coves–north
aspect; 25, colluvial slopes, benches, and coves–south aspect; 26, surface mines; 27, narrow shale ridges,
points, and convex upper slopes; 28, broad shale ridges and convex upper slopes. Source: Smalley (1984) /
Depending upon the nature of the specific site, many individual soil and physiographic
factors may be correlated with the site index of the desired forest species. Soil-site studies involve
measuring many soil and site variables (e.g., soil depth, texture, and drainage class; slope posi-
tion; aspect) and relating them with multiple regression analyses to tree height or site index.
Combining these with other soil and topographic factors has produced useful formulas to estimate
site index. In a classic study, Zahner (1958) restricted regressions to soil groups within a limited
geographical region and related the site index of two southern pines to surface soil thickness, the
proportion of clay and sand in the subsoil, and the slope percentage. Equations derived from such
soil-site studies were then used to develop tables and graphs for estimating site index in the field.
Similarly, combinations of soil and site variables in successful soil-site studies explained 65–85%
of the variation in tree height or site index (Carmean 1975). A disadvantage of this approach,
however, is that many soil and site variables are difficult or tedious to measure in the field, and
thus somewhat less precise equations are often developed using variables that are most eas-
ily measured.
Two important caveats are applicable to soil-site studies. First, many investigators have read
causal relationships into regressions based upon soil and topographic factors when the data only
support correlation. Importantly, such correlations may reflect a causal relationship attributable to
another unmeasured soil characteristic rather than the data at hand. Second, the dependent vari-
able of the regression equations is often site index as read from harmonized site-index curves. As
discussed previously, such values are suspect by the very nature of the method used to construct
the site-index curves, and thus caution is necessary in using soil-site equations.
The growth potential of forest trees appears to be chiefly affected by the soil volume
occupied by tree roots and the availability of water and nutrients in this soil space. Therefore, the
effective depth of the soil—the depth of the portion of the soil that is capable of being occupied
by roots of the tree—is of primary importance. The effective soil depth may be limited by a layer
that prevents further downward growth of roots, such as bedrock near the surface (Green and
Grigal 1979, 1980), the position of the water table during the growing season, a coarse, dry soil
stratum, or a highly compact and impervious stratum such as a well-developed hardpan. Soil
factors have been correlated with site quality using many measures of effective soil depth
(Coile 1952; Carmean 1975). The most important soil factors are the depth of the A horizon above
a compact subsoil, the depth to the least permeable layer (usually the B horizon), the depth to mot-
tling (indicative of the mean depth to restricted drainage), and thickness of the soil mantle over
bedrock. All these measures are important when soils are shallow but are less important for deep
soils where downward root development is unimpeded. In addition to effective soil depth, other
soil profile characteristics that affect soil water, soil drainage, and soil aeration are also important
for site quality, particularly soil texture and structure of the least permeable horizon (again usu-
ally the B).
The physical properties of the soil that govern soil-water and aeration are often closely related
to topographic position of the site (slope position) and other landform factors, as is microclimate.
Landform is the chief correlate of the patterns of soil and vegetation and thus site productivity,
because it subsequently influences local climate near the ground, belowground climate induced by
soil water, and nutrient regimes via parent materials and soil-water availability. When the relation-
ships between landform and soil are well known, topographic site can be quickly recognized and
evaluated using aerial photographs and topographic maps without the necessity of soil measurement.
For example, an index of forest site quality was developed for use in the Ridge-and-Valley physio-
graphic province of the central and southern Appalachian Mountains using only aspect, slope
percentage, and slope position (Meiners et al. 1984). Subjectively ranking field sites from 1 to 5 for
each of these three variables provided a simple index (3 = lowest quality; 16 = highest quality) that
provides a rapid evaluation of relative site quality. This physiography-based index is highly corre-
lated with site index of oak in the Allegheny Mountains of West Virginia and Maryland and with
m.a.i. of trees on steep slopes in the Ridge-and-Valley province of Virginia (Ross et al. 1982).
Relative elevation, aspect, slope position, and degree of slope have repeatedly produced useful
site relationships. Studies of site quality for oak forests in the Appalachian Mountains and the
Appalachian Plateau (Trimble and Weitzman 1956; Doolittle 1957; Carmean 1967; McNab 1987)
have shown close relationships between site quality and relative position between ridge top and
cove, aspect, and degree of slope. Equations based on topographic features alone explained more
than 75% of the variation in black oak height in southeastern Ohio (Carmean 1967). Likewise, topo-
graphic position was a better predictor of site quality than height–age site-index curves for mixed-
oak forests of western Maryland (Sturtevant and Seagle 2004). Topography and site quality are
closely related because topography is correlated with important soil features such as A horizon
depth, subsoil texture, stone content, organic-matter content, and nutrient availability (Figure 11.7).
The relationship of topography with microclimate is also critical, especially in hilly terrain. Northeast
aspects and lower slopes usually have cool, moist microclimates and thus are more productive sites;
southwest aspects and upper slopes and ridges have dry and warm microclimates and hence are
usually less-productive sites. In the Appalachian Mountains, the relationship of site quality to aspect
for mixed upland oak forests resembles a cosine curve (Figure 11.8; Lloyd and Lemmon 1970).
Notably, the predictive value of landform and topographic variables for tree height and site
quality is fairly inconsistent among studies, partly because of different methods of scoring or
measuring the variables. McNab (1987, 1991, 1993) developed indices to quantify slope position
and local landform shape. Site index of tulip tree was significantly correlated to landform index at
four sites in the Blue Ridge physiographic province (McNab 1989), suggesting that microsites may
be an important source of spatial site variation. Such microsites represent local ecosystem types
where the beneficial effects of leaf litter accumulation and decomposition (Stone 1977; Welbourn
80
75
Site index, feet

75
70
Site index, feet
70
65
65
60
45
NE
60
90 E
0 &
N
5
31 SE .
Slo 20 & az
pe
ste NW ect,
40 0 p
ep
ne 27 S As
ss, &
% W
60 225
SW
F I G U R E 11 .7 Relationship between aspect, slope steepness, and site index for black oak growing on
medium-textured, well-drained soils. Site index increases from southwest-facing slopes to northeast-facing
slopes. These increases are very pronounced for steep slopes, but site index increases related to aspect are
relatively minor on gentle slopes. For southwest-facing slopes, site index decreases drastically with increased
slope steepness, whereas it increases slightly on northeast slopes as slopes become steeper. Source: Carmean
(1967) / with permission of John Wiley & Sons. Reprinted from Soil Science Society of America Proceedings,
Vol. 31, p. 808, 1967 by permission of the Soil Science Society of America.
110
Site index ratio, %
44
78 58
100
72 60
92
73
90 83
N NE E SE S SW W NW
Aspect
F I G U R E 1 1 .8 Productivity curve with a site-index ratio over aspect; based on 560 soil-site index plots on 27
soil series in mixed upland oak forests of the Appalachian Mountains. The site-index ratio is the ratio of the
plot site index to the average site index of all plots of that soil series. Source: Lloyd and Lemmon (1970) /
with permission of Oregon State University. Reprinted with permission from Tree Growth and Forest Soils,
© 1970 by Oregon State University Press.
et al. 1981), combined with gravitational movement of subsurface water on mountain slopes
(Hewlett and Hibbert 1963; Dwyer and Merriam 1981), create especially favorable nutrient and
soil-water conditions.
Minor variations in topography may be highly important in very flat locations for various rea-
sons depending on the nature of the site. On the dry, flat, sandy, nutrient-poor glacial outwash plains
of northern Lower Michigan, small changes in elevation (as little as 15 m) or variations in landform
shape create subtle changes in soil texture because of the way the parent material was deposited by
fluvioglacial processes (Kashian et al. 2003). These subtle soil differences result in variations in
nutrient and soil-water conditions that produce marked variation in the height growth of jack pine
that is directly predictable by landforms across the region. In areas where the water table is close to
the soil surface, small changes in topography reflect the effective depth of the soil over poorly aerated
lower horizons. On southern river bottomland oak sites, topographic features varying only centime-
ters in elevation are correlated with differences in the silt and clay content of the soil, with flooding
conditions, and with soil aeration, and thus are highly related to site quality (Beaufait 1956).
SOIL SURVEYS
Soil surveys, which emphasize soil series and phases, were developed for application in agriculture
rather than forest ecosystem management (Chapter 9). Soil surveys provide maps and a taxonomic
classification of forest soils, but have not been precise in estimating forest site quality (Grigal 1984). In
part, these failures are attributed to the basic differences between soil taxonomic units and soil map-
ping units. Typically, soil taxonomic units include too much variation in forest productivity, as esti-
mated by site index, to be useful (Rowe 1962; Jones 1969; Carmean 1970a, 1975; Grigal 1984; Nei-
man 1988). For example, loblolly pine site index ranged from 59 to 105 on soils of the Ruston series
(Covell and McClurkin 1967). Similar variation in site index within soil taxonomic units also has been
reported for numerous species in eastern hardwood forests (Carmean 1970a, 1975). Soil series prove
more satisfactory when they incorporate specific soil, landform, and local topographic factors that are
closely related to forest productivity (Richards and Stone 1964; Carmean 1967, 1970a; Grigal 1984).
MULTIPLE-FACTOR METHODS OF SITE AND ECOSYSTEM CLASSIFICATION

Evaluating site quality with single-factor approaches, such as site index and vegetation indicators,
physiography, and soil, uses only individual components of the ecosystem. Such an approach is
troublesome because site quality is the sum total of factors affecting the land’s ability to produce
forests. Site productivity is always better estimated by including more factors in the process, and a
multifactor approach improves our ability to understand ecosystem processes as the basis for
managing forests. This is why intensive multifactor methods have been developed and used suc-
cessfully in Europe for over 85 years, and similar methods have more recently proliferated widely
in Canada and the United States.
Many multifactor methodologies have been developed, resulting in a wide variety of termi-
nology used to describe the ecosystem units within them. In practice, a common feature of multi-
factor methods is the consideration of a given ecosystem in two parts: the physical ecosystem factors
of a specific geographic place (site) and the plants and animals occurring there (biota). The geo-
graphic space may be described with terms such as land type, site type, site unit, geosystem, physio-
graphic site, or forest site. The biota is often described with terms such as vegetation type or cover
type. Ecosystem units themselves (i.e., the volumetric land units integrating site and biota) are also
described with various names, from the broadest (ecodomain) to the finest (ecosystem type). Nota-
bly, geographic site terms such as site unit, site type, and total site are often used to indicate landscape
ecosystems because these spatial units have the same or similar properties and potential vegetation
regardless of present vegetation. Despite the names of the various methods and their terminologies,
the common focus among the examples presented in the following text is on geographic area or site.
Multiple-Factor Methods of Site and Ecosystem Classification 257
ECOSYSTEM CLASSIFICATION AND MAPPING IN BADEN-W ÜRTTEMBERG

The first operational and systematically applied multi-scale, multi-factor system of ecosystem
classification and mapping was developed in the southwestern German state of Baden-Württemberg
(Barnes 1984, 1996). The area is extremely diverse in climate and geology, including the Rhine
River Basin, adjacent Black Forest, a limestone plateau (Swabian Alb), and glaciated terrain
stretching nearly to the Swiss Alps. Much of the original forest of beech and oak was replaced by
Norway spruce monocultures now often several generations old. An ecological approach integrat-
ing geology, physiography, climate, soil, vegetation, and forest history was developed to classify
and map forest ecosystems as a way to diversify and improve multiple-use land management. This
detailed and systematic approach is based upon classification and mapping at multiple scales
(Figure 11.9). Once mapped, differences in local ecosystem types (termed “site units”) are evalu-
ated with the growth and biomass production of major tree species. Detailed silvicultural recom-
mendations are then made for ecosystems at regional and local scales. This system, applied for
almost 75 years, emphasizes natural regeneration of native species (Mühlhäusser and Müller 1995)
and is used in long-term planning, forest and wildlife management, and for preservation of unique
ecosystems (Dieterich et al. 1970; Arbeitsgruppe Biözonosenkunde 1995). Amazingly, the interdis-
ciplinary team who developed this multi-scale and multi-factor approach integrated ecosystem
components at a time in the 1940s when disciplinary fields of geology, climatology, soil science,
Regional classification
in
Growth Areas subdivided
in Growth Districts
Local classification
in
Ecosystem Types
(Site Units)
Accumulation of
local silvicultural
experience
Ecosystem mapping
Basic
scientific
investigations
Evaluation of
growth and productivity
Silvicultural evaluation
Comprehensive summary
for each Growth Area
and
comparisons between
different Growth Areas
F I G U R E 1 1 .9 Model of the multi-scale and multi-factor system used in Baden-Württemberg, southwestern

Germany to classify, map, and evaluate regional and local landscape ecosystems.
and vegetation science were pursued for their own sake. This team approached the practical appli-
cation of these disciplines in forestry as the simultaneous integration of ecosystem components in
classification and field mapping (Schlenker 1964; Mühlhäusser et al. 1983).
A major emphasis of this approach lies in regionalization, where a regional framework is
provided that allows subdivision of broad ecosystems into successively smaller ecosystem units. In
Baden-Württemberg, the synthesis of factors at the regional level leads to a division of the state
into major landscapes (termed “growth areas”), which are in turn subdivided into minor land-
scapes (“growth districts”) (Figure 11.9). This regional framework limits sweeping generalizations
and prescriptions that were often made for species over wide areas having vastly different environ-
ments and histories. At the local level (within districts), individual ecosystem types (site units) are
identified, mapped at a spatial scale of 1:10 000, and described. Ecosystem types are distinguished
and mapped in the field using local differences in physiography, microclimate, soil factors of tex-
ture, structure, pH, depth, and water and nutrient status, and overstory and ground-cover vegeta-
tion. Ecosystem types include individual sites that are similar in physical site characteristics, silvi-
cultural and management potential, disturbance regimes, incidence of disease and insect attack,
and growth rates and tree biomass.
The German team also pioneered the use of ecological species groups (described in the ear-
lier text) to distinguish and characterize local ecosystem types (Schlenker 1964; Sebald 1964; Diet-
erich 1970). In the Upper Neckar growth district of Baden-Württemberg (Sebald 1964), 24 ecolog-
ical species groups were used to differentiate 30 ecosystem types along soil-water gradients
(Table 11.1). Several plant species comprise each group, which indicate certain site factor complexes
because of similar environmental requirements or tolerances. Some species groups have a wide
ecological amplitude, such as the Milium effusum group, whereas others have a narrow amplitude,
such as the Aruncus sylvester group (Table 11.1). Ecosystems are distinguished by the presence or
absence of groups and the relative abundance of the species in the respective groups. Units are
clearly differentiated along a gradual trend of differences in species groups when they are arranged
along two soil-water gradients (from moderately fresh to moderately dry, units 1–10, and from
fresh to wet, units 11–15). Importantly, the indicator value of each group is reliable only within the
rooting zone of the species in the group.
In field mapping, ecological species groups are used simultaneously with physiographic and
soil characteristics to delineate ecosystem boundaries. In part, this is because some units are well
defined by vegetation in the species groups alone, such as those units at opposite ends of the gra-
dients. On the other hand, adjacent site units (e.g., units 1 and 2 in Table 11.1) may have similar
species groups and would need to be differentiated in the field using soil and topographic features.
For example, unit 1 is a podzolized loamy sand on level terrain, whereas unit 2 is a podzolized sand
on a moderately steep slope of south to southwest aspect. The combined technique using soil and
physiography as well as vegetation is always faster and more reliable.
APPLICATIONS OF MULTI-FACTOR METHODS IN THE UNITED STATES

AND CANADA
Ecosystem Classification and Mapping in Michigan The Baden-Württemberg approach
has been modified and applied in Michigan at both regional and local levels. Regional ecosystems
were classified based mainly on climate and physiography, which are the dominant physical factors
driving ecosystems at regional scales. First, climates in Michigan were classified using separate
classifications for growing season temperature, winter temperature, and growing season precipita-
tion variables (Denton and Barnes 1988). The climate classifications were then combined with
physiographic and soils data to produce a general hierarchical classification and map of regional
landscape ecosystem regions and districts in the state at three spatial scales: regions, districts, and
subdistricts (Albert et al. 1986; Figure 2.10 in Chapter 2). This approach has been extended to
Wisconsin and Minnesota at the regional and district levels (Albert 1995).
Local landscape ecosystem types have been classified and mapped in Michigan for both old-
growth forests (Barnes et al. 1982; Pregitzer and Barnes 1984; Spies and Barnes 1985a; Simpson
et al. 1990) and highly disturbed landscapes (Archambault et al. 1990; Zou et al. 1992; Pearsall 1995;
Walker et al. 2003). A physiographic diagram illustrating the correspondence of landform, soil, and
vegetation for each ecosystem type in part of the McCormick Experimental Forest in Upper Michi-
gan is presented in Figure 11.10. Ecosystem types in this old-growth forest are relatively homoge-
neous in site conditions and groundflora. A map of the area illustrates the fine-scale occurrence of
ecosystems in formerly glaciated landscapes (Figure 11.11), where ecosystems recur in intricate but
predictable patterns. For example, rocky and fire-prone ridges support white pine and northern red
oak; outwash plains and slopes support northern-hardwood forests, and wetlands of various kinds
are found along the Yellow Dog River. The major ecosystem types occupy over 95% of the area and
Sugar Maple Black Spruce

Gymnocarpium Ledum
10 Maple-YB-Conifer
2 7
EXPOSED Clintonia
DEEP SHALLOW ROCKY RIDGE 8
VW-DR. SAND VW DR. SAND SHALLOW
AND GRAVEL ON ON EXPOSED SITES W-DR. SAND
LEVEL TERRAIN ON PROTECTED SITES
4 18
DEEP W-DR. SAND INFERTILE SWAMP
ON MODERATELY ON PEAT
SLOPING TERRAIN 3
DEEP IMPF-DR. SAND 17
BEDROCK ON GENTLY
2 SLOPING TERRAIN VERY
cm Oe, a 10 INFERTILE
E Oe, a SWAMP
cm ON PEAT
Bh E 4
20
cm Oe, a
20 Bs E
40 3
BC
40 20 Bh Oe, a 17
cm E cm
60 Bs Oi
2BC 60 40
Bh
BC 20 20 Oe
80 Bs
80 60
2E´ 40 40
100 BC
100 80 Oa
60 60
120
120 100 2C
2C
80 80
140
140 120
100 100
Water
140
120 120
140 140
F I G U R E 1 1 .1 0 Physiographic diagram of selected landscape ecosystem types of the McCormick Experi-

mental Forest, Ottawa National Forest, in Upper Michigan. Ecosystem types (identified by number) are
characterized here by late-successional overstory species and characteristic ecological species groups, soil
profile, and their physiographic position in and along the landscape gradient. Source: Barnes et al. (1982).
Reprinted with the permission of the Society of American Foresters.
16
Yellow Dog
River 19
17 19
17
19 3 3
16 18 6 4
16 16 19 3
18
4 3 3 17 3
6
2 16 19
4 2 4
4 18 2 2
4 3 19
11 16 17 2 6
10 4 3
5 5 17 17
3 12 6
11 10 5 16
2 6 17 4 11
10 5 10 4 3
Transect
10 11
8 3 2 3 3 16
3 4 6
10 5 10 19
10 19
4 16 10 19 2
10
11 10 5 4 2
6 7 11 12 4 3 12 6
11
15 7 13
4 3 4 3
11 10 4
2 5
6 8 14 11
4 2 5
12 12
6
N
1/4 1/2 MILE Natural Area Boundary
1°
Ecosystem Type 3 8
20 40 CHAINS Section Lines
River
.40 .80 KM.
Stream
F I G U R E 1 1 .1 1 Map of local landscape ecosystem type for part of the McCormick Experimental Forest,
Ottawa National Forest, in Upper Michigan. A transect line running roughly west to east illustrates the
pattern of ecosystems; see also Figure 11.10 for their relative physiographic position to one another. On the
west, an extensive flat, infertile outwash plain (type 2) is dominated by stunted sugar maple. Just to the east,
a crystalline-rock ridge runs approximately northwest to southeast. Ecosystem type 10 occurs along the ridge
top, and thin-soil types 7 and 8 are associated with it. On the northeastern slopes of the ridge, types 4 and 5
occupy mid and lower slope positions, respectively. Adjacent to the Yellow Dog River are wetland ecosystems
of acid (type 18) and circumneutral (type 19) soils. On lower slopes, wet-mesic ecosystem type 3 often occurs
adjacent to the swamps. Across the river on hilly, ice-contact terrain occur alternating steep sandy southwest
slopes dominated by eastern white pine and hemlock (type 6), and steep, sandy northeast slopes dominated
by sugar maple (type 4). One chain equals 66 ft. or 20.1 m. Source: Barnes et al. (1982) / with permission of
Oxford University Press. Reprinted with permission of the Society of American Foresters.
can be grouped into four ecosystem groups for management purposes: sandy outwash and ice-contact
terrain (types 2, 4, and 6), fertile lower slopes and valleys (types 5 and 11), rocky sites with shallow
soils (types 7, 8, 9, and 10), and wetlands and wet-mesic adjacent slopes (types 3, 16, 17, 18, and 19).
Mapping and distinguishing ecosystems at the resolution shown in Figure 11.11 are often
perceived as too expensive for development in North America, where management that might jus-
tify its cost is less intensive than in Europe. Long traditions of intensive land use are much different
in Europe than those in North America. Nevertheless, this fine level of detail is also useful in
understanding the ecological diversity of landscapes, to which the diversity of plants and animals
(biodiversity) is closely related (Lapin and Barnes 1995; Pearsall 1995). As biodiversity is an
increasingly important driver of forest management (Chapter 14), particularly as land use inten-
sifies in North America, understanding the fine-scale resolution of landscape ecosystems will
inevitably prove valuable. Notably, however, the ecosystem approach developed in southwestern
Germany and applied in Michigan is applicable at several scales to meet appropriate management
objectives (e.g., see Kashian et al. 2003; Walker et al. 2003), and fine-scale landscape ecosystems
represent only one part of the classification system.
Ecosystem Classification in the Southeastern United States In the southeastern

United States, landscape ecosystem classification is recognized as critical to managing ecosystems
(Jones and Lloyd 1993). A multi-factor approach similar to that described in the earlier text has
been applied in the Blue Ridge province of the southern Appalachian Mountains (McNab 1991)
and in South Carolina (Jones 1991). Using a physiographic map of South Carolina (Meyers
et al. 1986) as the regional hierarchical framework, landscape ecosystems were distinguished and
mapped using a combination of landform, soil, and vegetation (Jones 1991). A physiographic dia-
gram of the model for the Interior Plateau subregion of the Midlands Plateau region of the Pied-
mont province (Figure 11.12) illustrates five ecosystems named by their landscape position in rela-
tion to soil-water availability (xeric to mesic) and the vegetative associations occurring on these
sites. Maps of these ecosystem types provide the basis for interpreting productivity and silvicul-
tural management, but they also provide a framework for predicting potential habitat for
endangered species, wetlands delineation, and ecological restoration.
A similar landscape ecosystem classification was developed for a portion of the Gulf Coastal
Plain in southwestern Georgia (Goebel et al. 2001). Most of the upland forested area in the region
is on karst topography dominated by fire-dependent longleaf pine, with mixed hardwoods along
riparian areas. Seventeen ecosystems were identified and classified into six hierarchical levels,
using physiographic systems (fluvial or karst) and specific landforms (topography, landforms, and
terrain shape) at the highest three hierarchical levels and soil drainage class, soil texture, and veg-
etation at the lowest three levels. The classification was effectively tested by using it to map ecosys-
tems in the field (Palik et al. 2000). Mapping the landscape ecosystems illustrated that less than
half of the landscape supported plant communities that could be considered reference communities
SUB-
SUB- XERIC
XERIC
INTER-
XERIC SUB-
MEDIATE
XERIC
30 cm
SUB-
INTER-
MEDIATE SUB-
MESIC C 60 cm
MESIC
30 cm 90 cm
MESIC C SC
60 cm C
SC
90 cm
C
SC
SCL
SCL CL
SL
CL alluvium SCL N
SCL CL
CL
MESIC → AMERICAN BEECH - NORTHERN RED OAK - CHRISTMAS FERN
SUBMESIC → NORTHERN RED OAK - WHITE OAK - WILD GERANIUM
INTERMEDIATE → WHITE OAK - NORTHERN RED OAK - FALSE SOLOMON’S SEAL
SUBXERIC → WHITE OAK - SCARLET OAK - DEERBERRY
XERIC → POST OAK - BLACK OAK - LOWBUSH BLUEBERRY
F I G U R E 1 1 .1 2 Physiographic diagram and landscape ecosystem classification model for the Interior
Plateau subregion of the Midlands Plateau region of the Piedmont Province, South Carolina. Soil designa-
tions: C, clay; CL, clay loam; SC, sandy clay; SL, sandy loam; SCL, sandy clay loam. Source: Jones (1991) /
Restoration Index
High Priority
Medium Priority
Low Priority
Developed
Reference Condition
Water
4 0 4 kilometers
F I G U R E 1 1 .1 3 Landscape ecosystem map (left) and restoration priorities (right) developed for a portion of
the Jones Ecological Research Center in southwestern Georgia. Despite the fine-grain occurrence of
ecosystems, about 80% of the area found to be of highest priority for restoration occurred within 3 landscape
ecosystems. Source: Palik et al. (2000).
for restoration, but it also provided an appropriate framework for prioritization of the landscape
for restoration based on rarity and levels of disturbance (Figure 11.13). Notably, 80% of the high-
priority sites occurred within only three of the identified ecosystems (Palik et al. 2000).
Ecosystem Classification in the Southwestern United States In portions of the

southwestern United States, exclusion of historically frequent surface fires since European coloni-
zation in forests dominated by ponderosa pine has resulted in the accumulation of fuels and a
subsequent shift to stand-replacing fires (Covington et al. 1994; Chapter 16). Fire exclusion has
also led to the increase in dense stands of small-diameter trees and the reduction of the historically
common open stands dominated by old trees, which has altered understory plant diversity and
wildlife habitat. A study in northern Arizona ponderosa pine forests identified 10 landscape
ecosystems in the region with the objective of providing an ecological framework for restoration
(Abella and Covington 2006b). Most ecosystem types occurred across broad areas on this very flat
landscape and were differentiated by differences in soil texture and carbon content rather than by
physiographic variables, which nevertheless provides important guidance for restoration and
conservation. Although not developed within a regional framework, this work further shows the
value of identifying and understanding landscape ecosystems in a forested landscape heavily
altered by humans.
National Classification Systems in the United States A national hierarchical system

for ecological land classification (ELC) has been developed for the United States using many of the
principles described in the earlier text. With the goal of developing a consistent framework for
application nationwide, the National Hierarchical Framework of Ecological Units (NHFEU;
Cleland et al. 1997) was devised to identify areas at multiple scales that are similar in their capabil-
ities and potential for management. Similar to the Baden-Württemberg approach, the ecological
units were devised based on climate, physiography, soils, and potential natural vegetation, with cooperation
among the various disciplines representing each component encouraged as the system is utilized
(Avers and Schlatterer 1991). The system includes eight hierarchical ecological units based on
decreasingly smaller scales of climate, physiography, soils, and potential natural vegetation: ecore-
gional units (domain, division, and province), subregional units (section and subsection), landscape
units (landtype association, or LTA), and land units (landtype and landtype phase). Strategic
planning and assessment are the main activities at the level of ecoregional units, statewide or
multi-National Forest at the subregional unit level, individual National Forest planning at the
landscape unit level, and specific project management at the land units. Thus, most planning and
management activities within a given National Forest occur at the spatial scales of the LTA and
land unit, within the regional framework of the broader-scale units. In addition to ecosystem map-
ping, land managers in the United States are encouraged to use the NHFEU, at the appropriate
ecological scale, for activities such as resource assessment and management, environmental
analyses, watershed analyses, determining desired future conditions (restoration goals), and mon-
itoring (Cleland et al. 1997).
Ecological Land Classification in Canada Canada has been a leader in the use of
multi-scale and multi-factor ecosystem classification. The Canadian versions of the approach, as
in Europe, are characterized by many different systems and complex terminologies (e.g., Ru-
bec 1992; Wicken 1986), but their broad scale of application sets them apart. Establishment of a
common national system utilizing digital ecosystem mapping databases was completed in 1995 as
part of the country’s official ELC (Marshall et al. 1999). The classification was last updated in 2017,
with hierarchical levels including ecozones (15 units across Canada), ecoprovinces (53 units),
ecoregions (194 units), and ecodistricts (1027 units). For clarity, Canada also has an official
classification for forests called the Canadian Forest Ecosystem Classification System (CFEC 2010),
but it is a classification emphasizing plant associations and communities rather than landscape
ecosystems. The CFEC is part of the broader Canadian National Vegetation Classification System,
which is analogous to the US National Vegetation Classification System described in Chapter 2.
Prior to publication of the ELC, the enormous diversity of Canada’s forests had produced many
different approaches and terminologies for its ecosystem classification and mapping. The diversity
and detail of nine of these approaches, representing eight provinces and northern Canada, are
presented in a single issue of the Forestry Chronicle (Canadian Institute of Forestry 1992). Despite
Canada’s adoption of an official national classification system, we focus in the following text on a
few of the preceding systems to illustrate the diversity of approaches to ELC.
Hills’ physiographic approach Cajander’s method and that of the Zürich-Montpellier School of
Plant Sociology (Lemieux 1965; Burger 1972) heavily influenced early attention to site quality in
Canada, placing the main emphasis on ground vegetation. The pioneering work of Angus Hills in
Ontario in the 1950s and 1960s shifted the focus toward physiography in evaluating site quality, in
particular a total site or ecosystem approach integrating climate, soil, and vegetation on a landform
basis. The idea that landform influenced the distribution of plants and animals as well as local cli-
mate, drainage, and soil formation was novel and unusual at the time (Rowe 1992a). This total site
system (Hills 1952, 1960; Hills and Pierpoint 1960; Burger 1972; Hills 1977; Burger 1993) was
undertaken originally to provide accurate, descriptive resource maps for land-use decisions. Hills
provided a regional ecosystem framework by dividing Ontario hierarchically into 13 regional
landscape ecosystems he called site regions (Figure 11.14; Burger 1993). A site region was further
subdivided into site districts on the basis of relief and type of bedrock or parent material. Hills
used physiographic features as a basic frame of reference because “they remain most easily recog-
nizable in a world of constant change” (Hills 1952). Thus, physiographic site types and forest
types (characterized by both overstory and groundflora) were combined to form total site types
(Figure 11.14). Hills’s basic approach continues to guide ecological classification today specifically
in Ontario (Baldwin et al. 2000; Crins et al. 2009), and also at a national level (Marshall et al. 1999).
Other approaches used in Canada Development of a series of goal-oriented forest ecosystem

classifications intensified beginning in the early 1980s for a broad geographic zone across north-
central Ontario (Jones et al. 1983; Sims et al. 1989; Sims and Uhlig 1992). These classifications
were developed for use as stand-level planning tools and for a specific and predefined range of
uses. They are designed to help managers develop strategies and alternatives, such as prescribed
burning guidelines, estimating susceptibility of a site to erosion, predicting productivity of specific
tree species, and rating habitat for white-tailed deer. Other examples of application are given by
Sims and Uhlig (1992).
Other Canadian systems focus primarily on vegetation types and are especially useful where
human disturbances are minimal, for example, in Newfoundland and Labrador (Meades and Rob-
erts 1992) and British Columbia (MacKinnon et al. 1992). The detailed floristic research of Dam-
man (1964) facilitated forest site classification in Newfoundland. The phytosociological approach
has remained strong in British Columbia, although Krajina and associates have developed a more
inclusive biogeoclimatic approach that also considers climate, soil, and topography (MacKinnon
et al. 1992).
In summary, multiscale and multifactor landscape ecosystem classification and mapping
provide the appropriate units required for extensive and intensive management of forest ecosys-
tems. German methods illustrate that ecosystem classification and mapping can meet the most
intensive (and changing) management and conservation practices even in areas of highly variable
landforms, soil, and climate, and for sites that have a long disturbance history. Today, remote sens-
ing, geographic information systems, and computer-based techniques facilitate systematic ecosys-
tem classification and mapping to provide a basic framework for land management and planning
around the world.
Regional Landform Plants Animals Chance Humans

climate
Natural forest
succession
Depth of Soil Local climate Lesser Forest

material moisture vegetation cover
Physiographic Forest
site type type
(land-type component) (forest component)
Total
site type
(ecological unit)
F I G U R E 1 1 .1 4 Model of the classification of total site types (landscape ecosystems) by Hills’ method in
Ontario, Canada. Source: Hills and Pierpoint (1960) / CAB International.
SUGGESTED
R E A D I N G S
Barnes, B.V., Pregitzer, K.S., Spies, T.A., and Spooner, Rowe, J. S. (1984). Forestland classification: limitations
V.H. (1982). Ecological forest site classification. J. of the use of vegetation. In J. G. Bockheim (ed.),
For. 80: 493–498. Symp. Proc. Forest Land Classification: Experience,
Carmean, W.H. (1975). Forest site quality evaluation in Problems, Perspectives. NCR-102 North Central For.
the United Sates. Adv. Agronomy 27: 209–269. Soils Com., Soc. Am. For., USDA For. Serv., and
USDA Soil Cons. Serv.
Sims, R.A. 1992. Forest site classification issue. Forestry
Chronicle, 68(1):21–120. Rowe, J. S. (1991). Forests as landscape ecosys-
tems: implications for their regionalization and
Grigal, D. F. (1984). Shortcomings of soil surveys for forest
classification. In D. L. Mengel and D. T. Tew (eds.),
management. In J. G. Bockheim (ed.). Symp. Proc.
Symp. Proc. Ecological land classification: applica-
Forest land classification: Experience, problems, per-
tions to identify the productive potential of southern
spectives. NCR-102 North Central For. Soils Com., Soc.
forests. USDA For. Serv. Gen. Tech. Report SE-68.
Am. For., USDA For. Serv., and USDA Soil Cons. Serv.
Southeastern For. Exp. Sta., Asheville, NC.
Palik, B.J., Goebel, P.C., Kirkman, L.K., and West, L.
Rowe, J.S. and Sheard, J.W. (1981). Ecological land
(2000). Using landscape hierarchies to guide restora-
classification: a survey approach. Environ. Manag.
tion of disturbed ecosystems. Ecol. Appl. 10: 189–202.
5: 451–464.
Pfister, R.D. and Arno, S.F. (1980). Classifying forest
habitats based on potential climax vegetation. For.
Sci. 26: 52–70.
Forest Communities PA R T 4
F orest ecosystems are characterized by a layered structure of functioning parts—

forest biota sandwiched between physiography and soil below them and the
atmosphere above them. Though they are often the main emphasis, forest communities
of trees and associated plants and animals form only one key structural component of
whole forest ecosystems.
Evolving from these structured, ecological environments in response to the abi-
otic factors therein, plants are integral parts of all ecosystem processes. Their individual
forms and their community physiognomy, therefore, reflect the vertical stratification of
the air above and the soil and physiography below. Plants of the forest community, indi-
vidually and collectively, are instrumental in myriad of changes that occur in the soil and
atmosphere of forest ecosystems.
Animals are also an indispensable part of forest ecosystems, and they are considered
in Chapter 12. We emphasize the role of animals in affecting the life cycles of plants in
this chapter. Animals, small and large, affect all aspects of the life history of plants from
reproduction to death, and they even promote “life-after-death” in clonal plants. Because
animals attack and consume plants at all stages in their life cycle, we consider in detail a
forest plant’s ability to defend itself from animals. We also consider the mutualisms that
develop between animals and plants, including the critical processes of pollination, seed
dispersal, germination, and establishment. Finally, the enormous effects of large animals
associated with humans (livestock) on forest ecosystems receive special treatment.
The forest community is considered in Chapter 13 as an integral part of the forest
ecosystem, but we work to ground communities in the context of their specific environ-
ments. Communities are aggregates or collections of organisms in a particular place at a
given time. We review traditional views of the forest community, their development, and
their shortcomings. Interactions among organisms are emphasized, including the concepts
of mutualism, competition, and niche, and we use these concepts to discuss the classic
processes of forest structural development. Finally, because of its importance in community
dynamics, the concept of understory tolerance of forest trees is examined in detail.
This part also includes a chapter on the diversity of both ecosystems and organ-
isms (Chapter 14). The diversity of whole ecosystems determines biodiversity itself, but
267
268 PA RT 4 Forest Communities
as with many ecological characteristics, the biota is often a main focus. Therefore, in
Chapter 14, both biodiversity and ecosystem diversity are considered at regional and local
scales. The value of biodiversity and its measurement is treated in detail, along with its
association with ecosystem function, and the maintenance and conservation of ecosys-
tems and their biotic diversity. These considerations of forest biota together set the stage
for the chapters in Part 5 on forest ecosystem dynamics.
Animals in Forest
CHAPTER 12
Ecosystems
A nimals of all sizes form indispensable parts of forest ecosystems, influencing forest community
composition and ecosystem processes. Animals are likewise strongly affected both by the
physical environment and by the plants with which they associate.
Plants provide shelter and food for animals. Green plants form the base of most trophic
systems—the food webs that comprise plant and animal relationships within ecosystems. Trophic
levels in a system vary but may consist of plants, the animals (herbivores) that eat them (including
browsers and grazers), the animal predators (carnivores) and parasites that feed on the herbi-
vores, and the scavengers and detritivores that eat animal remains and excrement. A trophic
system is completed by decomposers, most typically fungi and bacteria but also some animals,
that degrade and mineralize plant litter and animal residues. Trophic systems are complex, and a
detailed treatment is beyond our scope. Instead, we emphasize here other interactions of animals
and plants, including plant defense, the role of animals in regulating plant life history and produc-
tion, and the effects of large animals on forest ecosystems.
We focus in this chapter on understanding the role of animals in forest ecosystems. Humans,
of course, are animals with overwhelming effects on forest ecosystems, and those effects are
discussed in detail in Chapters 20, 22, and 23, among others. A long history by humans of empha-
sizing timber production and wildlife habitat has focused much attention on the activities of forest
animals perceived by humans to be destructive, such as insects that cause outbreaks or livestock
and native herbivores that reduce tree species regeneration. We emphasize from an ecosystem per-
spective that animal activities are only destructive as a human construct, and that such activities
may be critical and beneficial even when they appear to negatively impact forest resources. The
ecological contributions of native forest animals in the evolution of plants and their indispensable
roles in ecosystem processes are immensely significant but often underappreciated.
PLANT DEFENSE
There are many observations of woody species–animal relationships that we assume to be mutual
adaptations, though often without rigorous demonstration of cause and effect. Perhaps the best
examples of plant–animal co-adaptations are found in the area of plant defense to herbivory. The
existence of plant defense is often evidenced by the fact that herbivores consume only a fraction of
the plants available to them. All parts of woody plants are subject to herbivory at any stage of their
life cycle, and it follows that they exhibit many defense mechanisms that presumably evolved due
to the presence of herbivores and seed predators. In all cases, the nature of plant defense has devel-
oped within the particular site conditions of the ecosystem (Mattson and Haack 1987; Herms and
Mattson 1992).
Plant defenses may be physical in nature or may involve specialized plant chemistry. Physical
defenses employed by woody plants include structures or surface texture that may prevent herbiv-
orous injury or destruction. Such defenses may include leaf toughness, the presence of trichomes
269
270 C h a p t e r 1 2 Animals in Forest Ecosystems
(small hair-like outgrowths on shoots that deter herbivores, as in oaks; Hardin 1979), or the
presence of specialized organs or tissues such as thorns or resin ducts. Many tree species, including
black locust, honey locust, osage-orange, hawthorns, junipers, and some hard pines, as well as
many woody shrubs and vines (roses, greenbriers, blackberries, raspberries), have prickles, spines,
thorns, or sharp needle-leaves that deter browsing. Physical defenses in tree species tend to be con-
centrated in the juvenile phase when foliage and stem feeding by rodents, rabbits, and other herbi-
vores is most likely. A review of physical defense is provided by Hanley et al. (2007).
Chemical defenses include the presence of secondary chemicals, compounds utilized for
purposes other than metabolism, which tend to reduce the palatability or digestibility of plant tis-
sues to herbivores. Qualitative defense chemicals, or toxins, disrupt herbivore metabolism or
development even at low concentrations. Examples of toxins include alkaloids, terpenes, and cya-
nogenic glycosides, and they act as the main defense against mammalian herbivores (Bryant
et al. 1991). Quantitative defense chemicals, or digestibility reducers, inhibit a herbivore’s
absorption of nutrition from the plant tissues they have consumed. Examples include resins,
cellulose, and phenolics including tannins and lignins, all of which must be produced in high con-
centrations to be effective. An enormous number and diversity of secondary chemicals have been
described in higher plants, including thousands of alkaloids alone. Overall, herbaceous plants
have more qualitative than quantitative defense chemicals, and woody plants have a mix of both
types. Early-successional tree species are likely to have fewer defensive chemicals in general than
late-successional species, as evidenced by their higher susceptibility to herbivory (Coley et al. 1985;
Coley 1980).
The investment into plant defense changes as plants age, mainly because of age-related
changes in resource allocation (Boege and Marquis 2005). Secondary chemicals are costly for a
plant to produce because the plant is expending energy (and carbon) on a process other than
growth or reproduction. It is proposed that plants should therefore invest in defense only when
the benefit of avoiding herbivory outweighs the cost of reduced growth or reproduction. If resource
availability is a main control of plant defense, then plants newly emerged from a seed should have
higher defense because they can allocate stored resources in seeds and cotyledons (Herms and
Mattson 1992; Stowe et al. 2000). Once a seedling is well developed and established, defense is
likely to be lowest because plants have a large root-to-shoot ratio and most resources are allocated
to growth. By the sapling stage, plants can produce more photosynthate and allocate these resources
to defense. At reproductive age, allocation to defense may decrease if the plant has heavy investment
into reproduction (as in early-successional species), but it may continue if reproduction is delayed
or intermittent (as in late-successional species). Finally, plant defense is likely to decrease with the
reduction in most metabolic functions that occurs with old age (Figure 12.1). Empirical evidence
for trends in defense with plant age is inconsistent, however. In a meta-analysis of 116 published
studies, Barton and Koricheva (2010) found a general trend of increasing chemical defense during
the seedling stage followed by an increase in physical defenses during the vegetative juvenile stage.
INVESTMENT IN PLANT DEFENSE

The general concept of resource allocation generally dictates why plant defense may be empha-
sized in some plant species more so than others. A series of hypotheses, all of which consider
resource allocation, has developed to explain these differences (Stamp 2003). Critically, the mix of
plant genotypes that survive on a given site depends on pressures exerted by biotic interactions
including competition, mutualisms, and herbivory. The relative importance of these biotic factors,
however, is determined by physical site factors, emphasizing the influence of the abiotic as well as
the biotic components of landscape ecosystems.
The apparency hypothesis (Feeny 1976; Rhoades and Cates 1976) suggests that plants
difficult for herbivores to find will experience less herbivory, and such plants should therefore
Plant Defense 271
(a) (b) (c) (d) (e)

Resources Large root:shoot As root:shoot ratio As the mature stage is reached, resistance Reduced
stored in ratio; plants need to decreases, plants can decrease or be maintained, photosynthetic
cotyledons produce more acquire more depending on the reproductive strategy activity and
and seed photosynthetic area; resources to store of the plant hormone
reserves not enough resource and use for production;
for resistance resistance water stress;
loss of vigor
Defense and/or tolerance
II
Cotyledon Seedling Sapling Pre-reproductive / reproductive Over mature

stage
F I G U R E 1 2 . 1 Hypothesized relationship of plant investment into defense at various life stages. Plant
defense is moderate at (a) the cotyledon stage when the plant relies on stored resources, but decreases at
(b) the seedling stage when the plant emphasizes growth over other functions. Defense reaches its highest
point at (c) the sapling stage when excessive photosynthate is produced, then either continues (II) or
decreases (I) at (d) the reproductive stage depending on the species’ reproductive strategy. Plant defense
decreases in (e) old age with other metabolic functions. Source: Boege and Marquis (2005) / with permission
of Elsevier.
invest less into plant defense (typically qualitative defense chemicals). Likewise, plants—and plant
tissues—that are more apparent to herbivores, such as the dominant tree species in a forest or the
long-lived, evergreen needles on a conifer, should have more defense chemicals (quantitative
defense chemicals). Thus, young leaves of trees tend to be lower in defense chemicals compared to
older leaves. A study of downy birch in the boreal forest of Russia found that larger individuals of
downy birch suffered higher insect herbivory; background losses of foliage to insects increased
from small saplings to large saplings and then to mature trees (Zverev et al. 2017). Moreover, many
of the small individuals escaped from herbivory altogether. Notably, the apparency hypothesis has
been criticized for inconsistent support for its predictions (Smilanich et al. 2016).
The resource-availability hypothesis (Coley et al. 1985; Coley 1987) emphasizes plant
growth rather than apparency as a determinant of investment in plant defense. This hypothesis
suggests that fast-growing plants are better able to tolerate herbivory compared to slow-growing
plants, and thus fast-growing plants invest less into plant defense. This is because slow-growing
plants typically grow on sites where resources are limited, such that replacing tissue lost to her-
bivory would be more costly. There is much evidence for higher defense investment in slow-
growing species (Endara and Coley 2011), although differences in defense investment are generally
higher in tropical forests compared to temperate and boreal ecosystems (Van Zandt 2007).
The growth–differentiation balance hypothesis (Herms and Mattson 1992) suggests
that resource allocation within plants balances growth with defense mechanisms that limit her-
bivory (Figure 12.2). On favorable sites, plants allocate photosynthate to growth rather than
defense so that they may maximize their acquisition of readily available resources of light, mois-
ture, and nutrients. Plants therefore tolerate and compensate for herbivores with vigorous growth
Source-limited Sink-limited
range range
NAR
Response variable
RGR
Constitutive
secondary
metabolites
A B C
Low High
Available Nitrogen
F I G U R E 1 2 . 2 Relationship of allocation to net assimilation rate (NAR), relative growth rate (RGR), and
secondary metabolism along a gradient of low-to high-nutrient availability. The vertical lines at A, B, and C
represent high, moderate, and low levels of nitrogen, respectively. At B and C, net assimilation is constant,
whereas the relative growth rate and secondary metabolism are inversely correlated. The physiological
trade-off between growth and secondary metabolism is apparent. Source: From Hattas et al. (2017), as
adapted from Herms and Mattson (1992).
that keeps them competitive. On poor sites, plants allocate photosynthate to structural and
chemical traits that enhance the structure and function of existing cells, including those involved
in plant defense, because the loss of growth (and the subsequent need for re-growth if plant tissues
or organs are lost or damaged) is particularly costly where resources are scarce. Thus, plants must
either outgrow herbivory or defend against it, but not both (Herms and Mattson 1992). The growth–
differentiation hypothesis is difficult to test rigorously (Stamp 2004), and only limited evidence
exists to support it (e.g., Glynn et al. 2007).
PLANT DEFENSE AGAINST INSECTS

Plants exhibit a variety of defense strategies against insects, from fully compensating for any and
all injuries (complete tolerance) to chemical and physical defenses that provide complete or near-
complete immunity (Mattson et al. 1988). Even immunity may be overcome during periods of
stress or under outbreak conditions, however (Mattson and Haack 1987), such that most plants
have evolved partial tolerance whereby they regulate and thus minimize damage with chemical
and morphological defenses. The type of plant defense system depends on multiple factors,
including the type of tissue targeted (meristematic tissue, phloem, etc.), timing of the attack (early
versus late in the growing season), and the type of attack itself (single insect or mass attack). Matt-
son et al. (1988) ranked 13 insect-feeding guilds according to their potential injury to woody plants.
Insects least damaging to trees and shrubs are leaf and twig gall-formers and those affecting
late-season or prior year’s leaves, and phloem and sapwood borers and root feeders are intermediate
in their effects. Most damaging are phloem and cambium borers on main stem and roots because
they can destroy the plant’s conductive tissue faster than the plant is able to replace it.
Plant Defense 273
Examples of Injury and Plant Defense In conifers, and particularly the pine family,
leoresin (“pitch”) is a terpene-based compound exuded from resin ducts in needles, shoots, and
o
bark to deter foliage feeders and bark beetles that excavate galleries in the inner bark tissues.
Though used as a mechanism for defense, however, insects use the vapors from oleoresins to
search for host trees. When a tree is attacked by insects and begins to increase its production of
oleoresin, the resin vapor becomes concentrated and serves to attract additional bark beetles to
the attacked tree. Thus, trees damaged by lightning or previous attack, or even freshly cut trees,
tend to suffer more from mass attack by bark beetles than vigorous, standing trees (Hanover 1975).
For example, western larch trees are strongly resistant to bark beetle attack while standing and
healthy, but are immediately attacked after being felled (Furniss 1972). Likewise, moisture-
stressed loblolly pines that abnormally lose needles and exude resin have been shown to attract
insects (Heikkenen et al. 1986).
Plants require an unwavering defense against bark beetles that feed on the cambium because
their girdling of the conducting phloem and xylem has devastating effects (Mattson and
Haack 1987). When one insect becomes established, it uses pheromones to rapidly attract others,
contributing to mass attack. Healthy trees are generally unharmed by insects at endemic population
levels, but become attractive to beetles when the trees are stressed (Larsson 1989). Oleoresin
pressure is high in young and vigorously growing pines, physically repelling beetles entering the
bark (“pitching out”) or rendering the beetles physiologically impaired by chemical properties of
the resin, thus preventing them from reproducing (Figure 12.3). The terpenes myrcene and limo-
nene in healthy ponderosa pine may actually kill western pine beetles feeding on its needles or
bark (Smith 1966). Conversely, the severity of bark beetle attack is greatest when resin defenses are
low due to the natural aging process or from major stresses caused by competition, drought,
F I G U R E 1 2 . 3 Lower bole of a lodgepole pine

in the initial stages of attack by mountain pine
beetle in Grand Teton National Park, Wyoming.
Resins in the tree have “pitched out,” or
physically expelled, beetles that have bored into
the bark such that each bore hole is marked by
“plugs” of hardened sap (six such plugs are
denoted with white arrows). Additional and
continued attack by beetles will eventually
overcome the tree’s defenses and kill it.
pollution, logging damage, disease, or defoliation. Bark beetles exhibit various degrees of tolerance
to resin toxicity, and as a group are host-specific and more tolerant to resins of their own host
species than those of other hosts.
Anti-herbivory resins and other defense chemicals are also found in hardwood species.
Certain hardwood trees such as cottonwoods have resins that deter insect herbivory (Curtis and
Lersten 1974), as do young leaves of the desert shrub creosote to avoid defoliation (Rhoades 1976).
Juglone is a compound produced by walnuts and shagbark hickory that deters feeding by some
bark beetle species. Tannins also tend to deter herbivory when present in high concentrations,
especially in oaks. A study of acorns (oaks) showed a high concentration of tannins around the
embryo at the apical end, thereby confining weevil activity to the less-protected basal end
(Steele et al. 1993). Oak leaves are most susceptible to insect herbivory during their development
in the spring when tannins are absent or scarce, but the late and rapid flushing of preformed leaves
and shoots of oaks minimizes the time insects may feed and reproduce using these tissues. This
rapid leaf-flushing trait in north temperate forests has apparently elicited reciprocal adaptations of
insects and their hosts. For example, flushing time may differ by as much as 3 weeks among trem-
bling aspen clones (Barnes 1969), and populations of tortricid caterpillars predominantly infest
leaves of early flushing clones (Witter and Waisanen 1978). A similar relationship was also reported
for larvae of tortrix moths on oaks in Russia (Sukachev and Dylis 1964) and Europe (DuMerle 1988).
In sugar maple, early flushing buds suffered greater damage by pear thrips than trees with late
budburst (Kolb and Teulon 1991).
There are many studies that provide examples of how site conditions may affect the intensity
and nature of insect attack and reciprocal plant defenses. Pinyon pines growing on cinder fields in
Sunset Crater of northern Arizona live in a highly stressful environment with limited moisture and
low nutrient status. The pines suffer unusually high levels of chronic and severe herbivory by
many insects, to the point that tree architecture is altered, growth is reduced, and female reproduc-
tive function is eliminated (Whitham and Mopper 1985; Mopper et al. 1991a). Pinyons growing on
less-severe sandy loam soils adjacent to the cinder fields, however, are rarely attacked and exhibit
normal reproduction and growth, despite producing significantly less resin than the Sunset Crater
pines. Notably, insects at Sunset Crater not only reduced tree growth directly by consuming above-
ground tissue but also indirectly by reducing the amount of ectomycorrhizal fungi associated with
roots of susceptible individuals (Gehring and Whitham 1991; Del Vecchio et al. 1993). These
studies therefore show that the incidence of herbivory may influence how photosynthate is differ-
entially allocated to growth, defense, or to maintain mycorrhizae.
Nutrition Insects assimilate nutrients from plant tissue with very low efficiency, such that
variations in the nutritional value of plants have large implications for insect herbivores. Plants
growing on poor sites (low moisture and nutrients) are likely to produce lower-quality tissues,
which may be detrimental to insect performance and thus effectively inhibit herbivory (Herms and
Mattson 1992). The plant-stress theory of herbivory, however, predicts that plant stress induced by
limited moisture or nutrients actually benefits insects by increasing the concentrations of usable
substances in foliage and creating a more favorable balance of nutrients (Mody et al. 2009; Mattson
and Addy 1975; White 1978). Evidence for both situations comes from experimental studies of
sawflies on pinyon pines at Sunset Crater, Arizona (Mopper and Whitham 1992; see pp. 337–338).
This seeming paradox, whereby abiotic stresses may decrease or increase performance, may be
explained by the type of stress considered. Mopper and Whitham (1992) contrast sustained plant
stress, such as poor soil-nutrient conditions or prolonged drought prevailing while insects are both
active and inactive, with “simultaneous” plant stress, such as low precipitation while the insect is
feeding or ovipositing (Figure 12.4). They predicted that under simultaneous stress, insect
performance will be high when plant stress is low and drops rapidly as stress increases. At the
highest level of plant stress, the plant becomes an inadequate food source. By contrast, under
Plant Defense 275
F I G U R E 1 2 . 4 Hypothetical relation-
ship between insect performance and Simultaneous
simultaneous and sustained plant
Sustained
stress. Conditions at Sunset Crater,
High
Arizona are indicated. Source: Mopper
Insect Performance
and Whitham (1992) / John Wiley &
Sons. Reprinted with permission of the Sunset
Ecological Society of America. Crater
Low
Low High
Plant Stress
sustained stress, insects are relatively successful at sites of low chronic plant stress, even more suc-
cessful at sites of intermediate stress, but unsuccessful when high stress makes the plant an inad-
equate food source. More recent studies have also suggested that intermittent stresses rather than
sustained stresses improve plant quality for some herbivores (Huberty and Denno 2004; Sconiers
and Eubanks 2017), particularly those that feed on sap rather than leaves or other plant tissues. For
example, aphids were found to be more abundant on intermittently stressed Sitka spruce com-
pared to those continuously stressed (Major 1990).
In summary, woody plants exhibit a wide variety of defenses against insect herbivores
including those associated with different individual genotypes. Even the likelihood of somatic
mutations in modular shoots of long-lived plants or clones may represent a mosaic of genotypes
that might effectively prevent herbivores from evolving metabolic pathways to overcome plant
defenses (Whitham and Slobodchikoff 1981). These new and diverse plant defenses force herbi-
vores to co-evolve and find new ways to avoid or resist them. These intricate interrelationships
among site conditions, vegetation, and herbivores further cement the important role of animals in
the structure and function of ecosystems on the landscape.
Plant Hybrid Zones as Reservoirs for Insect Diversity Plant hybrid zones have been
found to be critical centers of insect abundance and diversity (Kearsley and Whitham 1989;
Whitham 1989; Floate et al. 1993; Whitham et al. 1994). In Weber Canyon, Utah, 85–100% of the
Pemphigus betae gall aphid population occurs on less than 3% of its host population in a 13-km
hybrid zone of Fremont and narrowleaf cottonwood (Figure 12.5b; Whitham 1989). Hybridization
has altered the well-engineered defense of each parent species such that aphid populations are
more viable on the hybrid. The concentration of aphids on such a small segment of the host
population suggests that susceptible hybrid plants not only act as insect reservoirs in ecological
time, but they may also have inhibited the aphids from adapting to the more numerous parent
hosts in evolutionary time.
The same hybrid zone in Weber Canyon is also superior habitat for the free-f eeding
beetle, Chrysomela confluens (Figure 12.5a), in part because the early leaf flush of narrowleaf
cottonwood and the hybrid provides the first source of abundant food for beetles in spring. In
addition, staggered leaf phenology in the hybrid zone allows beetles to shift onto newly flushed
Fremont cottonwoods as foliage of the hybrid and narrowleaf trees declines in quality. Thus, the
hybrid zone is a phenological reservoir that increases beetle fecundity and leads to chronic
(a)
60
Fremont Hybrid Narrowleaf Zone
Zone and
50 Overlap
No. Beetles/2 min Census
Zone
40
30
20
10 Chrysomela confluens
0
(b)
1,200
No. Galls/10,000 Leaves
1,000
800
600
Pemphigus betae
400
200
0 20 40 60 80 100 120 140 160
Distance from River Mouth (km)
F I G U R E 1 2 . 5 Occurrence of two insect herbivores with similar distributions in the hybrid zone between
Fremont and narrowleaf cottonwoods along the Weber River, northern Utah. (a) Free-feeding beetle,
Chrysomela confluens, values are 3-year means (±1 SE). (b) Galling aphid, Pemphigus betae. Source: Floate
et al. (1993) / John Wiley & Sons. Reprinted with permission of the Ecological Society of America.
herbivory every year. This beetle species may also be an example of the hybrid bridge hypo-
thesis (Floate and Whitham 1993), which suggests that hybrid plants help herbivores to switch
or adopt additional plant host species, thus allowing them to evolve a larger range of host species
when hybrids are present. A study of two red oak species and their hybrid in central Mexico
found that 32% of insects in the study were specific to one of the parental oaks, 23% to the other
parent, and 9% to the hybrid (Tovar-Sanchez and Oyama 2006), and that the hybrid oak sup-
ported intermediate levels of herbivory between the parents. The increased genetic diversity
among the vegetation of the hybrid zone appears to support greater genetic diversity among the
herbivores.
PLANT DEFENSE AGAINST MAMMALS

Plants must also defend against browsing mammals as well as insects, and plants employ chemical
defenses against mammals in stems and leaves as well as physical defenses of prickles, spines, and
thorns. Boreal mammals follow a well-studied pattern of woody plant food selection in browsing a
variety of evergreen and deciduous species, growth phases, and plant parts, primarily as a strategy
to avoid secondary chemicals (Bryant 1981; Bryant et al. 1985; Huntly 1991). These chemicals
strongly influence the palatability and food value of plant tissues during the time they are dormant
Roles of Animals in Plant Life History 277
in the winter. Birches and willows are heavily browsed by voles and snowshoe and mountain hares
in boreal ecosystems, with heavy preference for juvenile rather than adult growth. These herbi-
vores have relatively little access to adult growth, such that their herbivory would be expected to
select for heavy chemical defenses in juvenile growth (Barton and Koricheva 2010). Small droplets
containing resin with secondary chemicals such as papyriferic acid are located on the surface of
birch twigs and young stems. The palatability of birch seedlings and saplings (1–8 years old) to the
mountain hare is negatively correlated with the number of resin droplets, and hares are keenly
able to detect palatable plants (Rousi et al. 1989, 1991). Birches seem to be especially low-quality
browse species. For example, when a previously multispecies site on Alaska’s Kenai Peninsula
became white birch-dominated after a forest fire, the moose population starved (Oldemeyer
et al. 1977).
Various woody plants produce juvenile-phase sucker shoots heavy in secondary chemi-
cals when browsing is severe. Even fast-growing boreal plants that emphasize compensatory
growth to replace browsed tissues allocate significant resources to defend against winter brows-
ing (Bryant et al. 1985). For example, juvenile sprouts of trembling aspen contain a chemical
that deters browsing by beavers such that they avoid juvenile saplings in favor of large trees that
have low concentrations of a specific phenolic compound (Basey et al. 1988, 1990). However,
where juvenile trees are uncommon, as in areas newly occupied by beavers, beavers select
smaller, non-juvenile trees to maximize their net energy intake. Woody plants found in boreal
or subalpine regions with harsh winter conditions may be more likely to experience herbivory
by mammals that preferentially browse adult over juvenile growth compared to those found in
moderate climates (Swihart and Bryant 2001).
ROLES OF ANIMALS IN PLANT LIFE HISTORY

Invertebrates and vertebrates affect many ecosystem processes, including nutrient cycling, water
cycling, and the regeneration and succession of forest trees. Animals influence these processes by
dispersing pollen and seeds, decomposing organic matter, mixing the soil, and damaging and
killing trees. In many instances, animals may even regulate forest composition and growth. In the
following sections, we examine some of the key roles of animals in affecting plants from
birth to death.
POLLINATION
Animals play a critical function in the life cycle of woody plants via their role in pollination.
Temperate and boreal forests are dominated mostly by species that are wind-pollinated, but animal
pollination is widespread among tropical woody species. Pollination is accomplished by insects
(bees, wasps, flies, beetles, butterflies, and moths), birds (especially hummingbirds in the New
World), and bats (Baker et al. 1983). Animal pollinators are mainly attracted to plants using nectar;
fragrance; flower color, shape, and size; and, in the case of birds, insects visiting the blossoms.
A comprehensive account of animal–plant interactions in pollination ecology is given by
Willmer (2011).
In North America, many families and genera of woody species are dominantly or wholly
insect-pollinated. Understory species are primarily insect-pollinated, whereas most upper tree
canopy species are mainly wind-pollinated. Major insect-pollinated groups include most species of
the Ericaceae, Fabaceae, and Rosaceae, as well as some species of the genera Acer, Aesculus,
Catalpa, Cornus, Magnolia, Liriodendron, Nyssa, Salix, Sassafras, Tilia, and Zanthoxylum. In the
case of tulip tree, insect pollination is inefficient (Boyce and Kaeiser 1961), and only about 10% of
the seeds may be viable. Nevertheless, enough viable seeds are produced per tree that natural
regeneration is not limited.
SEED DISPERSAL
Animals are instrumental in maintaining and spreading woody plant populations by connecting
site and plant through seed dispersal. The body of literature on seed dispersal is enormous, pre-
sented most often in a biotic context of animal–plant mutualisms, co-adaptations, and co-evolution
(Janzen 1983). We emphasize here that dispersal is part of plant regeneration, and plant regenera-
tion is a process that links plants and animals to physical site factors where the success of dispersal
is played out. Some woody species produce highly nutritious fruits adapted for a small group of
specialized frugivores (“specialists”) that reliably disseminate the seeds (Howe 1993). One example
of a woody plant catering to specialists would be various parasitic mistletoes, which have nutri-
tious, sticky fruits with seeds that tend to pass through the disperser’s gut, have peak availability
when other food sources are scare, need to be dispersed to locations at the tops of host trees, and
are dispersed by birds from only eight families (Watson and Rawsthorne 2013). Other woody
species offer less-nutritious fruits in very high abundance, relying on common and numerous fru-
givores that are individually less reliable (“generalists”), but collectively effective in dispersing the
seeds (Howe 1993). Most animal-dispersed temperate woody species, with some exceptions, fall on
the continuum between these extremes. Seed dispersers of woody plants tend to be vertebrate ani-
mals (birds, mammals, fish, and reptiles).
Three key features characterize an animal–plant–site dispersal system that is successful in
the establishment and persistence of the plant species. First, the fruit or seed must be attractive to
the disperser, either by sight, smell, or taste. Second, the fruit must be at its most attractive at the
same time that the seed matures, because premature ingestion destroys the developing seed.
Finally, a sufficient number of viable seeds must escape predation or digestion by the dispersal
agent and be deposited where the chances of successful establishment are high. Some escape
mechanisms include burying or regurgitation of the seeds by the disperser; hard, smooth, seed
coatings that assure undamaged passage through digestive tracts; and darkly or inconspicuously
colored fruits coupled with brightly colored accessory parts. In the latter case, the animal is
attracted to the fruit by a red or orange aril, peduncle, or bract, and if the dark-colored fruit is
dropped, it is not readily found (Janzen 1969). Regarding dispersal of seeds via vertebrate guts,
Janzen (1983) concluded that virtually all traits of seeds and fruits have probably been modified for
protection of the seed in passing through the guts of dispersing animals.
Fish and Reptiles Fish eat pulpy seeds of various woody species growing along rivers and
in recent decades have been recognized as important dispersers. Many riparian trees and shrubs
tend to have fruits and seeds dispersed by water, and thus dispersal by fish would allow upstream
transport that is not otherwise possible. Most of the evidence for dispersal by fish is in tropical
regions, although across wide biogeographical regions (Correa et al. 2007; Horn et al. 2011). Seed
dispersal by fish has also been documented in North America for channel catfish in the Mississippi
River (Chick et al. 2003). Catfish were observed to consume fruits of red mulberry and swamp
privet during high water periods, and consumed seeds had higher rates of germination than
unconsumed seeds when harvested from fish guts. It is assumed that fish defecation would even-
tually deposit the seeds on floodplain sites suitable for germination once floodwaters receded.
Reptiles, particularly turtles and tortoises but also alligators, lizards, and snakes, have a
keen sense of smell and may eat fruits after they have dropped off trees or when borne close to
the ground. The fruits of iguana hackberry are eaten by climbing iguanas. Although most
modern reptiles are not vegetarians, an increasing number of cases of reptilian frugivory have
been documented. For example, box turtles were found to be important dispersers of two palm
species and one understory shrub in dry, fire-dependent pine forests of southern Florida (Liu
et al. 2004).
Birds Birds are primary dispersal agents and have developed many adaptive interrelation-
ships with plants. Birds disperse seeds either by disgorging fruits or seeds carried in the mouth, or
by excreting seeds contained in fruits that have been eaten. Rarely do birds carry fruits on the
outside of their body (except for sticky fruits of mistletoe, referenced in the earlier text).
Seeds are often destroyed when they are eaten and digested by birds, but certain dispersers
do not immediately consume the fruits and seeds they collect. Many wood pigeons, thrushes,
nutcrackers, crows, and waxwings can disgorge whole fruits and seeds from their beak. An
important group of avian dispersers attempt to store or hide their food either above or below
ground but then neglect to recover all or part of it (Pesendorfer et al. 2016). Birds that disperse
and cache acorns and other nuts above ground, for example, usually do so in tree cavities and
bark crevices, such as the acorn woodpecker of California (Stacey and Koenig 1984). The acorn
woodpecker imbeds thousands of acorns, almonds, and hickory nuts into small holes it drills in
the bark of communal storage trees (Figure 12.6) known as “granaries” (Pavlik et al. 1991).
Squirrels and other rodents may then carry them to a germination site, completing dispersal.
Birds that routinely cache nuts and seeds below ground include jays and nutcrackers. Placement
of the seeds in the ground increases the probability of successful establishment for many woody
angiosperms dispersed in this manner (oaks, beeches, hickories, chestnuts, hazelnuts). In the
eastern deciduous forest, the blue jay disperses nuts more than several hundred meters and
caches them in the ground (Johnson and Webb 1989); documented dispersal flights are up to
1.9 km in Virginia and up to 4 km in Wisconsin (Darley-Hill and Johnson 1981; Johnson and
Adkisson 1985). From a single Wisconsin woodlot, jays were estimated to have dispersed 150 000
viable beechnuts in only 27 days (Johnson and Adkisson 1985).
The mutualism between wingless-seeded white pines of semi-arid and subalpine envi-
ronments in North America, Asia, and Europe and nutcrackers, jays, and woodpeckers
(Lanner 1981, 1990) may be one of the most striking. The birds disperse millions of pine seeds over
Photo by Lorraine Bruno, Sonoma County, CA
F I G U R E 1 2 . 6 This California acorn woodpecker (Melanerpes formicivorus) has imbedded thousands of

acorns in the bark of this tree near Sebastopol, California.
long distances and bury them in favorable microsites, but the seeds also are an abundant and
highly nutritious food source for the birds. Germination and establishment of unrecovered seeds
is likely the only way for species such as limber and whitebark pines to become systematically
established (Lanner 1980, 1982). Dispersal is more than simply a bird–pine interaction; insects
may indirectly affect the dispersal process by both birds and mammals, both of which avoid pines
with reduced seed crops due to insect infestation (Christensen and Whitham 1991, 1993). Notably,
specific site conditions, whether xeric woodlands or subalpine mountain sites, affect the life his-
tories of both plant and animals and their mutualistic association.
Note that gathering seeds and caching them in distributed locations away from where they
were collected, a behavior known as scatter-hoarding, only becomes an important method of seed
dispersal when some of the seeds are left unrecovered by the disperser. Thus, plants have evolved
various strategies that reduce the likelihood that cached seeds will be recovered (Vander Wall 2010).
First, the production of attractive, large, nutritious seeds and fruits is useful when it stimulates
dispersers to hoard or collect and cache more seeds than will be recovered for food. Second, if seeds
are difficult to eat because of physical or chemical barriers that take handling time to overcome,
seeds are more likely to be hoarded than immediately eaten. A third strategy is masting, where all
the individuals of a population of trees or shrubs produce a large crop of fruit (usually nuts) at the
same time (Chapter 4), which causes animals to collect and store more food than is needed. Finally,
odorless seeds are less likely to be detected once cached. Each of these traits is common and may
have evolved in response to scatter-hoarding (Vander Wall 2010).
The framework of seed-dispersal effectiveness describes the contribution of a disperser
to the fitness of a plant species (Schupp et al. 2010). Components of seed-dispersal effectiveness
include visitation rate and the number of seeds acquired per visit (dispersal quantity), how the
seeds are treated in the mouth and gut, and factors of seed deposition that affect seedling emer-
gence, survival, and growth (dispersal quality). In the southwestern United States, pinyon jays
and Clark’s nutcrackers disseminate pine seeds and bury them 2–3 cm deep in loose soil at com-
munal caching areas. Caching usually occurs on south-facing slopes that are free of snow by late
winter (Vander Wall and Balda 1977), such that cached seeds provide a critical food source to
initiate the breeding season (Ligon 1978). When pinyon seeds are abundant, entire flocks of 200–
300 pinyon jays gather and store them day after day over a period of months. Using conservative
figures of 30 seeds per trip and 4 trips per day, Ligon (1978) estimated a flock of 250 jays would
store 30 000 seeds per day and approximately 4.5 million seeds over a 5-month period. The Clark’s
nutcracker may carry an average of 55 and as many as 95 seeds of the smaller-seeded whitebark
pine in its sublingual pouch per trip (Vander Wall and Balda 1977). A single Clark’s nutcracker
could disperse 1225 seeds per day for 80 days or 98 000 seeds per individual per year (Hutchins
and Lanner 1982). Clearly, both nutcrackers and jays have very high seed-dispersal effectiveness
in this system.
Though not explicitly considered within the seed-dispersal-effectiveness framework, long-
distance dispersal (best exemplified by birds) is also considered to benefit plant fitness because it
allows for colonization of new areas and eventually tree migration and the expansion of species’
ranges (Johnson and Webb 1989; Cain et al. 2000). Pinyon jays can cache seeds up to 22 km from
pinyon stands, and so their role of disseminating pinyon seeds over long distances is enormous. In
a study of ponderosa pine, which has winged seeds for short-distance dispersal by wind but is also
dispersed over long distances by Clark’s nutcrackers and pinyon jays, four isolated Wyoming popu-
lations of ponderosa pine were found to have been established by long-distance dispersal (Lesser
and Jackson 2013). Ages and genotypes of each individual tree in the populations showed that
long-distance dispersal was responsible for initial colonization of each of the four sites. However,
the initial populations were unable to reproduce via on-site reproduction while young and required
additional long-distance dispersal events to continue population growth until they were able to
expand on their own.
Species such as limber pine and the stone pine group (including whitebark pine of western
North America, Swiss stone pine of Europe, and Korean, Japanese, and Siberian stone pines of
Asia) are found at the highest forest elevations and at timberline in harsh environments. The
integrated bird–plant–site dispersal system involving these species largely determines their mor-
phology, successional status, population age structure, and tree spacing. The morphology of trees
and cones appears to be influenced evolutionarily by nutcracker species (Lanner 1980). Cones are
displayed in a highly visible position on steeply upswept limbs, rigidly attached and retained in
the crown where seeds are visible but cannot fall out even when shaken or rotated—ideal for for-
aging by nutcrackers. Furthermore, the population genetic structure of pine stands is largely
shaped by the nutcracker filling its sublingual pouch with seeds from only one or a few trees
before caching. As a result, the multiple seedlings of whitebark pine from a single cache are more
closely genetically related than individuals from distant clumps (Furnier et al. 1987). Thus, the
local population structure and possibly the mating system of the pines are strongly related to
avian behavior in seed harvest and caching.
In addition to the white pines described in the earlier text, Juniperus is the other exceptional
group of bird-dispersed conifers. Eastern redcedar has cones that effectively resemble bluish
berries, thereby enticing cedar waxwings, sparrows, robins, warblers, and mockingbirds to con-
sume entire cones with seeds. A study of eastern redcedar in southwestern Virginia found that 65%
of cones were dispersed away from the parent trees by birds, and 61% of cones were dispersed over
long distances (Holthuijzen et al. 1987; Figure 12.7). As with the white pines, physical site factors
supporting eastern redcedar are again a key part of the dispersal system. Birds dispersed redcedar
Predation of
seeds by Remaining ripe cones
birds 0.9
Avian dispersal
3.1 65.3
TOTAL
CONE CROP
42 Short-distance 61.3 Long-
100
avian dispersal distance
avian dispersal
1.0 Predation
by insects
Dispersal of ripe cones by

17.9 Unripe 11.6 wind and gravity
cones Germination Predation
<0.1 ? ? <0.1 3.5 1.7 ? 24.5 ?
<0.1 0.3 0.1 2.2
0.9 7.3 Seed bank 2.5 36.8
44.8
F I G U R E 1 2 . 7 Descriptive model of eastern redcedar cone-crop dispersal from June through May of the
following year. Numbers in circles are percentages of the total cone crop and are means of four sample trees
in southwest Virginia. Source: Holthuijzen et al. (1987). / with permission of Canadian Science Publishing.
Reprinted with permission of National Research Council Canada.
cones to barrens and rock outcrops that were open, dry, and relatively fire-free, forming “cedar
glades” on the pre-European colonization landscapes of the Midsouth and Appalachian Moun-
tains. Combined with the ability of redcedar to germinate quickly and survive on dry sites, the
species was able to persist in regions otherwise overwhelmingly dominated by deciduous forests.
These same attributes also make the species today a successful colonizer of old fields and aban-
doned pastures.
Mammals Mammals such as rodents, ungulates, bats, and certainly humans are also impor-
tant dispersal agents. Tropical forests include a diverse suite of mammal dispersers because they
have fruits adapted for mammal dispersal year-round, and many fruits are eaten by both mammals
and birds. Rodents are the primary agents in temperate forests. Many deciduous woody plants in
northern temperate forests have their seeds dispersed by scatter-hoarding rodents as well as birds.
Many of these species produce nuts, such as oaks, hickories, walnuts, chestnuts, horse-chestnuts,
hazelnuts, and beech (Vander Wall 1990). As with many birds, rodents often destroy the seeds, but
many are cached and stupidly forgotten, allowing the unrecovered seeds to germinate and estab-
lish. Acorns and hickory nuts, for example, are dispersed by squirrels for distances up to 50 m from
the parent tree, appreciably farther than what might be accomplished by wind or gravity. This
longer-distance dispersal is important for hickory species that rely on gaps in the oak forest or the
forest edge for their eventual development into the overstory.
Secondary chemicals in acorns are effective in deterring rodent feeding so that if other more
palatable foods are available, acorns may be cached rather than eaten on the spot. Oaks are known
to concentrate tannins in the apical part of the seed where the embryo is located, such that the
apical part is less palatable and the probability of embryo survival of seeds discarded by the con-
sumers is higher (Steele et al. 1993). Squirrels prefer acorns of the white oak group to those of the
red oak group because tannins are three to four times higher in the red oak group, but tannins in
both groups deter rodent feeding (Short 1976). Moreover, acorns of the white oak group germinate
quickly after falling and diminish in palatability after sprouting (Smith and Follmer 1972). By con-
trast, acorns of the red oak group lie dormant over winter, with low palatability, and germinate the
following spring. Compared to legumes and other seeds, or dried and fleshy fruits, acorns have
relatively low levels of protein and phosphorus and insufficient nitrogen that do not supply ade-
quate nutrition to squirrels (Short 1976). The relatively large size of the nuts of many temperate
oaks and hickories is probably an adaptation to entice dispersers with a valuable energy source,
although nut size is probably related to both latitude and site quality. Many oaks and hickories
grow on dry or dry-mesic sites where energy from seed reserves is required for initial development
of a tap root during the seedling phase to penetrate a surface mat of organic matter and reach sub-
surface moisture during the growing season.
As dispersal agents, mammals differ from birds in that they have a keener sense of smell and
have teeth that masticate seeds. They are also typically larger, spend a significant portion of their
lives on the ground, are most active between dusk and dawn, and are color-blind feeders (van der
Pijl 1972). The characteristics of fruits eaten by mammals reflect these differences and include a
favorable smell; a hard skin; protection of the seed itself against mechanical destruction (often a
stone-like covering of the seeds as in all drupes), often assisted or replaced by the presence of
secondary chemicals in the seed; non-essentiality of color; and in many cases large size that causes
them to drop to the ground.
Also notable is the individual behavior of mammals and birds during the dispersal pro-
cess. Scatter-hoarding requires decisions about seed selection, whether to consume the seed or
cache it, how far to disperse it, and where to cache it (Lichti et al. 2017). Differences in these
decisions among individuals within an animal species have been shown to affect many important
ecological processes and have been examined as animal personalities (Wolf and Weissing 2012;
Carere and Maestripieri 2013). A lab study of nearly 650 voles, mice, and shrews captured from an
experimental forest in Maine showed that 90% of behaviors were repeated, indicating animal
ersonality (Brehm et al. 2019). Personality affected the size of seeds selected, the distance seeds
p
were dispersed, where the seeds were cached, and the probability of consuming a seed. For
example, the distance of seed dispersal was influenced by differences in anxiety (indicated by time
spent grooming) in mice and timidness (indicated by time spent in the open) in voles (Figure 12.8).
Moreover, the proportion of mice and voles with a given personality trait differed among forests
with different structure; mice, for example, were more active and less timid in even-aged forests
compared to reference forests (Brehm et al. 2019). Reviews of the emerging field of animal person-
ality are provided by Zwolak and Sih (2020), Carere and Maestripieri (2013), and Smith and Blum-
stein (2008).
Ungulates consume a wide variety of vegetation and thus are likely to disperse seeds,
although their influence on seed dispersal is probably strongest in tropical forests and savannas.
For example, many African savanna ruminants consume considerable amounts of tree fruits,
particularly legumes. Many acacias provide leathery, nutritious pods containing extremely hard
and smooth seeds, which evade or resist strong molars. Large grazing mammals extinct since the
Pleistocene (such as extinct horses, gomphotheres, glyptodonts, and ground sloths) may have also
been important in seed dispersal of certain plants in Central American lowland forests (Janzen
and Martin 1982). Even some temperate deciduous species with large, sweet-fleshed fruit, for
example, Kentucky coffeetree, osage-orange, honey locust, pawpaw, persimmon, and ginkgo, may
have had denser populations and much wider ranges in the past compared to today because they
were formerly dispersed by now-extinct megamammals. Delcourt and Delcourt (1991) speculated
that osage-orange, which is a widely planted but relatively rare tree with fleshy fruits approxi-
mately the size of a softball, may have formerly been dispersed by mastodons!
Other mammals are also prominent but less-influential seed dispersers in temperate for-
ests. Bats are important dispersal agents of woody species, primarily in tropical Asia and Africa
Mice Voles
6.3 6.3
Dispersal distance (m)
2.5 2.5
1.0 1.0
High 0 1 2 Low 0 1 2 3
Timid Bold
anxiety anxiety
Anxiety (Prop time grooming) Boldness (Prop time center)
F I G U R E 1 2 . 8 Effects of personality (x-axes) on seed dispersal distance (y-axes). Anxiety (quantified as

the proportion of time spent grooming) decreases dispersal distance in mice (left) and boldness (quantified
as the time spent in the open) increases dispersal distance in voles (right). Y-axes are on a log-10 scale.
Source: Brehm et al. (2019).
(van der Pijl 1957), eating fruits that are of drab color, have a musky odor, and are often large and
exposed outside the foliage. They transport seeds within about 200 m of the fruit source. Primates
such as monkeys and apes are mostly destructive, eating everything edible, ripe or unripe, with
an apparently limited dispersal role.
Finally, we cannot ignore the role of humans as significant seed-dispersal agents, spreading
plants (as well as insects and pathogens) into diverse areas outside their native ecosystems, often
as horticultural and forest introductions. The rate and distance of seed dispersal of many plant
species by humans are likely unmatched by any other animal group. The role of humans in seed
dispersal of many plants considered to be invasive species, as well as introduced and destructive
insects and pathogens, is discussed in detail in Chapter 21.
GERMINATION AND ESTABLISHMENT

Many animals primarily influence germination of temperate tree species by caching seeds in the
forest floor where dormancy requirements (if any) may be satisfied. Many seeds are adapted to pass
through the digestive tracts of birds and mammals unharmed, and digestive juices may weaken the
seed coat, favoring the absorption of water and increasing eventual germination.
Animals aid in plant establishment and plant growth, especially on nutrient-poor sites, by
transporting and “fertilizing” the soil with mycorrhizae (see Chapter 19) that help the plant obtain
water and nutrients (Maser et al. 1978; Maser and Maser 1988). Although some mycorrhizal fungi
produce aboveground fruiting bodies with wind-dispersed spores, many produce belowground
fruiting bodies that do not discharge spores and thus depend on animals for dispersal. When ani-
mals eat mycorrhizal fruiting bodies, they consume fungal tissue that contains nutrients, viable
fungal spores, nitrogen-fixing bacteria, yeast, and water. The fruiting bodies are digested and the
undigested material containing these components is excreted in the forest and in disturbed areas.
Under favorable conditions, roots of seedlings or established trees in contact with fecal pellets may
be inoculated with the mycorrhizal fungus when spores germinate. Fungal spores may also be dis-
persed by adhering externally to the disperser’s body using spines, ridges, hooks, and other struc-
tures (Halbwachs et al. 2015). Dispersal of mycorrhizal spores in temperate forests has been
documented for a diverse set of animal dispersers, including springtails, insects, mites, millipedes
and centipedes, slugs, salamanders, rodents, birds, and ungulates (see review in Vašutová
et al. 2019). Vertebrates and certain fungus-eating flies disperse mainly ectomycorrhizal fungi by
ingesting fruiting bodies, but these fungi may be also transported externally by springtails and
mites. Invertebrates are the main dispersers of arbuscular mycorrhizae, particularly earthworms
and millipedes. Slugs, snails, and beetles are known to be important dispersers of fungal propa-
gules in both mycorrhizal groups.
The digging activities of various animals turn up mineral soil that provides a seedbed suit-
able for plant establishment. Sukachev and Dylis (1964) reported that the germination of oak and
maple seeds was twice as high where 2–3% of the surface was covered with molehills compared to
undisturbed soil. Rooting by wild boars may also remove thick moss, organic matter, or other veg-
etation that inhibits germination and establishment, although rooting animals may also destroy
young seedlings and cause widespread damage. Rooting by wild boars in Russia may cause the
replacement of hardwoods by Norway spruce. Feral pigs in the eastern United States are huge con-
sumers of mast crops, which may shift forest species composition and tree density by interfering
with seed dispersal.
On the other hand, digging activities may either inhibit or favor tree species depending on
their specific mode of regeneration. In northern Arizona, belowground herbivory by pocket
gophers on the roots of trembling aspen prevented its colonization of mountain meadows (Cantor
and Whitham 1989). In 32 aspen-meadow associations, the distributions of aspens and pocket
gophers overlapped 93% of the time. However, in the mountain plateaus of central and northern
Utah, gophers may favor aspen regeneration by stimulating vigorous sprouting via belowground
herbivory. Notably, site conditions are key to the presence of this process: limestone-derived soils
in the West support gopher burrowing and nest building, but extensive animal burrowing is limited
on sandy Michigan sites such that aspen sprouting is less common.
DECOMPOSITION, MINERAL CYCLING, AND SOIL IMPROVEMENT

The relative contribution of soil animals to decomposition varies with climate, being greatest at mid-
latitudes and decreasing toward the poles (Swift et al. 1979). Where they are important, soil meso-
fauna (springtails, mites, earthworms) play a significant role in organic matter decomposition by:
1. Physically crumbling and breaking up tissues; increasing the surface area available for
bacterial and fungal action.
2. Selectively consuming material such as sugar, cellulose, and even lignin.
3. Converting plant residues into humus.
4. Mixing decomposed organic matter into the upper layer of soil.
5. Forming complete aggregates between organic matter and the mineral fractions of soil
(Edwards 1974).
Notably, climate appears to have a very strong effect on the importance of animals in terres-
trial decomposition. Soil animals increase the rate of decomposition in temperate and wet tropical
climates, but have a lesser effect where temperature or moisture limits their activity (Wall et al. 2008).
Other aspects of the physical environment, forest composition, and the amount and kind of organic
matter present strongly affect the number and diversity of microfauna and mesofauna. Soil inverte-
brate community composition depends on the vegetation present, such that soil fauna will change
with the sequential development of vegetation on a site.
Animals, particularly insects, may indirectly affect litter decomposition via their effects on
the canopy. Severe insect defoliation of the overstory usually results in more light, warmth, and
moisture reaching the forest floor, all of which increase organic-matter decomposition and hence
the rate of nutrient cycling. Insect defoliation also may regulate nutrient cycling by increasing the
rate of litter fall, influencing the rate of nutrient leaching from foliage, stimulating the redistribu-
tion of nutrients within plants, increasing light penetration through foliage, and stimulating the
activity of decomposer organisms (Mattson and Addy 1975). Herbivores can also change litter
chemistry, which in turn affects decomposition, by disproportionally removing high-quality forage
(Ritchie et al. 1998; de Mazancourt and Loreau 2000) or by inducing changes in the nutrient or
secondary-chemical composition of infested plants that carry over to the litter. The effect of her-
bivory on decomposition rate may depend on the type of canopy being defoliated. Chapman et al.
(2006) hypothesized that herbivory should increase decomposition of evergreen litter fall because
leaves would reach the ground prematurely before nutrients could be translocated into the plant,
but should decrease decomposition of deciduous litter fall because the insects would induce
secondary compounds in infested plants.
The physical and chemical properties of soils are altered and often improved by soil-dwelling
invertebrates and burrowing mammals. In addition to mixing organic matter with mineral soil, soil
mesofauna improve soil aeration, moisture-holding capacity, and nutrient content. Soil structure
is altered by soil mesofauna (insects, centipedes and millipedes, isopods, earthworms) when they
ingest soil particles and mix them with finely ground and digested organic matter, forming small
but durable aggregates that improve soil structure when excreted. In addition, soil animals may
return considerable amounts of nutrients to the soil through their excreta and carcasses. Small
animals leave their feces and fragments of litter at different depths of the soil as they travel through
it. They improve aeration, porosity, and moisture-holding capacity by mixing humus into the sur-
face mineral soil.
Many of the same soil-animal processes that alter the physical properties of soil also influence
soil water. The various actions of soil-dwelling invertebrates and burrowing vertebrates increase
water percolation into the soil, and organic matter mixing into soil increases its water-holding
capacity. Severe defoliation by canopy insects limits interception and evapotranspiration and
thereby increases soil-water inputs from precipitation. Finally, the drainage system of lands adja-
cent to streams and small lakes may be significantly changed by dam-building activities of beavers
(Hammerson 1994). Animal effects on soil and nutrient cycling is treated in Chapters 9 and 19,
respectively. The effects of earthworms in particular, which are invasive species in North America,
are discussed in Chapter 21.
DAMAGE AND DEATH

Despite their positive influences on forest ecosystem structure and function, nearly every animal
group inhabiting forest land has been cited for some kind of damage to forests, be it by browsing,
chewing, gnawing, stripping, debarking, girdling, felling, or trampling woody plants. Animals
were natural thinning agents prior to European colonization whose populations fluctuated with
climate, food supply, and the populations of predators and parasites. Managers became particu-
larly concerned about damage by animals in North America in the early twentieth century in the
midst of major reforestation efforts. High levels of damage to plantations and naturally regenerated
stands typically followed European colonization and deforestation, which had disrupted the
natural site–plant–animal interrelationships of forest ecosystems. Management devoted to artificial
regeneration by planting must often contend with problems initiated by humans, such as lack of
predators and resulting high populations of herbivores. For example, ungulate browsing is prob-
ably the most common source of animal damage to forest regeneration in Europe and North
America. In many areas of eastern North America, white-tailed deer populations exploded follow-
ing the height of logging in the nineteenth or early twentieth century (Figure 12.9). Under such
conditions, stands may become park-like, a distinct browse line is maintained, and forest
100
Deer population – in percent of peak
75
50
Heavy forest cutting

25
Deer
laws
0 enacted
1900 1910 1920 1930 1940 1950 1960 1970
F I G U R E 1 2 . 9 Change in white tailed deer populations in Pennsylvania from 1900 to 1970. Source: Marquis
(1975) / U.S. Department of Agriculture / Public Domain.

F I G U R E 12 .1 0 Elk browsing has completely eliminated aspen regeneration outside a fenced exclosure on
the National Elk Refuge, Jackson Hole, Wyoming.
regeneration is sparse or may fail entirely (Tilghman 1989). A similar situation exists in western
North America with the predominance of elk, which have converted many forested areas in semi-
arid climates to grass- or sagebrush-dominated vegetation, as shown by experimental ungulate
exclosures (Figure 12.10). Regeneration failure of hardwood forests in Pennsylvania (Marquis 1975;
Tilghman 1989), lack of sapling regeneration of eastern hemlock in the western Great Lakes region
(Frelich and Lorimer 1985; Mladenoff and Steams 1993), and regeneration failure of trembling
aspen clones in the Intermountain West (Kay 1997) are all attributed in part to high ungulate popu-
lations. In central Europe, high deer populations have severely limited forest regeneration and
stand development for centuries. Most natural forest regeneration and plantings must be fenced
from deer for several years during and after establishment. Several well-known examples of the
drastic effects of ungulates on forest communities (often detailing the importance of predator con-
trol of ungulates) include deer on the Kaibab Plateau in Arizona (Binkley et al. 2006), elk in Yel-
lowstone National Park in Wyoming and Montana (Ripple et al. 2015), and moose in Isle Royale
National Park in Michigan (Wilmers et al. 2006).
Every forest tree is subject to animal damage, from seed and seedling to mature tree. Cones,
fruits, and seeds are destroyed by various insects that may eliminate regeneration completely in non-
mast years. Squirrels cut and collect conifer cones and cache them by the hundreds, storing them for
2 years or more. Mature seeds of many species may be consumed by birds and small mammals. In a
classic study of a western Oregon clear-cut, birds and mammals caused a 63% loss of Douglas-fir
seeds, whereas only a 16% loss was sustained by the smaller-seeded western hemlock, and seed loss
for western red cedar was negligible (Gashwiler 1967). In a 2-year study of the fate of seeds of the
relatively large-seeded ponderosa pine, only 4% of all seeds reaching maturity were available for ger-
mination (Table 12.1). Only the production of millions of seeds in excellent seed years may ensure
that enough survive even heavy losses to germinate and live.
Once seedlings are established, the threat of seed predation changes to the threat of clipping
and bark removal of seedlings by hares and rodents, stem and root girdling by beavers and various
rodents (Crouch 1976; Teipner et al. 1983), and trampling and browsing by deer and other
Table 12.1 Effect of animal predation on seed availability of ponderosa pine in western Montana.
Fate of seeds (per 100 reaching maturity) Number of seeds Percent

Seeds used by animals before dispersal 66 66
Seeds dispersed 34
Seeds used by animals after dispersal 30 88
Seeds available for germination 4
Source: Schmidt and Shearer (1971) / U.S. Department of Agriculture / Public Domain.
ungulates. Saplings and pole-sized trees are subject to browsing by ungulates including deer,
moose, and elk; girdling by beavers and porcupines; and defoliation by a variety of insects (Mattson
et al. 1991). Larger trees are subject to damage by sapsuckers that drill large holes in the bark,
killing or damaging a wide variety of orchard, shade, and forest trees.
Mature and overmature trees are subject to severe attack by a variety of insect feeders.
Endemic levels of insect herbivory of 5–30% of annual foliage typically do not impair tree growth
or total annual plant production (Franklin 1970; Mattson and Addy 1975). Under epidemic condi-
tions, however, some insects may consume 100% of the foliage. In North America, about 85 species
of free-feeding and leaf-mining insects (mostly caterpillars of butterflies and moths in the order
Lepidoptera) can exhibit outbreaks and cause serious and widespread defoliation of forest trees
(Mattson et al. 1991) exceeding 1000 contiguous hectares. Interestingly, these 85 species represent
only 1–2% of forest Lepidoptera. Insect outbreaks are most likely in very old stands having low net
primary production. For example, spruce-budworm outbreaks in the eastern and western United
States tend to occur in overmature balsam fir and spruce stands (Mattson and Addy 1975), annu-
ally affecting 5.1 million hectares of commercial forests in the United States (Haack and Byler 1993).
Insects typically act as one of many important disturbance agents (Chapter 16) and alone do not
usually cause the regeneration of senescent forests. Instead, insect effects work in conjunction
with fungi, fire, climatic stress, and windthrow to recycle aging forests, which then may be replaced
by fast-growing, productive young stands.
In summary, animals interact with physical site characteristics and forest trees throughout
their complete life cycle, beginning with pollination, seed dispersal, and establishment, through
stand development and thinning, to the death of old trees and stands (and in clonal plants, life
after death).
INFLUENCE OF LIVESTOCK ON FOREST ECOSYSTEMS

Humans have had the most far-reaching influence on forests of any animal through a variety of
direct and indirect activities, and many of these are covered in detail in Chapters 20-23. One such
activity, the introduction of hoofed grazing and browsing animals such as cattle, sheep, and other
large animals has substantially changed forest ecosystems. At least 34% of all the forested area in
the United States experiences grazing (Clason and Sharrow 2000). Around the world, severe graz-
ing by livestock has probably been more important than any other factor in reducing the produc-
tive capacity of uncultivated land. Despite its human influence, we have chosen to include the
effects of livestock in this chapter because of its pervasive influence on forests around the world
and its similar effects as those of native grazers and browsers discussed in the earlier text.
Heavy grazing pressure from a single livestock species can cause entire changes in the struc-
ture of the plant community. Most broadly, these changes often alter the dominance of woody
species compared to grasses. After European colonization of the Intermountain West in the United
States in the nineteenth century, overgrazing from cattle and sheep depleted the bunchgrass veg-
etation, reducing competition for woody species and the amount of fine fuels for fires critical to
Influence of Livestock on Forest Ecosystems 289
grassland persistence (Van Auken 2009). As a result, woody species such as big sagebrush, juni-
pers, bitterbrush, mesquite, and serviceberry increased greatly and converted grassland into brush-
land (Wagner 1969). In the twentieth century, however, widespread increases of native mule deer
placed heavy browsing pressure on these shrubs, often causing them to disappear and return the
dominant vegetation to the original bunchgrass type in many areas.
In the United States, grazing has greatly affected forest regeneration in heavily grazed wood-
lands of the East, central States, and the South, and throughout most of the open ponderosa pine
forests of the West. Though forests are not typically the intended targets of grazing compared to
rangelands, domestic ungulates inevitably wander into nearby or adjacent forests. In the West,
severe overgrazing by livestock in the nineteenth century affected forest composition, fire regimes,
and site conditions. For ponderosa pine forests of the Intermountain West, grazing altered both the
composition and quantity of ground-cover vegetation. Removal of grassy and herbaceous compe-
tition and exposure of mineral soil by livestock helped prepare the ground for dense thickets of
pine reproduction, and their removal reduced fine fuels that historically increased the likelihood
of fire. Together with the lack of fires to thin dense stands, many ponderosa pine forests soon grew
into almost impenetrable sapling thickets (Cooper 1960; Belsky and Blumenthal 1997). Grazing
has had similar effects on forest regeneration in riparian areas of the West. For example, in southern
Arizona, small seedlings of Fremont cottonwood were virtually eliminated by cattle grazing,
threatening the future of the species’ dominance in riparian areas (Glinski 1977). Riparian forests
dominated by willow and black cottonwood in Oregon had little if any cottonwood or alder regen-
eration due to grazing, effectively altering future species composition (Kauffmann et al. 1983). A
recent detailed treatment of livestock grazing effects on forest vegetation is given by Öllerer
et al. (2019).
In part through its effects on plant communities, livestock grazing has had documented
effects on animal biodiversity in many forest ecosystems, particularly in the West. For example,
intensive cattle grazing in riparian areas of southcentral New Mexico affected populations of the
acorn woodpecker (Ligon and Stacey 1995). The population decline was correlated with the loss of
nearly all large granary trees of narrowleaf cottonwood and a lack of middle-aged trees to take
their place. This “hole” in the tree-age distribution was attributed to a period of intensive cattle
grazing during which regeneration of young cottonwoods was suppressed. Similar effects of graz-
ing on animal populations are evident in small mammals, reptiles, fish, insects, and other birds
(see review in Fleischner 1994).
Direct effects of heavy grazing on physical site factors result largely from the action of animal
hoofs in compacting the surface soil. Consistent pounding from livestock hoofs breaks down soil
aggregates, creating a crumb structure to the surface soil. As a result, the pore space of the surface
soil is greatly reduced, greatly decreasing aeration and infiltration of rainwater, thereby causing
surface runoff and potential erosion (Ohmart and Anderson 1982; Kauffman and Krueger 1984;
Orodho et al. 1990). Erosion typically occurs by sheet erosion, although gullying has been docu-
mented for heavily grazed areas compared to ungrazed areas (e.g., Kauffman et al. 1983). For
example, surface soil bulk density increased 7–17% after 26 years of livestock grazing in an oak
woodland of the Sierra Nevada foothills in California (Tate et al. 2004) and by 6% for pine planta-
tions after 7–8 years of grazing in British Columbia (Krzic et al. 1999). In a controlled grazing
experiment of Douglas-fir forests and adjacent pastures in eastern Oregon, the pastures had higher
bulk density and lower porosity compared to the forests, and water infiltration was 38% less (Shar-
row 2007). Notably, these differences disappeared after 2 years when grazing was removed. Forests
are therefore likely to better withstand the negative impacts of grazing than rangelands, though
grazing effects in forests are still substantial.
Heavy livestock grazing is problematic for soil and water in the eastern United States just as
it is in the West. For example, in a classic study in southern Wisconsin, highly compacted soils of
heavily grazed woodlots had lower initial moisture content in the spring and dried out faster in the
summer and late fall because of lowered soil permeability and increased runoff (Steinbrenner 1951).
Water permeability of grazed soils averaged about only one-tenth of those of the ungrazed wood-
lots. However, the effects of livestock grazing on soil and water appear to be less in the eastern
United States compared to the drier regions of the West. Despite the typical problems of heavy
grazing pressures, livestock grazing in moderation can have negligible effects on forest soil and
water (Patric and Helvey 1986), and there is less evidence that woodland grazing in the East, as
typically practiced, has substantial adverse effects on water quality or on flooding in streams drain-
ing grazed woodlands.
A final point is that the effects of livestock grazing are not universally accepted as being neg-
ative (in contrast to overgrazing), and that many studies suggest that at least some of the often-
accepted ideas may be inconsistently supported with empirical data. For example, a study of
forest–grassland boundaries in Montana was not able to identify a relationship between livestock
grazing and tree encroachment into the grasslands (Sankey et al. 2006). Moreover, some studies
have argued that careful management of grazing may favor desirable vegetation structures and
compositions (Darabant et al. 2007; Kaufmann et al. 2013), suppress woody invasive species
(Chauchard et al. 2006; Mayerfeld et al. 2016), or aid in fuels reduction and fire mitigation (McEvoy
et al. 2006; Varela et al. 2018). It is notable that many such studies are European, emphasizing
potential cultural differences in attitudes and perceptions toward grazing, as well as how it is
applied, that may be difficult to overcome and/or inapplicable in North America.
SUGGESTED
READINGS
Chick, J.H., Cosgriff, R.J., and Gittinger, L.S. (2003). Fish on temperate forest vegetation–a global review. Biol.
as potential dispersal agents for floodplain plants: Conserv. 237: 209–219.
first evidence in North America. Can. J. Fish. Aquat. Stamp, N. (2003). Out of the quagmire of plant defense
Sci. 60: 1437–1439. hypotheses. Q. Rev. Biol. 78: 23–55.
Herms, D.A. and Mattson, W.J. (1992). The dilemma of Vander Wall, S.B. (2010). How plants manipulate the
plants: to grow or defend. Q. Rev. Biol. 67: 283–335. scatter-hoarding behaviour of seed-dispersing ani-
Janzen, D.H. (1971). Seed predation by animals. Annu. mals. Philos. Trans. R. Soc. B 365: 989–997.
Rev. Ecol. Syst. 2: 465–492. Whitham, T.G. and Mopper, S. (1985). Chronic herbiv-
Mattson, W.J. and Addy, N.D. (1975). Phytophagous ory: impacts on architecture and sex expression of
insects as regulators of forest primary production. pinyon pine. Science 228: 1089–1091.
Science 190: 515–522. Willmer, P. (2011). Pollination and Floral Ecology.
Öllerer, K., Varga, A., Kirby, K. et al. (2019). Beyond Princeton University Press 832 pp.
the obvious impact of domestic livestock grazing
Forest Communities CHAPTER 13
I n Chapter 1, we focused on organisms and the ecosystems to which they belong. Communities
and populations, however, are categorically different from ecosystems. Communities are
aggregates of organisms and are contained within ecosystems as one compositional compo-
nent, but they themselves are not ecosystems. In this chapter, we consider aggregates of plants
as one such ecosystem component. These collections of plants form the vegetative cover of forest
ecosystems—from the overstory trees to the successively lower vertical layers of shrubs, herbs,
ferns, and bryophytes, as well as the non-plant layers of lichens, fungi, and algae. We exam-
ine the community concept, communities as parts of landscape ecosystems, historical views of
community, mutualisms and competitive relationships among plants, and the vertical structure
of forests. These considerations lead to a discussion of disturbance in Chapter 16 and ecosystem
succession in Chapter 17.
COMMUNITY CONCEPT
A community, simply stated, refers to the assemblage of organisms within a particular area at a
given time. If only plants are considered, it is a plant community; if all organisms are considered,
it is a biotic community which at broad scales is a biome. Community is a very generic term of
convenience used to designate aggregates of organisms regardless of how broad or complicated
(Cain and de Oliveira Castro 1959). Aggregates of plants over a broad area may be considered a
community (the plant component of a biome), as may a local assemblage of plants associated with
a specific site. Beyond the co-occurrence of plants that occupy space and have a spatial boundary,
definitions of a community vary widely.
Early concepts of plant communities emphasized co-occurrence or mixtures (Clements 1905,
p. 316) of species, physiognomy (life form), organism-like properties, and classification as idealized
“types.” In their detailed study of the community concept, Shrader-Frechette and McCoy (1993)
observed that the community concept included many disparate ideas by the middle of the twenti-
eth century, and focused on communities as “units of interacting species and habitats” in the
second half of the century. Communities at this time were also defined as distinct ecological units,
and units of dynamic stability. Whittaker (1975, p. 359) even defined plant communities as living
systems. Nearly 70 years ago, Cragg (1953) observed that community had limitless meaning,
ranging “from a piece of shorthand denoting an assemblage of organisms to something endowed
with the attributes of organization which, in the absence of factual support, rivals the daydreams
of the alchemists.” McIntosh (1993) also observed that a clear definition of community has largely
avoided many ecologists. Why is the concept of a plant community so contentious?
An important key to confusion about the nature of “communities” is that they are not
ecosystems, ecological units, or living systems, but simply aggregations of organisms that happen
to occupy a common geographic space. Communities are composed of individuals and species
whose ecological amplitudes, mutualistic relationships, and competitive abilities allow them to
291
292 Chapter 13 Forest Communities
coexist (Rowe 1984a). They are inseparable parts of landscape ecosystems, but they are not systems
themselves as are organisms and ecosystems.
The problem of community definition is traceable to a primary focus on vegetation rather
than landscape ecosystems. Plants and animals evolve together within site-specific environments
(Chapter 3) as integral parts of ecosystems. Plant communities, therefore, appear to not only have
plant properties but those of environmental and ecological conditions, past and present, in which
the plants evolved. Community composition, physiognomy, vertical structure, and population
fluctuations result as much from the effects of site factors on plants and their reciprocal interac-
tions as they do from direct plant-to-plant interactions. We may regard plant life forms, patterns of
species distribution, and competitive abilities as “plant” properties, but they have been determined
in large part determined through interactions with the sites that support them.
It is well known that certain patterns of forest composition characterize extensive areas of
the Earth. A regional forest ecosystem dominated by spruce and fir trees as well as pine and larch
extends around much of Earth’s boreal zone. This forest vegetation is heterogeneous, yet has a
characteristic physiognomy and composition that make it immediately recognizable as a “spruce–
fir” community or forest type. Furthermore, many of the smaller plants and many of the animals
will be common to spruce–fir forests in different geographic locations. Notably, regional and local
climates differ from place to place, and because site conditions are not always the same, the tree,
shrub, and herb associates of spruce and fir will certainly differ from place to place, as may the
relative proportions of spruce and fir and the races of these species (Chapter 3). Obviously, no two
communities or ecosystems are exactly alike. Nevertheless, it is clear that an “acceptable likeness”
exists (Rowe 1966), so that we convey an immediately recognizable concept and mental image to
others when we speak of the boreal spruce–fir forest.
An important point is that this mental image is named by tree species, but is not simply veg-
etative cover. It includes regional ecosystems characterized by harsh climate and cold soil, support-
ing complicated biotic communities dominated by spruce and fir in the overstory. Both regional
and local ecosystems are conventionally named by the dominant trees (oak–hickory forest, beech–
maple forest), and thus we may lose sight of the more inclusive landscape ecosystem of which the
community is only one part. Moreover, assigning a community to a class may suggest a uniformity
that doesn’t really exist on the ground.
GROUNDING COMMUNITIES
As communities are integral parts of ecosystems that form landscapes at broad and local scales, it
is useful to examine community occurrence, composition, and structure in relation to their sup-
porting ecosystems. These broad vegetative units may seem discrete and with sharp boundaries,
but ecosystems and their communities form more or less continuous patterns or gradients in
nature, typically following closely those of changing site conditions.
Florida Keys Regional ecosystems and their communities contain considerable heterogeneity,
particularly viewable at local scales where variation is evident with the close match between site
and the vegetation it sustains. In the Florida Keys, community structure and composition reflect the
complex of factors and factor gradients, including physiographic position, hydrology, soil depth,
salinity, frequency of tidal inundation, fire regime, and hurricane effects. The plant communities of
recently undisturbed dry tropical forest, for example, are conceived as parts of mapped landscape
ecosystems (here termed ecological site units [ESUs]) along a physiographic diagram (Figure 13.1;
Ross et al. 1992). Such a map emphasizes that the pattern of communities is intimately related to the
physical configuration of the island. Although this view shows spatially patterned ecosystems as if
they were all discrete units, the concrete systems and their communities vary from place to place
and are not homogeneous. In addition, each ecosystem and its community may grade gradually or
Community Concept 293
ECOLOGICAL SITE UNIT MAP OF UPPER SUGARLOAF KEY

WITH LOCATION OF TRANSECT A–B
ECOLOGICAL
SITE UNITS
BO
W
1
C
H
2
AN
3
N
4
EL
7
8
B 9
10
WATER
A
N
KILOMETERS
PERKY LAKE
0 0.2 0.4 0.6 0.8
PROFILE OF TRANSECT A–B

UPPER SUGARLOAF KEY
ESU 9 ESU 9
ESU 10
ESU 10 ESU 1
ESU 10
ESU 7 ESU 2
ELEVATION (m) ESU 8 ESU 8
2.00 ESU 4
ESU 3
1.00
0
0 200 400 600 800 1000 12000
A Distance (m) B
Groundwater
Unsaturated soil Salinity 0–10‰
and/or rock Salinity 10–20‰
Salinity >20‰
F I G U R E 13 .1 Ecological site units (landscape ecosystem types), geomorphology, and ground water
characteristics for Upper Sugarloaf Key, Florida. Upper diagram is a map of the ecological site units, and
the lower diagram is a cross section of part of the island along the transect line A–B shown on the map.
Source: Ross et al. (1992) / Association for Tropical Biology and Conservation.
abruptly into those surrounding it. The nature of the transition zone or ecotone between ecosys-
tems depends on site, disturbance, and community factors. In Figure 13.1, abrupt transitions are
evident between ecosystems (ESUs 1 and 2, 7 and 9, 3 and 8 on the lower physiographic diagram) as
well as gradual changes (ESUs 9 and 10, 8 and 10). Notably, the number of plant species in the
respective communities varies greatly, from the relatively species-poor mangrove units (units 1–3)
with only 5 species, to the species-rich pine rockland forest (unit 9) with 35 species.
Interior Alaska Viereck et al. (1984) studied communities and ecosystems in the taiga of interior
Alaska (Figure 13.2). They first developed a vegetation classification, then an ecosystem type
classification including physiography, soil, and other site factors associated with the vegetation.
A group of communities near Fairbanks illustrates the diversity of early-and late-successional stands
ranging from river floodplains and relatively warm southern slopes to the coldest north slopes at tree
line. The communities are clearly related to physiographic (landforms of river valley and upland flats,
and slopes, aspects, and elevation) and edaphic conditions (Figure 13.2). The warmest sites are occu-
pied by deciduous angiosperms (balsam poplar, quaking aspen, paper birch) that form early-
successional stands following fire. Aspen and birch communities (stands 15, 13, and 16 in Figure 13.2)
eventually succeed to white spruce on well-drained, loess soil. Both lowland and upland black spruce-
dominated ecosystems are evident. As in the Florida Keys, some but not all communities are abruptly
separated due to site conditions (for example, balsam poplar on alluvial soil and the adjacent white
spruce on permafrost; the black spruce muskeg versus the droughty bluff with quaking aspen). By
contrast, the transition from forest to tree line and alpine tundra is more continuous and gradual.
Southern Illinois In a midcontinental region further different from Florida or Alaska in geol-
ogy and macroclimate, local communities remain closely related to the interactions of geologic
substrate, slope position, soil, microclimate, and disturbance regime. In the Shawnee Hills of
N UPLAND RIVER FLOOD PLAIN

Alpine Tundra
North aspect, shallow soil
Southeast aspect,
Well drained
Flood plain, well-drained

Tree line
Flood plain, well-drained

Tree line
Saddle, poorly drained

poorly drained
Well drained
Muskeg, poorly drained

Bluff, droughty soil
Well drained
alluvial soil
alluvial soil
Loess
Permafrost
River
Permafrost Loess
B E D R OC K
Black Spruce - Intensive Site No. 33, 34
Permafrost
Gravel
White Spruce - Black Spruce No. 23
White Spruce - Black Spruce No. 28
Balsam Poplar No. 24, 25, 32

Black Spruce No. 20, 29, 30
White Spruce No. 26, 27, 31

White Spruce No. 12, 22
Paper Birch No. 13, 16
Quaking Aspen No. 11

Quaking Aspen No. 15
Black Spruce No. 19

Black Spruce no. 18
Black Spruce no. 21
Black Spruce no. 14
Black Spruce White Spruce Paper Birch

Aspen Balsam Poplar Alder, Willow
Dwarf Birch
F I G U R E 1 3 .2 Forest communities of landscape ecosystems in interior Alaska. Shown is a physiographic

diagram of the topography, landforms, vegetation, and parent material in the Fairbanks, Alaska area with
locations of sampled stands. Source: Viereck et al. (1984) / U.S. Department of Agriculture / Public Domian.
Community Concept 295
southern Illinois, communities and local landscape ecosystems may be distinguished on a physio-
graphic cross section of ridge and valley terrain (Figure 13.3; Fralish et al. 1991). Dominant forest
cover types during both pre-European colonization (prior to 1800; Figure 13.3a) and “old-growth”
forest cover types of 1988 that developed during European colonization characterized by fire sup-
pression (Figure 13.3b) were examined in the same physiographic setting. The ecosystem types of
the pre-European colonization period include: (i) shallow soil over bedrock, (ii) rocky upper slope,
(iii) loess-covered ridge, (iv) upper north slope, (v) lower north slope, (vi) alluvial floodplain, (vii)
terraces, and (viii) mid-south slope (Figure 13.3a). Geologic substrate, landform, and local topog-
raphy of this area change relatively little over hundreds to several thousands of years and form the
basis for distinguishing plant communities that reflect site conditions. Both gradual and abrupt
changes are evident between ecosystems and their cover types. Eastern redcedar and post oak
dominate the driest sites, white and black oaks dominate the fire-prone ridges and south slopes,
and river birch (Bn in Figure 13.3) and American sycamore (Po) dominate alluvial floodplains. The
white oak–black oak community was found in three ecosystems due to a relatively high fire fre-
quency: loess-covered ridge, upper north slope, and lower north slope.
The old-growth forest cover types that exist today on the area (Figure 13.3b) are, for the most
part, the same as those present prior to European colonization. The spatial positions of landform and
soil are relatively unchanged such that ecosystem types have not changed. However, lack of fire has
allowed more mesophytic species such as red oak and sugar maple to replace the white oak–black oak
overstory on upper north slopes and lower north slopes, respectively (Figure 13.3b). Red oak has also
replaced black oak in the loess-covered ridge ecosystem. Changes in these forest cover types illustrate
(a) SOUTH SLOPE RIDGE NORTH SLOPE (b) SOUTH SLOPE RIDGE NORTH SLOPE
WHITE WHITE & BLACK WHITE WHITE & RED
OAK OAKS OAK OAKS
RED POST RED POST
CEDAR OAK CEDAR OAK
WHITE & BLACK Loess RED
OAKS OAK
Sandstone Sandstone
Bedrock Bedrock
POST OAK POST OAK

WHITE & BLACK
WHITE
OAKS
WHITE & BLACK OAK
COVE SUGAR MAPLE COVE
OAKS
WHITE & RED OAKS,
MIXED HARDWOOD SUGAR MAPLE
Po Po
Bn Bn
Terraces Terraces
Sandstone Fragipan Sandstone Fragipan

Bedrock Sandstone Bedrock Sandstone
fragments fragments
Pre-European colonization forests prior to 1800 Old-growth Forests of 1988
F I G U R E 13 .3 Physiographic diagram of the pattern of (a) pre-European colonization and (b) old-growth
ecosystem types and forest cover types in the Shawnee Hills region of southern Illinois. Ecosystem types
(cover types) include: (i) shallow soil over bedrock (eastern redcedar), (ii) rocky upper slope (post oak and
white oak), (iii) loess-covered ridge (white and black oaks), (iv) upper north slope (white oak), (v) lower
south slope (white oak), (vi) stream terraces (white and northern red oaks, sugar maple, basswood),
(vii) alluvial floodplain (American sycamore and river birch), and (viii) mid-south slope (white and black
oaks). Pre-European colonization communities before 1800; old-growth communities as of 1988. Source:
Fralish et al. (1991) / with permission of The University of Notre Dame. © 1988, American Midland
Naturalist. Reprinted with permission of the American Midland Naturalist.
the dynamic nature of vegetation in relation to disturbance, or lack thereof in the case of fire (Fralish
et al. 1991). Notably, tree communities of some ecosystems are virtually unchanged. Thus, the compo-
sition of forest communities is highly dependent on the interaction of specific site conditions and
disturbance regimes that influence the tree species’ competitive ability on the Shawnee Hills landscape.
In summary, plant communities are visually distinctive features of the landscape. Under-
standing their composition, stand structure, dynamics, and spatial distribution for any landscape
requires a solid understanding of the site conditions, disturbance regimes, history, and processes
of the landscape ecosystems of which they are an inherent part.
VIEW FROM THE PAST: COMMUNITY CONCEPTS

The paradigm of the community as one component of landscape ecosystems differs from other
community concepts because of its first-order focus on ecosystems. The historical focus of plant
ecologists was on plant species and communities, their distribution, composition, and classification,
and less on their relationship to site factors. Many of the early terms and underlying concepts of
plant ecologists remain entrenched in the literature and used today.
Early plant geographers, such as Humboldt (see Botting 1973; McIntyre 1985), Schimper (1898),
Gradmann (1898), and Warming (1909) emphasized the geographical distribution of plants (i.e., phy-
togeography) and classification. For example, Warming (1909), a critical figure in shaping modern
ecology, developed a classification based on water plants and land plants, with the latter subdivided
further into 12 primary groups. Henry Cowles, the noted American ecologist and a keen field ecolo-
gist and botanist, pioneered a physiographic classification of plant communities of the Chicago area
(Cowles 1901). Cowles recognized that physiography influenced climate, soil, and geomorphic change
(erosion, deposition), and therefore best accounted for the distribution of species and communities.
Cowles purported to examine communities, but his emphasis was essentially on ecosystem units and
their geomorphic change to which characteristic plants and communities were dynamically related.
SCHOOLS AND TERMINOLOGY

The field of community ecology was dominated by two schools in North America in the first half or
two-thirds of the twentieth century. Frederic Clements (1905, 1916; Weaver and Clements 1929) first
developed a school based on plant succession, followed by the Curtis–Whittaker group who empha-
sized an abstract vegetative continuum (Curtis 1959; Whittaker 1962, 1975; Fralish et al. 1993). Clem-
ents had a broad, comprehensive understanding of the historical development of the time, innova-
tive ecological insights (especially regarding physiography and site–plant relations), and a complex
emphasis on succession and vegetation dynamics despite the fact that his theories of vegetation as an
organism, rigid classification, and monoclimax later became unacceptable (see Chapter 17).
Like many of his predecessors, Clements focused on the plant formation as the major unit
of vegetation, characterized by a given physiognomic form that recurs on similar sites (i.e., grass-
land, temperate deciduous forest, tropical rain forest). Each formation was composed of various
other distinctive community types, termed associations. Thus, the deciduous forest formation
was composed of several different associations such as beech–maple, oak–hickory, and others.
Braun (1950) mapped and described such associations as “forest regions” (Figure 2.6, Chapter 2).
Küchler (1964) published a non-hierarchical classification of the “potential” vegetation units of
the conterminous United States. In Canada, Rowe (1972) described forest regions, and Strong et al.
(1990) summarized its vegetation zones. A worldwide classification of plant formations was pub-
lished by Mueller-Dombois and Ellenberg (1974).
Europeans and Americans followed different courses in the use of the term “association”
(Mueller-Dombois and Ellenberg 1974). Partly because the term was strongly identified with its rigid
European definition and usage, and partly because Clements related it only to “climax” communities,
View from the Past: Community Concepts 297
“community” was soon used to replace “association” among American ecologists. The accuracy of
“community” is still questioned today, and alternatives such as plant assemblage are often used. Its
particular meaning in a given study depends on the context in which it is used and the modifiers
applied to it—oak–pine community, early-successional community, etc. American ecologists often
further subdivide the community or association into “layers.” Thus, the spruce–fir formation of north-
eastern United States and eastern Canada includes the red spruce–balsam fir community, which may
in turn be subdivided into an overstory or tree layer, an understory or shrub layer, and a groundcover
layer. Notably, the additional term forest type (or community type) refers to a forest community
defined only by overstory composition. Because the community is or should be defined by the total
plant complement, its name sometimes takes into account characteristic groundcover plants.
Concepts of Clements and Gleason Plant ecology in the early twentieth century was con-
ceptually dominated by the prevailing ideas of the plant association as a natural unit for study and
classification. Clements’s version of ecology was very tidy and orderly—far more orderly than na-
ture itself (Egler 1968). Although today we understand that all populations have inherent hetero-
geneity, at that time populations were thought to be more or less uniform such that it was easy to
believe in uniformity of plant formations and communities. Clements’s extreme view likened the
formation and association to a complex super-organism that arises, grows, matures, reproduces,
and dies. The final, stable, self-maintaining, and self-reproducing state in the development of veg-
etation was the climax (climatic maximum). Clements believed that only one true climax occurred
in a given climatic region, although common sense forced him to recognize and name the other
units that actually occurred with even more terms such as pre-climax, post-climax, and dis-climax.
A less extreme view held by plant ecologists such as Cooper (1913, 1923), Nichols (1923), and many
European phytosociologists was that the association was in fact a series of separate similar units
rather than a super-organism, variable in size but repeated in numerous examples, and analogous
to a species.
By sharp contrast, Henry Gleason developed an “individualistic concept” of ecology and the
plant association (Gleason 1917, 1926, 1939). Gleason objected to an organismal concept of vege-
tation as well as the classification of vegetation into rigid “pigeon-holes” of seemingly uniform
types. He emphasized the variability of communities that were supposed to be of the same type,
arguing that the variation occurred throughout both space and time due to chance and environ-
mental effects. He described the vegetation unit as a temporary and fluctuating phenomenon, its
origin, its structure, and disappearance dependent on the selective action of the physical environ-
ment and on the nature of the surrounding vegetation (Gleason 1939).
Gleason was a field botanist by training, and he advocated a floristic, individualistic approach
to vegetation (Nicolson 1990). Gleason used “individualistic” at two levels of organization: (i) for
individuals that make up a species, and (ii) for species as individualistic components of a
community (Egler 1968). He emphasized that the mixture of species in a community results from
both migration and environmental selection, and that a species’ spatial distribution depends upon
its individual peculiarities of these two major factors. Gleason (1936) noted that the true nature of
the association is determined by “the physical environment, which decides what kinds of plants
may exist in it, and the living plants themselves, which tend to control and to modify the physical
environment. Individuals of a species grow wherever they find favorable conditions, disappear
from areas where the environment is no longer endurable, and occur in company with any other
species of similar environmental requirements.” Notably, Gleason did not recognize or consider
the many mutualisms pervasive in communities, which suggests that individuals and species
aren’t the laws unto themselves to the degree he may have assumed.
Gleason’s argument for individuality of communities is strongest at regional scales where
differences are most evident. He asserted that similar environments within a region result in
similar floras that could be classified together and noted that abrupt and gradual changes
in environmental factors were often what caused similar changes among different communities.
Gleason’s floristic approach was revolutionary in stressing variation, the importance of chance
events in plant colonization, and the influence of vegetation on the site. However, Gleason’s
approach continued a primary focus on species and communities rather than the ecosystems to
which these organisms belong. From a landscape ecosystem perspective, communities aren’t
necessarily individualistic because they are influenced by the communities that surround them.
For example, forest communities of dry-mesic ecosystems protected from fire by adjacent
swamps (Chapter 10) have different composition and structure compared to those on unpro-
tected sites. Likewise, communities on mesic sites are often burned and changed by fires
spreading from adjacent flammable communities. Communities of upper-slope ecosystems
influence those on adjacent lower slopes through movement of wind-blown leaf litter and by
water and nutrients flowing downslope. Many such exchanges, from one community to another
through landform and soil-based linkages, provide a new perspective on the individualistic con-
cept of the community.
Phytosociology in Europe About the same time Clementsian ideas of the community were
taking root, the “association” was technically defined by ecologists at the Zürich-Montpellier
School of Phytosociology and adopted at the International Botanical Congress of 1910 as “a plant
community of definite floristic composition, presenting a uniform physiognomy, and growing in
uniform habitat conditions.” The school of phytosociology identified with Braun-Blanquet
(1921, 1964) had a primary objective of hierarchical vegetation classification, and the association
was the basic unit of this system. Braun’s approach was established by 1921 when he urged a flo-
ristic rather than an ecological classification system (van der Maarel 1975), with plant commu-
nities conceived as vegetation types recognized by their floristic composition. In this system, the
association has a type specimen complete with author, date, and description analogous to that of a
plant species: for example, “Abieti-Fagetum Oberdorfer 38”; and the “Galio-Carpinetum (Buck-
Feuct 37) Oberdorfer 57 em. Th. Müll. 66.” Interestingly, Braun never adopted an organismal
approach (van der Maarel 1975), and there is little justification for the naming of communities as
if they were species. Such a floristic approach has made important contributions to the description
and classification of vegetation, with hundreds of thousands of communities described (Ellen-
berg 1988). The approach is widely used in eastern and central Europe (Oberdorfer 1990) and by
many plant sociologists throughout the world. Excellent descriptions and perspectives of the
Braun-Blanquet approach and the Zürich-Montpellier School of Phytosociology are available
(Poore 1955; Becking 1957; Whittaker 1962; Shimwell 1971; Westhoff and van der Maarel 1973;
Mueller-Dombois and Ellenberg 1974; van der Maarel 1975; Podani 2006; Pott 2011; Mirkin
et al. 2015; Guarino et al. 2018).
CONTINUUM CONCEPT
The continuum concept of vegetation is the second major school of plant ecology in the United
States, including both the continuum approach of John Curtis and associates and the gradient
analysis approach of Robert Whittaker (Curtis and McIntosh 1951; Curtis 1959; Whittak-
er 1962, 1967; Cottam and McIntosh 1966; McIntosh 1967, 1968, 1993). The continuum approach
is an extension of Gleason’s floristic–individualistic concept of the community, which was over-
shadowed by the Clementsian and European association concepts by the 1940s. The concept of a
vegetation continuum was developed in reaction to the concept of the association as a relatively
discrete unit. The basis for the continuum concept is that vegetation varies continuously across a
landscape (Curtis 1959; McIntosh 1968). Notable plant ecologists such as Daubenmire (1966)
accepted that vegetation varied continuously but struggled with whether the transitional areas
(ecotones) between obvious communities could themselves be classified.
Community as a Landscape Ecosystem Property 299
There is little question, even by continuum critics, that vegetation or floras are continua
(Daubenmire 1966). Vegetation varies continuously because (i) changes in environmental condi-
tions vary, thus affecting the establishment and composition of the vegetation; (ii) genecological
variation of tree species is typically clinal, and a given genotype often has wide phenotypic plastic-
ity to tolerate a number of environments (Chapter 3); (iii) historical and chance events may rein-
force points 1 and 2 so that no communities are exactly alike; and (iv) a continuum of successional
change in time is superimposed upon changes in composition in space. Furthermore, whether
discontinuity or relative continuity may be demonstrated for a tract of vegetation in the field may
depend on methods of field sampling or analysis.
The opposition between those who favored classification and those who favored continua
was cast as a dichotomy of two extremes, suggesting that a middle ground would soon prevail. Log-
ically, studies of communities in nature would reveal that both gradually changing species aggre-
gations and relatively abrupt changes and acceptably similar compositional types exist (just as
Gleason had observed). This was not the case, however, because the continuum concept empha-
sized the arrangement of communities in abstract space, as models, not necessarily in nature. Lam-
bert and Dale (1964) recognized that many ecologists had become confused with continuum and
continuity because the terms were applied both to the actual continuity on the ground and to the
continuity in abstract models. Both Curtis (1959) and McIntosh (1967, 1993) recognized the dis-
continuities or abrupt changes of vegetation type but held them as irrelevant and not at odds with
the abstract continuum.
In retrospect, the continuum approach changed the viewpoint of many plant ecologists,
encouraged countless compositional studies of vegetation, and introduced increasingly sophisti-
cated multivariate analyses to classify and ordinate species and communities. Importantly, the
approach also generated controversy and confusion (Dansereau 1968; Austin 1985), and a
subsequent loss of interest in generalizations about plant communities. The approach contributed
to the notion that plant communities are too intricate and complicated to observe any general pat-
terns (Noy-Meir and van der Maarel 1987). Another perspective, however, is that as a subject of
study the plant community is unrewarding because it is incomplete, and, like climate, physiogra-
phy, or soil, is only one part of the landscape ecosystem. Communities are not functional systems
and therefore have no processes in isolation, but in fact function within and as parts of ecosystems
from which they derive their resources. It is these communities as ecosystem parts that ecologists
and managers seek to manage, conserve, and restore.
COMMUNITY AS A LANDSCAPE ECOSYSTEM PROPERTY

Austin and Smith (1989) reinforced the idea that communities are properties of ecosystems in re-
formulating the continuum concept. Figure 13.4 shows a landscape with four species present. Five
species associations may be recognized due to their frequency of occurrence along a transect from
low to high elevation: A, AB, B, C, and D (seen along x-axis in Figure 13.4). The combinations BC
and CD are regarded as transitions or ecotones. The “communities” composed of coexisting species
result from the site-specific conditions and spatial pattern of this landscape. However, the eleva-
tional gradient shows a continuum of species regularly replacing one another in the sequence A, B,
C, and D with increasing elevation (right side of diagram, Figure 13.4). Thus, what appear as
relatively distinct species associations horizontally along the x-axis of distance can be viewed as a
continuum along the abstract elevational gradient.
If the transect in Figure 13.4 were taken through an adjacent area where the first bench or
terrace were 30 m lower, then the combination AB would become rare because only species
A would be present. Similarly, if the second bench was 30 m higher, the communities recognized
would be A, B, BC, C, and D with ecotones AB and CD and the combination BC would become
common. Thus, communities are a function of the whole landscape and specific landscape
600 C
B
ENVIRONMENTAL
Altitude (m)
400 DISTRIBUTION
OF
A SPECIES
200
0
A
B Distance
C
D
SPATIAL DISTRIBUTION OF SPECIES
F I G U R E 1 3 .4 Pattern of co-occurrence of four species on a landscape along an indirect environmental

gradient of elevation. Note the distribution of species along the x-axis (distance) illustrating distinctive
species associations of A, AB, B, C, and D that occur at characteristic positions on the diagram. Then observe
the continuous variation of composition (regular replacement of A, B, C, and D) along the elevational
gradient on the right side of diagram. Elevation is an indirect gradient because its effect on growth is through
the site factors of temperature, water, and nutrients. Source: Austin and Smith (1989) / Springer Nature.
© 1989 Kluwer Academic Publishers. Reprinted by permission of Kluwer Academic Publishers.
e cosystems, and either abrupt changes or gradual transitions may occur depending on the pattern
of recurring ecosystems in a given landscape. Co-occurring groups of species can be recognized for
any particular area and may be recurring. Labeling or classifying these communities is useful for
communication, management, and research, but extrapolation of these communities to other
regions or landscapes will be accurate only if the regions have similar patterns of site factors (Aus-
tin and Smith 1989).
The elevation continuum in this example would be valid and applicable within similar tran-
sects of the same regional ecosystem, but not in a different region where the growth-influencing
variables of climate (temperature, water) relate differently to elevation. Austin and Smith (1989)
concluded that: (i) the concept of co-occurring species is only relevant to a particular landscape
and its pattern of combinations of site factors, and (ii) the continuum concept applies only to
abstract environmental space, not necessarily to any geographic distance on the ground or to any
indirect environmental gradient (such as elevation). Further details of this model of the contin-
uum concept are provided by Austin (1990).
EXAMPLES OF SPATIAL VARIATION IN FOREST COMMUNITIES

DISCRETE FOREST COMMUNITIES
Forests are probably most accurately considered in terms of gradient patterns. What appear to
be sharp boundaries between forest communities are likely due to abrupt local changes in
environmental gradients or to a completely different vegetational history of the two commu-
nities. Discrete communities provide evidence of the existence of discrete differences in past or
present environmental conditions. The boundary between two communities is more commonly
Examples of Spatial Variation in Forest Communities 301
a belt or zone which may vary in width. In the forest–grassland transition, for example, there
will always be an outer belt of forest that will be modified by the adjacent open areas, and an
inner belt of grassland that will be modified by the adjacent forest. As mentioned in the earlier
text, the transition zone between two communities is termed an ecotone, which usually
embodies some of the ecological features of the two communities, but often has a specific site
characteristic of its own.
Persisting, abrupt site differences often give rise to sharp forest type boundaries. Examples
of such site differences include: (i) a sharp boundary between two geological formations produc-
ing two contrasting soil types that differ in water and nutrients available to vegetation; (ii) a sharp
boundary between landforms that results in abrupt differences in soil drainage conditions, such
as between a poorly drained swamp and a well-drained upland; (iii) a sharp boundary in topo-
graphic position that affects local climate, such as a knife-edged ridge separating a north and a
south slope; and (iv) a sharp boundary in vegetation structure that affects the local climate and
soil conditions, such as a forest edge-facing grassland, a shrub community impinging upon open
rock surfaces, or a logging boundary between a clear-cut and a standing forest. Sharp boundaries
between plant communities may also arise from historical events such as fires, tornadoes and
other windstorms, salt spray from the sea, fumes from smelters, logging, and agricultural
development of land.
Coastal California: Giant and Pygmy Forests Soil scientist Hans Jenny and others (1969,
1980) provided a remarkable example of diverse ecosystems and communities related to landforms
along the Mendocino coast of north-central California. Rising seas during the Pleistocene cut ter-
races into the sandstone, and retreating seas covered them with beach sands, gravels, and clays.
The terraces were elevated by tectonic forces and today occur at elevations of approximately 30, 53,
91, 130, and 198 m above sea level (Figure 13.5). Hillsides and steep canyon slopes are dominated
by redwood and Douglas-fir, whereas the three upper, oldest terraces have distinctive hardpan soils
and support a sparse, dwarf or pygmy forest of slender, stunted cypresses and dwarfed pines
(bishop pine and Mendocino White Plains lodgepole pine). The soil is a spodosol (see Chapter 9)
Bi
Rw, Df Py
Py
3
Wave-cut
platforms
Gr
Dunes
2
Sea Beach deposits
1 Nickpoint Hardpan
Sandstone
F I G U R E 13.5 Physiographic arrangement of four marine terraces (1, 2, 3, and 4) in the Fort Bragg,
California area, with a young dune on the second and a very old dune on the fourth terrace. Gr = grassland;
Rw, Df = redwood–Douglas-fir forest; Bi = bishop pine forest; Py = pygmy forest. Horizontal distance is
4.8 km, and vertical distance is 152 m above sea level. Source: Jenny et al. (1969) / California Botanical Society.
characterized by extreme acidity and a thick hardpan layer of iron or clay. Water accumulates
above the hardpan layer in late fall, forming a perched water table that may flood the entire sur-
face. The surface water table disappears in late spring by evapotranspiration and seepage, and the
hardpan dries out, hardens, and imparts extreme xeric conditions in summer. The distinctive veg-
etation probably results from many physical and chemical factors associated with soil development
in the parent material of the terrace.
The lowermost terrace supports grass and pine forests, and the second terrace supports a
pattern of different communities from redwood, Douglas-fir, and western hemlock to dwarf pines,
depending on soil development. Neither of the lower two terraces has experienced soil development
as advanced as the upper terraces, because they are of more recent origin. The communities of the
terraces are therefore strongly dependent on physiography—the flat terrace form and its parent
material on which soil development occurs.
Sand dunes are present on all terraces, exhibiting three relatively distinct plant communities
depending on age and soil development. Recent dunes on the lowest terrace are still moving inland
and are stabilized by lupines and lodgepole and bishop pines. Young dunes on the second terrace
are thousands of years old and even older on the third terrace; these dunes have developed favor-
able water and nutrient conditions to support magnificent forests of redwood, Douglas-fir, and
grand fir. Very old dunes of the fourth and fifth terraces are highly acidic and infertile, but have no
hardpan formation. Very old dunes support bishop pines with a dense understory of ericaceous
species, wax myrtle, and chinquapin. Thus, three kinds of dune ecosystem types and their commu-
nities are evident due to their age and soil development.
Forest–Grassland Ecotone Abrupt changes between forest and grassland in the tropical
and temperate zones may or may not be associated with abrupt changes in Site. Once the forest
edge has been established, such as by fire or land clearing, contrasting climatic and soil conditions
between the forest and grassland may itself perpetuate the forest border. Grassland often originates
in forested areas following fire that kills trees and thereby creates an environment at the ground
more suitable for the development of grasses than for tree regeneration. The grassland persists
once established because of the inability of the adjacent forest trees to colonize the site—whether
due to the recurrent incidence of fires (Wells 1965, 1970; Rowe 1966; Veblen and Lorenz 1988);
failure of tree seedlings to penetrate the sod and reach a suitable medium for establishment (Jaku-
bos and Romme 1993); excessive root competition with grasses for soil water; or the absence of
mycorrhizae (Langford and Buell 1969). Examples of fire-created grasslands that have persisted for
hundreds of years include the alpine meadows of the western American mountains, the fingers of
prairie extending into the Black Hills of South Dakota, and the extension of the Prairie Peninsula
east into Indiana and Ohio, all under climates suitable for tree growth.
Importantly, forest–grassland borders are not static, and colonization of one type occurs by
the other, as climatic conditions have fluctuated over geologic time. For example, subalpine
meadows in Yellowstone National Park have been experiencing colonization by lodgepole pine
from adjacent forests for over a century (Jakubos and Romme 1993), probably due to a regional
climatic trend toward warmer and wetter growing seasons since about 1870. In many other parts
of the world, grasslands within forested regions are being colonized by forest both because of more
efficient fire suppression and because of changing climate.
Alpine Tree Lines The tree line of forests in mountain ranges, which results from the interac-
tion of trees and the site over a long period of time, is an obvious example of an abrupt forest edge.
Ecologists distinguish three kinds of altitudinal tree lines (Huggett 1995): the forest line or tim-
berline, which is the upper limit of tall, erect trees growing at normal forest densities; the tree
line, or the limit at which individuals recognized as trees (>2 m) grow; and the tree-species line,
Examples of Spatial Variation in Forest Communities 303
or the point up to which tree species will grow only in deformed habit (Krummholz or “crooked
wood”). Tree lines are not always sharp (Armand 1992), but the discontinuity between trees and
low-growing vegetation is indeed abrupt when two or three of these lines coincide.
Many explanations have been proposed for the formation of timberline (Daubenmire 1954;
Wardle 1985; Huggett 1995). In the zone of stunted and recumbent trees, wind, wind-blown snow,
snowpack, and other factors produce a harsh and exposed zone near the ground through which
trees cannot grow normally. Timberline location (as contrasted with climatic factors that cause
dwarfing and recumbent growth) is primarily determined by heat deficiency during the growing
season at high altitudes, which limits growth and winter-hardening of shoots. Wardle (1968) found
this to be the case for timberline in Colorado, where timberlines are among the highest in the
world despite desiccating wind and low winter temperatures. He also found that wind blew
Colorado sites free of snow, leaving seedlings unprotected during the winter and an absence of
melt water in the spring to moisten the rocky, coarse-grained soils. The causes of tree lines are
much debated, but most ecologists believe that climate is their main determinant (Huggett 1995).
MERGING FOREST COMMUNITIES

As much as distinct plant communities reflect abrupt changes in site, major disturbance, and land
history, gradual changes in site or vegetational history are expressed as similarly gradual changes
in forest composition. Forests of a generally similar history may exhibit such differences over a
geographical stretch of many kilometers, or over gradual changes in elevation, aspect, soil water,
and soil fertility. Climate also affects forest composition, and the relative continuity of macrocli-
mate in space and time favors gradual change. In the absence of an abrupt change in site of vege-
tational history, the composition and structure of forests generally vary continuously.
Eastern Deciduous Forest—Southern Appalachians The forests of the eastern United

States, characterized by deciduous hardwood species but also containing evergreen hardwoods
and conifers (Barnes 1991), show many gradual changes at both broad and local scales. The greatest
complexity and size of individual trees is found in the southern Appalachian region and Cumber-
land Plateau to the west in Kentucky and Tennessee, in which old, eroded slopes, valleys, and
ridges provide many ecological gradients of forest composition.
Whittaker (1956) studied composition gradients with altitude and with “moisture gradients”
within altitudinal belts in the Great Smoky Mountains. His “moisture gradient” referred to the
complex gradient that occurs from valley bottoms to dry slopes, but not its causation. Along this
gradient the dominant trees varied from mesic species in the valley bottoms to xeric species on the
dry, exposed portions of slopes. For example, between 750 and 1050 m in elevation, eastern hemlock
was most abundant in the bottoms, with silver bell, red maple, chestnut oak, scarlet oak, pitch
pine, and table-mountain pine each becoming more numerous toward the drier end of the gradi-
ent. On mesic sites, hemlock and red maple were most abundant at low elevations (600–900 m)
with silver bell, yellow birch, sugar maple, and basswood reaching maximum abundance between
900 and 1220 m, and yellow buckeye, mountain maple, and American beech most common at
higher elevations. Whittaker thus found the broad forest pattern of the Great Smoky Mountains to
be one of continuous gradation of stands along these generalized gradients, yet with certain
relatively discontinuous types (Figure 13.6). Notably, he recognized the existence of sharp discon-
tinuities between cove forests and beech stands on south-facing slopes, between grassy balds and
forests, and between heath balds and spruce–fir forests. As described by Whittaker (1962):
The whole pattern was conceived to be a complex continuum of populations, with the relatively discontinuous
types confined to “extreme” environments and forming a minor part of the whole. Allowing for discontinuities
produced by disturbance and environmental discontinuity, the vegetation pattern could be regarded as a com-
plex mixture of continuity and relative discontinuity.
Meters Feet
1980 6500 F HB
GB
1680 5500 SF SF WOC
BG
S ROC
1370 4500 ROC H H
P ROC CF H
1070 3500
OCF OCF OCH
OCH P
OCF OH OCF
760 2500 OH OCF
OCH
OH
450 1500
Redrawn, by permission, from R. H. Whittaker. 1956. Ecological Monographs 26: 1–80
F I G U R E 1 3 .6 Topographic distribution of vegetation types on an idealized west-facing mountain and

valley in the Great Smoky Mountains. Vegetation types: BG, beech gap; CF, cove forest; F, Fraser fir forest;
GB, grassy bald; H, hemlock forest; HB, heath bald; OCF, chestnut oak–chestnut forest; OCH, chestnut
oak–chestnut heath; OH, oak–hickory forest; P, pine forest and pine heath; ROC, red oak–chestnut forest;
S, spruce forest; SF, spruce–fir forest; WOC, white oak–chestnut forest. Source: Whittaker (1956) / John
Wiley & Sons. Reprinted with permission of the Ecological Society of America.
In contrast to Whittaker’s sampling approach along an elevational gradient, Hack and Good-
lett (1960) emphasized local physiographic features in the central Appalachians (Chapter 8). In a
similar region to that studied by Whittaker, they concluded that species assemblages generally
coincided with landform units and often changed abruptly with changes in the form of the slope.
Thus, at least part of how we conceive communities—whether they occur along gradients or with
sharp boundaries—is heavily influenced by the scale and detail of a given study and by the sam-
pling methods employed.
New England Much of the forest landscape in central New England is occupied by forests in
different successional stages on old agricultural fields. Revegetation of old fields following wide-
spread land clearing and farming has occurred on poorly to excessively drained soils of glacial
origin. Composition of these forest communities therefore varies in response to a temporal gradi-
ent that includes the development of old-field succession, and a spatial gradient that includes the
range of site conditions from wet to dry. Absolute elevation above sea level and aspect are less
important in affecting the composition of the forest in this region of moderate elevations and
rolling topography.
In Harvard Forest in central Massachusetts, Spurr (1956b) classified all existing stands
according to their relative position on a successional gradient (early-successional, transitional, or
late-successional) and their relative position on a soil-moisture gradient (from somewhat exces-
sively drained to very poorly drained). Occurrence of individual tree species varied consistently
with the two gradients. Northern red oak and red maple were prominent in all successional stages,
and one or the other was prominent on all sites (see data for transitional stands in Table 13.1). Both
species’ occurrence and abundance were strongly related to soil water; northern red oak was most
frequent on the well-drained and red maple on the very poorly drained sites. Other species were
more specific in their site associations. White oak was most frequent on somewhat excessively
drained sites, paper birch on well-drained sites, and white ash on somewhat poorly drained sites.
In total, the forest communities of Harvard Forest appear to represent a continuous gradient cor-
related with successional stage and soil water.
Competition and Niche Differentiation 305
Table 13.1 Occurrence of species as major components in transitional middle-aged stands in the
Harvard Forest, central Massachusetts.
Site Northern red oak Red maple Paper birch White oak White ash
Frequency (percent)
Somewhat 67 0 0 33 0
excessively drained
Well drained 95 61 14 13 3
Somewhat 81 81 9 6 19
poorly drained
Poorly drained 42 100 8 0 8
Very poorly drained 20 100 0 0 0
COMPETITION AND NICHE DIFFERENTIATION

Ecologists acknowledge that communities are not composed of mutually exclusive sets of species.
Individual species have different genetically based physiological characteristics, such that different
species are effective competitors on many different sites. A given species may be highly competi-
tive and thus dominate in one community on a given site, but may be less competitive (though still
present) in adjacent communities having different environmental conditions. Within a given
climatic region, most forest species probably develop optimally under similar site conditions of
favorable water and nutrients. Thus, two main processes in space and time, competition and mutu-
alism, coupled with adaptation of species to particular site conditions, are what elicit differentiation
of what we recognize as relatively distinct communities.
For example, many forest species in much of central Europe reach their optimal development
under similar site conditions (Figure 13.7), but the superior competitive ability of European beech
and two oak species restricts the other species to only portions of their potential range. Two significant
factors, soil water and nutrients (typically correlated with acidity), are used as the axes in the “eco-
grams” of Figure 13.7. Tree growth is absolutely limited only by waterlogged soils and by extremely
shallow soils over rock, both represented by dotted lines in the diagrams. The physiological opti-
mum range of each species, or the conditions of maximum growth in the absence of competitors, is
hatched more densely (Figure 13.7) than their potential tolerance range, or physiological amplitude.
The optimum ranges of the species tend to coincide, but the most successful species are the best com-
petitors, having a long life span, moderate growth rate, and ability to both tolerate and produce deep
shade. Beech is able to exclude all or almost all other trees from most of its optimal range (Figure 13.7),
birches dominate on wet soils only when these are very acid, and the proportion of oaks becomes
greater the poorer and drier or the wetter the soil. In this humid, submontane climate, deciduous
trees flourish on almost all soils and essentially exclude conifers except toward the limits of existence
of forests where greater disturbance is likely. At such locations, Scots pine is able to dominate either
on very dry sites, whether calcareous or acid, or at the edge of raised, oligotrophic bogs.
Plants having similar ecological requirements and tolerances occur together in recognizable
communities because their genetic makeup and physiological responses allow them to be success-
ful competitors in varying ways. Tree, shrub, and herb species with seemingly similar require-
ments and tolerances are able to exist in a given arid, cold, or wet environment because they occupy
different niches. That is, the plants either occupy different microsites or they utilize a given site
differently. This might be accomplished by differences in their morphological–physiological
responses and life-history traits such as time of germination; amount and timing of shoot and root
growth; depth of rooting; differential allocation of photosynthate to leaves, stems, and roots; and
adaptive mechanisms for obtaining and retaining water. Thus, species become niche-
differentiated when they are co-adapted in their physical and temporal environment.
broad-leaved trees
common oak durmast oak winter lime summer lime
Quercus robur Qu petraea Tilia cordata Tilia platyphyllos
beech hornbeam sycamore maple ash
Fagus sylvatica Carpinus betulus Acer pseudoplatanus Fraxinus excelsior
silver birch pubescent birch black alder sweet cherry
Betula pendula Betula pubescens Alnus glutinosa Prunus padus

dominance range
potential range wetness limit of woodland

needle trees potent. optimum drought limit of w.
soil more and more dry
Scots pine Norway spruce silver fir yew
Pinus sylvestris Picea abies Abies alba Taxus baccata
Soil richer in lime
F I G U R E 1 3 .7 Ecograms showing the potential and optimum range of important tree species of central
Europe in the submontane belt of a temperate suboceanic climate as related to soil-water and nutrient
gradients. Broad-hatch = “physiological amplitude” or potential range of tolerance; narrow-hatch =
“physiological optimum” range or potential optimum; area with thick black border = range where the
species achieves a natural dominance under natural competition; broken border = species is co-dominant
with others, or in the case of Pinus, this co-dominance applies only in the south and east of central Europe.
In each of the ecograms, the y-axis represents the degree of wetness of the site. The x-axis covers the range
from extremely acid to very basic soils. Above the upper dotted line it is too dry for tree growth; below the
lower line it is too wet. The small circle in the center of each ecogram indicates average conditions. Source:
Ellenberg (1988) / Cambridge University Press. Reprinted from Vegetation Ecology of Central Europe, © 1988
by Cambridge University Press. Reprinted with permission of Cambridge University Press.
Interactions Among Organisms 307
INTERACTIONS AMONG ORGANISMS

Plant interactions, be they with animals or other plants, are ubiquitous and countless in nature.
Such interactions are quite diverse and may be grouped by net effect (positive, negative, or neutral)
that each species has on its associates (Table 13.2). The most important of these interactions in shap-
ing the way organisms and evolution are perceived to be the antagonistic (negative) interactions of
competition and predation. By contrast, interactions with no net effect on one or both partners
(commensalisms, amensalisms, or neutralisms) are rarely studied. The remaining type of
interaction—mutualism—benefits both species and is also referred to as symbiosis, cooperation, or
facilitation. Mutualisms were once somewhat neglected in ecological studies relative to their impor-
tance in nature, but an enormous amount of information has now been accumulated (e.g., Bron-
stein 2015; Frederickson 2017; Schupp et al. 2017; Chomicki et al. 2019). Although competition is
an important process, it is unlikely that the natural world is organized primarily around a single
process. Rather, the influences of symbioses, cooperation, and mutualisms are also pervasive and
underlie the evolution and ecology of most species. Ecological and evolutionary considerations of
mutualisms are considered by Bronstein (1994, 2009, 2015) and Boucher (1985).
MUTUALISMS IN FOREST ECOSYSTEMS

Mutualisms occur throughout the entire life cycle of woody plants, and their importance cannot be
underestimated (Janzen 1985). The four main types of mutualisms include: (i) nutritional, includ-
ing the breakdown and supply of nutrients; (ii) protection from extreme site conditions or from
enemies; (iii) dispersal of pollen, spores, or seeds; and (iv) supply of energy by plants to animal
associates through photosynthesis.
Symbiotic Mutualisms—Mycorrhizae The association of roots with fungal hyphae may be

the most important symbiotic mutualism for plants. Such an association is called a mycorrhiza
and may increase the water and nutrient availability to the plant and protect roots from soil path-
ogens. Recent research estimates that there are about 50 000 fungal species that form mycorrhizal
associations with about 340 000 plant species (van der Heijden et al. 2015; Genre et al. 2020). About
82% of angiosperms are mycorrhizal, as are most conifers except for a few pines (Brundrett 2002).
The association between plant roots and mycorrhizal fungi is clearly the rule rather than the
exception. Mycorrhizal relationships are discussed in detail in Chapter 19.
Nonsymbiotic Mutualisms Nonsymbiotic mutualisms between two species are those that
occur in the absence of a physical connection, such as pollination, seed dispersal, protection, and
decomposition of organic matter. Pollination of flowers by animals is critical to successful sexual
reproduction in many woody plants, as is seed dispersal. The mutualisms of pollination and seed
dispersal by animals are discussed in detail in Chapter 12. Protective mutualisms between woody
plants and ants and other insects are common, whereby the plant benefits from protection from
Table 13.2 Grid of species interactions, grouped according to their net effect for each of the two
interacting species.
Effect of Species 2
Positive effect Negative effect Neutral effect
of Species 1
Positive effect Mutualism

Effect
Negative effect Predation/parasitism Competition

Neutral effect Commensalism Amensalism Neutralism
predators, parasites, and diseases, and the insects benefit from shelter and sometimes food (Boucher
et al. 1982). Ants are extremely important protective mutualists in the tropics and to a lesser extent
in temperate forests (Huxley and Cutler 1991). Finally, an indispensable mutualism in all forests is
the decomposition of organic matter on the forest floor and in the soil by many kinds of animals,
fungi, and bacteria, thus ultimately making available nutrients and CO2 to plants (Chapter 19).
COMPETITION
Competition is an antagonistic relationship between two species attempting to utilize the same
resource when that resource is limited in supply. As such, competition is an important agent that
shapes communities. For plants, competing species are capable of occupying the same geographic
space, and competition is considered over the entire life cycle (Grubb 1985). Competition also
occurs when the species interfere with each other’s ability to utilize a resource that is not limited,
such that the yield of one plant is reduced as a result of another plant being present (Harper 1961).
Competition in plants occurs both above and below ground, and its outcome is influenced by the
plants’ inherent requirements or tolerances of light, temperature, water, nutrients, and other site
resources. In the earlier text, we defined these inherent tolerances or requirements as a species’
niche; the principle of competitive exclusion states that one of two species occupying the same
niche will eventually be outcompeted by the other.
The eventual death of the most dominant individuals in a community and the constant
demand of individual trees for crown and root space create changes in the structure and composi-
tion of forest communities. As the main canopy trees grow larger, competition for growing space
increases, with a few individuals gaining space, a few maintaining their space, and an increasing
majority losing space and eventually succumbing. As dominant individuals die due to lightning,
fire, wind, insects, or diseases, a portion of the site is released from the main canopy for occupancy
by a growing and developing understory. These types of competitive relationships occur in all for-
ests. Intraspecific competition (between the trees of the same species) typically does not affect the
composition of the forest community, but interspecific competition (among individuals of different
species) results in a change in species composition and thus has a direct effect on forest succession.
Forest Community Structure and Composition Even-aged stand structure results when
trees regenerate en masse at roughly the same time, as following fire or flooding. Almost all pines,
Douglas-fir, black spruce, cottonwoods, and many other species may form natural even-aged stands.
Individuals in such a stand are nearly the same age and are often of a similar height and diameter. By
contrast, uneven-aged stands are formed by various long-lived hardwoods and conifers and exhibit
great diversity in tree age and size. Uneven-aged stands are typical of naturally regenerated hemlock–
northern hardwood forests (comprised of eastern hemlock, sugar maple, beech, yellow birch, and
basswood, among others) and of some spruce–fir forests of central Europe. Here, the trees of a given
species may range 300 years or more in age. At broad scales in many forests, a mosaic of small even-
aged groups forms the uneven-aged forest; this structure was probably characteristic of upland oak
forests of the East and ponderosa pine forests of the West prior to European colonization.
Competition for light, water, and nutrients in even-aged stands depends largely on stem
density, or the number of stems per unit area. In time, canopy closure develops (i.e., the crowns
of the trees come together to exploit all available growing space) and crowns of individual trees
differentiate into classes (Figure 13.8). A combination of environmental and genetic factors allows
some trees to develop rapidly and exhibit large, well-formed crowns, while others grow more
slowly, and their crowns become more or less restricted. Eventually the slower-growing trees are
gradually overtopped and suppressed. While crowns utilize growing space to compete for available
light, roots utilize belowground growing space to compete for water and nutrients. Root competi-
tion is often severe in even-aged stands, though it is far less obvious than crown competition. The
F I G U R E 13 .8 Differentiation of
trees of a pure even-aged stand into
crown classes. D, dominant; C,
co-dominant; I, intermediate; O, 25 years D I C D D I
overtopped. Source: Smith et al. (1997) O
/ John Wiley & Sons. Reprinted from
The Practice of Silviculture by D. M.
Smith, B. C. Larson, M. J. Kelty, and
P. K. S. Ashton, © 1997 by John Wiley
& Sons, Inc. Reprinted with permis-
sion of John Wiley & Sons, Inc. 20 years D O I C I D O I D O C
15 years D I C C I D I C D I D
10 years D C D D C D C C D C D
5 years D D D D D D D D D D D
root system of a pine in an even-aged stand, for example, may compete with several hundred trees,
whereas its crown competes only with a few adjacent trees. Parsons et al. (1994) found that
significant loss of nitrogen to groundwater did not occur until root gaps resulting from the removal
of clusters of 15–30 lodgepole pines were created in even-aged stands in southeastern Wyoming.
Intense competition in even-aged stands can be illustrated using a simple classification of
crown classes (Figure 13.8). Crown classes are often used by ecologists and forest managers to
judge the vigor of the stand and for determining and planning thinnings and other harvest opera-
tions. The major crown classes are:
Dominant: Trees with crowns extending above the general level of the canopy and receiving full
light from above and partly from the sides. Dominant trees are larger and more vigorous than the
average stems in the stand and have well-developed crowns.
Co-dominant: Trees with crowns forming the general level of the canopy or somewhat below.
Co-dominant trees receive full light from above but only moderate or limited amounts from the
sides. They typically have medium-sized crowns that are more or less crowded on the sides.
Intermediate: Trees shorter than the preceding classes but with crowns extending into the can-
opy formed by the dominant and co-dominant trees. Intermediate trees receive little direct light
from above and virtually none from the sides, and typically have small crowns that are consider-
ably crowded on the sides.
Overtopped: Trees with their crowns entirely below the general canopy level, receiving no direct
light from above or from the sides.
Co-dominants
Intermediates
Overtopped
and
Number of Trees
Overtopped
Dominants
Co-dominants
Intermediates
Dominants
Tree diameter
F I G U R E 1 3 .9 Typical tree diameter distributions for a classic even-aged stand (left) and an uneven-aged
stand (right). Moderate-sized co-dominant trees are most abundant in even-aged stands, but small, over-
topped trees are most abundant in uneven-aged stands.
Crown classes of individual trees can be combined with tree diameters to better understand
the structure of a stand; structure is also illustrative of light competition. Even-aged stands typi-
cally have a tree diameter distribution resembling a bell-shaped curve, with many moderate-sized
stems representing the co-dominant trees, and far fewer small (intermediate and suppressed trees)
and large (dominant trees) stems (Figure 13.9). Uneven-aged stands have a diameter distribution
that follows a negative exponential curve, with many small, often young, overtopped stems and
only a few larger, older stems (co-dominants and dominants). These classical diameter distributions
represent the two extremes of stand structure, and many variations exist between them. For
example, the diameter distribution of even-aged stands may resemble that of an uneven-aged
stand if they contain multiple species that differ in shade tolerance or a have well-established layer
of regeneration beneath the canopy.
The distribution of tree ages may be used to further study competition among individuals
within a species population in an even-aged stand or among different species in an uneven-aged
stand. The distribution of tree ages in a stand may be used as a proxy for time, allowing ecologists
to infer changes that have occurred or may happen in the future. For example, in an old-growth
hardwood forest in New Hampshire, sugar maple and beech exhibit sharply declining populations,
typical of species that periodically establish many seedlings on the shaded forest floor (Fig-
ure 13.10; persistent seedlings, Chapter 4). Many seedlings survive and eventually some replace
the dominant overstory trees. By contrast, red spruce exhibits a lack of young seedlings and may
diminish in abundance over time. Red spruce is not competitive in the younger age classes with
beech and maple on this site in the absence of disturbance.
Vertical Structure Competition for light among forest species also results in the development of
vertical vegetation structure. For example, a multistoried forest may have an upper layer of over-
story trees, one or more subdominant layers of younger stems of the overstory trees and/or mature
trees of subcanopy species, layers of high and low shrubs, and a ground layer of grasses, herbs,
mosses, liverworts, and lichens. The species of each layer are effective competitors where they
occur because they are genetically adapted to make the best use of the space, light, and microcli-
matic resources of their respective vertical positions.
Considerable structure occurs in the tree layer alone depending on species composition and
site conditions. Generally, more vertical layers are found in the presence of more favorable site
conditions, especially soil water. For example, cove forests in the southern Appalachian Mountains
exhibit complex vertical structure. In addition to a diverse overstory, there may be several under-
story tree layers and a dense layer of tall ericaceous shrubs, primarily the rosebay rhododendron.
Tropical rain forests are renowned for even more pronounced vertical structure, including huge
F I G U R E 13 .1 0 Numbers of trees per hectare 1000 000

(log scale) in an old-growth northern hardwood
forest over midpoints of 20-year age classes (log
scale) for sugar maple, beech, and red spruce. Sugar maple
Source: Leak (1975) / John Wiley & Sons. Beech
100 000 Red spruce
Reprinted with permission of the Ecological
Society of America.
10 000
No. of trees, ha
1000
100
10
1
10 100 1000
Age class, years
emergent trees whose crowns extend 5–10 m above the general canopy level and are fully exposed
to sunlight. In such multistoried forests, each vertical stratum has a different microclimate and
often a distinct assemblage of insects and other animals.
The overstory canopy of forests intercepts much of the incoming irradiance, and thus species
of the lower layers must be physiologically and structurally adapted to use the continuously
decreasing amount of light available beneath the canopy and at the forest floor. The canopy is
never completely unbroken over large stretches, however, and various light-requiring subdomi-
nant trees, shrubs, and herbs utilize the well-lighted microsites and growing space to extend their
crowns into openings. The crown densities of overstory trees vary widely, so that considerably
more light may penetrate one forest overstory compared to another. As such, the degree of
development of vertical structure varies across forests of different composition. For example, the
shrub layer is particularly well developed in oak forests of eastern North America compared to
beech–sugar maple forests because oak crowns are less dense, and the canopy is more open than
that of beech–sugar maple forests. The increased light, together with more frequent fires in oak
forests, favors the growth, sprouting, and clonal spread of understory shrubs.
Many pioneer species that colonize areas following major stand-replacing disturbances such
as fire initially exhibit only a single tree layer. Near-pure natural overstory monocultures may be
formed by species such as lodgepole and jack pines, aspens, and the southern pines following
stand-replacing fire. The image of a plantation-like expanse of even-aged and uniform stems is
misleading, however. In the dense stands that arise following a crown fire, for example, trees are
typically patchily distributed over tens or hundreds of hectares (Kashian et al. 2004; Turner
et al. 2004; Kashian et al. 2017). Such stands may be dense and may lack a well-developed under-
story when young and middle-aged, but with increasing age, gaps form and regeneration begins to
develop. A classic example is lodgepole pine forests of the central and northern Rocky Mountains
that may exhibit even-aged stand structure over large areas but show increasing vertical structure
in older stands (Kashian et al. 2005b).
Evidence of considerable variation in stand structure is accumulating and may be associ-
ated with physiography and soil conditions, variability in fire regimes, or insect and pathogen
infestation patterns. Working in central Colorado, Parker and Parker (1994) identified two dis-
tinct kinds of stand structures in 120–140-year-old lodgepole pine: those with closed canopies,
high basal areas and stem densities, and low pine sapling and seedling densities of pine; and
those with open canopies, low tree densities, and significantly higher densities of pine saplings
and seedlings. The structure of closed stands result from dense regeneration developing rapidly
following crown fires, while nearly continuous recruitment of pine occurs in the open stands.
Kashian et al. (2005b) confirmed the presence of both of these stand structures in the lodgepole
pine forests of Yellowstone National Park and found that the structures tended to converge after
about 200 years with age-related development.
Stand Density The abundance and size of trees occupying a site have important implications for
the trees themselves, the site, and for a land manager responsible for controlling forest composi-
tion, growth, and reproduction. Stand density is typically expressed as the number of stems or
basal area per unit area.
Except at very low stand densities, tree mortality in most stands is mostly caused by density-
dependent competition (also called self-thinning) among trees and between tree seedlings and other
vegetation for growing space, specifically light, water, and nutrients. Mortality is highest in the
seedling stage when the number of seedlings per unit area is highest. In a sugar-maple-dominated
forest in Wisconsin, Curtis (1959) reported 99% mortality of sugar maple seedlings over a 2-year
period; this mortality rate is initially independent of age and declines as the seedlings mature
(Hett 1971). Surviving individuals become larger as the outcompeted plants die, and the smaller
plants are continually eliminated from the population. There is therefore a strong relationship bet-
ween plant size and stand density, whereby larger size (or biomass) of the stems is associated with
fewer stems per unit area, and vice versa.
The relationship between plant size and stand density has been extensively explored as a
“fundamental truth” or natural law of plant populations. Reineke (1933) showed that plotting the
logarithm of number of trees per acre against the logarithm of average diameter of fully stocked,
even-aged forest stands typically resulted in a negatively sloping, linear relationship. Plant ecolo-
gists have since designated this as the “−3/2 law of self-thinning” because the slope of the line of
mean volume or biomass of surviving plants versus stand density is approximately −3/2 for popu-
lations of many different kinds of plants (Drew and Flewelling 1977; White 1985; Silvertown and
Lovett-Doust 1993).
This relationship is illustrated in Figure 13.11 for stand density and volume for five stands of
loblolly pine of different initial densities over a 50-year period (Peet and Christensen 1987). Com-
petition increases as stand development and self-thinning occur such that stand density decreases
and mean tree volume increases. The slope of the thinning curve is −1 for a given stand when its
net biomass production equals its loss by mortality, and the slope will be greater (less negative) if
the rate of biomass loss exceeds the rate of net production. Competition is low in a young pine
stand prior to canopy closure, and the slope is flatter than −3/2, but stand mortality rate converges
on a −3/2 slope when the canopy closes and resources become limiting. The rate of convergence to
the −3/2 line is clearly dependent on initial density, with stands at high initial densities thinning
F I G U R E 1 3 .1 1 Relationship between tree

density and mean tree volume for a series Slope = –1.5
1.0
of loblolly pine stands over a 50-year
period. The different initial stand densities
are indicated by the number below each
line (79–1141 stems originally in the
0.1-acre plots). The straight line has a slope
Mean Tree Volume (m3)

of −1.5, as predicted by the law of 0.1
thinning. The tree populations approach an
upper limit of density defined by this
−3/2 line, but no stand closely tracks the
line. Source: Peet and Christensen (1987) /
with permission of Oxford University
Press. Reprinted from BioScience (vol. 37,
0.01
no. 8, page 591), © 1987 by the American 79
Institute of Biological Sciences.
149
233
430
0.001
1141
1000 10 000 100 000
Tree Density (ha ) –1
more rapidly and those at low densities taking a longer time to reach the line. No single stand
clearly defines the −3/2 line, yet all stands together do define such a line, which individual stands
approach but never cross. The line therefore defines the maximum stocking level. The relationship
holds for many stands of many species, of many ages on many sites (Long 1980), although some
exceptions have been reported (Lonsdale 1990; Osawa and Allen 1993). Notably, the intercept of
the line may vary widely depending on species and site conditions (Silvertown and Lovett-
Doust 1993). In forestry, the relationship has been widely applied to stand-density management of
Douglas-fir plantations (Drew and Flewelling 1979).
The relationship between tree diameter and stand density means that careful control of density
is extremely important for forest managers. If large trees are the objective, wide initial spacing
achieved via planting or by thinning is necessary to maximize diameter growth, yet still fully utilize
site resources. Low stand density may negatively affect wood quality for some uses, however, because
trees become branchy, the wood knotty, and/or growth rings may be too wide. In such cases, a balance
must be reached for a given species on a given site through appropriate density control, depending
on the end product desired. As long as the site is fully stocked, such that trees are making full use of
available resources, a species will produce the same amount of wood per year over a relatively wide
range of densities regardless of whether the trees are large or small. More trees can be grown per unit
area and a greater volume of wood can be obtained as site quality improves. Thus, understanding site
quality is of great importance in regulating stand density. Detailed considerations of stand density
from the standpoint of growth and yield and silvicultural management of forests are available in sev-
eral texts (Oliver and Larson 1996; Nyland et al. 2016; Ashton and Kelty 2018; Burkhart et al. 2018).
Competition and Overstory Composition Trees in the main canopy of a forest increase yearly
in height, bole size, length of each growing branch, and number of leaves. A tree must grow to remain
alive and vigorous, and thus it must increase its growing space—its utilization of the site. In doing so,
competition among growing individuals in the overstory ultimately eliminates some trees, particularly
those species genetically and physiologically less suited for survival under existing site conditions. The
result is a gradual (or occasionally an abrupt) change in the composition of the overstory.
Table 13.3 summarizes 60 years of growth of a middle-aged, mixed hardwood stand on an
island in the St. Mary’s River between Michigan and Ontario. Relatively short-lived aspens and
birches practically disappeared from the stand during this period, and sugar maple suffered steady
mortality, with basal area slowly decreasing after 1978. Red maple declined greatly in abundance,
but the growth of the remaining trees compensated for the loss of individuals to create a steady
increase in basal area. Sugar and red maples made up 72% of the basal area by 1993. There were
few northern red oaks on the island, but they grew relatively fast and made up a substantial
proportion of the basal area. The better competitive ability of sugar maple, red maple, and northern
red oak at the expense of aspen and birches brought about major changes in stand composition.
Notably, permanent plot data record how the overstory composition changed over time, but it
does not chronicle the cause of tree mortality in the interval between measurements. It might be
assumed that most of these trees simply were suppressed (leaf area and root feeding area were reduced)
to the point that they died for their inability to maintain a positive carbon balance under competition
for site-specific resources of light, water, and nutrients. However, death may have occurred from many
causes. Even “natural mortality” of a tree often occurs during some period of extreme stress, such as
in a hot and dry spell in late summer, in a severe unseasonable frost, or in an extremely cold winter. In
other cases, the tree becomes weakened to the point that it becomes susceptible to insects or disease.
Usually, several factors are interrelated and together account for death (Mueller-Dombois 1987).
Competition in the Understory Overstory competition involves relatively few individuals

and is spread out over many years, but competition near the forest floor involves many shrubs,
Table 13.3 Changing composition over 60 years in a northern hardwood forest on Sugar Island, St.
Mary’s River, Upper Michigan.
Species 1933 1944 1955 1971 1978 1989 1993

Stand density (trees/ha)
Sugar maple 803 741 652 502 474 403 381
Red maple 314 210 173 158 151 133 133
Paper birch 89 40 22 7 7 7 5
Bigtooth aspen 49 7 7 7 7 5 5
Northern red oak 44 44 44 37 37 35 35
Yellow birch 35 17 15 10 10 5 2
All trees a
1364 1077 927 734 697 598 571
Basal area (m /ha)
2
Sugar maple 9.4 11.6 12.7 13.2 13.3 12.4 11.8

Red maple 7.1 7.5 7.9 9.0 9.7 10.1 10.7
Paper birch 2.8 1.6 1.3 0.5 0.5 0.6 0.4
Bigtooth aspen 1.4 0.4 0.5 0.8 0.9 0.8 0.9
Northern red oak 2.0 2.8 3.6 5.0 5.4 6.2 6.7
Yellow birch 1.0 0.7 0.7 0.6 0.6 0.2 0.1
All treesa 24.0 25.0 27.3 29.7 31.0 31.0 31.2
a
Includes some individuals of species not listed in the earlier text.
Understory Tolerance 315
herbaceous plants, and grasses as well as trees, often in very large numbers and changing on an
annual basis. Many plants are adapted to life in the understory, such as herbs that complete a major
portion of their annual growth before the canopy trees fully flush; herbs, many shrubs, and small
trees that are generally tolerant of understory site conditions; and trees that can both thrive in the
understory and also occupy the overstory itself when the opportunity occurs.
Wendel (1987) estimated the sheer numbers of potentially competing understory stems and
plants in north-central West Virginia oak forests. Forest floor samples from each of four different
sites were collected and the plants allowed to develop in a greenhouse during the following growing
season. Nine species of tree seedlings and sprouts averaged 981 000 stems/ha, 8 species of woody
shrubs averaged 243 000 stems/ha, and 25 species of herbs averaged 220 000 stems/ha. Most of the
newly germinating seedlings in this study, grown free from competition, would not have survived
in the forest understory, but the potential for competition in the forest understory is nevertheless
enormous in these ecosystems.
Similar to overstory trees, several factors are usually at play in the death of understory plants.
Unsuccessful competitors lose vigor and die in part due to starvation resulting from inadequate
photosynthesis due to inadequate irradiation, resulting in insufficient carbon fixation for the plant
to survive. Lack of sufficient photosynthate leads to a “death spiral,” as the root system is weak-
ened and unable to obtain adequate water to support a vigorous cambium and crown. Inadequate
soil water is another factor that contributes to the death of outcompeted understory plants, as are
insects and diseases, particularly when the plant has reached a weakened condition (Decker 1959).
The site factors relating to understory survival are discussed in the following text.
UNDERSTORY TOLERANCE
An important factor shaping the composition of forest communities is the physiological capacity
of plants to persist, or not, in the low light conditions of the forest understory. Those that can per-
sist in these conditions have the potential to reach overstory status, whereas those that cannot are
destined to die. These basic attributes therefore shape forest composition over time. Plants that
become shaded in open habitats may avoid shade by growing taller or otherwise toward the light
(called shade avoidance), but those found in the forest understory do not have such a luxury.
Survival of understory plants is central to an understanding of forest succession, because forest
trees able to both survive as understory plants and then respond to release to reach overstory size
will eventually represent a major component of the forest community. A forest tree that can sur-
vive and thrive under a forest canopy is said to be understory-tolerant (or simply tolerant),
whereas one that can thrive only in the main canopy or in the open is classified as understory-
intolerant (or intolerant). There is a continuum of tolerance from one extreme to the other.
The general biological meaning of “tolerance” refers to the capacity of a plant to be genetically
adapted and to have the physiological ability to be a good competitor under conditions generally
unfavorable to other plants. A salt-tolerant plant is therefore one that is adapted to grow in soil
with a high salt concentration, and a drought-tolerant plant is one that can grow at low soil-water
levels that would be fatal to most other plants. Many pioneer species are intolerant of low light but
tolerant of high irradiance, extreme temperature, drought, and low nutrient status often found on
exposed, open sites. However, in this chapter, and elsewhere in this book, our unmodified use of
the term “tolerance” refers to a plant’s vigor in the forest understory due to its genetic adaptations
and physiological attributes to this complex environment. Although this is often described in the
literature as “shade tolerance” because light is typically a limiting factor in the forest understory,
“understory tolerance” is the more acceptable term because of the prevalence of many other abi-
otic and biotic factors that contribute to understory conditions, such as soil-water stress, tempera-
ture, browsing, and disease. The term “shade tolerance” should be applied specifically to situations
where light is clearly the dominant factor.
Notably, the concept of tolerance is somewhat misleading and has been described as
inadequate (Grubb 1985). Many plants do not simply tolerate the conditions with which they are
associated in the field; they may actually require them for maximum growth. For example, some
plants associated with shade require it because they perform poorly in full sunlight. Likewise,
flood-tolerant trees such as red ash, eastern cottonwood, sweetgum (Broadfoot and Willis-
ton 1973), and water tupelo (McKevlin et al. 1995) may actually require flooding to outcompete
other plants in establishment and for optimum growth.
CHARACTERISTICS OF UNDERSTORY-T OLERANT AND -I NTOLERANT SPECIES

A tolerant tree is one that grows and thrives under a forest canopy, and its vigor under such condi-
tions may be characterized in various ways. An intolerant tree is usually characterized by opposite
extremes of the same traits. Many characteristics have been cited as criteria or indicators for the
determination of tolerance:
Tolerance of Suppression in the Understory: Tolerant trees may live for many years under
dense forest canopies. Many species are able to germinate and survive a few years under such
conditions, but only tolerant trees will persist and continue to grow for decades or centuries.
Response to Release: The vigor of tolerant trees is evident when they are able to initiate “release,”
or immediate and substantial growth following removal of the overstory. A tolerant tree may sur-
vive for years or even decades in the understory with very little growth, and yet grow extremely
rapidly soon after release.
Crown and Bole Development: The lower branches on tolerant trees maintain foliage longer
and to a greater extent than those on intolerant trees, such that tolerant trees will have deeper and
denser crowns. Tolerant trees will therefore naturally prune at a slower rate, will maintain a
greater number of leaf layers in the crown, and will maintain healthy and vigorous leaves deeper
into the crown. The greater depth of crown of tolerant trees results in more tapered boles com-
pared to those of intolerant trees (Larson 1963b).
The shoot growth and branching pattern of a species are adapted to provide the amount of light
required of its foliage. The relatively fast branch growth and spreading form of American elm, syc-
amore, and silver maple, for example, create a well-lighted crown necessary to maintain the leaves
of these species. By contrast, sugar maple and American beech leaves tolerate considerable shade
and thrive in the interior of a crown that is more compact and has slower branch growth than that
of intolerant species.
Stand Structure: Tolerant trees persist over long periods of time in natural mixed stands and tend
to be successful when mixed with other species of equal size. Consequently, tolerant trees may
form denser stands with more stems per unit area compared to intolerant trees.
Growth and Reproductive Characteristics: Tolerant trees inherently grow more slowly than intol-
erant trees, especially in moderate to high light conditions (in dense shade, a slow growing, tolerant
species may appear to grow faster than a dying intolerant species). Tolerant deciduous angiosperms
(beeches, maples) tend to have seasonally determinate shoots, whereas intolerants (willows, aspens,
cottonwoods, birches) have seasonally indeterminate shoots (Marks 1975). Tolerant trees typically
mature later, flower later and more irregularly, and live longer than intolerant trees.
Populations of the respective extreme types of tolerance have evolved under differing
selection pressures of environment and plant competition such that they belong to complex
and contrasting adaptation systems. Intolerant species are typically early-successional species
that may colonize a wide variety of sites, successful because of two major types of adaptations.
First, they are able to rapidly establish on disturbed sites, grow quickly in the open, produce
seed early in their lives, and disseminate seeds widely. These characteristics have enabled intol-
erant species to perpetuate themselves wherever fire, windstorm, flooding, cultivation, or other
disturbances have eliminated or reduced the existing vegetation, and they may depend on such
disturbances for persistence in many climatic regions. Intolerant species also exhibit important
adaptations to extreme site conditions, such as dry or infertile sites as well as to the temperature
extremes of initially disturbed sites. They may form relatively permanent communities on
extreme sites where they may outcompete more tolerant species. For example, willows in annu-
ally flooded bottomlands, jack pine on nutrient-poor sands of the Lake States, table-mountain
pine on the driest and least fertile soils of the Appalachian Mountains, and sand live oak on
sands of the southern Coastal Plain are all examples of intolerant species occupying
extreme sites.
Tolerant species establish, grow, and have become adapted to conditions markedly different
from those of intolerants, typically more mesic (moist, sheltered, and fertile sites). They replace intol-
erant species during succession and perpetuate themselves through adaptations favoring survival and
growth in a shaded understory, establishment in relatively undisturbed litter layers, and long life spans.
Their presence tends to reinforce the shaded, moist, humid, and relatively disturbance-free environ-
ment favorable to their own regeneration once they attain dominance in the canopy (Chapter 4). Nota-
bly, although it is instructive to compare the two extremes of tolerant and intolerant species, species
typically have varying degrees of intermediacy from very intolerant to extremely tolerant.
TOLERANCE RATINGS OF TREE SPECIES

Tolerance varies with genetically different individuals and races of a given species, and with
different regional climates, local site conditions, plant associates, vegetative conditions (seed-
lings versus sprouts), and especially with age. A given species may appear to be more tolerant
in one part of its geographic range than another, one site in comparison to another, and/or in
one forest type compared to another. For example, eastern white pine is thought to be mid-
tolerant in Minnesota but intolerant in New England. In central New England, it is more toler-
ant on dry, sandy soils than on moister sandy loams. Like most species, eastern white pine is
more tolerant as a juvenile compared to the adult phase. For those species that vigorously
sprout, such as oaks, hickories, ashes, and tulip tree, the sprouts are more tolerant than seed-
lings of the same species in part because they are better able to absorb water and nutrients from
the well-established root system of the parent. Thus, understory tolerance is not only a relative
matter, as a species’ tolerance rating will depend upon the regional ecosystem, specific site, age,
and its associates. Tolerance ratings must be interpreted with great care and with these
points in mind.
At the same time, the approximate tolerance rating of more common forest tree species
should be known and understood by ecologists and forest managers as a general frame of refer-
ence. We present such a frame of reference as an estimate of the relative understory tolerance of
tree species assuming representative site conditions in Table 13.4. Notably, such estimates can and
do vary among observers and researchers. Niinemets and Valladares (2006) attempted to assign a
numerical value for shade tolerance on 806 woody species across the Northern Hemisphere using
the 5-class ranking scale of Baker (1949). While this protocol appears useful for many species, in
some cases the attempt to generalize across all regional ecosystems and sites provided relative
values contrary to those observed by some ecologists in the field. For example, in the forests of the
Lake States, black oak is an intolerant species that grows on dry and nutrient-poor sites, forming
savannahs prior to European colonization. Northern red oak is a mid-tolerant tree found on dry-
mesic to mesic sites, often found as a minor component of mesic and nutrient-rich sites dominated
Table 13.4 Relative understory tolerance of selected North American forest treesa.
Eastern North America

Gymnosperms Angiosperms Western North America
Very tolerant
Eastern hemlock Flowering dogwood Western hemlock
Balsam fir American beech Pacific silver fir
Sugar maple Pacific yew
Tolerant
White spruce Basswood Spruces
Black spruce Hop-hornbeam Western redcedar
Red spruce White fir
Grand fir
Subalpine fir
Redwood
Mid-tolerantb
Eastern white pine Yellow birch Western white pine
Slash pine Silver maple Sugar pine
Red maple Douglas-fir
White oak Noble fir
Northern red oak
Hickories
White ash
Elms
Intolerantb
Red pine Black cherry Ponderosa pine
Shortleaf pine Tulip tree Junipers
Loblolly pine Sweetgum Red alder
Eastern redcedar Sycamore Madrone
Black walnut
Black oak
Red ash
Scarlet oak
Sassafras
Very intolerant
Jack pine Paper birch Lodgepole pine
Longleaf pine Aspens Whitebark pine
Virginia pine Black locust Digger pine
Tamarack Eastern cottonwood Western larch
Pin cherry Cottonwoods
Survival in the understory is related to light irradiance, moisture stress, and other factors. As a general guide to the light irradi-
ance component, we estimate the range of minimum percentage of full sunlight for a species to survive in the understory at each
of the five arbitrary levels of tolerance.
Very tolerant species may occur when light irradiance is as low as 1–3% of full sunlight. Tolerant species typically require 3–10%
of full sunlight. Intermediate species 10–25%. Intolerant species, 25–50%. Very intolerant species, at least 50%. For example,
an intolerant species competing in the understory is unlikely to survive with less than about 25% of full sunlight (unless other
compensating factors are favorable).
a
Based on representative site conditions for the respective species.
b
Many mid-tolerant and intolerant species are observed to be more tolerant in the juvenile phase than in the adult phase. Nota-
ble examples include black cherry, white ash, and elms.
by sugar maple and beech. White oak is also a mid-tolerant tree, but is less tolerant than northern
red oak and is found on dry to dry-mesic sites, rarely with mesic species (Barnes and Wagner 2004).
By contrast, Niinemets and Valladares (2006) characterize black oak and northern red oak as
equally shade-tolerant, but black oak as the more drought-tolerant, and white oak as the most
shade-and drought-tolerant of the three species. We therefore emphasize that all forest ecologists
must learn to understand the relative tolerance and competitive ability of trees using their own
observations and field experiences and not from a textbook table. Nevertheless, a “standard
condition” or general starting point is appropriate, given an understanding of the nature of under-
story tolerance.
EXAMPLES OF UNDERSTORY TOLERANCE IN FOREST ECOSYSTEMS

We provide four examples of understory tolerance using photographs from contrasting forest eco-
systems. In a young, second-growth stand of intolerant tulip tree in a cove site of the southern
Appalachian Mountains, sufficient light and water reach the understory to favor development of a
conspicuous and diverse community of mid-tolerant and tolerant species, such as oaks, maples,
and beech, as well as a rich shrub and herbaceous flora (Figure 13.12). A stand of mid-tolerant
oaks in the Missouri Ozarks (Figure 13.13) illustrates a dominant oak overstory and an open,
partially shaded understory of oaks and other vegetation. In contrast to the understory of the tulip
tree forest, growth of the oak understory is curtailed by the shade cast by the overstory and low soil
water during the growing season. In a stand of sugar maples in northern Michigan, dense shade
from the overstory favors the very-tolerant sugar maple seedlings over all other species in the
understory (Figure 13.14). In deep shade, sugar maple seedlings survive but grow slowly; in small
openings where more light is available (background of Figure 13.14), they respond with acceler-
ated growth and, barring disturbance, will perpetuate the dominance of sugar maple. Finally, the
western white pines of a stand in northern Idaho originally colonized the site following a fire,
forming a dense stand two centuries later with a canopy that intercepts most of the incoming light
(Figure 13.15). The pines will gradually be replaced by the very tolerant species in the understory,
such as western hemlock, grand fir, and western redcedar, in the absence of fires, windstorms,
or logging.
NATURE OF UNDERSTORY TOLERANCE

The relative tolerance of a given species growing in a given ecosystem can be estimated reasonably
accurately, but understanding the nature of its tolerance is much more difficult. The nature of
tolerance may be examined at the species level in terms of adaptations, but the specific trait of
survival in the understory is usually investigated by studying the site factors and the physiological
processes involved.
Two major hypotheses have been proposed to investigate the traits determining a species’
understory tolerance as it relates specifically to light, and understanding these perspectives is use-
ful in framing further discussion of the nature of understory tolerance. The carbon gain hypothesis
emphasizes that plants in low light maximize the capture of light while minimizing respiration
costs needed for tissue maintenance (Givnish 1988). Thus, shade tolerance may be increased by
any trait that helps a plant more efficiently use light and more efficiently increases photosynthesis
and net carbon gain. Recent evidence has suggested that the better performance of shade-tolerant
species in low light results more from minimizing carbon losses during dark respiration than by
increasing carbon gain (Chapter 18; Walters and Reich 2000b; Craine and Reich 2005). A relation-
ship arising from the carbon gain hypothesis is that plants with high growth rates in high light
have low survivorship in low light, and vice versa (Kobe et al. 1995). A second hypothesis related
FIGURE 13.12 Young second-

growth stand of tulip tree in a cove
of the Appalachian Mountains,
North Carolina. Source: U.S. Forest
Service photo.
F I G U R E 13.13 Uncut, old-growth

oak stand in the Ozark Mountains
of southeastern Missouri. Source:
U.S. Forest Service photo.
F I G U R E 1 3 .1 4 Sugar
maple stand in the
Upper Peninsula of
Michigan. Unbrowsed
sugar maple seedling
shown on left and
browsed seedling on
right. Source: U.S. Forest
Service photo.
F I G U R E 1 3 .1 5 The old-growth western

white pines will eventually be replaced,
barring disturbance, by very tolerant conifers
that have become established in the
understory; northern Idaho. Source: U.S.
Forest Service photo.
to shade tolerance is the stress tolerance hypothesis, which suggests that plant survival in low light
is explained best by its capability to resist biotic and abiotic stresses in the understory (Kitaji-
ma 1994). One process for such stress tolerance may be a higher allocation to storage, such that a
plant is able to tolerate periods of low light by utilizing stored photosynthate produced during
periods of high light (Kobe 1997; Canham et al. 1999; Myers and Kitajima 2007). Many shade-
tolerant species exhibit determinate growth and terminate annual growth earlier in the growing
season compared to the indeterminate growth of intolerant species, allowing shade-tolerant
species more time to allocate photosynthate toward storage rather than growth (Kikuzawa 2003).
A review of the evidence supporting and refuting each of these hypotheses is presented in Val-
ladares and Niinemets (2008).
Environmental Factors Relating to Understory Tolerance The most obvious ecologi-

cal feature of the understory is low light irradiance, and as suggested in the earlier text, many ecol-
ogists associate a plant’s ability to survive in the understory solely based on its tolerance to shade.
Notably, abundant observational and experimental evidence has shown that light irradiance is
sufficient under most forest canopies to permit most forest trees to photosynthesize at rates high
enough to compensate for respiration losses. Moderate canopy cover, as in pine and oak forests,
allows ample light to reach the forest floor, and even dense canopy cover like that formed by
spruce–fir and sugar maple–beech forests permits sunflecks and light at the edges of small gaps
that support plant survival and growth in the understory. In general, light is the single limiting
factor in understory survival only where light on the forest floor is less than about 2% full sunlight
(Chapter 6).
Light is not the only environmental factor greatly modified by the forest canopy. When a
forest canopy is dense, almost all climatic and soil factors differ from those on similar but open
sites. Soil water in particular is among the environmental factors affected. The importance of soil
water in regulating understory establishment and growth was demonstrated in a classic study in
1904 by Fricke, who cut the roots of competing understory trees by trenching around small, poorly
developed seedlings in a Scots pine stand in Germany. The seedlings responded vigorously, indicat-
ing that their growth had been inhibited by soil-water limitation created by the competing roots of
the overstory trees rather than by low light. These experiments were repeated with similar results
under eastern white pine in New Hampshire (Toumey and Kienholz 1931), loblolly pine in North
Carolina (Korstian and Coile 1938), and others.
Understory tolerance should not be attributed solely to soil water just as it should not be
attributed to light alone. Both soil water and light as well as nutrients (Chapter 19) contribute to
understory survival. Often soil water and light availability occur on opposite gradients, with dry
sites correlated to open areas and shaded areas more mesic, leading many ecologists to the idea
that shade and drought tolerance are inversely related to one another (Niinemets and Val-
ladares 2006; Figure 13.16). The logic behind such an idea is that tolerance to one environmental
factor incurs a physiological cost to the plant, making it difficult or unable to also tolerate addi-
tional stressors. Trade-offs between shade tolerance and flood tolerance have also been document-
ed in floodplains of the Southeast (Battaglia and Shari 2006). Given that other factors of the under-
story environment, such as temperature and humidity, are also likely to influence understory
tolerance as they become limiting, it is clear that understory tolerance includes much more than
the effects of light availability.
The intricate association between the crown and roots of understory plants illustrates how
shade and drought tolerance are related. Loblolly pine in North Carolina and elsewhere, for
example, has relatively intolerant seedlings that seem to photosynthesize more than enough to
counterbalance respiration losses. However, seedlings beneath loblolly pine canopies do not pho-
tosynthesize enough to permit their root systems to expand into the deeper soil. Without a root
system of sufficient extent and depth, the seedlings struggle to compete with the roots of overstory
F I G U R E 1 3 .1 6 Inverse relationship 6
0
of shade tolerance to drought tolerance A North Europe East
for 806 woody species from three America Asia
continents. Insets provide the slopes of –0.4
Slope
the regression lines for each continent,
with the steepest slope found for East 4 –0.8
Shade tolerance
Asian species. Source: Niinemets and b
Europe a
Valladares (2006) / John Wiley & Sons. –1.2
North America
c
2
East Asia
0
0 2 4 6
Drought tolerance
trees and understory plants. As a result, they often develop somewhat weakened tops over time
and may die during a period of hot, dry, midsummer conditions when soil-water stress becomes
unusually severe. By contrast, mid-tolerant oaks and other hardwoods under similar conditions
develop root systems extensive and deep enough for them to survive droughts.
While loblolly pine seedlings may require a rate of photosynthesis high enough to develop a
sufficient root system in the moderate light conditions of the understory, either light or root competi-
tion may be relatively more important under other conditions. For example, light is at extremely low
levels under dense Sitka spruce and western hemlock in temperate rain forests of the Pacific Northwest,
but the site is almost always wet or at least moist such that light is obviously the more important factor.
Under open oak woodland types or under dry ponderosa pine forests, light availability in the under-
story is usually well above any critical level but soil water is always in short supply, such that soil water
is the more important factor. Always, it is the ecosystem-specific interaction of light, water, tempera-
ture, and other environmental factors as well that together determine understory survival and growth.
Physiological Processes Relating to Tolerance Basic genetic differences among species

dictate that some understory plants can survive and grow under the site conditions of the under-
story while others cannot. Such genetic differences are expressed in species’ physiological responses
to these conditions. Based on the carbon gain hypothesis, any trait that improves a species’ capa-
bility of capturing carbon will improve its understory (specifically, shade) tolerance (Givnish 1988),
but the importance of many such traits in improving survival under low light is disputed. Val-
ladares and Niinemets (2008) summarized 40 such traits related to the physiology, biochemistry,
anatomy, and morphology of leaves; crown architecture; and whole plant characteristics. Notably,
the expected values for shade-tolerant plants of 26 of the 40 traits have been challenged by at least
one study. Logically, we would expect that in having adapted photosynthetic mechanisms to low
light intensity, a shade-tolerant plant would exhibit traits to efficiently harvest and utilize limited
available light (crown geometry and photosynthetic rate) and maintain low rates of respiration.
For the purposes of this discussion, examples of characteristics of surviving plants compared to
failing plants in the understory include: (i) greater photosynthesizing leaf area, (ii) more efficient
photosynthesis per unit leaf area, (iii) lower rates of dark respiration and thus lower compensation
points and light saturation points, and (iv) lower relative growth rates.
As discussed in the earlier text, tolerance to one environmental factor often leads to a trade-off in
tolerance to another. A similar physiological trade-off is present in intolerant species, which have evolved
highly adapted photosynthetic and respiration mechanisms to maximize their production in full sunlight
at the cost of lowered efficiency under shaded conditions. The differences in rates of dark respiration
between tolerant and intolerant species are probably the most important and least debated (Valladares
and Niinemets 2008) determinants of success or failure in forest shade (Loach 1967; Walters and
Reich 2000b; Craine and Reich 2005; Baltzer and Thomas 2007). Given that daily net carbon gain is the
difference between photosynthesis and respiration, plants in a shaded understory may spend many more
hours below than above even their lower light compensation point. Shade-tolerant plants are thus able
to perform well in low light by minimizing their loss of carbon dioxide. By contrast, the high photosyn-
thetic rates of intolerant species are offset to some extent by high rates of respiration (Chapter 18), and
thus are less adapted to shaded conditions where photosynthetic rates are inevitably lower.
Photosynthetic rates have long been posited as an important mechanism for successful plant
performance and survival in the shade (Logan and Krotkov 1969; Logan 1970). Tolerant species typ-
ically exhibit lower photosynthetic rates, presumably because of reduced photosynthetic constitu-
ents in the leaves such as RuBisCO, ATP synthase, and electron carrier per unit of leaf surface, which
translates into lower carbon fixation (Givnish et al. 2004; Pallardy 2008). As such, shade-tolerant
plants have a greater carbon gain and correspondingly better survival in low light, but tolerant plants
tend to grow slowly relative to intolerant plants regardless of whether they are in the shade or in full
sunlight. Notably, photosynthetic capacity (the maximum rate at which leaves are able to fix carbon
during photosynthesis) in intolerant species is higher than that of tolerant species in both high and
low light (Kitajima 1994; Walters and Reich 1996; Reich et al. 2003), but higher rates of respiration in
intolerant plants reduce their net carbon gain. Thus, selection for a high rate of photosynthesis and
high growth rate at high light irradiances may inevitably limit the plant in the shade.
Northern red oak was examined in a study of growth and CO2 exchange of several northern
hardwood species in controlled environments (Walters et al. 1993) that bear out the trade-offs bet-
ween photosynthesis, growth rate, and respiration. The intolerant paper birch had a photosyn-
thetic rate over 3 times that of the tolerant sugar maple, but also had respirations rates 1.5–2 times
those of sugar maple (Table 13.5). Paper birch also had high relative growth rates compared to the
tolerant sugar maple. Northern red oak is fire-dependent and mid-tolerant and may grow in both
open and moderately shaded environments. Northern red oak was intermediate in relative growth
rate, remarkably low in respiration, and had the highest ratio of root weight to total plant weight.
Thus, the higher understory tolerance of red oak relative to paper birch incurs a cost of a lower
relative growth rate. The relationship of photosynthesis and total plant respiration in determining
the carbon balance of individual plants and whole ecosystems is considered in Chapter 18.
Table 13.5 Comparison of photosynthesis and respiration for tolerant and intolerant species.
Photosynthesis
Respiration (nmol CO2 g−1 s−1)
(nmol CO2 g−1 s−1)
Species Leaf Leaf Stem Root
Paper birch High light 225 22.9 23.2 37.5
Low light 102 25.3 22.3 42.1
Northern red oak High light 94 11.1 10.4 12.7
Sugar maple High light 69 14.8 10.0 20.5
Photosynthesis is net leaf CO2 assimilation expressed on a leaf mass (nmol CO2 g−1 s−1) base of seedlings grown in high and
low light conditions in growth chambers. Respiration is in mass-based values (nmol CO2 g−1 s−1) as overall seedling means.
Source: Walters et al. (1993) / Springer Nature. Reprinted from Oecologia. © Springer-Verlag Berlin Heidelberg 1993. Reprinted
with permission of Springer-Verlag.
SUGGESTED
R E A D I N G S
Austin, M.P. and Smith, T.M. (1989). A new model for Kitajima, K. (1994). Relative importance of photosynthetic
the continuum concept. Vegetatio 83: 35–47. traits and allocation patterns as correlates of seedling
Bronstein, J.L. (2009). The evolution of facilitation and shade tolerance of 13 tropical trees. Oecologia 98: 419–428.
mutualism. J. Ecol. 97: 1160–1170. McIntosh, R.P. (1993). The continuum continued: John
Gleason, H.A. (1926). The individualistic concept of the T. Curtis’ influence on ecology. In: Fifty Years of
plant association. Bull. Torrey. Bot. Club 53: 7–26. Wisconsin Plant Ecology (ed. J.S. Fralish, R.P. McIn-
tosh and O.L. Loucks). Madison: University of Wis-
Grubb, P.J. (1985). Plant populations and vegetation in
consin Press.
relation to habitat, disturbance and competition:
problems of generalization. In: The Population Struc- Niinemets, Ü. and Valladares, F. (2006). Tolerance to shade,
ture of Vegetation, Handbook of Vegetation Science, drought, and waterlogging of temperate northern hemi-
Part III (ed. J. White). Dordrecht: W. Junk. sphere trees and shrubs. Ecol. Monogr. 76: 521–547.
Diversity in Forests CHAPTER 14
T he diversity of organisms, how to measure it, and hypotheses about its causes have long
interested researchers studying all types of ecological systems. Exploring natural forests,
inventorying species, searching for rare species, and posing questions such as why a species is pre-
sent, rare, or abundant at some sites but not at others are fundamental to forest ecologists. “Diver-
sity” simply means “variety” in ecology, but the concept quickly becomes complex with attempts
to make it quantitative or comparing it among areas. At its simplest level, studying diversity is a
descriptive pursuit in which species are listed or counted.
Diversity may be studied at many different levels of organization, from genes to ecosystems.
In this chapter, our treatment of diversity is primarily at organism and ecosystem levels. Diversity
is closely intertwined with many interacting topics covered individually in this book, including
paleogeology and ecology, ecosystem geography, plant physiology, and human activities. Moreover,
we again emphasize the physical and ecological processes of forests whose ultimate result includes
the diversity of ecosystems and organisms that inhabit them. Our overview of diversity is necessar-
ily limited, and readers are encouraged to delve deeply into books that treat organismal diversity
(Wilson 2010; Ricklefs and Schluter 1993; Huston 1994; Rosenzweig 1995; Gaston and Spicer 2014),
measurement of species diversity (Krebs 1998; Stohlgren 2007; Magurran and McGill 2011; Ma-
gurran 2013), and conservation biology (Meffe and Carroll 1997; Groom et al. 2006; Primack 2010;
Van Dyke and Lamb 2020; Hunter and Gibbs 2021).
CONCEPTS OF BIOLOGICAL AND ECOSYSTEM DIVERSITY

The natural world contains many different types of diversity. For example, the diversity of physical
features of bedrock, landform, geomorphic processes, parent material, and climate are strong
drivers of the growth and distribution of plants. The diversity of biota—plants and animals, their
gene pools, and their aggregation in populations and communities—is of great interest and con-
cern. We define biological diversity (biodiversity) as the types and numbers of organisms and
their patterns of distribution. In addition to the organisms themselves, this definition also includes
the genetic diversity of gene pools and other types of diversity related to groupings of biota, such
as the diversity of families and genera, populations and communities, and “structural” properties
of communities (vertical layering, plant density, patchiness, etc.).
Biota are not ecosystems, and thus biodiversity should be distinguished from ecosystem
diversity, although landscapes and ecosystems are often included as part of “biodiversity.” We
define ecosystem diversity as the kind and number of ecosystems in an area, including the pat-
terns of association of certain ecosystems with one another and the recurrence of these patterns on
a given landscape. Ecosystem diversity can be estimated at multiple spatial scales using the same
diversity indices used to measure species diversity (Romme 1982). Maintaining and conserving
biodiversity depend on our understanding of ecosystem diversity because species belong to
ecosystems.
327
328 Chapter 14 Diversity in Forests
THE VALUE OF SPECIES DIVERSITY

Scientists today have relatively little knowledge of the Earth’s biodiversity (Solbrig 1991), and esti-
mates of the Earth’s biota seem to be continually updated. For example, Wilson (2010) estimated
that 10–100 million species existed, only 1.4 million of which had been described and identified.
He estimated that 73% of the Earth’s biota were animals, mostly insects (53% of the total), and 18%
plants, mostly angiosperms. The fact that the combined diversity of insects and flowering plants is
so large reflects the intricate symbioses that unite these groups (Wilson 2010), and in fact a large
fraction of plants depend on insects for pollination and dispersal (Chapter 12). Wilson’s estimates
have been refined in recent decades; one well-cited study estimates that 8.7 million eukaryotic
species exist on Earth, with animals representing an even higher proportion (91%) of the 1.2 mil-
lion species identified, though still mostly insects (Mora et al. 2011). This estimate remarkably
suggests that 86% of the Earth’s species have yet to be discovered and described, which would
require almost 500 years to complete at the current pace of scientific progress (May 2011). An even
more recent estimate suggests that the world’s biodiversity is dominated by bacteria (78%), with
animals (mostly insects) representing only 7% and plants less than 1% (Figure 14.1; Larsen
et al. 2017). These estimates were in part made based on estimates of the number of parasitic
species, many of which are bacteria, likely to be found to be specific to each insect species. As the
technology for genome sequencing refines the boundaries between species, new estimates of bio-
diversity will surely come to light.
Where does biodiversity exist in forests? Admittedly in this book we draw the most attention
to plants, and in particular to woody plants. Biodiversity in forests is found mainly in understory
plants (rather than the dominant trees) and animals that use forests as habitat. However, most of
the diversity of forests is found in soil organisms. Soil biodiversity per unit area is often far greater
than aboveground biodiversity (Nielsen et al. 2015). Soil microbial organisms (most fungi and
bacteria) probably represent a pool of diversity of tens of thousands of different species per gram
of soil (Roesch et al. 2007), depending on soil type, climate, and the researcher’s definition of what
makes a microbial species. Soil invertebrates such as insects, springtails, arachnids, nematodes,
earthworms, and protists also add to the belowground biodiversity pool. Despite our focus mainly
on plants and other aboveground diversity in this chapter, we acknowledge the enormous diversity
of life below ground.
Despite estimates of a seemingly extraordinary number of organisms, the vast majority of
plant and animal species that have ever lived on the Earth are extinct (Raup 1986). As many as four
billion species of plants and animals alone are thought to have lived at some point in the geologic
0.4% 1%
2.8% 0.1%
4.1% 7.3%
5.6%
7.4%
17.6% 90.5% 0.02%
Animals
7.3%
Fungi
4.9% 78% Plants
73.1% Protists
Bacteria
Wilson (1992) Mora et al. (2011) Larsen et al. (2017)
F I G U R E 1 4 . 1 Three estimates of the relative proportion of different groups of organisms. Earlier estimates
suggest that animals (mostly insects) dominate Earth’s biodiversity, but recent estimates suggest that bacteria
are the dominant group. Source: From Larsen et al. (2017) / with permission of University of Chicago Press.
The Value of Species Diversity 329
800
CAMBRIAN
ORDOVICIAN
SILURIAN
DEVONIAN
CARBONIFEROUS
PERMIAN
TRIASSIC
JURASSIC
CRETACEOUS
TERTIARY
600
Number of Families
400
200
0
600 500 400 300 200 100 0
Millions of Years Ago
F I G U R E 1 4 . 2 Example of the increase of biological diversity over geological time using data from families
of marine organisms. A slow increase is seen, with occasional setbacks through mass global extinctions.
There have been five such extinctions so far, indicated here by lightning flashes. A sixth major decline is now
underway as the result of human activity. Source: THE DIVERSITY OF LIFE by Edward O. Wilson,
Cambridge, Mass.: The Belknap Press of Harvard University Press, Copyright © 1992 by Edward O. Wilson.
Used by permission. All rights reserved.
past (Simpson 1952), but far fewer are thought to exist today. Thus, species extinction has been
almost as common as origination. Much evidence exists that major extinction events, like that at
the end of the Cretaceous, are regularly spaced in geologic time (Raup 1986), but smaller turnovers
occur as well. The trend for biodiversity has been upward for the past 600 million years even in the
face of such extinctions (Figure 14.2). Global biodiversity of plants (Figure 14.3) peaked in the
Cenozoic because: (i) the aerobic environment was created, (ii) land masses became fragmented,
and (iii) species were increasingly pushed into regional and local ecosystems of the developing
landscape (Wilson 2010). The number of plant species in local floras has more than tripled in the
past 100 million years (Figure 14.3). Despite this seemingly exploding biodiversity, a sixth great
extinction event is thought by many to be currently occurring (Wilson 2010). In this context, we
consider the value of biodiversity, as well as the increasing role of biodiversity in shaping the sci-
ences of ecosystem conservation, management, and restoration.
VALUE OF BIODIVERSITY
There are many reasons for conserving, promoting, and managing biodiversity (Berry et al. 2018;
Burton et al. 1992), and a review of the extensive literature confirming the value of biodiversity is
well beyond our scope here. Biodiversity maintains or provides important ecosystem services, or
the benefits that ecosystems provide to humans (Cardinale et al. 2012; Chapter 23). Biodiversity
has many ecological benefits for ecosystems and their functioning (Lefcheck et al. 2015; Cardinale
et al. 2006; Chapin et al. 2000), some of which we discuss later in this chapter. A sample of such
benefits might be renewable resources (Cardinale et al. 2012; Piotto 2008), reducing plant invasion
(Quijas et al. 2010; Levine et al. 2004), reducing insect pests (Letourneau et al. 2009), improving
soil quality (Quijas et al. 2010), increasing carbon sequestration (Cardinale et al. 2012), and indi-
cating the loss of ecosystem integrity (Burton et al. 1992), among many others.
100
Local Diversity of Plants

Average Number of Species per Flora
80
60
40
20
DEVON. CARBON. PERM. TRIAS. JURASSIC CRETACEOUS TERTIARY
PALEOZOIC (in part) MESOZOIC CENOZOIC
408 360 320 286 248 213 144 66 16

Millions of Years Ago
F I G U R E 1 4 . 3 The average number of plant species found in local floras over geologic time. The number
of plants found in local floras has risen steadily since the invasion of the land by plants 400 million
years ago. The increase reflects a growing complexity in terrestrial ecosystems around the world.
Source: THE DIVERSITY OF LIFE by Edward O. Wilson, Cambridge, Mass.: The Belknap Press of
Harvard University Press, Copyright © 1992 by Edward O. Wilson. Used by permission. All rights reserved.
Biodiversity has been directly related to the critical ecosystem service of human health.
Opportunities for identifying natural products for future pharmaceutical value increase with
species diversity (Chivian and Bernstein 2008; Mendelsohn and Balick 1995). Pharmaceutical
compounds derived from wild plants are often of global importance. The Madagascar periwinkle,
for example, is used in the treatment of childhood leukemia and Hodgkin’s disease, and is report-
edly worth millions in sales each year (Shiva 1990). Likewise, the western yew tree is the source of
taxol, a compound found to have strong activity against a number of cancers (Wani et al. 1971).
High biodiversity is also thought to be associated with reduced transmission and emergence of
diseases that affect humans (Keesing and Ostfeld 2015), such as zoonotic diseases (Keesing and
Ostfeld 2021; Keesing et al. 2010) and lyme disease (Ostfeld and Keesing 2000). Moreover, green
space and wilderness that support and maintain biological diversity can promote human well-
being (Easley et al. 1990; Kaplan 1992; Thompson and Barton 1994; Gaston et al. 2007), an impor-
tant part of human mental health.
Economic value may be placed on the ecosystem services that biodiversity provides (Paul
et al. 2020; Hanley and Perrings 2019; Edwards and Abivardi 1998; Gowdy 1997), although the
valuation of biodiversity remains an important discussion among environmental economists.
Costanza et al. (1997) estimated the market value of the world’s ecosystem services to be about
US$33 trillion/year (1012), which was more than the global gross national product at the time
(US$18 trillion/year). At that time, forests contributed 14% of the $33 trillion, and wetlands (in part
forested swamps and floodplains) contributed another 15%. A revised estimate in 2011 showed a
global loss in ecosystem services due to land use and management changes (Costanza et al. 2014).
Moreover, the future value of ecosystem services may increase by as much as US$30 trillion/year
or decrease by as much as US$51 trillion/year depending on land use and management scenarios
(Kubiszewski et al. 2020). This bottom-line approach provides a powerful incentive for conservation
of the natural capital stock. Notably, many ecologists and others would additionally, or alterna-
tively, cite very real and intrinsic reasons for sustaining the Earth’s ecosystems, reasoning that the
Measuring Diversity 331
existence of species and ecosystems has value irrespective of humans (e.g., Leopold 1949;
Regan 1981; Norton 1982; Naess 1986; Rowe 1990; Ghilarov 2000; Alho 2008; Justus et al. 2009;
Fearnside 2021).
COMMON THREATS TO DIVERSITY

Species extinction is the most well-known human impact upon biodiversity, and it has escalated
with increasing human populations and their need to utilize natural ecosystems for agriculture,
forest management, and urbanization. The main threat to diversity in North America and around
the world is therefore the human population, particularly as it increases in size. All threats to
diversity stem from humans either directly or indirectly. A detailed discussion of threats to diver-
sity is well beyond our scope, but the loss of species has occurred largely because of four main
human impacts:
Direct exploitation of species by humans, either by removing every last individual of a species or by
reducing their populations to the point that they have little resilience to events that might cause
their extinction. Trees or other woody plants are less commonly exploited to extinction by hu-
mans than are animals. However, many conservationists have raised concerns about fuelwood
collection in developing countries whose increasing demand is likely to lead to species extinc-
tions in areas where it is concentrated (Gaston and Spicer 2014).
Habitat loss and fragmentation is perhaps the most widespread threat to diversity in forests. Forest
loss via conversion to agriculture and development is extensive worldwide, estimated as a 29%
decrease since significant human disturbance at the turn of the twenty-first century (Klein Gold-
ewijk 2001). Similar land-use changes are likely to occur in the future not only because of continued
human population growth but also due to climate change (Chapter 20). Remaining forests are
likely to support fewer species due to their fragmentation into small, isolated patches (Chapter 22).
Invasive species that have the potential to outcompete native species or alter the ecosystems in
which native species occur are an increasingly important threat to diversity (Chapter 21). Both
native and introduced species respond to similar site factors in successfully establishing in an
ecosystem (Sax 2002), and thus ecosystems with higher native species richness also tend to have
more invasive species. In a summary of the 100 worst invasive species across the world based on
their impacts on biodiversity and human activities (Baskin 2002), 25 of the species were woody
land plants and 4 were insects or pathogens of forest trees.
Climate change is a human impact likely to have enormous consequences for biodiversity in the
next 50–100 years. Changing climate is predicted to alter the distribution of many species at a rate
too high for species to adapt to the changing environmental conditions. Species extinctions due to
climate change have been predicted to reach as high as 54% (Urban 2015). Although some species
are likely able to disperse to new suitable habitat, a recent study of 538 plants and animals by
Román-Palacios and Wiens (2020) suggested that 57–74% of the species would not disperse
quickly enough to avoid extinction. However, they predicted that species adaptations were more
likely to occur than previously thought, such that only 16–34% of the species would go extinct in
the next 50 years. The effects of climate change on forests are described in detail in Chapter 20.
MEASURING DIVERSITY
LEVELS OF DIVERSITY
Diversity includes two important components: richness and evenness. Richness refers to the num-
ber of units (alleles, species, families, communities, ecosystems) per unit area, and evenness refers
to their abundance or dominance relative to one another. These two concepts can be applied and
measured at many levels of organization: ecosphere and continental-level ecosystems, regional

ecosystems, local ecosystems, species, and the genetic diversity of allele frequencies within popu-
lations of a species. Species are typically the foci of biodiversity measurement because they are
easily observed, and the most common taxonomic unit used in ecological studies of forests. The
complexity of forest ecosystems is of note, however, because the patterns of diversity for trees may
not reflect those of ground-cover vegetation, birds, or microorganisms.
Biodiversity is typically studied at the local level in areas of <25 ha, but may also be exam-
ined at broader spatial scales. Diversity at broad scales is often referred to in the literature as
landscape or regional diversity. Here, the focus may be on relatively broad vegetative cover types
such as those in Yellowstone National Park (Romme 1982; see Chapter 22) or the White Mountains
of New Hampshire (Reiners and Lang 1979). At broad spatial scales, diversity results from the
interaction of multiple patterns of: (i) species distributions along gradients of limiting factors, (ii)
recurrent spatial patterns of local ecosystems, and/or (iii) communities in different stages of
succession following disturbance. Diversity of taxonomic units above the species level (i.e., genera,
family, order) are often examined in continental-level studies (Currie and Paquin 1987; Latham
and Ricklefs 1993a, b). At this spatial scale, diversity is related to macroclimate, physiography,
continental drift, and the historical development of plant floras. By contrast, observed patterns of
richness and evenness at local scales (<25 ha) are determined by interspecific interactions and
local site factors of microclimate, landform, and soil.
Whittaker (1960, 1972, 1975, 1977) was among the first ecologists to recognize that species
diversity may be measured in several ways and at several spatial scales. First, he recognized four
increasingly larger spatial scales at which inventory diversity could be measured:
Point diversity: at a microsite scale with samples taken from within a homogeneous site, that is,
a microsite within a local landscape ecosystem type (e.g., 100–500 m2).
Alpha diversity: at a homogeneous site, that is, for a given landscape ecosystem type or community
in a homogeneous site (micro-ecosystem level, e.g., <1 to several hundred hectares). Simply, it is
defined as the variety of organisms at a particular site. It is sometime referred to as local diversity
in the literature.
Gamma diversity: at a larger landscape unit, that is, for an island or a regional ecosystem. Gamma
diversity includes the diversity of all ecosystems within a region (e.g., 625–2500 ha).
Epsilon diversity: total diversity for an area encompassing a group of areas of gamma diversity,
that is, macro-ecosystem level (e.g., >2500 ha).
As an analogy, Magurran (1988) described a leaf as point diversity, a plant as a unit of alpha
diversity, a group of plants as gamma diversity, and the entire forest containing the group of plants
as epsilon diversity.
Whittaker also recognized that species composition changes across the landscape, both along
environmental gradients, and from one generally homogeneous site to the next. This change from
one area to the next, regardless of scale, he called beta diversity. Beta diversity measures how dif-
ferent two or more areas are from each other based on the variety of species present. Beta diversity
may compare the similarity of sites using species distinction, or it may relate species richness across
different spatial or temporal scales. Jurasinski et al. (2009) proposed that the first of these versions
of beta diversity be termed differentiation diversity, while the second should be termed propor-
tional diversity.
Diversity may also be used to describe the structural diversity of communities as well as their
composition. In forested systems, this analysis focuses on the spatial pattern of species occurrence
(random, regular, or patchy), stand density, diameter class distribution, and the vertical layering of
a forest. Structural diversity is of considerable importance to wildlife in that the number of animals
capable of using a forested system depends on the presence of appropriate areas for nesting,
feeding, resting, and hiding. Structural diversity for old-growth, mixed-age or -size, and mixed-
species stands is typically high, whereas that for plantations or single-age, single-species stands is
low. Storch et al. (2018) developed an index of structural diversity for forests in Baden-Württemberg,
Germany using 11 variables relating to tree diameter, tree height, deadwood decay classes, dead-
wood diameter, bark diversity, diversity of flowering and fruiting, and tree species richness. Ex-
pressed as a value between zero and one, the index may be used to examine structural diversity
across large-scale forest inventories to support biodiversity monitoring. Similarly, McElhinny et al.
(2006) developed a structural diversity index for dry forests of Australia that included 13 variables
related to vegetation cover of forest strata, perennial species richness, life-form richness, diameter
and basal area of live trees, number of hollow or dead trees, length of deadwood, and litter dry
weights. The index was then used to compare and differentiate among forest structures across dif-
ferent sites.
MEASUREMENT
Inventory Diversity: Alpha Diversity There are various ways to measure species diver-
sity. Perhaps the simplest measurement is to count the number of species present in a designated
area. This number is species richness: the oldest, most fundamental, and least ambiguous of the
diversity measurements (Peet 1974). Richness varies with the area sampled, so it is often expressed
as number of species per unit area. Richness will increase with the area sampled up to a point
where most or all species have been captured in the sample. Increasing the sample area at this
point does not increase the species number. In using sample plots to determine the richness for a
given ecosystem, it is therefore important to use plots large or numerous enough to encompass
most of the species. The total number of species for most major organism groups is difficult to
determine for large, forested areas. The species–area relationship is consistent for homogeneous
areas, such as a single ecosystem type with relatively little microtopographic variation, relatively
small gaps or disturbance features, and uniform climate. Once the sampled area expands beyond a
single local ecosystem, however, environmental heterogeneity increases, and different ecosystems
with different species are included in the sample. It is therefore important to recognize ecosystem
boundaries if reliable estimates of alpha diversity of species are to be obtained.
Diversity also depends on the distribution of species across an area, termed evenness, or the
degree to which all species share dominance in an area. For example, forest stands A and B in
Figure 14.4 both have eight species and thus the same richness. Stand A is dominated by one or
two species, with the other species being relatively uncommon, whereas stand B exhibits a more
equal abundance of all species and thus exhibits greater evenness. In this case, stand B is more
diverse because it has the more equal abundance of the species with the same number of species.
Two measures of diversity in Figure 14.4 that combine richness and evenness into one index—
Shannon’s and Simpson’s diversity indices—are higher for stand B than for stand A.
Many diversity indices summarize richness and evenness into a single number, each with its theo-
retical advantage under specific circumstances. Two of the most commonly used indices for
measuring species diversity are Simpson’s index and the Shannon–Wiener (or Shannon’s) index.
The Shannon–Wiener index is probably the most commonly used diversity index, and is
computed as:
s
H pi ln pi ,
i 1
where S is the number of species in the sample, and pi is the proportion of individuals that belong
to species i. H′ generally ranges from a low of around 1.5 to a high of 4.5, rarely exceeding this
value (Margalef 1972). The Shannon–Wiener index is sensitive to the number of species in a
0.8 F I G U R E 14.4 Comparison of

Stand A species’ alpha diversity exhibited by
0.7 two hypothetical stands having the
Richness = 8 same number of tree species. Species
0.6 richness for each stand is eight. Two
Shannon’s H’ = –Σpi In pi = 1.03 diversity indices are given, where pi is
Relative Abundance
the proportional abundance of each

0.5
species, i, and the summation is
Simpson’s D = 1 – (Σpi2) = 0.50
performed over all eight species.
0.4
Because stand B exhibits greater
evenness, with less dominance by any
0.3
one species, it has greater overall
diversity than stand A, as measured
0.2 by the indices. Source: Reprinted with
permission by the Forestry Chronicle
0.1 from Burton et al. (1992).
0
#1 #2 #3 #4 #5 #6 #7 #8
Species (in Descending Order)
0.8
Stand B
0.7
Richness = 8
0.6
Shannon’s H’ = 1.84
Relative Abundance
0.5
Simpson’s D = 0.87
0.4
0.3
0.2
0.1
0
#1 #2 #3 #4 #5 #6 #7 #8
Species (in Descending Order)
sample, so it is biased toward measuring species richness (Figure 14.4). The evenness component
of H′ is computed as (Magurran 1988):
H H
E .
H max ln S
In this computation, the Shannon–Wiener index is scaled by its maximum possible value (H′ max),
which would occur if all species had equal abundance and happens to be the natural log of S.
Simpson (1949) developed an index of diversity that is computed as:
S ni ni 1
D ,
i 1 N N 1
where ni is the number of individuals in species i and N is the total number of species in the
sample. Diversity is inversely related to D, so Simpson’s index is usually expressed as its complement,
1-D or 1/D. Simpson’s index is sensitive to the most common or dominant species in a community
and is relatively insensitive to rare species. For this reason, Simpson’s index, or its complement
(1-D), is sometimes used as a measure of evenness (Figure 14.4).
Given reasonable standardization, multiple diversity indices computed for the same data
tend to be correlated, and sampled areas will rank similarly in diversity as measured by various
indices. However, diversity indices vary in how they are computed and interpreted, so it is inappro-
priate to compare diversity measured by one index with that measured by another. Diversity values
will also vary depending on which organisms or forest strata (overstory; ground cover, all layers)
are sampled, the specific area sampled, and the method of sampling. Thus, the diversity index
sensitive to the factors of interest in the study should be selected.
Diversity may also be assessed with species abundance models, which express commonly
observed abundance distributions of species. Examples of such models include the log normal dis-
tribution, the geometric series, the logarithmic series, and MacArthur’s broken-stick model (Magur-
ran 1988). A common technique is to plot species abundance or dominance against their rank order
of abundance, dominance, or coverage. These species-abundance plots can be used to compare not
only the relative richness of communities but also to provide insight into evenness. Figure 14.5
shows species-abundance curves for ground-cover plants sampled in two very different ecosystem
types on glaciated terrain in northern Lower Michigan: (i) an excessively drained, infertile, sandy
outwash plain supporting a sparse, short overstory of small bigtooth and trembling aspens (ecosys-
tem type 1), and (ii) a moderately well-drained, loamy, fertile moraine supporting a moderately
dense canopy of tall bigtooth and trembling aspens, white ash, paper birch, and basswood (ecosys-
tem type 116). In this example, species abundance is measured as percent coverage. Ecosystem type
1 shows lower ground-cover species richness (37 versus 69 species in ecosystem type 116) and a
sharp decrease in dominance from the most dominant species (mean coverage = 44%) to the second
most dominant species (mean coverage = 2%). Such a plot is characteristic of ecosystems extreme
in factors such as microclimate, fire frequency or severity, light level, or fertility, and in an early-
successional stage. By contrast, ecosystem type 116 has nearly twice the species richness as type 1
and its plot exhibits a more gradual drop in dominance, from about 3.3% to <1%, from the most to
least dominant species (Figure 14.5b). This plot is more typical of ecosystems that are not at the
extremes of microclimate, fertility, light, or disturbance, and are in a mid-successional stage.
Differentiation or Beta Diversity If the interest is in measuring how different or similar

the variety of species is across a range of samples (Magurran 1988), then beta diversity is esti-
mated. Communities of a given area that have very few species in common collectively have high
beta diversity, and vice versa. This difference in species variety may be determined for ecosystems
or communities that occur along an environmental gradient or within an area supporting diverse
sites. Because the term “beta diversity” gives little clue to the spatial or ecological context of the
diversity that is measured, beta diversity may be usefully understood as differentiation diversity
(Whittaker 1977; Magurran 1988)—a measure of species diversity along a gradient containing dif-
ferent ecosystems or throughout an area where multiple ecosystems occur.
Beta diversity is typically measured by determining the similarity in species composition
between pairs of communities or ecosystems. Six methods of measuring beta diversity are assessed
and discussed by Wilson and Shmida (1984) and Magurran (1988). Two basic approaches are sim-
ilarity coefficients and measuring change along gradients using a variety of indices. Percentage
similarity (PS) and the Jaccard and Sorensen similarity coefficients are widely used where the
presence or absence of species is known. These methods relate the number of shared species to the
total number of species in two samples. For example, if two samples are taken and each has the
same 10 species, the similarity is 100%, but if only five species are shared, then the similarity is only
(a)
45.0
40.0
Mean percentage cover
35.0
30.0
25.0
20.0
15.0
10.0
5.0
0.0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37
Dominance rank
(b)
3.5
3.0
Mean percentage cover
2.5
2.0
1.5
1.0
0.5
0.0
1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69
Dominance rank
F I G U R E 1 4 . 5 Comparison of ground-cover species diversity of two ecosystems using plots of species

abundance: (a) Ecosystem type 1 occurs on an outwash plain landform with excessively drained, infertile,
sandy soil. (b) Ecosystem type 116 occurs on a moraine with moderately well-drained, fertile, loamy soil.
Abundance is measured by percentage cover. Ground-cover species for each ecosystem are ranked in order
by their percent coverage along the x-axis (1 = species with greatest abundance). Richness is indicated by the
number of species shown at the far right on the x-axis. Evenness is indicated by the shape of the curve, with
ecosystem 116 being more “even” in distribution than ecosystem type 1. Source: Pearsall (1995) / Pearsall.
50%. This example of percentage similarity is a qualitative (presence/absence) estimate of beta

diversity; the Sorensen index contains a quantitative estimate that can be used when abundance
data are available. Another quantitative measure, the Morista–Horn coefficient, is independent of
sample size and thus is preferred among all similarity measures (Wolda 1981).a
DIVERSITY OF LANDSCAPE ECOSYSTEMS

Most diversity studies have focused primarily on species, especially the alpha and beta diversities
defined by Whittaker (1977). Rather than strictly biota, ecosystem diversity is concerned with local
and regional landscape ecosystems. Diversity of landscape ecosystems can be a surrogate for biodi-
versity (Bailey et al. 2017), and conservation of “geodiversity” is considered as important as
conservation of biodiversity (Shaffer 2015; Schrodt et al. 2019) as exemplified in the “Conserving
the Stage” approach (Anderson and Ferree 2010) that is now the basis for The Nature Conservan-
cy’s conservation planning for the continental US (Anderson et al. 2016). Diversity exists and can
be treated at a variety of levels of ecological organization, but applying the concepts of diversity to
the hierarchically nested series of ecosystems or to vegetation types throughout a large landscape
is relatively uncommon. Exceptions are Romme’s (1982) study of forest ages across the landscape
in Yellowstone National Park (Chapter 22), Loehle and Wein’s (1994) study of forest vegetation
types in western Tennessee, and Pearsall’s (1995) study of ecosystem diversity in northern Lower
Michigan. Recent work by Anderson et al. (2014, 2016) has incorporated ecosystem diversity—
defined as “the variety of landforms created by an area’s topography, together with the range of its
elevation gradients”—into a landscape model of climate resilience. Distinguishing and mapping
landscape ecosystems at multiple scales provide the spatial framework for determining ecosystem
diversity at multiple spatial scales.
EXAMPLES OF DIVERSITY
Ground-Cover Species Diversity in Northern Lower Michigan In the following text,
we examine measures of species diversity on a large landscape in northern Lower Michigan.
The purpose is not only to compare various measures of alpha and beta diversity but also to
examine diversity at several ecosystem spatial scales (major and minor landforms) and soil
types, by groups of ecosystems, and by individual ecosystem types. We also consider ecosystem
diversity.
Ecosystem Groups Landscape ecosystems of multiple scales were distinguished and mapped at
the 4000-ha landscape of the University of Michigan Biological Station in northern Lower Michi-
gan (Lapin and Barnes 1995; Pearsall 1995; Zogg and Barnes 1995). The terrain was shaped by two
late-Wisconsin glacial advances and retreats between 15 000 and 13 000 years ago. The area was
highly disturbed by logging and repeated post-logging fires in the late nineteenth and early twen-
tieth centuries. Early-successional species, primarily bigtooth and trembling aspens, now domi-
nate the overstory of upland ecosystems, having replaced many kinds of upland and lowland coni-
fers and hemlock–northern hardwood forests common in the pre-European colonization forest
over 200 years ago.
Landscape ecosystems were distinguished using physiography, microclimate, soil, and
ground-cover vegetation, and mapped at four hierarchical scales: major landforms, minor land-
forms, ecosystem groups, and local ecosystem types. The three major landforms of outwash plain,
ice-contact terrain, and interlobate moraine were initially subdivided into 21 ecosystem groups
(Lapin 1990). Six of these 21 groups were used to analyze alpha and beta diversity of ground-cover
vegetation (Figure 14.6). The key site features of the six ecosystem types are given in Table 14.1.
Ground-cover species richness ranged from 27 species to 16 per 100 m2 plot, and Shannon–
Wiener heterogeneity ranged from 2.21 to 0.61 (Table 14.2). Ecosystem group 20 (moist and nutrient-
rich) was the richest and most heterogeneous, and ecosystem group 4 (dry and nutrient-poor) was
the least diverse. Ecosystem groups were ranked in alpha diversity as follows, from more diverse to
less diverse: (i) group 20 (moist and nutrient-rich), (ii) group 17 (moderately moist and moderately
nutrient-rich), (iii) group 11 (dry and nutrient-rich), (iv) group 2 (moist, nutrient-poor, and climati-
cally extreme), (v) group 1 (dry, nutrient-poor, and climatically extreme), and (vi) group 4 (dry and
nutrient-poor) (Tables 14.1 and 14.2). Multiple-comparison methods often showed significant pair-
wise differences (Table 14.2).
Species richness and diversity tended to increase with moisture and nutrient availability, as
long as factors such as temperature, light, and herbivory were not limiting. Fire-regenerated aspens
that currently dominate the overstory of all ecosystem groups provide a favorable light environment
for many vascular plants. Pre-European colonization forests—dominated by eastern white pine,
American beech, and eastern hemlock (Kilburn 1960)—were probably denser than today’s relatively
open canopies of aspens. As such, ground-cover diversity measures applied to forests as they were
prior to the arrival of Europeans are likely to yield very different results compared to the same
areas today.
(a)
West East
Low-level High-level
Moraine outwash plain Moraine outwash plain
Eco Group 17
Eco Group 20
Eco Groups
1 and 2 Eco Group 4
(b)
North South
Ice-contact High-level
terrain Douglas Lake outwash plain
Eco Group 11 Eco Group 4
F I G U R E 1 4 . 6 Physiographic diagrams illustrating the location of six ecosystem groups on major landforms
(outwash plain, moraine, ice-contact terrain) of the University of Michigan Biological Station, Emmet and
Cheboygan counties, northern Lower Michigan: (a) West–east transect south of Douglas Lake showing
location of ecosystem groups 1, 2, 17, 20, and 4. The low-level outwash plain (Pellston Plain) is a huge frost
pocket that is over 3 km wide between high interlobate moraines; vertical scale exaggerated. (b) North–south
transect showing the location of ecosystem groups 11 and 4. Source: Lapin and Barnes (1995). Reprinted by
permission of Blackwell Science, Inc.
Table 14.1 Comparative summary of site conditions of six landscape ecosystem groups of the University
of Michigan Biological Station, Emmet and Cheboygan Counties, northern Lower Michigan.
Ecosystem groupa Site condition Physiography Soil and drainage
20 Moist, nutrient rich Interlobate moraine; flat Loamy sand over clayey
and moderate slopes calcareous till; well drained
17 Moderately moist, Interlobate moraine; flat Sandy soil with many
moderately nutrient rich heavy-textured bands;
noncalcareous; well drained
11 Dry, nutrient rich Ice-contact terrain Calcareous, gravelly medium
sand, somewhat
excessively drained
2 Dry to seasonally moist; Low-lying outwash Deep, non-calcareous medium
nutrient poor; wide daily plain between two sand; moderately well drained
and seasonal tempera- moraines; flat
ture extremes
1 Dry, nutrient poor; wide Low-lying outwash Deep, non-calcareous medium
daily and seasonal plain between two sand; excessively drained
temperature extremes moraines; flat
4 Dry, nutrient poor High-level outwash Deep, non-calcareous medium
plain; flat sand; excessively drained
a
Pre-European colonization cover type—groups 20, 17, and 11: hemlock–northern hardwoods; groups 1, 2, and 4: eastern white
pine, red pine, and northern red oak.
a
Present cover type—groups 20, 17, and 4: bigtooth aspen; groups 1 and 2: trembling and bigtooth aspens; group 11: northern
hardwood species and bigtooth aspen.
Source: Lapin and Barnes (1995). Reprinted with permission of Blackwell Science, Inc.
Table 14.2 Comparison of alpha diversity indices for ground-cover plant species among six ecosystem
groups of the University of Michigan Biological Station, Emmet and Cheboygan Counties, northern
Lower Michigan.
Ecosystem groupa
Index b
20 17 11 2 1 4
SGC 27.2 18.4 17.2 23.6 19.4 15.8
a bc bc ac bc b
SH 14.2 7.8 9.0 10.2 9.4 7.2
a b b b b b
SW 13.0 10.6 8.2 13.4 10.0 8.6
a abc c ab ac c
H′ 2.21 1.66 1.15 0.99 0.70 0.61
a ab b b
1-D 0.827 0.704 0.537 0.420 0.268 0.241
a ab bc cd cd d
E 0.68 0.57 0.41 0.31 0.23 0.22
a a b bc c c
a
Number of plots = 5 for all groups. There was no significant difference between groups that share a letter, alpha = 0.1.
b
SGC = ground-cover species richness, SH = herbaceous species richness, SW = woody species richness, H′ = Shannon–Wiener
heterogeneity, 1-D = Simpson’s heterogeneity, E = Shannon–Wiener evenness.
Diversity measures based on ground-cover vegetation show a much higher level of diver-
sity than would be seen by examining only the overstory vegetation. On this landscape, there are
significant differences in ground-cover composition with an overstory dominated only by two
closely related tree species—bigtooth and trembling aspens. This relationship emphasizes that
alpha and beta species diversity to some extent are driven by how sample plots are defined and
bounded. If plot boundaries are set by community type (i.e., by the bigtooth aspen cover type), it
may be assumed that any one plot will be representative of the whole cover type. If plot boundaries
are determined by ecosystem type (distinguished based on landform, soil, and climatic factors),
then patterns of diversity emerge within the cover types from which the idea of representativeness
can be better gauged.
Analysis of beta diversity further demonstrates the differences in ground-cover vegetation
among ecosystem groups using Jaccard’s coefficient and percentage similarity (PS; Table 14.3). Jac-
card’s coefficient is a qualitative measure based on species presence/absence, whereas PS is the
similarity of the percentage coverage of the ground-cover species (Krebs 1989). Jaccard’s coeffi-
cient shows that ecosystem groups 20 (moist, nutrient-rich) and 11 (dry, nutrient-rich) have the
most species in common (Cj = 56%) and contain several species that are not present in other
groups. Three nutrient-rich groups (20, 17, and 11) characterized by pre-European colonization
vegetation of hemlock–northern hardwood forest are the most similar ecosystem groups based on
presence–absence data, but the most similar two groups are only slightly over 50% alike. The least
similar groups (less than one-third similar) are the various combinations of nutrient-rich (20, 17,
and 11) and the nutrient-poor, climatically extreme ecosystem group (1).
PS indicates extreme differences in species coverage among the ecosystem groups (Table 14.3).
Percentage similarity ranged from 2 to 92%, with most pairs of ecosystem groups ranging from 12 to
36% similar. Ecosystem group 11 (dry, nutrient-rich) is very dissimilar to the three nutrient-poor
groups (2, 4, and 1; PS = 2–3%). The pairs with greatest PS (92) are groups 1 and 4, both dry and
Table 14.3 Comparison of beta diversity measures for ecosystem groups of the University of Michigan
Biological Station, Emmet and Cheboygan counties, Northern Lower Michigan. Jaccard’s coefficienta for
all combination pairs of six ecosystem groups is shown outside parentheses, and the percentage
similarityb (PS) is shown within parentheses.
Group 20 17 11 2 1 4
20 —
17 53 (57) —
11 56 (28) 47 (36) —
2 42 (23) 36 (14) 31 (2) —
1 33 (20) 30 (12) 28 (2) 39 (56) —
4 36 (23) 46 (16) 34 (3) 39 (55) 37 (92) —
a
Jaccard’s coefficient (species presence/absence): Cj = j/(a + b – j), where j equals the number of species present in both sam-
ples, a equals the number of species present in sample 1, and b equals the number of species present in sample 2
(Magurran 1988).
b
Percentage similarity, PS = minimum (p1i, p2i), based on the coverage of a species in a sample where p1i equals the proportion
of species i in sample 1 and p2i equals the proportion of species i in sample 2 (Krebs 1989).
nutrient-poor, although they occur on different landforms and are not spatially adjacent (Figure 14.6).
Notably, ecosystem groups may share many species, but the species’ relative coverage may differ
widely across the ecosystems. For example, ecosystem groups 20 and 11 share 56% of the species (Ta-
ble 14.3), but their PS in coverage is only 28%. Conversely, groups 1 and 2 share only 39% of the same
species, but are 56% similar in species coverage.
Ecosystem Types The classification and mapping of 125 local landscape ecosystem types
(finer units within ecosystem groups) for the 4000-ha tract facilitated the study of alpha and beta
diversity of ground-cover species within and among diverse major and minor landforms. For
example, several measures of alpha diversity in Table 14.4 illustrate differences in diversity for
major landforms (outwash plains and moraines), groups of ecosystem types within these landforms,
and ecosystem types with calcareous versus non-calcareous soils. Notice also in Table 14.4 that the
diversity values are presented in two ways: for all ecosystems and plots within a given unit and the
average per unit area (within parentheses). For example, total species richness for outwash plains
with 41 ecosystem types is 182, whereas the average per 150 m2 plot is 20.2. For major landforms,
moraines exhibit greater richness and heterogeneity of ground-cover species per unit area than out-
wash plains (26 versus 20, respectively) despite the greater number of ecosystem types in the out-
wash plains. Ecologically, this pattern relates to the greater soil-water and nutrient availability on
moraines compared to the outwash plains.
Notably, the differences in ground-cover diversity between major landforms at the Biological Station
have persisted for 25 years since the data shown in Table 14.4 were collected. Ricart et al. (2020) re-sampled
ground-cover vegetation at the major landforms in 2015 and found that species richness and Shannon–
Wiener heterogeneity remained higher on moraines relative to outwash plains. Values of both metrics
declined over the 25-year period for both major landforms, but these results show the close relationship
between diversity and physical site factors even after a quarter-century of forest succession.
Within moraines, diversity is higher in the bigtooth aspen-dominated interlobate moraine
compared to the hemlock–northern hardwood-dominated Colonial Point moraine (Table 14.4).
These landforms are very similar in their high soil-water and nutrient availability, but species rich-
ness is low on the Colonial Point moraine due to the dense overstory and understory of beech,
sugar maple, hemlock, and other species that shade the forest floor. By contrast, open crowns of
Table 14.4 Alpha diversity measure of ground-cover diversity for upland ecosystem groups on outwash
plain and moraine landforms of the University of Michigan Biological Station, Emmet and Cheboygan
Counties, northern Lower Michigan.
Number Indexb
Number
Ecosystem group of ecosystem
of plotsa S H′ 1-D E
types
Major landforms
Outwash plain 41 175 182 (20.2) 1.77 (1.22) 0.54 (0.48) 0.34 (0.41)
Moraine 9 48 141 (25.8) 3.06 (1.84) 0.93 (0.71) 0.62 (0.58)
Minor landforms
Interlobate moraine 3 26 123 (31.5) 3.05 (21.3) 0.93 (0.80) 0.63 (0.62)
Colonial 6 22 76 (19.1) 2.29 (1.49) 0.81 (0.60) 0.53 (0.52)
Point moraine
Low-level 11 34 132 (23.6) 1.53 (1.16) 0.46 (0.45) 0.31 (0.36)
outwash plain
High-level 21 121 142 (18.9) 1.75 (1.17) 0.56 (0.47) 0.35 (0.40)
outwash plain
Soil type in
outwash plains
Calcareous outwash 18 70 142 (22.4) 1.89 (1.33) 0.60 (0.52) 0.38 (0.43)
Non-calcareous 23 105 149 (18.3) 1.63 (1.05) 0.50 (0.43) 0.32 (0.36)
outwash
a
Plot size = 150 m2.
b
S = ground-cover species richness, H′ = Shannon–Wiener heterogeneity, 1-D = complement of Simpson’s index, E = Shannon–
Wiener evenness. Diversity values outside parentheses are based on all ecosystem types and plots for a given unit. Diversity
values inside parentheses are expressed on a per unit area basis, in this study 150 m2. For example, there is a total of 182 different
ground-cover species in the 175 plots and 41 ecosystem types in outwash plain landform, whereas the average number of
species per unit area (150 m2) is 20.2. See text for discussion.
Source: Courtesy of Douglas Pearsall.
bigtooth aspens dominate ecosystems of the Interlobate moraine (Figure 14.6). In general, for
these two landforms, richness is highest when canopy cover is intermediate, but richness is lowest
where canopy cover is high and the low light at the forest floor is limiting to many species.
Two contrasts are shown in Table 14.4 for outwash plains—between (i) outwash plains
of different elevation, where low-lying outwash plains are climatically extreme (giant frost
pocket, Figure 14.6), and (ii) non-calcareous versus calcareous ecosystem types. The contrast
of the colder low-level outwash plains with warmer high-level plains shows their similar
values of H′. Note that the high-level outwash contains nearly twice as many ecosystem types
as the low-level outwash, resulting in its higher diversity measures. The contrast of calcareous
versus non-calcareous ecosystems in outwash plains (with similar overstory coverage of
aspens) shows that the more nutrient rich calcareous outwash exhibits greater diversity. Cal-
careous ecosystems are able to support both nutrient-requiring and non-nutrient-requiring
plants, whereas some plants requiring relatively high nutrient levels are excluded in the dry,
non-calcareous outwash soils.
Beta diversity of ground-cover vegetation among paired comparisons of ecosystems in
upland landforms is essentially the same for ecosystem types of outwash plains and moraines. Beta
diversity is relatively low in the Interlobate and Colonial Point moraines where there are relatively
few ecosystem types and species richness and coverage are similar.
ECOSYSTEM DIVERSITY
Thus far we have considered species diversity, but similar analyses of richness and evenness may
be completed for ecosystems on landscapes where they have been distinguished and mapped at a
given scale (see Figures 2.11 and 11.11 in Chapters 2 and 11, respectively). For example, the Pine
River flows through a large expanse of Lake Superior beach in the Huron Mountains of Upper
Michigan (Figure 14.7). Six riverine ecosystem types are tightly clustered along the river, and their
abundance, size, and spatial distribution contrast with the single, flat, fire-prone beach ecosystem
dominated by jack pine. The riverine ecosystems have diverse soil conditions ranging from
extremely acid and infertile peat to very fertile muck, as well as a remarkable richness of plant
species of different life forms (Simpson et al. 1990). Similar “hotspots” of ecosystem diversity were
identified at the 4000-ha Biological Station tract described in the earlier text (Pearsall 1995). Impor-
tantly, high biodiversity is associated with these hotspots of ecosystem diversity in the glacial ter-
rain of Michigan. Thus, land managers may efficiently prioritize areas of high biodiversity for
preservation by identifying areas of high ecosystem diversity.
Conventional alpha and beta diversity indices may be used to quantitatively assess the diver-
sity of landscape ecosystems and landforms. Pearsall (1995) used this approach to utilize the num-
ber of ecosystems, multiple characteristics of their physiography, soil, and vegetation, and their
pattern in the landscape to determine the ecosystem diversity of landforms and ecosystems of the
Biological Station tract. Ecosystem diversity was highest in outwash plains, and the centers of
highest diversity apparent from ecosystem maps were confirmed by this multivariate diver-
sity measure.
Lake Superior
N
2
Kilometers
Roads 0 .1 .2 .3 .4 .5
5
3
Pine
River 43
38 13 2
18 13 41 50
13 18 19
50
18 50 13
18 13 18 43
50
44
43 43
18
44
18
18
50
18 50
50
2
44
3
2
Pine 5
Lake
3
F I G U R E 1 4 . 7 Landscape ecosystem diversity. Comparison of the number, size, and pattern of landscape
ecosystem types of an area along the south shore of Lake Superior, Huron Mountain Club, Marquette
County, Upper Michigan. In contrast to the extensive beach terrace (ecosystem type 2), six small ecosystem
types (13, 18, 41, 43, 44, and 50) border the Pine River where it flows through the beach ecosystem.
Source: Simpson et al. (1990) / Simpson.
Causes of Species Diversity 343
CAUSES OF SPECIES DIVERSITY

In the following text, we examine biological diversity from the broad scale of continents to the fine
scale of ecosystem types, with the objective of exploring the factors working together to create the
patterns of diversity we observe today.
DIVERSITY AT CONTINENTAL AND SUBCONTINENTAL SCALES

Paleogeography and Continental Relationships The differences in tree species diversity
that exist among regions of the Earth have interested biogeographers and ecologists for at least two
centuries. Wet tropical forests have 10 times as many tree species as moist forests of the northern
and southern temperate zones. No single mechanism has gained general acceptance as an explana-
tion for this pattern, and several factors may contribute. Notably, ecologists have often, perhaps
inappropriately, invoked local-scale deterministic processes of competition and other interactions
to explain global-scale diversity patterns (Latham and Ricklefs 1993b). Nevertheless, differences in
diversity among regions with similar climate suggest that local and regional processes have impor-
tant implications for regional species richness, as do historical and evolutionary events.
Latham and Ricklefs (1993b) examined present-day taxonomic diversity among major moist
temperate forest regions of the Northern Hemisphere: (i) northern, central, and eastern Europe,
(ii) east-central Asia, (iii) Pacific slope of North America, and (iv) eastern North America. A total
of 1166 species comprise the characteristic tree flora of these regions, approximately in the ratio
2:12:1:4 (Table 14.5); eastern Asia has by far the most species and genera. A total of 63 genera occur
in either two or three of these regions (i.e., those neither endemic nor cosmopolitan), and 59 of
these 63 genera (94%) occur in eastern Asia. Eastern Asia and eastern North America share the
most genera (20), reflecting the well-known range disjunction of many plants inhabiting both
regions (Boufford and Spongberg 1983). However, the temperate forests of eastern and western
North America share few genera and none uniquely, despite their proximity.
The high diversity of temperate tree flora in eastern Asia compared to other regions of the
Northern Hemisphere probably arose from differences in colonization history and perhaps in
subsequent rates of proliferation of endemic taxa (Latham and Ricklefs 1993b). Colonization
may have occurred more frequently in Asia, giving rise to new native taxa and their geographical
spread among the temperate regions. Asia is characterized by many more species at the local
level than found in temperate forests elsewhere, reflecting this species proliferation. Consider-
able site differentiation resulting from mountainous terrain and the relative lack of continental
glaciation are reasons for such proliferation. Notably, most cosmopolitan taxa of temperate trees
Table 14.5 Summary of moist temperate forest trees in the Northern Hemisphere by taxonomic level
and geographic region.
Number of tree taxa characteristic of moist temperate forests in:

Taxonomic Northern, central, East-central Pacific slope of Eastern North Northern
level and eastern Europe Asia North America America Hemisphere (total)
Subclasses 5 9 6 9 10
Orders 16 37 14 26 39
Families 21 67 19 46 74
Genera 43 177 37 90 213
Species 124 729 68 253 1166
Source: Latham and Ricklefs (1993b) / with permission of University of Chicago Press. Reprinted with permission from Species
Diversity in Ecological Communities, Ricklefs and Schluter, eds., © 1993 by the University of Chicago. Reprinted with permission
of the University of Chicago Press.
are likely to have originated in eastern Asia and dispersed to Europe and North America. Diver-
sity differences between temperate and tropical flora reflect a physiological barrier to coloniza-
tion of temperate zones that can only be overcome by the evolution of freezing tolerance. The
important point is that latitudinal gradients in richness are best explained by historical and evo-
lutionary factors affecting land masses and floras rather than present-day ecological interactions.
Glaciation Continental glaciation has importantly affected species diversity, and some of these
effects are described in Chapter 15. Species displaced by glaciation may have gone extinct when
they were unable to migrate to their former locations. In central Europe, the Alps prevented
southerly migration of many species resulting in their extinction. Moreover, continental ice sheets
altered the climate of areas even at great distances from the advancing or retreating ice fronts;
these changes are discussed for North America in Chapter 15. Glaciation has also affected global
sea levels; low sea levels during the Pleistocene Era formed a land bridge between Asia and North
America, allowing immigration of humans into the Americas that facilitated elimination of many
large North American mammals and altered disturbance regimes of fire and flooding. The extent
of human influence on diversity in the Americas is unclear, but their introduction during the
Pleistocene had a remarkable effect on the structure and diversity of North American forests.
Latitude and Elevation Plant species richness decreases with both increasing elevation
and increasing latitude toward the poles (Billings 1973, 1995). This decrease results from evolu-
tionary and geographic processes over the very long term as well as local ecological conditions
such as decreasing mean annual and growing season temperatures, drought stress, and ultraviolet-
B irradiation at high elevations (Billings 1995).
The latitudinal gradient in diversity of plants from the tropics to the poles has generated
great scientific interest. The pattern has been demonstrated to have been present for flowering
plants throughout most of the Cretaceous (Crane and Lidgard 1989). Overall diversity for organ-
isms does not occur along a particularly smooth gradient from the poles to the equator, however;
the diversity gradient is much steeper in the Northern Hemisphere than south of the equator,
where diversity declines more slowly from the equator to the South Pole (Platnik 1991, 1992). The
diversity gradient is complex and not well understood, but the reasons for its existence are related
to (i) tropical areas having more land area than temperate and polar areas, such that speciation will
occur at higher rates and extinction at lower rates (Rosenzweig 1992); (ii) areas nearer to the
equator have higher available energy, providing more resources to therefore support more species
(Turner et al. 1996; Wright et al. 1993), and (iii) evolutionary rates may have been higher in the
tropics without interruption by glaciations and/or associated drying climates (Rohde 1992).
Clearly, factors other than latitude affect richness. For example, the geographical pattern of
tree species richness for North America shows that richness is indeed greatest in the southeastern
United States and decreases steadily northward (Figure 14.8). However, richness also decreases in
the arid parts of the continental interior and is near its lowest values in North America at the same
latitude as the southeastern United States. Tree species richness is strongly related to annual actual
evapotranspiration and mean annual temperature using large geographic quadrats (Figure 14.8,
Currie and Paquin 1987), but no significant relationship exists between actual evapotranspiration
and local tree richness of broad-leaved deciduous trees at finer scales using 0.5–10-ha plots (Latham
and Ricklefs 1993a,b). The tree richness pattern in North America is thought to reflect the evolu-
tionary history of broad-leaved trees and the relative newness of continental arctic climates com-
pared to the unglaciated areas of the southeastern United States (Latham and Ricklefs 1993a,b) as
well as local ecological interactions.
Plant diversity is strongly related to elevational gradients as well as latitudinal gradients.
Plant diversity generally decreases with elevation (Billings 1987), especially when elevation is
10
20 0
30
10
40
40
60 30
60
80
80
30 100
120
20 140
160
180
F I G U R E 1 4 . 8 Species richness of North American trees north of the Mexican border. Contours connect
points with the same approximate number of species per quadrat. Data are based on 620 native tree species
in quadrats each 2 1/2° × 2 1/2° south of 50° N, and 2 1/2° × 5° (longitude) north of 50° N. Source: Currie and
Paquin (1987) / Springer Nature. Reproduced with permission from Nature [Currie, D. J., and V. Paquin.
Large-scale biogeographical patterns of species richness.], ©1987 Macmillan Magazines Limited.
combined with increasing latitude as illustrated in Figure 14.9 for high mountain ranges in Europe
and Asia. Declining diversity is not always simple with increasing elevation; often diversity
increases from low to mid elevations, then decreases toward high elevation, although diversity at
low elevations typically exceeds that at high elevations (Grytnes and Vetaas 2002). Similar to lati-
tude, diversity gradients related to elevation are typically explained by land area (reduced land area
at higher elevations), and energy availability (peaking at mid elevations), but also by the isolation
of high-elevation areas that limits immigration of species and thus speciation.
Whittaker (1977) contrasted patterns of plant species diversity in forests of two mountain
ranges and two different regional ecosystems of the western United States (Figure 14.10). In the
relatively humid, maritime climate of the Siskiyou Mountains of southwestern Oregon, richness of
woody plant species decreased with increasing elevation (Whittaker 1960). In addition, herb diver-
sity was highest at middle elevations and lower in dense stands with a sclerophyll tree stratum at
500 F I G U R E 14.9 Relationship of the number

N 29 °N of plant species with increasing elevation in
the European Alps (left) compared with
400
high mountain ranges in central Asia (right)
Number of species
having many peaks above 6000 m. The

300 letters show the increasing number of
T 33 °N
species with decreasing latitude: A = Alps,
P = Pamirs, H = Hindu Kush, T = Tibet,
200 N = Nepal Himalaya. Source: Breckle (1974)
H 36 °N / Nordic Publications in Botany. Reprinted
100
39° with permission by the Council for Nordic
P Publications in Botany, Copenha-
gen, Denmark.
46° A
3,000 4,000 5,000 6,000
Elevation (m)
low elevations. Maximum richness was found in mesic forests at middle elevations (approximately
1200 m), with a secondary maximum at lowest elevations (Figure 14.10). At all elevations, richness
increased from xeric (ridges and south–southwest aspects) to mesic and wet ravines. By contrast,
the drier, continental climate of the Colorado Front Range of the Rocky Mountains supports more
open forests without sclerophyllous trees. Forests give way to shrubland and grassland at low ele-
vations and to alpine meadows at high elevations. Plant richness is low in forests of middle eleva-
tions and middle topographic positions of exposed, open slopes, but is higher in ravines, alpine
meadow transitions, and the more open stands of driest topographic sites and low elevations
(Figure 14.10). Again, richness increases from xeric aspects to mesic and wet ravines. These exam-
ples emphasize the importance of considering biodiversity through studies of regional ecosystems
(southwestern Oregon mountains versus Colorado Rocky Mountains) and the local ecosystems
(mesic ravines, open mid-elevation slopes, xeric ridges, subalpine woodland, etc.) that occur at
positions along gradients of temperature and moisture in the respective mountain systems.
DIVERSITY AT LOCAL SCALES

It remains difficult to provide many generalities or principles regarding biodiversity, but two
concepts seem reasonably clear: ecological heterogeneity promotes diversity, and extreme condi-
tions constrain diversity. For example, a heterogeneous spatial pattern of ecosystems or their com-
ponents (climate, physiography, soil, hydrology), disturbance regimes, and stand structures pro-
mote diversity. By contrast, the most extreme physical conditions of temperature, water, and/or
soil chemistry limit biodiversity. Above all, biodiversity is closely related to the spatial pattern,
diversity, and physical site characteristics of regional and local ecosystems. It is beyond the scope
of this chapter to examine the enormous array of factors and multiple-factor determinants of
species diversity, but we discuss selected ecological relationships in the following sections.
Physiography and Soil At regional and local scales, physiography and soil, as described
in Chapters 8 and 9, influence the distribution and diversity of plants. Physiography in
particular affects plant diversity by its effects on microclimate and soil development. Diversity
in mountainous regions in the temperate zone and in unglaciated areas is often very high
because of the dissected terrain and the resulting diverse sites and niches available for plant
occupancy. Even in less dissected terrain, microtopography provides highly localized environ-
mental heterogeneity (Chapter 8, Figure 8.16; Foster 1988a), thereby resulting in increased
(a) Siskiyou Mountains (b) Colorado Front Range
Forest Pattern and Plant Diversity on Diorite Forest Pattern and Plant Diversity
Alpine Krummholz 40.3 Alpine
21.5 21.2 22.5 20 19.2 17.0
38.3
Tsuga mertensiana L Woodlands H
37.0 ne 11,000
Subalpine Forests
nobilis Subalpi 26.
1
Abies 34.8
30 25 21.4
Bog Forests
3,200 35.1 21.3
35.0 30.4 21.3 21.5 24.4 24.8 6,000 rpa ni
1,800
lasioca elman Pinus
Abies eng
onco lor 21.5 icea flexilis
A bies c P us 10,000
3,000 - Pin L
1,600 Montane -Picea 18.6 20 30
sts
s
40.6 41.1 34.3 32.5 33.5 29.2 5,000 Abie
torta
Mesic Fore on 31.8
Wet Forests
H Forests ga 34.1 us c
1,400 u d otsu 2,800 25.3 Pin
Pse ta
9,000
c o lor - tor .9
A. c
on on 26
ens
35 sc
tsuga
Elev. ft.
Elev. ft.
Elev. m
Elev. m
32.2 3 .8
2 Pin u
34.9 34.0 36.8 24.3 24.0 4,000 2,600
1,200 34.6 rosa
Picea pung
hy
ll H a- onde
tsug - P. p
seudo
l e rop eudo u ga 38.8

s
Sc P ots 8,000
a- ud 33.0
aris - P
1,000 ug 2,400
P se
d ots 46.8
32.4 eu 29.7 23.4 20.8 3,000 60.2 45.0 ds 40.2
28.9 Ps 25.0 lan
aecyp
od
Ravine Forests
Mixed Evergreen Forests 2,200 a Wo

800
os ds
7,000
28.0 er
Cham
nd lan
L
po rub
Riparian Forests
s Sh
2,000 nu a
600 34.4 24.3 17.4 2,000 os 44.8
33.7 31.9 25.5 Pi er
n
d
44.7 po H 6,000
Sclerophyll - Pseudotsuga P.
1,800 Foothill
400
30 25 20 40 Shrubland
1,000
Ravines Sheltered Open Slopes Ridges Ravines Sheltered Open Slopes Ridges
NE W S SSW NE W S SSW
Wet, Mesic NNE N SE SW Xeric Wet, Mesic NNE N SE SW Xeric
F I G U R E 14.10 Patterns of plant species’ richness in the Siskiyou Mountains of southwestern Oregon and the eastern slope of the Front Range of the central Rocky
Mountains in Colorado. Numbers are for vascular plants species in 0.1-ha quadrats averaged for several stands representing a given combination of elevation and
physiographic position. Source: Whittaker (1977) / Plenum Publishing Corp. Reprinted from Environmental Biology with permission of the Plenum Publishing Corp.
species diversity. Regeneration of some understory tolerant species depends on microtopogra-

phy formed by treefall or coarse woody debris mounds. For example, yellow birch and eastern
hemlock establish preferentially on moss-covered logs and mounds in hemlock–northern hard-
wood forests (Chapter 4), as do western hemlock and Sitka spruce in old-growth forests of the
Pacific Northwest (Harmon 1987; Harmon and Franklin 1989).
Soil-water, nutrients, and oxygen availability also are responsible for many observed differ-
ences in diversity. Soils extreme in one or more attributes—low water or nutrient availability, high
specific mineral content (e.g., soil derived from serpentine parent material), or very shallow—
often support forests with low tree diversity (e.g., pine barrens, black spruce swamps). Notably,
such soils may support many herbs or rare and unusual herbs that have adaptations making them
competitive on such sites. Conversely, soils high in nutrient availability may support high tree
species richness. For example, vascular plant richness was strongly related to soil nutrient avail-
ability for a broad range of ecosystems in the Piedmont Province of North Carolina (Figure 14.11).
High tree species richness was also associated with abundant moisture, and moister sites had more
species where soil fertility was low.
Deep, moist, or high-nutrient soils may support either high-or low-diversity forests depend-
ing on history, amount of disturbance, and successional stage. Dominance by one or a few species
often occurs more rapidly under favorable site conditions, where all organisms can potentially
grow rapidly (Huston 1993, 1994). Under such conditions, the best competitors are able to elimi-
nate most other species by monopolizing an essential resource such as light. Under poorer soil-
water and nutrient conditions, superior competitors are unable to dominate as rapidly, and more
species are likely to coexist, leading to smaller and/or fewer individuals. For example, canopy cov-
erage in late-successional stands may be very dense in mesic beech–sugar maple and hemlock–
northern hardwood forests of the western Great Lakes region. Few species survive in the under-
story and ground-cover layers of these forests other than young individuals of the canopy species.
Ecosystems on the fertile and moist moraine referenced in Figure 14.6, for example, exhibit
relatively low ground-cover diversity (Table 14.4, Colonial Point moraine). However, as discussed
in the earlier text, a bigtooth aspen community on a similarly fertile and mesic soil (Table 14.4,
Interlobate moraine) has approximately 40% greater ground-cover richness due to favorable light
F I G U R E 14. 11 Relationship between species

Mesic Eutrophic richness (average number of plant species in a
Alluvial
0.1-ha sample) and average exchangeable
cation (Ca, Mg, K) for a diverse range of
Species Richness (Species in 0.1 ha)
70
ecosystem types, representing wet, mesic, and
dry sites of the North Carolina Piedmont.
Montmorillonite Source: Peet and Christensen (1980b) /
60 Eidgenössische Technische Hochschule Zürich.
Swamp Dry Reprinted with permission of the ETH Zürich,
Warm Eutrophic Finanzdienste, CH8092 Zürich, Switzerland.
Monadnock
50
Mesic Mesotrophic
Dry Mesotrophic
40 Dry Oligotrophic
Cool Monadnock
1 2 3 4 5 6 7
Soli Cation Concentration (meq/100 g)
conditions resulting from the aspens’ relatively open crowns. Ground-cover diversity observed in
summer in ecosystems with a closed canopy may be somewhat misleading because of a relatively
diverse flora of early spring ephemerals that complete much of their aboveground growth before
the canopy trees leaf out in the spring.
Community Composition and Structure The diversity of birds, mammals, and other
forest animals may be affected by: (i) forest species composition, and (ii) structure/age differences
within and among communities, including vertical layering of vegetation, horizontal arrangement
of trees and other vegetation within and among stands (e.g., random versus patchy stem distribu-
tion), and the distribution of standing and downed dead trees. A classic case is the relationship of
bird species richness and structural complexity of forest vegetation. Bird species diversity has been
shown to be positively related to foliage height diversity in northwestern North America (MacAr-
thur and MacArthur 1961), but negatively related to it in northern Patagonia, Argentina (Ralph 1985;
Huston 1994, p. 41). In the Pacific Northwest, the northern spotted owl inhabits old-growth forests,
needing both large hollow trees for nesting and open subcanopy areas for foraging. Controversy
over the spotted owl was a watershed event of American environmental policy in the 1980s and
1990s (Yaffee 1994). The controversy involved conserving structurally heterogeneous old-growth
forests, which are lost following logging, to retain habitat of the owl and many other organisms of
these unique forests (Franklin et al. 1981; Franklin et al. 1997). By contrast, the rare red-cockaded
woodpecker of the southeastern United States breeds in old-growth longleaf, slash, or loblolly pine
communities with low structural diversity. This bird nests in colonies in cavities carved over gener-
ations in large, living, fungal-infected, resin-rich pines (Lennartz 1988; McFarlane 1992). The birds
preferentially reside in ecosystems with large trees and having little or no mid-story, few tall shrubs,
and generally sparse ground cover. The woodpecker’s preference for simplified vertical structure is
probably related to protection from predators, hot fires, and availability of insect prey.
Disturbance and Succession It is difficult to provide generalized textbook examples

about the effect of disturbance on species diversity given the wide range of ecosystem variation in
composition, structure, and function. The concept most often cited relating to the association bet-
ween diversity and disturbance is the intermediate disturbance hypothesis (Connell 1978).
This hypothesis suggests that species richness is low at very high disturbance frequencies because
only the most resistant species are able to survive, but also at very low disturbance frequencies
because the most competitive species exclude other species. Richness is thereby theorized to be
maximized at intermediate frequencies (Figure 14.12). The intermediate disturbance hypothesis
has received mixed support based on empirical studies (Bendix et al. 2017; Huston 2014; Svensson
et al. 2012, Mackey and Currie 2001), and the general validity of the model is still debated. Recent
work suggests that disturbance intensity likely interacts with disturbance frequency to determine
species diversity (Yeboah and Chen 2016; Miller et al. 2012; Hall et al. 2012).
Disturbance is ecosystem-specific and thus maintains a combination of trees, ground cover,
and animals consistent with the geographic setting of regional and local ecosystems (Chapter 16).
Disturbance will maintain either high or low species diversity depending on ecosystem type, and
thus patterns of diversity in hurricane-prone ecosystems will differ from those in fire- or flood-
prone ecosystems. For example, frequent fires prior to European colonization maintained jack
pine as the only dominant species on some dry, outwash-plain ecosystems in the Lake States, but
tree species richness sharply increases on adjacent mesic moraines where fires are virtually absent.
Simultaneously, frequent fire in outwash-plain ecosystems maintains high herbaceous species
richness by sustaining high light conditions via a lack of jack pine canopy closure.
Fire is the disturbance most commonly studied in North America for its effects on diversity
(Thom and Seidl 2016). A universal example of fire disturbance reducing temporal species richness
is observed in ecosystems where mesophytic, fire-sensitive species colonize the site when fire is
DIVERSITY HIGH
LOW
DISTURBANCES FREQUENT INFREQUENT
DISTURBANCES SMALL LARGE
SOON AFTER DISTURBANCE LONG AFTER
B
A C CANOPY
b c c c
b c UNDER-
STORY
A COLONIZING B MIXED C CLIMAX
F I G U R E 1 4 .1 2 General relationship between tree species diversity and disturbance related to frequency of
disturbance, size of disturbed area, and temporal scale corresponding to early-(colonizing), mid-(mixed),
and late-successional stages of succession in the wet tropical Budongo rain forest of Uganda. The diagram
represents the “intermediate disturbance hypothesis” where a mixture of early-and late-successional tree
species exhibits the greatest tree-species diversity at an intermediate position in disturbance frequency and
size and time after disturbance. Source: Connell (1978) / with permission of American Association for the
Advancement of Science. Reprinted with permission from Science 199, p. 1303, © 1989 by the American
Association for the Advancement of Science.
absent, increasing richness, but richness then decreases at the occurrence of the next fire. For
example, dry-mesic oak forests of the Lake States and northeastern United States were character-
ized by frequent, light surface fires prior to European colonization. Shade-tolerant, mesophytic
species that were able to invade during the fire-free interval were eliminated before they were re-
cruited to the canopy. Fire exclusion in these forests following European colonization has today
removed the disturbance necessary to eliminate mesophytic invaders such that open-canopy, oak-
dominated forests are currently more species-rich. In the further absence of fire, however, oak
forests will eventually be replaced by low-diversity mesic forests with dense canopies and a low-
diversity, shade-tolerant understory as the intermediate-tolerant oak species are outcompeted
(Nowacki and Abrams 2008).
Over longer temporal scales, however, frequent, less-severe disturbances are often necessary
for the maintenance of diversity in ecosystems characterized by periodic fire. For example, lack of
periodic fires in prairie or savanna ecosystems allows the encroachment of woody plants (Cur-
tis 1959) and creates a reduction in herbaceous diversity. In an attempt to restore pre-European
colonization of northern pin oak savanna structure and composition in central Minnesota, annual
prescribed burning was begun in 1964 (White 1983, 1986). After 13 years of burning, oak basal area
was significantly reduced, and shrub cover eliminated in burned versus unburned plots. The rich-
ness of herbs subsequently increased in burned plots by about 60%.
Disturbance also affects species diversity in successional sequences. Total plant species
diversity is associated with the four general stages of secondary succession following disturbance
(Chapter 17; Figure 17.4): (i) stand initiation (establishment), (ii) stem exclusion (self-thinning),
Focal Species In Conserving Diversity 351
Number of species
max.
Animals
Plants
Fungi
min.
Stand Stem Understory Old-growth
initiation exclusion re-initiation
F I G U R E 14 .1 3 Predicted number of species along forest succession for animals, plants, and fungi for beech
forests of central Europe. Successional stages along the x-axis approximate those described in the text and
are not to scale. Source: Hilmers et al. (2018) / John Wiley & Sons.
(iii), understory re-initiation (transition), and (iv) old growth. Vascular plant diversity may be char-
acterized for the respective stages as follows: (i) diversity is high in an open site soon after distur-
bance as species resprout or colonize by seeding; (ii) diversity is low once canopy closure occurs
and the overstory “shades out” all but the most understory-tolerant species; (iii) diversity is low but
increasing once gaps begin to appear in the overstory; and (iv) diversity ranges from low to high
(depending on canopy density and number of gaps) as minor and major disturbances begin to
grossly change the structure of the declining old-growth forest. This general pattern of diversity
with successional stage has been demonstrated for Rocky Mountain forests (Peet 1978) and old-
field mixed conifer–hardwood forests of the North Carolina Piedmont (Christensen and Peet 1981;
Peet and Christensen 1987). In a study of European beech forests of central Europe, Hilmers et al.
(2018) confirmed this “U-shaped” response of species richness for plants and animals over the
course of forest succession, but an increasing trend of fungi that reached its maximum during
understory re-initiation (Figure 14.13). Species assemblages were most similar to each other
among early- and late-successional stages rather than in mid-succession. For plants, changes in
species richness over forest succession were strongly driven by changes in abundance, that is, more
individual stems—and thus more species—were present early and late in succession due to lower
canopy cover. Obviously, disturbances occurring during succession would affect diversity in differ-
ent ways depending on ecosystem type and disturbance characteristics—kind (wind, fire, insect,
etc.), origin (natural or human), and severity.
FOCAL SPECIES IN CONSERVING DIVERSITY

Ecologists, land managers, and the public have increasingly focused on maintaining the local,
regional, and global diversity of species, resulting in the emergence of an entire science of
conservation biology within the last half-century. Maintaining diversity in forests requires under-
standing the status of all species that occur there, which is rarely realistic given available resources.
Focal species, or those whose measurement provides information beyond that about its own
status, are therefore often used in efforts to conserve diversity. There exist many types of focal
species used for this purpose which may be distilled to five categories (Miller et al. 1999; Day-
ton 1972): (i) keystone species, or those with an ecological effect disproportionate to their abun-
dance; (ii) umbrella species, or those requiring large areas whose protection would encompasses
many other species; (iii) flagship species, or those with broad public appeal whose plight is likely to
stimulate public support for conservation efforts; (iv) indicator species, or those associated with
particular environmental conditions, and (v) foundation species, or those dominant in an ecosys-
tem that influence structure and function through their abundance. We have treated indicator
species in detail in Chapter 11; in the following sections, we discuss foundation and keystone
species for their relevance to forests.
FOUNDATION SPECIES
Ecologists recognize that the most abundant species play major roles in controlling the rates and
directions of many ecosystem processes. These species, termed foundation species (Dayton 1972),
dominate an ecosystem in abundance and influence, and are often crucial for maintaining com-
munities because they typically provide the major energy flow and the physical structure that sup-
ports and shelters other organisms (Gentry and Dodson 1987; Ashton 1992). Foundation species
are typically plants and are usually trees in forested ecosystems. As we emphasize repeatedly
throughout this book, trees in a forest shape its structure, microclimate, and light availability as
well as its biomass and carbon dynamics and biogeochemistry. The potential loss of foundation
tree species therefore will strongly affect the local environment on which a variety of other species
depend (Ellison et al. 2005). The loss of foundation tree species is increasing across the world,
which will inevitably influence diversity.
Several examples of the effects of foundation species loss on biodiversity have been docu-
mented and summarized by Ellison et al. (2005). For example, eastern hemlock grows in pure
stands in the eastern United States, where its dense canopy intercepts light and precipitation, and
its acidic, slowly decomposing litter slows nutrient cycling and creates infertile soils. Plant species
richness in hemlock-dominated ecosystems is therefore typically very low. Hemlock is in rapid
decline due to the hemlock woolly adelgid, an invasive insect that often results in the replacement
of hemlock by deciduous species. Loss of hemlock may result in increases in plant species richness
at local scales (e.g., Martin and Goebel 2013), but it is also likely to reduce broadscale variability in
species richness (Ellison et al. 2016). Perhaps most importantly, hemlock in the southern
Appalachian Mountains is notable for providing a shaded, cool environment along streams, such
that its loss may impact the diversity of aquatic communities (Adkins and Rieske 2015; Webster
et al. 2012).
Foundation species in forests are typically considered to be the dominant species, but recent
research suggests that more than one species may be foundational in an ecosystem at some scale
(Angelini et al. 2011). Specifically, Thomsen et al. (2018) reviewed 140 published studies examin-
ing secondary foundation species, or those species provided habitat by primary foundation
species which themselves provide structurally complex habitat that alters environmental condi-
tions for additional species in the community. They found that the presence of secondary
foundation species increased species abundance and richness in many different ecosystems. For
example, a study of Spanish moss, a flowering plant epiphytic on live oak in southeastern Georgia
showed that oaks (primary foundation species) improved epiphyte survival by shading and
reducing temperature on the Spanish moss (secondary foundation species). In turn, Spanish moss
reduced drying effects and predation on insects, supporting communities that were larger and
more species-rich than those found on live oaks without Spanish moss (Angelini and Silli-
man 2014). Studies such as these show that the biotic drivers of diversity in forests involve much
more than just the dominant species, and that efforts to conserve diversity is therefore a very com-
plex endeavor.
KEYSTONE SPECIES
In contrast to foundation species, a keystone species is one whose impact on its community
or ecosystem is disproportionately large relative to its abundance (Power et al. 1996; Paine 1969).
The concept of a keystone species is based on less abundant species that have major community
and ecosystem effects, rather than those species that have high influence because they are very
abundant or dominant. Identifying influential but less-obvious species is important for under-
standing how losing species will affect ecosystem function. Keystone species typically have
high trophic status, and can exert effects through consumption (e.g., pocket gophers, elephants,
Focal Species In Conserving Diversity 353
wolves, trout, kangaroo rats, etc.) as well as by interactions and processes such as competition,
mutualism, dispersal, pollination, disease, and by modifying abiotic factors (Bond 1993; Mills
et al. 1993). One example of a keystone species is the badger, whose mounds maintain diversity
in prairie floras (Platt 1975). Relatively few woody plants are identified as keystone species; fig
trees that provide food for animals were the only woodland keystone species identified by
Power et al. (1996).
The keystone species concept is helpful in identifying the most suitable areas for biodiver-
sity preserves, understanding complex linkages among ecosystem biota and site–biota interac-
tions, and in managing for single species. Power et al. (1996) gave three useful insights from the
keystone species concept: (i) land managers should carefully consider the consequences of the
loss of species for which no obvious role in ecosystems has been discovered, (ii) introduced alien
species may, like keystone species, have potential strong effects disproportionate to their bio-
mass, and (iii) there is no well-developed protocol of identifying keystone species; the field is
littered with far too many untested anecdotal “keystone species.”
ENDEMICS AND RARE AND ENDANGERED SPECIES

Much of the concern about diversity involves species considered rare, in danger of extinction, or
endemic. Very often such species serve as flagship species for conserving diversity. Endemic
species are those whose ranges are extremely limited in size, such as those restricted to an ecolog-
ical region, a single mountain range, serpentine soils or other spatially limited sites, or to politically
bounded units such as states or provinces. Endemics occur in both diverse and non-diverse areas,
but they often contribute substantially to regional diversity.
Endemism is not uniform across North America. Areas of high endemism are usually corre-
lated with unusual environmental conditions, relative isolation of an area, or a history of changing
environmental conditions that have limited some species to a restricted area (Gaston and Spic-
er 2014). There are an estimated 186 endemic species of trees in the United States, representing
about 29% of tree species richness (Jenkins et al. 2015; Figure 14.14). The highest numbers of
endemic species occur in California, Texas, and Florida (Gentry 1986), many of which (though not
all) occur in forest ecosystems. The highest concentration of endemic trees is found in the south-
eastern United States. Most of Florida’s endemic plants are associated with swamps, flatwoods,
and hardwood forests in northern and central Florida. A large number of Floridian species are
Richness Endemic
641 species 186 species
138 74
1 1
F I G U R E 14 .1 4 Tree species richness and endemic tree species of the lower continental United States. Total
richness is the number of all tree species; endemics are species whose entire range is within the lower 48
states. Source: Jenkins et al. (2015).
associated with a small area known as “scrub” (Myers 1990). Many plants with very limited
distributions apparently arose on isolated islands when most of the current peninsula was below
sea level and thus persist today in isolated scrub “islands.” Animals also evolved on these islands,
including the Florida scrub jay and the sand skink, the latter of which is a lizard that “swims” just
below the sand surface. These species are well adapted to survive on very specific sites and on an
archipelago of scrub islands, but they are regionally rare, and many are considered to be threat-
ened or endangered due to encroachment of agriculture and urbanization into the ecosystems that
support them.
DIVERSITY AND THE FUNCTIONING OF ECOSYSTEMS

Ecologists have begun to uncover important roles for diversity in the way ecosystems function.
That is, biodiversity has become increasingly recognized not only as a description of the biota in
ecosystems, but as an important driver of the way ecosystems work. Biodiversity has been shown
to be strongly correlated with several important attributes of ecosystem function (Lefcheck
et al. 2015; Naeem et al. 2012; Cardinale et al. 2006; Chapin et al. 2000), although causal relation-
ships remain the subject of much debate. In the following text, we briefly discuss two of these
relationships with reference to forests: the relationship between biodiversity and productivity, and
the importance of biodiversity for ecosystem stability.
BIODIVERSITY–PRODUCTIVITY RELATIONSHIP
Species richness tends to be correlated with primary productivity (Oehri et al. 2017; Balvanera
et al. 2006; Waide et al. 1999). Most of the data supporting this relationship have surfaced from
grasslands and other herbaceous communities rather than forests, but the relationship is most often
described as positive but decelerating as species richness increases (Figure 14.15a; Liang et al. 2016b;
Zhang et al. 2012). Notably, the form of the relationship may change as a result of many different
factors, including the method of characterizing productivity (Sheil and Bongers 2020; Groner and
Novoplansky 2003), the type of organisms studied (Mittelbach et al. 2001), the scale of the analysis
(Mittelbach et al. 2001; Waide et al. 1999), and many other factors. For example, in a study of more
than 115 000 forest plots across the United States, Fei et al. (2018) found an important effect of cli-
mate on the biodiversity–productivity relationship, with a hump-shaped relationship characterizing
mesic climates and a linearly positive or non-significant relationship in dry climates. Meta-analyses
have revealed strong biodiversity effects on productivity that may be as large as effects of other
factors such as drought, fire, or herbivory (Tilman et al. 2012; Hooper et al. 2012).
The mechanisms driving the relationship between biodiversity and productivity are well
studied but remain unclear. Both variables depend on climate and physical site factors, and there-
fore are likely to be positively correlated; that is, physical site characteristics tend to determine the
growth (productivity) of organisms and the number of species populations (Allen et al. 2002). The
positive effect of biodiversity on productivity is generally explained by two main mechanisms.
First, communities with higher species richness typically have species that are able to differentiate
their niches (Chapter 13) in a way that they become complementary to one another. This comple-
mentarity may provide for more efficient use of resources and nutrient retention, reduced com-
petition (Hooper et al. 2005), and relief from herbivory and pathogens (Civitello et al. 2015) that
may lead to increased productivity. Similarly, the positive effects of one species on another (facili-
tation; see Chapter 17) may increase the productivity of a community as a whole, and thus a
diverse community is more likely to be more productive because it contains species that are able to
facilitate others (McIntire and Fajardo 2014; Hooper et al. 2005). Second, more diverse commu-
nities are more likely to contain dominant species with high productivity—often called the “sam-
pling effect” (Hooper et al. 2005). Complementarity has been shown to be less prevalent in
Diversity and the Functioning of Ecosystems 355
F I G U R E 14 .1 5 Relationship (a)
between tree species richness and
productivity. (a) Effect of tree species 12
Productivity (m3ha–1yr–1)
richness on productivity using a
global dataset from 44 countries.
Productivity increases at a decreas- 8
ing rate with species richness.
Source: After Liang et al. (2016b) /
with permission of American 4
Association for the Advancement of
Science. (b) Effect of productivity on
tree species richness in the north-
0
western United States, using gross 0 20 40 60 80 100
photosynthesis per square meter as a
Tree species richness (%)
proxy for productivity. Species
richness peaks at intermediate levels (b)
of productivity. Source: Swenson and 25 Douglas fir absence
Waring (2006) / John Wiley & Sons.
Douglas fir presence
20
Tree species richness
15
10
0
0 500 1000 1500 2000
Pg, g C m–2
temperate forests, where species interactions tend to lead to competitive exclusion, compared to
the harsher environments of boreal forests where facilitation among species may be more impor-
tant (Paquette and Messier 2011).
A great deal of research has also examined how the biodiversity–productivity relationship
may also work in the opposite direction; that is, whether productivity is an important driver of
species richness. This relationship is commonly described as hump-shaped or unimodal, with
richness peaking at intermediate levels of productivity (Figure 14.15b; Liang et al. 2016a; Waide
et al. 1999). Competitive exclusion theory would explain this relationship as stress and a lack of
resources dominating at low productivity, supporting few species, and competition dominating at
high productivity, limiting species richness (Tilman 1982; Huston 1979; Grime 1979). High pro-
ductivity could also simply translate into more individuals, which eventually translates into more
species (Evans et al. 2005). Excellent reviews of this direction of the biodiversity–productivity rela-
tionship are given by Grace et al. (2016), Gillman and Wright (2006), and Loreau et al. (2001).
The importance of the biodiversity–productivity relationship for forests is not simply academic.
If forest productivity truly is driven by biodiversity, then the negative effect of biodiversity loss could
have serious economic implications for global forest resources. Moreover, decreases in productivity
would reduce the ability of forests to absorb and sequester carbon, likely resulting in consequential
effects on the global carbon cycle. The biodiversity–productivity relationship further emphasizes the
“value” of biodiversity discussed at the beginning of this chapter; the economic value of biodiversity
in maintaining commercial forest productivity across the globe is estimated to be as much as $490
billion per year (Liang et al. 2016b). It also suggests that biodiversity plays a key role in how ecosys-
tems function.
THE ROLE OF BIODIVERSITY IN ECOSYSTEM STABILITY

Ecosystems with communities higher in species richness are thought to be more stable. There are
many ways to define ecosystem stability (Ives and Carpenter 2007), but perhaps the simplest defi-
nition of a stable system is one changing relatively little from its average state, even in the face of
environmental fluctuations (also called ecosystem resistance; Chapter 16). Stability may be consid-
ered from the perspective of species or from the perspective of ecosystems, and this frame of refer-
ence has implications for how the biodiversity–stability relationship is interpreted. From a species
perspective, higher species richness may destabilize ecosystems because competition for resources
among species will limit the number of individuals in a system, reducing species’ population sizes
and increasing the likelihood that some species will go extinct (May, 1974). From an ecosystem
perspective, higher species richness creates a higher likelihood that an increased abundance of one
species can compensate for the decreased abundance of another if the species vary in how they
respond to environmental fluctuations—a process known as compensatory dynamics. Because
both types of relationships are present in natural systems, diversity may stabilize ecosystems while
simultaneously destabilizing the abundances of individual species (Hector et al. 2010; Tilman
et al. 2006; Tilman 1996). As with the biodiversity–productivity hypothesis, the relationship bet-
ween biodiversity and stability emerged from studies of grasslands, and fewer studies of the rela-
tionship have been conducted in forests.
Most recent attention to the biodiversity–stability relationship examines the apparent positive
influence of biodiversity on ecosystem-level functions such as resource capture and biomass produc-
tion (Cardinale et al. 2011). Indeed, a large body of research suggests that such functional attributes
are more stable in diverse communities (Hector et al. 2010; Jiang and Pu 2009; Cottingham et al. 2001).
Loreau and De Mazancourt (2013) synthesize the mechanisms driving the relationship between diver-
sity and ecosystem stability as follows:
1. Differences in species’ fundamental niches drive their differential preference of environ-

mental factors such as temperature, rainfall, and resources. As a result, species’ populations
respond differently to environmental fluctuations, creating a more stable ecosystem as a
whole. Differing environmental responses among species is always stabilizing and is
thought to be a major factor causing compensatory dynamics between species.
2. Species populations differ in their intrinsic rates of natural increase, such that they
respond at different speeds to environmental changes. Similar to the first mechanism
mentioned in the earlier text, the differences in response time tend to generate asynchro-
nous population dynamics and thereby promote ecosystem stability. This mechanism is
also compensatory.
3. Reducing the strength of competition increases stability through the overyielding effect, where
mean total biomass of a diverse community increases compared to a non-diverse community
or monoculture. Increased abundance or biomass reduces the strength of demographic
stochasticity in population dynamics, which again contributes to ecosystem stability.
As described in the earlier text, the biodiversity–stability relationship is well studied but not
as well understood, particularly for forests. Although the literature continues to evolve rapidly,
thorough reviews and syntheses of the relationship are provided by Loreau and De Mazancourt
(2013), De Mazancourt et al. (2013), Ives and Carpenter (2007), and McCann (2000), and Mori
et al. (2017) review the concept relative to applied research in forest ecosystems. An important
Forest Management and Diversity 357
point is that a stabilizing effect of biodiversity on ecosystem functions would imply yet another
implicit value of biodiversity that would clearly justify its conservation.
FOREST MANAGEMENT AND DIVERSITY

Silviculture and forest management include many operations, such as regenerating and tending
forests, that are very different from the ecological processes occurring in natural stands. Creation
of even-aged, even-sized stands using clear-cutting and artificial reforestation (Oliver and Lar-
son 1996; Smith et al. 1997) has been the most common management system for many decades.
This system creates highly simplified stands that lack many aspects of community structure and
complexity, such as snags, coarse woody debris, multiple canopy layers, irregular tree spacing, and
gaps (Franklin 1995). Forest ecosystems may be simplified at the genetic, stand, and landscape
levels, as well as temporally (succession). The economic criteria of efficient forest management
(e.g., whole-tree harvesting) and high productivity of commercially important tree species have
largely driven forest simplification because it was held that the structural complexity found in
natural stands was not essential for sustained tree productivity of a site. Research on the impor-
tance of stand structure and biological legacies as they affect ecosystem function, biodiversity, and
long-term site productivity has been conducted by Franklin (1995, 1997) and Franklin et al.
(1997, 2018). In the following sections, we briefly discuss the impacts of traditional forest
management techniques on diversity and present alternative silvicultural approaches designed
specifically to accommodate diversity as much as possible.
EFFECTS OF TRADITIONAL FOREST MANAGEMENT ON DIVERSITY

Clear-cutting is the forest management method most extensively studied because of the enormous
changes it brings to ecosystems and their biota (Boyle 1992). Despite common assumptions to the
contrary, the effects of clear-cutting on diversity are not always clear-cut (so to speak). Diversity of
animals and plants may increase or decrease depending on the specific sites, scales, and patterns
of clear-cuts, as well as the methods employed, including post-harvest burning or other site prepa-
ration. The effects of clear-cutting typically affect species differently, and therefore do not usually
create a simple reduction in species diversity. Species dependent on old-growth or near-natural
forests are often lost, presumably until suitable environmental conditions are restored, whereas
colonizing species or those dependent on disturbance may thrive.
In a study of natural and managed Douglas-fir forests of the Pacific Northwest, Halpern
and Spies (1995) found that changes in understory diversity were fairly short-lived following
clear-cutting and slash burning. Populations of most vascular plants recovered quickly to original
levels prior to canopy closure, although diversity may remain depressed for decades on severely
burned sites, and some species may experience local extinction (Halpern and Spies 1995). Nota-
bly, the effects of clear-cutting may vary by ecosystem, even within the same region. For example,
plant species richness did not differ between cut and uncut plots 50 years after clear-cutting in a
sugar-maple-dominated ecosystem in Upper Michigan (Albert and Barnes 1987), but clear-
cutting an eastern-hemlock-dominated ecosystem resulted in higher richness in the cut plots
after 50 years (Hix and Barnes 1984). The differences in the hemlock system reflect the sudden
change from a low-moisture, acidic, nutrient-poor site beneath a dense hemlock canopy to an
open, less acidic, more nutrient-rich environment facilitated by regeneration of hardwoods on
the logged site (Albert and Barnes 1984). In any case, clear-cutting obviously creates the poten-
tial for strong effects on species richness, and thus silvicultural prescriptions that maintain
diversity of stand structure and site conditions will be most effective in maintaining diversity.
Likewise, any management application that precludes or delays the re-establishment of
old-growth attributes (such as intensive, short-rotation plantation forestry) may result in long-
term loss of diversity.
Diversity in forests is also affected by patterns of forest management at broad spatial scales.
Fragmentation of forests—or the breakup of forests into progressively smaller and isolated
patches—by roads, agricultural fields, and clear-cut patches (Chapter 22) also influences diversity.
The consequences of ecosystem fragmentation are not only the creation of “islands” of various
sizes but also include large changes in the physical environment that markedly influence species
and gene pool diversity. In many ways, the effects of fragmentation resemble those of clear-cutting
or other activities that reduce or eliminate forest interior conditions. The significance of
fragmentation and edge effects has been considered by many (With 2019; Fahrig 2003; Harris 1984;
Saunders et al. 1991, Ledig 1992), and we examine it in detail in Chapter 22.
Our main focus in this chapter has been on ecosystem and species diversity, but genetic
diversity of forests (i.e., species diversity and gene diversity within species) has also been greatly
influenced by human activities. Forest management may affect genetic diversity via its manipula-
tion of the evolutionary processes of extinction, selection, drift, gene flow, and mutation
(Ledig 1992). Clear-cutting, for example, decreases population size and connectivity and thus
increases differentiation, genetic drift, and inbreeding in remaining adult trees, but not necessarily
in regenerating trees (El-Kassaby et al. 2003). By contrast, management that retains some degree
of stand structure and regeneration is less likely to affect genetic diversity of adult trees but could
promote inbreeding and genetic drift of the regeneration (Sagnard et al. 2011). Because managers
select the trees for harvesting and for regeneration, diversity may be increased (e.g., exposing reces-
sive genes and increasing the occurrence of novelties and mutants), but these practices can also
reduce diversity. Forest harvest practices that cause fragmentation, alter population sizes, or sim-
plify age structure all affect the breeding systems of forest trees. Ledig (1992) describes how habitat
alteration, environmental deterioration, and domestication of forest species may all lead to reduced
biodiversity, and Ratnam et al. (2014) review the effects of forest management on genetic diversity
in forests across the world.
PRESERVING DIVERSITY IN MANAGED FORESTS

The most common method of diversity preservation is the establishment of preserves—an area of
land set aside to protect the species and ecosystems within it. The concept of biodiversity preserves
has been heavily influenced by island biogeography theory of MacArthur and Wilson (1967;
Chapter 22), whereby preserves are viewed as islands of suitable habitat within an ocean of unsuit-
able habitat. Based on island biogeography theory, preserves function better to maintain diversity
when they are large such that core habitat remains; smaller preserves work better if they are
connected. A great deal of research has been dedicated to optimal preserve design, much of which
we discuss in Chapter 22.
The success of a preserve system for diversity protection is usually assessed based upon the
proportion of area set aside. Worldwide, protected areas have higher abundances of individual
species and higher species richness compared with areas having alternative land uses (Coetzee
et al. 2014). In 2020, the International Union for Conservation of Nature (IUCN) estimated that
17% of the Earth’s land area was set aside as preserves based on their own set of guidelines (UNEP-
WCMC, IUCN, NGS 2020). For comparison, about 8% of the lower 48 states in the United States
was within an IUCN categorized protected area as of 2015, mostly in the western half of the
country where a large proportion of land is in public ownership (Jenkins et al. 2015). Relatively
few preserves are available in the southeastern United States, where species richness and ende-
mism are the highest in the country (Figure 4.14), and most land is in private ownership. These
trends emphasize the limitations of preserves in conserving diversity: preserve design is often
subject to political and economic pressures that minimize their effectiveness. Many researchers
argue that the solution is to either simply increase the amount of area in preserves or set aside
higher-quality areas into preserves, but these strategies are often not practical.
It became clear almost three decades ago that a preserve system alone is not sufficient to
effectively sustain forest diversity (Hansen et al. 1991; Marcot 1997; Westman 1990; Wilcove 1989).
The world’s forested lands are dominated by semi-natural ecosystems and cut-over land such that
conservation of biological and ecosystem diversity rests primarily on innovative management of
the land outside of protected areas. To a large degree, the ecosystem services described in the ear-
lier text and in Chapter 23 are provided by the biological and ecosystem diversity found on all
landscapes regardless of their history of human activities, such that conserving diversity is a
fundamental need in all forests (Lindenmayer and Franklin 2002). As such, a variety of silvicul-
tural procedures have been developed for the conservation of biodiversity on managed lands. A
detailed discussion of forest management techniques is beyond our scope, but we review the basics
of forest management that incorporate diversity in the following sections.
ECOLOGICAL FORESTRY: INCORPORATING BIODIVERSITY

INTO FOREST MANAGEMENT
Traditional silvicultural harvest and regeneration methods—clear-cut, seed-tree, and shelterwood
in even-age management systems and selection used in uneven-aged systems (Smith et al. 1997)—
were created solely for regeneration and subsequent growth of commercially important tree species.
These traditional methods have typically been applied without concern for diversity and usually
lack the structural complexity of natural stands. Lindenmayer and Franklin (2002) summarize that
sustaining diversity in managed forests requires shifting the focus from what is extracted from a
forest, such as timber or other resources, to what remains in place after the resources are harvested.
As biodiversity depends on factors such as structural complexity, special habitats or niches that
might result from that complexity, keystone and foundation species, and biological legacies, retain-
ing these factors in a post-harvest forest is critical for maintaining diversity. Hunter and Schmie-
gelow (2011) simplify this process into a “micro approach” that provides structures and complexity
necessary for habitat (e.g., snags or downed logs, or microclimatic conditions) within forest stands,
and a “macro approach” that provides a diversity of stand conditions on the landscape (e.g., mix-
tures of open-and closed-canopy forests, or forests at different successional stages). This multi-scale
process may be relatively easily incorporated into modern silvicultural systems that manage forests
for multiple values, including wood production as well as diversity (Franklin et al. 2018).
Broadly speaking, managing forests from the perspective of conserving biodiversity and eco-
logical productivity as guided by natural disturbance regimes is known as ecological forestry
(Palik et al. 2021; Franklin et al. 2018; Seymour and Hunter 1999). General components of ecolog-
ical forestry include retaining portions or features of the pre-harvest stand in the managed stand,
using knowledge of forest succession to manipulate established stands, using whatever silvicul-
tural return intervals are necessary to achieve desired structure or ecological processes, and using
knowledge of large-scale processes to plan management on landscapes (Palik et al. 2021). Com-
pared to traditional forestry focused on timber production, ecological forestry maintains a diversity
of structures and processes, values complexity and heterogeneity, and emphasizes ecosystem
diversity. By contrast, production forestry utilizes only the subset of structures and processes that
support economic goals, values simplicity and homogeneity, and emphasizes optimizing tree
growth (Franklin et al. 2018).
The structural complexity of old-growth Douglas-fir–western hemlock forests of the western
Cascades that range in age from 250 to over 600 years has been studied in detail (Franklin et al. 1981;
Harmon et al. 1986; Franklin et al. 1987; Maser et al. 1988; Spies and Franklin 1991; Franklin 1993;
Spies 1997). These forests provide an effective model from which the principles of ecological for-
estry have been developed. Aspects of structural complexity here include varied sizes and condi-
tions of live trees, multiple canopy layers, and presence of canopy gaps, as well as snags and
downed logs and woody debris. These structural features markedly influence biotic composition of
the forest, ecosystem processes such as biomass production, nutrient cycling, and succession, and
regeneration following natural disturbances. In particular, biological legacies—surviving organ-
isms and organically derived structures such as snags, logs, and organic soil horizons (Chapter 17)—
are essential for the rapid reestablishment of forests that have highly diverse composition, function,
and stand structure (Franklin and MacMahon 2000; Franklin et al. 2000). Creating or maintaining
structural complexity in managed stands in this region is an effective example of managing forests
for multiple and complex objectives, including both wood production and biodiversity (Linden-
mayer and Franklin 2002; Franklin et al. 1997). Palik et al. (2021) and Franklin et al. (2018) provide
excellent overviews of the variety of methodologies in ecological forestry; we provide an example
of one such method in the following text.
Variable-Retention Harvest System The variable-retention harvest system attempts to retain

elements of the rich structural, compositional, and functional diversity in managed stands that are
characteristic of old-growth forests. Variable-retention harvest prescriptions focus upon providing
the structures and complexity necessary for habitat within stands as well as enhancing connectiv-
ity in the managed landscape (Franklin et al. 1997). The term “variable retention” is used because
the structures retained may vary widely in amount (percent of area covered), kind, and spatial
pattern (Figure 14.16). Structural retention focuses largely on biological legacies—both living and
dead organic structures, including living trees of various species, sizes, and conditions and their
derivatives such as snags and logs on the forest floor. Such structures provide habitats for many
species (Carey and Johnson 1995) and ecosystem functions (Franklin et al. 1981; Spies 1997). Re-
taining biological legacies acts to (i) perpetuate species and genotypes (lifeboating); (ii) enrich
F I G U R E 14. 16 Variable-
Photo courtesy of Jerry F. Franklin
retention harvest system in

old-growth Douglas-fir forest,
western Oregon. Silvicultural
prescription is for dispersed
structural retention of large
live trees (135-year old),
snags, and logs (20–25 trees
per hectare) to meet minimal
goals for coarse woody debris.
H. J. Andrews Experimental
Forest, western Oregon.
structure; and (iii) reduce the contrast in structure between managed and unmanaged areas (“soft-
ening the matrix”; Palik et al. 2021).
Lifeboating provides a way for genetic and species diversity to be carried from the pre-
harvest stand to the post-harvest stand. Plant species may be specifically lifeboated by retaining
mature individuals as seed sources or potential asexual reproducers by sprouting. Retention of
biological legacies may perpetuate biota in general by (i) providing habitat or refugia that might
otherwise be lost from the harvested area; (ii) ameliorating microclimatic conditions in relation
to those that would be encountered without retention, and (iii) providing energy and nutrients to
the post-harvest stand (Palik et al. 2021). For example, Figure 14.16 shows dispersed retention of
living trees and woody debris following harvest in a 135-year-old Douglas-fir forest. Retention of
trees in this manner retained microclimatic conditions critical for survival of some organisms. In
addition, retention of live trees and shrubs provides critical habitat to maintain populations of
soil organisms including mycorrhizae (Louma et al. 2006; Amaranthus et al. 1994).
Besides serving as refugia immediately after logging, structural retention enriches the com-
plexity of managed forest stands for an entire rotation. Suitable conditions for species can thereby
be re-established much faster than would otherwise be possible. Retention may provide suitable
habitat for species that are generally rare or absent in young stands of simple and homogeneous
structure. For example, many forest stands 80–200 years old in the Douglas-fir region provide suit-
able nesting and foraging habitat for the northern spotted owl and other species associated with
late-successional forests. By retaining some old-growth Douglas-fir trees, managed stands may
provide suitable nesting and foraging habitat for spotted owls within 50 or 60 years of harvest.
Without retention, it may take 120 years or more to create the necessary structural elements
(Franklin et al. 1997). Notably, lifeboating and structural enrichment may be accomplished simul-
taneously using the same biological legacies (e.g., live trees that both perpetuate plant species but
also provide habitat).
Facilitating the movement of organisms within a managed landscape is a third value of
structural retention. Connectivity is enhanced in the harvest unit by retaining structures in the
harvested area that facilitate the dispersion of some organisms, in addition to creating corridors
between intact stands. The objective here is to make the traditionally non-habitat, managed stands
(“the matrix”; see Chapter 22) more hospitable for dispersion by retaining, for example, well-
spaced logs, trees, and shrub patches for protective cover or habitat.
Designing a Variable-Retention Harvest System Developing a variable-retention harvest system

requires that managers identify which and how much structures to retain, and the spatial pattern
for the retention (i.e., dispersed or aggregated, or some combination of these). As described in the
earlier text, structures to retain may include live trees, especially large-diameter trees, snags in
varying states of decay, logs and other debris on the forest floor, forest understory, and undisturbed
layers of forest floor. Dispersed retention creates an even distribution of legacies (Figure 14.16),
which effectively provides wildlife habitats and future sources of energy across a large area. Aggre-
gated retention retains small patches of forest (“green trees”), which tends to provide more micro-
climatic modification and retain more diversity by canopy structure and species composition. For
example, in one of the first trials of aggregated retention (Figure 14.17), foresters of the Plum
Creek Timber Company retained 15% of the green trees as aggregates in a clear-cut unit. Similarly,
the City of Seattle experimented with aggregated retention using teardrop-shaped islands of green
trees left between radiating clear-cut strips (Figure 14.18). In both cases, these compact aggregates
are better at modifying microclimate and may be less susceptible to windthrow, whereas linear
aggregates can provide better visual screening, wildlife corridors, and stream protection.
How Well Does Variable Retention Conserve Biodiversity? Approximately three decades have
passed since the initial variable-retention systems were established, providing ample opportunity
for assessing their effects on diversity. Meta-analyses of variable-retention studies have suggested

F I G U R E 1 4 .1 7 One of the first examples of aggregated retention in which 15% of green trees were left in
the clear-cut area which lies in the white polygon shown. Aggregated retention is often advantageous from
standpoints of logging costs, safety concerns, and management efficiency. Plum Creek Timber Company
lands in southwestern Washington State.
F I G U R E 1 4 .1 8 Variable-retention harvest system in Douglas-fir forest illustrating the aggregated method

of retaining live trees following harvest. Teardrop-shaped patches of trees are left, providing corridor-like
connections of live-tree groups to the intact stand. Note individual trees left between the islands of green
trees. City of Seattle, Cedar River Watershed, western Washington State.
Epilog: Conserving Ecosystem and Biological Diversity 363
that the system is a valuable way to conserve biodiversity while simultaneously achieving timber
production (Beese et al. 2019a). An analysis of 39 studies that compared variable-retention har-
vesting to clear-cutting in North America and Europe found that retaining structures in the
harvested area reduced species loss in 72% of the studies, and typically improved habitat for early-
successional insects, birds, and woody plants (Rosenvald and Lõhmus 2008). Over the 39 studies,
species richness and abundance of birds and ectomycorrhizal fungi and the abundance of woody
plants increased with tree retention. Notably, the coverage of grasses and herbs generally decreased
in the presence of retained trees, but total species richness did not decrease because of the higher
presence and survival of woody plants (Rosenvald and Lõhmus 2008). A more recent analysis of 78
studies showed that species richness and abundance were higher in retention areas compared to
clear-cuts; the richness and abundance of species preferring open habitat were also higher in
retention areas compared to unharvested areas (Fedrowitz et al. 2014). Retaining a higher
proportion of trees generally increased species richness, but diversity does not appear to respond
to the spatial arrangement of retained trees (aggregated or dispersed).
Effects of variable-retention harvesting are not all positive, although most negative impacts
are on tree regeneration and growth rather than species richness and abundance. Some forest-
interior species found in unharvested forests are unable to tolerate harvesting even when trees and
other structures are retained, and some species that prefer open habitats benefit more from clear-
cuts than from variable-retention harvests (Fedrowitz et al. 2014). Several studies also suggest that
fungi and lichens, having many species associated with old forests, are particularly susceptible to
losses from harvesting (Trofymow et al. 2003; Sillett et al. 2000). Negative effects for timber produc-
tion are also non-trivial, as regeneration is often reduced in retention harvest units compared to
clear-cuts. For example, increasing the proportion of retained trees resulted in a decline in aspen
sucker density and volume 9 years after harvest in boreal mixed forests in Alberta, Canada (Grad-
owski et al. 2010); sucker density declined by 50% even when only 20% of the original basal area
was left in the stand. In the Pacific Northwest, growth of 5–12-year-old seedlings of Douglas-fir and
western white pine was reduced at higher levels of tree retention, and natural regeneration density
was higher in dispersed treatments compared to aggregated treatments (Urgenson et al. 2013). Re-
tained trees exposed to wind when adjacent trees are harvested may be damaged by wind at rates
as high as 50% and increase with reduced retention and aggregate size (Beese et al. 2019a, b). These
results suggest that managers need to be opportunistic with the flexibility provided by variable
retention harvesting in order to best balance the management objectives of timber production and
biodiversity conservation.
In general, ecological forestry is likely to play an important role in sustaining biodiversity
where timber production is an important objective (Beese et al. 2019a). Importantly, both
variable-retention harvest systems and preserves in combination are likely to be critical for con-
serving biodiversity rather than either approach alone (Lindenmayer and Franklin 2002).
Regardless of the improvement of variable-retention harvesting over clear-cutting, some species
will require forest interior conditions that cannot be provided in a harvested landscape regardless
of the proportion of trees retained. Nevertheless, variable-retention harvesting and other
methods encompassed by ecological forestry (such as variable density thinning; Palik et al. 2021;
Churchill et al. 2013) are important innovations in sustaining biodiversity in forests on human-
dominated landscapes.
EPILOG: CONSERVING ECOSYSTEM AND BIOLOGICAL DIVERSITY

The number of species in danger of extinction continues to increase as landscapes become increas-
ingly dominated by humans (Vitousek et al. 1997). Currently, more than 1600 species or subspe-
cies are considered to be endangered or threatened in the United States, 57% of which are plants,
representing an increase of 78% since the last edition of this book. Concern for diversity stretches
beyond endangered and threatened species, however; it is for the immense number of unknown
species and for ecosystem processes that are basic to long-term sustainability of ecosystems. As
such, targeting whole ecosystems for biodiversity conservation rather than individual species may
be an appropriate approach (Rowe 1990, 1997; Anderson and Ferree 2010; Shaffer 2015).
It is obvious from our discussion of diversity that species tend to be the focus of
conservation, which typically proceeds on a species-by-species basis. Ecologists have long real-
ized that such an approach is not sufficient, and that a species-by-species approach will fail
because it will quickly exhaust available time, financial resources, societal patience, and
scientific knowledge. To even come close to attaining the goal of preserving biodiversity, broad-
scale approaches—at the levels of regional and local ecosystems—are the only way to conserve
the overwhelming mass—millions of species—of existing biodiversity (Franklin 1993). Most
forethinking biodiversity conservation today combines a “coarse-filter” approach of protecting
broadscale or regional ecosystem processes and structures with a “fine-filter,” species-by-species
approach (Franklin et al. 2018). As we close this portion of the textbook that has focused largely
on the biota of forests, we re-emphasize a focus on the spaces—ecosystems that provide the
support system for the biota they contain—to facilitate protection of the species (Barnes 1993).
SUGGESTED
R E A D I N G S
Franklin, J.F., Johnson, K.N., and Johnson, D.L. (2018). Lindenmayer, D.B. and Franklin, J.F. (2002). Conserv-
Ecological Forest Management. Long Grove, IL: ing Forest Biodiversity: A Comprehensive Multiscaled
Waveland Press 646 pp. Approach. Washington, D.C.: Island Press 352 pp.
Huston, M.A. (1994). Biological Diversity. Cambridge Magurran, A.E. (2013). Measuring Biological Diversity.
University Press 681 pp. New York: Wiley 272 pp.
Latham, R.E. and Ricklefs, R.E. (1993). Continental com- Palik, B.J., D’Amato, A.W., Franklin, J.F., and Johnson,
parisons of temperate-zone tree species diversity. In: K.N. (2021). Ecological Silviculture. Long Grove, IL:
Species Diversity in Ecological Communities: Histori- Waveland Press 343 pp.
cal and Geographical Perspectives (ed. R. Ricklefs and Wilson, E.O. (1992). The Diversity of Life. Cambridge,
D. Schluter). Chicago: University of Chicago Press. MA: Belknap Press, Harvard University Press 424 pp.
Forest Ecosystem PA R T 5
Dynamics
A ll ecosystems are dynamic entities that change through time and over space, be
they aquatic or terrestrial. In the previous sections of this book, we have consid-
ered how the physical environment, in combination with life-history traits of forest trees and
associated plants and animals, influences the composition and structure of forest ecosystems
from place to place at a given time. In particular, we have discussed how climate, physiog-
raphy, and soil influence the geographic distribution of forest ecosystems at local, regional,
and global scales. In this part, we consider the mechanisms and ecological implications of
changes that occur over time in ecosystems as influenced by the physical environment.
Changes that occur in ecosystems over very long temporal scales reflect the rela-
tionship of biota with physical factors that form ecosystems after changes in vegetation
distribution and genetics occur over millennia in the form of plant migration, speciation,
and evolution. In Chapter 15, we discuss quaternary biogeography and paleoecology
to study changes in tree and forest distributions and environmental changes, respec-
tively, as an attempt to understand the ecological foundations of the ecosystems we see
today. Current ecosystems are impacted by disturbances of many kinds and in many
ways. Fire, flooding, insects, windstorms, and harvesting can remove one complement
of plants and animals and, over time, replace it with another. In Chapter 16, we treat
disturbance as an ecosystem process, recognizing that the events that disrupt forest
organisms are fundamental to ecosystems. The treatment of disturbance leads natu-
rally to the consideration of forest ecosystem succession. In Chapter 17, we discuss
the concepts, causes, mechanisms, and models of succession and provide examples
to illustrate how the composition and structure of different forest ecosystems change
following a particular type of disturbance.
Disturbance and subsequent changes in forest communities fundamentally direct
the pattern in which carbon and plant nutrients are cycled and stored within forest eco-
systems. In Chapter 18, we discuss the physiological processes controlling the capture,
storage, and loss of carbon (C) by forest trees, a set of processes that directly control
growth. We then build upon these principles to understand the capture, storage, and loss
of C by entire forest ecosystems, and the extent to which disturbance alters this ecosystem
365
366 PA RT 5 Forest Ecosystem Dynamics
process. Changes in the growth of forest ecosystems following a disturbance also changes
the demand for nutrients. Nutrient cycling is considered in Chapter 19, in which we trace
the flow of growth-limiting nutrients into, within, and out of forest ecosystems. We place
particular emphasis on the extent to which both natural and human disturbances alter
the pattern in which nutrients are cycled and stored within forests.
Long-Term Forest Ecosystem
CHAPTER 15
and Vegetation Change
E arly in this book, we described how landscape ecosystems exist at multiple spatial scales as
nested geographic and volumetric segments of the ecosphere. In this chapter, we consider
how forest ecosystems change over long temporal scales. We can easily observe important changes
in some ecosystem properties and processes at scales of days, seasons, and years, but noticeable
changes in landforms, soils, and biota occur over time frames of decades, centuries, and millen-
nia. Succession is the forest ecosystem change most commonly considered by forest ecologists.
It occurs over decades to centuries and creates differing vegetation composition, structure, and
biomass (Figure 2.2 of Chapter 2; Chapter 17). At even longer temporal scales, changes in vege-
tation distribution and genetics occur over thousands and millions of years in the form of plant
migration, speciation, and evolution (Figure 2.2 of Chapter 2), also within an ecosystem context.
Here, we emphasize long-term distributional changes in north temperate forest trees over the
last 20 000 years as they reflect associated climatic, physiographic, and soil changes in landscape
ecosystems.
What we know about long-term ecosystem change comes from geological studies in for-
merly ice-covered areas, from plant fossils, and from the forest trees themselves. Plants are among
the best indicators of historical landscape ecosystems for which there lacks a written record of
climate, physiography, and soil conditions. In other words, we are able to track and interpret the
migration of a tree species, but much less is known about the changes in regional and local climate,
physiography, soil, and disturbance regimes that largely controlled these migrations. We lean
heavily on using forest tree migration to provide insights into ecosystem change related to macro-
climate. In turn, we also expect major changes in tree migration to occur with future changes in
climate (Chapter 20).
Plants migrate with changing environments and may establish totally new spatial ranges, all
the while leaving evidence of extinction in the fossil pollen record at places of former dominance.
The genus Sequoia, for example, is today restricted to a single species distributed locally in
California, but fossil records show former distributions in Europe, central Asia, North America,
Greenland, Spitsbergen (midway between Norway and the North Pole), and the Canadian Arctic.
Tundra and boreal species in North America no longer exist where they once did 20 000 years ago,
except for small pockets at high elevations in areas such as the southern Appalachian Mountains.
When species persist in a given geographic site for several thousands of years, the genetic makeup
of those populations change as they adapt to regional and local climate (Chapter 3). For example,
the American beech of a given locality in eastern North America 9000 years ago differs genetically
from the beech population occupying the same place today.
CHANGE BEFORE THE PLEISTOCENE AGE

Flowering plants (angiosperms) and mammals have dominated the Earth since the beginning of
the Cenozoic (66 million years ago or 66 Ma). Long-term variations in global temperatures and dis-
placement of the land masses relative to their modern locations altered the prevailing climate,
367
368 Chapter 15 Long-Term Forest Ecosystem and Vegetation Change
which in turn shaped the geographical distribution of forests in ways different than today. Tree
taxa ancestral to many modern-day taxa were present in these forests, but it is highly unlikely that
trees were associated in regional-level and local-level ecosystems in the same communities that we
observe today. Nevertheless, plants occurred in spatial patterns that corresponded to climatic,
physiographic, and soil patterns of the time. For example, tropical rainforests covered much of the
southeastern United States 55–65 Ma, but high-elevation forests in the southern Appalachian
Mountains supported temperate communities of alders, birches, and hickories. Elsewhere, broad-
leaved evergreen forest dominated coastal regions, while inland forests at northern latitudes and
high-elevation forests consisted of more temperate, deciduous woody plants such as ginkgos,
viburnums, sycamores, birches, and elms (Graham 1999). By 16 Ma, as the global climate had dried
and cooled, the Southeast was dominated by a dry tropical forest with scrubby pines on sandy sites
and deciduous forests elsewhere (oak–chestnut, oak–hickory, and southern mixed hardwoods).
Montane coniferous forests of spruces and hemlocks developed at high elevations in the
Appalachians, and mixed hardwood–coniferous forests formed to the north featuring maples,
hickories, chestnuts, and beeches, with cooler forests of larches, pines, hemlocks, and birches.
Western forests had developed as woodland–savanna, mixed hardwood–conifer, and western
coniferous forests at high elevations (Graham 1999).
We do not expect that historical plant communities would exactly resemble those of modern
forests, nor do we expect that past species–site relationships would remain today. However, where
the physical components of physiography, microclimate, and soil were generally similar to those
today, certain ancestral genera of the canopy and groundflora layers may have been present in
combinations similar to those that exist today. Such appears to be the case at least by the end of the
Cenozoic (1.6 Ma), with various genera grouped together into associations that are familiar today.
Among many important events at this time are included the development of a familiar deciduous
formation of maple–basswood; a refinement of an Appalachian montane forest that featured firs,
spruces, and hemlocks; and the appearance of a near-modern boreal coniferous forest
(Graham 1999). Similarly, remnants of the wide-ranging dawn redwood forests of the Tertiary
Period forests 50 Ma still exist today, discovered only in 1949 in a remote valley in the northern
Hubei Province of China. Many genera of the Tertiary forests of northeastern Oregon (maples,
alders, Ailanthus, hackberries, beeches, hornbeams, pines, oaks, hemlocks, elms, etc.), where the
dawn redwood once grew, are also found today in the location in China where the dawn redwood
presently grows (Tallis 1991).
PLEISTOCENE GLACIATIONS
Ice accumulates on the Earth’s surface as the climate becomes progressively colder. The “Pleisto-
cene Ice Age,” beginning about 2.6 Ma, consisted of multiple glacial–interglacial cycles (currently
estimated at about 17) of about 100 000 years duration. Data from a series of ice cores at Vostok,
East Antarctica dating into the late Pleistocene suggest that cold glacial stages lasted 70 000–
90 000 years, as glacial ice sheets built up on continents. These glacial stages were directly followed
by a short and warm stage of about 10 000 years, though interglacial periods also varied in length
(Figure 15.1). The last continental cold–warm cycle in North America, the Wisconsin Glaciation,
reached its maximum extension 25 000–21 000 years ago. Thereafter, the climate warmed, and the
sudden warming about 11 650 years b.p. (Before Present) marks the beginning of the present
Holocene Epoch.
Previous glaciations eliminated forests when ice sheets overrode the land, with the exception
of a few small refugia (Figure 15.2; see discussion in the following text). The present forest vegeta-
tion of these now-deglaciated lands therefore dates from the last continental retreat of Pleistocene
ice sheets. Importantly, major changes in whole regional and local landscape ecosystems occurred
throughout the Pleistocene. New landscapes were created by the dynamics of the glaciers and
Pleistocene Glaciations 369
Depth (m)
0 500 1000 1500 2000 2500 2750 3000 3200 3300
Temperature (°C)
0
–2
–4
–6
–8
0 50 000 100 000 150 000 200 000 250 000 300 000 350 000 400 000
Age (yr BP)
F I G U R E 15.1 Temperature data determined from the Vostok ice core for the last 400 000 years incorporating
five interglacial periods (shown as shaded columns). Source: Petit et al. (1999).
Prudhoe
Bay C
B
Bethel Fairbanks
Whitehorse Yellowknife
Edmonton
Vancouver
Winnipeg
Seattle
Minneapolis
A
Chicago
Denver
St. Louis
F I G U R E 15 .2 Maximum extent of continental glaciers in North America during the last ice age. Major
refugia are marked by letters: A (southern refugium), B (Beringia), and C (Canadian Arctic Archipelago).
Source: Levsen and Mort (2009) / with permission of Canadian Science Publishing.
through the action of waters as the ice disintegrated. These new features affected climate at regional
and local scales in ice-covered and adjacent lands. The detailed consideration of these features is
beyond our scope, but we discuss the significance of physiography in glacial landscapes for forests
in Chapter 8.
ECOSYSTEM AND VEGETATIONAL CHANGE SINCE THE LAST

GLACIAL MAXIMUM
As glacial ice receded, new physiographic features and their soil parent materials were created and
colonized by organisms. Sediment-transporting meltwater formed new outwash plains, and lakes
formed and reformed, reworking the lands they covered. Forests migrated onto the deglaciated
land as the ice melted, and tree genera and certain species with distinctive pollen morphology in
the prehistoric record clearly indicate dramatic changes that occurred across landscapes and wa-
terscapes over the last 25 000 years. Pollen has been used to track and map the directions and rates
of tree population migrations using fossil pollen and plant parts and by reconstructing forest com-
position (Delcourt and Delcourt 1987). The study of prehistoric plants is known as Quaternary
biogeography if the emphasis is on geographical distribution of the flora, and as Quaternary
paleoecology if environmental changes are of primary concern.
The composition of forests in times past was determined by the combination of geologic
processes and interrelated climatic events. Climatic cooling and mountain building in the Tertiary
Period were major factors in western North America (Graham 1999), but Pleistocene glaciation
and postglacial climate have been the primary influences on ecosystem change and forest distribu-
tion and composition in eastern North America. Vegetation in temperate regions has been in an
almost constant state of instability and adjustment due in part to an almost continuously changing
climate over the past 20 000 years and even more fundamental changes in the configuration of the
land and the composition of its surficial materials (Davis 1986). Overall, vegetation changed slowly
during the glacial maximum until 17 000 years b.p., quickly between 16 000 and 8000 years b.p.,
then slowly again after 7000 years b.p. until vegetation change accelerated at about 500 years b.p.
due to the impacts of European activities (Williams et al. 2004). Importantly, however, climate var-
ied at every temporal scale important to forests, from years to decades to centuries to millenia
(Jackson 2004). Present vegetation patterns are closely related to events in recent geological his-
tory, of which climatic change is one important part.
EASTERN NORTH AMERICA

The major features of climatic history in eastern North America during the end of the Pleistocene
were summarized by Webb et al. (1993) and confirmed with the Vostok ice cores (Petit et al. 1999):
(i) climate changed continuously from 18 000 years b.p. to the present, (ii) a major increase in mean
annual temperature occurred between 15 000 and 9000 years b.p., (iii) the greatest seasonal con-
trast between cold winters and hot summers took place between 12 000 and 9000 years b.p., (iv) the
modern spatial gradients for July temperature and annual precipitation first appeared at 9000 years
b.p., (v) the time of maximum warmth was about 6000 years b.p. when temperatures were not more
than 1–2 °C higher than today; and (vi) the time of maximum dryness in the last 12 000 years was
6000 years b.p. in the Midwest and 9000 years b.p. in the northeastern United States.
In addition to the prevailing climate, key characteristics of tree species affect their migration
patterns and community composition. The potential range of environments in which trees can col-
onize and grow is far greater than what we see in their present-day natural occurrences for several
reasons. Tree species persisting today have evolved a balance of genetic and non-genetic mecha-
nisms to maintain fitness in a given environment or the flexibility to change with changing condi-
tions (Chapter 3). Trees that have persisted for millions of years in temperate and boreal environ-
ments have evolved mechanisms for long-distance dispersal using wind, water, and mutualisms
with birds and mammals. Competition may severely restrict forest species to a relatively small part
of their potential geographic range, that is, their realized niche (Chapter 13). Finally, many forest
trees have adaptations enabling them to colonize ecosystems with already-established commu-
nities and maintain themselves, depending on the specific site conditions (Chapter 21).
Ecosystem and Vegetational Change Since the Last Glacial Maximum 371
Overall Migration Sequence and Patterns An overview sequence of the location of

various broad forest types that followed the last glaciation is presented in Figure 15.3. In eastern
North America, the climate during glacial times was sufficiently cool even to middle latitudes, and
forests were displaced to the south of a narrow band of tundra vegetation at the ice margin. During
the late Wisconsin Glaciation, 18 000 years B.P., a boreal-like coniferous forest dominated by white
spruce occurred on both sides of the Mississippi River Valley south to about 31° N latitude
(Figure 15.3a). Just as today, structure and composition of the full-glacial boreal forest varied
across the region, and forest structure resembled an open woodland moving westward
(Adams 1997). Boreal species such as jack pine and white spruce were displaced as far to the south
as the coastal plain between northcentral Georgia and northern Florida (Watts 1970; Watts and
Stuiver 1980; Graham 1999; Jackson et al. 2000). Temperate genera such as walnuts, maples, oaks,
and beeches were likely locally abundant on moist soils in this region (Delcourt 1980; Jackson
et al. 1997). Temperate hardwoods grew locally in the Lower Mississippi Valley as far north as 35°
N latitude (Jackson et al. 2000), but white spruce, fir, and tamarack extended southward to nearly
31° N latitude, occupying extensive sand flats of braided streams that carried glacial meltwater
sediments into the Gulf of Mexico (Delcourt and Delcourt 1977; Royall et al. 1991). Evidence of
shifting sand dunes east of the Appalachians also suggests forests in this area were much more
open than those today (Wells 1992; Adams 1997). Forests dominated by pine (mainly jack pine
with local red and white pines) were found in this region, possibly as far south as the Florida Pan-
handle (Jackson et al. 2000).
The Gulf Coastal Plain was more forested than areas to the north, with a higher proportion
of deciduous trees (mainly oaks and hickories) mixed with southern pines (Watts and Stuiver 1980;
Delcourt and Delcourt 1981). Oaks and hickories may have been favored by dry summers, partic-
ularly on fire-prone sites in the area (Adams 1997). Some researchers have speculated that a
narrow transition zone of cool-temperate conifers and northern hardwoods existed from northern
Mississippi to South Carolina between the spruce and jack-pine-dominated forest to the north and
this oak–hickory–southern pine region on the Gulf Coastal Plain (Delcourt and Delcourt 1981).
Mixed mesophytic forest may have existed along the bluffs east of the Mississippi Valley and
within protected valleys of major river systems of the Southeast. These bluffs are composed of
loess soil (fertile, mesic, often calcareous wind-deposited silt), which is readily eroded to form
steep ravines and banks that would have provided protected sites for the species-rich mesophytic
communities (Delcourt and Delcourt 1979).
Tree populations began to migrate northward as the climate slowly became warmer and
wetter beginning about 16 500 years ago. By 14 000 years b.p., spruce–jack pine forest had colonized
lands formerly occupied by tundra as the ice sheet retreated, and mixed hardwoods–conifer forest,
now dominated by deciduous species, expanded northward and eastward across the southeastern
United States (Figure 15.3b; Watts 1979; Delcourt 1980; Whitehead 1981; Adams 1997). Oaks, in
particular, expanded quickly northward from the southeastern United States sometime before
14 000 years b.p. (Williams et al. 2004). Mixed mesophytic forest also expanded to the north and
east into the Appalachian Highlands at this time (Delcourt and Delcourt 1981). A period of cooling
and drying at about 13 000 years b.p. caused severe winters in the Northeast and a reversion from
a mixed hardwood–conifer forest (oak, ash, spruce, and fir) to one more closely resembling a boreal
community (spruce, fir, larch, birch, and alder). Pines rapidly increased in abundance during this
period in the Great Lakes region (Williams et al. 2004), largely replacing spruces probably due to a
regional decrease in moisture availability (Shuman et al. 2002). The colder and drier climatic phase
ended sometime near 11 500 years b.p. resulting in a decline in boreal vegetation in the northeast-
ern United States toward a forest dominated by white pine and oak, and deciduous genera
continued to increase in abundance and range elsewhere in eastern North America (Adams 1997).
Red spruce and Fraser fir populations were stranded at high elevations in the southern Appala-
chians and persist today on some of the higher montane peaks.
(a) (b)
Polar desert
Polar desert
Dry
Tundra
Tundra
Ice
Ice
Forest-
tundra Ice
Dry tundra
Temperate
semi-desert
Tundra
Mid taiga
Semi-desert
Open boreal Grass- Cool temperate/mixed
‘Mid taiga-like’ woodland Medtn land forest
Temperate open scrub
woodland Scrub Warm temperate
Temperate open forest types
Semi-desert woodland Temperate
Temperate scrub
conifer/mixed
Dry grassland
About 22,500 yrs BP About 12,900 yrs BP
Tropical
Savanna Tropical
rainforest
rainforest
(c) (d)
Ice
Polar
desert
Forest-tundra
Tundra
Tundra Open
boreal Mid
Ice woodland
Mid taiga
Forest-steppe
South taiga
South taiga
Grassland
Grassland Cool temperate/

Cool temperate/ and mixed
mixed
Medtn Semi-
Medtn Warm temperate
Semi- Warm temperate scrub desert
scrub forests
desert forests
Open
Temp.te woodland
Wdld Tropical rainforest
Tropical rainforest
About 8,900 yrs BP About 500 yrs BP
F I G U R E 1 5 .3 Maps of vegetation change for North America. (a) Map of vegetation for 22 500 years b.p. (b)
Map of vegetation for 12 900 years b.p. (c) Map of vegetation for 8900 years b.p. (d) Map of vegetation for
500 years b.p. Source: Adams (1997) / Adams.
Major changes in vegetation continued throughout the Holocene, most of which was slightly
warmer than at present. With increasing warmth and aridity in the Great Plains and a general
increase in the prevailing westerly winds, oak–hickory forest, oak savanna, and prairie shifted east-
ward between 9000 and 7000 years b.p. (Figure 15.3c). Fire, fanned by the westerly winds across
the predominantly flat landscape, was a major factor in pushing back and maintaining the prairie-
Ecosystem and Vegetational Change Since the Last Glacial Maximum 373
f orest border. Reaching into Minnesota, Illinois, Indiana, southern Michigan, and Ohio, prairie
openings gradually replaced mixed pine–deciduous forests to the north and elsewhere deciduous
forests and oak savannas. This Prairie Peninsula (Transeau 1935; Stuckey 1981) is evidenced by
the widespread occurrence of prairie soils, remnants of oak savannas, and by the persistent frag-
ments of the prairie until the initiation of farming that occurred with European colonization
(Whitney 1994). Pines and temperate hardwoods gradually replaced prairies, although prairie
openings probably survived in eastcentral Minnesota until 4000 years b.p. (Wright 1971). The
Prairie Peninsula posed a barrier to the northward migration of upland mesophytic species such as
beech, hemlock, and tulip tree. These species migrated into the Midwest from the east.
After 7000 years b.p., mixed conifer–hardwoods forest had moved into southern Canada, the
mixed mesophytic forest species were restricted to moist coves and slopes of the Cumberland and Al-
legheny Plateaus, and oak–chestnut communities dominated the southern and central Appalachians.
Southern pines replaced oak–hickory forests on sandy uplands of the Gulf and Atlantic Coastal Plains,
and oak–hickory–pine forests were restricted to the Piedmont Province and the Ozark and Ouachita
Mountains. Throughout this interglacial period, regional physiographic features of plains, plateaus,
and mountains were important in shaping the diverse vegetation types of modern regional ecosystems.
The present potential (pre-European colonization; Figure 15.3d) vegetation is similar to that at
5000 years except for a minor southward shift of the boreal forest and a retreat to the west of the
prairie-forest boundary. These changes were caused by a long-term cooling trend at high latitudes and
increased precipitation across the Midwest that began at about 6000 years b.p. In general, the present
potential vegetation boundaries are similar to those of Braun’s (1950) forest regions based on broad
physiography (Figure 2.6, Chapter 2). This correspondence illustrates the importance of regional
physiographic features in understanding the distribution of broad vegetation types. Macroclimate is
an important but not isolated driver of vegetation distribution since the last glacial maximum. More-
over, multiple geologic and soil factors as well as disturbances such as fire, flooding, and windstorms
acted at all spatial and temporal scales to shape the vegetation patterns of regional and local ecosystems.
Ecosystem Change in the Southern Appalachians Delcourt and Delcourt (1987)

describe ecosystem change in both time and space for Mt. LeConte (elevation 2000 m) in Great
Smoky Mountains National Park, eastern Tennessee to western North Carolina (Figure 15.4). At
20 000–16 500 years b.p., mean annual temperatures at elevations above 1500 m in the southern
Appalachians were well below 0 °C, creating perennially frozen ground (permafrost), and
permanent snowpacks. Intensive freeze–thaw churning of soil prevented establishment of trees,
but the montane environments were suited for alpine tundra. Coniferous krummholz dominated
lower elevations, at and below 1500 m. From 16 500 to 12 500 years b.p., rising temperatures to near
0 °C led to the development of a patchwork cover of tundra and krummholz above 1500 m. Boreal
and cool-temperate forests became established at intermediate and lower elevations, respectively.
Climatic amelioration after 12 000 years b.p. made temperate stream processes more influen-
tial than other geomorphological processes. By 10 000 years b.p., conifers and northern hardwoods
occupied mountain summits and upper slopes, and temperate deciduous hardwoods (including
mixed mesophytic forest) dominated the lower elevations. Today, a great diversity of ecosystems
comprises the landscapes of Mt. LeConte at all elevations (Figure 15.4), determined largely by the
physiographic diversity in elevation, aspect, and slope characters of the mountain landforms.
WESTERN NORTH AMERICA

In western North America, the continental ice sheet did not extend nearly as far south as it did in
eastern North America (Figure 15.2). Species were nevertheless displaced to the south, and
mountain glaciation was widespread in the Rocky Mountains and other ranges south of the
continental ice border. This mountain glaciation complicated the postglacial history in the Pacific
(d) TODAY F I G U R E 15. 4 Diagrammatic

BOREAL SPRUCE - FIR illustration of vegetational changes on
HEATH FOREST
BALD TEMPERATE Mt. LeConte, Great Smoky Mountains
FORESTS. National Park: (a) full-glacial interval,
MIXED OAK
CO
VE
HD 20 000–16 500 years b.p.; (b) late-glacial
WD PINE
S
interval, 16 500–12 500 years b.p.;
TULIPTREE
(c) early-Holocene interval, 12 500–
N 10 000 years b.p.; (d) today. Source: After
Delcourt and Delcourt (1987) /
Springer Nature. Reprinted from
(c) 12,500 to 10,000 yr B.P. Long-term Dynamics of the Temperate
KRUMMHOLZ
Zone, © 1987 by Springer-Verlag
New York, Inc. Reprinted by permis-
BOREAL FOREST
sion of Springer-Verlag New York, Inc.
TEMPERATE
DECIDUOUS
FOREST
(b) 16,500 to 12,500 yr B.P.

ALPIN TUNDRA
AND KRUMMHOLZ
BOREAL FOREST
TEMPERATE
DECIDUOUS FOREST
(a) 20,000 to 16,500 yr B.P.

SNOWFIELDS AND
ALPINE TUNDRA
KRUMMHOLZ
BOREAL FOREST
Northwest (from Oregon north to Alaska) because the ice radiated east and west from the north–
south-oriented mountain ranges rather than simply moving southward, as in eastern North
America. In addition, maritime climatic conditions originating from the Pacific Ocean moderated
at least some of the postglacial climatic fluctuations compared to those found further east. There-
fore, refugia for various plants and animals along the Pacific Coast, including Sitka spruce and
western hemlock, served to restock much of the glaciated terrain rather than migration from the
south as in eastern North America (Hansen 1947, 1955; Heusser 1960, 1965, 1983;
Barnosky 1987).
Near the last glacial maximum (22 500 years b.p.), most of the Pacific Northwest was alpine
tundra with only scattered areas of spruce (Whitlock and Bartlein 1997). During this period, the
central Rockies included a mosaic of open conifer forests and dry shrubland, depending on
altitude, and sparse temperate forest in the Southwest and along the Pacific Coast (Figure 15.3a;
Adams 1997). A late-glacial cold period about 14 500 years ago was followed by a warmer and
drier trend that ultimately gave way to a cooler, more humid climate. This period was also impor-
tant for the retreat of the ice such that a narrow ice-free corridor linked Alaska to the continental
Patterns of Tree Genera and Species Migrations 375
United States (Figure 15.3b). Boreal forest was present in the lowlands of Washington State, and
birches became common in the tundra regions of Alaska and northwest Canada. As the ice-free
corridor widened by about 13 000 years ago, spruce parkland and forest developed in western
Canada. This conifer-dominated area spread to the eastern foothills of the Canadian Rockies by
about 11 000 years ago. Some studies have suggested that a lodgepole pine parkland developed in
the Pacific Northwest and northwestward as far as southeastern Alaska (MacDonald and Cwyn-
ar 1985). Lodgepole pine reached its current most northward locality possibly a century ago
(Cwynar and MacDonald 1987) and migrated at an average speed of about 180 m per year over a
distance of 200 km in about 12 000 years (18.3 km per century). As the climate ameliorated,
Douglas-fir became the dominant species of the Pacific Northwest during the warmer and drier
times of the early-to-mid Holocene. Where rainfall was heavy, western hemlock replaced pine
and mountain hemlock, and Sitka spruce became well represented from British Columbia to
southeastern Alaska. In the Willamette Valley, the succession was to Douglas-fir and Oregon oak,
and grassland achieved dominance on the dry plateaus east of the Cascades. As in the East, the
approximate boundary dates of this period of maximum warmth and dryness are 8000–
4000 years b.p.
PATTERNS OF TREE GENERA AND SPECIES MIGRATIONS

Pollen analyses have been used to describe the individual migration pathways, directions, and rates
of various genera of forest trees of eastern North America rather than the location of broad vege-
tation types. Classic studies by Davis (1983a, b) examined migration routes and arrival times of
spruces, white pine, beech, and chestnut (Figure 15.5), among others. Many trees migrated gener-
ally from southern refuges northward along a broad front as seen for the spruces (Figure 15.5a). In
contrast, white pine (Figure 15.5b) and hemlock migrated north and west from refugia on the
central coastal plain or on the exposed continental shelf (Davis 1983a, b). American beech
(Figure 15.5c) migrated northward, and after about 9000 years ago expanded westward through
the Great Lakes region. American chestnut (Figure 15.5d) migrated through the Appalachian
Mountains in a northeasterly direction and arrived in New England after most of the other tree
species with which it was associated.
Early-successional species with many tiny wind-and water-dispersed seeds migrated at the
highest rate (Godwin 1956). For example, aspens expanding into tundra areas averaged a rate of
advance of 263 m per year from 14 000 to 8000 years b.p. The comparable migration rates for
heavy-seeded species such as oaks and hickories were 126 and 119 m per year, respectively. In
general, availability of suitable sites for establishment and subsequent growth was more limiting
than dispersal rate for migration rates for boreal and temperate trees in eastern North America
(Delcourt and Delcourt 1987). Physiography and soil therefore played important local roles in
glacial–interglacial range expansion and late-Quaternary community dynamics.
Birds and other animals were important seed dispersal agents and aided in establishment
for many plant species during post-glacial periods. These mutualisms are significant in tree
species of the beech family, especially for oaks, American beech, and American chestnut (John-
son and Webb 1989), all of which are known to be dispersed and cached by blue jays. Blue jay
acorn dispersal may explain the rapid expansion of the range of oaks in late glacial times relative
to beech and chestnut. The slow migration rate of chestnut may be partially explained by the
preference by jays for small nuts relative to chestnut. Beech migration was probably limited more
by climate than by dispersal (Johnson and Webb 1989), as well as by its narrow site requirements
for establishment and late seed-bearing age. When dispersal occurs over long distances (typically
by birds), plants may reach suitable sites for establishment far ahead of the leading edge of their
main advance. Similarly, strong winds can place small seeds in outlying sites far ahead of the
advancing front. For example, by 12 000 years ago aspen populations had established in Maine,
(a) (b)
8 5 6
7
8 10 10
9 99 7 6 7 8
10 8 11 11 7 9
12 11 2 10
10 82 8 8
11
14 12 5 8 8 10
11 12
9 9 10 10
12 12 12 6
11 0 10
13
12 12 0 10 10 11 12
12 12 12 ? 11 12
12 10? 10 11
12 10
10
21 14 >13
0
23 12 >11
0
25 (14)
14 10
22–16 0
20–13 17 0
Picea spp Pinus
strobus
>14 Spruce
0
White
Pine
0 400 km 0 400 km
(c) (d)
4
4
5 54 6
6
46
5 6
8 8
9 2 2
7 7 8 8 2
78 7 4
5
6 6 4 2
5 6
6 5
6 8 10
10 8
10 8
13 10 8
15 11 15
12 12 5
Castanea
>13 14 14 Fagus dentata
grandi-
Chestnut
folia
Beech
0 400 km 0 400 km
F I G U R E 1 5 .5 Migration routes of tree species in eastern North America during the late-glacial and early
Holocene. Small numbers on the map indicate the time of arrival at individual sites. Contours show the
leading edge of population advance at 1000-year intervals. Shaded area is the range for the genus or species
prior to European colonization. Source: After Davis (1983b) / Missouri Botanical Garden Press. Reprinted by
permission of Margaret B. Davis and the Missouri Botanical Gardens.
Patterns of Tree Genera and Species Migrations 377
then 700 km east of the continuous range margin for aspen in the Lake States (Davis and Jacobson
1985; Delcourt and Delcourt 1987). In the western Great Lakes region, hemlock was advanced by
long-distance dispersal of seeds, forming isolated populations ahead of the main front (Davis 1987).
MIGRATION IRREGULARITIES AND DISTURBANCE

In imagining the re-vegetation of recently deglaciated terrain, we envision smooth waves of trees
and other organisms establishing and migrating over the newly available sites to reach their pre-
sent areas of distribution. Importantly, migration appears “smooth” only at the temporal scale of
millennia, and migration over centuries would appear to respond more directly to barriers or
corridors such as major rivers or mountain ranges. The pollen record has recognized numerous
irregularities from smooth advances even at long temporal scales, such as the migration of
American beech around or across Lake Michigan (Davis et al. 1986; Webb 1986; Davis 1987;
Delcourt and Delcourt 1991) and patterns of hemlock expansion in the western Great Lakes
region (Davis et al. 1986; Davis 1987). In New England, early colonization of sites by elms is
noted 300–700 years earlier than the main “explosion” in elm pollen values (Webb et al. 1983;
Davis and Jacobson 1985).
Once initial re-vegetation occurred, disturbances by wind, flooding, and fire were undoubt-
edly very important in providing open sites for establishment of migrating species. Both severe
and minor disturbances were important in incorporating late-arriving species into existing forest
communities. Fire, in particular, provided suitable seedbeds, temporarily reduced competition,
and generally favored the migration and maintenance of all boreal species, pines, oaks, and
many other temperate species. Even a species such as eastern hemlock, although perceived as
fire-sensitive, establishes readily on fire-prepared seedbeds. For example, eastern hemlock colo-
nizes burned areas at the same time as trembling aspen, paper birch, or white pine, but grows far
more slowly, reaching the overstory only after 100 years or more. This pattern has been recog-
nized in the pollen record (Swain 1973, 1978) as initial high proportions of aspen and birch
pollen, and increases in white pine pollen after 50 years, and hemlock after 100 years
(Swain 1973, 1978). The relationship of climate change and fire regimes and their collective
influence on temperate forest composition have been demonstrated in several paleoecological
studies (Green 1981; Clark 1989, 1990) and reviewed by Patterson and Backman (1988) and Del-
court and Delcourt (1991). Again, macroclimate is a main driver of ecosystem and vegetation
change, but geomorphic processes also change in time and interact with climate at regional and
local levels. These geomorphic processes also interact and drive disturbances. Thus, long-term
vegetation change is a complex ecosystem process involving many interrelated factors.
MIGRATION FROM GLACIAL MICROREFUGIA

When considering migration of species during Pleistocene deglaciation, pollen studies are unlikely
to detect small or outlying populations. Reconstruction of the late-glacial distribution of vegetation
relies on the amount of pollen produced and deposited in sediments. A high proportion of pollen of
a species in a sediment suggests that the species was not only present, but present at an abundance
high enough to produce most of the pollen found in the sample. The converse is not necessarily true.
When pollen of a species is lacking within a sample, it is only be concluded that there was not
enough pollen produced by the species to leave a substantial record in the sample. Often the
conclusion is that the species was absent at that location, but it might also mean that the species was
present in small pockets distant from where the sample was taken, or that the species was present
but not abundant enough to leave a pollen record. In short, pollen data are very good at indicating
the distribution of species that reach higher densities, but they are far less effective at identifying
small populations, or even widespread, sparser populations (McLachlan and Clark 2004).
Understanding the details of paleoecological studies is important for how we estimate the
rates of migrating taxa. Accurately assessing the migration rates of forest taxa during the rapid
changes in climate during deglaciation has become increasingly important in the current era of cli-
mate change, because it provides some perspective on how quickly these taxa will adapt as the con-
temporary climate continues to change (Chapter 20). As described earlier in this chapter, many
temperate species and genera present in eastern North America today are thought to have origi-
nated mostly from the southeastern United States during the last glacial maximum, with those
species having moved northward as the ice sheet retreated (Figures 15.3 and 15.5). This notion of
Pleistocene vegetation, first advocated by Edward Deevey (1949), reasoned that the colder climate
that accompanied the ice sheet displaced temperate species far south of their current distributions.
These regions of refugia generally include central Texas, the Lower Mississippi River Valley, the
Florida Peninsula, the Atlantic and Gulf Coastal Plains, and the southern Appalachians (Soltis
et al. 2006; Jaramillo-Correa et al. 2009). An alternate argument is that many temperate eastern taxa
maintained smaller populations closer to the ice margin during the last glacial maximum (Jackson
et al. 2000; Delcourt 2003), as initially evidenced by the presence of plant fossil specimens discov-
ered farther north. As a counter to Deevey’s hypothesis of southern refugia, E. Lucy Braun (1950)
argued that more northern regions, particularly the Cumberland Plateau of western Kentucky and
Tennessee, served as a smaller refugium for temperate taxa (today recognized as those of the mixed
mesophytic forest), although the surrounding region was dominated by boreal vegetation. These
northern microrefugia would have been small populations surviving on suitable sites in an other-
wise unsuitable climate. If microrefugia, particularly those located closer to modern species’ ranges,
contributed to post-glacial re-vegetation, then ignoring them may overestimate the species’ migra-
tion rates (McLachlan et al. 2005). Such populations would have had to persist throughout the last
glacial maximum to be likely sources of re-vegetation, so their importance is often ambiguous, and
the focus remains on the displacement of taxa into southern geographical areas.
The advent of molecular and genomic methodologies in the early twenty-first century
provided strong evidence for refugia that pollen studies were unable to detect. These techniques
can be used to compare the genetic structures of living trees from across their geographic range
and thus reconstruct the migration paths of populations as they expanded. In a classic study, chlo-
roplast DNA was used to examine the migration routes and rates of two temperate deciduous
species in eastern North America: red maple and American beech (McLachlan et al. 2005). Pollen-
based studies had concluded that red maple originated from a refugium near the Lower Mississippi
River Valley, migrating northward and westward into the Great Lakes region and the Northeast.
Molecular evidence, however, suggested that more northerly populations from the Appalachian
Mountains contributed most to the spread of red maple to the north. Likewise, American beech
was assumed to have originated from a refugium on the Gulf Coastal Plain based on pollen evi-
dence (Davis 1981; Delcourt and Delcourt 1987), but the genetic structure of most beech popula-
tions in the Upper Great Lakes region did not resemble that of beech found in traditional southern
refugia. Thus, low-density populations of red maple and American beech likely persisted near the
ice sheet rather than in southern refugia, and at least some of these populations were influential in
post-glacial re-vegetation. It is possible that additional small populations that made little or no con-
tribution to re-vegetation also existed during this period, which are not detectable using these
methodologies.
Modern genomic evidence does not suggest that every species existed in microrefugia north
of their southern refugia during the last glacial maximum. For example, hickory species are
thought to have originated from the Gulf Coastal Plain based on reconstructions from pollen data
(Watts and Stuiver 1980; Delcourt and Delcourt 1981). Genomic evidence, however, suggests that
bitternut hickory expanded from a northern microrefugium near the confluence of the Mississippi
and Ohio Rivers where habitat was otherwise inhospitable (Figure 15.6a), whereas pignut hickory
indeed expanded from a main southern refugium in central Mississippi on the eastern Gulf Coastal
Independent Migration and Similarity of Communities Through Time 379
(a) (b)
50° 50° likelihood of
origin (relative
to maximum
likelihood)
1
0.9
40° 40°
0.8
0.7
0.6
0.5
0.4
30° 30°
0.3
0.2
0.1
–100° –90° –80° –70° –100° –90° –80° –70°
F I G U R E 15 .6 Expansion origins for (a) bitternut hickory and (b) pignut hickory in eastern North America
during the last glacial period based on genomic evidence. The red cross represents the maximum likelihood
of the point of origin, with the color ramp indicating the probability of alternate points of origin. Blue
shaded area represents glaciated regions. Source: After Bemmels et al. (2019) / with permission of PNAS.
Plain (Figure 15.6b; Bemmels et al. 2019). Increasing molecular evidence will undoubtedly show
that both northern microrefugia and traditional southern refugia were important in the re-
vegetation by temperate taxa of deglaciated terrain during the late Pleistocene and early to mid
Holocene (Soltis et al. 2006; Jaramillo-Correa et al. 2009).
INDEPENDENT MIGRATION AND SIMILARITY OF COMMUNITIES

THROUGH TIME
Paleoecologists have made enormous advances over the years in determining the general patterns
of response of a variety of tree genera to the changing conditions of the last 20 000 years. Most
paleoecological literatures of the past two decades acknowledge that plant taxa have migrated
independently of each other in response to climatic changes, and that plant communities arise
fortuitously (e.g., Graham 1999). This individualistic nature of plant taxa migration includes not
just the rate of migration, but also the timing of migration and its direction of range expansion and
taxa abundance (Williams et al. 2004). As a result, many of the tree genera and species that we
observe growing together today did not co-occur in the past, and many associations or commu-
nities recognized during the Holocene have no modern analog (Williams et al. 2004; Williams and
Jackson 2007). An example of an “extinct” community is one that featured spruces, sedges, ash,
and hornbeams. This community was very common in eastern North America between 17 000 and
12 000 years ago, but today the geographic ranges as well as the local site requirements of these taxa
overlap far less, if at all.
By contrast, many modern forest communities date at least to the early Holocene, such as
several boreal assemblages, oaks and hickories, and American beech and eastern hemlock. In
some cases, such as with oaks and hickories, the historical formation can be detailed only to genera
by the pollen record, and it remains unclear whether the same species were associated together
during the Holocene as are associated today. Modern genomic data have the capability to phyloge-
netically differentiate among species within genera (e.g., Bemmels et al. 2019) and should soon
address many of these issues. Other taxa associated today had similar geographic ranges in the past
and had the opportunity to co-occur geographically and ecologically. Notably, the range expansion
of some tree species, such as hemlock and beech, lagged behind those of other temperate species
with which they are associated today, and did not migrate at the same rate or along the same paths
during the Pleistocene (Davis 1983a). Using pre-European colonization ecosystems of 300 years
ago as a baseline, ecosystem and community similarity would tend to be least similar in the early
stages of migration and become progressively similar as these species colonized established
communities.
Although we often recognize plant formations present during the glacial period as resem-
bling modern forest communities (e.g., spruce–jack pine, mixed conifer–northern hardwoods,
mixed mesophytic, oak–hickory–southern pine), their occurrence does not necessarily imply that
the local ecosystems and their communities of the past are the same as those of today. This is
because the similarity of communities, past and present, depends on the similarity of the physical
environment and disturbance regimes of the site where the plants grow. Some species are similar
enough in their site requirements and tolerances that they respond similarly to disturbances of
fire, flood, windstorm, and specific site conditions of soil water and nutrient availability. For
example, we would expect combinations of “fire species” such as black spruce, tamarack, and
leatherleaf; jack pine and red pine; eastern white pine and eastern hemlock; and loblolly and short-
leaf pines to be more often associated together (with characteristic groundflora species) in fire-
prone ecosystems of their respective regional ecosystems rather than growing in floodplain ecosys-
tems with willows and cottonwoods or elms and silver maples. Thus, where regional gradients or
local environments of the past were similar to those of today, it is probable that certain species
occurred in characteristic site-species patterns not unlike those of today. There is no exact
correspondence between present and past communities, nor can entire intact communities migrate
together because they are structurally and functionally linked to physiographic and soil features of
landscape ecosystems that do not migrate.
SUGGESTED
READINGS
Davis, M.B. (1983). Holocene vegetational history of the McLachlan, J.S., Clark, J.S., and Manos, P.S. (2005).
eastern United States. In: The Late Quaternary Envi- Molecular indicators of tree migration capacity
ronments of the United States, The Holocene, vol. 2 under rapid climate change. Ecology 86: 2088–2098.
(ed. H.E. Wright and S. Porter). Minneapolis: Uni- Williams, J.W. and Jackson, S.T. (2007). Novel climates,
versity of Minnesota Press. no-analog communities and ecological surprises.
Delcourt, P.A. and Delcourt, H.R. (1987). Long-Term Front. Ecol. Environ. 5: 475–482.
Forest Dynamics of the Temperate Zone. Springer- Williams, J.W., Shuman, B.N., Webb, T. III et al. (2004).
Verlag 439 pp. Late-quaternary vegetation dynamics in North
Delcourt, H.R. and Delcourt, P.A. (1991). Quaternary America: scaling from taxa to biomes. Ecol. Monogr.
Ecology, a Paleoecological Perspective. New York: 74: 309–334.
Chapman and Hall 242 pp.
Graham, A. (1999). Late Cretaceous and Cenozoic His-
tory of North American Vegetation. New York: Oxford
University Press 350 pp.
Disturbance CHAPTER 16
D isturbances affect and often drive ecosystem composition, structure, and function by killing
organisms and changing site conditions. Disturbances selectively disrupt biota and initiate
succession, the course and rate of which are often directed by disturbance-caused changes in com-
position, species interactions, and spatial patterns of organisms. These changes in species compo-
sition occur in the context of disturbance effects on parent material, soil, and hydrology. All natural
ecosystems have been shaped by disturbances in some form, from plate tectonics and volcanism to
local avalanches, floods, and fires. Disturbances are important functional components and a vital
part of all ecosystems over long periods of time.
Particular kinds of natural disturbance are characteristic of particular kinds of landscape
ecosystems—such as avalanches in mountains, fires in dry plains, hurricanes along coastal regions,
or floods in river valleys. Disturbances influence changes difficult to observe, such as the genetic
differentiation of plants of these ecosystems, as well as the obvious vegetation changes in ecosys-
tem structure and composition. Many kinds of natural disturbances were prevalent in the forest
ecosystems of North America long before the first migrants arrived from eastern Asia. Humans
thereafter significantly disturbed environments in new ways as they developed cultural landscapes.
For many decades, researchers assumed pre-European colonization forests to be undisturbed old-
growth, Clemenstian “climax” forests (Whitney 1994). Moreover, most forest ecologists agree that
widespread and localized disturbance patterns of wind, fire, and other disturbance types associ-
ated with regional ecosystems and landforms were commonplace (Frelich 2002). Whitney (1994)
observes: “If the primeval forest did not consist of stagnant stands of immense trees stretching
with little change over vast areas (Cline and Spurr 1942), neither was it an amalgamation of pioneer
species recovering from one form of disturbance or another.”
Human-caused disturbances are superimposed on those of natural ecosystems, and human
alteration of original landscape ecosystems has been enormous throughout the world. Humans
significantly affect the soil, microclimate, and biota wherever they occupy a region as they develop
a culture. Trees have been the principal fuel and building material of almost every society for over
5000 years, accompanied by deforestation (Perlin 1989), and specific evidence of widespread
human disturbance has been documented for North America as early as the sixteenth century
(Denevan 1992). Working in the eastern United States, Abrams and Nowacki (2019) used tree
pollen and charcoal records to compare pre-European colonization and modern forests, finding
that modern forests contain tree species more shade-tolerant and drought-intolerant and less fire-
adapted than early forests. As such changes could not be explained by climate alone, they concluded
that vegetation changes were likely driven by reduced burning, associated with shifts in human
population and land use well prior to European colonization. Whitney (1994) has described distur-
bances of all kinds, natural and human, and their consequences from 1500 AD to the present for
eastern North America. Overall perspective and details of disturbance are provided by Pickett and
White (1985), Oliver and Larson (1996), Coutts and Grace (1995), and Johnson and Miyanishi
(2001). Here, we consider the concept of disturbance and the kinds of disturbances that bring
381
382 Chapter 16 Disturbance
about secondary succession—fire, wind, insects and pathogens, logging, land clearing, biotic
changes—as an introduction to a detailed examination of succession as an ecological process in
Chapter 17.
Forest ecologists and managers are concerned mainly with how disturbances initiate ecosys-
tem change, specifically regeneration of forest species and changes over time in composition, stand
structure, and biomass. These changes to the existing forest are often described as forest dynamics
or secondary succession, in contrast to primary succession, which occurs on previously unvegetat-
ed sites such as water or bare soil or rock. Disturbances of one sort or another in any forest region
are constantly altering the course of forest dynamics and initiating or restarting secondary
succession. Understanding how disturbances shape ecosystems remains one of the most exciting
fields of forest ecology.
CONCEPTS OF DISTURBANCE
As humans increasingly alter natural systems in novel and unprecedented ways, the concept of
disturbance in ecology has become somewhat more complex than it appears at first consideration.
DEFINING A DISTURBANCE
The word “disturbance” in ecology is generally interpreted to mean any relatively discrete event in
time that disrupts the composition, structure, and function of ecosystems. Disturbances typically kill
some but not all of the organisms in an ecosystem, and in the process resources such as nutrients,
water, growing space, or light are made available to new or surviving organisms (Pickett and
White 1985). When disturbances are natural (not caused by humans), such as wildfires, wind-
storms, hurricanes, or insect outbreaks, they are typically part of a well-defined historical and
repeating disturbance regime (Pickett and White 1985; Turner 2010). That is, natural distur-
bances are usually neither unique nor novel events, but part of a recurring series of events that
have long shaped the ecosystems and landscapes where they occur. Human disturbances on the
other hand, such as logging or land clearing, may be particularly unique to a given ecosystem in
terms of their severity, frequency, or seasonality, and therefore may elicit very different ecological
responses.
An important component in defining disturbances is that they are discrete in time whether
natural or human-caused, meaning they have a beginning and an end, and thus are absolute and
not relative (Pickett and White 1985; White and Jentsch 2001). Non-discrete events that affect eco-
systems are more aptly categorized as perturbations rather than disturbances, where some defin-
ing parameter of the system changes and the system departs from its normal behavior or trajectory.
Perturbations are typically gradual changes and are often caused by humans; climate change is
perhaps the best example of a perturbation. Keeley and Pausas (2019) identified common pertur-
bations to fire regimes (see description in the following text) as human-caused changes in specific
fire parameters, such as alterations in the frequency of fire, the pattern of burning across land-
scapes, the amount and types of fuels accumulated in some ecosystems, or the intensity of fires.
Ecosystems adapted to fire are not necessarily threatened by fire itself but by changes in the way
fire burns in ecosystems.
The definition of disturbance given in the earlier text is not perfect and is not synchronous
with all other definitions of disturbance. For example, disturbances have also been defined primar-
ily based upon their duration. Many ecologists today think about disturbances in terms of “pulse”
and “press” disturbances (Bender et al. 1984). Pulse disturbances are typically short-term,
abrupt, easily delineated disturbances, such as fires, floods, or even timber harvest, after which the
ecosystem responds and recovers. By contrast, press disturbances are disturbances of a pro-
longed intensity that have long-term impacts on an ecosystem, often resulting from changes to its
physical structure. Examples of press disturbances in forests might be clearing of forests for
Concepts of Disturbance 383
agriculture, chronic browsing by deer or elk, or climate change. This categorization of disturbances
is simple, and it clearly incorporates long-term stressors and gradual changes into the definition.
The duration of a disturbance has important implications for how an ecosystem recovers. Under-
standing the length of time that has passed since the disturbance ended is equally important for
understanding the structure and function of ecosystems.
Generally, an individual natural disturbance event occurs within a sequence of similar events
over space and time for a particular ecosystem, known as a disturbance regime (White and
Jentsch 2001). Disturbance regimes are typically characterized using descriptors of average distur-
bance frequency (how often the disturbance occurs in the ecosystem), size (the typical area
disturbed by an individual disturbance event), intensity (the physical energy of the event, such as
wind speed or the heat released by a fire), and severity (the effect of the event on the ecosystem).
Often disturbance regimes are described in terms of one type of disturbance (e.g., a fire regime or
flood regime), and this is characteristically the most pervasive disturbance type for a particular site.
However, a disturbance regime may include different types of disturbance agents. For example, the
fire regime of Yellowstone National Park is characterized by large (>5000 ha), infrequent (occurring
every 100–300 years), stand-replacing wildfires (most or all trees are killed in a burned patch; Rom-
me and Despain 1989). However, small (1–5 ha) stand-replacing fires occur between large fires
almost every year (Despain 1990), widespread bark beetle outbreaks occur every 20–40 years (Cole
and Amman 1980), and individual trees are uprooted by wind at least weekly. Each of these
disturbances occurs at its own spatial scale (from a single tree to a forest >5000 ha) and temporal
scale (from weekly or daily events to those that occur every few centuries). Notably, these different
disturbance types may interact and affect the frequency and intensity of the others, often in novel
ways. For example, bark beetle outbreaks have been shown to affect the probability of subsequent
burning by large, infrequent fires in Yellowstone (Lynch et al. 2006).
Subjectively, the most important effect of disturbances on ecosystems is that they alter avail-
able energy, substrate, and community composition, creating spatial and temporal variability
within and among ecosystems. The way ecosystems respond to disturbances has two primary com-
ponents. Resistance is the ability of an ecosystem to remain unchanged when subjected to some
driver of disturbance; generally, a more resistant system requires a stronger force to change it
(Carpenter et al. 2001). For example, Sánchez-Pinillos et al. (2019) measured resistance in the
boreal forest of Quebec, Canada as the similarity in forest structure and composition between the
pre-disturbance state and the state immediately after disturbance. Black spruce forests were highly
resistant to spruce budworm outbreaks because spruce bud burst and budworm larval emergence
were not synchronized. Outbreaks that were severe or of long duration, however, had the potential
to overcome this resistance and cause an acute disturbance event, leading to collapse of black
spruce forests into treeless systems. By contrast, balsam fir forests showed low resistance to spruce
budworm outbreaks, easily shifting toward dominance of paper birch, but the disturbed forests
recover to the pre-disturbance composition quickly (Sánchez-Pinillos et al. 2019). This ability of an
ecosystem to recover from a disturbance and return to its pre-disturbance state is called its resil-
ience. Resistance and resilience together determine an ecosystem’s stability, a term which has
become ambiguous and cumbersome in the ecological literature (Grimm and Wissel 1997; May
and McLean 2007).
The conceptual bases of resistance and resilience are often illustrated using “ball-in-cup”
diagrams (Figure 16.1; Holling and Meffe 1996; Lamothe et al. 2019). Ecosystems—particularly
biota—can persist in a stable state, a condition defined by the combination of the biotic
community and the disturbance regime. For example, a lodgepole pine forest can be maintained
through time by fire; the forest and disturbance regimes together constitute a stable state, though
at any point some areas may be burned or in an intermediate stage of recovery.
Many ecosystems contain multiple stable states toward which the system may move after a
disturbance (Beisner et al. 2003), with unique species compositions and disturbance processes. In
Black spruce forests Balsam fir forests
Fir
Birch
Spruce Spruce
Treeless Treeless Fir–Birch Fir–Birch
(low BA) (high BA)
F I G U R E 1 6 .1 Ball-in-cup diagrams representing resistance, resilience, and multiple stable states in the
boreal forest of Quebec. Black spruce forest has high resistance, represented by narrow deep cups, implying
that a significant disturbance would be necessary to move spruce forests to treeless systems (though only a
small disturbance would be needed to shift between high-and low-density forests). Balsam fir forests have
low resistance but high resilience, represented by broad, shallow cups that would require less disturbance for
movement among stable states. Notably, pure fir forests are extremely easy to shift toward a different species
composition, thus the “ball” is located at the top of a narrow peak. Source: Sánchez-Pinillos et al. (2019).
a ball-in-cup diagram, the bottom of the cup represents a given stable state. In the black spruce
forest example described in the earlier text, alternate stable states include high-density spruce
forest, low-density spruce forest, and treeless former forests. Having high resistance, each stable
state in the black spruce forest is represented as a ball in a steep-sided cup, because it would take a
large force, or a severe disturbance, to move the ball out of such a position (Figure 16.1). By
contrast, the balsam fir forest, with high resilience but low resistance, is represented by balls in
broader cups with shallower walls, where transitions between stable states (in this case fir forest,
birch forest, or mixed forest) occur easily (Figure 16.1).
DISTURBANCE AS AN ECOSYSTEM PROCESS

Importantly, any truly ecological definition of disturbance encompasses the disruption of all ecosys-
tem components, such as landforms, soil, vegetation, and any of the processes that link these com-
ponents. However, disturbance definitions often have an organismal focus, such as “a force that kills
at least one canopy tree” (Runkle 1985), “the mechanisms which limit the plant biomass by causing
its partial or total destruction” (Grubb 1988), or events that make growing space and other resources
available to survivors or new colonists (Glenn-Lewin and van der Maarel 1992). These are certainly
disturbances, but the concept does not end here. Disturbance is often conceived as representing
death and destruction when organisms are the primary focus, but they are in fact fundamental
natural events that occur in all ecosystems and, in many cases, are responsible for the stability of
those organisms through time. Even recent recognition of the importance of disturbances for eco-
system structure and function tends to have an organismal focus in terms of recovery.
Flood, fire, or wind that kills forest organisms allows the study of recovery mechanisms and
the sequence of organisms that replace those destroyed. Doing so, however, requires at least an
understanding of the underlying ecosystem processes that mediate the changes following distur-
bance, and such an understanding results from conceiving disturbance itself as an ecosystem
process. In doing so, we find that certain disturbances are characteristic of physiographic features
(Chapter 8), for example, and disturbances may be expressed by rates of occurrence per unit of
time (flooding and avalanche frequency, fire return interval) just as the rates of change of other
ecosystem processes are expressed.
Concepts of Disturbance 385
An ecosystem perspective of disturbance brings several insights worth considering:

Disturbance types are specific to regional ecosystems. Most natural disturbances are not ubiquitous;
that is, their occurrence is not randomly located but rather is shaped by macroclimate and phys-
iography at broad spatial scales. For example, some parts of North America are more susceptible
to hurricanes (Gulf and Atlantic coasts) than other areas because of their location relative to the
ocean and the movement of tropical storms. Severe thunderstorms and tornadoes occur in the
mid-continent because of the relatively flat topography and their location where dry cold air from
the north meets warm humid air from the south. Lightning-caused fires occur in the Intermoun-
tain West because of a dry climate resulting from Pacific storm precipitation blocked by the
Cascade and Sierra Nevada ranges to the west. At finer spatial scales within these regions, specific
ecosystem characteristics (physiography, soil, vegetation) and position of an ecosystem in rela-
tion to other ecosystems strongly influence the kind, frequency, and severity of disturbance
(Chapter 8).
Disturbances affect sites as well as biota. The dramatic effects of disturbance on individual plants
and communities are often the focus, but disturbances have many major effects on forest sites
that impact subsequent regeneration and forest succession. Certain disturbances create new
landforms, rearrange parent material, or otherwise change soil properties. These include volca-
nism, earthquakes, and glaciation at broad spatial scales and land movement (debris flow, mud-
slide, soil creep), floodwater, wind, and fire at finer scales. Wind creates new landforms such as
dunes and loess deposits, and floodwater creates new landforms by cutting riverbanks and form-
ing point bars, levees, and terraces. Wind also rearranges soils and creates diverse microtopogra-
phy by uprooting trees (Chapter 8). Removal or killing of vegetation also changes the light and
temperature regimes of the site. Thus, disturbances have far-reaching effects on regeneration
and forest succession beyond their direct effects on biota through both severe and subtle
site changes.
Disturbances vary in their effects. Different disturbances, such as fire, wind, and flooding, each
affect site and biota differently, resulting in different processes of regeneration and succession.
For example, wind primarily damages large, old, overstory trees rather than young stands or
the understory of old stands. By contrast, fire almost always affects the understory—whether it
is crown fire that kills the overstory and kills or even consumes the understory, or surface fire
that kills or severely damages only the understory. Unlike wind and fire, flooding has more
significant direct effects on landform and soil and more indirectly affects the biota, although
understory plants are much more vulnerable to flooding than overstory trees. Fire more com-
monly maintains and regenerates early-successional communities compared to wind, whereas
wind disturbance may hasten succession to more tolerant species if they are well established in
the understory. Flooding maintains those plants tolerating or requiring its effects.
Disturbances are vital to ecosystems. Ecosystems well-adapted to the frequency and intensity of a dis-
turbance regime may actually require the disturbance in order to maintain a normal, resilient system
(Waring 1989). Perturbation of such a system then typically results from the suppression or exclusion
of disturbances such as fire and flooding, often with new and unexpected consequences. Fire, for
example, is a fundamental feature of many ecosystems, from southern pine forests to subarctic black
spruce–jack pine forests. Fire suppression or exclusion in these systems becomes the perturbation to
those systems with accompanying changes in ecosystem structure, composition, and function. In
many fire-dependent systems, fire-dependent species, including many oaks and pines, fail to regen-
erate, allowing more fire-sensitive species to take advantage of the changed site conditions.
Disturbances leave both biological and physical legacies that shape subsequent forest dynamics.
Natural disturbances such as fires, floods, windstorms, avalanches, or insect and disease out-
breaks kill trees and other organisms, but they leave behind biological legacies in the form of
snags (standing dead trees) and downed logs. The patchy nature of disturbances also leaves
behind large living legacies of the pre-disturbance forest, including trees. Biological legacies help
to initiate and shape secondary succession following the disturbance. Studies of post-disturbance
regeneration have emphasized the importance of biological legacies for the rapid reestablish-
ment of forests with high diversity of composition, function, and stand structure. By contrast, old
agricultural fields and large clear-cuts offer few legacies. Physical legacies following distur-
bance include bare or rearranged mineral soil (after fires, floods, or landslides), new microtopog-
raphy resulting from tree uprooting by wind, soil movement via debris flows, and the cutting and
depositing actions of floodwater.
SOURCES OF DISTURBANCE
Where do disturbances originate? For decades, ecologists assumed that disturbances are external
and are initiated from outside the plant community (e.g., Grubb 1988). Lightning strikes a tree;
floodwaters kill plants intolerant of low O2; spongy moths (formerly gypsy moths) or spruce
budworms arrive en masse. In each of these cases, disturbances appear to originate from the
outside because of an organism or community focus. When the focus is on ecosystems rather than
organisms, disturbance frequency, extent, severity, and consequences are each closely related to
the physical site conditions where the plants are growing rather than simply the plants themselves:
the lightning-struck tree is on a ridgetop, not in the protected valley; the flooded plants are on the
first bottom of the floodplain rather than the elevated second bottom. An ecosystem context is
therefore paramount in understanding disturbance effects.
Examples are innumerable. In determining the frequency and severity of fire, site and biota
are inseparable: dry, nutrient-poor sites favor conifers with flammable foliage, whereas moist,
fertile sites favor species with rapidly decomposing leaves and reduced fuel accumulation. Wind-
throw is prevalent on sites where rooting depth is limited but where trees may grow tall and are
susceptible to wind bursts, such as sites with high water tables or hardpan development. The
interrelationship of water table and tree crowns is revealed whenever trees in or near wetlands
and swamps are cut or die from disease: the water table rises due to lessened transpiration, risking
injury or death to the root systems of other tree species. In Lower Michigan, the death of American
elms in swamp ecosystems due to Dutch elm disease in the 1960s and 70s facilitated a rise of the
water table and the subsequent death of basswood and red maple trees, initiating an even further
raising of the water table (Barnes 1976). The roots of northern white-cedar were eventually over-
whelmed as well, and only the most tolerant tree species of a high water table, black ash,
remained—until it was mostly killed by emerald ash borer in the 2000s (Kashian and Witter 2011).
Many disturbances caused by humans are more impactful to site or biota than natural distur-
bances that occurred before human occupation of North America (about 12 000 years ago). Clear-
ing of forests for agriculture drastically changed both site and biota. Widespread logging and the
repeated severe fires that followed it destroyed forest understories of many areas of the upper Great
Lakes in the late nineteenth and early twentieth centuries. These disturbances drastically changed
conditions for regeneration and hence the structure and composition of forests. Modern commercial
clear-cutting and timber extraction, followed by planting, generate simplified forests in structure
and composition, and different in function from naturally regenerated ecosystems (Franklin 1995).
Ongoing recovery of herbaceous layers in Appalachian cove forests following clear-cutting may
require centuries (Duffy and Meier 1992). Fire and flood exclusions since the early 1900s have
dramatically changed forest ecosystems, as have increasing urbanization and fragmentation of
landscapes by roads, fields, and power lines and the introduction of invasive exotic species.
In summary, disturbance is ecosystem-specific, and its kind, source, frequency, severity, and
consequences are best understood in this context. In the sections that follow, we examine several
major disturbances that affect the structure, composition, and processes of forest ecosystems.
Major Disturbances in Forest Ecosystems 387
MAJOR DISTURBANCES IN FOREST ECOSYSTEMS

The most obvious disturbances in forest ecosystems are those that alter sites or affect the forest
community by killing either the trees in the overstory or the trees and other plants in the understory.
FIRE
It was stated in the first edition of this textbook that fire is the dominant fact of forest history
(Spurr 1964). Most forest ecosystems of the world—excepting only the perpetually wet rainforests
such as those in southeastern Alaska, the coast of northwestern Europe, and the wettest belts of the
tropics—have burned at fairly regular intervals for many thousands of years. Fire continues to be a
major disturbance factor in most North American forests even under present-day conditions when
awareness of forest fires is high and forest fragmentation interrupts the continuity of fuels across
most landscapes. Organized fire suppression activities were mounted beginning in about 1900,
becoming increasingly effective after 1945 and thereby greatly reducing the number and size of fires.
The incidence of fire was quite different prior to the twentieth century. Humans from many
different societies and cultures throughout the world had until relatively recently no compunction
about burning forests, and no desire, intent, or ability to put out either human-or lightning-ignited
fires. On the contrary, fires have been set deliberately around the world for thousands of years to
clear underbrush, improve grazing, drive game, improve game habitat, combat insects, without
thought, or just for the hell of it. As more research is conducted into the ecological history of fire,
the more it is obvious that frequent burning has been the rule, not the exception, for the vast
majority of the forests of the world as far back as we have any evidence (Chapter 10).
Fires have been influential in the development of heathlands and moors of western Europe
and the British Isles, for many savannas within tropical forest belts, for high-elevation meadows
within the mountainous forests of the Americas, and, in general, for the persistence of grassland
areas on upland sites within forest regions. Dominance of many pine and oak forests around the
world is due to fire, as are the vast areas of Douglas-fir in the Rocky Mountains and Pacific Northwest
and eucalyptus in Australia. The extensive areas of spruce in the boreal forest of North America and
Eurasia are also structured to a great extent by past fires (Bloomberg 1950; Sirén 1955; Viereck 1973).
Even hemlock–northern hardwood forests of Upper Michigan, mesic and dominated by sugar
maple, have been regularly burned by fires, though very infrequently (Frelich and Lorimer 1991).
Though sweeping and perhaps overstated, these statements reflect the feeling of many forest
ecologists who have increasingly realized the great importance of fires and subsequent succession
in framing forest composition and structure (Cooper 1961; Agee 1993; Whelan 1995). Fire is clearly
a disturbance that alters the composition and structure of an existing stand, but it is also a natural
ecosystem factor whose effects have long been incorporated in species’ adaptations and ecosystem
dynamics (Chapter 10). In the following text, we highlight the importance of fire in forest ecosys-
tems by examining the similar roles it plays in many different fire-dependent systems.
Roles of Fire in Forest Ecosystems Fire has played important and similar roles in fire-
dependent conifer and hardwood forests around the world prior to significant human coloniza-
tion. Many ecosystems depend strongly on fire; existence for many forest species literally begins
and ends with fire, and thus the spatial mosaic of species and structures created by fire is important
to maintaining many forest systems. The following major functions and processes are regu-
lated by fire:
Regeneration and Reproduction: Severe fires kill all or most of the trees in existing stands (often
called “stand-replacing” fires) and begin the process of regeneration for the forest that follows. As
a selective force, fire also elicits the following reproductive characteristics of forest trees:
1. Asexual reproduction, mainly via sprouting, occurs in all fire-dependent angiosperms and
in some conifers (Chapter 10).
2. Sexual reproduction by light, wind-blown seeds is characteristic of many fire-dependent tree
species. Moreover, self-fertility in pines may have evolved as a result of fire, although only a
small percentage of self-fertilized seeds are viable in many pines. Red pine is the exception
and is highly self-fertile, producing viable seeds and seedlings showing no growth depres-
sion, unlike most other tree species (Fowler 1965a,b). A single isolated red pine that survives
a fire may therefore self-fertilize and perpetuate itself by establishing a colony of seedlings.
Preparation of Seedbeds and Dry-Matter Accumulation: Fire sharply reduces the amount of
litter, sometimes baring mineral soil, and thus greatly enhances seedling establishment. Seed
germination is favored when seeds are partially buried in the ash layer. Fire also regulates dry-
matter accumulation, influencing the severity of burning.
Reduction of Competing Vegetation: The elimination of trees, shrubs, and herbs by severe fires
favors the reestablishment of a new stand. Light surface fires at regular intervals diminish
encroaching vegetation, reducing competition for light, soil water, and nutrients. This effect is
typical in many mixed conifer–hardwood stands in the southern United States in which under-
story hardwood regeneration is reduced by periodic burning. Similarly, in oak forests, periodic
surface fires kill seedlings of shade-tolerant species that continuously establish on the forest floor
and grow into the understory.
Nutrition: Recurrent surface fires in a forest reduce organic matter to its basic components of
water, CO2, and mineral nutrients that are otherwise unavailable to forest vegetation
(Chapter 19). Nutrients made available by fire are particularly important in nutrient-poor
conifer forests and wherever site conditions are unfavorable for decomposition. Although
nitrogen is lost from soil after burning, post-fire soil conditions often favor nitrogen-fixing soil
organisms.
Thinning: Periodic surface fires reduce the density of pure stands of seedlings that establish after
fire, particularly in conifer-dominated ecosystems. Larger, faster growing seedlings with thicker
bark are favored, temporarily reducing competition for soil water and nutrients and increasing
resistance to mortality that might occur due to drought conditions that often accompany fires.
Sanitation: Dense, even-aged stand conditions may be conducive to disease and insect outbreaks.
Outbreaks that kill many trees increase fuel concentrations that in turn lead to intense fires that
terminate the outbreaks (such as those of bark beetles and the spruce budworm) and create con-
ditions for the establishment of a new even-aged stand that is initially resistant to disease and
insect attacks. The maturing stand will eventually be again susceptible to outbreaks, and a self-
perpetuating cycle is established. Fires also eliminate plant parasites such as mistletoes on pon-
derosa pine, lodgepole pine, and black spruce.
Succession: Depending on site conditions and the frequency and intensity of recurring fires, fire
tends to retain fire-dependent species on an area as long as the species are adapted to these fire
characteristics. Ponderosa pine, the southern pines, red and white pines, upland oaks, and many
other species are often regenerated generation after generation, even when they are not the most
shade-tolerant or late-successional species of their respective regions. Changes (particularly
increases) in the fire frequency or severity may result in the replacement of a given species by a
species more able to take advantage of post-fire conditions. For example, in the Rocky Mountains,
Engelmann spruce is replaced by Douglas-fir, Douglas-fir is replaced by ponderosa pine, and
ponderosa pine is replaced by grasses with decreasing elevation, due to differences in water avail-
ability as well as the frequency of fire.
In addition to differences in species composition that varied with fire frequency, a mosaic of
different age classes in each forest type was typical of many North American upland forests prior
to European colonization. This mosaic resulted in part from the erratic burning patterns of fires
and the resulting mosaic of fire severities, which in turn was controlled by heterogeneous soil-site
conditions and fuel buildup in the different ecosystems (Williamson and Black 1981). All or most
fires create a spatial mosaic of severities, from unburned or hardly burned areas to areas where all
trees are killed, resulting in a corresponding mosaic of unburned surviving trees to a gradient of
low to high tree mortality, and an eventual mosaic of tree age classes across the burned area.
Very shade-tolerant species of both hardwoods and conifers, such as sugar maple, red maple,
beech, hemlocks, firs, northern white-cedar, and western redcedar, tend to be poorly adapted to
fire. These species occupy protected or moist sites that are least susceptible to burning and did not
develop adaptations to cope with fire. During fire-free intervals, these species colonize adjacent
areas, establishing and eventually replacing fire-dependent species in the absence of fires intense
enough to kill them. Exceptions to this process exist, however: although its shallow roots make
eastern hemlock susceptible to fire, it regenerates vigorously following burns due to highly suitable
post-fire seedbeds for its tiny, wind-dispersed seeds.
Wildlife Habitat: Fire universally creates a mosaic of habitats and niches for wildlife of many
kinds (Chapter 10). Depending on how they affect and shape a particular species’ habitat, fires
clearly benefit some wildlife species more than others. However, overall animal species diversity,
as with plants, tends to increase following fire until crown closure occurs, and then declines.
In the following text, we present a few of the many examples that demonstrate these roles and sig-
nificance of fire. This selection represents important North American trees and regions as an effort
to illustrate the importance of fire in forest ecosystems.
Pines in New England and the Lake States In the northeastern United States and adjacent Canada,
the occurrence of the two- and three-needled pines (red, jack, and pitch pines) as well as that of
pure even-aged stands of white pine (not to be confused with individual white pine in mixed
stands) is strongly related to the past occurrence of forest fires. During the early colonization
period, many of the fires resulted from burning by European colonists in land-clearing operations,
but much evidence suggests that fires were commonly set by indigenous people for many hundreds
of years before this period (Little 1974; Denevan 1992; Munoz et al. 2010; Pinter et al. 2011; Marlon
et al. 2013; Abrams and Nowacki 2015, 2019). Lightning apparently played a lesser role in causing
fires in eastern forests compared to the West, probably because of the heavy precipitation that often
accompanies summer thunderstorms in the East.
Many ecological studies of relict old-growth pine stands have shown that fire was impor-
tant in their formation. In both northwestern Pennsylvania (Lutz 1930; Hough and Forbes 1943)
and southwestern New Hampshire (Cline and Spurr 1942), more or less pure even-aged stands
of old-growth eastern white pine originated from past forest fires. By contrast, nearby mixed
forest types with occasional dominant white pine were relatively free from evidence of past
burns (Abrams 2001). Recent studies suggest that white pine is favored by a mean interval bet-
ween severe fires of 150–300 years (Frelich 1992); longer fire intervals allow mesic hardwoods to
invade such that fires are less common, and shorter fire intervals shift the dominant vegetation
toward paper birch or aspen (Frelich 2002). In northwestern Minnesota, extensive even-aged
stands of old-growth red pine clearly date from a series of forest fires, many of which antedate
the advent of the colonists (Spurr 1954; Frissell 1973). In Canada, the abundance of red pine was
found to be directly related to fire intensity and was most favored by moderate-intensity fires
rather than severe crown fires (Flannigan 1993). In many regions of the northern Lake States,
frequent (25–35 years), low-intensity fires supported even-aged stands of red pine (Drobyshev
et al. 2008).
Jack pine in the Lake States and Canada and pitch pine on sandy soils near the mid-Atlantic
coast (Little 1974) are virtually completely fire-dependent (Chapter 10). Jack pine grows in nearly
pure stands on dry, sandy soils, forming a highly flammable vegetation type. Severe crown fires burn
jack pine forests during hot, dry periods at intervals of a few decades, killing all or most of the vege-
tation in a patchy distribution. Following fires, the subsequent generation of trees arises from multi-
ple sources. Jack pine regenerates from seeds stored within many years’ serotinous cones in the tree
crowns, and thereby benefits from severe crown fires. Red pine regenerates from seeds of trees with
bark of sufficient thickness and lacking enough lower branches to permit them to survive surface
fires. Hardwoods such as bigtooth and trembling aspen, northern red oak, and red maple regenerate
by sprouts arising either from roots or the lower part of the stem not killed by the fire. Some pioneer
hardwoods regenerate by the dissemination of seeds into the area from afar either by wind for aspens
and birches, or by birds for the cherries. Infrequently, some hardwoods such as pin cherry and black
cherry regenerate from seeds stored in the seed bank of the forest floor (Marks 1974).
Western Pines and Trembling Aspen The relationship between fire and pines in the northeastern Unit-
ed States is applicable to most of the other pine species in the United States. Lodgepole pine in the
Intermountain West plays an analogous role to that of its close relative, jack pine, in the Lake States. It
regenerates on recent burns, largely from seeds stored in serotinous cones of trees killed by crown fires
or severe surface fires, to form dense, even-aged, post-fire stands (Figure 16.2). In 1988, fires burned
(a) F I G U R E 16. 2 Lodgepole pine

Photos by Dan Kashian
regeneration following fire in

Yellowstone National Park. (a)
Lodgepole pine revegetation in a
burned patch approximately 20 years
after fire. Seedling density in this patch
is nearly 500 000 stems per hectare,
with height averaging only 1–1.5 m.
(b) Lodgepole pine thickets develop
following fire when initial seedling
regeneration is dense. This
130-year-old stand of about 5400 stems
per hectare in Yellowstone averages
only 10 m high and 9 cm in diameter at
breast height. The numerous naturally
thinned dead trees suggest the stand
density was at least 11 000 stems per
(b) hectare within the last few decades.
570 000 ha of the Yellowstone Plateau in northwestern Wyoming, including approximately 240 000 ha
of forested land in Yellowstone National Park (Schullery 1989). The fires began in mid-July and were
essentially uncontrollable until snow and cool temperatures arrived in mid-September (Ellis
et al. 1994). Regeneration of lodgepole pine at varying levels of density blanketed the area in succeed-
ing years in patterns related to fire severity, pre-fire stand density and cone serotiny, and site conditions
(Kashian et al. 2004; Turner et al. 2004).
The Yellowstone fires of 1988 raised important questions as to whether the Park’s policy of
fire suppression from 1872 to 1972 led to an abnormal buildup of fuels and therefore to abnormal
fire spread, intensity, and severity. Romme and Despain (1989) concluded that the 1988 fires should
not be viewed as an abnormal event for this area, and were in fact similar in total area burned, rate
of spread, and severity to those that occurred in the early 1700s. Crown fires similar in extent and
behavior probably occurred on this landscape every 100–300 years and are driven primarily by
weather conditions rather than fuels. Although a century of fire suppression may have had some
influence on size and behavior of the 1988 fires, suppression is difficult and generally ineffective in
forests characterized by crown fires, and the 1988 fires had undergone normal successional
dynamics and fuel buildup following the last major fires approximately 280 years ago. Notably, Yel-
lowstone National Park has experienced several years of large area burned (>15 000 ha) since 1988,
including 1994, 2003, 2007, and 2016 (https://famit.nwcg.gov/applications/FireAndWeatherData,
accessed August 19, 2021).
Depending on site conditions, lodgepole pine forests are gradually replaced by shade-
tolerant Engelmann spruce and subalpine fir if the period between fires reaches several hun-
dreds of years. White and black spruce also play a part as late-successional species in the
northern range of lodgepole pine in Alberta, and Douglas-fir and other more shade-tolerant
western conifers become prominent toward the Pacific Northwest. In subalpine and high foot-
hills in Alberta, Horton (1956) estimated that succession from pine to spruce and fir will occur
if free from fire from 225 to 375 years, but it takes much longer on drier southern slopes, if
indeed it ever takes place.
Pure stands of ponderosa pine result from a long and complex fire history at lower elevations
and in warmer, drier portions of western forests (Figure 16.3; Cooper 1960; Fischer and Brad-
ley 1987). Ponderosa pine is a pioneer species following fire in the cooler and moister portions of
its range and is gradually replaced by more shade-tolerant conifers such as Douglas-fir, incense
cedar, and white fir. In the warmer and drier parts of its range, those species cannot generally out-
compete ponderosa pine and it dominates the forest. Fire regimes and stand development of pon-
derosa pine are discussed for a variety of regional and local ecosystems by Fischer and Bradley
(1987), Agee (1993), Covington and Moore (1994), Arno et al. (1995), Veblen et al. (2000), and
Baker and Ehle (2001), and are reviewed in detail in Chapter 10.
Trembling aspen is one of the few hardwood species able to coexist with conifers in the
fire-dominated western mountains. In the northern Rocky Mountains and in western Canada, it
competes as a pioneer via wind-disseminated seeds on higher, cooler, and wetter sites. In warmer
and drier regions such as the Great Basin and the central and southern Rocky Mountains, aspen
persists mainly via root suckers and may be maintained as a late-successional type by fire in some
localities (Mueggler 1985). In contrast to the eastern and northern parts of its range, aspen clones
of the central and southern Rockies may become very large; a single aspen clone in Utah is 43 ha
in extent and contains thousands of genetically identical stems (Kemperman and Barnes 1976;
Mock et al. 2008; Chapter 4, Figure 4.14). Some researchers have speculated that individual
clones may have been perpetuated for thousands of years following rare events of seedling estab-
lishment by a combination of fire and low ungulate browsing intensity (Barnes 1975; Kay 1997).
However, recent molecular analyses suggest that many very small, even single-stem clones exist
in and among the larger clones, meaning that sexual reproduction (by seed rather than root
suckers) is not a negligible process and likely contributes to aspen persistence (Mock et al. 2008).
F I G U R E 1 6 .3 Fire and ponderosa pine. (a) A surface fire burns grass and litter of this ponderosa pine stand
in central Idaho. (b) Fires prior to European colonization time maintained open, grassy, parklike stands of
ponderosa pine such as this one in western Montana. Source: U.S. Forest Service photos.
In the moister northern Rocky Mountains, aspen seedlings may establish readily following fire
in appropriate microsites. Seedling establishment is also possible, although rare, further south in
the species’ range given a very narrow window of conditions dictated by sustained moist weather
conditions immediately following the fire. For example, following the major fires in Yellowstone
Park in 1988, thousands of aspen seedlings were observed over a broad range of vegetation types
(Renkin et al. 1994; Turner et al. 2003a; Romme et al. 2005), in part because the growing season
in 1989 was unusually cool and wet and allowed extensive seedling establishment.
Many aspen clones of the central and southern Rocky Mountains are now deteriorating.
Concern about aspen decline in the western United States, which features clones characterized by
aging mature trees without regeneration, began in the 1970s (Schier 1975) and reached its height
beginning in the 1990s (Kay 1997). Aspen decline was most evident at local scales (Kashian
et al. 2007), and was attributed primarily to fire suppression and excessive ungulate browsing
(Baker et al. 1997; Kaye et al. 2005; Kulakowski et al. 2013). In the 2000s, aspen in the West began
to experience a phenomenon termed sudden aspen decline, characterized by an episode of
unusually high overstory crown thinning and mortality (in addition to the previously recognized
lack of regeneration) that occurs at broad scales over several years or more (Worall et al. 2008, 2013).
Sudden aspen decline has been linked to drought-related moisture stress but is also associated
with multiple interacting factors including climate, land-use history, and successional dynamics
(Singer et al. 2019).
Southern Pines The dependence of pine forests upon recurring fires is most evident in the southern
pine belt of the southeastern United States. Of the four most common southern pines, longleaf pine
depends the most upon recurring surface fires for its persistence (Chapman 1932). Longleaf pine dom-
inated much of the Coastal Plain prior to European colonization (Christensen 1988a) but is now con-
sidered an endangered forest type in the United States (Noss et al. 1995). Pre-European colonization
longleaf pine forests occurred on three fairly different kinds of sites—dry sandhills, intermediate
areas, and wetlands that included shrub-bogs, savannas, and flatwoods—each with different fire
effects and successional pathways (Marks and Harcombe 1981). Longleaf pine regeneration is pre-
vented by annual burning and summer fires, and thus the proper ratio of winter fires to fire-free years
is critical if pure longleaf pine forests are to persist. Furthermore, longleaf pine stands require burning
during the “grass stage” to control brown spot disease, as well as repeated burning following canopy
formation to avoid the establishment of and replacement by understory hardwoods in the absence of
fire. After many attempts at complete fire exclusion by southern foresters in the first half of the twen-
tieth century, prescribed burning has been heavily used to achieve these ecological objectives.
Longleaf pine is extremely shade-intolerant and regenerates most successfully in full sun-
light on mineral seedbeds created by forest fires that destroy and/or consume all but scattered
longleaf seed trees. It is one of the most fire-resistant of all pines. Seedlings appear like dense
bunches of grass (Figure 16.4), and this stage is known as the grass stage. Longleaf pine seedlings,
unlike all other North American pines, remain in the grass stage for about 6 years (range 3–12 years)
while they develop a deep tap root. A surrounding sheath of dense needles in the grass stage pro-
tects the bud from fire, and if burned away are replaced by a new set from stored carbohydrates and
nutrients in the large tap root. Rapid stem elongation, called “bolting,” eventually occurs, quickly
elevating the crown above the level of periodic surface fires.
The other common southern pines—loblolly pine, slash pine, and shortleaf pine, as well as
sand pine, Virginia pine, pitch pine, and other species of local occurrence—are pioneer species
that establish after fires, giving way to shade-tolerant hardwood mixtures in their absence (Little
and Moore 1949; Wahlenberg 1949; Campbell 1955; Myers 1990). The dependence of these species
upon fire has given rise to practical use of fires as a silvicultural tool in these systems that will
maintain natural succession in the pioneer pine stage or to restore forests after long periods of fire
suppression (Stanturf et al. 2002; Ryan et al. 2013).
F I G U R E 1 6 .4 A dense stand of natural longleaf pine seedlings in the Coastal Plain of South Carolina
following a prescribed burn to prepare the seedbed. Source: U.S. Forest Service photo.
Douglas-Fir in the Pacific Northwest Characteristic summer droughts in the Douglas-fir region of
northern California, western Oregon and Washington, and southern British Columbia create
highly flammable conditions during the hottest period of the year. As a result, extensive forest fires
in the region have been common since pre-European colonization. Douglas-fir is a pioneer species
following fires on the west side of the Cascades and over much of the Coastal Range, so long as
adjacent unburned Douglas-fir stands are able to supply a source of seed. More shade-tolerant
conifers, such as western hemlock, Sitka spruce, and western redcedar, later colonize the under-
story of Douglas-fir forests and may eventually dominate after 500 or more years when the domi-
nant Douglas-fir matures and senesces (Figure 16.5; Franklin et al. 1981). Many fire-history studies
of Douglas-fir forests suggest that stand-replacing fires occurred every 200 to 400+ years, but
others reconstruct a regime of multiple patchy fires of varying severity recurring every 20–100 years
(Morrison and Swanson 1990). Fire history and stand development of Douglas-fir throughout its
Pacific Northwest range are discussed by Agee (1993). An excellent example of a fire history study
in Douglas-fir forests is provided by Taylor and Skinner (1998), and Perry et al. (2011) provide a
treatment of mixed-severity fire regimes in the region.
Giant Sequoia The native giant sequoia-mixed conifer forests of the Sierra Nevada of California
are also fire-dependent (Kilgore 1973; Agee 1993; Swetnam 1993). Fire in sequoia forests prepares
a soft, friable, mineral seedbed upon which the tiny seeds of sequoia fall and are lightly buried,
favoring germination (Hartesveldt and Harvey 1967). Seedling establishment is limited by water
availability (Rundel 1972), and litter that is partially burned may contain more available water than
unburned litter (Stark 1968).
Periodic surface fires, some of which were presumably set intentionally by indigenous people,
maintained an open and parklike setting prior to European colonization (Figure 16.6a; Biswell 1961).
Fires in these ecosystems probably burned every 15 years (Swetnam et al. 2009). Such fires eliminated
small shade-tolerant trees, particularly white fir, which continually colonized the understory. Burning
by indigenous people was gradually eliminated, and fire suppression became increasingly efficient
F I G U R E 1 6 .5 The dominant, old-growth Douglas-fir individual (left) in this western Washington stand
will eventually be replaced in the absence of fire by more tolerant firs and western hemlocks in the under-
story. Source: U.S. Forest Service photo.
upon European colonization. As a result, surface fuels built up, and white fir and other shade-tolerant
species often formed thickets under the sequoias (Figure 16.6b). A vertical sequence of “ladder fuel”
was thus formed from ground level to low-hanging fir branches to the top of understory crowns, 10–
30 m above the surface (Kilgore and Sando 1975), such that fire in surface fuels was more likely to
pass through understory crowns and reach sequoia crowns. Unlike redwood, giant sequoia is a poor
sprouter and will die once the crown is killed. Fires have occurred more severely and more frequently
in these forests since 2015 (Stephenson and Brigham 2021); high-severity fires in 2020 and 2021 are
estimated to have killed up to 15% of all existing large sequoias across its range (Stephenson and
Brigham 2021). Prescribed burning every 5–8 years is needed to reduce surface fuels, remove small
understory trees and the lower crowns of large understory trees, and prevent continued encroach-
ment of a shade-tolerant understory.
Fire History and Behavior Fire-history studies in forests increase our awareness of fire as a
dominant force over virtually all of the northern and western landscapes. Such studies are fantas-
tically abundant in the literature. Three classic studies illustrate fire history, behavior, and effects.
Northern Lake States Heinselman’s (1973) classic study in the Boundary Waters Canoe Area in
northern Minnesota documents 400 years of fire occurrence and effects on a landscape where fire
has driven forest composition and structure for nearly 10 000 years. Recurrent fires had maintained
nearly three-fourths of the region in recent burns and stands of small-diameter trees prior to the
arrival of logging in about 1895. Thus, relatively little of the 215 000 ha tract, now reserved as
wilderness, was subjected to the predominant harvesting method of the time: “high-grading” (cut
the best and leave the rest). Effective fire control began about 1911.
Heinselman developed a fire chronology based on the age of existing stands, tree-ring counts
from sections or cores from fire-scarred trees, and historical records. Nearly all of the forest burned
one to several times between 1595 and 1972, but much of the burned area is accounted for by just
a few major fires. In the European colonization period, fires in 1875 and 1894 accounted for 80% of
F I G U R E 1 6 .6 1890–1970: 80 years of fire exclusion in the giant sequoia-mixed conifer forests of Yosemite
National Park (Confederate Group, Mariposa Grove). (a) 1890: A parklike stand as the result of periodic fires.
Source: Kilgore 1972; Historical documentation by Mary and Bill Hood. Photo by George Reichel. Courtesy
of Mrs. Dorothy Whitener. (b) 1970: The parklike stand was invaded by thickets of white fir. By 1970 firs
obscured all but the fire-scarred sequoia on the left. Such thickets provide ladder fuels that could support a
crown fire fatal even to mature sequoias. Source: National Park Service photo by Dan Taylor.
92°20' 92°00' 91°00' 90°10'

Lac La Croix Lake
48°20'
CA Saganaga Lake
NAD 1863–1864
A
48°10'
Bo
un
da
ry Area
48°00'
Waters
Canoe
Ely
Lake
47°50'
Superior
0 8 16 Miles
Grand Marais
0 12 24 Kilometers Study area
F I G U R E 16.7Severe fires of 1863 and 1864 burned (shaded area) over 1800 km2 (700 mi2) of the Boundary
Waters Canoe Area, northern Minnesota. Source: Heinselman (1973) with permission of Cambridge University
Press. Reprinted with permission of the Quaternary Research Center, University of Washington, Seattle.
the total area burned; but just five brief fire periods accounted for 84% of the total prior to European
colonization. These major fire periods coincided with prolonged, subcontinental-scale summer
droughts. Climate combined with fuel accumulations (driven by dry matter accumulation, spruce
budworm outbreaks, and blowdowns) to create optimum burning conditions at long intervals. The
fires of 1863 and 1864, associated with the major droughts of those years, burned 44% of the area
(Figure 16.7). Fire in 1864 also burned over 400 000 ha in Wisconsin.
Heinselman estimated the natural fire rotation (the average number of years necessary to
burn the entire area) since 1595 to be about 100 years, although the natural rotation for different
ecosystems varied. For example, ecosystems dominated by aspen–birch forests burned more
frequently (50 years or less) than those dominated by red and white pines (150–350 years). In red
and white pine ecosystems, fire typically occurs as light surface fires that reduce the organic layer
and understory competition, with occasional severe fires that kill the overstory and provide open-
ings for regeneration. A second severe fire that occurred only a few years after the first could
eliminate many conifers, being too small to withstand the heat and needle scorch and too young to
bear cones. Regeneration in that scenario comes from hardwood sprouts of trembling aspen, paper
birch, red maple, and northern red oak.
Fires over the entire area produce a mosaic of stands of different composition and ages
driven by the complex of site factors and fire size and severity, generating diversity in stand struc-
ture and species composition. Both plants and animals are adapted to these ecosystems and
community patterns and are directly or indirectly adapted to fire. Moose, beaver, black bear,
snowshoe hare, woodland caribou, small mammals, and birds are all adapted and dependent on
the mosaic of different habitat conditions created by fire. Notably, widespread fire exclusion is now
restructuring the entire regional ecosystem, gradually creating changes in forest composition and
eliminating the niches of many wildlife species. Restoring the natural vegetation mosaic in this
now human-dominated wilderness will be difficult but will surely depend on the use and
management of fires.
Boreal Forest and Taiga Wildfires burning over diverse sites create a vegetation mosaic in the
boreal forest and taiga of North America and have a major influence on the region’s plant and
animal diversity (Lutz 1956; Slaughter et al. 1971; Rowe and Scotter 1973; Viereck 1973, 1983;
Bonan and Shugart 1989). The patchy pattern of vegetation is maintained by fires and provides
diverse wildlife populations. Seven of the ten major boreal tree species are pioneers on disturbed
sites and have adaptations that favor rapid invasion after fires (Chapter 10). Only balsam and sub-
alpine fir, whose cones disintegrate at maturity, and whose seeds are not retained in the crown, are
not well adapted for immediate post-fire regeneration. The understory and groundcover layers of
the boreal forest are dominated by clonal species whose underground stems and roots are protected
from fire and regenerate rapidly by sprouting.
The taiga lies north of the boreal forest and is a broad, open area of slow-growing spruces
interspersed with scattered, well-developed stands of trees and treeless bogs, locally called mus-
keg (Viereck 1973; Van Cleve et al. 1986). Vegetation pattern is closely related to fire history and
permafrost, or permanently frozen ground with water often incorporated as lenses of pure ice.
Permafrost perches the water table and thus cools the ground by creating a deep, long-lasting
ice layer on undisturbed sites (Bonan and Shugart 1989). Thick layers of moss insulate the
forest floor during summer months, limiting soil thawing to depths of less than 1 m (termed the
active layer). Fire may increase this annual depth of thaw via removal of this insulating layer
and other organic matter; historically the upper permafrost layers recover as coniferous vegeta-
tion regrows. In many areas, permafrost and vegetation are in delicate balance and altering the
depth of summer thaw can significantly affect the presence and composition of the vegetation.
Increasing the depth of thaw following fire may cause the areas over ice lenses to subside,
creating a polygonal mound-and-ditch pattern. This pattern has been documented in paper
birch stands at least 40–50 years after fire. These sites may succeed to black spruce and stabilize
as permafrost reforms, or small ponds may develop and alternate in the cycle with black spruce.
Northern Rocky Mountains Fire has a major ecological influence in the semi-arid, mountainous,
coniferous forests of northern Idaho, western Wyoming and Montana, eastern Washington, and
adjacent portions of Canada (Habeck and Mutch 1973; Arno 1976; Agee 1998; Baker 2009;
McKenzie et al. 2011; Pyne 2019). Fire in this region (as well as insect outbreaks, discussed in the
following text) was second only to precipitation as the major factor shaping forests until fire sup-
pression and exclusion became effective in the 1940s. Large stand-replacing fires were common
due to the extremely dry summer climate and the buildup of fuels in areas that were densely for-
ested. Fuel buildup was exacerbated by forest insects and diseases, which killed many trees in
dense stands over extensive areas. Between 1 and 2 million hectares burned in 1910 alone follow-
ing prolonged drought.
Many forests in the regions may be “reburned,” or burned multiple times at short intervals.
In some cases, the first fire will reduce the severity of a second fire, if it occurs, by consuming fuels
and reducing the abundance of fire-sensitive vegetation (Parks et al. 2014). However, where the
first fire results in abundant growth of flammable vegetation or accumulation of hazardous fuels,
the probability and severity of the second fire may increase. Famously, parts of the 1910 fires
reburned in either 1917 or 1919, or both, and even again in subsequent years. Snags were particu-
larly important for ignition by lightning, and wind-blown embers from tall snags could start new
fires far in advance of the burning front. A recent study by Harvey et al. (2016a) in northern Idaho
and western Montana and Wyoming found that initial high-severity fires facilitated high-severity
reburns when the interval between fires was longer than a decade. Successive stand-replacing fires
were most common in high-elevation forests where abundant tree regeneration tends to follow
burning. Coastal systems in the West tend to reburn at shorter intervals between fires when con-
ditions are wetter, and interior forests when conditions are drier, suggesting that burn intervals are
largely determined by vegetation productivity and the accumulation of fine fuels (Buma
et al. 2020).
Unstoppable, severe fires occur in regional ecosystems that are characterized by the
climatic and vegetational features conducive to crown fire. An excellent example with a long,
documented fire history is the Yellowstone Plateau that includes Yellowstone National Park
mentioned in the earlier text (Romme 1982; see also Chapter 22). Most of the Yellowstone
Plateau is above 2300 m (7500 ft) in elevation, and most of the forested area is dominated by
relatively even-aged lodgepole pine forests between 100 and 350 years old. The 1988 Yellow-
stone fires burned in one of the driest summers on record in the Rocky Mountains with
extremely dry fuel conditions. In addition, twice as many lightning storms were recorded in
1988 than in the average year (Wuerthner 1988). Following the fires, lodgepole pine regenera-
tion density varied across the burned area from zero to 535 000 stems per hectare (Turner
et al. 1994), with variation explained mostly by elevation and, secondarily, distance to unburned
areas. An earlier study of lodgepole pine regeneration in moderately burned, severely burned,
and unburned stands found the highest seedling densities in moderately burned areas (Ellis
et al. 1994). Where burn severity was high, seeds stored in serotinous cones in the tree crowns
were likely consumed by the fire, resulting in relatively low post-fire seedling densities. The
varying or patchy pattern of burn severity allowed seeds from lightly or moderately burned
areas to supply much of the seed source for adjacent severely burned areas. Local levels of cone
serotiny, in addition to burn severity, were also likely to have affected post-fire lodgepole pine
regeneration (Turner et al. 1997, 1999).
Fire is also a key factor in other parts of the Rocky Mountains and has been examined in
detail by Peet (1981), Veblen (Veblen and Lorenz 1986, 1991; Veblen et al. 1991a; Veblen 2000),
Baker (2009), McKenzie et al. (2011), and Knight et al. (2014), among many others.
Fire Suppression and Exclusion At a temporal scale of centuries, fire is an integral part of
the evolution and perpetuation of forest species as parts of landscape ecosystems. It exists as an
internal factor to ecosystems rather than an external disruptive force that ecologists once consid-
ered to be a “disturbance.” From this perspective, increasingly efficient alteration of the timing,
frequency, intensity, or severity of fire via its suppression or exclusion by humans, if continued
indefinitely, constitutes a perturbation that will greatly change ecosystems (Smith et al. 1997).
Many such changes are already evident.
Fires were common before European colonization, set by lightning and indigenous people,
and remained common during colonization, even increasing in frequency in many places. Unless
they threatened human life, livestock, or other property, fires were little regarded except as a local
nuisance. Public opinion arose against fire in the early 1900s, fueled by unrestrained burning and
the large destructive fires of the late 1800s and 1910. Such public opinion led to the development
of rigid suppression policies whereby fire was considered a disastrous threat to the forest and to
human life and property rather than a natural phenomenon. Even today the public is exposed to
verbiage about fires in forests too often described as “catastrophic,” “destructive,” or “devastating”
without context of its ecological importance. Fire damage can be seen immediately and thus
encouraged anti-fire “propaganda;” its beneficial effects, on the other hand, might take years to
materialize (Harper 1962).
Today, technological advances and a century of experience have led to efficient fire control,
making it increasingly possible to extinguish fires very early in their development. Effective sup-
pression is not possible in all forest types, particularly where crown fires dominate the fire regime
such as the boreal forest and high-elevation forests in the Rocky Mountains, because fires in
these regions are generally driven by extreme weather conditions. Even crown fire-dominated
forests have experienced a significant reduction in the area burned since the middle of the twen-
tieth century, however, often increasing the incidence of wind damage and insect and disease
attacks as the proportion of mature and late-successional forest increases and spatial connectiv-
ity of those forests is high (due to a loss of the fire-created mosaic of the past). Where surface fires
are dominant, fire suppression is more easily accomplished, increasing the intervals between
fires and permitting a buildup of fuel and a decline in animal and plant diversity. Prolonged
buildup of fuel over longer fire-free intervals in these systems often leads to severe fires, even
shifting surface-fire regimes to crown fire regimes, resulting in fires more impactful to the
ecosystem than what occurred in the original forest, and at intensities for which the ecosystem
may not be adapted.
Many federal agencies and other organizations in the United States, such as the National
Park Service, the United States Forest Service, and The Nature Conservancy, have instituted fire
management programs in recent decades to allow fires to resume their natural role in forest eco-
systems, where possible. Prescribed burning is an important tool in restoring fire to ecosystems
having experienced fire suppression. Fire suppression and exclusion permit some thin-barked,
slow-growing species to grow into size classes that are fire-resistant (Harmon 1984), such that res-
toration of the original fire interval alone will not restore the original forest composition and struc-
ture. Thus, prescribed burning is often combined with efforts such as thinning or other harvesting
methods to reduce stand density in restoring formerly fire-prone forest ecosystems. Fire
management and prescribed burning have been considered from many standpoints and in relation
to many different forest ecosystems (Slaughter et al. 1971; Wright and Heinselman 1973; Kayll 1974;
Kilgore 1975, 1976a,b,c; Mutch 1976; Martin et al. 1977; Arno et al. 1995; Knight et al. 2014; North
et al. 2015; Thompson et al. 2018).
Because fire is an important ecosystem process, it is also an ecosystem-specific process and
thus fire suppression or exclusion has varying effects depending on the physical site factors and
forest species that make up the ecosystem. Perhaps the most well-known example of the negative
effects of fire suppression is found in low-elevation ponderosa pine forests in the Rocky Mountains
(Covington and Moore 1994; Veblen et al. 2000). These forests were historically characterized by
short-interval surface fires that maintained low fuel availability, minimal undergrowth, and low
stand densities. However, the relatively light surface fires were easily and successfully suppressed
following European colonization. As a result, fire suppression and exclusion had major effects
upon these forests, allowing fuel buildup, increased tree density, and altered tree species composi-
tion, as well as high-intensity fires when fires do occur. Suppression-induced increases in fuels and
tree density were blamed for several major fires in the West in the 2000s. In response, some recent
forest management in the West—often regardless of forest type—has focused upon proactively
reducing fuels and stand density as an attempt to reduce fire intensity. However, increasing evi-
dence suggests that ponderosa pine forests varied in structure prior to European colonization, with
some naturally dense and burned by crown fires rather than light surface fires (Williams and
Baker 2012). These data further support the idea that most fires consist of a mosaic of burn sever-
ities ranging from low to high, making proactive fuels management difficult to carry out appropri-
ately. Moreover, high-elevation forests in the Rocky Mountains dominated by lodgepole pine or
spruce and fir were historically characterized by dense stands, long fire intervals, and crown fires
that generally occur during prolonged dry periods. Such crown fires were limited by weather con-
ditions rather than by fuel accumulations and are very difficult to suppress, such that the effect of
suppression in these forests is restricted primarily to reducing the area burned. It therefore makes
little ecological or practical sense to attempt to reduce fire severity by altering forest structure,
reducing fuels, or incorporating prescribed burning in high-elevation forests as would be appropri-
ate in ponderosa pine forests. These differences again highlight the site-and species-specific nature
of fire as an ecosystem process (Schoennagel et al. 2004). Effects of fire suppression on western
forests are summarized in recent reviews and studies by Hagmann et al. (2021), Haugo et al. (2019),
Keeling et al. (2006), Keane et al. (2002), and Arno et al. (2000), among others.
WIND
Wind disturbances have enormous consequences in forest ecosystems, affecting forest structure,
composition, and function at regional and local scales. The occurrence of wind disturbances is
much more difficult to identify and track compared to more dramatic disturbances such as fire,
and thus its prevalence in forests across the world is probably understated. Like fire, wind
disturbances occur within a particular regime at a frequency, severity, and size specific to a given
ecosystem or landscape. Varying regimes have varying effects as wind encounters forested
landscapes. In this section, we examine the ecological effects of wind in forest ecosystems.
Widespread and Local Effects Certain regional ecosystems of North America and the
world are characterized by distinctive wind patterns. Hurricanes occur predictably along the Gulf
and Atlantic coasts of North America, where sustained winds may reach velocities of over
200 km h−1 (1989 hurricane Hugo, 216 km h−1; 1992 hurricane Andrew, 242 km h−1; 2005 hurricane
Katrina, 280 km h−1). Hurricanes may cause forest destruction over huge swaths across a landscape
(50-km wide for Andrew; Pimm et al. 1994). Tornadoes are most prevalent in the central part of the
United States with “tornado alleys” extending from Texas to Nebraska and Kansas east across
central Indiana. Tornadoes have higher wind velocities than hurricanes, but their damage is more
localized (typical width is 100–200 m, maximum width 2 km). Wind speeds of >430 km h−1 were
reported for the 1993 Kane tornado in northwestern Pennsylvania (Peterson and Pickett 1995),
although only about 3% of tornadoes have winds that exceed 320 km h−1. Straight-line winds in the
form of downbursts during severe thunderstorms are common in the Lake States (Canham and
Loucks 1984). A single thunderstorm may produce multiple downbursts as it moves across the
landscape; this family of downbursts is known as a derecho. A derecho that occurred in northern
Wisconsin in 1977 included 26 separate downbursts and created a damage path 20-km wide and
200-km long across the forested landscape (Frelich 2002).
As opposed to wind events that affect large portions of a landscape, local wind gusts (90–150 km h−1),
associated with cyclonic storms that break off or uproot one or more trees, may affect only small areas
(20–100 m2) in the forest. These canopy gaps or exposed patches on the ground occur over vast areas in
North America, and their accumulated ecological significance may be equal to or greater than that
caused by large wind events (Bormann and Likens 1979). Notably, winds capable of causing forest
disturbances occur over a broad range of velocities, and they encounter regional and local ecosystems
with many different site conditions and vegetation at different times of the year. Despite the many gra-
dients, the nature and amount of wind disturbance are somewhat predictable by regional ecosystem,
specific landform–soil drainage conditions, site exposure, and tree species and height.
Principles of Wind Damage Wind is turbulent near the Earth’s surface due to the nature of
the ground over which the air is moving, with higher turbulence along rougher surfaces. Forests
form a particularly rough surface such that wind over them is much more turbulent than over a
field. Turbulence is organized into coherent gusts that move widely across the forest (Figure 16.8).
Wind becomes unstable as it moves across a forest, and each gust consists of a rapid increase in
wind speed and a downward movement of air into the canopy (see #5 in Figure 16.8). The stron-
gest gusts exert a force on trees up to 10 times larger than does the mean wind velocity (Quine
et al. 1995), making them the most consistent cause of windthrow in forests.
Changes in vegetation height or composition induce additional turbulence and wind
acceleration. Forest edges, roads, and other open areas that cause wind to deflect upward tend to
increase damage to trees. In addition, the force of the wind on trees nearer to a forest edge is sub-
stantially greater than that on interior trees, and trees that grow on edges have significantly higher
physical resistance to wind. By contrast, trees suddenly exposed by the formation of a new edge or
the opening of a patch lack such a physical resistance and are more vulnerable to wind damage
than those that grow on a forest edge. The edges between cutover and uncut areas tend to deflect
wind currents, particularly at high elevations in mountainous areas, producing increased wind
velocities where the deflected currents join together (Figure 16.9). Wind damage in mountainous
forested country is often due to local acceleration of wind by either topography or forest edges
(Gratkowski 1956; Alexander 1964).
u(z)
h
x
5. Break up and
production of
smaller-scale
turbulence
4. Streamwise vortices
y
3. Instability
2. Transverse vortex
1. Kelvin–Helmholtz waves
F I G U R E 1 6 .8 Idealized formation of coherent gusts over a forest. (1) The rapid change of wind speed (u) at
the top of the canopy (z = h) is unstable and leads to the emergence of Kelvin–Helmholtz waves. (2) The
waves become transformed into across-wind vortices. (3) These vortices are unstable and begin to distort.
(4) The distortion produces coherent gusts aligned in the direction of the wind. The gusts propagate across
the forest and if strong enough, lead to wind damage. (5) Eventually the gusts become distorted and break
up. Source: Finnigan and Brunet (1995) / Cambridge University Press. Reproduced with permission of the
British Forestry Commission.
N
Uncut
Cutover
Unaccelerated winds
Accelerated winds
WIND
Turbulent area
Uncut
F I G U R E 1 6 .9 Windthrow of standing trees following a harvest cut. A change in the direction of the cutting
boundary that was parallel to the prevailing winds acts to funnel wind into standing timber, causing a pocket
of blowdown. Routt National Forest, northwestern Colorado. Source: Alexander (1964) / with permission of
Oxford University Press. Reprinted with the permission of the Society of American Foresters.
Topography has a strong influence on wind damage to forests. Wind velocity is often assumed
to be at its highest near the tops of slopes and ridges, but this is the case only when the wind
direction is perpendicular to the slope. When the wind direction is at a sharp angle to the slope, the
wind velocity is highest at the midslope position, and at the valley floor when the wind is parallel
to the slope (Frelich 2002). Local wind acceleration can also take place on gradual and smooth lee
slopes during severe windstorms, in gaps and saddles of main ridges, and in narrow valleys or V-
shaped openings in the forest that constrict the wind flow. Forests subject to hurricanes or other
windstorms where wind direction is lateral and in the same direction often experience damage
related to slope aspect (Chapter 22; Boose et al. 1994). Where damaging winds result from vertical
downbursts, however, there may be little correspondence between wind damage and topography
(Frelich and Lorimer 1991).
Windthrow (uprooting of trees) and snapping is most likely where air currents are concen-
trated to form high wind velocities at a particular spot. Windthrow most often occurs in shallow-
rooted species on shallow or poorly drained soils. Soil provides tree stability, and thus a tree whose
root system is in contact with more soil mass is less likely to be uprooted by wind. Rooting depth,
rather than lateral root spread, is strongly determined by soil conditions. Rooting depth is most
often restricted by inadequate oxygen supply (poorly drained soils) or impenetrable layers (bed-
rock, ortstein, etc.). Tall trees with large crowns are especially susceptible to wind damage, and
conifers are more vulnerable than deciduous trees in winter, when deciduous trees are leafless. It
is possible to identify species at high risk for windthrow—black spruce, white pine, and Norway
spruce, for example—and high-risk sites.
Broadscale Disturbance by Hurricanes Occasional severe storms, particularly hurricanes,

may destroy the overstory canopy across hundreds or thousands of hectares of forest, initiating
secondary forest succession over large areas.
Gulf and Southern Atlantic Coasts More than 50 hurricanes have made landfall along the south-
eastern Louisiana coast since 1851 (Blake and Gibney 2011). Hurricane Katrina made landfall in
southeast Louisiana and the Mississippi coastline on August 29, 2005, with hurricane-force winds
that extended at least 100 km inland. It was at the time the third most intense storm on record to
make landfall in the United States, killing or severely damaging 320 million trees over 84 000 km2
(Chambers et al. 2007). Wind damage to forests in the region was spatially variable, however, with
60% severely disturbed, 35% moderately disturbed, and 7% only lightly disturbed in the Lower
Pearl River Valley and its surrounding area (Wang and Xu 2009). Overall, post-hurricane damage
patterns were closely related to stand conditions and site characteristics (Chapman et al. 2008;
Kupfer et al. 2008). Forests closest to rivers and streams in the region were more severely disturbed
than areas further away from bottomlands, related to the increased susceptibility of trees to wind
on poorly drained sites. In addition, older stands with a higher percentage of hardwoods were
more susceptible to wind damage. In general, wetland forests were more susceptible to damage
than other forest types with the exception of those having high proportions of cypress and/or
tupelo, which were resistant to wind damage in wetlands because of their buttressed roots. Oaks
and sweetgum were more susceptible to strong winds because of their shallow rooting. On uplands,
evergreen pines (especially longleaf pine) were more resistant to wind damage compared to oaks.
Shorter trees with smaller diameters in the shrub/scrub ecosystems of the region were the most
resistant to wind damage.
Hurricane Andrew struck on August 24, 1992, in south Florida and severely affected ecosys-
tems over an area 50-km wide by 100-km long (Loope et al. 1994; Pimm et al. 1994). From 25–40%
of overstory trees in pinelands were uprooted or snapped at heights of 1–6 m; 2–3 times as many
pines were broken as were uprooted. The understory was relatively unaffected. Tall, large-diameter
hardwood trees of hammock ecosystems (small islands of trees within broad wetland areas) were
extensively damaged (20–30% downed or large branches broken off), but forests dominated by bald
cypress showed only modest damage. Mangroves suffered heavy damage, reaching 80–95% trunk
snapping and uprooting in the most vulnerable sites (Smith et al. 1994). Interestingly, small patch-
es of surviving mangroves were saplings that had previously regenerated in gaps caused by
lightning-induced fires. These small gaps provided nuclei for recolonization of destroyed forests,
given that the Florida mangroves only reproduce viable propagules on stems less than 1 m in
height. A primary concern during post-hurricane succession is the acceleration of introduced
plant species spread that was already occurring before Hurricane Andrew. Propagule dispersal and
facilitation of establishment and further spread in new canopy gaps of exotic tree species, such as
the Chinese tallow, have proven to be problematic over the two decades since the event occurred
(Conner et al. 2002; Henkel et al. 2016; Smith et al. 2020).
New England—1938 Hurricane Hurricanes also regularly and historically affect forests of the northern
Atlantic coast and New England region. Seven major hurricanes along different tracks from 1620 to
1950 have inflicted light to severe damage (Chapter 22, Figure 22.7) at intervals of a few to more than
a 100 years. A great hurricane in 1938 severely damaged more than 240 000 ha of central New Eng-
land’s forests along a 100-km swath from central Connecticut and Massachusetts to the northwest
corner of Vermont (Foster and Boose 1992; Whitney 1994). Foster (1988b) found that damage to
species and stands from this severe windstorm occurred quite predictably and specifically; damage
was not heavy everywhere. Similar to patterns found after Hurricane Katrina, post-hurricane damage
was a mosaic of differentially damaged stands controlled by the physiography, wind direction, soil
type, and nature of the pre-hurricane vegetation, especially species composition, tree height, and
density (Spurr 1956a; Foster 1988a,b; Foster and Boose 1992). A case study of disturbance across the
New England landscape is presented in Chapter 22.
Wave-Regenerated Fir Species Wave-regeneration of subalpine fir species is an example of

specific broadscale, wind-induced disturbance patterns. In high-elevation balsam fir forests of the
northeastern United States (Sprugel 1976; Sprugel and Bormann 1981) and fir forests of central Japan
(Sato and Iwasa 1993), persistent wind (rather than wind velocity per se) and other factors maintain
fir regeneration and other ecosystem processes over long time periods. The fir canopy is broken by
numerous crescent-shaped strips of standing dead trees, with mature forest on the leeward side of the
dead trees and young, vigorously regenerating forest on the windward side (Figure 16.10). This
“wave” of death and regeneration moves slowly through the forest, with seedlings replacing the
dying trees at the wave’s leading edge. The waves move in the general direction of the prevailing wind
at constant speeds of 0.37–3.3 m per year. The effects of prevailing winds, including winter desicca-
tion, summer cooling, increased rime deposition, and breakage of branches and roots, are presum-
ably the causes of tree mortality of the dieback zone. Shallow, rocky soils of these ecosystems also
contribute to death of the fir trees, as strong winds induce tree sway and cause root breakage on sharp
rock surfaces, thereby reducing water conducting tissue and encouraging root disease fungi (Har-
rington 1986). Seedlings established under the canopy trees are released upon canopy tree death and
eventually become canopy trees themselves in 60–80 years (Figure 16.10). The wave-regenerated
forest is unusually orderly due to an omnipresent and highly predictable nature of the prevailing
wind disturbance (Sprugel and Bormann 1981).
FLOODWATER AND ICE STORMS

Floodwaters have shaped and rearranged physiographic features of streams of all sizes for tens or
even hundreds of thousands of years (Chapter 8). Through inundation, sedimentation, and
influence on local hydrology, floodwaters control the regeneration of riverine species and thus
shape the distinctive pattern of forest vegetation associated with riverine landforms. Flood
Dying– Young–
Mature regenerating intermediate Mature
Wind
FIGURE 16.10 Diagrammatic cross section through a regeneration wave. Fir regeneration is initiated in and
along the edges of the mature stand and released as the trees die. Source: Sprugel and Bormann (1981) / American
Association for the Advancement of Science. Reprinted with permission from Science 211 : 390–393, © 1981 by the
American Association for the Advancement of Science.
regimes—the long-term duration, frequency, and temperature of floodwater—have elicited genetic

differentiation of river valley species. The many different physiological adaptations of river valley
species to lack of oxygen for root metabolism enable particular species and races to dominate
diverse riverine ecosystems (Gill 1970; McKnight et al. 1981; Hook 1984).
Wind rapidly and visually impacts the overstory, but floodwaters slowly and pervasively
change the forest site itself by bank cutting and soil deposition. Topographic sites available for
establishment of vegetation are constantly created and destroyed by the flux of sedimentary
deposits (Malanson 1993). Flooding also controls seedling regeneration by continuously and sub-
tly eliminating upland invaders, favoring physiologically adapted plants and providing water and
nutrients, and thus significantly determines the composition and dynamics of riverine vegetation.
By altering ecosystem processes, flood control measures and human development of floodplains
have changed the successional patterns associated with flood regimes. Many upland species are
currently found in bottomland sites where flooding is excluded, emphasizing the enormous
selective force of floodwaters on seedling establishment. Damage to overstory and understory trees
alike by ice floes and woody debris moved downstream by floodwater can be considerable (Lindsey
et al. 1961).
Ice or glaze storms occur when rainfall freezes on contact with leaves, branches, or boles of
trees, and affects tree growth when branches break under the weight of the accumulating ice. Ice is
particularly damaging when ice-coated branches are subsequently affected by wind. Damaging ice
storms occur regularly throughout the eastern United States; Kashian and Barnes (2021), for
example, noted ice storm damage in southeastern Michigan that occurred at least twice per decade
over the last 60 years. As with many disturbances, some species are more susceptible to ice damage
than others (Downs 1938; Whitney and Johnson 1984), particularly fast-growing species such as
black cherry, aspen, willow, and basswood. Oaks, sugar maple, white ash, and hickories are less
likely to suffer ice damage, and conifers are particularly resistant to damage even with a constant
presence of foliage. Severe ice storms may create canopy gaps, providing an opportunity for suscep-
tible species such as black cherry to grow into the canopy. In addition to direct effects of ice storms
on the overstory by severe branch or stem breakage, ice damage may indirectly affect forest tree
species by providing entry for pathogens or weakening trees, so they are more subject to insect attack.
INSECTS AND DISEASE

Insects and disease-causing pathogens that kill or reduce the growth of trees affect millions of
hectares of forests each year in North America and are major regulators of forest ecosystems
(Chapter 12; Romme et al. 1986; Haack and Byler 1993; Castello et al. 1995). Insects and patho-
gens create widespread disturbances and influence succession by killing or weakening trees singly
or in patches. These disturbances reciprocally affect insect populations in many ways (Schowal-
ter 1989). Unlike fire, wind, or even logging (see in the following text), insects and disease affect
overstory trees directly without changing the understory, forest floor, or soil. Root fungi, bark
beetles, and inner-bark borers are the primary tree killers, tending to attack one particular species
or genus of trees, and strongly control the rate and direction of succession (Franklin et al. 1987).
Small gaps resulting from the death of overstory trees killed by pathogens or insects often favor
shade-tolerant species already established in the understory. Larger areas of overstory tree
mortality have been shown to result in increases of understory plant diversity and productivity
(Pec et al. 2015).
Epidemic levels of insects or disease may occur, especially those due to bark beetles, and
create a much larger effect on forested landscapes than endemic or background herbivory or
disease. For example, the return interval for mountain pine beetle outbreaks in lodgepole pine
forests of the northern Rocky Mountains is about 20–40 years, with an average duration of
about 6–10 years (Cole and Amman 1980; Taylor et al. 2006; Alfaro et al. 2010). These compo-
nents of the disturbance regime vary with climate and tree density, such that outbreak severity
will vary based on these factors. An unprecedented outbreak of mountain pine beetle occurred
in British Columbia between the mid-1990s and about 2017. Peaking in 2005, the rate of spread
and severity of the outbreak were the largest recorded in Canadian history, resulting in the loss
of 58% of the province’s merchantable pine volume over >18 million hectares (Figure 16.11;
Canadian Forest Service 2021). The size and duration of the outbreak have been attributed to a
Photo courtesy of Colorado State Forest Service
F I G U R E 1 6 .1 1 Lodgepole pine trees killed by mountain pine beetles near Granby, Colorado. The red-
needled trees in the photo are those recently killed by the beetle.
combination of factors, primarily the simultaneous maturation of lodgepole pine stands across
the landscape (due to past logging and effective fire suppression) and warming winter temper-
atures. Other native insect species have a similar influence on forests. In some Colorado subal-
pine forests, the effects of spruce beetle outbreaks appear to be as great as those of fire (Veblen
et al. 1991b). Having coevolved with the natural disturbance agents that affect them, forest
ecosystems are generally resilient to insect and disease outbreaks, but exotic or invasive insects
and disease may have more drastic effects on forests. The effects of some exotic insects and
diseases are described later in this chapter and in Chapter 21. Finally, insects and disease often
interact with other natural disturbances, such as fire and wind; these disturbance interactions
are also discussed in the following sections.
CATASTROPHIC AND LOCAL LAND MOVEMENTS

Catastrophic events such as volcanism, earthquakes, and landslides create new ecosystems and
shape species occurrence and distribution. Volcanic activity on several continents has provided
new substrate for forest development, such as in the Cascades of the Pacific Northwest, the Andes
in Chile (Veblen 1987), the Hawaiian Islands, and the Changbai mountains of northeastern
China (Barnes et al. 1992). The 1980 eruption of Mt. St. Helens has provided abundant evidence
of the overwhelming effects on existing vegetation and the rapidity of succession on newly
formed substrates as a result of biological legacies of diverse kinds (Franklin et al. 1985; Franklin
et al. 1988; Franklin 1990; Dale et al. 2005c; Del Moral and Chang 2015). Landslides in mountain-
ous terrain provide new substrates for plant occupancy (Pabst and Spies 2001). For example, a
1960 earthquake in the southern Chilean lake district triggered volcanic activity and thousands
of landslides that provided excellent regeneration sites for southern beeches (Nothofagus spp.)
(Veblen 1987). Such broadscale disturbances occur frequently enough to influence forest compo-
sition and structure over long temporal scales. Landslides and rock avalanches initiated by pre-
historic earthquakes also shaped landscapes and changed forests in mountains of the Puget
Sound area of Washington State about 1000 years ago (Jacoby et al. 1992; Schuster et al. 1992).
Catastrophic and local debris avalanches—rapid downhill flows of soil, rock, and vegetation
almost always preceded by heavy rains—are major disruptive and land-shaping forces in the
Appalachian Mountains (Hack and Goodlett 1960; Eschner and Patric 1982; Jacobson et al. 1989).
LOGGING
Lumbering was a major factor in the development of the United States, and major changes in
forest composition followed it. Post-logging fires, fueled by woody debris left after logging, also
contributed significantly to change in forest composition (Whitney 1994). Aside from large-scale
clear-cutting, logging as a disturbance may be similar to wind in that tree harvesting removes the
overstory and releases the understory. Many understory hardwood species cut in logging-related
operations resprout to form vigorous and fast-growing stems that then compete for overstory space.
As a general rule, the competitive ability of post-harvest regenerating trees will depend on
the intensity and pattern of the cut. Light, partial cuts such as thinning and improvement cutting
favor tolerant species, particularly those already established in the understory, and tend to acceler-
ate forest succession rather than return it to an earlier stage (Figure 16.12). Moderately heavy
partial cuts or cutting of small groups of trees to mimic the formation of canopy gaps (group selec-
tion cutting) favor mid-tolerant species. Clear-cutting or the creation of large gaps with a diameter
of at least twice the height of the stand favors the invasion of pioneer species, particularly if min-
eral soil is exposed by the harvest operation. A forest manager can therefore greatly influence
subsequent forest composition and the rate of succession by regulating the intensity and pattern
of cutting.
F I G U R E 1 6 .1 2 Partial cutting in this eastern white pine stand in Maine has favored advanced regeneration
of tolerant red spruce. Source: U.S. Forest Service photo.
The differential effect of partial cutting versus clear-cutting on forest succession is clear in
the Douglas-fir region of Washington and Oregon. Older stands that are cut are more or less
even-aged Douglas-fir dating to forest fires that occurred 150–500 years ago (Franklin et al. 1981).
Clear-cutting in such stands during or immediately after a seed year provides the reestablishment
of even-aged Douglas-fir. Clear-cut patches ranging from 1 to 15 ha in size will normally regenerate
to Douglas-fir within a few years with seed dispersed from surrounding uncut forest. On any site,
the new forest that follows a clear-cut is composed of pioneer species that range from shade intol-
erant to mid-tolerant in their competitive ability in the understory in that region. On clear-cuts on
the lower and wetter sites in the Coastal Range, red alder colonizes prolifically as a pioneer species.
Although large-area clear-cutting and planting of Douglas-fir was a conventional forestry practice
for several decades, an increasingly common view is that Douglas-fir can be regenerated and grown
under a partial overstory created with a variable retention harvest system or shelterwood
(Chapter 14, Figures 14.16–14.18).
Logging also affects forest succession by the differential removal of one species over
another—thus changing the composition of the forest. Such effects are exemplified in the mixed-
wood forests of Maine where logging has occurred nearly continuously for almost 150 years (Whit-
ney 1994). Early logging focused on large white pines suitable for masts for wooden sailing ships
and house construction, but later shifted to the large-scale cutting of red spruce to build and supply
sulfite and groundwood pulp mills. Removal of the red spruce from a community that included
tolerant balsam fir, sugar maple, and beech created a mixed forest of these species. Balsam fir was
more utilized as the supply of red spruce decreased. Partial cutting was soon replaced by clear-
cutting, resulting in a rising water table that precluded forest in many of the wetter flats; pioneer
species such as trembling aspen and paper birch colonized the drier sites. Thus, forest composition
of central and northern Maine has been markedly shaped by successive waves of logging, each
concentrating upon different species.
LAND CLEARING
Many forests occur in humid regions that are suitable for the raising of agricultural crops. The
great majority of the world’s population is found in forested regions, and thus much of the world’s
forests have inevitably been cut and the land cleared for agriculture. The massive forest clearance
that occurred in the eastern United States and its effects are described by Whitney (1994). In the
tropics, forest land use for agriculture is often (but not always) transitory, with fields being allowed
to revert to forest again after a few years. Even in the temperate zone, much land cleared and for-
merly used for farming has been found unsuitable for continued cropping or has been supplanted
by bringing better lands into production. This land has been planted to forest or allowed to revert
naturally to forest; many currently forested lands belonging to State and National Forests in the
Upper Midwest have followed this process. Secondary forest succession following land abandon-
ment is therefore an important process, occurring over many hundreds of thousands of hectares in
well-settled forest areas.
DISTURBANCE INTERACTIONS
As described in the earlier text, most ecosystems are subjected to more than one type of distur-
bance. The result of those interactions may be cumulative (increasing impact) or negative (e.g., the
second disturbance is less impactful or less likely to occur), or the disturbances may interact to
produce a novel or unexpected result (Paine et al. 1998). Interacting disturbances—specifically
that the effects of one disturbance may either change the likelihood or alter the effects of another—
have long been recognized by forest managers and ecologists. For example, extensive blowdown of
subalpine forest in 1939 led to an epidemic of spruce beetles in the 1940s that devastated 290 000 ha
of the White River National Forest alone (Hinds et al. 1965). The largest and/or most severe
wildfires typically occur during periods of prolonged drought, as in the infamous Peshtigo Fire of
Wisconsin in 1871, the massive western fires of 1910, and the 1988 fires in and around Yellowstone
National Park, all of which burned through prior burn locations. Fire-scorched or drought-stressed
trees are often sentinels to successful attack by bark beetles and other insects (Mattson and
Haack 1987). Mudslides often follow fire in steep-slope, chaparral-dominated ecosystems of
California and in other ecosystems of the western United States. The pattern and nature of multi-
ple disturbances are characteristic of regional ecosystems with their distinctive climates
(Mediterranean, northern continental, coastal/marine), physiographic features (mountains,
plains), and soil/hydrology (excessively drained, poorly drained).
Interacting disturbances have been studied extensively over the last two decades and may be
characterized as two types of interactions (Buma 2015). Disturbances that alter ecosystem resis-
tance, by changing the likelihood, intensity, severity, or size of the subsequent disturbance, are
called linked disturbances (Simard et al. 2011). Linked disturbances occur when the first distur-
bance creates a legacy, such as reduced availability of host trees for an insect or altered distribution
of fuels, that mechanistically affects the second. In a study of a spruce beetle outbreak in the 1940s
in western Colorado, Kulakowski et al. (2016) found that young stands of Engelmann spruce and
subalpine fir that regenerated from fires in the late-nineteenth century were not susceptible to the
spruce beetle outbreak. Thus, the location and severity of the spruce beetle outbreak were altered
by legacies of the fires in the late 1800s (in this case, lack of mature, susceptible trees). Other exam-
ples of linked disturbances include those presented by Kulakowski and Jarvis (2013), Kulakowski
et al. (2012), Simard et al. (2011), Lynch et al. (2006), Bigler et al. (2005), and Veblen et al. (1994).
When disturbances alter the resilience of an ecosystem to a subsequent disturbance such
that the interacting disturbance changes the likelihood or rate of recovery of the ecosystem, the
interaction is the result of compounded disturbances. In a study of American beech dominance
in coastal New England, Busby et al. (2008) found that intensive forest harvesting in the nineteenth
century favored regeneration of oaks and beech, but repeated hurricanes thereafter favored beech
persistence in the understory over oaks. A severe hurricane in 1944 and heavy deer browsing
released beech saplings (typically avoided by deer) while preventing further oak establishment.
The reduced resilience of oaks as a result of compounded disturbances in this example has created
long-term changes in forest composition and diversity (Busby et al. 2008). Other examples of com-
pounded disturbances include those presented by Andrus et al. (2021), Carlson et al. (2017), Kula-
kowski et al. (2013), and Girard et al. (2009).
Interestingly, the same events can simultaneously result in both linked and compounded
disturbance interactions (Buma 2015), as illustrated by two independent studies that examined the
interaction of a severe wind event with subsequent fire in the subalpine forests of northwestern
Colorado (Figure 16.13). Kulakowski and Veblen (2007) found that the changes in fuels resulting
from the blowdown altered the severity (but not the size) of the fire that followed, suggesting that
wind and fire in this case were linked disturbances. In a separate study in the same area, Buma and
Wessman (2012) found that the blowdown affected resilience of the system by altering seed dis-
persal into the areas disturbed by wind and severe fires. In particular, wind-damaged areas were
too large to be revegetated by spruce and fir seeds dispersed from adjacent unburned forests, and
Initial blowdown event
~5 years
Altered spatial
distribution of Altered vertical
cy
downed woody fuel structure

ga
debris
Le
Subsequent fire event
Altered distribution of Low regeneration by

Low regeneration by
high-severity burn patches seed dispersal species
serotinuous species due
to due altered spatial fuel due to more
to altered fire
structure, Kulakowski and homogenous burns,
characteristics, Buma
Veblen 2007 Buma and Wessman
ct and
and Wessman 2011

2012
im nism
pa
a
ch
Burn temperature
Me
Burn duration
Linked interaction Compound interactions
F I G U R E 1 6 .1 3 The interaction of a large wind event and subsequent fire in Colorado show that linked and
compound interactions may occur from the same events. Burn severity was higher in the areas affected by
wind (linked disturbance interaction), but the blowdown reduced post-fire resilience by limiting seed
dispersal into the large and severely burned patches from adjacent unburned areas and from serotinous
cones within the area (a compound disturbance interaction). Source: Buma (2015) / John Wiley & Sons.
Biotic Composition Changes 411
the increased severity of the fires in the wind-affected areas consumed the serotinous cones of
lodgepole pine and the seeds therein. These mechanisms reduced the ability of the ecosystems to
recover, suggesting a compounded disturbance interaction (Figure 16.13). An excellent review of
disturbance interactions and their ecological implications is provided by Buma (2015).
BIOTIC COMPOSITION CHANGES

Secondary succession originates primarily from fire, windstorms, logging, and land clearing, but a
second major kind of disturbance—the addition or subtraction of plant or animal species—will
alter the trajectory of forest succession. These changes in the flora or fauna may create consider-
able disturbance to existing ecosystems and thus may well be considered as initiating secondary
succession in a very real sense.
ELIMINATION OF SPECIES
When a species is eliminated or greatly reduced in abundance, its place in an ecosystem is typically
taken by other organisms. In many instances, species loss is the result of the introduction of exotic
insects or pathogens to which native tree species have not adapted. There are numerous examples
of species’ elimination or near-elimination that have occurred over the last 100 years as a result of
exotic insect or pathogen introduction. The American chestnut, once a major overstory dominant
in eastern North America (Braun 1950), was virtually eliminated from the eastern deciduous forest
by the 1940s by a pathogen introduced from Asia in 1904. American chestnut was replaced primar-
ily by its major associates in New England, the mid-Atlantic states, and the southern Appalachians
(Keever 1953; Woods and Shanks 1959; Good 1968; Stephenson 1974). Dutch elm disease, caused
by two related pathogens introduced from Europe in the early-twentieth century, drastically
reduced the abundance of mature American elm trees in swamp and river floodplain forests of the
eastern and midwestern United States, replaced by a variety of different species depending on site
conditions. In southeastern Michigan, American elm has not been eliminated from deciduous
swamps but remains as an understory and subcanopy overstory species rather than a dominant
overstory species (Barnes 1976). It will likely persist by seeds from young elm trees with an average
life span that is drastically reduced. Emerald ash borer is an invasive insect introduced into the
metropolitan Detroit area from Asia in the 1990s. Specific to the ash genus, the insect had spread
to 36 states and five Canadian provinces by 2022, killing nearly all ash trees larger than about 1 in.
in diameter once it reaches high population levels. Where ash was found in mixed stands,
succession has mainly featured replacement by associates depending on the species and site con-
ditions (Kashian and Witter 2011). However, pure stands of black ash, typically found on poorly
drained soils, may be potentially converted to shrub wetlands or other cover types if ash mortality
is high (Kolka et al. 2018; Palik et al. 2021). Like American elm, most ash species are likely to per-
sist, but only as seedlings, saplings, and sprouts from top-killed overstory trees and the seeds they
produce (Kashian 2016). Effects on forests of invasive species such as chestnut blight, Dutch elm
disease, and emerald ash borer are discussed in detail in Chapter 21.
Reduction of tree species in a community is not always the result of the introduction of
exotic species. For example, in the spruce–fir forests of eastern Canada and the northeastern Unit-
ed States, spruce budworm epidemics have greatly reduced the abundance of overmature and
mature balsam fir (together with lesser amounts of black and white spruces) over large areas. The
budworm apparently has played a major role in these mixed softwood forests in suppressing the
proportion of the very-tolerant balsam fir compared to that of the somewhat less-tolerant spruces
(Bognounou et al. 2017). Notably, the response of balsam fir in this case is a local reduction in
abundance rather than a widespread elimination because it has coevolved and adapted to the long-
term presence of the native spruce budworm.
In the animal portion of forest ecosystems, the virtual elimination of many predators from
most forests in North America—particularly wolves, bears, cougars, and lynx—has played a role in
increasing the populations of their prey: deer, rabbits, and other herbivores. This in turn results in
greater browsing of the understory, including tree regeneration. Such disturbance precludes the
regeneration of black cherry in areas of Pennsylvania (Marquis 1974, 1975, 1981; Tilghman 1989)
and hemlock in many places in the northern United States (Frelich and Lorimer 1985; Peterson
and Pickett 1995). Reintroduction of wolves into Yellowstone National Park in 1995 had significant
effects on elk populations, both in their population size (reduced by predation) as well as their for-
aging behavior. As a result, browsing was reduced and recruitment to the overstory was increased
for trembling aspen and various species of cottonwoods and willows (Ripple and Beschta 2003, 2012;
Beschta and Ripple 2016).
ADDITION OF SPECIES
The species occupying a given site may simply be those that have access to that site at the time of
its availability rather than those best adapted to compete and grow on that site. Invasion of the site
by better competitors often results in substantial changes in forest succession. The chestnut blight
fungus, Dutch elm disease fungus, and emerald ash borer cited in the previous section are exam-
ples of accidental introductions that have greatly modified forests. Other organisms may be trees,
other higher plants, fungi, bacteria, and animals of all kinds (Chapter 21). Important deliberate
forest-tree introductions in Europe include Douglas-fir, eastern white pine, northern red oak, and
black locust; in the northeastern and north-central United States: Scots pine, Norway spruce,
Norway maple, black pine, white mulberry, Chinese tree of heaven, Siberian elm, and European
larch; and in temperate zones of the Southern Hemisphere: Monterey pine, patula pine, and slash
pine. These and many others have become vigorous and often dominant species in the local flora.
The effects of additions of such species to forests is the subject discussed in detail in Chapter 21.
SUGGESTED
R E A D I N G S
Agee, J.K. (1993). Fire Ecology of Pacific Northwest Paine, R.T., Tegner, M.J., Johnson, E.A., and E.A. (1998).
Forests. Washington D.C.: Island Press 493 pp. Compounded perturbations yield ecological sur-
Chapters 5 and 7–12. prises. Ecosystems 1: 535–545.
Covington, W.W. and Moore, M.M. (1994). Postsettle- Pickett, S.T.A. and White, P.S. (1985). Natural Disturbance
ment changes in natural fire regimes and forest and Patch Dynamics. New York: Academic Press 472 pp.
structure: ecological restoration of old-growth pon- Sprugel, D.G. and Bormann, F.H. (1981). Natural distur-
derosa pine forests. J. Sustain. For. 2: 153–181. bance and the steady state in high-altitude balsam fir
Heinselman, M.L. (1973). Fire in the virgin forests of the forests. Science 211: 390–393.
boundary waters canoe area, Minnesota. Quat. Res. Whitney, G.G. (1994). From Coastal Wilderness to Fruited
3: 329–382. Plain. Cambridge: Cambridge Univ. Press 451 pp.
Forest Succession CHAPTER 17
A ll components of ecosystems change over time, including climate, landforms, soil, and biota,
and we have discussed the basis of dynamic change in forest ecosystems in previous chap-
ters. Change may be gradual or rapid depending on the particular disturbances that characterize
regions and local sites (Chapter 16). Ecosystem development and change are most easily observed
in vegetation, especially the forest trees, which may change dramatically over time following
disturbances such as fires or windstorms.
One manifestation of ecosystem change is forest succession, which is often characterized
as the change in species composition or the replacement of the biota of a site by one of a different
nature. It is notable that other aspects of ecosystem change, such as soil development, have also
been documented (Olson 1958; Miles 1985; Matthews 1992), and changes involving biomass
accumulation and nutrient cycling are described in Chapters 18 and 19. In this chapter, we focus
upon changes in species composition. Importantly, forest succession is what happens in one place
over an extended period of time, measured in tens to thousands of years. In other words, succession
happens with space fixed and time changing (Rowe 1961b). It is characterized by a sequential
change in the relative structure, kind, and abundance of the dominant species. We focus here on
the intervals between disturbances and within a time period of the same order of magnitude as the
life span of the longest-lived organisms in the successional sequence (tens to thousands of years).
Forest succession progresses in nearly infinite ways and is driven by many different factors,
and many other processes occur simultaneously with it. However, succession can only have
meaning in the context of a particular geographic framework defined by regional and local ecosys-
tems. The heart of the matter is that location-specific ecosystems undergo development, change, and
succession. The way a given ecosystem changes when disturbed is displayed more or less visibly
according to the resilience and inertia of its component parts. Our attention is quickly drawn to the
visibly changing vegetation, but attempts to explain and predict its successional path separate from
its geographic context are unlikely to be rewarding.
Because vegetation is most easily observed, plant succession has historically been defined in
the context of plant species or communities. Definitions include those of Grime (1977): “a progres-
sive alteration in the structure and species composition of the vegetation” and Finegan (1984): “the
directional change with time of the species composition and vegetation physiognomy of a single
site.” Forest succession obviously deals with changes in biota, but the way we conceive its context
(plant or ecosystem, part or whole) is critical in understanding the patterns and processes of
succession over the complexity of regional and local ecosystems. Troll (1963b) was the first to
emphasize and explicitly describe succession as a landscape process involving all ecosystem com-
ponents, including climate, physiography, and soil as well as vegetation.
Adopting vegetation as a primary focus of succession contributed to the development of an
elaborate and often criticized concept, with literally thousands of published papers introducing
additional terminology and controversy. A perceptive review by Pickett et al. (1987) clears up much
413
414 Chapter 17 Forest Succession
confusion, and a landscape ecosystem perspective provides the spatial framework for under-
standing how succession works. All successions have these key characteristics that affect their
course and rate: (i) a sequence of concomitant environmental and vegetational change that may be
characterized by stages for convenience and practical applications; (ii) disturbance regimes; and
(iii) mechanisms (processes) occurring at different times in the sequence. Above all, the possible
successional sequences of interacting organisms, mechanisms, and disturbances are controlled or
mediated by the regional and local landscape ecosystems where succession takes place.
In this chapter, we examine successional concepts, the evolution of the concept of forest
succession, causes of ecosystem and forest change, and the mechanisms involved. The focus is of
course on vegetation change but placed in an ecosystem context as far as possible. Examples of
primary and secondary succession are presented to illustrate succession in diverse settings. For
details beyond the scope of our treatment, the following books and reviews provide diverse view-
points: Clements (1916, 1936); Tansley (1929, 1935); Drury and Nisbet (1973); Connell and Slatyer
(1977); Golley (1977); Miles (1979, 1987); White (1979); West et al. (1981); Finegan (1984); Pickett
et al. (1987); Glenn-Lewin et al. (1992); Matthews (1992); Peet (1992); Worster (1994); Schmitz
et al. (2006); Swanson et al. (2011); Walker and Wardle (2014); and Prach and Walker (2020).
BASIC CONCEPTS OF SUCCESSION

PRIMARY AND SECONDARY SUCCESSION
Primary succession is ecosystem change that occurs on previously unvegetated terrain, uniniti-
ated by a disturbance. For example, primary succession on recently deglaciated terrain in Alaska
eventually supports spruce forest after first being colonized by communities of willow or alder.
Secondary succession follows a disturbance to an existing forest that disrupts ecosystem
processes and destroys or disrupts existing biota. The distinction between primary and secondary
succession is often more arbitrary rather than real, but it is followed here as a matter of convenience
in understanding the dynamics of forest ecosystems.
BIOLOGICAL LEGACIES
Biological legacies that follow disturbances such as fires, floods, and windstorms distinguish the
course of ecosystem change during secondary succession. Biological legacies are biologically
generated elements of a pre-disturbance ecosystem that persist in some form after a disturbance
occurs (Franklin et al. 2018). Examples of biological legacies include: (i) living organisms or
portions of organisms that survive a disturbance, (ii) organic debris, and (iii) biotically derived pat-
terns in soils and understories (Franklin 1995). Living legacies include intact plants and animals,
rhizomes, and dormant spores and seeds. Important biotically derived structures include standing
dead trees (snags) and fallen logs (coarse woody debris), large soil aggregates, and dense mats of
fungal hyphae. These diverse biological legacies, remaining from pre-disturbance ecosystems,
strongly influence the paths and rates of forest recovery and ecosystem function by shaping
post-disturbance community structure and/or providing substrate for new colonizers. In fact, the
types and abundances of biological legacies are among the most important ecological outcomes of
disturbances (Franklin et al. 2018), although they differ sharply among different kinds of distur-
bances (Table 17.1).
SUCCESSIONAL PATHWAYS, MECHANISMS, AND MODELS

Pickett et al. (1987) define the terms successional pathway, mechanism, and model. A successional
pathway is the temporal pattern of vegetation change, which typically shows change in community
Basic Concepts of Succession 415
Table 17.1 Biological legacies generated by different types of severe disturbances.
Disturbance type
Clearcut
Biological legacy Severe wildfire Wind Insect outbreaks Volcanoes
harvesting
Live trees Few Few or absent Depends on forest Few, confined Absent
composition to margins
Snags Abundant Few Abundant Abundant Absent
Downed Common Abundant Eventu- Abundant Absent
woody debris ally abundant
Uproots Absent Abundant Absent Few Absent
Intact understory Rare Abundant Abundant Absent Very little
Mineral seedbed Abundant Abundant Absent Absent if buried Abundant
with uproots
Source: Adapted from Franklin et al. (2018) and Swanson et al. (2011).
Trembling aspen
Jack pine
Fire severity (y)
Black spruce
50 Times since fire (x) >250
F I G U R E 17.1 Alternative successional pathways for aspen–conifer forests in the clay belt of northwestern
Quebec, Canada. Forest composition and structure are strongly determined by the severity of the last fire
(y-axis) and the time elapsed since the last fire (x-axis). High-severity fires initially result in forests ranging
from aspen-dominated to a mixture of jack pine and black spruce. Low-severity fires result in black spruce-
dominated forests across a range of density. With time, forests generally converge toward open, multi-aged
black spruce forests. The rate and direction of forest succession on this landscape are strongly influenced by
site factors, especially soil type. Source: From Franklin et al. (2018).
types and may describe the decrease of particular species populations. Figure 17.1 is an excellent
example of successional pathways because the initial site conditions in conjunction with distur-
bance are major factors in determining ecosystem change. A successional mechanism is an
interaction or process that contributes to successional change. Examples include general ecolog-
ical processes such as dispersal, competition, and establishment, and physiological processes such
as biomass allocation (Chapter 18) and nutrient uptake (Chapter 19) that affect plant form and
reproductive ability. A successional model is a conceptual map that explains a successional
pathway by identifying and specifying the relationship among the mechanisms and the various
stages of the pathway.
AUTOGENIC AND ALLOGENIC SUCCESSION

Successional change has traditionally been characterized as either autogenic (by endogenous
factors) or allogenic (by exogenous factors) depending whether successive changes are brought
about by the action of the plants themselves on the site or by external factors (Tansley 1929). Pri-
mary succession is an autogenic process in the classical view because it is plant-driven, and
secondary succession is allogenic because it is driven by periodic disturbances. The dichotomy
between autogenic and allogenic factors is actually quite artificial: the causal factors don’t fit
strictly into one category or the other (White 1979; Matthews 1992). Tansley (1929, p. 678) recog-
nized this problem when he proposed the concept, cautioning that real successions show a mixture
of the two classes of factors.
Today, the ecosystem rather than the plant is or should be the focus, because change is elic-
ited by a combination of factors both within the ecosystem and external to it. Allogenic factors
(physical processes and disturbances such as fires and windstorms) may be initiated outside a
given ecosystem, but their effects often strongly depend upon site conditions (e.g., topography,
aspect, soil depth, drainage) and the associated vegetation within a given ecosystem. For example,
the aspect, topography, soil properties, and vegetation of an ecosystem, and its position in relation
to fire breaks, strongly affect fire frequency and severity. Windstorm effects are significantly
related to soil depth, drainage, rooting ability, and the height of tree species of the affected eco-
system. Furthermore, autogenic factors of litter accumulation and plant-induced changes in
microclimate, soil water, and nutrients all affect the outcome of allogenic disturbances. In turn,
autogenesis is almost always influenced by allogenic factors. Thus, allogenic and autogenic
factors are closely interrelated.
HOW IS SUCCESSION DETERMINED?

In either primary or secondary succession, ecosystem and vegetation change should ideally be
tracked at one place over time such that an actual successional pathway for a given ecosystem
might be determined. However, such monitoring is rarely possible due to the long life span of
forest trees, and much of what we know about succession therefore comes from short-term obser-
vations at different places. The typical approach used to characterize long-term succession is to
infer successional trends from stands of different ages, all of which are assumed to have developed
on the same site and under the same disturbance conditions. This approach has come to be known
as a chronosequence (Major 1951) or space-for-time substitution (Pickett 1988). If constructed
carefully and replicated, chronosequences can be indispensable for inferring how forest ecosystem
structure and function change over centuries (e.g., Kashian et al. 2013). A chronosequence
approach is not without its problems, however (Matthews 1992; Johnson and Miyanishi 2008).
Chronosequences are generally most appropriate when successional pathways are likely to
converge, biodiversity is low, and disturbance is infrequent and/or not severe (Walker et al. 2010).
Chronosequences are less suitable when successional pathways are likely to diverge, or the
ecosystems are species-rich, highly disturbed, or arrested in time. The actual chronosequence of
ecosystem and forest change is one that is site-specific over time, and ideally this should be the
process that is observed. Several examples of chronosequences are described later in this chapter.
Evolution of the Concept of Forest Succession 417
EVOLUTION OF THE CONCEPT OF FOREST SUCCESSION

The dynamic nature of forests was recognized by naturalists in North America and Europe as early
as the eighteenth century in a form that approximates a plant-based (bioecosystem) concept of
forest succession (Spurr 1952a). Such ideas gained significant traction in the nineteenth century.
In Europe, beginning with Hundeshagen in 1830, observed changes in forest composition were the
subject of specific articles by professional foresters and botanists. Hundeshagen pointed out
instances of spruce replacing beech and other hardwoods in Switzerland and Germany, and of
spruce and other species taking the place of birch, aspen, and Scots pine. The first detailed North
American report of species composition changes was apparently that of Dawson in 1847. Working
in the Maritime Provinces of eastern Canada, Dawson recognized the effects of windthrow and fire
in the forests found by European colonists and distinguished between successional trends in small
clearings following cutting, following a single fire, following repeated fires, and as a result of agri-
cultural use of land.
As early as 1860, the writer Henry David Thoreau recognized that pine stands on upland
soils in central New England were succeeded after logging by even-aged hardwood stands that
today constitute the principal forest type of the region. He named this trend “forest succession”
and published an essay on “The Succession of Forest Trees.” His late natural history writings, pub-
lished long after his death (Thoreau 1993), are remarkably filled with insights on ecological
succession: pines are “pioneers” to oaks and “lusty oaken carrots” that are a “remarkable special
provision for the succession of forest trees.” In articles published in 1875 and 1888, Douglas dis-
cussed at some length the concepts of forest succession and pioneer species, and presented an
explanation of how short-lived, light-seeded pioneer species formed the first forest types on
burned-over pine land.
FORMAL ECOLOGICAL THEORY

A general and formal theory of plant succession evolved slowly but was well established by the
beginning of the twentieth century when several American ecologists systematized its study.
The foundations of plant ecology as a study of vegetation dynamics and succession were initi-
ated by Henry C. Cowles (1899, 1901), who analyzed succession on sand dunes of Lake Michigan,
beginning with uncolonized sand and ending with a mature forest. Although dunes were formed
and destroyed by wind, he found that “vegetation profoundly modifies the topography.” In 1911,
Cowles noted that the original plants in any habitat give way “in a somewhat definite fashion to
those that come after.” Furthermore, he recognized that changes in vegetation caused by climatic,
topographic, and biotic changes occurred concurrently, at different rates, and at times in differ-
ent directions.
A contemporary, Frederic E. Clements, next developed an elaborate philosophical structure
of plant succession (1916, 1949) that attempted to categorize and formalize all eventualities of
plant community change. Specific examples of forest succession were documented by William
S. Cooper, with his studies of Isle Royale in Michigan (1913) and the colonization by plants follow-
ing glacial retreat in Glacier Bay, Alaska (1923, 1931, 1939). Clements, in particular, developed a
complex nomenclature to describe plant succession in a meticulous, tidy organization of nature
that has both facilitated and greatly complicated the efforts of his successors. Some of his terms
have taken a permanent place in the vocabulary of ecologists (vegetation dynamics, biome), others
have persisted but with broadened and changed meanings (climax), while many have been dropped
from general usage (cyriodoche, consocies). The work of Clements is described in detail in the
following text.
HOW DOES SUCCESSION WORK?

Succession involves many factors and processes operating within a geographic framework at mul-
tiple scales (Table 17.2). Understanding the key characteristics of ecosystem change and what they
might tell us about succession is critical for examining how succession works. Key characteristics
determining actual vegetation change (with space fixed and time changing) include the following:
The physical site conditions of the Earth’s regional and local ecosystems provide a geographic
framework that determines the organisms involved and the rate and trajectory of succession.
Disturbance factors and processes associated with particular regional and local ecosystems provide
substrates for primary and secondary succession and periodically affect the rate and trajectory of
succession.
Colonizing organisms have an enormous range of life-history traits (genetic adaptations of struc-
tural architecture, longevity, and physiological processes) and regeneration strategies that are
tightly linked to site conditions and disturbance regimes.
Changes in site conditions and organisms occur reciprocally, with one affecting the other and
vice versa.
Organisms interact through competition and mutualisms.
With these characteristics we can expect that: (i) successional pathways are likely to be unique, and
no two pathways will be the same; (ii) different successional pathways will occur where regional
and local ecosystems are also diverse; (iii) there are a great number of different and simultaneously
operating mechanisms that control the rate and trajectory of succession; (iv) distinguishable
patterns and stages of succession are evident in ecosystems not severely affected by humans;
(v) species occurring early in succession modify the environment in ways that may limit their own
regeneration and may favor or inhibit the regeneration and growth of species occurring late in
Table 17.2 Causes of succession: Hierarchical summary of factors and processes.
Landscape ecosystem characteristics

Regional climate, physiography, vegetation patterns
Local climate, physiography, landform, soil, species distribution
Site availability
Disturbance (kind, periodicity, size, severity, dispersion, disturbance history of target site and
adjacent ecosystems)
Differential species dispersal and establishment

Landscape position in relation to seed source availability
Reproductive mode of species (sexual, asexual) and fecundity
Dispersal (dispersal agents, landscape configuration)
Propagule pool for establishment (time since land disturbance, land use treatment)
Differential species performance (in regeneration and post-establishment

development)
Genetic adaptations and plasticity (germination and establishment traits, understory
tolerance, assimilation and growth rates, biomass allocation, plant defense)
Competition and mutualism (presence, kind, amount)
Environmental stress (climatic cycles, site history, prior occupants)
Allelopathy (soil chemistry, neighboring species)
How Does Succession Work? 419
succession; and (vi) although succession proceeds at various rates, there is no end point of
the process.
With these key characteristics, what are the significant features of how succession works?
We may summarize the general trends of how succession works as follows:
1. Following a disturbance that creates either (i) new land (unvegetated substrate for primary
succession) or (ii) a disrupted forest with a vegetated site having remnants of the
pre-disturbance forest (with secondary succession following), substrate is available for
recolonization or continued growth by organisms.
2. Species from surrounding areas or within the disturbed area itself occupy the disturbed site
in varied patterns and temporal sequences. Where a disrupted forest undergoes secondary
succession, new colonists may be dispersed into the site or are already present in surviving
vegetation, the soil seed bank, or the vegetative propagule bank (Figures 4.2 and 4.13,
Chapter 4).
3. The colonizing plants develop on the site and either become established or die.
4. Plants interact through competition and mutualisms and thereby affect further plant
establishment and development over time.
5. Plants change site conditions, such as light, temperature, or soil properties, and processes
which may or may not affect them as they continue to interact with other plants
and animals.
6. All of the abovementioned processes are modified by multiple disturbances that are
ecosystem-specific in type, frequency, and severity.
7. Some species persist, thrive, and dominate at their time during succession, whereas others
drop out or diminish in abundance.
The nature and rate of change of vegetation depend on the spatial and temporal scales under
consideration, the site conditions of regional and local ecosystems, species’ life-history traits and
regeneration strategies, disturbance regimes, chance effects, and other processes (Table 17.1). If it
is known which species are present, which are in the vicinity of a particular land area, and the
ecology of the species in the context of the various land-vegetation patterns, then it is possible to
make an educated guess as to what is likely to happen in the future (Rowe 1961b).
CLEMENTSIAN SUCCESSION
It is instructive to compare the scenario just described with Clements’ model of succession, the
first to describe the vegetative approach and a legacy for our current understanding and the basis
for opposing ideas. Clements (1905; Weaver and Clements 1929) declared what he saw as the rules
of succession in no uncertain terms. The basic processes were clear cut, and stages proceeded in a
systematic, stepwise, directional, and predictable process (Figure 17.2). In this model, ecesis (ger-
mination, establishment, and growth) occurs immediately after plants reach the site (migration).
Dominant plants change the environment (reaction) to favor or “facilitate” the dominance of the
next community that can better compete at the site. The end point of a given succession is reached
as the climax formation or association is attained. The plants that dominate at the climax stage
modify their environment such that they perpetuate themselves, resist change, and therefore cre-
ate the community of maximum stability. Clements also assumed that all primary successions in a
climatic region eventually converge to the same climax community from multiple starting points
(monoclimax). A critical flaw of Clements’ was that in emphasizing the influence of plants upon
DISTURBANCE F I G U R E 17. 2 Classical (Clementsian)

MIGRATION model of autogenic plant succession.
Disturbance brings about nudation followed
ECESIS
by arrival of propagules (migration) that
Biotic Reactions germinate and grow (ecesis) and modify the
Favor New Migrants site (biotic reaction) in competition with one
BIOTIC REACTION
another. Ecesis, reaction, and competition
often occur simultaneously. Competition and
reaction continue in cycles (left) that favor
COMPETITION new immigrants and hence new commu-
nities; ultimately (lower right) the reactions
Biotic Reactions Favor lead to the climax or end stage. Source: From
Already Established Christensen (1988b) / University of
Dominants Washington Press.
CLIMAX
the habitat (i.e., biotic reaction in Figure 17.2); he saw the driving force of succession to be changes
caused by internal factors (i.e., autogenic succession). It is therefore important to consider the
stages of succession as a basis for examining insights, problems, and misconceptions about orderly
succession.
STAGES OF SUCCESSION
In describing plant succession, we begin with a hypothetical unvegetated substrate or disturbed
forest and describe the successive plant communities that will occupy this site. Doing so requires
assumptions that (i) the regional climate will remain unchanged, and (ii) disturbances such as
severe windstorms, fires, or insect outbreaks are absent. Such assumptions are unrealistic given
the hundreds or thousands of years involved in forest successions. However, these assumptions
provide the view of succession as a more or less orderly and predictable sequence that depends
upon the character of the physical site and associated biota. The assumptions also provide the
framework to examine some problems of this perspective.
The recognition of stages is a matter of convenience rather than of their actual occurrence.
At best, stages are simply wave-like replacements of species with similar ecologies, and at worst,
they are arbitrary divisions in a continuum (Matthews 1992). It is notable that depending on site
and species availability, plant communities vary in their rate of change as new species colonize the
site and existing species either reproduce or disappear. The arbitrary classification of this continuum
into stages, characterized by the dominance or presence of certain species and plant life-forms, is
certainly useful and a convenience worth maintaining.
Primary Succession Unvegetated sites range from pure mineral material (rock or soil) to
water; plant colonization and growth are most favored by mixtures of soil and water (i.e., moist,
well-drained mineral soil). Primary plant succession beginning with dry rock material (either as
rock or as mineral soil) is termed a xerarch succession; that beginning with water is termed a
hydrarch succession; while that beginning with moist, aerated soil is a mesarch succession.
The major stages of primary succession are generally consistent, but many types of primary
successions exist (Matthews 1992). Both the specific successions and the vegetational stages within
each are arbitrarily chosen. In Table 17.3, a series of 10 stages is given for each representative type
of primary succession. Notably, some stages are omitted under conditions where the next
successional life-form (tree, shrub, herb) is capable of directly colonizing an earlier vegetational
Table 17.3 Stages of primary succession.
Stage Xerarch Mesarch Hydrarch

1 Dry rock or soil Moist rock or soil Water
2 Crustose lichens (Usually omitted) Submerged water plants
3 Foliose lichens and mosses (Usually omitted) Floating or partly floating plants
4 Mosses and annuals Mostly annuals Emergents
5 Perennial forbs and grasses Perennial forbs and grasses Sedges, sphagnum, and mat plants
6 Mixed herbaceous Mixed herbaceous Mixed herbaceous
7 Shrubs Shrubs Shrubs
8 Intolerant trees Intolerant trees Intolerant trees
9 Mid-tolerant trees Mid-tolerant trees Mid-tolerant trees
10 Tolerant trees Tolerant trees Tolerant trees
type. In the Clementsian paradigm, the series of stages at a given site is termed a sere and each
stage a seral stage.
There exist several potential misconceptions in relating the possible stages to real-world
succession. First, succession does not necessarily begin with stage 1, proceeding neatly through
each successive stage unidirectionally. Second, stages do not typically proceed discretely one after
another in relay fashion, but instead exhibit considerable overlap between them. Third, there is no
predetermined time period for each stage to begin and end; depending on the site, one stage may
occupy the site for a long time and seemingly terminate succession until site conditions change. In
sum, real-world succession typically has stages that are far less uniform and rigid than those listed
in Table 17.3.
The various life-forms and developmental stages in succession may be characteristic of more
than one stage, and indeed, many persist through many stages. Some mosses, for instance, may col-
onize a site early in succession and persist until the later stages characterized by tolerant trees. Tree
seedlings often establish themselves very early in succession along with annuals and grasses or
beneath shrubs, particularly in secondary succession when tree seed sources surround the disturbed
area. Thus, all plants of each stage do not necessarily appear and die out abruptly at an appointed
time; rather they may overlap in various sequences according to the life-history traits of each species.
For example, in the Appalachian Mountains many clonal shrubs and alder, rhododendron, and
mountain laurel thickets persist for 20–40 years or more, preventing the establishment of tree seed-
lings within them (Niering and Goodwin 1974; Damman 1975; Vose et al. 1994).
The actual species composition of the different stages depends upon those species that have
access to the site in question, either in terms of the relative location of seed sources or their ability
to be dispersed to the site (Clements 1905; Abrams et al. 1985; McCune and Allen 1985;
McClanahan 1986; Matthews 1992; Fastie 1995; Buma et al. 2017). Thus, the actual plant commu-
nities present on a given site depend upon the available plants as well as upon the site itself. Such
plants may be already on the site as buried seeds and propagules (roots, rhizomes) or may be
transported by wind, water, or animals from adjacent or distant ecosystems.
The importance of species availability and dispersal in succession was clearly borne out in a
study by Buma et al. (2017), who resampled Cooper’s original permanent plots in Glacier Bay,
Alaska after 100 years of succession. The plots were located by Cooper at particular distances from
a retreating glacier to represent time since glaciation, and thus various stages along the successional
sequence. The first 50 years of succession seemed to confirm the basic hypotheses of plant succession
that plant coverage, in general, and willow coverage, particular to this location, would be higher in
“older” plots. The second 50 years of succession, however, suggested that the spatial location of seed
sources and species dispersal into the plots began to become more important than time since glaci-
ation. Moreover, the plots were highly variable in canopy coverage, species composition, and soil
development. In particular, predicted late-successional dominant species (Sitka spruce) did not
establish in plots where willow arrived early and dominated. These results highlight that, in some
cases, early species arrival has an enormous influence on successional pathways (Buma et al. 2017).
Egler (1954) described the dichotomy of two models, “relay floristics” and “initial floristic
composition” (IFC), that contrast the strategies of species as they colonize old fields. Relay flo-
ristics (Figure 17.3a) refers to the appearance and disappearance of stages of species at a site.
This model suggests that one group of plants favors or facilitates the establishment of the next
(a)
Abandonment
Crop Weeds Grassland Shrubland Forest

Species abundance
Time
Abandonment
(b)
Crop Weeds Grassland Shrubland Forest
Species abundance
Time
F I G U R E 1 7 .3 Models of succession from studies of abandoned agricultural fields (old-field succession).

(a) Relay floristics—colonization of successive groups of species occurs in sequential phases. (b) Initial
floristic composition (IFC) —all species establish at or shortly after initiation of succession but still dominate
sequentially in stages. Source: Egler (1954) / Springer Nature. Vegetation science concepts I. Initial floristic
composition, a factor in old-field vegetation development. Vegetatio 4:414, © 1954 by Uitgeverij Dr. W. Juink,
Den Haag. Reprinted with kind permission from Kluwer Academic Publishers.
group in the sequence by modifying the environment. Some attribute this rigid scheme to
Clements, but, in fact, he acknowledged the less discrete nature of the stages as they appeared.
In 1905, Clements observed that many migrants appeared into a new, denuded, or greatly modi-
fied habitat to then be sorted into three groups by the establishment process: (i) those that arrive
but are unable to germinate or grow and soon die, (ii) those that grow normally under the con-
ditions present, and (iii) those that arrive but remain dormant through one or more of the earlier
stages to appear at a later stage. Matthews (1992, p. 294) also observed that the relay floristics
model, as shown here, does not realistically represent the lack of discreteness of the species
groups in primary succession.
In contrast to the relay floristics model, the IFC model (Figure 17.3b) suggests that all species
are present at or shortly after succession begins, as acknowledged by Clements, and then occur as
separate stages over time. The IFC model incorporates the idea that succession is a function of
species’ life-history traits rather than simply facilitation by those species arriving earliest. Perhaps
the most questionable assumption in Egler’s IFC model is that all or most species are present at the
start of succession and additional species arrivals do not occur. For example, Finegan (1984) argued
that shrubs and small trees (cherries, juniper), some pines, and red maple are early colonizers in
old fields in the eastern United States, but oaks and hickories colonize significantly later and are
not present at the start. Finegan also demonstrates many other situations where succession does
and does not resemble the IFC model. Matthews (1992) examined the IFC model in studies of pri-
mary succession on recently deglaciated terrain in Alaska and found an appreciable lag between
colonization by pioneer species and that of loose groupings of later colonizers. These differential
patterns of establishment draw the IFC model into question, as more sequences resemble the relay
floristic model without the discreteness shown in Egler’s diagram (Figure 17.3a). In reality, there
simply lacks a model of forest succession that is universally applicable to all sites and situations.
Secondary Succession To some degree, secondary succession mimics primary succession in

its basic principles. Most forest ecosystems regenerate rapidly following disturbances, and so
secondary succession is often described in terms of the redevelopment of forest tree populations.
Ecologists working in even vastly different forest regions have produced similar conceptual models
of post-disturbance forest succession featuring four phases or stages (Table 17.4). In European
beech forests, Watt (1947) described “upgrading” and “downgrading” periods that he divided into
four phases, of which the gap phase has had the most application elsewhere. In general, we rec-
ognize two basic stages, known as aggrading (also upgrading or building) versus senescing (also
downgrading, degenerating, or declining). Among the best-known models in North America are
those featuring stand development (Oliver 1981; Table 17.4) and populations (Peet and
Christensen 1980a, b; Table 17.4). The stand development model was developed in New England
and the Pacific Northwest (Figure 17.4a). The population model applies particularly to the
Table 17.4 Stages of secondary succession.
Forest maturation model Stand development Population model (Peet

Beechwood model (Watt 1947)
(Franklin et al. 2018) model (Oliver 1981) and Christensen 1980a, b)
Disturbance and Gap
legacy creation Upgrading
Preforest Stand initiation Establishment Bare
Forest canopy closure
Young forest Stem exclusion Thinning Oxalis
Mature forest Understory reinitiation Transition Downgrading
Old forest Old growth Steady state Rubus

(a)
Species A Species B Dominates Species B, C, & D
Dominates Dominate
B
C
D
D
A C
C
A D
Time Since Disturbance

Disturbance
Stand Stem Understory Old

Initiation Exclusion Reinitiation Growth
Stage Stage Stage Stage
B B B
B C
B B D
C
D C D
C
C C D C D
C
C D D
C
A A A C D
C C C
C C C B BB D C D D
B C B D D C C D C C C C
(b)
Stand Stem Understory Disturbance Accelerated
Initiation Exclusion Reinitiation Succession
Stage Stage Stage Stage
A A A
A C
A A A
B B C
C
A A AA A A A C BB C B C C B
B
A
C B B C B C B
A A C
F I G U R E 1 7 .4 Stages of stand development following major disturbance. (a) All trees forming the forest are
already present in the stand or colonize soon after disturbance. However, the dominant overstory tree species
changes as stem number decreases and vertical stratification of the species progresses. Height attained and
duration of each stand varies with species, site conditions, and disturbances. Barring intervening distur-
bances, the “old growth” stage may be reached in less than 200 to over 500 years. Source: After Oliver (1981) /
with permission of Elsevier. (b) Alternative diagram of disturbance-mediated accelerated succession in a
pioneer forest community. Species A is a pioneer tree, whereas B and C are later successional species.
Disturbances may include logging, windthrow, ice storm, fire, and insect/disease epidemic. The old-growth
stage of species B and C is not shown. Source: After Abrams and Scott 1989. Part (a) is reprinted from Oliver,
C. D. 1981. Forest development in North America following major disturbances. Forest Ecology and
Management 3(3): 156, © 1981 with kind permission of Elsevier Science-NL, Sara Burgerhartstrast 25,
1055 KV Amsterdam, The Netherlands. Part (b) after Abrams and Scott 1989. Reprinted with the permission
of the Society of American Foresters.
Piedmont Plateau in the southeastern United States, where abandoned agricultural land reverted
to pine-mixed hardwood forest. Both models also apply widely to forests in other regions with the
following characteristics: (i) developing following a major disturbance, (ii) having single or several
age classes, and (iii) having stems that regenerate during a relatively short period following distur-
bance. These models are described in the following text. By contrast, Watt’s beechwood example
was developed for forests on mesic sites and is most applicable in forest regions where small gaps
are characteristic. Franklin et al. (2018) developed a model we term the forest maturation model
that focuses heavily on the importance for biodiversity of early-successional forests, old forests,
and biological legacies (Table 17.4). It was developed for forests that experience infrequent, high-
severity disturbances such as those in the Pacific Northwest but is broadly applicable to forests
disturbed more frequently and/or less severely.
All models of secondary succession are initiated by a disturbance that occurs at fine to broad
scales. Such disturbances create marked structural changes in forests by killing trees, but their eco-
logical function is to free resources such as light, nutrients, and soil moisture as well as to create a
suite of biological legacies that will direct the rate and course of succession and forest redevelop-
ment. Described as the disturbance and legacy creation event by Franklin et al. (2018), such an
event is transitional between the undisturbed forest and the beginning of secondary succession
(Table 17.4).
In the stand initiation or establishment stage (also called reorganization), plants
regenerate the area, and new individuals and species continue to appear for several years. Regen-
erating plants may develop from (i) newly dispersed or buried seeds (e.g., species A in Figure 17.4a),
(ii) sprouts from stems and roots remaining from disturbed plants, and (iii) advanced regeneration
(species B, C, D)—saplings and seedlings of the previous forest understory that accelerate growth
when released by disturbance. Stand initiation may last from 5 to 100 years depending on the type
and severity of disturbance, species’ regeneration strategies, weather, herbivory, and many other
factors (Oliver and Larson 1996). During this period, invasion continues until resources (particu-
larly growing space) become limiting, and plant species diversity is often maximized.
Stand initiation is analogous to the preforest stage (Franklin et al. 2018), which precedes
the redevelopment of a closed-canopy forest. At this point in succession, diversity of both plants
and animals is relatively high, biological legacies are common, habitat niches and food sources are
numerous, and food webs and trophic relationships are complex relative to later points in
succession. Trees are typically present and are developing in the new forest, but do not yet domi-
nate the ecosystem. Incorporating such a non-forest stage into a conceptual model of forest
succession is critical because the lack of trees results in an open, structurally complex stage that is
often the most species-rich of the successional process (Swanson et al. 2011). The ecological impor-
tance of early-successional forests is often overlooked (Swanson et al. 2011; King and Schloss-
berg 2014), in part because such forests are perceived as recently disturbed ecosystems in the midst
of a “recovery” to a more recognizable, familiar forested condition.
The preforest or stand initiation stage is followed by another important transitional stage,
the forest canopy closure event, which creates sudden and dramatic changes in environmental
conditions (Franklin et al. 2018). This event represents the reassertion of strong dominance of a
site by trees, although the degree of dominance (and the speed at which canopy closure occurs)
varies among forest types depending on the density of the canopy. The most obvious environ-
mental changes that occur with canopy closure include the limitation of light at the forest floor
(Chapter 6) and the moderation of the microclimate (Chapter 7). However, canopy closure also has
significant influence on other ecosystem factors, such as competition and diversity. Ecosystem
processes such as biomass accumulation and carbon storage are also affected, because canopy leaf
area approaches its maximum at about the time of canopy closure, thus determining the rate of
photosynthesis, carbon allocation, and growth (Chapter 18). For example, in lodgepole pine forests
in Yellowstone National Park, Kashian et al. (2013) found that tree density strongly controlled
carbon storage only until canopy closure occurred, after which tree size and stand age became
more important because leaf area had been maximized.
Once canopy closure occurs, the processes that occur in the next stages are generally related
to competition or biomass accumulation. In the stem exclusion or thinning stage (also called
aggradation or young forest), virtually all growing space is utilized, crown closure occurs, over-
topped seedlings die, and tree regeneration is limited or ceases (Figure 17.4a), although herbs and
shrubs are typically still present in the understory. In addition, more vigorous trees grow larger and
capture the space from weaker trees, and severe self-thinning of the initial cohort takes place (Peet
and Christensen 1987; Figure 13.11, Chapter 13). Figure 17.4a illustrates that the pioneer species
A has been excluded and other species (B, C, D) assort themselves vertically depending on growth
rate and biomass allocation. Many examples of development in single-and multiple-cohort stands
are described in detail by Oliver and Larson (1996). In general, these competitive effects result in a
forest with relatively uniform structure. Biological legacies (mostly snags and fallen logs) remain
from the pre-disturbance forest, and many trees killed due to thinning are also present. Plant
species diversity is often minimized in this stage.
In the understory reinitiation, transition, or mature forest stage (Figure 17.4a), the
overstory begins to thin out or “break up” due to insects, pathogens, or other disturbances, thereby
increasing the light reaching the understory. Gaps may begin to form in the canopy, releasing some
of the smaller trees previously suppressed in the understory or subcanopy layers. Most overstory
trees have maximized their height growth and crown spread. Tree regeneration begins, and indi-
viduals of advanced regeneration may live for a few years or decades and then die. Longevity of
regeneration less than a meter in height in the understory varies substantially, for example, from
3 or 4 years in cherry bark oak to over 100 years for Pacific silver fir (Oliver and Larson 1996).
Together, these features significantly increase the structural complexity of the stand relative to the
stem exclusion stage. Pre-disturbance biological legacies have significantly decayed and declined,
but snags and logs from the post-disturbance forest begin to accumulate during this stage.
The old-growth, steady-state, or old forest stage (i.e., the late-successional forest)
develops as overstory trees age and increase in height and biomass (Figure 17.4a) while invasion
and regeneration continue. Senescence, chronic insects and pathogens, and minor and major dis-
turbances cause tree crowns to further disintegrate or entire trees to die and disintegrate singly and
in groups. As a result, some advanced regeneration and understory trees are able to reach the over-
story. Second- and even third-generation trees grow into the canopy, and thus canopy trees begin
to vary in their size. Very old forests may lack any representative trees from the original
post-disturbance cohort, such that a 500-year-old forest may lack trees older than 300 years! Struc-
tural diversity is therefore maximized in the old-growth stage, both in terms of the abundance of
diverse tree structures as well as the patchiness of structure (Franklin et al. 2018). For example, in
old-field Piedmont forests, hardwoods dominate over declining loblolly pines, and relatively even-
aged patches resulting from previous gaps now undergo a miniature version of the three previous
stages. Importantly, the old-growth stage is defined not only by tree age and size but also by abun-
dant biological legacies (such as snags and coarse woody debris) and ecological processes that are
very different from those in earlier stages (Franklin et al. 1981; Harmon et al. 1986; Oliver and
Larson 1996).
The importance of the old-growth stage for forest ecosystems has been emphasized by many
researchers and is reviewed extensively by Franklin et al. (2018). Old-growth forests are ecologi-
cally unique in North America in large part because twentieth-century forest management deemed
them unsuitable or “overmature” for timber production and other resources, resulting in their
harvesting and conversion to earlier successional stages. The structural diversity provided by the
old-growth stage is critical for its provision of many habitats and ecological niches, as well as for
providing critical living space for rare or endangered species. The old trees that persist within
many old-growth forests have ecological significance of their own, for their provision of habitats in
their cavities, large branches, and complex crowns; resistance to future disturbance because of
their thick bark; and their contributions to genetic diversity (Franklin and Johnson 2012). Old
trees that die continue to have ecological value as snags, logs, and coarse woody debris, all of which
are common in the old-growth stage. In general, the ecological importance of individual trees
could be summarized as increasing slowly as they age and increase in complexity, and then gradu-
ally declining once they die (Franklin et al. 2018).
There are many possible variations of the four-stage models presented in Table 17.4
depending on the site, vegetation, and disturbance characteristics specific to the ecosystem. One
alternative is illustrated in Figure 17.4b, where pioneer species A dominates in the first two stages
only to be disturbed in the midst of the understory reinitiation stage, causing accelerated succession
of species B and C. Thereafter, a B–C old-growth stage would be characterized either by areas lack-
ing disturbance, or perhaps by minor or major disturbance which, in turn, might lead to replacement
of B by C, C by B, or B and C by A! Logging or windthrow of the overstory when advanced regen-
eration is present (Figure 16.12, Chapter 16) is one example of the Figure 17.4b alternative, and
other specific examples are given by Abrams and Scott (1989).
Again, for both primary and secondary succession, complete, stage-by-stage successions
rarely, if ever, occur in nature. Disturbances constantly disrupt the gradual internal changes that
occur in an ecosystem and may set back, accelerate, or permanently change successional
trajectories. Moreover, we emphasize that succession is ecosystem-dependent, and a predictable
sequence of vegetation stages can be expected within the context of most ecosystems. Disturbances
often only temporarily alter the sequence unless they act to restart succession or permanently
change the site, as with erosion, deposition, fire, and windstorm. In such cases, the disturbance
may initiate an entirely new successional sequence. For example, on well-drained sand soils,
nutrients are retained primarily in the accumulated soil organic matter. A severe fire that destroys
the stand and the organic matter may actually change soil conditions so drastically that a new
succession is initiated (Damman 1975).
Even a vegetation stage itself can change the site to the degree that there is a resulting change
in succession. In Newfoundland, for example, Damman (1975) reported a relatively predictable
succession following fire on most sites unless a Kalmia heath became established. If such a heath
became firmly established after fire, it initiated soil changes leading to thin, iron-pan formation,
water logging, and peat bog formation, preventing the recovery of forest vegetation and creating a
whole new successional sequence. Although succession does not always follow predicted patterns,
disturbance-or vegetation-mediated changes in site conditions may strongly determine the trajec-
tory of the new pattern.
SUCCESSIONAL CAUSES, MECHANISMS, AND MODELS

This section describes some of the key characteristics, causes, mechanisms, and models of
succession. These features gained attention as the role of disturbance in forest ecosystems was
increasingly appreciated, as the number of ecosystems in which succession was observed increased,
and in reaction to the classical model.
Key Characteristics and Regeneration Strategies Succession functions as the

i nteraction of species with different genetically determined life histories and regeneration strat-
egies. There are likely limitless possibilities for differences between species in their regeneration
requirements, allowing replacement of the plants of one generation by those of the next
(Grubb 1977). Ecologists often attribute these traits and strategies to the species themselves, but
they are intimately related to the site conditions that elicited them (Chapter 3). The characteristics
of forest trees paramount to succession include: understory tolerance; seed production, size, and
dispersal; growth rate and longevity; tree crown architecture; resistance to insects and pathogens;
biomass production; allocation of carbohydrates; and nutrient requirements. It has been

hypothesized that the competitive ability of species is regulated during succession by the changing
ratio between nutrient availability and light (Tilman 1985, 1988), and that community composi-
tion should therefore change whenever the relative availability of the two limiting resources
changes. Known as the resource-ratio hypothesis, this idea has been criticized for its lack of general
applicability (Glenn-Lewin and van der Maarel 1992, p. 31) and because its predictions have not
been robustly tested (Miller et al. 2005, 2007).
The vital attributes concept (Noble and Slatyer 1977, 1980) identifies the characteristics of a
species that are vital in determining its role in successional sequences. Such attributes include
means of recovery from disturbance, ability to establish in the face of competition, and time needed
to reach critical life-history stages. The basic trait of differential longevity is also critical for some
successional sequences (Egler 1954, 1977; Drury and Nisbet 1973). For example, short-lived pio-
neer species may be replaced by long-lived, late-successional species that simply outlive them, and
they are not regenerated even by subsequent periodic disturbances.
Each of these characteristics (e.g., seed size, nutrient requirements, vital attributes, lon-
gevity, etc.) evolved in response to site-specific ecosystem conditions (Chapter 3), and broad gener-
alizations without an ecosystem context can be misleading. For example, in certain ecosystems
small seed size and a short-lived pioneer role are key factors in explaining successional pathways
where such species are replaced by heavier-seeded, more understory-tolerant, long-lived species.
However, not all small-seeded species are short-lived pioneers. For example, hemlocks (Tsuga) and
cedars (Thuja) are small-seeded, very tolerant, long-lived groups that persist in certain kinds of
ecosystems whose seedbed conditions (e.g., moist to wet sites; bare mineral soil) and disturbance
regimes (e.g., fire, erosion, and deposition) favor them. By contrast, walnuts are heavy-seeded but
intolerant and fast-growing; likewise, their vital attributes evolved and are expressed in specific
kinds of ecosystems wherever they persist today in North America and Asia. In summary, the key
life-history attributes and regeneration strategies are inherent parts of many overlapping, sequen-
tial, ecosystem-specific processes, such as regeneration (Chapter 4), competition and mutualisms
(Chapter 13), and biomass accumulation and growth (Chapter 18). These processes, involving both
site and plants, explain succession on an ecosystem-by-ecosystem basis. Walker and Chapin (1987)
describe this process model for primary succession in Alaska, and details are available in a number
of sources (Walker and Chapin 1986; Walker et al. 1986; Matthews 1992).
Availability and Arrival Sequence of Species Successional trajectories are strongly

affected by the availability of species and their arrival at a disturbed site. The timing and magni-
tude of species invasion vary depending on their occurrence in adjacent ecosystems, the spatial
position and distance of the disturbed site in relation to seed sources, the site conditions and dis-
turbance history of the disturbed area, and species life-history traits. Differential arrival often
explains why different successional pathways occur on similar sites. In secondary successions fol-
lowing windstorms or fires, surviving understory plants, clonal plants, and those germinating from
the soil seed bank have immediate access to the site. For example, the differential arrival of species
largely determined three pathways of primary succession in Alaska where a 220-year glacial retreat
progressively exposed parent material for colonization (Fastie 1995). The variability of species
arrival was related to the distance from each site to the closest seed source of Sitka spruce at the
time of deglaciation. In secondary succession following logging and fire in the Great Lakes region,
some eastern white pine seed trees were able to survive, and the variable successional pathways on
a given area today demonstrate the influence of these remnant seed sources (Palik and
Pregitzer 1994).
Facilitation, Tolerance, and Inhibition Connell and Slatyer (1977) proposed three
alternative models of succession based on the mechanisms of facilitation, tolerance, and inhibition
A disturbance opens up
a relatively large space
FACILITATION
Individuals of any species
Only certain ‘pioneer’
in the succession could
species are capable
establish and exist as
of becoming established
adults under the prevailing
in the open space
conditions
TOLERANCE INHIBITION
Modification of the Modification of the Modification of the

environment by early environment by early environment by early
occupants makes it more occupants has little or occupants makes it less
suitable for recruitment no effect on subsequent suitable for recruitment
of ‘late-successional’ recruitment of ‘late- of ‘late-successional’
species successional’ species species
In time, earlier species As long as earlier colonists

are eliminated through persist undamaged and/or
competition for resources continue to regenerate
with established ‘late- vegetatively, they exclude
successional’ adults or suppress subsequent
colonists of all species
The sequence continues

until the current resident
species no longer
facilitate the invasion If external stresses are
and growth of other species present, early colonists
and/or may be damaged or killed
no species exists that can and be replaced by
invade and grow in the species which are more
presence of the resident resistant
F I G U R E 1 7 .5 Three autogenic mechanisms—facilitation, tolerance, and inhibition—that determine

successional trends. Source: After Connell and Slatyer (1977) / University of Chicago Press, as modified by
Begon et al. 1990. Reprinted with permission from American Naturalist, © 1977 by the University of
Chicago. Modification reprinted with permission of Blackwell Science.
(Figure 17.5). Notably, many ecologists have found these models restrictive (Huston and
Smith 1987) and oversimplified (Christensen and Peet 1981) because each was developed to
represent a single, opposing, alternative pathway (Pickett et al. 1987). In reality, nature features an
enormous variety of pathways and mechanisms that may simultaneously influence successional
change in complex ways.
Facilitation is analogous to the model of relay floristics (Figure 17.3a), whereby early
successional species modify the environment to favor later successional species. In addition to the
amelioration of environmental stress, facilitation may operate by increasing the availability of
resources and enhancing colonization ability (Pickett et al. 1987). Thus, a species with one
combination of life-history traits may establish and modify environmental conditions in a manner
that favors individuals with other combinations (Huston and Smith 1987). Facilitation in this sense
has been reported in many instances (e.g., Cooper 1923, 1931, 1939; Jenny 1941; Christensen and
Peet 1981; Walker and Chapin 1986; Wood and del Moral 1987; Matthews 1992; Fastie 1995; Bau-
meister and Callaway 2006; Calder and St. Clair 2012). Chapin et al. (1994) investigated facilitation
during succession following deglaciation at Glacier Bay, Alaska (Figure 17.6). Life-history traits
determined the pattern of succession, and initial site conditions and facilitation (where present)
influenced the rate of succession as well as the composition and productivity of the late-successional
community. Facilitative effects were strongly related to the increase in nitrogen that is enhanced
by all plants but especially by alder.
Inhibition, where species established early inhibit subsequent colonization of other species
by pre-empting space and other resources, was also an important process at Glacier Bay (Figure 17.6).
A common example of inhibition is when shrubs resist colonization by trees (Niering and Egler 1955;
Webb et al. 1972; Niering and Goodwin 1974; Damman 1975; Chapin et al. 1994).
The mechanism of tolerance relates to the ability of a species to tolerate low resource levels.
In such situations, the initial species modify the environment, but this change has little or no effect
on subsequent recruitment and growth of later successional species (Figure 17.5). Instead,
late-successional species replace early-successional species either by active competition or by
simply being longer-lived. Active and passive concepts of tolerance are discussed by Pickett
et al. (1987).
Importantly, facilitation, inhibition, and tolerance describe relative, rather than absolute,
processes (Huston and Smith 1987). The three processes almost always occur simultaneously
with varying degrees of importance and are likely to occur in most primary and secondary
successions. Various combinations of the three processes are reported for primary succession on
glacier forelands (Matthews 1992; Chapin et al. 1994; Fastie 1995), in old-field succession in
New York (Gill and Marks 1991), and on the North Carolina Piedmont. On Alaskan floodplains,
STAGE: Pioneer Dryas Alder Spruce
FACILITATIVE Survivorship N (weak) SOM Germination

EFFECTS: Growth (weak) N
Mycorrhizae
Growth
INHIBITORY Germination Germination Germination Germination

EFFECTS: (weak) Survivorship Survivorship Survivorship
Seed predation Seed predation Seed predation
and mortality and mortality and mortality
Root competition Root competition
Light competition Light competition
N
F I G U R E 1 7 .6 Summary of facilitative and inhibitory effects (weak effects noted) for four successional
stages on establishment and growth of spruce seedlings at Glacier Bay, Alaska, as determined from field
observations, field experiments, and greenhouse studies. SOM = Soil organic matter. Source: After Chapin
Change in Ecosystems 431
many different processes and factors (e.g., life history, chance, facilitation, competition, her-
bivory, etc.) affect how alder and spruce interact during succession, and no single successional
process or model adequately describes successional change in these ecosystems (Walker and
Chapin 1987).
CHANGE IN ECOSYSTEMS
Ecosystems are constantly changing over time and space: from night to day, seasonally, and year to
year as climate and soil change, and as the cycle of organisms’ activity changes. As such, a regional
or local ecosystem never can and never does reach a complete balance or permanence except for
something arbitrarily chosen by humans for their own understanding.
END POINT OF SUCCESSION?

Forest succession has been recognized from the earliest days of natural history study, but whether
there is a last stage of succession has been debated for decades. Cowles observed in 1901: “The
condition of equilibrium is never reached, and when we say that there is an approach to the meso-
phytic forest, we speak only roughly and approximately. As a matter of fact, we have a variable
approaching a variable rather than a constant.” On the other hand, Clements defined the term
“climax” relatively rigidly as the end point of succession and used it as the driver of his vegetation
classification. Over decades many descriptors were added to the term to improve it conceptually,
such as “polyclimax,” “polyclimatic climax,” “site climax,” “climax pattern,” or position of relative
stability. Regardless of these modifications, the term implies a terminus of succession, and epito-
mizes succession within a vegetation context rather than an ecosystem context. In fact, the term
“climax” is inappropriate within an ecosystem context. It is more useful to recognize pioneer or
early-successional, mid-successional, and late-successional species or groups for position in char-
acterizing vegetation along the temporal scale for site-specific ecosystems. These terms (early, mid,
late) are not absolute but are relative to the specific ecosystem in the landscape.
In any case, climax terminology permeates the literature such that it is useful to examine
how the terms are used. Macroclimate is the dominant factor forming communities in Clements’
monoclimax theory, whereas other factors (physiography, soil, fire, biota) are of secondary impor-
tance. When other factors are of equal importance, a different climax community is possible or
even likely for different physiographic settings, soil types, and disturbance factors. This is the
polyclimax theory (Tansley 1939), which holds that biotic succession will move toward a climax
for any combination of environment and organisms, but that the specific nature of the climax will
vary with environmental factors and biotic conditions. Cowles (1899, 1901), Moss (1913), and espe-
cially Nichols (1923) contributed heavily to polyclimax theory. In fact, Nichols argued for a different
physiographic climax on each site that differed more or less from the regional climatic climax.
For example, in his classic studies of plant succession on the sand dunes of lower Lake
Michigan, Cowles at first thought that succession on all sand dune sites would eventually reach
the same forest stage of mixed mesophytic hardwoods. In a later study of the same area, Olson
(1958) concluded that the drier, lower fertility, and more exposed sites would never support such
a community but would instead be occupied more or less permanently with a black oak–blueberry
type. Furthermore, Olson suggested that although vegetation changes would continue to become
progressively slower with time, but they would never reach complete stability. Over the years, the
accepted concept of climax became broader, less rigid, and more closely approximated the poly-
climax approach. Tansley (1939) was instrumental in broadening the concept by adopting a
dynamic viewpoint in recognizing that “positions of relative equilibrium are reached in which
the conditions and composition of the vegetation remain approximately constant for a longer or
shorter time.”
Polyclimax theory is also mired in its own terminology: climatic, edaphic, topographic,
topoedaphic, fire, zootic, salt, etc. (see descriptions in Oosting 1956 and Daubenmire 1968). Many
of these terms approximate or are merely more specific distinctions of Clements’ climaxes. For
example, in the polyclimax approach the Clementsian climax was renamed “climatic climax;”
Clements’ subclimax may be either an “edaphic” or “topographic climax;” Clements’ disclimax
becomes a “fire” or a “zootic climax.” In addition, like Clements, only one climatic climax was
recognized. A polyclimatic climax theory was later developed (see Meeker and Merkel 1984) that
recognized more than one climatic climax for a macroclimatic region.
Whittaker (1953, 1975) recognized a gradational climax pattern of vegetation corresponding
to environmental gradients—a continuum of climaxes. Whittaker (1953) wrote that climax
composition was determined by
all “factors” of the mature ecosystem—properties of each of the species involved, climate, soil and other
aspects of site, biotic interrelations, floristic and faunistic availability, chances of dispersal and interaction,
etc. There is no absolute climax for any area, and climax composition has meaning only relative to position
along environmental gradients and to other factors.
The “climax pattern” continuum has clear problems in application to management, and so
Dyksterhuis (1949, 1958) proposed the use of site units to characterize succession and climax for
range classification. This approach is termed site climax by Meeker and Merkel (1984). Other
viewpoints and details of the climax concept have been presented by many authors
(Phillips 1931, 1934–35; Tansley 1935; Clements 1936; Cain 1939; Whittaker 1953, 1975; Dauben-
mire 1968; Langford and Buell 1969; White 1979; Matthews 1992).
SUCCESSION AS AN ECOSYSTEM PROCESS

A consistent thread throughout the abovementioned discussion is that a general model of
succession based on species or communities alone is unrewarding. It is somewhat useful to develop
a framework of stages based on the development of stands or populations, but specific site condi-
tions are the primary determinant of the duration, distinctness, and predictability of those stages.
Species’ life-history traits are inseparably linked to the sites and disturbance regimes to which they
are adapted. An ecosystem approach to forest succession provides a functional and geographical
framework for understanding these multiple mechanisms and processes that drive succession.
An ecosystem approach to forest succession requires that the first-order focus is the regional
and local ecosystem rather than component parts, such as vegetation or soil. Succession is a mul-
tidimensional process with many factors and processes occurring at multiple scales (Table 17.2)
that can be assessed for specific regional and local ecosystems. A “top-down” approach that
begins with an understanding of the geographic hierarchy of regional and local ecosystems, their
physical factors, biotic interrelationships, and past histories (Chapter 2) is often most appropriate
in this regard. Each regional ecosystem will exhibit unique successional pathways and
mechanisms, but their character and pattern will be further different among climatic–
physiographic regions.
For example, forest succession in the southern Appalachian Mountains is different than that
of the Yellowstone Plateau, the slopes of the Sierra Nevada, or the southeastern Coastal Plain. Like-
wise, the four ecosystem regions in Michigan (Figure 2.10, Chapter 2) exhibit different patterns of
succession due to differences in climate, physiography, soil, and vegetation. Within a given region
in Michigan, ecosystem districts (Figure 2.10, Chapter 2) have their own characteristic succes-
sional patterns for local ecosystems occurring as sand plains, acid swamps, calcareous swamps,
rocky ridges, or moist and fertile lower slopes. Variation in succession may even be expected for
local ecosystems (Simpson 1990) due to past disturbance or management (logging, drainage), dis-
turbance frequency and severity, seed source availability, and microsite variation, among other
Succession as an Ecosystem Process 433
Paludification
Thermal erosion
No fire
Balsam fir
Long period Bog

Fire
no fire
(Eastern Canada) Drying
Long period
? Bog filling
no fire
Fire
Black
White spruce
spruce
Fire
Floodplain No Wet, mesic, and dry sites
es
sites-no fire
s it
fire s
ic
Me
Paper birch
es
es
s it
s it
Dr y i
n
Aspen la
Flo o d p
es
it
ys
Dr
Jack pine
F I G U R E 1 7 .7 General example of successional pathways involving black spruce on different ecosystems,

defined by their site conditions, of the boreal forest in Alaska and Canada. Source: After Viereck (1983) /
John Wiley & Sons. Reprinted with permission from SCOPE 18, The Effects of Fire in Northern Circumpolar
Ecosystems, edited by Wein, R. W. and D. A. MacLean 1983 © SCOPE, John Wiley & Sons, Ltd.,
Chichester, UK.
factors. The factors that obscure successional trends are significantly reduced by specifically
defining the geographic framework and working with more or less homogeneous site conditions.
A general example of multiple, ecosystem-specific successional pathways is presented in
Figure 17.7 based on the site characteristics and disturbance regimes of the boreal forest of Alaska
and Canada. On many wet, mesic, and dry sites, black spruce replaces itself with or without fire
(Viereck 1983). In the southern regions of the boreal forest, colonization of burned black spruce
forest by birch and aspen is more common than in the north. Long fire-free periods in eastern
Canada enable balsam fir to replace black spruce (Figure 17.7). Many other ecosystem-specific
pathways are described by Viereck (1983).
A specific example of an ecosystem approach to succession comes from research on biomass
accumulation and nutrient cycling in northwestern lower Michigan (Figure 17.8; Zak et al. 1986;
Host et al. 1987; Host et al. 1988). Site-specific ecosystems were first distinguished at two levels:
landforms (outwash plain, ice-contact drift, and moraine) and ecosystem types within landforms
(named by overstory and ground-cover species). Successional trends for landforms and ecosystem
types were then estimated based on an understanding of site conditions and life-history traits of
the species in the understory and ground-cover layers (Figure 17.8). Importantly, the oak cover
types failed to distinguish the landforms and ecosystem types as well as the predicted successional
pathways, suggesting that an understanding of entire ecosystems—not just communities defined
by dominant species—is necessary to ascertain successional pathways. Moreover, different
pathways were predicted for ecosystem types within each landform, and these pathways, based on
multiple ecological factors, are the best estimate of successional trends available for making
Landform Society of Current Apparent

(Surface Parent American Foresters Dominant Future Compositional
configuration) Materials Cover Type Ecosystem Understory Changes
Loss of Oaks in Overstory

White Pine
Black Oak- Conversion to White Pine
Outwash Plain
Well-Sorted White Oak/ ?
(Level to gently
Medium Sand Vaccinium
undulating) White Oak- White Oak Persistence of White Oak
Black Oak- (Suppressed) Formation of Oak-Sadge Savanna
Well-Sorted Northern Red Oak
Medium Sand Mixed Oak/ Loss of Oaks in Overstory
With Loamy Red Maple
Trientalis Conversion to Red Maple
Ice-contact Subsurface Textural Northern
Terrain Bands Red
(Hilly) Medium Sand Oak Loss of Oaks in Overstory
Mixed Oak/ Red Maple
With Clay Loam Conversion to Red Maple
Viburnum
Subsurface Textural or Northern Hardwoods
Bands
Sugar Maple- Loss of Red Oak in Overstory
Sugar Sugar Maple
Fine or Loamy Sands Maple Red Oak/ Replacement by Sugar Maple or
Maianthamum Sugar Maple and Beech
Moraine
(Hilly) Fine or Loamy Sands Loss of Basswood in Overstory
With Sandy Clay Loam Sugar Maple Sugar Maple-Basswood/ Sugar Maple Replacement by Sugar Maple or
Subsurface Textural Basswood Osmorhiza
Sugar Maple and Beech
Bands
F I G U R E 1 7 .8 Estimated future compositional change associated with two levels of landscape ecosystems
(landforms and ecosystem types) in northwestern lower Michigan. Source: After Host et al. (1987) / Oxford
University Press. Reprinted with the permission of the Society of American Foresters. Site conditions,
overstory cover type (Eyre 1980), and current dominant understory species are also shown.
management decisions. Being specific to ecosystems that recur throughout the landscape allows
the proposed trends to be monitored and tested over time. Such an approach is applicable for
improved management, conservation, and restoration of ecosystems on many or most public and
large private lands in North America.
EXAMPLES OF FOREST SUCCESSION

PRIMARY SUCCESSION ON RECENTLY DEGLACIATED TERRAIN
Primary succession is perhaps best studied on the land newly formed in front of a glacier, and such
detailed studies of the glacier foreland have been underway since the late 1800s. Matthews (1992)
provides a comprehensive review of primary succession on recently deglaciated terrain, examining
physical and ecological processes and successional stages. From his geoecology viewpoint,
succession in front of the retreating glacier is a process of the developing landscape when all parts
of the ecosystem develop simultaneously. The most important visible features of the landscape in
the glacier foreland are the landforms and their associated parent materials, and the developing
vegetation is a function of the dominant factor terrain age. Other factors are subordinate: climate,
physiography, parent material, organisms, as well as physical site change and disturbance.
Matthews found that vegetation chronosequences are generally limited by differences among sites,
both the initial environment and their subsequent environmental histories, but their careful use
can provide approximations of succession at particular sites. For example, the chronosequence at
Glacier Bay, Alaska is a series of eight stages that occur along a continuum:
I. Early pioneer stage: mountain avens (Dryas drummondii) and prostrate willows, mosses,
and herbs.
II. Dryas mat: coalescing mats of Dryas.
III. Late pioneer stage (after 15–20 years): young Sitka alder and black cottonwood shorter
than 2–3 m.
Examples of Forest Succession 435
IV. Open thicket stage (after 20–25 years): clumps of alder.

V. Closed thicket stage (after 25–30 years): dense alder with individual erect specimens of
balsam poplar, shrubby willows, and Sitka spruce.
VI. Balsam poplar line stage (after 30–40 years) where poplars reach a height of 4–5 m and
form a canopy and visible line above the alder.
VII. Spruce forest stage (at 75–90 years) when spruce replaces alder.
VIII. Spruce–western hemlock stage (speculative because the forest may remain dominated by
spruce even after 200 years).
The horizontal and vertical structure of vegetation on the glacier foreland of the Grand
lacier d’Aletsch in the Swiss Alps has long been studied with permanent plots (Figure 17.9). The
G
sequence of herbs to shrubs to trees typically reflects the sequence of plant size, longevity, and
competitive power, but trees may colonize much earlier in the sequence depending on seedbed
conditions and seed availability. In the Canadian Rocky Mountains, establishment time ranges
from 10 to 15 years to 80 years (Luckman 1986).
Matthews’ review of primary succession in the glacier foreland reveals that succession may
proceed in trajectories that are parallel, but also in those that are strongly divergent or convergent.
Generally, divergent pathways are most clear during early stages and are associated with relatively
severe physical environments; convergence is favored by relatively strong biotic controls. Biomass
and species richness tend to increase during primary succession, although patterns of diversity
vary considerably from site to site (Matthews 1992).
Matthews notes that most of the general successional models (as previously described) are
not well applied to recently deglaciated terrain. Such land instead is better described by a geoeco-
logical model that emphasizes physical environmental processes as well as biological processes. In
addition, models of succession as a landscape process must incorporate spatial variation in site
F I G U R E 17 .9 Coverage of bare ground and

five vegetative strata on recently deglaciated 150
terrain, which is on terrain of increasing age,
Grand Glacier d’Aletsch, Swiss Alps.
Source: After Lüdi (1945). Reprinted with
permission of the 17, Stiftung Rübel, Zurich, 100
Switzerland.
% cover
50
0
4–6 25 45 70 85
Terrain age (years)
Bare ground Dwarf shrubs (<0.5 m)
Mosses and lichens Shrubs (>0.5 m)
Herbs Trees (>2 m)

characteristics through time. In sum, Matthews strongly emphasizes the properties (landforms
and glacial sediments, climate, soil) and processes (erosion by water; frost weathering, heaving,
and sorting; solifluction; wind deposition and erosion; and soil development) of the physical
landscape.
Matthews’ summary points concerning succession on recently deglaciated terrain provide
real-world insights into the mechanisms of succession discussed previously in this chapter:
Allogenic and autogenic processes interact during succession and are often inseparable.
Establishment is more important than migration (species arrival) in determining species composi-
tion in the harsh environment of glacial forelands. Here, it is likely that pioneers require some
special physical, chemical, or microbial conditions as well as lack of competition.
Multiple processes—including physical as well as biological processes—are characteristic in
replacement of pioneers by later colonizers. Pioneer species may decline more due to physio-
chemical changes associated with leaching (nutrient depletion, balance of available nutrients,
toxicity associated with pH changes) and other physical changes than to the rise of species of the
next successional stage.
The action of allogenic factors (physical processes and disturbances) as compared with auto-
genic action of the plants (the importance of life-history traits) influences both the nature and rate
of succession in the harsh physical environment of the glacier forelands. In favorable environ-
ments (high in resources; low in stresses), there is a rapid increase in the importance of autogenic
processes through time (Figure 17.10, see “1”). This increase may reflect the rapid accumulation of
biomass that occurs as soil water and nutrients become more favorable in a favorable climate or a
relatively undisturbed site. By contrast, allogenesis, as a determinant of vegetational change, is
reduced sharply (1′) as the intensity of physical processes and disturbances decrease plant biomass
and vegetative control of the environment increases. As the environment becomes less severe,
biological production is constrained, allogenic processes become relatively more important in any
successional stage, and the control of succession by allogenic processes becomes longer. The
relative importance of allogenesis in Figure 17.10 (curves 2′ and 3′) is highest in more severe envi-
ronments. That is, on very severe sites, plants themselves modify the environment to a lesser extent
than on favorable sites. This relationship can be applied at both the broad scale of regional
100 F I G U R E 17. 10 Relative importance of

autogenic (1, 2, 3) and allogenic (1′, 2′, 3′)
3´
processes under increasing environmental
severity (1 and 1′ with solid line = least severe; 2
Relative importance [%]
2´ and 2′ with dashed line = moderately severe; 3

and 3′ with dotted line = most severe) in the
1´
context of recently deglaciated terrain. The model
50 assumes species/community context with
1 autogenic processes relating only to vegetational
processes; allogenic processes are a combination
2 of physical processes and disturbance. Source:
After Matthews (1992) / Cambridge University
3 Press. Reprinted from The Ecology of Recently
Deglaciated Terrain, © 1992 by Cambridge
0 University Press, Reprinted with the permission
Successional time of Cambridge University Press.
ecosystems (increasing latitude, for example, boreal forest, tundra, polar desert) and the local scale
within the different landforms of the glacier foreland. As described in the earlier text, however,
autogenic and allogenic processes are not mutually exclusive, and their relationship to site severity
applies only to the extent that these processes can be distinguished. It appears likely that they are
most distinct in the harshest environments.
SUCCESSION FOLLOWING THE ERUPTION OF MOUNT ST. HELENS

On May 8, 1980, Mount St. Helens erupted in the Cascade Mountains of Washington in the Pacific
Northwest. The eruption produced a catastrophic disturbance to approximately 600 km2 of
old-growth and planted forests dominated by western hemlock, Douglas-fir, and Pacific silver fir.
The disturbance created by the eruption was extremely heterogeneous due to the diversity of geo-
logical processes that occurred. These processes included pyroclastic flows (a landslide and flows
of hot gases and pumice fragments) that eliminated ecosystems and landforms over about 60 km2,
a lateral blast zone that blew down or scorched forests over 500 km2, and the deposition of more
than 5 cm of ash over 1000 km2 (Dale et al. 2005a). Whether organisms survived the eruption
depended heavily on which of these zones they resided in prior to the disturbance, with little
survival in the area of pyroclastic flows, some survival of underground plant parts and burrowing
animals or organisms protected by snow in the blast zone, and greater survival of taller organisms
in the ashfall zone. The presence of snow was a major factor in organism survival and the
subsequent rate and trajectories of succession (Figure 17.11).
Much like we observe in areas of severe fires or windstorms, the heterogeneity of distur-
bance intensity resulting from the eruption had important consequences for post-eruption
succession. Unlike fires or windstorms, the pyroclastic zone generally buried or scoured away any
organic material toward its terminus (Dale et al. 2005b), and succession there was therefore mostly
primary (Dale and Denton 2018). However, secondary succession characterized the blast zone and
the ashfall areas because there was survival of some to many organisms. Remarkably, some indi-
viduals of most plant species present before the eruption survived the disturbance in refuges of
various forms (Del Moral 1983; Halpern and Harmon 1983; Adams et al. 1987; Zobel and
Antos 1992), and survivors represented most life-forms and successional stages. For example,
snowbanks that had persisted into early May of 1980 protected the tree seedlings and shrubs that
they had buried, allowing them to immediately resume growth upon snowmelt (Franklin and
MacMahon 2000). Together with populations dispersed into the disturbed area, these survivors
facilitated and shaped secondary succession (Dale and Crisafulli 2018), and succession occurred
most rapidly from areas with surviving organisms (Franklin and MacMahon 2000).
Succession following the eruption of Mount St. Helens has been extensively studied, and a
series of important lessons have emerged. The concept of biological legacies emerged from the
aftermath of the Mount St. Helens eruption, and there they were found to be extremely important
in directing the rates and trajectories of succession (Dale et al. 2005a). Pre-eruption legacies
included dead organic material (mainly trees blown over or left standing but scorched in the blast
zone), but also organisms that survived the blast but died before they were able to reproduce (Dale
and Crisafulli 2018). Ecologists had largely ignored the importance of such legacies for ecosystem
structure prior to 1980, but they were found to be important in shaping post-eruption erosion and
deposition processes as well as supporting many organisms with protective cover, food, habitat,
and nutrients (Franklin and MacMahon 2000).
An additional important lesson provided by succession at Mount St. Helens is that succession
is extremely complex and does not always follow ecological theory. Heterogeneity in organism
survival and in the distribution of biological legacies created many different points of initiation of
succession across the landscape that resulted in many different pathways of succession that did not
necessarily converge. In some cases, adjacent sites undergoing primary succession supported very
F I G U R E 1 7 .1 1 Top: Photo chronosequence taken from 1983 to 2004 showing changes in plant community
development in the blowdown zone of Mount St. Helens at a site that was snow-covered at the time of the
eruption. Snow cover facilitated the survival of numerous plants, including conifer saplings. Bottom: Photo
chronosequence showing changes in plant community development from 1983 to 2016 in the blowdown
zone at a site that was not snow-covered at the time of the eruption. Source: From Dale and Crisafulli (2018);
photos by Peter Frenzen, USDA Forest Service.
different plant communities, emphasizing the importance of random or chance events in the suc-
cessional process (Dale and Crisafulli 2018). Specific to forests, several late-successional coniferous
species (e.g., western hemlock, Pacific silver fir) were found to colonize the primary succession
sequence on the debris flow area within the first 3 years after the eruption, although seedling
density greatly decreased by 5 years post-disturbance (Dale et al. 2005b). These results highlight
the differences between ecological theory and the reality of ongoing succession in real-world eco-
systems following complex disturbance events. Other succession research completed at Mount St.
Helens through 2015 is summarized in detail by Crisafulli and Dale (2018).
SECONDARY SUCCESSION FOLLOWING FIRE IN PONDEROSA

PINE FORESTS OF WESTERN MONTANA
Fire leads to ecosystem change throughout forests of the western United States (Chapters 10
and 16). Fischer and Bradley (1987) described the major role of fire in determining successional
pathways in grasslands and forests in western Montana. For example, ponderosa pine of the sub-
montane zone in western Montana competes with grasses in dry ecosystems at low elevations and
with Douglas-fir in the cooler and moister ecosystems of the montane zone. In this submontane
zone, fire acts to (i) maintain grasslands, (ii) maintain open ponderosa pine stands, and
(iii) encourage ponderosa pine regeneration. Frequent fires maintain the grassland community by
killing pine seedlings (stage 1; Figure 17.12a). Over occasional long fire-free periods, ponderosa
pine seedlings become established gradually as the result of a seedbed-preparing fire (stage 2). The
seedlings develop into saplings in the absence of further burning, and fires during this period may
kill young trees (stage 3) or thin them (stage 4). In time, light surface fires produce an open stand
of mature trees (stage 5) that may accumulate enough fuel to produce a stand-replacing fire
(stage 6) which is followed again by grassland (Figure 17.12a).
Multiple successional pathways are evident in these forests using a starting point of open,
park-like, old-growth ponderosa pine forest (Figure 17.12b). The interaction or absence of fire
(timing and severity) and the life-history traits of pine and grasses result in various states of pine
stands (e.g., A, B2, C3, E2, etc.) and pathways. For example, starting with an open, park-like,
old-growth pine stand (A), frequent fires (pathway 33) maintain grassland (K), but moderate-
severity fire (pathway 5) leads to a closed-canopy, multistoried pine stand (stage D1). A dense stand
(El) occurs in the absence of fires and is either maintained by low-severity fire (pathway 7) or suc-
ceeds to grassland (F) after severe fire (pathway 9). This example illustrates the complexity of
succession and the lack of any end point of the successional process. Similar examples of fire
ecology for the diverse range of ecosystems of eastern Idaho and western Wyoming are also avail-
able (Bradley et al. 1992).
Quantitative succession models are beyond the scope of our treatment, but many simulation
models have been developed for succession (Huston et al. 1988; Botkin 1993). A mechanistic,
ecological process model simulating fire succession of coniferous forests of the northern Rocky
Mountains is also available (Keane et al. 1996), as is a commonly used forest succession simulator
with versions specialized for individual forest regions of the United States (Crookston and
Dixon 2005). Recent model development also includes a spatially explicit landscape-scale
succession model capable of incorporating climate change and additional disturbances (Scheller
et al. 2007).
SECONDARY SUCCESSION AND GAP DYNAMICS

Forest trees may establish and successfully reach the overstory canopy via three modes
(Veblen 1992): (i) following severe disturbance, (ii) by establishing in overstory gaps, or (iii) rarely,
by continuous regeneration in a disturbance-free understory and recruitment into an overstory
without discrete gaps. Understory intolerant species typically reach the overstory following severe
disturbances, whereas mid-tolerant and tolerant species often utilize relatively small treefall gaps.
Tolerant species may recruit without disturbance if the canopy is not too dense, but they usually
establish in a disturbance-free understory (Lorimer et al. 1988; Busing 1994) and then require at
least periodic episodes of overstory crown thinning. We examine regeneration in canopy gaps in
this section.
When an opening or gap occurs in the forest overstory, advanced regeneration and new
stems colonizing the gap from the understory form a patch of vegetation that responds to the newly
changed environmental conditions—leading to a type of secondary succession termed gap
dynamics (Pickett and White 1985). Watt (1925, 1947) described four phases (Table 17.4) in the
development of pure European beech forests in England. One of these four phases was the “gap
phase” that occurred in old-growth beech stands. Gaps are initiated by major and minor
(a)
Fire-
maintained
grassland
2
6
Grass/forb Prepare seedbed
Stand-destroying fire
(rarely occurs)
Seedlings
PIPO
Crowded, two-
3
storied stand
(rarely occurs) Return to grassland
PIPO
Saplings 4
PIPO Thinning fire
Pole stand
PIPO
Fire-
maintained
open, park- 5
like stand
PIPO Generalized forest
succession
Cool fire
Severe fire
F I G U R E 1 7 .1 2 The role of fire in forest succession in warm and dry ponderosa pine (PIPO) ecosystems.
(a) Generalized diagram of succession illustrating situations where grassland is maintained or where
ponderosa pine is favored by a less-frequent fire regime. (b) Specific diagram of multiple successional
pathways starting from an open, park-like old growth ponderosa pine stand. Numbers (1, 2, etc.) indicate
successional pathways; letters (A, B) indicate states of pine forest or grassland. Source: After Fischer and
Bradley (1987) / U.S. Department of Agriculture / Public Domain.
(b)
M
M
od
od
Se
er
er
5
ve
3 8
at
at
Most
re
e
e
A 1 B1 C1 D1 E1
Open, Open PIPO Scattered Closed Crowded
park-like, stand; uneven- PIPO overstory; canopy multi- PIPO stand ;
old growth aged regen. dense PIPO storied PIPO structure Severe
PIPO understory stand varies 9
2 4 7
Se
Low Low Low
ve
6
re
33
M
An
os
t
10 Severe
C2
14 Severe
B2 E2
frequent fires
Open
Scattered D2 Broken, 16
PIPO over- Stagnant F
PIPO stand; story; dense decadent
PIPO pole Grass
evenaged PIPO PIPO
understory stand stand
regen.
e
re 13 15 ver
ve Se
12 Low Low 20
Se 11 Low-
Moderate
C3
Scattered E3
Severe PIPO over- D3 Open
story; sparse Open PIPO park-like Go to A
17 PIPO pole stand PIPO stand
understory
K 18 Low 19 Low 21 Any

Fire-
maintained
grassland
Succession in absence of fire

Response to fire
Low Cool or light surface fire
Mod. Fire of intermediate (moderate) severity
Severe Hot, stand-destroying fire
F I G U R E 17 .1 2 (Continued)
disturbances, and most emphasis has been placed on severe disturbances such as windstorms,
fires, flooding, and landslides (Chapter 16). However, fine-scale or tree-fall gaps also significantly
affect the structure and composition of mesic forests around the world, especially those dominated
by understory tolerant trees.
Small canopy gaps are characteristic of many mesic forests of eastern and western North
America. Larger, major disturbances occur at very long intervals in these forests, such that small
gaps and slight but chronic disturbances to overstory crowns necessarily play a major role in tree
regeneration. Lorimer (1977, 1980) found that mesic pre-European colonization forests in eastern
North America and the Appalachian Mountains were characterized by all-aged stands—stands
with an inverse J-shaped size-class distribution (Figure 13.9, Chapter 13). In mesic old-growth for-
ests of eastern North America, Runkle (1982) estimated average annual canopy disturbance rates
of less than 1% per year, suggesting that it requires 110–125 years on average for the entire canopy
to be disturbed. For both tropical and temperate forests, Veblen (1992) estimated an average canopy
rotation time of approximately 100 years and a range of 50–575 years.
Late-successional species can perpetuate themselves more or less indefinitely in and near
small canopy gaps in the absence of larger or more severe disturbances. Regeneration of species
such as beeches, sugar maple, western and eastern hemlocks, firs, and basswoods occurs when
persistent seedlings in the understory fill the gaps, excluding less-tolerant species that need to first
establish in the newly formed gap. A canopy profile through a single-tree gap in a sugar
maple–beech forest is illustrated in Figure 17.13a. The gap is 50 m2 in the tree canopy and affects
250 m2 at ground level and formed from the breakage of a single stem 13 m above the ground about
3 years before the profile was made. Such small gaps may be filled by (i) adjacent canopy trees
expanding branches into it or (ii) already-established saplings filling it from below. This process
takes place in temperate forests in gaps smaller than approximately 0.04 ha; the abundance of
intolerant species sharply increases in larger gaps (Runkle 1985; Stewart et al. 1991; Busing 1994).
Many authors report that two or more gaps are required for most saplings of tolerant species to
reach the overstory. Mean residence time in the understory for tolerant species is often long; for
sugar maple, it is often over 100 years and can be as long as 230 years (Barden 1981; Canham 1985).
Mesic Douglas-fir–western hemlock forests of the Pacific Northwest appear to exhibit slower
gap dynamics than other forests (Spies et al. 1990). Gap formation is typically less than the range
of 0.5–2.0% reported for temperate deciduous forests (Runkle 1985). Nevertheless, vertical forest
structure provides an important context for gap dynamics in these forests. In Figure 17.13b, a small
gap occurs in an old-growth Douglas-fir–western hemlock forest whose overstory trees are over
twice as tall as those in the beech–sugar maple forest in Figure 17.13a. Because of their great height
and narrow crowns, the death of one or a few conifers in such forests may not transmit enough
light for the regeneration of mid-or intolerant species (Canham et al. 1990; Spies et al. 1990). Gaps
may remain important for the regeneration of Douglas-fir and other early-successional species
where trees are shorter and canopies less dense, such as mixed-conifer forests (Spies and
Franklin 1989).
Larger gaps may allow mid-tolerant and intolerant species (opportunists or gap-phase
species) to establish and reach the overstory. Such species have light, wind-disseminated seeds
with the ability to germinate on small patches of exposed mineral soil, such as those created by the
uptorn roots of windthrown trees. Juvenile growth under conditions of partial shade and partial
root competition is rapid so that at least some seedlings can outgrow and overtop already-established
tolerant species in the openings. Typical gap-phase replacement species include yellow birch
and white ash, and with increasingly larger gaps include tulip tree, northern red oak, black cherry,
and sweet birch. For example, Webster and Lorimer (2005) found that yellow birch was able to take
advantage of gaps as small as 100–400 m2 in hemlock–northern hardwood forests in northern Wis-
consin, although those most successful in reaching the canopy were already present at the time of
gap formation.
Small gaps create a shifting mosaic during the transition and old-growth stages of mesic for-
ests (see Table 17.4), but the forest canopy is not a simple closed-canopy matrix punctuated by gaps
(Lieberman et al. 1989). Tree crowns thin with age as their branches die, are sheared off by wind,
or are broken by colliding with the branches of adjacent trees. The intensity of light reaching the
forest floor therefore changes over time, and gaps represent an extreme condition where light may
travel unimpeded to the forest floor. Latitude and the height of trees determine how much light
penetrates beyond the perimeter of the gap itself and under the branches of surrounding overstory
trees (Canham et al. 1990). In addition, succession in canopy gaps depends on many factors in
addition to light, because of spatial variability in the amount of vegetative cover, soil water and
nutrients, occurrence of coarse woody debris, and bare soil. Other factors include those associated
with different regional and local ecosystems: latitude, physiography, drainage, nutrient cycling,
tree species composition, disturbance regime, mutualisms with animals, and mycorrhizae. Of pri-
mary consideration in assessing gap dynamics is the disturbance type that creates the gap. For
example, the specific type of wind damage can affect succession; trees uprooted by wind may form
larger gaps and provide bare soil for colonizing species, whereas stem breakage may favor existing
advanced regeneration in a smaller, more shaded gap.
Runkle (1981, 1982, 1985, 1990) discovered in a variety of mesic forests of the eastern United
States that small gaps 50–200 m2 favor tolerant species but allow opportunists to persist in low den-
sities. Gaps caused by the death of one species that favor seedlings of another species (i.e., reciprocal
F I G U R E 1 7 .1 3 Examples of tree-fall (a)

gaps in eastern and western North 30
American ecosystems. (a) Overstory
canopy profile through a single-tree
FA
gap. The gap, 50 m2 in the tree canopy 25
FG
and affecting 250 m2 at ground level, AS
was formed when one stem broke at a 20 AS
height of 13 m. Crowns with solid
Distance (m)
AS AS
lines are within 5 m of the observer,
and crowns with broken lines are 15 AS AS
between 5 and 10 m of observer. FA, FG AS
Fraxinus americana (white ash); FG, 10
Fagus grandifolia (American beech); AS
AS, Acer saccharum (sugar maple). FG
FG
5
The vertical lines in the understory
are FG; thick lines are AS. (b) Vertical
and horizontal profile through a 0
450-year-old Douglas-fir/western 0 5 10 15 20 25 30
hemlock stand in western Oregon. Distance (m)
Source: After Spies et al. (1990) / (b)
Canadian Science Publishing. 60 70
Reprinted with permission of PM
The small gap in the transect center
(8–10 years old) was created when a 60
50
large Douglas-fir broke off and fell on
a smaller hemlock, breaking its
crown. Tree regeneration consists
entirely of western hemlock seedlings 50
40
less than 0.5 m tall. PM, Pseudotsuga
menziesii (Douglas-fir); TH, Tsuga
heterophylla (western hemlock); G,
Meters
40
canopy gap. Source: After Runkle 30
TH
(1990). Reprinted with permission of
30
20
20
10
10
N
10 m
G
TH
PM
Snag
replacement; Fox 1977; Woods 1979, 1984; Runkle 1981) are more common in low-diversity forests
such as those dominated by beech and sugar maple stands, although exceptions exist. In the
southern Appalachian Mountains, for example, four mid-tolerant to intolerant species—northern
red oak, white ash, black cherry, and tulip tree—occupied 3% of the canopy area via infrequent
success in gaps, especially those formed by multiple trees (Barden 1981).
Gap Specialists: American Beech and Sugar Maple American beech and sugar maple,
two of the most common species in eastern North America, are long-lived, very tolerant, late-
successional species of mesic forests where severe disturbances are very infrequent. These two
species are the most likely of all deciduous species to perpetuate themselves over many centuries
because they are able to regenerate in shaded understories and utilize treefall gaps to gain the
canopy. Where the two species co-occur, beech occurs over a wider variety of sites because it is less
nutrient-demanding, tolerates wetter sites (Crankshaw et al. 1965; Leak 1978), and is more shade
tolerant than sugar maple. For example, beech occurred on 256 soil types in pre-European
colonization forests of Indiana, second only in distribution to white oak (Crankshaw et al. 1965).
However, many investigators have sought to explain the co-existence of these late-successional
species in mesic forests by their performance in relation to the light relations of forest understories.
A long-term study of beech and sugar maple coexistence is a 16-ha beech–sugar
maple-dominated ecosystem that is part of Warren Woods in southwestern Michigan. Beech was
the main overstory species in 1933 as it was prior to European colonization in the surrounding area
(Kenoyer 1933). At that time, beech had 10 times the number of trees >15 cm diameter at breast
height (DBH) and represented 82% of the basal area (Cain 1935). Three-fourths of the treefalls
from 1946 to 1976 were of beech (Brewer and Merritt 1978). Meanwhile, sugar maple was more
common as seedlings and saplings. Long-term studies of the understory show that both species
maintained themselves over the entire area without marked replacement of one by the other, a
relationship termed allogenic coexistence (Poulson and Platt 1996). In this case, “allogenic” is
emphasized because wind creates the treefall gaps that appear to explain the observed pattern of
species interaction. Notably, autogenic ecosystem factors, such as old (300 years), tall (30–40 m)
trees with broad crowns, hollow trunks, brittle wood, and shallow rooting, also are significant in
gap formation, which has increased markedly since 1975 in this aging stand.
Beech and sugar maple understory individuals are able to coexist in this ecosystem due to
their differences in understory tolerance, tree architecture, and how they respond to light levels
both in the understory and in treefall gaps (Runkle 1985; Canham 1988; Poulson and Platt 1996).
Beech replaces sugar maple in the deeply shaded understory and in small gaps where its saplings
exhibit horizontal (plagiotropic) growth to obtain light. Sugar maple is favored where multiple
treefall gaps occur; sugar maple has more rapid vertical (orthotropic) extension growth enabling it
to quickly colonize the gap. Where a mix of treefall gaps of different sizes occurs at appropriate
frequencies, both species may coexist at near-equal abundances over the forest as a whole
(Figure 17.14).
Despite the importance of species trade-offs based on light relations, the Warren Woods
example is incomplete without consideration of soil properties and nutrient and water relations
specific to this local ecosystem. Soils at Warren Woods are sandy and strongly acid, and the high
spring water table tends to restrict rooting to the acidic topsoil. The less nutrient-demanding
understory beech should be favored over sugar maple on the low nutrient status of this site and
contribute significantly to beech dominance with minimum canopy disturbance. This relationship
was reflected by the overwhelming dominance of beech overstory in 1933 (Cain 1935). As wind-
storm disturbances increase in frequency within the forest as they have over the last several
decades, understory dominance may shift to sugar maple (Figure 17.14, panel 4). The proportion
of the species in the understory and overstory may reflect soil and nutrient differences in other
ecosystems where sugar maple is favored, provided sufficient light is available for sugar maple
regeneration. For example, sugar maple dominates only in two favorable ecosystems in the White
25 YR BEFORE TREEFALL 25 YR AFTER TREEFALL
no gaps nearby
local gaps every 10 × 10

100–200 yr metres
<1% full sun (1)
BC 6–8 MC 1
canopy tree ratio
huge B canopies cast deep shade; multiple M winners in a huge B treefall
rare sapling B are suppressed
nearby gaps every

50–100 yr
local gaps every

60–100 yr (2)
1–3% full sun
BC 3–4 to MC 1
many B in understory; M usually B replaces almost every treefall
die during first suppression
nearby gaps every

20–50 yr
local gaps every

40–60 yr (3)
1–6% full sun
BC 2–3 to MC 1
more B than M in understory; M or B replace B;
M endure 1–4 suppression cycles B almost always replaces M
nearby gaps every

0–20 yr
local gaps every

0–40 yr (4)
1–30% full sun
BC1 to 1–2 to MC
more M than B in understory; gaps grow with time; M replaces most
M rarely completely suppressed B and some M; B replaces some M
F I G U R E 1 7 .1 4 A conceptual model of coexistence of beech (B) and sugar maple (M) at Warren Woods,
southwestern Michigan. The sequence of four panels shows that infrequent and small treefall gaps (panel 1)
favor beech (open silhouettes), whereas frequent and large gaps (panel 4) favor sugar maple (closed
silhouettes) in the understory. The left panel of each pair, 25 years before treefall, shows how the understory
percentages of full sunlight levels, due to branch gaps and nearby treefall gaps, affect forest structure and the
relative success of beech and sugar maple. The right panel, 25 years after a treefall (x marks the former
position of the crown top of a fallen tree), shows responses of species to gaps. Vertical and horizontal axes
are drawn to a 10-m scale, and only individuals >1 cm DBH and >l–2 m tall are shown. Bc = beech canopy
tree; Mc = sugar maple canopy tree. Source: After Poulson and Platt (1996) / John Wiley & Sons. Reprinted
with permission of the Ecological Society of America.
Mountains of New Hampshire, whereas beech is more abundant in eight (Leak 1978). In addition,
in the heavy-textured, calcareous morainal soils that supported beech–sugar maple forest prior to
European colonization in central and southeastern Michigan (Quick 1923; Veatch 1953, 1959),
sugar maple appears to be much more vigorous than beech and the tree replacement patterns differ
from those at Warren Woods. For example, in 1997 in an old-growth forest on a clay-rich, calcar-
eous, nutrient-rich moraine in southeastern Michigan, sugar maple dominated the overstory (49%
sugar maple to 23% beech for basal area), and its saplings overwhelmingly dominated the under-
story (62–13% for number of stems). Notably, sugar maple has become even more dominant on this
site over the last 20 years in the understory (now 66–14% for number of stems). Experimental evi-
dence demonstrates that growth of sugar maple seedlings under low light is enhanced 2 to 4 times
on rich sites in part because of higher nitrogen and water availability (Walters and Reich 1997).
FIRE AND OAK DOMINANCE—O AKS AT RISK

Oaks dominate the overstory of dry and dry-mesic sites in much of eastern North America today,
but the current generation may be their last on all but the most extremely dry sites. Succession of
oak forests to more mesophytic and understory tolerant species has been documented for many
different regions and ecosystems (Christensen 1977; Lorimer 1984; McCune and Cottam 1985;
Host et al. 1987; Pallardy et al. 1988; Hammitt and Barnes 1989; Abrams and Downs 1990; Nowacki
et al. 1990; Nowacki and Abrams 1991; Abrams 1992; Shotola et al. 1992; Abrams 2003; Aldrich
et al. 2005; Nowacki and Abrams 2008; Johnson et al. 2019). Although oaks dominate in the over-
story, seedlings are relatively few and there is virtually no recruitment to the sapling understory.
Such a trend, if continued, would likely create major differences in ecosystem structure and
function, including shifts in both animal and plant communities.
Upland oaks evolved in dry, fire-prone ecosystems and are characterized by a suite of fire and
dry-site adaptations (Chapter 10; Abrams 1990). Depending on regional and local site conditions,
fire seasonality, frequency, intensity, and severity are instrumental in maintaining oak regenera-
tion and determining the distribution and occurrence of oak species. Widespread fire exclusion
throughout much of the East is perceived as the major reason for lack of oak regeneration, as is
herbivory by deer in many areas (Chapter 12). Red maple and sugar maple are the understory tol-
erant, mesophytic species replacing oaks in the understory on all but the driest sites. Opportunistic
species such as black cherry, sweet birch, blackgum, and sassafras can also establish and develop
in the absence of fire where oak advance regeneration fails and logging or wind disturbance to the
oak overstory provides sufficient light.
Successful oak regeneration depends on an abundance of oak saplings in the forest under-
story before disturbance, that is, the presence of advanced regeneration. There are many factors
that drive oak regeneration, such as seed production and dispersal or animal damage to seeds and
seedling (Lorimer 1993). However, fire is critically important for successful oak regeneration
because it kills competitors and promotes resprouting of young oaks. In addition, fire provides a
source of readily available nutrients for developing oak sprouts when it consumes organic matter.
As mid-tolerant species, young upland oaks are at a competitive disadvantage from many early-
successional species in open understories as well as from the species that tolerate shade. Thus,
mesophytic species are always invading the understories of oak forests by wind and animal dis-
persal, but oak dominance is maintained by fire.
Light and moderate surface fires historically occurred at all stages of stand development in
oak forests, indiscriminately killing tree seedlings, saplings, and shrubs. Few species have the
resprouting ability exhibited by oaks (and their usual associates, the hickories) in the seedling or
sapling stage, such that repeated fires act to purge the understory of mesophytic competitors in
oak–hickory forests. Young oak stems killed by fire resprout vigorously from well-established root
systems, although the thick bark of many young saplings allows them to survive surface fires. The
density of the oak overstory may be high enough to deter or significantly slow the invasion of
mesophytic species into the understory during stand initiation and stem exclusion. Fire is necessary
to kill invaders and reduce shrub competition from the understory reinitiation stage onward, when
oaks are themselves becoming established and forming advanced regeneration. When crown
breakage or gaps form during the old-growth stage, the advanced regeneration of oaks and hick-
ories is well positioned to dominate the understory (as in Figure 13.13, Chapter 13). On the driest
sites of ridges and fire-prone barrens, oaks (blackjack, post, chestnut, black, northern pin) may be
self-replacing or only very slowly replaced by other species (Abrams 1992). In these ecosystems,
fire, wind, soil movement, or other disturbances, together with other species adaptations to
extreme sites, may maintain the intolerant oak species and associates of hickory or pine.
Understanding a complex process such as forest succession is a tall order. Almost all ecolog-
ical research (at least on secondary succession) has been conducted within the life span of existing
forests, and therefore successional pathways must always be inferred or modeled rather than
observed. It is notable that macroclimate has been considered to be the grand driver of succession
since the concept was first articulated, and thus nearly everything we think we know about how
forests are likely to change over time is likely to be altered in the current era of rapid climate
change. From this perspective, despite the fact that forest succession has been studied in detail for
well over a century, ongoing and future research on successional processes should remain at the
forefront of forest ecology. We examine climate change and its effects on forests in the final part of
this book.
SUGGESTED
READINGS
Abrams, M.D. (1992). Fire and the development of oak Crisafulli, C. and Dale, V. (ed.) (2018). Ecological
forests. BioScience 42: 346–353. Responses at Mount St. Helens: Revisited 35 Years
Buma, B., Bisbing, S., Krapek, J., and Wright, G. (2017). After the 1980 Eruption. New York: Springer 351 pp.
A foundation of ecology rediscovered: 100 years of Glenn-Lewin, D.C. and van der Maarel, E. (1992). Pat-
succession on the William S. Cooper plots in Glacier terns and processes of vegetation dynamics. In: Plant
Bay, Alaska. Ecology 98: 1513–1523. Succession: Theory and Prediction (ed. D.C. Glenn-
Chapin, F.S. III, Walker, L.R., Fastie, C.L., and Sharman, Lewin, R.K. Peet and T.T. Veblen). London: Chap-
L.C. (1994). Mechanisms of primary succession fol- man and Hall.
lowing deglaciation at Glacier Bay, Alaska. Ecol. Matthews, J.A. (1992). Chapter 6. Plant succession:
Monogr. 64: 149–175. processes and models. In: The Ecology of Recently-
Christensen, N.L. (1988). Succession and natural dis- Deglaciated Terrain, 250–316. Cambridge: Cam-
turbance: paradigms, problems, and preservation of bridge University Press.
natural ecosystems. In: Ecosystem Management for Pickett, S.T.A., Collins, S.L., and Armesto, J.J. (1987).
Parks and Wilderness (ed. J.K. Agee and D.R. John- Models, mechanisms and pathways of succession.
son). Seattle: University of Washington Press. Bot. Rev. 53: 335–371.
Carbon Balance of Trees
CHAPTER 18
and Ecosystems
A lthough forest ecosystems cover only 21% of the Earth’s land surface, they constitute a
disproportionately large share of terrestrial plant mass (75%) and its annual growth (37%).
Because plants are largely constructed of carbon (47%), the extent to which forests are altered
by human activities (i.e., harvesting or conversion to other types of vegetation) has a substantial
influence on the pattern in which carbon (C) is cycled and stored at local, regional, and global
scales. Recent attention has focused on factors influencing the fixation and release of C by trees
and forest ecosystems (i.e., their C balance), because the rising CO2 concentration of the Earth’s
atmosphere and projected changes in global climate have the potential to alter the present-day
geographic distribution of forests and the rate at which they sequester CO2 from the atmosphere
(Pastor and Post 1986; Melillo et al. 1993). Such a change is of both ecological and economic impor-
tance, because the production of plant matter in forest ecosystems not only influences the global
cycling of C but also represents an indispensable source of food, fuel, fodder, and fiber for human
populations around the Earth. Globally, forests also are home to countless species of animals, as
well as vascular and non-vascular plants. Importantly, aspects of climate change, such as rising
atmospheric CO2, altered temperature and precipitation regimes, and the deposition of human-
derived nitrogen (N), have the potential to alter the C balance of forest trees, forest ecosystems,
and also that of the entire Earth.
The C balance of plants (i.e., growth) is controlled by the quantity of CO2 fixed through
photosynthesis and the rate at which photosynthetically fixed C is returned to the atmosphere
by respiring plant tissues. Light, temperature, and the availability of soil water and nutrients
constrain these physiological processes and consequently influence the productivity of plants
and the ecosystems in which they occur. One needs to only examine desert, grassland, and
forest ecosystems to understand that the processes controlling the productivity of individual
plants also have a substantial influence on the productivity of terrestrial ecosystems. This
situation is particularly true in forests, in which the growth and metabolism of plants far out-
weigh that of any other organism.
From an ecological perspective, the flows of C and energy are somewhat interchangeable;
that is, plants transduce solar energy into C-based biochemical energy via photosynthesis. Further,
it is plant-derived biochemical energy that drives the subsequent flow of energy through entire
terrestrial ecosystems. In this chapter, we focus on the physiological processes of plants that con-
trol the C balance (i.e., the flow of energy) of terrestrial ecosystems, especially forests. Importantly,
the principles we present here apply equally to the C dynamics of all terrestrial ecosystems,
whether they are tundra, desert, or grassland. We discuss how climate and soil nutrient availability
influence the C balance of plants, and hence their growth. We then build upon this information to
understand how environmental factors influence the C balance of ecosystems at local, regional,
and global scales.
449
450 Chapter 18 Carbon Balance of Trees and Ecosystems
CARBON BALANCE OF TREES

Only a small fraction of the radiant energy reaching the Earth’s surface (2%) is used by green plants
to assimilate atmospheric CO2 into organic compounds via photosynthesis. These compounds are
used to construct new plant tissue, maintain existing tissue, create storage reserves, or provide
defense against insects and pathogens. Simply stated: trees grow (or gain C) when the amount of
CO2 fixed through photosynthesis exceeds the amount of CO2 lost from respiring leaves, branches,
stems, and roots. Understanding the factors influencing the balance between photosynthesis and
respiration is of ecological importance, because plants with the greatest net C gain under a specific
set of environmental conditions (i.e., light, water, and nutrient availability) are often the best com-
petitors. It follows that fast-growing plants require rapid rates of photosynthesis, but the converse
is not always true. As you will learn, the net C gain of plants is not influenced by photosynthetic
rate alone (Ceulemans and Saugier 1991). In the following section, we trace the flow of photosyn-
thetically fixed CO2 to the construction and maintenance of plant tissue, focusing on how environ-
mental factors influence the net C gain of plants (i.e., growth).
PHOTOSYNTHESIS, DARK RESPIRATION, AND LEAF C GAIN

The processes of photosynthesis and respiration occur in the leaves of all plants, and the balance
of these more or less opposing physiological processes controls the net C gain of leaves. Gross pho-
tosynthesis is the total amount of C plants assimilate from the atmosphere. However, a portion of
that fixed C is returned to the atmosphere from leaves as CO2 during dark respiration. This pro-
cess is so termed because leaf respiration can only be determined in the absence of C fixation and
is hence measured on unilluminated leaves. In leaves, the process of dark respiration provides the
biochemical energy (i.e., ATP) to repair proteins and membranes mediating photosynthesis. Net
photosynthesis is the balance between gross photosynthesis and leaf dark respiration, and it rep-
resents the amount of photosynthate available for the growth and maintenance of non-
photosynthetic tissue, as well as that allocated to storage and defense. Light intensity, temperature,
and the availability of water and soil nutrients (especially N) all influence photosynthesis, respira-
tion, and the C gain of leaves.
Rates of net photosynthesis measured under saturating light, optimum air temperature, low
vapor pressure deficit, and ambient atmospheric CO2 represent the maximum photosynthetic
capacity of plants (see Table 18.1). Care should be taken in interpreting Table 18.1, because
maximum photosynthetic capacity varies with leaf canopy position (shade versus sun leaves), as
trees develop from a juvenile to an adult phase (see ponderosa pine), soil nutrient availability, and
measurement technique. Also be aware that maximum photosynthetic capacities are often poorly
correlated with the C gain of forest trees (Ledig and Perry 1969), because environmental factors
(light, temperature, water, and soil nutrients) constrain photosynthesis under field conditions,
often well below its maximum. Nevertheless, Table 18.1 illustrates several important physiological
differences among forest trees that have important ecological implications.
The maximum photosynthetic capacity of forest trees ranges from 2 to 25 μmol m−2 s−1
(Table 18.1), generally lower than that of most agronomic crops (Ceulemans and Saugier 1991).
Shade-intolerant deciduous species have some of the highest photosynthetic capacities of all
trees (up to 25 μmol m−2 s−1; Table 18.1), whereas those of shade-tolerant deciduous species are
much lower (3–6 μmol m−2 s−1; Table 18.1). Recognize that shade tolerance is only one of several
characteristics enabling a species to persist in the understory (i.e., understory tolerance;
Chapter 13). Although some conifers have high photosynthetic capacities (see Scots and
Monterey pine), most are modest when compared to those of deciduous trees.
Canopy architectures efficient at intercepting solar radiation can outweigh moderate
photosynthetic rates, allowing some trees to attain large C gains despite the relatively low net
Carbon Balance of Trees 451
Table 18.1 The photosynthetic capacity of trees under conditions of saturating light, optimum temperature
and water regimes, and ambient atmospheric CO2 (350 μmol mol−1). Rates are expressed in micromoles of
CO2 assimilated by 1 square meter of leaf during 1 second.
Shade-understory tolerance Maximum photosynthetic capacity μmol CO2 m−2 s−1

Tolerant
Deciduous trees
sugar maple 3–4
American beech 4–5
hop-hornbeam 6
American basswood 3
flowering dogwood 4–6
Coniferous trees
European silver fir 2–4
white fir 8
grand fir 4–5
balsam 2
fir seedling
Sitka spruce 3–9
Norway spruce 2–7
subalpine fir 9
Mid-tolerant
Deciduous trees
white oak 2
northern red oak 4–5
red maple 3–5
white ash 4–5
yellow birch 9
Coniferous trees
Engelmann spruce 3–5
Douglas-fir 2–6
Intolerant
Deciduous trees
black oak 7
sassafras 6
black cherry 4
sweet gum 6–11
tulip tree 7–17
paper birch 10
black locust 13–17
red ash 20–25
bigtooth aspen 14
cottonwood 15–19
(Continued )
Table 18.1 (Continued)
Shade-understory tolerance Maximum photosynthetic capacity μmol CO2 m−2 s−1

trembling aspen 20–22
black poplar hybrid 20–25
Coniferous trees
lodgepole pine 3–9
Ponderosa pine
seedlings 25
mature trees 5–14
old-growth 4
Scots pine 10–16
Monterey pine 17
loblolly pine 3–6
Source: All values are expressed on a one-sided leaf area basis and are summarized from Bazzaz (1979), Ceulemans and
Saugier (1991), Jurik et al. (1988), Larcher (1969), Wallace and Dunn (1980), and Walters et al. (1993).
photosynthetic rates of individual leaves. The rapid growth and modest net photosynthetic
capacities (2–10 μmol m−2 s−1) of some conifers (e.g., Douglas-fir) illustrate this point (Table 18.1).
Coniferous canopies disperse incoming solar radiation over a larger number of leaves than broad-
leaved canopies, enabling them to capture more solar energy at high light intensities. The
orientation of coniferous leaves and branches also produces less self-shading of leaves lower in the
canopy, enabling them to photosynthesize at a relatively high rate. Also note that the photosyn-
thetic capacities in Table 18.1 are expressed on an area basis (μmol CO2 per square meter of leaf
surface per second), a factor that can be compensated for by canopy architecture as described in
the earlier text. Additionally, coniferous species often maintain higher leaf areas than deciduous
species, creating a larger surface area to capture solar radiation. For example, Douglas-fir and
many true firs maintain leaf areas 2–6 times greater than that of some broad-leaved species. As
such, low net photosynthetic rates in some conifers are offset by an efficient light-gathering canopy
architecture, enabling them to achieve relatively large C gains, despite relatively low rates of
maximum net photosynthesis.
LIGHT AND LEAF C GAIN

Net photosynthesis varies substantially with light intensity, and the response of trembling aspen
(shade intolerant) and northern red oak (mid-tolerant) leaves to increasing light intensity illus-
trates this important point (Figure 18.1). Sun leaves of trembling aspen attain a greater light-saturated
rate of net photosynthesis (i.e., maximum photosynthetic capacity; see right-hand portion of
Figure 18.1) than do the sun leaves of red oak. The shade leaves of trembling aspen also have
greater light-saturated rates of net photosynthesis than do the shade leaves of northern red oak,
but rates in the shade leaves of both species are lower than those of sun leaves. Although shade
leaves photosynthesize at a lower rate than sun leaves, they can contribute as much as 40% of the
C assimilated by tree canopies (Schulze et al. 1977).
Plants attaining high rates of net photosynthesis, such as trembling aspen and other shade
intolerant species, typically contain large amounts of chlorophyll and the enzymes used for CO2
fixation (Field and Mooney 1986), which represent the biochemical machinery plants use to con-
vert light energy into biochemical energy. High respiration rates are associated with plant tissues
F I G U R E 18 .1 The photosyn- 30
thetic response to changing
light intensity for trees of
contrasting tolerance. Trem- 25
bling aspen is an intolerant
species, whereas northern red
Net photosynthesis (mg CO2/dm2•h)

oak is a mid-tolerant species Aspen sun leaf
that can persist in the forest 20
Light-saturated photosynthetic rates

understory. Source: After Loach
15
Oak sun leaf

10
Aspen shade leaf
5
Oak shade leaf
0
0.2 0.4 0.6 0.8 1.0
Compensation Relative light intensity
points
F I G U R E 1 8 .2 The relationship between y = 0.0913x + 0.455

photosynthetic capacity and leaf dark 4
Dark respiration (μmol•m–2•s–1)
r = 0.750
respiration for deciduous broad-leaved
(closed circles), evergreen broad-leaved
(half-open circles), and coniferous (open 3
circles) tree species. Source: After Ceule-
mans and Saugier (1991) / John Wiley &
2
Sons. Reprinted from Physiology of Trees, ©
1991 by John Wiley & Sons, Inc. Reprinted
with permission of John Wiley & Sons, Inc. 1
0
0 5 10 15 20 25 30
Photosynthetic capacity (μmol•m–2•s–1)
containing large amounts of metabolically active enzymes (Amthor 1984); such high rates of res-
piration are needed to replace and repair enzymes and membranes involved in the biochemical
processes mediating photosynthesis. Consequently, leaves with high net photosynthetic rates also
respire rapidly. Leaf dark respiration represents a relatively constant proportion (7–10%) of the
maximum photosynthetic capacity of many forest trees (Figure 18.2; Ceulemans and Saugier 1991).
However, rates of net photosynthesis rarely attain their maximum under field conditions, because
light, soil nutrients, or water are often in short supply. Under these conditions, dark respiration can
account for 10–20% of annual net photosynthesis in a wide range of coniferous and deciduous tree
canopies (Ryan et al. 1994).
Because rates of net photosynthesis vary with light intensity, whereas dark respiration
remains constant across a range of light intensities, the ability of a species to persist in low light
environments is related to the balance of these physiological processes. In Figure 18.1, the light
intensities at which net photosynthesis is zero (i.e., light compensation point) for the sun and
shade leaves of trembling aspen are substantially higher than those of northern red oak. This rela-
tionship results from high dark respiration rates that maintain rapid rates of photosynthesis in
aspen leaves. Dark respiration is lower in shade tolerant species, thereby enabling them to main-
tain positive rates of net photosynthesis (i.e., positive leaf C gain) at much lower light intensities.
This ability illustrates the point that high photosynthetic rates alone do not insure a high C gain in
all light environments.
TEMPERATURE AND LEAF C GAIN

Air temperature directly controls leaf C gain by influencing both rates of gross photosynthesis and
dark respiration (Figure 18.3). The difference between these two physiological processes (i.e., net
photosynthesis) reaches a maximum between 15 and 25 °C for most temperate trees (Kozlowski
and Keller 1966; Mooney 1972). Rates above and below this range rapidly decline because low and
high temperatures limit the process of CO2 fixation. In Figure 18.3 we see that gross photosynthesis
increases with rising temperature to an optimum and then rapidly declines with a further increase
in temperature. Dark respiration also increases with rising temperature, attains a maximum, and
then declines markedly as temperature continues to rise. However, the response of dark respira-
tion to temperature differs from that of gross photosynthesis in several important ways (Figure 18.3).
First, the maximum rate of dark respiration occurs at a much higher temperature than does the
maximum rate of gross photosynthesis. Also, the slope of the gross photosynthesis and respiration
curves differs at high temperatures (i.e., 35–50 °C). In this range of temperatures, gross photosyn-
thesis declines rapidly, whereas dark respiration continues to increase. Note also that leaf C gain
approaches zero as temperature rises between 35 and 43 °C. At temperatures above 43 °C, rapid
rates of dark respiration combined with slow rates of gross photosynthesis result in a negative C
Optimum temperature range

for net photosynthesis
Net CO2 balance = 0

Gross
photosynthesis Respiration
Net loss in
Net gain in carbohydrate
carbohydrate production
production
0 10 20 30 40 50 60
Temperature, °C
F I G U R E 1 8 .3 The temperature response of gross photosynthesis, dark respiration, and leaf C gain. Gross
photosynthesis and dark respiration in leaves are temperature-dependent processes, and leaf C gain is
maximized at the temperature where the difference between these processes is greatest. Source: After Daniel
et al. (1979) / McGraw-Hill. Reprinted from Daniel, T., J. A. Helms, and F. S. Baker, Principles of Silviculture,
2nd ed., © 1979 by McGraw-Hill, Inc. Reprinted with permission of The McGraw-Hill Companies.
balance and the net loss of C from the leaf. Thus, you can see that changes in rates of gross
photosynthesis and dark respiration in response to temperature have a great impact on leaf C gain
(Berry and Bjorkman 1980; Berry and Downton 1982), controlling whether it is positive (net C
gain) or negative (net C loss).
WATER AND LEAF C GAIN

Water is important for plant growth and its availability directly influences photosynthesis and leaf
C gain. Plants are faced with a fundamental trade-off because CO2 fixation and water loss occur
simultaneously in leaves. Trees open their stomata allowing CO2 to diffuse from the atmosphere to
enzymatic sites of CO2 fixation in the leaf mesophyll. At the same time, water vapor diffuses out of
stomata in response to differences in water potential between the leaf and atmosphere (see
Chapter 9). Declines in leaf water content from 5 to 10% often have little influence on photosynthe-
sis (Hanson and Hitz 1982), but a drop in leaf water potential beyond this level can sharply reduce
CO2 fixation. Low leaf water potentials cause stomata to close, restricting the loss of water to the
atmosphere and the flow of CO2 into the leaf. At leaf water potentials of −1.0 to −2.5 MPa
(Figure 18.4), stomatal closure begins to limit the diffusion of CO2 into the leaf, thereby slowing
rates of photosynthesis. As leaf water potentials fall below a critical level, stomata completely close
and photosynthesis ceases.
Although the species in Figure 18.4 inhabit a wide range of sites (dry-mesic uplands for
white pine and northern red oak; wet lowlands for speckled alder, red ash), leaf water potential
limits photosynthesis over a surprisingly narrow range of values. Plants occurring in dry sites have
several physiological and morphological adaptations enabling them to maintain high leaf water
potentials at relatively low leaf water contents (percent water). Physiological changes in cell wall
elasticity, membrane permeability, and the concentration of solutes within plant cells can main-
tain high leaf water potentials (i.e., less negative), allowing photosynthesis to occur at relatively
low leaf water contents (Bradford and Hsiao 1982). Trees inhabiting dry environments also have a
higher rate of gas exchange, lose turgor at a lower leaf water potential, and have greater root:shoot
ratios, leaf thicknesses, guard cell lengths, and stomatal densities than trees from mesic sites
(Abrams et al. 1994). Such adaptations provide dry-site species with increased water-use efficiency
(μmol CO2 fixed/mmol of H2O transpired), allowing them to fix more CO2 per unit of water lost.
F I G U R E 1 8 .4 The relationship 12
between leaf water potential
and net photosynthesis for Alnus rugosa
10
deciduous and coniferous trees
Net carbon uptake by leaves
occurring in different habitats.

(mg CO2 dm–2 h–1)
Source: After Hinckley et al. 8

(1981) / with permission of
Fraxinus pennsylvanica
Elsevier. Reprinted from Water 6
Deficits and Plant Growth, ©
Quercus rubra
1981 by Academic Press, Inc.
4
Reprinted with permission of Pinus rigida
Academic Press, Inc.
2 Pinus strobus
0
0 –1.0 –2.0 –3.0
Leaf water potential, MPa
SOIL NITROGEN AVAILABILITY AND LEAF C GAIN

Photosynthesis as well as other biochemical processes in the leaf are influenced by the supply of
soil nutrients. Nitrogen is a constituent of chlorophyll, other light-harvesting pigments, photosyn-
thetic membranes, and the enzymes mediating photosynthesis. Their presence makes leaves rich
in N compared to most other plant tissues. Thus, the supply of N can have a large influence on the
ability of a leaf to fix CO2 from the atmosphere. For example, leaf N concentration is directly
related to the photosynthetic rate of a wide array of herbaceous plants (Field and Mooney 1986).
Net photosynthesis in forest trees also increases with leaf N concentration, but coniferous and
deciduous species respond much differently (Brix 1971; Linder et al. 1981). In deciduous species,
net photosynthesis is particularly responsive to leaf N concentration (Linder et al. 1981). For
example, the highest rates of net photosynthesis for European white birch occur at the greatest leaf
N concentration (5%), and rates sharply drop as leaf N concentrations decline (Figure 18.5). Conif-
erous species, on the other hand, lack the ability to increase photosynthetic enzyme contents to the
same extent as deciduous trees, making photosynthesis much less responsive to changes in leaf N
concentration, which often reflects soil N availability. Photosynthetic rates in conifers, such as
Douglas-fir, initially increase up to 25% following N fertilization, but rates rapidly decline with
time (Brix 1971). However, conifers often increase whole-tree C gain following N fertilization by
producing more foliage, thus creating a larger photosynthetic surface area to capture CO2 (Brix and
Ebell 1969; Brix 1971). In deciduous trees, N fertilization often increases both the photosynthetic
rate of individual leaves as well as leaf production, resulting in a larger and more photosyntheti-
cally active canopy.
CONSTRUCTION AND MAINTENANCE RESPIRATION

Forest trees are unique compared to other types of terrestrial vegetation due to their size and lon-
gevity. As a consequence of their large mass, forest trees require substantial amounts of photosyn-
thate (i.e., carbohydrates produced by photosynthesis) to maintain the function of living cells
located in the non-photosynthetic tissues of branches, stems, roots, and reproductive structures.
Photosynthate also is required for the construction of new leaves to capture light, roots to forage
for water and nutrients, and reproductive structures to insure regeneration. The construction and
maintenance of these tissues require the products of photosynthesis to generate cellular energy
(i.e., ATP). Respiration during the biosynthesis of new tissue is construction respiration,
whereas respiration used to biochemically sustain already existing living tissue is maintenance
F I G U R E 18. 5 Leaf N
concentration is an important
30 factor controlling the rate of net
photosynthesis in deciduous
species. The highest rates of net
Net photosynthesis
(mg CO2 /g leaf /h)
photosynthesis for European

20
white birch occur at saturating
light intensities and high leaf N
concentrations. Source: Linder
10 5 et al. (1981) /U.S. Department of
4 Energy / Public Domain.
3
2 %
1 Nitrogen
0
0 in leaf
0 110 220 230 440
Photon flux density, μE/m2/s
respiration. These physiological processes, which return fixed CO2 to the atmosphere, have an
important influence on the C balance of all plants.
Construction Respiration: Respiration provides the cellular energy (ATP) used to
convert the products of photosynthesis into the biochemical constituents of new plant tissue.
Thus, photosynthetically fixed CO2 is lost to the atmosphere (as CO2) during the synthesis of
new plant tissue. This loss (i.e., construction respiration) is calculated knowing the biochemical
constituents of a particular plant tissue and the biosynthetic pathways by which they are
formed. This information can be used to determine the amount of CO2 produced during the
biosynthesis of a particular plant compound or organ. Glucose, a product of photosynthesis, is
often considered the substrate for biosynthesis when calculating the cost of construction respi-
ration during the biosynthesis of a plant compound or a tissue (Penning de Vries 1975; Chung
and Barnes 1977).
Look at Table 18.2 and note that the needles and shoots of loblolly pine contain large propor-
tions (%) of carbohydrates, lignin, and phenolic compounds, whereas nitrogenous compounds,
lipids, and organic acids are present in smaller amounts. The biosynthetic pathways giving rise to
these compounds require different amounts of substrate (i.e., glucose) and produce different
amounts of CO2 as they are biochemically synthesized. For example, the synthesis of 1 g of carbo-
hydrate during cell wall construction requires 1.2 g of glucose as substrate, of which 11% is lost as
CO2 (Table 18.2). Although lipids are present in relatively small amounts, their construction
produces relatively large amounts of CO2 compared to other biosynthetic pathways.
By knowing the constituents of a particular plant tissue and the amount of CO2 lost during
their biosynthesis, one can calculate construction respiration for plant tissues. For example, during
the construction of 1 g of loblolly pine needles, approximately 19% of substrate glucose is respired
during biosynthesis. Similarly, construction respiration consumes approximately 16, 18, and 17%
of the glucose used to synthesize loblolly pine shoots, cambium, and roots, respectively (Chung
and Barnes 1977). Clearly, accounting for these losses of C is essential for understanding the
C balance of whole plants, as well as the C balance of forest ecosystems. It is important to realize
that construction respiration is a fixed cost to the plant. Regardless of temperature or other envi-
ronmental conditions that a plant faces, the amount of photosynthate used to synthesize a protein,
carbohydrate, or for that matter a leaf, is invariant. The process of construction respiration may
occur at a faster or slower pace depending on temperature, but the amount of photosynthate used
to construct a compound or tissue remains the same. This differs dramatically from maintenance
respiration, which is highly response to environmental conditions.
Table 18.2 The biochemical constituents of the needles and shoots of loblolly pine, and the proportion
of carbon respired during their biosynthesis.
Needles Shoots Construction respirationa

Compound ----------------Percent -------------------------------------------------
Carbohydrates 35.7 38.0 11
Lignin 23.1 23.3 14
Phenolic compounds 21.0 20.0 29
Nitrogenous compounds 9.3 8.4 25
Lipids 5.6 5.3 49
Organic acids 3.8 3.5 32
a
Calculated as the percent of glucose lost as CO2 during the synthesis of a compound in the left-hand column.
Source: After Chung and Barnes (1977) / Canadian Science Publishing. Reprinted with permission of National Research
Council Canada.
Maintenance Respiration: Carbohydrates produced by photosynthesis are used to sustain

the function of living plant tissues during the process of maintenance respiration. Protein syn-
thesis and replacement, membrane repair, and the maintenance of ion gradients across cell mem-
branes all require the expenditure of energy initially derived through photosynthesis (Ryan 1991).
The rate of maintenance respiration, like all other enzymatic processes, increases predictably with
temperature. For most plants, maintenance respiration roughly doubles for every 10 °C
increase in ambient temperature. That is, maintenance respiration increases exponentially with a
linear increase in temperature. This fundamental fact has important implications for the C balance
of plants, as well as entire ecosystems, as the Earth’s climate continues to warm from fossil
fuel burning.
Maintenance respiration also is strongly influenced by tissue N concentration (Waring
et al. 1985), wherein tissues with high N concentrations also have high rates of maintenance respi-
ration. In fact, leaf maintenance respiration for a wide range of plant species can be predicted from
N concentration using the following equation (Ryan 1991):
RM 0.0106 N
where RM is leaf maintenance respiration (mmol of CO2 mol tissue C−1 h−1) and N is the total
nitrogen concentration of that tissue (mmol N mol tissue C−1). This relationship arises because
the majority of N present in plant tissue occurs in proteins, and a large proportion (60%) of
maintenance respiration is used for their repair and replacement (Penning de Vries 1975). Such
a relationship also contributes to the high rates of leaf dark respiration that accompany rapid
photosynthetic rates (see Figure 18.2). Herein lies an important link between soil N availability,
leaf N concentrations, and the physiological processes that control plant C gain. Greater soil N
availability can increase the N concentration of plant tissues. This can lead to greater rates of
photosynthesis in leaves and more rapid rates of maintenance respiration in non-photosynthetic
tissue. However, increases in maintenance respiration of non-photosynthetic tissues (i.e., plant
roots) are surpassed by increases in net photosynthesis in leaves with higher N concentrations.
In combination, these responses often lead to higher plant C gains when soil N availability
increases.
As trees develop from juvenile to adult phases, the ratio of photosynthetic to non
photosynthetic tissues decreases, a factor that substantially influences maintenance respiration
and the C balance of whole trees. Canopy leaf area reaches a maximum relatively early in tree
development, thereby imposing a limit on canopy photosynthesis and the supply of photosynthate
for tissue construction and maintenance. Although the supply of carbohydrate is constrained by
canopy photosynthesis, the demand for carbohydrate rises as the proportion of non-photosynthetic
tissue continues to increase during tree ontogeny. For example, stemwood accounts for a large
proportion of total tree mass (approximately 50–70% in mature trees) and contains both living
(sapwood) and dead tissue (heartwood). Although the proportion of living, non-photosynthetic
tissue in stems decreases as trees mature, the absolute mass of living non-photosynthetic tissue
continues to increase. Consequently, the amount of photosynthate allocated to the maintenance of
living cells in stemwood dramatically increases as trees develop from a juvenile to an adult (Agren
et al. 1980; Waring and Schlesinger 1985; Ryan 1988). Remember that the product of photosyn-
thate used to maintain the mass of live, non-photosynthetic tissue in mature trees cannot be used
for the construction of new tissue. Consequently, the greater cost of tissue maintenance in mature
trees results in a decline in annual growth. Annual growth also can be slowed by reductions in the
net photosynthetic rate of adult trees. As you will read later in this chapter, increases in mainte-
nance respiration and declines in growth as trees mature have important implications for the C
balance of forest ecosystems.
ALLOCATION TO STRUCTURE, STORAGE, AND DEFENSE

In the previous paragraphs, we have discussed physiological processes influencing the fixation and
loss of C by forest trees. Accounting for these processes, one can construct a C budget illustrating
how plants partition C into the growth and maintenance of leaves, stem, roots, reproductive struc-
tures, as well as the production of defense and storage compounds. Such an approach can shed
light onto how light, water, and nutrients influence the allocation of photosynthate into organs
that aid in gathering light or the acquisition of water and soil nutrients.
Summarized in Table 18.3 is the annual C budget for a young Scots pine tree (Agren
et al. 1980). This budget quantifies the partitioning of photosynthetically fixed CO2 into the growth
of various plant tissues and accounts for their respiration during construction and maintenance.
Approximately 88% of fixed C in this young tree was used to produce new tissue, whereas only 10%
was consumed during construction and maintenance respiration. The production of new roots to
forage for water and nutrients in soil represented a large investment of photosynthate, accounting
for 56% of annual net photosynthesis. In combination, root growth and respiration comprised
almost 62% of annual net photosynthesis, indicating a substantial proportion of fixed C was allo-
cated to belowground growth and metabolism to acquire soil resources.
The amount of photosynthate plants allocate to growth, storage, and the production of
defense compounds is under strong genetic control. Patterns of C allocation vary among species,
within a species seasonally as well as with increasing age. Environmental factors also play an
important role in influencing the allocation of C into structure, storage, and defense. Figure 18.6
illustrates the C allocation priorities in lodgepole pine. Allocation of photosynthate to new foliage
and buds represents the highest priority for carbohydrate, followed by the production of new roots.
Tree growth often is limited by light, water, and nutrients, making it necessary for the plant to first
meet the carbohydrate demand of tree crowns and roots. The storage of carbohydrate in leaves,
stem, and coarse roots has a lower priority than allocation to new leaves and roots. Allocation of
Table 18.3 The carbon budget of a 14-year-old Scots pine tree.
Assimilation Allocation Percent of total

g C year−1
Net photosynthesis 1723
Growth
Current needles 286 16.6
Branch axes 132 7.7
Stem 145 8.4
Roots 960 55.6
Total growth 1523 88.4
Construction and maintenance respiration
Stem 49 2.8
Branch axes 15 0.9
Roots 109 6.3
Total respiration 173 10.0
Growth + respiration 1696 98.5
Unaccounted net photosynthesis 27 1.5
Total 1723 100.0
Source: After Agren et al. (1980).

Pr
ote
cti
ve
Ch
e
mi
ca
ls
5
New Leaves 1 Buds 1
Canopy
Storage 3
Diameter Growth 4
Stem Storage 3
Storage 3
New Roots 1-2
F I G U R E 18.6 The allocation of photosynthate into structure, storage, and defense in lodgepole pine. Tree
components with low numbers (e.g., one) represent strong priorities for allocation, whereas components with
high numbers indicate a low priority for allocation. Source: Waring and Pitman (1985) / John Wiley & Sons.
Reprinted with permission of the Ecological Society of America.
carbohydrate to stem growth follows that to storage and the production of chemicals to defend
against herbivory, which is the lowest priority.
Patterns of C allocation vary substantially over the growing season, wherein certain tissues
function as sinks or sources for carbohydrate. Nowhere is this more evident than during the phe-
nological development of deciduous trees. Storage reserves in buds, branches, stems, and roots
function as a source of carbohydrate to build new leaves early in the growing season. Once the
expanding canopy attains a positive C balance, the priorities of allocation shift. In sugar maple
seedlings, for example, large amounts of photosynthate are allocated to the production of new
leaves early in the growing season (Table 18.4). The proportion of C allocated to stems and coarse
roots increases following canopy development (i.e., mid growing season), and represents the
greatest sink for photosynthate late in the growing season. The amount of carbohydrate stored in
stems, coarse roots, and fine roots also increases over the growing season, attaining the highest
priority late in the growing season. These carbohydrate stores are then used to construct new
leaves, as well as new fine roots, at the start of the next growing season.
Table 18.4 The seasonal C allocation to leaves, stem, and roots in sugar maple seedlings.
Percent of growth allocated to:

Leaves Stem Coarse rootsa Fine rootsb
Early growing season 85 5 0 10
Mid growing season 30 30 25 15
Late growing season 0 5 60 35
a
Coarse roots are 2–5 mm.
b
Fine roots are <2 mm.
Source: After Burke et al. (1991). / Canadian Science Publishing. Reprinted with permission of National Research Council Canada.
LIGHT AND C ALLOCATION

By knowing the C allocation pattern of trees, one can begin to understand how the availability
of light, water, and nutrients influences allocation to structure, storage, and defense. Carbohy-
drate produced during photosynthesis is first allocated to new leaves and then new fine roots
(<2 mm in diameter; Figure 18.6). These are the energy, water and nutrient gathering structures
of trees, and they have the highest priority for photosynthate. If the carbohydrate demand of
new leaves and fine roots is satisfied, then remaining carbohydrate can be allocated to storage,
the growth of stemwood, and the production of chemicals to defend against herbivory. Such a
pattern has been observed for both coniferous (Waring and Pitman 1985) and deciduous species
(Burke et al. 1991).
In very dense stands of lodgepole pine, low light levels limit photosynthesis and carbohy-
drate production, even when soil resources are abundant. In such a situation, carbohydrate allo-
cated to meet the demand of new leaves and roots occurs at the expense of allocation to lower
priorities of storage, stem growth, and the production of defense chemicals. As a consequence,
stem diameter growth declines and the susceptibility to mountain pine beetle attack increases;
trees with low diameter growth are those most susceptible to beetle attack (Waring and Pit-
man 1985). Thinning these forests increases light availability, decreases intraspecific competition
for light, and increases the photosynthesis of remaining trees. Therefore, carbohydrate limitation
within the remaining trees lessens, leading to increases in diameter growth and a reduction in sus-
ceptibility to mountain pine beetle attack. Nitrogen fertilization produces a similar response,
wherein greater soil N availability increases the production of new foliage, which, in turn, increases
canopy photosynthesis and the supply of carbohydrate within the tree. This is an interesting link
between environmental conditions, changes in the allocation of photosynthate, and the suscepti-
bility of a tree to an insect attack.
SOIL NITROGEN AVAILABILITY AND C ALLOCATION

The availability of soil nutrients is a potent modifier of C allocation in forest trees, substan-
tially influencing the proportion of carbohydrate allocated to leaves, stems, and roots. Fine
roots (< 1 mm in diameter) are the water- and nutrient-absorbing organs of forest trees, and
their biomass usually declines as water or nutrient availability increases. By contrast, the
amount of foliage often increases when soil water and nutrients are in abundant supply, sug-
gesting that a trade-off occurs between aboveground and belowground plant growth as soil
resource availability increases. That is, the root to shoot ratio (root:shoot) of plants declines
along a gradient of increasing soil nutrient availability. In sugar maple, for example, N fertil-
ization increased the mass of leaves and stem by 12%, whereas root mass (coarse + fine)
declined by an equivalent proportion (calculated from Burke et al. 1991). Scots pine and a
black poplar hybrid respond in a similar manner to N fertilization, but these species exhibit an
even more dramatic shift from roots to leaves (Linder and Axelsson 1982; Pregitzer et al. 1995).
Herbaceous species also apparently decrease root and increase leaf mass along gradients of
increasing soil N availability (Tilman 1988). Such a pattern suggests that the plants “invest”
smaller amount of carbohydrate into roots when soil resources are relatively abundant. As a
result, carbohydrate is available to produce relatively more foliage, a response that allows the
plant to capture more light energy on N-rich soils.
Nevertheless, there is considerable debate regarding the extent to which soil N availability
alters the total allocation of carbohydrate to the production of belowground plant tissues. This
debate stems from our limited understanding of fine-root production (i.e., birth) and mortality
(i.e., death), factors that greatly influence the total amount of carbohydrate that plants allocate to
belowground growth. Unlike the leaves of deciduous trees, which are initiated in spring and
abscise during autumn, the birth and death of fine roots occurs simultaneously throughout the
growing season (Hendrick and Pregitzer 1992). Consequently, relatively little can be inferred about
patterns of C allocation by comparing the standing crops (i.e., biomass) of leaves and roots across
gradients of soil nutrient availability.
Theoretically, the proportion (i.e., percent) of plant mass composed by fine roots could
decline in N-rich soil by several alternative responses: (i) fine-root production could remain
constant while mortality increases (Figure 18.7a), (ii) production could decline while mortality
remained constant (Figure 18.7b), or (iii) both production and mortality could increase with rates
(a) F I G U R E 18. 7 Three alternative mecha-

1.0 100
nisms by which allocation to fine roots
could decline with increasing levels of soil
Production N availability. Each alternative has very
different implications for the total amount
50 of carbohydrate allocated to fine-root
Mortality
production.
Root
B iomas
s
Root production or mortality (relative change)
(b)
1.0 100
Production
50
Mortality
Root
B iomas
s
(c)
1.0 100
Production
lity
Mor ta 50
Root Bio
m ass
Low High
Soil N availability
of mortality surpassing those of production (Figure 18.7c). The first mechanism suggests that the
life span of fine roots decreases, but root production (i.e., birth) remains constant; such a response
would not alter the total amount of carbohydrate allocated to fine roots. The second mechanism
argues for a decline in the total amount of carbohydrate allocated to fine roots, wherein fewer roots
are produced with the same life span (i.e., no change in mortality). In the third alternative, fine
root production increases, but greater N availability decreases the life span of fine roots, which
thereby increases the total amount of carbohydrate “invested” into these structures. In other
words, fine roots “live fast and die young” in the third alternative. Clearly, determining which of
these alternative responses is correct will rest on our understanding of the processes controlling
fine-root production and mortality.
Attaining such an understanding has been difficult, largely because of the numerous
problems associated with studying live plant roots in soil. Most methods of studying the produc-
tion and mortality of fine roots modify the soil in ways that may invalidate the observations.
However, technology has enabled ecologists to actually view and record the birth, growth, and
death of individual tree roots. Clear plastic tubes (mini-rhizotrons) can be placed in the soil of
a forest stand, and a miniaturized video camera can be used to capture images of individual
roots from their birth to death (Figures 18.8 and 18.9; Hendrick and Pregitzer 1992; Iversen
et al. 2012). The root images can be digitized, processed by computer, and production can be
Clear Tube Soil Surface
Video Camera
F I G U R E 1 8 .8 Diagram of a mini-rhizotron system used to capture video images of actively growing plant
roots. The clear plastic mini-rhizotron tube (A) is placed into the soil using an auger, and the micro-video
camera (B) is passed down the tube to record video images of the plant roots with the use of a video cassette
recorder (VCR). The controller allows the operator to focus the camera and adjust light levels in the
mini-rhizotron tube. Recorded images are subsequently digitized with the use of computer software.
Source: After Hendrick and Pregitzer (1992) / John Wiley & Sons. Reprinted with permission of the
Ecological Society of America.
2009 2010
28 May 2 July 5 Aug. 7 Sept. 9 Oct. 19 May 23 June 28 July 31 Aug. 6 Oct.
5 cm
10 cm
F I G U R E 1 8 .9 Images of the Scots pine fine-root growth over two growing seasons in surface soil (0–10 cm).
Notice that fine roots are produced throughout the growing season and experience a pulse of mortality from
autumn to spring (9 Oct 2009 to 19 May 2010). Source: Image obtained from Iversen et al. (2012).
calculated as the cumulative increase in root length over one growing season. Similarly, mortality
can be determined as disappearance of roots that had previously appeared in the video image
(i.e., a negative change in length). The results of such a study are shown in Figure 18.10, which
illustrates the cumulative change in root production and mortality for a black poplar hybrid (a fast-
growing intolerant tree) growing in soils of low and high N availability (Pregitzer et al. 1995).
The leaf mass of black poplar hybrid in the high N soil of this experiment increased by 35%,
and fine-root mass decreased by 25%, relative to plants in low N soil. The decline in fine roots and
increase in leaves are consistent with traditional views of changes in plant C allocation and
with increases in soil N availability (a decline in root:shoot ratio). In the high-N soil, however,
fine-root production increased, whereas fine-root life span decreased (i.e., increase in mortality;
Figure 18.10a,b), supporting the idea that more, not less, photosynthate was actually allocated to
fine-root growth in N-rich soil (Pregitzer et al. 1995).
The underlying mechanisms for an increase in fine-root mortality can be found in the rela-
tionships among soil N availability, the N concentration of fine roots, maintenance respiration, and
the longevity of plant tissue. Tissue N concentrations in the black poplar cultivar also increased in
response to greater soil N availability. Conceptually, fine roots growing in N-rich soil should have
relatively high tissue N concentrations (i.e., ion uptake proteins) and high rates of maintenance
respiration, compared to fine roots growing in N-poor soil (Ryan 1991). As a consequence, plants
growing on N-rich soil should allocate proportionately more carbohydrate to fine roots, which
have a greater maintenance “cost.” Although we know little about the influence of tissue N
concentration on the life span of fine roots, leaf life span is inversely related to photosynthetic rate,
Carbon Balance of Ecosystems 465
F I G U R E 18.10 The influence of soil N availability

800
on fine-root production and mortality of a black Low N
poplar hybrid. Note that both fine-root production
Cumulative fine root-production (mm)

High N
and mortality increase with greater levels of soil N
availability. Although production increases in the 600
high-N soil, the standing crop of fine roots is kept
small by rapid rates of fine-root mortality. Such a
pattern suggests that the total amount of carbohydrate
400
allocated to fine roots increases in N-rich soil. Source:
After Pregitzer et al. (1995) / John Wiley & Sons.
200
0
100
Cumulative fine-root mortality (mm) 75
50
25
0
11 7 21 4 18 31
14 27
June July August
tissue N concentration, and metabolic activity (Larcher 1969; Chabot and Hicks 1982; Reich
et al. 1991, 1992). If the same relationship holds for fine roots, then one would expect fine-root life
span to decline as soil N availability increases. Taken together, these results support the view that
the allocation of C to fine-root growth and maintenance increases in N-rich soil, but greater rates
of fine-root mortality (i.e., decreased life span) produce a proportionately smaller mass of fine roots.
We still have a great deal to learn about the C allocation patterns and fine-root dynamics of
forest trees. Certainly, the processes controlling fine-root production and life span are of great
importance for understanding the C allocation patterns of plants. Without a clear understanding
of plant C allocation, we cannot fully understand how changes in light and soil resources influence
the C balance of ecosystems. In the following section, we will build on the processes controlling
the C balance of plants to understand how climate, soil resources, and disturbance influence the C
balance of entire ecosystems.
CARBON BALANCE OF ECOSYSTEMS

The physiological processes by which plants photosynthetically fix CO2 from the atmosphere
and how plants allocate photosynthate to growth, storage and defense are important at a global
scale. For example, approximately 810 Pg of C presently reside in the Earth’s atmosphere, but
2000 Pg of C are stored in the vegetation and soil of Earth’s terrestrial ecosystems, the majority
(60%) of which occurs in forests. Because of this fact, the establishment and management of for-
ests have been looked to as a means to decrease the accumulation of anthropogenic CO2 in the
atmosphere. However, to do this effectively, we need to understand the physiological processes
by which CO2 fixed by forest ecosystems is stored within them and the processes that release this
stored C back to the atmosphere. To do this, we must consider the activities of organisms other
than trees to understand the flow of C into, within, and out of forest ecosystems. For example,
most soil microorganisms use dead plant tissue (i.e., leaves, branches, stems, and roots) as an
energy source. They incorporate a portion of the plant-derived C into microbial cells, release
microbially derived organic compounds that aid in humus formation (i.e., soil C storage), and
return the remainder to the atmosphere as CO2 during microbial respiration. Leaf feeding (phy-
tophagous) insects annually consume from 5 to 30% of the canopy leaves in most forests.
Although such a loss of photosynthetic surface area often has little influence on tree growth
(Mattson and Addy 1975), it represents a transfer of C from one ecosystem component to another.
That is, C once residing in plants has been transferred to insects, portions of which are lost dur-
ing their respiration and in excrement. In the sections that follow, we consider how climate, soil
nutrient availability, and disturbance influence the processes controlling the cycling and storage
of C in terrestrial ecosystems. Collectively, these processes are of global importance, due to the
large amount of C stored in the Earth’s forest ecosystems. We will later draw on this information
to understand the cycling and storage of plant nutrients within forests, a topic covered in
Chapter 19.
BIOMASS AND PRODUCTIVITY OF FOREST ECOSYSTEMS

Several concepts are important in understanding the C balance of ecosystems, whether they are
dominated by forest trees, prairie grasses, desert shrubs, or arctic tundra. Biomass is the dry
mass of living organisms and dead organic matter contained in a defined area, usually one
square meter or a hectare (g m−2 or kg ha−1, respectively). In some instances, biomass is reported
in its C or energy equivalent (i.e., g C m−2 or cal m−2), a necessary tool for understanding the flow
of C (or energy) within ecosystems. In forest ecosystems, biomass is located in 5 major pools: (i)
the aboveground and belowground tissues of overstory and understory plants, (ii) woody debris
consisting of dead, fallen tree stems, (iii) forest floor, (iv) mineral soil, and (v) the tissues of het-
erotrophic organisms (decomposers and consumers). Although biomass in woody debris, forest
floor, and mineral soil is composed of dead organic matter, it is convention to consider these
compartments when discussing the distribution of biomass within forest or other terrestrial
ecosystems.
Compare the amount and distribution of biomass among boreal, temperate, and tropical
forest ecosystems in Table 18.5. Notice that the total biomass of a tropical rain forest is almost five
times that of a black-spruce-dominated boreal forest; the temperate coniferous and deciduous for-
ests lay between these two extremes. Differences in the proportion of total biomass located in over-
story trees, forest floor, and mineral soil also occur among the forests in Table 18.5. For example,
forest floor and mineral soil compose the largest pool of biomass in the boreal forest, whereas the
majority of biomass in the tropical forest resides in overstory trees. The temperate coniferous and
deciduous forests in Table 18.5 also differ in biomass distribution, wherein overstory trees compose
a larger proportion of total biomass in the coniferous forest.
Caution should be used in interpreting Table 18.5, because there is substantial variation in
the amount and distribution of biomass within forests from boreal, temperate, and tropical regions.
For example, overstory biomass in tropical forests ranges from 10 Mg ha−1 in dry forests to
540 Mg ha−1 in moist forests (Brown and Lugo 1982), similar to the range of overstory biomass
found in temperate deciduous (2–578 Mg ha−1; Grier et al. 1989) and coniferous (5–810 Mg ha−1;
Grier et al. 1989) forests.
Table 18.5 The distribution of biomass in selected boreal, temperate, and tropical forest ecosystems.
Boreal Temperate Wet tropical

Alaska Washington New Hampshire Amazon
Location USA USA USA Brazil
Overstory
Dominant species black spruce Douglas-fir sugar maple–beech mixed speciesa
Age (years) 95 60 55 mature
Biomass pools (Mg ha−1)

Overstory 50 410 165 990
Woody debris – 9 29 18
Forest floor 76 15 48 7
Mineral soil 152 119 173 250
Heterotroph <1 <1 <4 <1
Total 278 553 419 1265
a
The overstory of this tropical rain forest is composed of 50 species, but the total number of plant species exceeds 600 species
per hectare. Species in the Leguminosae, Euphorbiaceae, and Sapotaceae families compose the majority of plant biomass.
Source: Data have been summarized from Bormann and Likens (1979), Harmon et al. (1990), Klinge et al. (1975), Schowalter
(1989), Van Cleve et al. (1983), and Zak et al. (1994).
Primary production is the annual growth of all living plants within an ecosystem, and it
can be subdivided into several basic components. Gross primary production (GPP) is the total
amount of C fixed in an ecosystem during the process of photosynthesis (Figure 18.11). In all ways,
GPP is the ecosystem equivalent of the gross photosynthesis of an individual plant; it is the summed
gross photosynthesis of all plants in an ecosystem. As with an individual plant, a portion of the
total C fixed by an ecosystem is lost to the atmosphere via leaf dark respiration and the construction
and maintenance respiration of non-photosynthetic plant tissue (Figure 18.11). The difference bet-
ween gross primary production and plant respiration is termed net primary production
(NPP = GPP – RA, where RA = total plant respiration consisting of leaf dark respiration, construction
respiration = RC, and maintenance respiration = RM). Gross and net primary productivity can vary
by an order of magnitude in forest and grassland ecosystems. However, the proportion of GPP lost
via plant respiration (RA) is surprisingly constant among the ecosystems in Table 18.6, varying
from 50 to 60%. The majority of that loss arises from the maintenance respiration (RM) of non
photosynthetic tissues (75–88% of RA), whereas construction respiration (RC) represents a much
smaller proportion of RA (12–24%). Much as the rate of net photosynthesis controls the C gain of
an individual plant, the relationship between GPP and RA directly influences NPP, and hence the
rate at which C accumulates within or is lost from terrestrial ecosystems.
The growth and metabolism of decomposing organisms in soil and that of animals foraging
within forest ecosystems are fueled by organic matter produced through NPP (Figure 18.11). The
proportion of NPP remaining after accounting for losses to herbivory and litterfall is the live bio-
mass accumulation of plants (Figure 18.11), and it can be either a positive or negative value. For
example, a forest that has suffered a devastating windstorm, an insect outbreak, or has been har-
vested can experience a negative live biomass accumulation.
Net secondary production is the annual increase in the biomass of all organisms obtaining
their energy from net primary production, including decomposers and consumers (i.e., heterotro-
phic organisms). Realize that not all organic matter assimilated by decomposers or consumers is
CO2
Gross Primary Production

(gross photosynthesis)
CO2
Leaf Dark Respiration
Net Photosynthesis
CO2
Construction &
Maintenance Respiration
Net Primary Production
Consumption
Excretion
Litter Production Assimilation

CO2 CO2
Heterotrophic Heterotrophic
Respiration Respiration
(decomposer) (consumer)
Net Secondary Production
Litter and Soil Organic Mortality

Matter Accumulation Predation
Live Biomass Accumulation Live Biomass Accumulation

of Plants of Consumers
Net Ecosystem Production
F I G U R E 1 8 .1 1 The processes controlling the C balance of terrestrial ecosystems. Gross primary production
is the process responsible for C entering terrestrial ecosystems, whereas the construction and maintenance
respiration of plants and animals is responsible for returning the C fixed through photosynthesis back to the
atmosphere as CO2. Also note that decomposing organisms in soil ultimately return C fixed by plants to the
atmosphere. Source: After Aber and Melillo (1991) / Saunders College Publishing. Figure from Terrestrial
Ecosystems by John D. Aber and Jerry M. Melillo, copyright © 1991 by Saunders College Publishing,
reprinted with permission of the publisher.
Table 18.6 Estimates of total plant (RA), construction (RC), and maintenance (RM) respiration in relationship
to gross primary production (GPP) and net primary productivity (NPP) in forest and grassland ecosystems.
Plant biomass GPP NPP RA RC RM RA/GPP Rm/RA

Ecosystem
(Mg C ha−1) Mg C ha y
−1 −1
Percent
Forest
Deciduous 88 21.7 7.3 14.4 1.8 12.6 66 88
Oak–pine 71 12.8 6.0 6.8 1.5 5.3 53 78
Grassland
Tallgrass 8 9.5 3.7 5.8 0.9 4.9 61 84
Shortgrass – 1.5 0.7 0.8 0.2 0.6 53 75
Recall that NPP = GPP–RA and RA = RC + RM; estimates of GPP do not include leaf dark respiration.
Source: After Ryan (1991) / John Wiley & Sons. Reprinted with permission of the Ecological Society of America.
used to build biomass; portions are lost during heterotrophic respiration (RH) and in excrement
(Figure 18.11). The live biomass of consumers accumulates only when the rate of net secondary
production is greater than the rate at which consumer biomass is lost via mortality and predation.
Litter production, both aboveground and belowground, and its decomposition directly con-
trol organic matter or biomass accumulation in the forest floor and mineral soil. Organic matter
accumulates in these ecosystems pools when its rate of production (leaves and roots) exceeds the
rate at which it is decomposed by soil microorganisms and returned to the atmosphere as CO2
(Figure 18.11). Clearly, changes in either the rate of litter production or decomposition have the poten-
tial to alter the accumulation of organic matter in soil, a topic we will consider later in this section.
The annual rate of biomass accumulation in live plants, live animals, and soil organic matter
is termed net ecosystem production (NEP), and it represents the summed change in all ecosys-
tem biomass pools (Figure 18.10), and it is defined in the following expression:
NEP GPP RA RH .
Net ecosystem production, or the C gain of ecosystems, can be either positive or negative
depending on the change in the biomass of live plants, live animals, and soil organic matter. Bio-
mass accumulates within terrestrial ecosystems only when the amount of C fixed during GPP
exceeds the rate at which C is lost via the respiration of plants (RA) and heterotrophic organisms
(RH). That is, NEP > 0 when GPP > (RA + RH).
In forests, the biomass and respiration of organisms that consume plant material and their
predators are relatively small compared to that of the overstory trees, forest floor, and mineral soil.
Moreover, the amount of NPP consumed by herbivores, primarily insects, in forests is relatively
small (3%), compared to grasslands, such as the Serengeti, in which large herds of grazing animals
(i.e., wildebeest, zebra, antelope) annually consume a significant proportion of aboveground NPP
(ANPP, 66%). Consequently, NEP in forests is largely controlled by the difference between NPP and
the rate at which CO2 is returned to the atmosphere during organic matter decomposition (RH).
Any forest disturbance that reduces or eliminates the live biomass accumulation of plants and
increases rates of soil organic matter decomposition could lead to a net loss of C from an ecosys-
tem. Under such a condition, NEP is negative and total ecosystem biomass declines. Ecosystems in
such a state are a source of CO2 to the atmosphere, rather than a sink, which has global implica-
tions for the amount of C stored by terrestrial ecosystems.
Understanding the C balance of forest ecosystems has clear management implications—any
management practice that tilts the balance toward negative NEP has the potential to alter the
amount of C sequestered by forest ecosystems. Nevertheless, it is important to realize that the C
balance of forest ecosystems is dynamic and can rapidly become positive following destruction of
plants by harvest, fire, disease, or insect outbreak.
MEASUREMENT OF BIOMASS AND PRODUCTIVITY

Ecologists have long been interested in understanding how and why the productivity of ecosys-
tems differs from one geographic location to another. Clearly, biomass and productivity differ
markedly among forest, agricultural, grassland, or desert ecosystems. Nonetheless, measuring the
biomass and productivity of an ecosystem is not an easy task, particularly in forests in which tre-
mendous amounts of biomass reside in the aboveground and belowground portions of over-
story trees.
Three approaches provide ecologists with insight into the biomass and productivity of forest
ecosystems. One frequently used approach is based on relationships that exist between the dimen-
sions of forest trees (diameter and height) and their weight. These species-specific, or allometric,
relationships are typically expressed in the form of a mathematical equation in which the diameter
at breast height (DBH) and height of a particular tree is used to estimate the weight of leaves,
branches, stem, and sometimes roots. Development of allometric biomass equations is a very
labor-intensive task in which individual trees of a particular species which span a range of DBH
and height are harvested, divided into components, dried, and weighed. Allometric equations have
been developed for most North American forest trees and a significant number of shrub species
(Whittaker 1966; Tritton and Hornbeck 1982; Smith and Brand 1983). Because of the great diffi-
culty of collecting tree roots from the soil, most allometric biomass equations are used to predict
the aboveground biomass and productivity of forest ecosystems.
With the use of allometric biomass equations, aboveground net primary productivity can be
estimated with the following expression:
ANPP B L
In this equation, the annual change in biomass (ΔB) of all aboveground plant tissue plus the
biomass of senesced leaves, seeds, and flowers (L) produced in that year is equal to the above-
ground net primary productivity of the ecosystem. The biomass of senesced leaves, seeds, and
flowers is estimated by collecting this material in a fixed area as they are shed by overstory trees.
Estimates of root biomass and root litter are often unavailable and hence are often absent from
estimates of net primary productivity.
Aboveground net primary productivity is estimated by making measurements of tree or
shrub heights and diameters in successive years, computing the biomass for each year using allo-
metric equations, and subtracting the biomass estimate in year 1 from that estimated in year 2.
Estimates of net aboveground primary productivity obtained using this technique are readily avail-
able for a wide array of forest ecosystems distributed throughout North America and Europe (see
Reichle 1981; Grier et al. 1989).
A second approach estimates net primary productivity by directly measuring the
physiological processes controlling the C balance of ecosystems. This approach entails
measuring photosynthesis and respiration for representative ecosystem components (i.e.,
leaves, branches, stems, soil) and then extrapolating their CO2 flux to the entire ecosystem.
Ecologists have used this approach to obtain estimates of gross primary productivity, ecosystem
respiration, and net primary productivity for a late-successional oak–pine ecosystem at the
Brookhaven National Laboratory on Long Island, New York (Botkin et al. 1970; Woodwell and
Botkin 1970).
In this study, leaves, branches, stems, and the soil surface were enclosed within small
chambers that were connected to gas analyzers which measured CO2 concentration over time
(Figure 18.12). Because the entire forest canopy cannot be practically enclosed within a cham-
ber, estimates of photosynthesis and dark respiration from a known quantity of leaves must be
scaled to the entire forest canopy. Care must be taken to obtain estimates of photosynthesis
and dark respiration for sun and shade leaves, thereby accounting for differences in photosyn-
thesis from the top to the bottom of the canopy. Similarly, estimates of branch, stem, and soil
respiration must be extrapolated to a flux from the entire ecosystem. Given that rates of CO2
exchange from a small portion of the entire ecosystem are actually measured, great care must
be taken to ensure that data are collected on representative samples of each ecosystem compo-
nent. Otherwise, estimates could be greatly in error when they are extrapolated to the entire
ecosystem.
Using this approach, gross primary productivity at the Brookhaven forest was estimated to be
14.0 megagrams of C per hectare (Mg C ha−1) during the growing season, whereas leaf dark respira-
tion (4.0 Mg C ha−1), stem respiration (5.5 Mg C ha−1), and soil respiration (4.5 Mg C ha−1) totaled
14.0 Mg C ha−1. These results indicate that net ecosystem productivity was 0 Mg C ha−1, suggesting
that this relatively late-successional forest is neither gaining nor losing C (i.e., GPP = RA + RH).
Exhaust
Leaf Chamber
Stem Chamber Air

CO2
Sampling
Analyzer
System
Exhaust
Exhaust
Soil
Chamber
Compressed Refrigeration
Air Supply Compressor
F I G U R E 18 .1 2 A CO2 exchange system used to measure the C balance of forest ecosystems based on
quantifying the physiological processes responsible for C fixation and loss. Source: After Woodwell and
Botkin (1970) / Springer Nature. Reprinted from Analysis of Temperate Forest Ecosystems, © by Springer-
Verlag Berlin, Heidelberg 1970. Reprinted with permission of Springer-Verlag New York, Inc.
One drawback of the small chamber approach is that, despite great effort, estimates of NPP
are often inaccurate because root respiration cannot be separated from the respiration of soil
microorganisms (i.e., soil respiration = root maintenance, root construction, and microbial respi-
ration). This currently remains a challenge for ecologists because it places uncertainty on the
actual belowground C budget of forest ecosystems, as well as other terrestrial ecosystems across
the Earth.
Recently, micrometeorological techniques have provided a third approach to study the
exchange of CO2 between terrestrial ecosystems and the atmosphere. The eddy covariance
method quantifies the net exchange of CO2 by measuring vertical gradients of CO2 from the forest
floor to above the canopy. In combination with wind speed, wind direction, and temperature, the
vertical CO2 profile can be used to estimate the flux of C into and out of ecosystems on hourly,
daily, and yearly time intervals (Figure 18.13). This technique has been used to study the C
balance of a wide range of terrestrial and aquatic ecosystems distributed around the Earth,
including forests, savanna, grassland, desert, tundra, and lakes. This micro-meteorological
approach has enabled scientists to gain understanding of the degree to which the Earth’s ecosys-
tems function as a sink or source of CO2 to the atmosphere, thereby improving our understanding
of the Earth’s C balance. In a second-growth northern hardwood ecosystem at the Harvard Forest
in Massachusetts, for example, net ecosystem production estimated using the eddy correlation
was 3.7 Mg C ha−1 y−1. Ecosystem respiration, calculated from the CO2 flux during nighttime
(a)
20
Meters
Micrometeorological CO2 Sensors Net Radiation

Sensors: Meter
Temperature, Humidity
(b)
F I G U R E 18.13 Instrumentation of a forest ecosystem to measure the net exchange of C with the
atmosphere. (a) The eddy covariance approach uses the vertical gradient of CO2 from above the canopy to the
forest floor to estimate the net uptake and release of CO2 by terrestrial ecosystems. (b) Photograph of the eddy
covariance tower located at the Harvard Forest in Massachusetts, United States. Source: J. William Munger.
hours, was 7.4 Mg C ha−1 y−1, suggesting that gross primary productivity was 11.1 Mg C ha−1 y−1
(Wofsy et al. 1993). In contrast to the late-successional forest at Brookhaven, New York, this
second-growth northern hardwood ecosystem is accumulating C based on the fact that NEP is
positive, that is, GPP > RA + RH.
It is important to realize that the eddy correlation technique also suffers from the fact that
soil respiration cannot be broken into separate contributions of plant roots and soil microorgan-
isms, making it difficult to estimate the proportion of net ecosystem production and net primary
production occurring below the soil surface. One cannot simply accomplish this task by separat-
ing roots from soil, and then measuring the respiration of each component; root respiration is
higher in the presence of rhizosphere microorganisms, and the respiration of rhizosphere micro-
organisms is higher in the presence of roots. This reciprocal relationship confounds our ability to
discern the separate contribution of roots and microorganisms to the flux of CO2 from soil. Glob-
ally, this flux is 8 times greater than the annual combustion of fossil fuels by humans, making it a
significant component of the Earth’s C balance. Clearly, the greatest uncertainty associated with
estimating the C balance of ecosystems, and consequently the Earth, occurs in measuring rates of
root production and the amounts of CO2 separately respired by plant roots and soil microorgan-
isms. This remains a frontier in the study of ecosystem C balance.
CLIMATE AND PRODUCTIVITY

Although there is still much to learn about the C budget of forests and other terrestrial ecosystems,
striking patterns and predictable differences in ANPP occur over the Earth’s surface. Geographic
differences in ANPP are not random and are strongly linked to patterns of climate. It is no surprise
that the productivity of desert or shrub steppe ecosystems is much less than that of prairie
or forests.
Differences in temperature and precipitation over the Earth’s surface influence the distribu-
tion of ecosystems and the processes controlling ANPP. Figure 18.14 summarizes the relationship
between temperature, precipitation, and the ANPP of a wide array of terrestrial ecosystems, includ-
ing tundra, deserts, grasslands, and forests from tropical, temperate, and boreal regions. Above-
ground net primary productivity is relatively low in cold (Figure 18.14a) and dry (Figure 18.14b)
climates, and it rapidly rises as both temperature and precipitation increase. This relationship is
F I G U R E 1 8 .1 4 (a) Mean (a)

annual temperature, (b) mean
annual precipitation, and the
aboveground net primary 25
productivity (ANPP) of
Productivity (Mg ha–1 y–1)
terrestrial ecosystems. Data 20

used to construct these
relationships were collected 15
from a wide array of terrestrial
ecosystems including deserts,
10
grasslands, forests, and tundra.
Increasing temperature and
precipitation generally result 5
in greater amounts of ANPP.
0
Source: After Lieth (1973) /
–13.0 –10.0 –5.0 0.0 5.0 10.0 15.0 20.0 25.0 30.0
Springer Nature.
Temperature (°C)
(b)
25
Productivity (Mg ha–1 y–1)
20
15
10
0
0 500 1000 1500 2000 2500 3000 3500 4000 4500
Precipitation (mm/y)
striking, given the wide diversity of plants and ecosystems used to compile this relationship. Why
are global-scale patterns of ANPP so closely related to climatic factors, such as temperature and
precipitation?
The connection between climate and ANPP arises due to a fundamental trade-off during
photosynthesis—plants lose water when they open their stomates to fix CO2 from the atmosphere.
At a global scale, the amount of CO2 fixed by plants and terrestrial ecosystems is limited by the
quantity of water available for transpiration. Actual evapotranspiration (AET) is an ecologically
important climatic attribute that incorporates precipitation, temperature, and the water-holding
capacity of soil to estimate the actual amount of water lost to the atmosphere from evaporation
(soil surface) and plant transpiration. AET often is compared with potential evapotranspiration
(PET), which estimates the amount of evapotranspiration that would occur if the supply of soil
water was unlimited. In regions where temperatures are high and precipitation is infrequent (i.e.,
deserts), PET greatly exceeds AET because low amounts of precipitation limit evapotranspiration
even when ample energy is available to evaporate water from the surfaces of soil and leaves.
Figure 18.15 illustrates the relationship between the ANPP of a wide array of terrestrial eco-
systems from different climatic regimes and AET. Desert ecosystems occur in regions where little
water is available for evapotranspiration and hence can be found at the low end of the AET axis in
Figure 18.15. Forests, by contrast, occur where annual amounts of precipitation exceed AET (upper
end of the axis); a situation in which the availability of water does not often limit evapotranspira-
tion. Notice that a positive relation exists between ANPP and AET—an increase in the amount of
water returned to the atmosphere via AET results in greater ANPP. Again, this relationship exists
because plants must open their stomates to fix CO2 from the atmosphere and simultaneously lose
water during transpiration. At a global scale, the amount of water available for transpiration sets
the upper bounds on plant C gain, and hence is responsible for differences in ANPP among ecosys-
tems from different climatic regions.
Mountainous regions exhibit dramatic temperature and precipitation gradients that create
differences in ecosystem productivity equivalent to those depicted in Figure 18.15, albeit at a much
smaller spatial scale. The Coastal and Cascade Mountain ranges in the Pacific Northwest help cre-
ate marked changes in climate, vegetation, and productivity across relatively small distances
F I G U R E 18. 15 Actual evapotrans-

piration (AET) and aboveground net
Net primary productivity (Mg ha–1 y–1) aboveground
20
primary productivity (ANPP) for
alpine, boreal, temperate, and
tropical terrestrial ecosystems. The
10 positive relationship between AET
Forest and ANPP emphasizes that climate
factors impose important constraints
5 on global patterns of ANPP. Source:
After Rosenzwieg (1968) / University
of Chicago Press. Reprinted with
20 permission from The American
Naturalist, © 1968 by the University
of Chicago. Reprinted with permis-
10 sion of the University of Chi-
cago Press.
0.5 Desert
100 200 500 1000 2000

Actual evapotranspiration (mm y–1)
(Chapter 8). At low elevations in the Coastal Range (200 m) near the Pacific Ocean, relatively warm
and wet (246 cm y−1) conditions give rise to tall-statured forests dominated by Sitka spruce and
western hemlock (Franklin and Dyrness 1973). Inland lies the Cascade Range where relatively
warmer and drier conditions at lower elevation give rise to forests ecosystems dominated by
Douglas-fir and grand fir. Moving eastward and up in elevation, orographic cooling produces pro-
gressively colder and wetter climatic conditions in which Douglas-fir, mountain hemlock, and
Pacific silver fir dominate the forest overstory. Rain shadow conditions occur on the east slope of
the Cascade Range, creating progressively warmer and drier climates as one moves further east-
ward and down in elevation. Ponderosa pine ecosystems occur on the east slope (870 m) of the
Cascade Range in regions receiving 40 cm y−l of precipitation; western juniper (25 cm y−1) and sage-
brush (20 cm y−1) occur in the drier climates at lower elevations further on the east slope.
Along this relatively broad climatic gradient in the Pacific Northwest, ANPP ranges from
0.3 Mg ha−1 y−1 in dry, sagebrush ecosystems to 15 Mg ha−1 y−1 in wet, coastal western hemlock eco-
system (Gholz 1982), a range of values equivalent to that in Figure 18.15. In this mountainous
region, patterns of ANPP are strongly related to differences in temperature, precipitation, and soil
water storage. Gholz (1982) combined these physical site factors to estimate the water availability
in Pacific Northwest ecosystems using the following expression:
WB P E SWC.
In this equation, P is monthly precipitation (cm), E is monthly potential evaporation (cm),

SWC is the average monthly soil water content (cm), and WB is the monthly water balance. Above-
ground annual net primary productivity is relatively low in juniper and sagebrush ecosystems in
which potential transpiration greatly exceeds precipitation and soil water content (i.e., a negative
water balance; Figure 18.16). As either precipitation increases or evaporation decreases, the annual
water balance in this region becomes less negative, which relates to a linear increase in ANPP
(Figure 18.16). Again, this reinforces the idea that water availability sets a fundamental limit on
ecosystem productivity by controlling the amount of CO2 fixed through photosynthesis.
Although evapotranspiration is strongly related to global gradients of productivity, it is not
the sole determinant of ecosystem productivity. In Figure 18.15, study the variation in ANPP at the
point where AET equals 500 mm; note that ANPP ranges from 3 to 10 Mg ha−1 y−1at this point in
Figure 18.15. Such variation is not uncommon in temperate forests in which ANPP differs by a
factor of 10 (Grier et al. 1989), even though AET varies to a lesser extent (500–700 mm). Also real-
ize that ecosystems with very low productivity can be found in warm, tropical regions receiving
relatively large amounts of precipitation, something that would not be predicted from the relation-
ships depicted in Figures 18.14 and 18.15. What factors cause productivity to differ so widely
among ecosystems within a particular climatic region? In the following sections, we explore the
extent to which soil properties, species composition, and stage of ecosystem development modify
NPP and NEP within a particular climatic region, somewhat independently of the amount of water
available to plants.
SOIL PROPERTIES, FOREST BIOMASS, AND ANPP

Within a particular climatic region, soil parent material can exert a substantial influence on the
productivity of forest ecosystems through its influence on soil water-holding capacity. Recall that
sandy-textured soils hold relatively lower amounts of plant available water compared to soils of
loam or clay loam texture (Chapter 9). In the Lake States, glacial activity (10 000–14 000 years BP)
created local landscapes in which parent materials range from coarse outwash sands to relatively
fine-textured glacial till; such changes occur over relatively short distances, for example, 100 m to
Coastal
western hemlock
15
Douglas fir
NPP (Mg ha–1 y–1)
10
Ponderosa pine
5 Sagebrush
0
+20 0 –20 –40 –60 –80 –100 –120
Water balance (cm)
F I G U R E 1 8 .1 6 The water balance and aboveground net primary productivity (ANPP) of ecosystems in the
Pacific Northwest. The relationship between water balance and ANPP emphasizes the idea that increases in
water availability along climatic gradients in mountainous regions relate to an increase in net primary
productivity. Source: After Gholz (1982) / John Wiley & Sons. Reprinted with permission of the Ecological
Society of America.
1 km. Because these contrasting parent materials occur within a particular climatic regime, they
present the opportunity to examine how landform-mediated differences in soil parent material
(i.e., plant available water) modify the productivity of forest ecosystems.
One example occurs in northern Lower Michigan, where productive northern hardwood
forests growing on loamy glacial till occur in the same landscape as lower productivity oak
dominated (black and white) ecosystems of the sandy outwash plains. Overstory biomass
(208 Mg ha−1) and productivity (3.2 Mg ha−1 y−1) of a northern hardwood ecosystem are twice that
of xeric oak ecosystem (105 Mg ha−1 and 1.5 Mg ha−1 y−1; Host et al. 1988). Clearly, soil texture can
greatly modify the influence of regional climate on the productivity of forest ecosystems.
Differences in soil parent material also give rise to differences in soil N availability, which, at a
regional climatic scale, controls the productivity of many boreal, temperate, and a some tropical for-
ests. Unlike agricultural ecosystems that receive N additions from chemical fertilizer, soil N availability
in forests and other less-intensively managed ecosystems is controlled by the activity of saprotrophic
soil microorganisms. Most soil microorganisms obtain energy and essential nutrients during the
decomposition of plant litter (senesced leaves and fine roots). During this process, only a portion of
the N contained in litter is used by soil microorganisms to maintain or build biomass. The remainder
is released into soil solution as ammonium (NH4+), where it can be assimilated by plant roots (i.e., N
mineralization; see Chapter 19 for details). In the Upper Lake States, variation in ANPP among forest
ecosystems is related to differences in soil N availability, wherein relatively high rates of ANPP (over-
story only) occur in soils with high rates of N mineralization (Figure 18.17). It is important to note
that fine-textured soils, which hold relatively large amounts of plant available water, also release the
greatest amounts of N for plant growth. The relationship between ANPP and soil N availability is not
unique to Lake States forests, because ANPP also is related to soil N availability in grassland (Schimel
et al. 1985), shrub steppe (Burke 1989), and desert (Fisher et al. 1987) ecosystems.
Soil texture can modify the influence of climate on ANPP by controlling the amount of water
and N that is available for plant growth. In other words, climate sets the regional potential for pro-
ductivity through patterns of temperature and precipitation, but that potential is constrained, or
realized, by differences in soil texture which in turn control the availability of water and N within
the local landscape.
4.0
3.5
ANPP (Mg ha–1 y–1)
3.0
2.5
2.0
1.5
1.0
40 60 80 100 120
Mineralization (µg N g–1)
F I G U R E 1 8 .1 7 In the Lake States region, aboveground productivity of forest ecosystems is well correlated
with the rate at which soil microorganisms release N from soil organic matter. Illustrated in this figure is the
ANPP (overstory only) and N mineralization rate of nine upland forest ecosystems in northern Lower
Michigan. The positive relationship between these variables suggests that soil N availability imposes
an important constraint on ecosystem productivity within the climate of the Great Lakes region.
Source: After Zak et al. (1989) / Canadian Science Publishing. Reprinted with permission of National
Research Council Canada.
BIOMASS ACCUMULATION DURING ECOSYSTEM DEVELOPMENT

Following an event that destroys vegetation, living plant biomass increases over time from near
zero to that of a mature forest ecosystem (Figure 18.18). Concomitant changes also occur in the
activities of decomposing organisms and in the biomass stored in forest floor and mineral soil. The
rate at which biomass accumulates in forest ecosystems is controlled by the amount of CO2 fixed
through photosynthesis (i.e., GPP) and the amount of C returned to the atmosphere via mainte-
nance and construction respiration of plants and saprotrophic soil microorganisms (i.e., RA + RH).
These processes are illustrated for an idealized forest following the destruction of living plant
biomass in Figure 18.18.
In the newly established forest, living plant biomass (Mg ha−1) is nearly zero and it rapidly
increases during the first 20–25 years of ecosystem development. After this period of time, the
increase in living plant biomass begins to slow and it eventually attains a maximum value, which
is near year 50 in Figure 18.18b. The biomass of leaves initially increases and attains a maximum
value after approximately 20 years (Figure 18.18b), the point at which canopy closure occurs.
Recall that net canopy photosynthesis, and hence the amount of photosynthate available for plant
growth and maintenance, is a product of leaf photosynthetic rate (μmol m−2 s−1) and total canopy
leaf area, as well as canopy geometry. Leaf biomass (Mg ha−1) and area (m2 ha−1) are well correlated
in most forest trees, indicating that the amount of leaf area (or biomass) at canopy closure sets the
upper limit on canopy photosynthesis and GPP. It follows that this relationship also sets a limit
on the amount of C that can be allocated to growth and maintenance. Notice that total biomass
(non-photosynthetic tissue) continues to increase well after leaf biomass has attained its maximum
(Figure 18.18b). This means that net photosynthesis attains a maximum set by environmental
conditions early in ecosystem development, but the further accumulation of non-photosynthetic
biomass causes losses from respiration to increase. This dynamic causes NPP to slow over time, the
ecological implications of which we discuss in the following paragraphs.
(a) F I G U R E 18. 18 Changes in the

GPP processes controlling (a) the C balance
1.0
of forest ecosystems and (b) the rate
of biomass accumulation following a
Carbon flux relative to GPP
0.8
RA disturbance that destroys living plant
biomass. Such a disturbance includes
0.6 NPP
intensive harvesting of overstory trees,
fire, or windstorm. Change in net
0.4 secondary production is not illustrated
NEP and accounts for a relatively minor
0.2 RH flux of C in forest ecosystems.
0
50 100 150
–0.2
(b) Total ecosystem
600
Living plant
400
Biomass (Mg ha–1)
200 Mineral soil
Forest floor
Leaves
0 50 100 150
Years following disturbance
The pattern of total biomass accumulation in this forest ecosystem indicates the balance bet-
ween GPP and RA + RH changes during ecosystem development. In Figure 18.18a, GPP is relatively
consistent after approximately 20 years, whereas RA continues to increase for a longer period of
time. This pattern suggests that an increasingly greater amount of fixed C is allocated to tissue
maintenance, diminishing the amount allocated to the construction of new tissue. This is consis-
tent with the increase in non-photosynthetic biomass (total) long after the biomass of photosyn-
thetic tissue has reached its maximum at canopy closure. The initial increase and subsequent
decline in NPP (i.e., GPP − RA) reflects this fundamental trade-off: C allocated to maintenance
cannot be allocated to growth. Clearly, the C balance of plants exerts direct control on the C balance
of entire ecosystems.
Changes in forest floor biomass following a disturbance reflect a balance between litter pro-
duction (i.e., abscised leaves, fine roots, twigs, reproductive structures) and the amount of litter
consumed by decomposing organisms (Figure 18.18b). Early in ecosystem development (0–15 years),
plant litter production is low, but activities of decomposing organisms are enhanced by the
relatively warmer, moister conditions at the soil surface, resulting from the absence transpira-
tion by overstory trees. Note that RH increases during the first several years following disturbance.
Low litter production in combination with enhanced decomposition rates results in an initial
decline in forest floor biomass. Biomass begins to accumulate in the forest floor when litter pro-
duction exceeds decomposition, a situation that occurs between 20 and 50 years after disturbance
in the example illustrated in Figure 18.17. The relatively constant biomass of the forest floor late
in ecosystem development suggests that litter inputs to forest floor are balanced by rates of decom-
position. That is, the forest floor is neither gaining nor losing biomass late in ecosystem
development.
The amount of organic matter stored in mineral soil reflects the long-term (1000–
1 000 000 years) balance between plant litter production and the rate at which decomposing organ-
isms return it to the atmosphere as CO2. Biomass or organic matter stored in mineral soil often
changes little following a disturbance that destroys a forest overstory (Figure 18.18b; Bormann and
Likens 1979). In some situations, the amount of organic matter lost from the forest floor and min-
eral soil is quickly regained by inputs from rapidly developing vegetation (Gholz and Fisher 1982).
However, soil erosion following disturbance can result in a substantial reduction in soil organic
matter, particularly in areas of high rainfall and steep topography.
Net ecosystem production measures the C balance of terrestrial ecosystems, and it can dra-
matically change during ecosystem development. Immediately following a disturbance, high rates
of heterotrophic respiration (RH) and low rates of NPP (GPP – RA) result in a net flux of C from the
ecosystem, as depicted in Figure 18.18a. This negative C balance is marked by a reduction in total
ecosystem biomass, which results in a negative NEP. Recall that any disturbance that diminishes
NPP and accelerates rates of organic matter decomposition (RH) has the potential to reduce NEP
below 0. In such a situation, ecosystems are sources of CO2 to the atmosphere, rather than func-
tioning as a sink. Additionally, total ecosystem biomass increases only when GPP exceeds the
respiratory loss of C from plants and heterotrophic organisms, a situation occurring from 5 to
75 years following disturbance in our example. Note that NEP declines to 0 late in ecosystem
development (130 years), as a result of the relationship GPP = RA + RH. Because the fixation and
loss of C on an ecosystem basis are in balance, total ecosystem biomass reaches a relatively
constant value late in ecosystem development (Figure 18.18b).
The generalized patterns illustrated in Figure 18.18 occur in all terrestrial ecosystems,
albeit at different magnitudes and over different periods of time. After an initial decline, for
example, total ecosystem biomass rapidly accumulates in young slash pine plantations in Florida
(Figure 18.19; Gholz and Fisher 1982). Maximum leaf biomass and canopy closure occur after
5 years, which sets an upper limit on canopy leaf area and the amount of photosynthate available
for growth and maintenance. Notice that total biomass continues to accumulate long after can-
opy closure, indicating that greater amounts of GPP are allocated to maintain the increasing
amount of non-photosynthetic tissue (Figure 18.19b). This trade-off in C allocation from growth
to maintenance is further illustrated by the peak in tree and understory biomass at 25 years
(GPP ≈ RA; Figure 18.19a). After 25 years of growth, overstory trees are commercially harvested
and seedlings are planted to reestablish a new plantation. These activities again reduce living
plant biomass to near zero and restart the process of biomass accumulation in the newly develop-
ing plantation.
A substantial amount of biomass is also found in dead trees either killed by a disturbance or
subject to long-term mortality as forests age. Severe wildfires, windstorms, hurricanes, and insect
and pathogen outbreaks kill trees but leave behind a majority of their biomass as deadwood, either
standing or fallen onto the forest floor. One notable exception to this pattern is harvesting, which
actively removes most large wood (note that Figure 18.17b lacks a pattern for deadwood and is
therefore representative of forest recovery following harvesting). Older forests produce less dead-
wood via chronic mortality than do disturbances, but mortality of individual old trees is an impor-
tant and long-lasting source of deadwood, particularly when the trees are large (Harmon and
Hua 1991). Deadwood is therefore typically most abundant immediately after a disturbance,
(a)
350
300
Total ecosystem
Organic matter (Mg ha–1)
250
200
Forest floor and soil
150
Tree and understory
100 (includes coarse roots)
50
0
0 5 10 15 20 25 30 35
(b) Stand age (y)
200
Tree total
Aboveground tree biomass (Mg ha–1)
150
Stem wood
100
Dead
50 Branches Stem bark Coarse roots
Live branches
Foliage
0 5 10 15 20 25 30 35
Stand age (y)
F I G U R E 1 8 .1 9 Changes in (a) ecosystem biomass and (b) the components of living plant biomass in a
developing slash pine ecosystem in Florida. Source: After Gholz and Fisher (1982) / John Wiley & Sons.
Reprinted with permission of the Ecological Society of America.
declines to a minimum in middle stand ages when the wood decays with only minor additional
inputs from the reestablished forest, and then increases again in old age due to mortality of older
trees (Sturtevant et al. 1997; Kashian et al. 2013).
The dynamics of deadwood in forests are often critical to NEP. When disturbances kill many
or all trees and create a substantial amount of deadwood, NEP is strongly shaped by the balance
between (i) the amount of carbon lost via heterotrophic respiration that occurs with the decompo-
sition of deadwood, and (ii) the amount of carbon gained in reestablishing vegetation. The high
rates of heterotrophic respiration and corresponding negative NEP immediately following a distur-
bance, as described in the earlier text, largely occur because of the decomposition of deadwood,
which continues for several decades (Crutzen and Goldhammer 1993). Biomass accumulation in
growing trees eventually compensates for C lost through decomposition as the forest reestablishes
and RH slows, and NEP becomes positive. In old forests, NEP eventually approaches zero as NPP is
reduced in older trees and tree mortality produces additional large deadwood. Deadwood is there-
fore an important, though often overlooked, component of ecosystem C dynamics. For example,
deadwood represented 36% of total ecosystem biomass in lodgepole pine forests in Yellowstone
National Park, ranging from 2% in a middle-aged stand with little deadwood to 80% immediately
after a stand-replacing fire (Kashian et al. 2013).
Natural disturbance regimes of fire and windstorms also alter the C balance of forest ecosys-
tems. Prior to European settlement, fire was an important component of ecosystem development in
the hardwood forests of southern Wisconsin (Figure 18.20; Loucks 1970). Repeated, random fires
altered the C balance of these ecosystems by destroying living plant and reducing GPP. Biomass oxi-
dized by fire to CO2 also represents a loss of C and causes a decline in total ecosystem biomass. In
combination, these responses cause NEP to fluctuate in a cyclic manner that reflects the magnitude
and frequency of fire. Notice that the period between some fires in Figure 18.19 is sufficient for RA
and RH to equal GPP, producing an NEP that is near 0 late in ecosystem development. In some eco-
systems, fire may occur during a period of total biomass accumulation (GPP > RA + RH), thus pre-
venting total biomass from attaining a “steady state” that might occur late in ecosystem development.
Natural disturbance that destroys living plant biomass can occur over relatively long periods
of time (i.e., 100s to 1000s years) in some ecosystems, thus allowing them to reach a “steady state”
in total biomass. In the Pacific Northwest, old-growth ecosystems dominated by Douglas-fir and
western hemlock can attain ages of 450 years or more. Although NEP is low (or 0) in these old
growth forests, they have accumulated large amounts of C relative to second-growth forests
(Table 18.7). Total ecosystem biomass of old-growth forest is over twice that of the second-growth
forest, suggesting that GPP exceeds RA + RH for a substantial period of time between 60 and
450 years after disturbance. The accumulation of biomass in stemwood and coarse woody debris
accounts for 77% of the increase in total biomass between the second-and old-growth forests.
Although NPP and NEP decline through time, substantial amounts of biomass can accumu-
late in old-growth forests. Clearly, the conversion of old-growth to second-growth forests has the
potential to greatly alter the cycling and storage of C in the Pacific Northwest.
Random Disturbance (Fire–windstorm)
Ecosystem
Biomass
1250 1450 1650 1850
+
NEP
0
F I G U R E 1 8 .2 0 The relationship between fire frequency, total ecosystem biomass, and net ecosystem
productivity for forests in southern Wisconsin. Fire was an important disturbance prior to European
settlement that functioned to alter ecosystem C balance in a cyclic manner. The present suppression of fire in
Lake States forests has greatly altered this relationship. Note the lack of fire following 1850 and the relatively
long period of time over which NEP is greater than 0. Source: Modified from Loucks (1970).
Table 18.7 Biomass pools in second-and old-growth ecosystems dominated by Douglas-fir and west-
ern hemlock in the Pacific Northwest.
60-year-old forest 450-year-old forest

Component Mg ha −1
Foliage 12 13
Branch 15 56
Stem 308 687
Coarse roots 62 151
Fine roots 12 12
Forest floor 15 55
Coarse woody debris 8 206
Mineral soil 119 119
Total 551 1299
Source: Based on Harmon et al. (1990). Effects on carbon storage of conversion of old-growth forests to young forests. Reprinted
with permission from Science, p. 700, © 1990 by the American Association for the Advancement of Science.
SOIL N AVAILABILITY AND BELOWGROUND NET PRIMARY PRODUCTIVITY

Allocation of C to belowground plant growth accounts for a substantial proportion of NPP in forest
ecosystems, as well as others. Studies conducted in coniferous and deciduous forests suggest 50%
of NPP can be allocated to the growth and maintenance of roots. Earlier in this chapter (see
C Balance of Plants), we observed that soil N availability substantially alters the proportion of
carbohydrate allocated to the leaves, stem, and roots of individual plants. In this section, we extend
this idea to an ecosystem level and summarize patterns of belowground NPP in forest ecosystems.
We focus on the extent to which soil N availability modifies the proportion of NPP allocated to fine
roots and associated mycorrhizae, which are the nutrient and water absorbing organs of forest
trees (Chapter 19).
Quantifying fine-root production in forest ecosystems remains one of the most important,
but difficult, challenges to ecologists. Unlike leaves that senesce and abscise during a defined time
of the year, root production and mortality are continuous processes occurring throughout the
entire growing season. Most often, belowground NPP is estimated by measuring the monthly
change in root biomass over 1 year. Other techniques include the use of mini-rhizotrons and sev-
eral methods that employ a nutrient budget. These methods of quantifying belowground produc-
tion have led ecologists to very different conclusions regarding the influence of soil N availability
on belowground NPP. Regardless of this uncertainty, it is clear that the production of plant tissue
below the soil surface in forest ecosystems is a substantial component of NPP.
Two alternative views regarding the influence of soil N availability on fine-root dynamics
have emerged (Figure 18.21). One hypothesis argues that fine-root life span (i.e., mortality) does
not change along gradients of soil N availability. It also predicts that fine-root production declines
in N-rich soil, enabling the plant to allocate proportionately more carbohydrate to foliage and
stems (Hypothesis 1 in Figure 18.21). Such a view contends that fewer roots are needed to forage
for soil resources when they are in abundant supply. Evidence for this view comes from a
comparison of two 40-year-old Douglas-fir ecosystems occurring on soils that differ in N availabil-
ity (Table 18.8). Sequential monthly measurements of fine-root biomass were used to estimate
fine-root production and allometric equations were used to estimate the change in coarse root bio-
mass. The proportion of NPP allocated to fine-root production declined from 36.4% in the N-poor
soil to 7.9% in the N-rich soil, but rates of fine-root mortality were unchanged (Keyes and Gri-
er 1981). At the same time, allocation to leaves rose from 33.1% in the N-poor soil to 55.5% in the
Soil N Availability and Belowground Net Primary Productivity 483
Relative Units Hypothesis 1
Mortality
Pe
rc en
St
an to
din f to
gB tal
iom NP
as P
s
Low High
Nitrogen Availability
Hypothesis 2
ity
r tal
Mo
Relative Units
Percent of Total NPP
St
an
din
gB
iom
as
s
Low High
Nitrogen Availability
F I G U R E 1 8 .2 1 Two alternative hypotheses describing the allocation of NPP to the production of fine roots
in forest ecosystems. Source: Reprinted from Hendricks et al. (1993) / with permission of Elsevier. With kind
permission of Elsevier Science—NL, Sara Burgerhart-straat 25, 1055 KV Amsterdam, The Netherlands.
N-rich soil (Table 18.8). In combination, these observations suggest that the proportion of NPP
allocated belowground declined in N-rich soil, facilitating a decline in fine-root biomass.
Another view contends that allocation to fine roots remains constant along a gradient of N
availability, but the life span of fine roots declines when soil N availability is relatively high
(Figure 18.21; Hypothesis 2). Therefore, the proportion of NPP allocated to fine-root growth remains
constant, but an increase in mortality facilitates a reduction in fine-root biomass when soil N avail-
ability is high. This view of fine-root dynamics has developed from the use of an N budget to estimate
the amount forest trees allocate to the annual production of fine roots. It assumes that the allocation
of N to fine roots equals the difference between the amount of N assimilated by forest trees and the
amount of N allocated to aboveground growth. This approach has been applied to a series of hard-
wood and coniferous forests in Wisconsin that occur along a gradient of soil N availability (Nadel-
hoffer et al. 1985; Figure 18.22) and has yielded results in direct contrast to the aforementioned
study in Douglas-fir. Using the N budget approach, the biomass N content (g N m−2) of fine roots
declined along an increasing gradient of soil N availability, but the total amount of N allocated to
Table 18.8 The allocation of net primary productivity to aboveground and belowground plant tissues
in 40-year-old Douglas-fir stands growing on soils of low and high fertility.
Low N soil Percent of High N soil Percent of

Component
(Mg ha−1 y−1) total (Mg ha−1 y−1) total
Stem 5.1 33.1 9.9 55.5
Branch 0.2 1.3 0.6 3.4
Leaves 2.0 13.0 3.2 18.0
Coarse roots 2.5 16.2 2.7 15.2
Fine roots 5.6 36.4 1.4 7.9
Total 15.4 100 17.8 100
Source: After Keyes and Grier (1981) / Canadian Science Publishing. Reprinted with permission of National Research Council Canada.
8.0 8.0
N Allocated to fine roots (N•m–2•y–1)
Fine-Root biomass N (gN•m–2•y–1)

o
dt
c ate ction
llo du
6.0 N A t pro 6.0
o
Ro
4.0 Root b 4.0

io mass
N
2.0 2.0
0.0 0.0
3.0 5.0 7.0 9.0 11.0 13.0 15.0
Soil N Availability (gN–2•y–1)
F I G U R E 1 8 .2 2 Changes in the N content of fine roots (dashed line) and the total amount of N allocated to
fine-root production (solid line) along a gradient of increasing soil N availability. Source: After Nadelhoffer
fine-root production increased; NPP allocated to fine roots varied little and averaged 27%. The
inverse relationship between fine-root biomass N and the total amount of N allocated to fine-root
production indicates that increased mortality at high soil N availability maintained a relatively low
fine-root biomass. Recall that plant tissues with high N concentrations often have high mainte-
nance respiration costs and a short life span. These attributes may partially explain the increase in
fine-root mortality, because roots in high-N-availability soil often have a high N concentration (i.e.,
proteins mediating ion uptake), which results in more rapid rates of maintenance respiration.
A third approach, based on a short-term C budget, has also produced evidence for a positive
relationship between aboveground and belowground NPP (Raich and Nadelhoffer 1989). It is
based on the principle that all aboveground and belowground litter production is respired by de-
composing organisms on an annual basis. Under that assumption, the allocation of C to root
growth and maintenance is proportional to the flux of C from soil respiration minus the C in
aboveground leaf litter production (for details, see Raich and Nadelhoffer 1989). Soil respiration
and leaf litter production have been quantified in many forest ecosystems, and Figure 18.23 sum-
marizes the relationship between C allocated to leaf litter and fine roots (fine-root production,
construction respiration, and maintenance respiration) for 30 forest ecosystems occurring in
tropical, temperate, and boreal regions. Notice that leaf litter and the amount of C allocated to fine
Soil N Availability and Belowground Net Primary Productivity 485
12
Root Allocation (Mg C•ha–1•y–1)
10
8 n
tio
oca
t All
o
6 Ro
0
0 1 2 3 4 5 6
Litterfall (Mg C•ha–1•y–1)
F I G U R E 1 8 .2 3 Leaf litter and fine-root production in tropical, temperate, and boreal forests. Estimates of
the total amount of C allocated to fine roots (production + respiration) were derived using a C budget
approach. Source: After Raich and Nadelhoffer (1989) / John Wiley & Sons. Reprinted with permission of the
Ecological Society of America.
roots appear to be positively related on a global basis. Such a relationship provides further evidence
that a relatively constant proportion of NPP is allocated belowground, which provides additional
support for Hypothesis 2 in Figure 18.21.
All the approaches discussed use indirect methods of quantifying fine-root dynamics. That is,
none directly observe the production and death of individual roots, which is the only approach that
has the potential to resolve our uncertainty of the belowground C budget of forest ecosystems. In
the next chapter, we provide further discussion on how the architecture, production, and mortality
of roots mediate the acquisition of soil resources. As you will learn, gaining insight into the below-
ground dynamics of plant growth remains a difficult endeavor for ecologists. Consequently, we still
have a great deal to learn regarding the ecosystem-level allocation of C to fine-root production in
forest ecosystems before we can resolve the disparity in observations of fine-root production and
mortality. This fact clearly limits the ability of ecologists to accurately estimate the NPP of forest
ecosystems, and to fully understand the function of forest ecosystem in the global C cycle. Despite
this limitation, it is clear that the growth and maintenance of plant tissues below the soil surface
account for a significant proportion of C fixed via photosynthesis in many forest ecosystems, and
therefore is a globally significant component of the cycling and storage of C in terrestrial ecosystems.
SUGGESTED
R E A D I N G S
Bormann, F.H. and Likens, G.E. (1994). Pattern and Methods and Instrumentation. New York: Chapman
Process in a Forested Ecosystem. New York: Springer- and Hall 457 pp.
Verlag 266 pp. Smith, W.R. and Hinkley, T.M. (ed.) (1995). Resource
Givnish, T.J. (1988). Adaptation to sun and shade: a whole- Physiology of Conifers: Aquisition, Allocation, and
plant perspective. Aust. J. Plant Physiol. 15: 63–92. Utilization. New York: Academic Press 396 pp.
Mooney, H.A. (1972). The carbon balance of plants. Raghavendra, A.S. (ed.) (1991). Physiology of Trees.
Annu. Rev. Ecol. Syst. 3: 315–346. New York: John Wiley 509 pp.
Pearcy, P.W., Ehleringer, J., Mooney, H.A., and Rundel,
P.W. (ed.) (1991). Plant Physiological Ecology: Field
Nutrient Cycling CHAPTER 19
A lthough carbon (C), hydrogen (H), and oxygen (O) form the basic building blocks of all
biological tissue, plants require a suite of 14 other elements, termed nutrients, in order
to maintain existing tissue and build new biomass (see Table 9.3, Chapter 9). These nutrients
mostly enter terrestrial ecosystems from the atmosphere (wet and dry precipitation) or through
the weathering of soil minerals. Plants assimilate nutrients from soil, incorporate them with
photosynthetically fixed CO2 to form living biomass, and eventually return these nutrients to soil
in dead leaves, roots, branches, and stems (i.e., plant litter). After dead plant material enters the
forest floor (leaves, twigs, etc.) and surface mineral soil (fine roots), it is subject to decomposition,
a microbially mediated process that releases organically bound nutrients into inorganic forms that
can again be assimilated by plant roots. The uptake of nutrients by plant roots, their incorporation
into living tissue, and the release of nutrients during organic matter decomposition cause nutri-
ents to flow or cycle within terrestrial ecosystems. Nutrient cycles are biogeochemical processes,
so named because they are controlled by the physiological activities of plants and soil microor-
ganisms, as well as by the geochemical processes in soil that control nutrient supply (Chapter 9).
F I G U R E 19. 1 A conceptual diagram of

Atmospheric the processes controlling the flow of
Deposition
nutrients into, within, and out of forest
ecosystems. Nutrients enter forest ecosys-
tems through atmospheric deposition, N2
fixation, and mineral weathering. The flow
of nutrients within forest ecosystems is
Translocation controlled by nutrient uptake, the translo-
Aboveground cation of nutrients from senescent tissue,
Litter
Production
Fixation the return of nutrients in leaf and root
Gaseous litter, the decomposition of this litter, and
Loss
the microbial release (i.e., mineralization)
of nutrients from organic matter. Leaching
and gaseous losses (i.e., denitrification) are
processes by which nutrients are lost from
forest ecosystems.
Organic Matter
Belowground Forest Floor and
Litter Mineral Soil
Production
Nutrient Uptake
and Assimilation Decomposition
and Mineralization
Nutrients
in
Soil Solution
Mineral
Weathering
Leaching
487
488 Chapter 19 Nutrient Cycling
Recall that net primary productivity (NPP) differs among boreal, temperate, and tropical
f orests (Chapter 18). Because nutrients are assimilated along with CO2 to form living bio-
mass, differences in net primary productivity among ecosystems suggest that rates of nutrient
cycling also must differ. In this chapter, we discuss processes controlling the input of nutri-
ents to forest ecosystems, the redistribution of nutrients within forest ecosystems by the met-
abolic activities of plants and soil microorganisms, and the loss of nutrients from forest
ecosystems to streams, groundwater, and the atmosphere. These processes are illustrated in
Figure 19.1, which outlines the organization of this chapter. Nutrient uptake by plant roots is
central to the cycling of nutrients in forest ecosystems, and we place emphasis on the
physiological and morphological mechanisms by which plants forage for nutrients in soil. We
also discuss the extent to which the C balance of terrestrial ecosystems influences nutrient
retention and loss within forests, particularly during forest development. Our treatment of
nutrient cycling focuses on nitrogen (N), because the supply of this nutrient in soil most
often limits the productivity of boreal, temperate, and several types of tropical forests (i.e.,
dry and montane tropical forests).
NUTRIENT ADDITIONS TO FOREST ECOSYSTEMS

Nutrient additions to forest ecosystems are
relatively small compared to the annual Atmospheric
Deposition
requirements of actively growing overstory
trees, typically contributing 5–30% of the
annual need for most plant nutrients. Never-
theless, input processes serve as important
conduits by which nutrients flow into forests Translocation
and other terrestrial ecosystems, and over Aboveground

Litter Fixation
relatively long periods of time (100–200 years), Production
Gaseous
they contribute significantly to the nutrient Loss
capital of an ecosystem. Nutrients enter ter-

restrial ecosystems through geological, hydro-
logical, and biological processes. These are
highlighted in the reduced version of
Figure 19.1, which lies next to this paragraph.
Organic Matter
In the paragraphs that follow, we discuss the Belowground Forest Floor and
Mineral Soil
magnitude of nutrient inputs to forest ecosys- Litter
Production
tems from mineral weathering, atmospheric Nutrient Uptake

deposition, and biological fixation, the pri- Nutrients
and Mineralization
mary mechanisms by which nutrients flow in

Soil Solution
into forest ecosystems. Mineral
Weathering
Leaching
MINERAL WEATHERING
Plant nutrients contained in the chemical structure of rocks and soil minerals are largely unavail-
able to plants. Through physical abrasion and chemical dissolution, nutrients are slowly released
from geologic materials into forms that plants take up from soil solution. The rate at which
geologic materials weather to yield plant nutrients is strongly influenced by climate, tempera-
ture, and precipitation, which directly control rates of chemical reactions in soil, especially the
solubilization of plant nutrients from minerals. Rates of mineral weathering are quite slow, and
Nutrient Additions to Forest Ecosystems 489
because not all geologic materials are chemically similar, they weather at different rates and
yield different plant nutrients. All plant nutrients are found in the chemical structure of rocks,
and therefore can be supplied to the plant from mineral weathering, albeit at an annual rate that
does not meet the annual nutritional needs of plants. In Chapter 9, we discussed mineral
weathering as it pertains to soil horizon development, soil acidity, and base saturation. Here, we
focus on aspects of mineral weathering that control the input of plant nutrients to forest
ecosystems.
The most well-known attempt at quantifying mineral weathering in ecosystems began in
1963, when a team of scientists quantified the nutrient cycles of northern hardwood forests in the
White Mountains of New Hampshire (Likens et al. 1977). In this region, a number of comparable
watersheds (i.e., similar in climate, size, geology, topography, soil parent material, and vegetation)
are underlain by a layer of impermeable bedrock, preventing rainfall from reaching groundwater.
Because these watersheds are sealed, streamflow represents the main process by which nutrients
can exit. Therefore, the difference between the atmospheric deposition of nutrients and streamwater
loss of nutrients should approximate the annual release of nutrients from mineral weathering.
Using this approach, a large number of watershed-level studies of mineral weathering have been
completed, allowing a comparison of weathering rates for different climates and parent materials.
In areas dominated by silicate rocks, plant nutrients weather from soil parent materials in the
following order:
Ca > Na > Mg > K > Fe
This sequence partially reflects mineral solubility and partially the degree to which
these elements participate in the formation of secondary minerals, such as phyllosili-
cate clays.
Table 19.1 summarizes nutrient inputs from mineral weathering for forest ecosystems occurring
in different climates and on different parent materials. Notice that the weathering of dolomitic
parent materials (CaCO3 and MgCO3) beneath limber and bristlecone pines yields relatively large
amounts of Ca compared to weathering of adamellite (SiO2:NaAlSi3O8:CaAl2Si2O8). The weathering
of Ca-and Mg-bearing minerals represents an important input of these nutrients in many terres-
trial ecosystems, satisfying 35% of Ca and Mg required for the annual growth of some temperate
Table 19.1 Rates of mineral weathering for forest ecosystems occurring in different climates and on
different soil parent materials.
P K Ca Mg
Ecosystem Parent material Location
kg ha y
−1 −1
Wet tropical Alluvium Venezuela − − 6 1

Douglas-fir Volcanic tuff Oregon − 2 47 12
Limber and Dolomite California − 4 86 52
Bristlecone pines Adamellite California − 8 17 2
Trembling aspen Glacial till Wisconsin 1 4 7 –
Northern hardwood Glacial till New Hampshire 13 7 21 4
Source: Data have been summarized from Boyle and Ek (1973), Fredriksen (1972), Hase and Foelster (1983), Likens et al.
(1977), Marchand (1971), and Wood et al. (1984).
forests (Table 19.1; Likens et al. 1977). Similarly, the weathering Ca-, Fe-, and Al-phosphate min-
erals provides an important source of P for plant growth. It had been thought that mineral
weathering was not a source of N to plants, because few rocks contained N in their chemical struc-
ture. Recently, a global analysis of rock chemical composition has revealed that more N resides
within them than previously thought. This realization suggests that N released from rock
weathering could compose 8–26% of annual N inputs to terrestrial ecosystems, prior to the anthro-
pogenic input of N from atmospheric deposition and fertilizer use (Houlton et al. 2008). Figure 19.2
summarizes the amounts of N contained in surface rock (Figure 19.2a), the amounts annually
released through weathering (Figure 19.2b), and the percent contribution of annual N inputs sup-
plied via mineral weathering. This summary has dramatically revised our thinking of how N enters
terrestrial ecosystems.
ATMOSPHERIC DEPOSITION
Over the past several decades, ecologists have become increasingly interested in understanding the
geographic pattern and amount of nutrients entering terrestrial ecosystems from the atmosphere.
In many portions of the Earth, human activity has altered rainfall chemistry and hence the atmo-
spheric deposition of nutrients in dramatic ways. In the northeastern United States, central Europe
and broadly across Asia, terrestrial and aquatic ecosystems receive enhanced inputs of NO3−, SO42−,
and H+ (i.e., acidic deposition or acid rain) from fossil fuel burning. Both NO3− and SO42− contain
essential plant nutrients, and elevated inputs of these ions have altered the rate and pattern by
which N and S are cycled and stored within terrestrial ecosystems.
Atmospheric deposition occurs through three processes: (i) wet deposition—the addition
of nutrients contained in rain or snow, (ii) dry deposition—the direct deposition of atmospheric
particles and gases to vegetation, soil, or water surfaces, and (iii) cloud deposition—the input of
small, nonprecipitating water droplets (in clouds and fog) to terrestrial surfaces (Fowler 1980;
Lovett and Kinsmann 1990; Lovett 1994). Wet and dry deposition occur in all terrestrial ecosys-
tems, but their importance as a source of nutrients varies greatly from place to place. By contrast,
cloud deposition is generally restricted to coastal and mountainous regions that are frequently
immersed in clouds or fog (Lovett 1994), like coastal giant redwood forests of northern California,
United States.
Wet deposition occurs when solid particles (0.2–2 mm diameter) and gases in the atmosphere
dissolve in water droplets that fall to Earth as rain or snow (Lovett 1994). This process is con-
trolled by several factors that vary markedly from one region to another and result in a broad
range of nutrient inputs from wet deposition. In North America, continental-scale patterns of wet
deposition are relatively well understood (Figure 19.3). For example, anthropogenic emissions of
N and S in the Midwest, Southeast, and Northeast result in widespread patterns of increased NO3−
and SO42− deposition over the eastern United States. By contrast, ammonium (NH4+) and Ca2+
inputs from wet deposition are relatively low in the eastern United States but are elevated in the
eastern Great Plains region and the Midwest. The plowing of agricultural fields and traffic on
unpaved roads cause Ca2+ associated with dust particles to be carried eastward from the Great
Plains region by the prevailing winds. Ammonium released into the air from heavily fertilized
agricultural fields and volatilized from animal manure in concentrated feeding operations is
transported eastward in a similar manner.
Wet deposition entering forest ecosystems can pass through the canopy as throughfall or
portions can flow down tree stems in stemflow. The chemical composition of wet deposition can
be greatly altered by the “wash off” of dry material accumulated on plant surfaces, the leaching of
nutrients from plant tissues, and the direct assimilation of nutrients into leaves. Potassium, Ca2+,
Rock N
(a) reservoir
(kg N m–3)
1.90
1.43
0.95
0.48
0.00
Chemical
(b) N flux
(kg N ha–1 yr–1)
10.0+
7.5
5.0
2.5
0.0
Increased modern N
(c) input from rock
(%)
100+
75
50
25
F I G U R E 1 9 .2 (a) The amount of N contained in surface minerals, (b) the annual release of N from
minerals, and (c) the percent of annual ecosystem N inputs supplied by the weathering of N from surface
minerals. Cool colors (blue) indicate low inputs and amounts, whereas warm colors (red) indicate the
opposite. Source: Houlton et al. (2008) / American Association for the Advancement of Science.
5
5 6
5 10
10 14 10
20 18
14 18 10
20 20
14
14
20
14
20
20
2– 5 – 10
SO 4 10 NO3
6
1.5 1.0
2.5 1.5
2.5 1.0
1.5 1.5
1.5
3.5
2.5
2.5
1.5
2.5 1.0
2.5 1.0 1.0
1.0 1.5
1.5 1.5
2.5 1.5 2.0 1.5 1.0
NH 4
+ 1.5 2.5 3.5 3.5 Ca
2+ 2.0 1.5
F I G U R E 1 9 .3 Geographic patterns of sulfate (SO42−), nitrate (NO3−), ammonium (NH4+), and calcium
(Ca2+) in wet deposition across the continental United States. Values are in kilograms per hectare per year
(kg ha−1 y−1). Source: Lovett (1994). Reprinted with permission by the Ecological Society of America.
and Mg2+ are easily leached from leaves (Tukey 1970), and their concentrations in throughfall can
be greater than that of the original wet deposition. This may be particularly important in tropical
forests in which species-specific differences in canopy leaching are related to the abundance of
epiphytes (Schlesinger and Marks 1977). In some cases, however, forest canopies directly assimi-
late water-soluble nutrients (i.e., N) from wet precipitation, thereby lowering their concentration
in throughfall (Olson et al. 1981). Stemflow is significant in that it returns a relatively concentrated
supply of nutrients directly to the base of the tree (Gersper and Holowaychuk 1971), where they
may accumulate and be taken up by the tree.
Determining the extent to which the “wash off” of dry material and leaching influence the
nutrient concentration of throughfall and stemflow can be difficult. However, it appears that 85%
of the SO42− in throughfall originates from the “wash off” of dry material deposited on leaves
(Lindberg and Garten 1988), making the dry deposition of nutrients an important process in forest
ecosystems. The high amount of leaf surface in forests makes them effective collectors of dry mate-
rials suspended in the atmosphere.
Dry deposition is a complex process involving atmospheric chemistry, wind velocity, and
canopy characteristics like leaf shape, orientation, spatial arrangement, and area (Lovett 1994).
Solid particles (< 5 μm diameter) enter terrestrial ecosystems through gravitational sedimentation,
the process by which the force of gravity pulls particles suspended in the atmosphere toward the
Earth’s surface. Some particles are small enough to be kept aloft by wind turbulence and enter
terrestrial ecosystems only when they impact the surface of vegetation. Even smaller dry particles
and gases directly enter the plant through stomata and lenticels.
The movement of atmospheric particles toward leaf surfaces is determined by the change in
particle concentration from the atmosphere to the leaf surface and the resistance to flow along that
path (Garland 1977). Atmospheric scientists estimate the input of particles to terrestrial ecosys-
tems by determining their deposition velocity:
Rate of particle dryfall mg cm 2 of leat s 1
Deposition velocity 3
.
Concentration in air mg cm
With knowledge of deposition velocity (cm sec−1), atmospheric concentration (mg cm−3), and leaf
area (m2 m−2), atmospheric scientists estimate the amount of nutrient-containing dry particles
entering a particular forest ecosystem. Total dry deposition (mg m−2) is simply the mathematical
product of atmospheric concentration, deposition velocity, and canopy leaf area.
Dry deposition of nutrients is somewhat more variable than wet deposition for several rea-
sons (Figure 19.3). The ratio of wet to dry deposition increases with distance from the source,
because readily dry-deposited materials are depleted from the air as it travels downwind (Ollinger
et al. 1993). Second, differences in canopy structure and leaf morphology among trees cause nutri-
ents in dry material to be captured with different efficiencies, depending on overstory composition.
Nonetheless, dry deposition can bring significant amounts of nutrients into forest ecosystems
relative to those contained in wet deposition.
Figure 19.4 summarizes wet, dry, and cloud depositions for a series of forest, prairie, and
agricultural ecosystems distributed across North America. Notice that the ratio of dry to wet depo-
sition is highly variable among these ecosystems, reflecting differences in the distance to nutrient
sources and canopy characteristics. The prairie ecosystem (AR) located downwind from Chicago,
Illinois, United States receives relatively greater inputs of S and N in dry deposition than would be
expected for an ecosystem in the Midwest (Figure 19.4) because it is relatively close to a major
source of anthropogenic NO3− and SO42− production.
Cloud deposition results when small, nonprecipitating water droplets containing dis-
solved nutrients come in contact with terrestrial surfaces, such as tree canopies. In contrast to
wet deposition, tree canopies exert a substantial influence on the amount of nutrients entering
forest ecosystems from cloud deposition. This process is a particularly important nutrient
input in ecosystems that lie in coastal and mountainous regions (WF and CD in Figure 19.4),
whereas in continental areas (e.g., Midwest), cloud deposition contributes relatively minor
amounts of nutrients to terrestrial ecosystems. For example, relatively large quantities of
SO42+, NO3−, and NH4+ enter spruce/fir ecosystems in the northeastern United States and red
spruce ecosystems at high elevation in the Appalachian Mountains through cloud deposition
(Figure 19.4). Over 50% of the atmospheric deposition of NH4+ in these ecosystems occurs
through cloud deposition.
BIOLOGICAL FIXATION OF NITROGEN

Biological inputs of N to forest ecosystems occur through symbiotic associations of soil bacteria
and tree roots, free-living bacteria in litter and mineral soil, as well as epiphytic lichens living on
the external surfaces of trees and decaying wood (Melillo 1981; Waughman et al. 1981; Boring
et al. 1988). Several genera of soil bacteria, actinobacteria (filamentous bacteria), and cyanobacte-
ria have the ability to reduce or “fix” atmospheric N2 into NH4+. Depending on the fixation process,
NH4+ can be directly used by the microbial cell or it can be transported from the microorganism to
the host plant as N-containing organic compounds. The process of N2 fixation is catalyzed by the
nitrogenase enzyme system, which requires substantial quantities of energy (i.e., ATP). The
following equation summarizes this process:
N 2 16 ATP 8 e 10 H Nitrogenase
2 NH 4 H 2 16 ADP Pi .
The high energy cost (i.e., ATP) associated with N2 fixation has important ecological implications
and gives rise to widely different rates of fixation depending on the fixation process and the micro-
organisms involved. Soil organic matter provides small amounts of energy for fixation by free-
living soil bacteria, whereas carbohydrates from photosynthesis can fuel rapid rates of N2 fixation
by symbiotic bacteria and actinobacteria inhabiting plant roots. As a result, quantities of N entering
terrestrial ecosystems from N2 fixation are highly variable and range from as little as 1 kg N ha−1 y−1
for free-living soil bacteria to 200 kg N ha−1 y−1 from symbiotic fixation in the roots of legumes,
alder, and other higher plants (Cole and Rapp 1981).
Rates of free-living N2 fixation by free-living organisms are low and equivalent to rates of
atmospheric deposition (Figures 19.2 and 19.3) in most terrestrial ecosystems. Free-living hetero-
trophic bacteria of the genera Azotobacter and Clostridium fix atmospheric N2 into forms that
plants can assimilate after their cells die and cellular constituents are decomposed by other soil
microorganisms. Because soil organic matter supplies the energy for free-living fixation, this pro-
cess is greatest in soils with relatively high organic matter contents (Granhall 1981). In the forests
of the Pacific Northwest and northeastern United States, free-living N2 fixation has been observed
in decaying logs (Roskoski 1980; Silvester et al. 1982), a relatively rich-energy substrate for hetero-
trophic bacteria. In most cases, the input of N to forest ecosystems from free-living N2 fixation is
relatively small, approximately 1–5 kg N ha−1 y−1 (Boring et al. 1988).
Free-living cyanobacteria (blue-green algae) are common soil microorganisms that can, in
some instances, be a source of N in some terrestrial ecosystems. Fixation by these photosynthetic
microorganisms is associated with high-light environments, and in most forests, fixation by cyano-
bacteria is limited by low light levels at the soil surface. However, algal crusts that form on the
surface of some desert soils can have exceptionally high rates of N2 fixation (Rychert et al. 1978).
Nitrogen-fixing cyanobacteria also enter into mutualistic associations with some fungi to form N2
fixing lichens that bring N into forest ecosystems. Nostoc, an N2-fixing cyanobacteria, is one of two
species of algae present in the lichen Lobaria oregana that inhabits the crowns of old-growth
Douglas-fir. Lobaria is estimated to fix approximately 8–10 kg N ha−1 y−1, an amount comparable to
the input of N via atmospheric deposition (see Figure 19.4).
Nitrogen-fixing actinobacteria and unicellular bacteria can infect the fine roots of certain higher
plants and induce the formation of root nodules, which are the location of symbiotic N2 fixation. These
small (< 5 mm), round, hollow structures are formed on individual roots and enclose large numbers of
bacterial or actinobacterial cells, the agents of N2 fixation. Nitrogen fixed by the microbial symbiont is
transferred to the plant, and carbohydrate supplied by the host in turn provides energy for the micro-
bial fixation of N2 in the root nodule. Because the cost of N2 fixation is subsidized by a supply of car-
bohydrate (i.e., energy) from the plant, symbiotic fixation can bring substantial quantities of N into
terrestrial ecosystems, far surpassing fixation by free-living bacteria and cyanobacteria.
Legumes are perhaps the most important plants with N2-fixing bacteria inhabiting their
roots. They occur as overstory species in both temperate and tropical forests. However, not all
legumes have the ability to fix N2 in association with Rhizobium, their bacterial symbiont. Black
locust, honey locust, acacias, and mesquite are trees of the legume family that occur in temperate
forests; there are hundreds of similar species in the tropics. Also, many herbs and shrubs of this
family inhabit the forest floor throughout much of the Earth, many becoming dominant following
forest harvesting or fire. Soil bacteria belonging to the Rhizobium genus exclusively infect the roots
of legumes and enter into a symbiotic N-fixing relationship.
(a) (b)
20
NO3-N Deposition Rate (kg•ha–1•yr–1)
40 16
S Deposition Rate (kg•ha–1•yr–1)
30 12
20 8
10 4
0 0
WF HF SC OR CW CD PN AR PW TH WF HF SC OR CW CD PN AR PW TH
(c)
10
NH4-N Deposition Rate (kg•ha–1•yr–1)
Cloud
8 Dry
Wet
6
0
WF HF SC OR CW CD PN AR PW TH
Annual
Code Site Location Vegetation Elevation Precipitation
(m) (cm)
WF Whiteface Mountain, New York Spruce/fir forest 1000 115
HF Huntington Forest, New York Beech/maple/birch forest 500 97
SC State College, Pennsylvania Corn/wheat/soybeans 393 121
OR Oak Ridge, Tennessee Loblolly pine forest 300 114
CW Coweeta, North Carolina White pine forest 725 144
CD Clingmans Dome, North Carolina Red spruce forest 1740 203
PN Panola, Georgia Meadow 212 130
AR Argonne, Illinois Meadow/forest 229 122
PW Pawnee site, Colorado Prairie 1641 28
TH Thompson site, Washington Douglas-fir forest 220 114
F I G U R E 19.4 Atmospheric deposition of S, NO3−-N, and NH4+-N in agricultural, prairie, and forest ecosystems in North America. The relative contribution of wet,
dry, and cloud deposition is depicted. Note that cloud deposition is an important process in coastal areas and at high elevations. Source: Lovett (1994). Reprinted with
permission of the Ecological Society of America.
In the southeastern United States, black locust readily establishes following a major distur-
bance, such as fire or clear-cut harvest. Fixation in the root nodules of this leguminous tree is a
substantial input of N to the forests of the southern Appalachians, particularly during the early to
intermediate stages of ecosystem development (Waide et al. 1988). Four years following clear-cut
harvest, rates of N2 fixation equaled 48 kg N ha−1 y−1, a substantial input of N relative to atmo-
spheric inputs (see Figures 19.2 and 19.3). Rates subsequently increased to 75 kg N ha−1 y−1 at
17 years following harvest, and after 38 years of ecosystem development rates declined to
33 kg N ha−1 y−1 (Waide et al. 1988). Compare these rates with mineral weathering and atmospheric
deposition, which are orders of magnitude slower.
Actinobacteria (filamentous bacteria) of the genus Frankia form a symbiotic relationship
with the roots of Alnus, Ceanothus, Casuarina, Elaeagnus, Comptonia, and Myrica. This symbiotic
relationship is particularly important in the Pacific Northwest, where N2 fixation by red alder and
Frankia can substantially increase the amount of N entering forest ecosystems. One example comes
from a comparison of 38-year-old red alder and Douglas-fir stands that initially established adjacent
to one another on the same soil type. Differences in N pools between them resulted directly from N2
fixation by the alder–Frankia symbiosis (Table 19.2). An additional 85 kg N ha−1 y−1 have accumu-
lated under red alder, an input of N far exceeding atmospheric deposition (see Figure 19.3). Much
of the N2 fixed by the red alder–Frankia resides in forest floor and mineral soil, where it is eventually
released during organic matter decomposition. The shrub Ceanothus velutinus also occurs in the
Pacific and Inland Northwest and enters into an N2-fixing symbiosis with Frankia that can con-
tribute as much as 100 kg N ha−1 y−1 following forest fire or harvest (Youngberg and Wollum 1976).
Greater N availability resulting from N2 fixation on N-poor soils can dramatically increase
rates of biomass accumulation and nutrient cycling during ecosystem development. In the Hawaiian
Islands, Myrica faya is an exotic N2-fixing tree that has greatly altered the N cycle of native forests
(Vitousek et al. 1987). This invasive, early-successional species establishes on volcanic ash flows
that were formerly colonized by Metrosideros polymorpha, a native, non-N2 fixing tree. In the
absence of M. faya, N accumulates in ash-flow soils at a very slow rate, constraining rates of eco-
system development and productivity (approximately 5 kg N ha−1 y−1 from free-living fixation and
atmospheric deposition). Nitrogen fixation by M. faya brings an additional 18 kg N ha−1 y−1 into the
ecosystem, increasing the annual rate of N accumulation by a factor of three. The N2 fixed by Myrica
eventually enters the soil via leaf and root litter, and during decomposition, relatively greater
amounts of N are released into soil solution where they are assimilated by plant roots.
Table 19.2 The accumulation of N in forest ecosystems dominated by nitrogen- fixing (red alder)
and non‑nitrogen-fixing (Douglas-fir) tree species. The overstory of each ecosystem is 38 years old
and differences in the amount of N contained in each ecosystem pool reflect the fixation of nitrogen by
red alder.
Nitrogen Annual accumulation

Ecosystem pool Douglas- fir Red alder
kg N ha−1 kg N ha−1 y−1
Overstory 320 590 7.1
Understory 10 100 2.4
Forest floor 180 880 18.4
Mineral soil 3270 5450 57.4
Total ecosystem 3780 7020 85.3
Source: Cole and Rapp (1981) / Cambridge University Press. Reprinted from Dynamic Properties of Forest Ecosystems,
© 1981 by Cambridge University Press. Reprinted with the permission of Cambridge University Press.
Nutrient Cycling within Forest Ecosystems 497
NUTRIENT CYCLING WITHIN FOREST ECOSYSTEMS

Nutrients entering forest ecosystems from
mineral weathering, atmospheric deposi- Atmospheric
tion, and biological fixation can enter soil Deposition
solution where they are absorbed by plant

roots. Within the plant, nutrients taken up
from soil solution participate in a wide
array of physiological processes, and in Translocation
some cases, growth-limiting nutrients are Aboveground
removed (i.e., translocated) prior to the Litter
Production
Fixation
shedding of some plant tissues. The pro- Gaseous

Loss
duction of plant litter above and below the
soil surface eventually returns nutrients to
forest floor and mineral soil, where they are
released from litter by soil microorganisms.
During the process of litter decomposition,
soil microorganisms incorporate organi- Organic Matter
cally bound nutrients (and C) into their Belowground Forest Floor and
Mineral Soil
Litter
biomass and release excess nutrients into Production
soil solution where they can again be taken Nutrient Uptake

up by plant roots. Thus, the rate at which

and Mineralization
Nutrients
nutrients flow within forest ecosystems is in
Soil Solution
controlled by the physiological activities of Mineral
Weathering
plants and soil microorganisms, and their Leaching
requirement for growth-limiting nutrients.

In the sections that follow, we trace the flow of nutrients within forest ecosystems. The
cycling of nutrients within ecosystems occurs through: (i) root nutrient absorption via the processes
of uptake and assimilation, (ii) nutrient allocation to biomass construction and maintenance,
(iii) nutrient translocation from senescent tissue, (iv) the return of nutrients in aboveground and
belowground litter, and (v) the microbially mediated release of inorganic nutrients into soils solu-
tion (i.e., mineralization) during organic matter decomposition. These processes are highlighted in
the reduction of Figure 19.1, which is located above. Central to this discussion is the ability of
plants to forage for nutrients in soil solution, a process mediated by physiological and morpholog-
ical adaptations of plant roots. Also of importance is the growth dynamics of soil microorganisms
during litter decomposition, because they control the release of organically bound nutrients from
plant litter and soil organic matter. As you will learn later in this chapter, soil microorganisms are
the “gate keepers” of N availability within forest ecosystems.
NUTRIENT TRANSPORT TO ROOTS

Nutrients must diffuse to the root surface from soil solution before they can be taken up and
incorporated into biologically active compounds. The uptake of nutrients by plant roots is initially
constrained by rates of organic matter decomposition, mineral solubility, cation/anion exchange
reactions, as well as diffusion through soil solution. Although these processes control nutrient
availability in soil, the fact that nutrients are present in soil solution or on exchange sites does not
ensure uptake by plant roots. In the paragraphs that follow, we discuss the processes controlling
the flow of nutrients to the root surface, uptake from soil solution, and their incorporation into
biologically active compounds within the plant.
Nutrients in soil solution move toward root surfaces in response to two processes: mass flow
and diffusion. Mass flow is a passive process in which ions move with the flow of water for tran-
spiration; they are physically carried along from soil solution to the root surface. Diffusion occurs
when ions move from a region of high concentration to one of low concentration, for example,
from soil solution to the root surface. In most soils, the concentration of nutrients in solution is low
enough that mass flow does not meet the demand of actively growing plants. Rapid uptake by
plant roots also causes some nutrients to be virtually absent near their surface, thereby forming
zones of depletion around individual roots. Consequently, the supply of nutrients to roots is con-
strained by the rate at which ions diffuse toward root surfaces from areas of higher concentration
in soil solution (Nye 1977). Nutrients such as PO43−, K+, and NH4+ diffuse slowly in water and rapid
uptake depletes their concentration near root surfaces so that concentrations increase as the dis-
tance from the root surface increases. Although NO3− is very mobile in soil solution and diffuses
rapidly toward roots, a high demand by most plants maintains low NO3− concentrations throughout
soil solution. Because of its high rate of diffusion, mass flow may meet the NO3− demand of some
forest trees.
The size of the zone of depletion surrounding an individual root is directly proportional to
its radius and the rate at which a particular ion diffuses in water. The radius of the zone of deple-
tion (Rdepletion in cm) for any ion is
Rdepletion a 2 Dt ,
where a is the root radius (cm), D is the diffusion coefficient for an ion (cm2 sec−1), and t is time in
seconds (Nye and Tinker 1977). One also can compare root size (cm3) to the zone of depletion
(cm3) it creates using the following expression:
2
a 2 Dt /a 2
Given this relationship, smaller roots exploit a greater volume of soil than larger roots (Fitter 1987).
Recall that large roots have higher construction costs than small roots, indicating that plants
“invest” relatively greater amounts of photosynthate into the production of large-diameter roots
(Chapter 18). By forming small-diameter roots, plants increase the volume of soil exploited for
water and nutrients, while reducing the amount of photosynthate allocated to root production. It
is likely that this is why roots generally less than 1 mm in diameter are the nutrient- and
water-absorbing organs of forest trees.
Once nutrients arrive at the root surface, they must be taken up from soil solution and incor-
porated into biologically active compounds. In some soils, the supply of some nutrients is exces-
sive, and they are actively excluded from uptake. It is not uncommon to observe accumulations of
CaCO3 adjacent to the roots of desert shrubs growing on highly calcareous soils (Klappa 1980). In
most instances, however, nutrient supply is low (especially for N and P), and plants actively forage
for nutrients in soil. Nutrient foraging occurs through the physiological mechanisms of nutrient
uptake and through morphological changes in root architecture that increase the absorptive area
of plant root systems.
NUTRIENT UPTAKE AND ASSIMILATION BY ROOTS

Nutrients are incorporated into biologically active compounds (i.e., amino acids, nucleic acids) in
a two-step process consisting of uptake and assimilation. Uptake is the physiological process by
which nutrients in soil solution are actively transported across cell membranes into roots.
Assimilation occurs when inorganic nutrients transported into plant cells are biochemically
incorporated into organic compounds such as amino acids, nucleic acids, lipids, or other biologi-
cally active compounds.
High rates of enzymatic uptake are one physiological means by which plants can maximize
acquisition of growth-limiting nutrients. For example, the uptake of nutrients from soil solution
is mediated by enzymes associated with the membranes of very-fine roots (< 0.5 mm diameter).
The rate of enzymatic activity mediating nutrient uptake increases with increasing nutrient con-
centrations in soil solution until the capacity of the enzyme system is saturated. Uptake capacity
for nutrients varies widely among tree species, and this process can be highly responsive to soil
temperature (Figure 19.5). In a comparison of trees of the taiga, rapidly growing species such as
6 Balsam Popular 60 Balsam Poplar 60 Trembling Aspen

5 3 Trembling Aspen 50 50
4
Ammonium absorption (µmol·h–1·g–1)

Phosphate absorption (µmol·h–1·g–1)
40 40
3 2
2 30 30
1 1 20 20
0
0 10 20 10 10
3 0
Paper Birch 0 10 20 0 0
0 2000 4000 0 2000 4000
40 Paper Birch 40 Green Alder
2 2
Green Alder 30 30
20 20
1 1
10 10
0 0 0 0
0 10 20 0 10 20 0 2000 4000 0 2000 4000
Phosphate concentration (µmol·L–1) Ammonium concentration (µmol·L–1)
10° 4 Balsam Poplar 4 Trembling Aspen

20°
Nitrate absorption (µmol·h–1·g–1)
3 3
2 2
1 1
0 0
0 2000 4000 0 2000 4000
4 Paper Birch 4 Green Alder
3 3
2 2
1 1
0 0
0 600 1200
0 2000 4000
Nitrate concentration (µmol·L–1)
F I G U R E 1 9 .5 Nutrient uptake by four taiga trees in response to nutrient concentration and soil tempera-
ture. Uptake rates of NH4+, NO3−, and PO43− were measured on excised roots of seedling growing under
laboratory conditions. Source: Chapin et al. (1986). Reprinted from Oecologia, © Springer-Verlag Berlin,
Heidelberg 1986. Reprinted with permission of Springer-Verlag, New York, Inc.
balsam poplar and trembling aspen have relatively high uptake capacities for NH4+, NO3−, and
PO43−, compared to the slower-growing paper birch and green alder (Chapin et al. 1986). Balsam
poplar and trembling aspen also occur on relatively warm soils in the taiga and exhibit rapid
increases in nutrient uptake with rising soil temperature (Figure 19.5).
Although the productivity of many temperate forests is constrained by soil N availability,
we understand relatively little regarding the ecological importance of NH4+ versus NO3− uptake
by overstory trees. Ammonium (NH4+) often is the dominant form of N in forest soils, and some
trees exhibit a physiological preference for NH4+ over NO3−. This relationship can be observed in
Figure 19.5 for taiga trees, and it also has been documented for several overstory trees in
temperate forests (Figure 19.6). In the fine roots of Douglas-fir and sugar maple, rates of NH4+
uptake far exceed rates of NO3− uptake, suggesting that these widely distributed trees have a
(a)
35 35
30 30
µmol 15NH 4·g–1·h–1
25 25
µmol 15NO–3·g–1·h–1
20 20
15NH+
+
15 4 15
15NO–
3 10
10
5 5
0 0
0 200 400 600 800 1000
µM 15NH+4 or 15NO–3
(b)
10 10
8 8
µmol 15NH+4·g–1·h–1
µmol 15NO–3·g–1·h–1
6 6
4 4
2 2
0
0 200 400 600 800 1000
µM 15NH+4 or 15NO–
3
F I G U R E 19 .6Ammonium and NO3− uptake in the fine roots of (a) sugar maple and (b) Douglas-fir. At any
given concentration, rates of NH4+ uptake exceed those of NO3−, suggesting that these overstory species have
a physiological preference for NH4+ over NO3−. Source: Kamminga-Van Wijk and Prins (1993) / Springer
Nature. After Kamminga-Van Wijk and Prins 1993 and Rothstein et al. 1996. Panel (a) reprinted from
Oecologia, © Springer-Verlag Berlin, Heidelberg 1996. Reprinted with permission of Springer-Verlag,
New York, Inc. Panel (b) reprinted from Plant and Soil, 1993, Vol. 151, “The kinetics of NH4+ and NO3−
uptake by Douglas-fir from simple N-solutions and from solutions containing both NH4+ and NO3−” by
C. Kamminga-Van Wijk and H. Prins, pp. 91–96, © 1993 by Kluwer Academic Publishers. Reprinted with
kind permission from Kluwer Academic Publishers.
physiological preference for NH4+, especially for sugar maple which has very low maximum
rates of NO3− uptake.
Western hemlock and jack pine commonly occur on acidic soils with extremely low NO3−
availability. These overstory species also have a limited capacity for NO3− uptake and appear to
satisfy their demand for N through rapid NH4+ uptake (Lavoie et al. 1992; Knoepp et al. 1993). In
jack pine, NO3− uptake enzymes are present at very low levels in fine roots (Lavoie et al. 1992), and
when seedlings are supplied with NH4+ or NO3− as a sole source of N, those supplied with NH4+
attain twice the biomass of seedling grown solely on NO3−. (Lavoie et al. 1992). This observation
suggests that jack pine has a low physiological capacity for NO3− uptake, even when concentrations
in soil solution are relatively high. The inability of this species to use NO3− suggests that it would
be a poor competitor in NO3−-rich soils, relative to plants that have rapid rates of NO3− uptake and
assimilation.
There are several reasons why NH4+ uptake is more rapid than NO3− uptake in many forest
trees. First, NH4+ is often the dominant form of N in soil solution and small amounts can strongly
inhibit enzymes involved with NO3− uptake and assimilation. Second, plants must reduce NO3− to
NH4+ before it can be assimilated into biologically active compounds, a process that requires sub-
stantial amounts of energy (ATP). When this reaction occurs in leaves, it is subsidized by energy
from the light-harvesting reactions of photosynthesis. However, many woody plants have the
ability to reduce NO3− in roots, which requires the transport of reducing compounds from the leaf.
Most often, plants with rapid rates of NO3− uptake and assimilation occur in high-light habitats in
which excess energy from photosynthesis can be allocated to NO3− reduction.
A second physiological mechanism for maximizing the uptake of limiting nutrients is via
the production of root enzymes that release nutrients from soil organic matter. This is particu-
larly true of PO43−, which diffuses slowly in soil solution. Phosphatases are enzymes produced
by plants and microorganisms that act upon soil organic matter to yield PO43− from organically
bound P. The release of these enzymes into soil appears to be an important mechanism for
increasing PO43− supply to roots in PO43−-poor soils. Phosphatase activity associated with the
roots of arctic tundra plants can supply up to 65% of their annual PO43− demand (Kroehler and
Linkins 1988).
ROOT ARCHITECTURE, MYCORRHIZAE, AND NUTRIENT ACQUISITION

Root Architecture In addition to physiological mechanisms, plants also forage for nutrients by
altering root architecture and growth. Ecologists have become increasingly interested in the struc-
ture and function of plant root systems, because plants appear to alter root branching patterns, or
architecture, in response to soil nutrient availability (Fitter 1987; Fitter and Stickland 1991). Several
analyses suggest that fine-root systems consisting of only a main axis and primarily laterals
(herringbone architecture; Figure 19.7) are the most efficient at nutrient acquisition (Fitter and
Stickland 1991; Berntson and Woodward 1992). The herringbone arrangement maximizes the
volume of soil exploited by roots, while minimizing the overlap between the zone of depletion cre-
ated by adjacent fine roots; however, the construction cost of this type of root system is high. It
requires the production of larger diameter roots (e.g., higher construction), each of which trans-
ports a greater proportion of the total flow of water and nutrients into the aboveground portion of
the plant (Figure 19.7). By contrast, the mean root diameter of a dichotomously branched root
system is much smaller than that of a herringbone architecture, resulting in relatively lower
construction costs. This spatial arrangement of roots allows the depletion zones of adjacent roots
to overlap, and thus it is less efficient at foraging for nutrients than the herringbone architecture.
Differences in the cost of constructing and maintaining these root systems, and the efficiency
with which they explore soil for nutrients, suggest that root-system architecture should vary in
response to soil nutrient availability. Plants growing in nutrient-poor soils should allocate relatively
more photosynthate to a highly efficient root system for nutrient uptake (i.e., herringbone
F I G U R E 19. 7 Examples of plant root

systems with different degrees of branching.
Source: Fitter (1987) / John Wiley & Sons.
Herringbone Intermediate Dichotomous
branching pattern), whereas those growing in nutrient-rich soil should invest relatively less C into
a dichotomously branched root architecture. These predictions hold for different herbaceous
plants (dicots) from nutrient-rich and nutrient-poor soils, and also hold for the same species of
herbaceous plants grown in soils of different fertility (Fitter and Stickland 1991; Berntson and
Woodward 1992; Taub and Goldberg 1996). The fine-root system of forest trees appears to respond
in a similar manner, but unfortunately, few studies have focused on these organisms. The fine roots
of pin cherry rapidly proliferate within experimental patches (125 cm3) of N-rich soil by increasing
rates of production and the degree of branching, that is, their fine-root system becomes more
dichotomous (Pregitzer et al. 1992). Although there are clear differences in the construction cost
of fine-root systems, it is also important to consider the maintenance cost of these structures,
because maintenance costs could exceed the cost of their construction. This is especially true for
fine roots with high N concentrations (enzymes for ion uptake), but, unfortunately, we presently
do not have sufficient observations to understand the relationships among construction costs,
maintenance costs, and fine-root longevity for a large number contrasting plant species.
Importantly, we still are learning how form and function of plant root systems are related, as
well as how this might be coordinated with aboveground plant form and function (i.e., canopy
geometry, maximum rates of net photosynthesis). Recent global analyses of plant root systems
indicate that plants effective in nutrient-rich soil have a lower specific root length (i.e., smaller
diameter per unit length) and higher N concentration than plants which are good competitors for
nutrients in nutrient-poor soils (Weigelt et al. 2021). These root traits correspond to aboveground
form and function, wherein leaves of plants in nutrient-rich soil are thinner, have high N concen-
trations, and more rapid rates of maximum net photosynthesis than those from N-poor soils. Like
leaves, fine roots with high N concentrations likely have high maintenance costs to fuel the rapid
uptake of nutrients from soil solution. Although this is an intriguing parallel, we still lack a
comprehensive understanding of how fine root form (architecture) and function (ion uptake) are
coordinated among forest trees or any other plant species. This remains an important gap in our
understanding of the belowground physiology and natural history of plants as well as how they are
coordinated, or not, with the aboveground plant attributes.
Mycorrhizae Forest trees increase their ability to forage for nutrients in soil by entering into a
symbiotic relationship with soil fungi. This form of mutualism, termed a mycorrhiza or literally
a root-fungus, increases the volume of soil exploited by an individual plant. Mycorrhizal fungi are
important for the successful establishment of some tree seedlings. Mycorrhizae also facilitate an
important nutrient cycling process—plant nutrient uptake and, to some extent, the decay of soil
organic matter. In this section, we further explore the extent to which mycorrhizal fungi increase
the ability of plant roots to forage for nutrients within soil.
Root hairs are not common in nature, because tree roots are almost universally colonized by
mycorrhizal fungi. In pines, spruces, firs (and all other genera of the Pinaceae), birches, beeches,
F I G U R E 1 9 . 8 Shown are the manner in Ectomycorrhiza Arbuscular mycorrhiza

which the hyphae of ectomycorrhizal fungi
(ECM) and arbuscular mycorrhizae (AM)
fungi penetrate plant roots. A notable
difference between is the fact that ECM
hyphae grow throughout root cell wall,
whereas AM hyphae oppress plant cell Hyphopodium
membranes and form arbuscules, which are Mycelium
the site of photosynthate and nutri-
ent exchange.
Mantle
Arbuscules
Hartig
net Spore
oak, basswoods, and willows, ectomycorrhizal fungi (ECM; ecto meaning outside) form a sheath or
mantle surrounding fine roots giving them a characteristic swollen appearance (Figure 19.8).
Fungal hyphae penetrate the space between the outer cortical cells, but do not enter individual root
cells. Over 2400 species of fungi are known to form ECM on North American trees (Marx and
Beattie 1977). A less conspicuous group, arbuscular mycorrhizae (AM), forms no sheath, but
individual hyphae grow within and between epidermal and cortical cells of roots. Many plant
species form AM, including cultivated crops, grasses, and most tropical tree species (Janos 1987).
Temperate species forming AM include redwood and many hardwoods: maples, ashes, tuliptree,
sweet gum, sycamore, black walnut, and black cherry (Marx and Beattie 1977; Harley and
Smith 1983).
Ectomycorrhizae and AM are effective accumulators of nutrients (and water) because they
increase the absorbing surface area of tree roots, thereby allowing the plant to forage for nutrients
in a greater volume of soil. Recall that small fine roots (approximately 0.1 mm in diameter) more
effectively exploit soil for nutrients than larger roots. Fungal hyphae (10 μm in diameter) are much
smaller than fine roots, which further increases the ability of plants to exploit soil for nutrients.
Additionally, the hyphae of mycorrhizae can extend great distances into the forest floor and min-
eral soil. For example, 1 mm3 of soil can contain 4 m of ECM hyphae that transport water and nutri-
ents back to the host plant. In a 450-year-old Douglas-fir ecosystem in Oregon, 5000 kg ha−1 of ECM
occur in the surface soil (0–10 cm), composing over 11% of the total root biomass (Trappe and
Fogel 1977). Moreover, a single Douglas-fir–Cenococcum ectomycorrhizae can form 200–
2000 individual hyphae, some of which extend over 2 m into soil and form more than 120 lateral
branches or fusions with other hyphae.
As compared to non-mycorrhizal roots, those infected with mycorrhizal fungi tend to be
more metabolically active, exhibiting rapid rates of respiration and ion uptake. Mycorrhizae are
widely known to increase the PO43− uptake of forest trees growing in P-poor soils; however,
mycorrhizae also facilitate the uptake of other plant nutrients such as N (Bowen and Smith 1981).
The mycorrhizae formed by the fungus Paxillus involutus are able to take up substantial amounts
of NH4+ from soil solution, assimilate the NH4+ into amino acids, and transfer the newly formed
amino acids to European beech (Finlay et al. 1989). Paxillus involutus also is able to assimilate
NO3−, but it does so at a substantially lower rate. In addition to the direct uptake of ions from soil
solution, mycorrhizae have the ability to produce extracellular enzymes which facilitate organic
matter decomposition (Antibus et al. 1981; Dodd et al. 1987) and also produce organic acids which
release plant nutrients from soil minerals (Bolan et al. 1984).
Ectomycorrhizae and AM fungi differ from one another in many important ways that have
implications for plant nutrient supply. Foremost, AM fungi are evolutionary ancient and are
thought to have mediated the transition of aquatic plants to those on land. They represent an evo-
lutionary narrow group of fungi (i.e., monophyletic), which assist plants with the acquisition of
both water and phosphorus. These organisms produce phosphatase enzymes which release PO43−
from organic matter, as well as organic acids that solubilize PO43− from soil minerals. By contrast,
ECM have evolved more recently and have arisen from multiple saprotrophic fungal ancestors
(i.e., polypheletic; approximately 80 independent times). While some ECM have retained genes
that encode enzymes mediating organic matter decay, others have lost them during their evolution
into root symbionts (Pellitier and Zak 2018). Isotopic evidence suggests that ECM who have
retained genes for saprotrophic function can modify soil organic matter and provide the N
contained within it to a host plant (Pellitier et al. 2021). This appears particularly important where
soil N supply is slow and large quantities of N are locked with soil organic matter. It is interesting
to note that a large number of boreal trees associate with ECM whose genomes retain genes with
decay function, likely a response to slow rates of organic matter decay in these cold ecosystems.
Because mycorrhizal fungi are heterotrophic microorganisms, they require a supply of
carbohydrate for growth, for uptake of nutrients from soil, and for translocation of nutrients to the
host plant. The cost of forming mycorrhizae is not inconsequential. Carbohydrate formed by the
plant that could otherwise be used for growth, maintenance, or other physiological functions must
be transferred to the mycorrhizal fungi in return for a greater nutrient supply. In a Pacific silver fir
ecosystem, ECM compose only 1% of total ecosystem biomass, but their growth and maintenance
consume over 15% of net primary production (Vogt et al. 1983). Because of the high cost of forming
mycorrhizae, plants growing in nutrient-poor soil tend to form more mycorrhizae than those
growing in nutrient rich-soils, because they are more dependent on them for nutrient acquisition.
PLANT LITTER AND THE RETURN OF NUTRIENTS TO FOREST

FLOOR AND SOIL
Plant litter production above and below the soil surface directly controls the amount of nutrients
returned to the forest floor and mineral soil, and therefore constitutes important processes controlling
the cycling of nutrients within forest ecosystems. Nutrients absorbed by roots and mycorrhizae are
used for a wide variety of physiological functions, including the growth of new tissue, the mainte-
nance of existing tissue, storage, and the production of defense compounds—fates similar to that of
photosynthetically fixed CO2 (Chapter 18). Prior to abscission, both coniferous and deciduous trees
withdraw nutrients from leaf tissue and transport them into the nearby branches for storage. These
materials are removed from storage when growth commences in spring and are used to form new
leaves and construct a canopy. However, it is not clear whether nutrients are withdrawn from fine
roots prior to their death. Consequently, both the production of litter and the concentration of nutri-
ents contained within it control the rate at which nutrients absorbed by plants are returned to the soil
from which they came. In the paragraphs that follow, we explore geographic patterns of aboveground
and belowground plant litter production and the extent to which plants withdraw nutrients prior to
leaf abscission, a process controlling the nutrient concentration of leaf litter.
Leaf and Root Litter Production On a global basis, leaves account for approximately 60–
75% of total aboveground litterfall in forest ecosystems; the remaining proportion is composed of
woody material (approximately 30%) and reproductive structures (1–20%). In Chapter 18, we found
that global patterns of aboveground net primary production (ANPP) were strongly related to
climate, particularly temperature and actual evapotranspiration. Leaf litter production in forest
ecosystems also is strongly related to global patterns of climate, because of the aforementioned
relationship. In Figure 19.9, leaf litter production is low at high latitudes where short growing sea-
sons limit plant growth; it increases toward the equator where plant growth can occur throughout
the entire year. Also notice that substantial variation occurs at any particular latitude (Figure 19.9).
This regional-scale variability in leaf litter production undoubtedly results from the modification
of climate by physiography (i.e., slope and aspect), differences in soil water and nutrient availabil-
ity, or disturbance.
Recall that leaf biomass (or area) places the upper limit on the amount of photosyntheti-
cally fixed CO2 available for NPP (Chapter 18). As a result, leaf biomass and ANPP are positively
related across a wide array of terrestrial ecosystems, including deserts, grasslands, and forests
(Webb et al. 1983). Because leaves are temporary plant tissues, even in evergreen coniferous for-
ests, one would expect that patterns of leaf litterfall should be related to ANPP. The solid line in
Figure 19.10 indicates a 1:1 relationship between ANPP and leaf litter production, wherein all
ANPP is allocated to leaf production. At low values of ANPP (i.e., < 7.5 Mg ha−1 y−1), note that leaf
litter production represents a large proportion of ANPP. At higher rates of ANPP, leaf litter con-
stitutes a smaller proportion, reflecting a greater allocation of photosynthate to other plant tissues
and functions.
Although we have a clear understanding of global patterns of leaf litterfall, our knowledge
of the production of plant litter below the soil surface is incomplete. It is clear, however, that the
death of fine roots represents a significant proportion of NPP in many forest ecosystems, often
exceeding leaf litter production. Notice that fine-root litter represents 40–330% of leaf litter in the
temperate coniferous and deciduous forests summarized in Table 19.3. In the Pacific Northwest,
the production and death of roots can account for 59–67% of NPP in Pacific silver fir ecosystems,
whereas leaf litter production ranges from 6 to 8% of NPP (Vogt et al. 1983). Fine roots often have
nutrient contents (N and P) greater than or equivalent to leaf litter, suggesting that fine-root litter
represents a substantial transfer of nutrients from plant tissue to forest floor and mineral soil.
F I G U R E 1 9 .9 Global patterns of leaf litter 18

production in forest ecosystems. Source:
16
Total litter production (Mg·ha–1·y–1)
Modified from Bray and Gorham (1964) and

O’Neill and DeAngelis 1981. Modified from 14
Advances in Ecological Research, ©, 1964 by
Academic Press, Ltd. London. Reprinted 12
with permission of Academic Press, Ltd. 10
London and from Dynamic Properties of
Forest Ecosystems, © Cambridge University 8
Press 1981. Reprinted with permission of
6
Cambridge University Press.
4
0
0 10 20 30 40 50 60 70
Latitude (°N)
20 F I G U R E 19. 10 The relationship between

aboveground net primary production (ANPP)
and aboveground litter production in
temperate and boreal forest ecosystems. The
15 solid line indicates a 1:1 relationship
Total litterfall (Mg·ha–1·y–1)
between ANPP and the production of

aboveground litter. At low levels of ANPP,
virtually all production is allocated to leaves,
10 and the departure from the 1:1 line at high
levels of ANPP indicates a greater allocation
of production to non-photosynthetic tissues.
Source: After O’Neill and DeAngelis (1981) /
5 Cambridge University Press. Reprinted from
Dynamic Properties of Forest Ecosystems,
© Cambridge University Press 1981.
Reprinted with permission of Cambridge
University Press.
0
0 5 10 15 20
Aboveground net primary production (Mg·ha–1·y–1)
Table 19.3 Leaf litter and root litter production in forest ecosystems from different geographic locations
in temperate North America. The ratios of fine-root litter to leaf litter indicate that the production of fine-
root litter in forest ecosystems ranges from 40 to 330% of leaf litter.
Litter
Ecosystem Leaf Fine-root Root:leaf
Mg ha−1 y−1
Temperate coniferous
Pacific silver fir 1.5–2.2 4.4–7.4 2.9–3.3
Douglas-fir 2.8–3.0 2.4–2.7 0.9
Scots pine 2.1 2.0 1.0
White pine 2.9 2.6 0.9
Mixed pine 3.1 2.6 0.8
White spruce 2.7 1.6 0.6
Red pine 2.5–5.3 1.3–2.0 0.4–0.5
Temperate deciduous
Black oak 4.1 5.9 1.4
Northern red oak 4.2 5.2 1.2
White oak 3.6 4.1 1.1
Sugar maple 2.9 4.0 1.4
Paper birch 2.8 3.2 1.1
Mixed hardwoods 4.4 2.7 0.6
American beech 3.2 1.3 0.4
Source: Vogt et al. (1986) / with permission of Elsevier and Nadelhoffer et al. 1985. Modified from Advances in Ecological
Research, © 1986 by Academic Press, Inc. Reprinted with permission of Academic Press, Inc. and from Ecology by the permission
of the Ecological Society of America.
Recall that the return of nutrients to the forest floor and mineral soil is controlled by the pro-
duction of plant litter and its nutrient concentration. From the previous paragraphs, it should be
clear that forest ecosystems dramatically differ in the production of litter, both above and below the
soil surface. In the following paragraphs, we discuss the withdrawal, or retranslocation, of nutri-
ents prior to the shedding of leaf litter. This process directly influences the concentration (i.e., per-
cent) of nutrients in litter, which together with litter production controls the total return of
nutrients in plant litterfall.
Nutrient Retranslocation The biochemical CO2-fixing machinery of leaves requires substan-

tial quantities of N and P to form photosynthetic proteins, as do the nutrient uptake and assimila-
tion mechanisms of fine roots. As a result, green leaves and live fine roots contain relatively high
concentrations of nutrients compared to other plant tissues, such as stems and branches. Because
low quantities of N and P in soil often limit plant growth, plants have evolved mechanisms to con-
serve growth-limiting nutrients within their tissues. As noted earlier in the text, in both coniferous
and deciduous trees substantial quantities of N and P (and other nutrients) are retranslocated from
leaves prior to abscission and are mobilized into storage. Much of the N and P removed from leaves
resides within adjacent branches, stem, and large structural roots, where they can be remobilized
to meet the demand of newly forming leaves when growth resumes. However, the extent to which
forest trees retranslocate nutrients from senescent leaves is highly variable (Killingbeck 1996); it
changes from year to year within an individual species and changes among species from differ-
ent habitats.
The retranslocation of many nutrients reflects their availability in soil and their demand
within the plant for subsequent growth. In Brazilian rainforest trees, 17–73% of N and 41–83% of P
are translocated from leaves prior to abscission. Although retranslocation efficiency (i.e., percent
removed) varies among rainforest species, they all reduce P to similar concentrations in abscised
leaves (i.e., 0.04–0.06% P; Scott et al. 1992). The retranslocation of N is not as complete and
relatively large amounts of N (>1.0%) remain in abscised leaves (Scott et al. 1992). Differences in
the amount of P and N remaining in abscised leaves of rainforest trees reflect the fact that P availability
in soil, not N, limits the productivity of wet tropical forests (Vitousek 1984).
The extent to which forest trees retranslocate nutrients from senescing leaves directly influences
the nutrient-use efficiency of litter, which is defined by the following expression:
Dry weight of leaf litterfall kg ha 1y 1
Nutrient use-efficiency .
Nutrient content of leaf litterfall kg ha 1y 1
Plants that retranslocate large quantities of nutrients prior to leaf abscission have high nutrient-use
efficiencies and return relatively few nutrients (per unit biomass) to the forest floor in litterfall.
On a global basis, there is a remarkable relationship between the N- and P-use efficiency
and nutrient availability in tropical, temperate, and boreal forests. In Figure 19.11, the amount
of nutrients returned in litterfall (x-axis) is used as a surrogate for nutrient availability—large
amounts of nutrients residing in litterfall occurs where their availability in soil is high. Notice
that N-use efficiency is inversely related to the amount of N contained in litterfall, with the
exception of tropical forests (Figure 19.9a). The inverse relationship between N-use efficiency
and litterfall N in temperate coniferous and deciduous forests indicates that N-use efficiency is
high in ecosystems in which relatively small quantities of N are annually cycled (i.e., those with
low litterfall N contents).
In tropical ecosystems, litterfall N varies widely (10–180 kg N ha−1 y−1), but N-use efficiency
is relatively constant (approximately 80; Figure 19.11a). The fact that N-use efficiency changes
little, whereas the amount of N annually cycled to the forest floor varies by a factor of three,
(a)
368 Tropical
Coniferous
Dry weight:N (ratio of litterfall)
280 Deciduous
N-Use Efficiency
240 Mediterranean Shrublands

200 N-Fixing Species
160
120
80 223
40 224
0
0 20 40 60 80 100 120 140 180 200
N in litterfall (kg· ha–1· y–1)
(b) 5076 4548
4000
3600
Dry weight:P (ratio of litterfall)
3200
2800
P-Use Efficiency
2400
2000
1600
1200
800 12
14
400
0
0 1 2 3 4 5 6 7 8 9 10
P in litterfall (kg· ha–1 · y–1)
F I G U R E 1 9 .1 1 The relationship between plant nutrient-use efficiency and the amount of nutrients
returned in leaf litterfall for tropical, coniferous, temperate deciduous, and Mediterranean shrublands.
Forests dominated by N-fixing plants are denoted with a diamond. Plant nutrient-use efficiency is estimated
as the ratio of litterfall mass to its (a) N or (b) P content. The amount of N and P returned in litterfall (kg N or
P ha−1 y−1) is a relative index of soil N or P availability. Source: After Vitousek (1982) / John Wiley & Sons.
Reprinted with permission from American Naturalist, © 1982 by the University of Chicago. Reprinted with
permission of the Chicago University Press.
indicates that N availability does not influence the N-use efficiency of tropical forest trees. In con-
trast to this pattern, P-use efficiency and the amount of P returned in litterfall are inversely related
in tropical forests (Figure 19.9b). This pattern suggests that P translocation from senescent leaves
is high (i.e., high P-use efficiency) where P availability is relatively low within the ecosystem. Also
notice that there is a weak relationship between P-use efficiency and litterfall P in other forest eco-
systems. These patterns are consistent with the observation that N generally limits the productivity
of boreal and temperate forests, whereas a limited supply of P constrains productivity in wet
tropical forests.
Taken together, these observations suggest that limited supplies of N or P result in greater
translocation prior to leaf abscission, thereby increasing N- or P-nutrient-use efficiency of leaf
litter. This is likely an evolved mechanism that permits plants growing in low N or P soils to con-
serve these growth-limiting nutrients. Forest trees that conserve nutrients through a high
nutrient-use efficiency should have a competitive advantage on nutrient-poor sites, and it is not
uncommon to observe species replacement along gradients of nutrient availability (Pastor
et al. 1984; Zak and Pregitzer 1990). Individual species also are able to alter nutrient-use efficiency
in response to nutrient supply, which may allow them to persist in soils of markedly different fer-
tility. Northern red oak, for example, occurs on soils with a wide range of N availability in the
northern Lake States; N-use efficiency in this species drops as soil N availability increases; similar
responses have been observed in sugar maple, basswood, and American beech (Zak et al. 1986).
Although fine-root mortality represents a significant input of litter to forest floor and mineral soil,
we do not yet understand the extent to which soil nutrient availability influences the nutrient-use
efficiency of fine roots. While it is relatively easy to collect leaves prior to and following abscission,
fine roots present a unique challenge to gain this same insight—they remain buried in soil and are
difficult to observe and analyze for nutrient content.
The combined influence of plant litter production and nutrient-use efficiency on the
amount of N returned to forest floor and mineral soil is summarized in Table 19.4. Notice that N
inputs from fine-root litter are equivalent to or greater than the amount of N contained in leaf
litter, regardless of whether forests occur in tropical, temperate, and boreal regions (i.e., root:leaf
N ratio >1). The combined amounts of N entering forest floor and mineral soil are generally
greatest in tropical forests, in which leaf decay is rapid. Recall rates of N input from atmospheric
deposition (see Figure 19.3) and free-living fixation (approximately 1–5 kg N ha−1 y−1) and realize
that leaves and fine roots annually contribute 10 times more N to forest floor and mineral
soil. Clearly, aboveground and belowground litter production and its nutrient-use efficiency
represent an important process influencing the internal cycling of nutrients within terrestrial
ecosystems.
NUTRIENTS IN THE FOREST FLOOR

Large amounts of organic matter and nutrients can accumulate on the soil surface in boreal and
cool temperate forests, emphasizing the importance of the forest floor in controlling the flow of
nutrients within these ecosystems (Table 19.5). Nevertheless, forest floor organic matter and
nutrient content do not vary in a predictable manner among tropical, subtropical, and warm-
temperate forests. In fact, the quantity of organic matter and nutrients in the forest floor of dry and
montane tropical forests differ greatly from wet tropical forests and are more similar to that in
temperate forests (Anderson and Swift 1983). Nevertheless, several important generalizations can
be drawn from the information in Table 19.5. First, the accumulation of organic matter and nutri-
ents in the forest floor of cold-temperate and boreal evergreen forests is much greater than that in
warm-temperate, subtropical, or tropical forests. In temperate and tropical regions, deciduous
forest floors contain half the organic matter of evergreen forest floors, regardless of whether they
are dominated by needleleaf or broadleaf evergreen species. Although the greatest forest floor N
and P contents occur in boreal evergreen forests, there is substantial variation among the N and P
contents of tropical, subtropical, and temperate forest floors.
The quantity of organic matter and nutrients contained in the forest floor of tropical,
temperate, and boreal forests reflect two opposing processes: the production and the decomposi-
tion of plant litter. Consequently, it is not surprising that patterns of leaf and root litter production
alone do not well reflect differences in forest floor mass among different forest ecosystems (com-
pare Tables 19.4 and 19.5). Temperature, precipitation, and the biochemical constituents of plant
Table 19.4 The nitrogen content of leaf and fine-root litter in forest ecosystems from different geographic
locations.
Litter
Ecosystem Leaf Fine root Root:leaf N
kg N ha−1y−1
Tropical
Broadleaf evergreen 119 255 1.9
Warm temperate
Broadleaf deciduous 36 44 1.2
Cold temperate
Needleleaf evergreen
White pine 21 40 2.0
Mixed pine 16 36 2.2
white spruce 28 22 0.8
red pine 12 19 1.6
Broadleaf deciduous
black oak 31 79 2.5
Northern red oak 30 62 2.0
white oak 26 47 1.8
sugar maple 23 47 2.0
paper birch 25 43 1.7
Boreal
Needleleaf evergreen 24 26 1.1
Compare the magnitude of nitrogen entering the soil from leaf and fine-root litter and notice that the N content of fine-root litter
is equivalent to or greater than that of leaf litter.
Source: Vogt et al. (1986) / with permission of Elsevier and Nadelhoffer et al. 1985. Modified from Advances in Ecological
Research, © 1986 by Academic Press. Inc. Reprinted with permission of Academic Press, Inc. and from Ecology by the permis-
sion of the Ecological Society of America.
litter all influence the rate at which leaf and root litter is decomposed by soil microorganisms;
differences in these factors contribute to the variation in forest floor accumulation among ecosys-
tems in Table 19.5. Earthworm activity is also of importance, because these organisms incorporate
fresh leaf litter into surface mineral soil horizons which accelerates the decomposition process
(Edwards and Bohlen 1996). In the absence of earthworms, relatively thick, distinct organic hori-
zons develop over thin A horizons. By contrast, discontinuous and thin organic horizons form over
thick, organic-matter-rich A horizons in the presence of earthworms.
A simple model considering plant litter production and decomposition has been used to gain
insight into forest floor dynamics (Olson 1963). This mass-balance approach assumes that the
annual rate of litter decomposition equals the annual rate of plant litter production, such that the
amount of organic matter in the forest floor is unchanged. This condition only occurs very late in
forest ecosystem development, and it is where this approach has been used to estimate the rate of
decomposition. Under this assumption, a constant proportion (k) of forest floor organic matter
decomposes on an annual basis:
Litter Production k Forest Floor Mass .

Table 19.5 Forest floor mass and nutrient contents in tropical, temperate, and boreal forest ecosystems.
These values have been summarized for a wide range of forest ecosystems occurring in tropical,
temperate, and boreal regions of the Earth.
Forest floor
Organic matter N P
(Mg ha−1)
Ecosystem kg ha
−1
Tropical
Broadleaf evergreen 22.6 325 8
Broadleaf semideciduous 2.2 35 –
Broadleaf deciduous 8.8 – 14
Subtropical
Broadleaf deciduous 8.1 – –
Warm temperate
Needleleaf evergreen 20.0 362 25
Broadleaf deciduous 11.5 163 12
Cold temperate
Broadleaf deciduous 32.2 624 50
Boreal
Source: Vogt et al. (1986) / with permission of Elsevier. Modified from Advances in Ecological Research, © 1986 by Academic
Press, Inc. Reprinted with permission of Academic Press. Inc.
In this equation, k represents the decomposition rate constant for the entire forest floor. With
knowledge of litter production and forest floor mass, values of k are derived using the follow-
ing equation:
1
Litter production Mg ha 1y 1
k y 1
.
Forest floor mass Mg ha
When decomposition is rapid, relatively small amounts of organic matter accumulate in the forest
floor, and values for k are typically greater than 1 y−1. This situation occurs in late-successional
tropical forests in which microbial activity has the potential to respire more organic matter than
that contained in aboveground and belowground litter production (Cuevas and Medina 1988). By
contrast, decomposition is slow in cold, boreal forests where substantial amounts of organic matter
and nutrients accumulate in the forest floor; values of k are approximately 0.01 y−1.
If the forest floor is neither gaining nor losing organic matter (i.e., it is in equilibrium), then
k can be used to estimate the mean residence of time of organic matter or nutrients (l/k), which is
the average time organic matter or nutrients reside in the forest floor. Values range widely among
the forest ecosystems listed in Table 19.6, primarily reflecting differences in temperature, precipi-
tation, and plant litter biochemistry. The mean residence time for organic matter, N, and P in the
forest floor of tropical forests is less than 1 year (Table 19.6), whereas these constituents can reside
within the forest floor of boreal forests for tens to hundreds of years.
Table 19.6 The mean residence time of organic matter, nitrogen and phosphorus, in the forest floor of
tropical, temperate, and boreal forest ecosystems. Mean residence time of forest floor organic matter is
calculated by dividing the forest floor mass (kg ha−1) by the annual leaf litterfall mass (kg ha−1 y−1); residence
times for nutrients are calculated in the same manner.
Mean residence time in yearsa

Organic matter P
N
Ecosystem
Tropical
Broadleaf evergreen 2 2 1
Broadleaf semideciduous 0.4 0.2 –
Broadleaf deciduous 0.9 – 2
Subtropical
Broadleaf deciduous 2 – –
Warm temperate
Needleleaf evergreen 5 14 11
Broadleaf deciduous 3 5 4
Cold temperate
Broadleaf deciduous 10 19 11
Boreal
a
Mean residence times are 1/k.
Source: Vogt et al. (1986) / with permission of Elsevier. Modified from Advances in Ecological Research. © 1986 by Academic
Press, Inc. Reprinted with permission of Academic Press, Inc.
Relatively large amounts of N and P (Table 19.6) in the forest floor of cool temperate and
boreal forests, combined with long residence times (1/k), emphasize the importance of the forest
floor as a site for nutrient storage in these forest ecosystems. Also, compare the mean residence
times of coniferous (i.e., needleleaf) and deciduous forests, and notice that organic matter and
nutrients reside in the forest floor of coniferous forests for relatively longer periods of time. This
pattern reflects the fact that conifers contain greater proportions of organic compounds that are
not easily metabolized by soil microorganisms, thus they have slower rates of litter decomposition.
Mean residence times in Table 19.6 are calculated using rates of aboveground litter produc-
tion alone, which comprise only 28–56% of total litter input to forest floor (see Table 19.4). Using
leaf litter production as the primary input of plant litter, the mean residence time of organic matter
in the forest floor of cool temperate forests ranges from 8 to 67 years. The range of values is much
lower when fine-root litter is included in the calculation of mean residence time (e.g., 5–15 years;
Vogt et al. 1986), further emphasizing the importance of belowground litter production in studying
the storage and cycling of nutrients in forest ecosystems.
ORGANIC MATTER DECOMPOSITION AND NUTRIENT MINERALIZATION

To fully understand the process of decomposition and the release of nutrients from plant litter and
organic matter, one needs to consider the factors that control microbial growth and maintenance in
soil. Once abscised leaves, dead roots, and other plant tissues enter forest floor and mineral soil, they
become a growth substrate for saprotrophic microorganisms. During the decomposition process,
soil bacteria, actinobacteria, and fungi assimilate the organic compounds contained in plant litter
into their cells for biosynthesis (i.e., growth and maintenance), albeit at different rates depending on
the types of compounds contained in plant litter. Whether soil microorganisms release nutrients
into, or assimilate nutrients from, soil solution depends on the biochemical constituents of plant
litter and their suitability for microbial growth and maintenance. In the paragraphs that follow, we
discuss how plant litter functions as a substrate for microbial growth and maintenance in soil, and,
in turn, how these microbial processes influence the supply of growth-limiting nutrients to plants.
As you will read later in this section, the amount of energy that soil microorganisms derive from
plant litter controls the processes of decomposition and soil N availability.
Biochemical Constituents of Plant Litter The leaves, stems, and roots of plants are con-
structed of a remarkably small set of organic compounds, regardless of differences in phylogeny
and growth form. Plant litter is largely composed of cellulose (15–60%), hemicellulose (10–30%),
lignin (5–30%), protein (2–15%), fats (1%), and soluble compounds such as sugars, amino acids,
nucleic acids, and organic acids (10%). These organic compounds carry out essential metabolic
functions, provide physical support, and defend plants against herbivores and pathogens. Upon
entering soil, however, they become substrates that yield different amounts of energy for microbial
growth and maintenance.
The process of decomposition is mediated by the microbial production of enzymes that
harvest the biochemical energy contained in compounds formed during the biosynthesis of plant
tissue. In most terrestrial ecosystems, microbial growth in soil is limited by usable forms of energy,
and amounts contained in annual litter production are only sufficient to maintain (i.e., no growth)
microbial populations in soil. The microbial production of enzymes that make the energy contained
in plant compounds available is energetically expensive. Consequently, energy produced during
the microbial metabolism of a particular plant compound must surpass the cost of the enzymes
used to harvest its energy in order for microorganisms to use it as a growth substrate.
The use of a plant-derived organic compound as a substrate for microbial growth is deter-
mined by: (i) the amount of energy released by breaking different types of chemical bonds, (ii) the
size and three-dimensional complexity of molecules, and (iii) its nutrient content (N and P). Look
at Figure 19.12 and notice that the complexity of chemical bonds, three-dimensional structure,
and nutrient content varies dramatically among the primary constituents of plant litter. Simple
sugars (i.e., carbohydrates), such as glucose, are among some of the first products of photosyn-
thesis, and although their concentration in fresh litter is low (<5%), these water-soluble compounds
yield substantial amounts of energy for microbial growth. In soil, glucose and other simple sugars
can be directly taken up and used to fuel energy-producing biochemical reactions (i.e., glycolysis
and tricarboxylic acid cycle which generate ATP) within microbial cells. Given these characteris-
tics, it should not be surprising that simple sugars are rapidly decomposed in plant litter.
Amino acids are the building blocks of proteins and molecules that carry out all biochemical
reactions. A single protein contains thousands of amino acid subunits, making them too large to
be directly taken up by microbial cells. Proteases released by microbial cells enzymatically cleave
proteins into smaller polypeptides and eventually into their amino acid subunits. These relatively
small, high-energy-yielding molecules can then be taken up by microbial cells and used for protein
synthesis or energy production. Notice that proteins and amino acids are the only plant com-
pounds in Figure 19.12 that contain N, an essential nutrient also required for microbial growth and
maintenance. This fact has important implications for the release of N from plant litter during the
process of mineralization, which we discuss later in this section.
Starch is a storage carbohydrate (Figure 19.12) and its synthesis occurs when the supply of
photosynthate exceeds the carbohydrate requirements for the construction and maintenance of plant
tissues. The concentration of this energy-storing glucose polymer is typically low in many types of
plant litter, because starch is used for maintenance respiration prior to the senescence of leaves and
CH2OH CH2OH CH2OH

H C O H H C O H H C O H
H C H C C H C C H C
CH2OH OH H OH H OH H
+
H3N C COO– O C C O C C O C C O
H C O H
H OH H OH H OH
C H C CH2
OH H Starch
OH C C OH CH2 A storage carbohydrate
H OH COO– CH2OH H OH CH2OH

H C O O C C H H C O O
Glucose Glutamate
H H H H
A simple sugar An amino acid building C OH H C C OH C C OH H C
block of protein O C C H H C O O C C H
H OH CH2OH H OH
Cellulose
A structural carbohydrate
CH2OH CH2OH CH2OH

H C O H OH C O H H C O H
C HOH H C C HOH H C C HOH HO C
OH C C OH H C C OH OH C C OH
H OH H OH H H
glucose galactose mannose
COOH COOH
H C O H OH C O H
C H C C H C
OH H OH H
OH C C OH H C C OH
H OH H OH
glucuronic acid galacturonic acid
H H
H C O H OH C O H
C HOH H C C HOH H C
OH C C OH H C C OH
H OH H OH
xylose arabinose
Four Forms of Hemicellulose
A Structural Carbohydrate
Building Blocks of Hemicellulose
H2COH H2COH
HC CO
HCOH CH2
H2COH
H2COH MeO OMe H COH HC
2
OMe
H2COH CO HC
O CH O
CH H2COH CH2
HC
CH CH CH
1/2 1/2 OMe
HC OMe HCO H2COH
OMe OH
MeO H2COH CH HC HC
H2COH O O
O (OMe)0.5 HC CO
CH HC OC CH2
HC O OH HCOH HC CH
MeO OMe HC HCOH
H2COH
H2COH OMe O CH
O
HC O HC O OMe OMe
H2COH CH
HCO(C6H10O5)nH HC CH OH O
HC CH2 MeO
O
H2COH OH
O CH H2COH OMe
HCOH HC O
HC [ O
MeO
O OMe
OH
Lignin
F I G U R E 1 9 .1 2 The chemical constituents of plant litter that influence the rate of organic matter decompo-
sition in forest floor and mineral soil. Notice that nitrogen is a component of amino acids and proteins, but it
is not found in the chemical structure of the other plant compounds illustrated in this figure.
fine roots. Starch is broken down by many soil bacteria, actinobacteria and fungi, which synthesize
extracellular enzymes (i.e., amalyses) that cleave starch into subunits consisting of several glucose
molecules; these subunits can then be taken up and assimilated by microbial cells where they are
used to produce energy. Although the decomposition of starch is somewhat slower than that of
simple sugars, its metabolism yields relatively large amounts of energy (ATP) for microbial growth.
Cellulose is the primary component of plant cell wall (Figure 19.12) and is the most abun-
dant compound in plant litter. It is often associated with hemicellulose and lignin, which also are
present in the plant cell wall. Cellulose is formed by linking glucose molecules into an unbranched,
semicrystalline polymer. Notice that starch and cellulose differ in the type of bond joining adjacent
glucose subunits (Figure 19.10). Unlike starch, the glucose subunits of cellulose cannot be mobi-
lized by plants for use in energy production once it is formed. Cellulose is a structural polymer,
whereas starch is an energy-storage polymer.
The decomposition of cellulose occurs as a two-step process. First, the bacterial production
of the enzyme cellobiohydrolase repeatedly cleaves off two glucose subunits from the end of the
molecule. Second, the resulting disaccharide, cellobiose, is degraded into glucose by the enzyme β
glucosidase. As with starch degradation, the glucose subunits can be directly assimilated by the
decomposing microorganism. Unlike proteins and amino acids that contain N, both starch and
cellulose are exclusively composed of C, H, and O.
Hemicellulose is a heterogeneous group of polymers composed of several types of plant
sugars (Figure 19.12). The linking of these sugar subunits produces branched as well as unbranched
molecules, which in pure state are often broken down quite rapidly. In nature, however, the
association of hemicellulose with other substances such as lignin in the plant cell wall makes their
breakdown more difficult. Pectin (polygalacturonic acid) is a type of hemicellulose molecule found
in the middle lamella of plant cell walls, and it is broken down by several enzymes collectively
known as pectinases. These enzymes appear to be primarily produced by soil fungi and actinobac-
teria, which use the sugar subunits as a source of energy. The initial entry of mycorrhizal fungi into
plant roots is thought to be facilitated by the production of pectinases. There is little difference
between cellulose and hemicellulose in energy yield and hence their rate of decay by microorganisms.
Lignin is by far the most biochemically complex constituent of plant litter, containing a
variety of chemical bonds and three-dimensional conformations (Figure 19.12). It is an abundant
component of leaf litter, and high concentrations occur in most woody tissues. Lignin molecules
surround cellulose and hemicellulose microfibrils in the plant cell wall, providing rigidity and
protecting them against pathogenic fungi and bacteria that break down plant cell walls. Because of
its biochemical complexity and the fact that most soil microorganisms cannot enzymatically
harvest the energy contained within its chemical bonds, lignin decomposition occurs at a very slow
rate. White-rot (Pleurotus, Phanerochaete) and brown-rot fungi (Poria, Gloeophyllum) are primarily
responsible for the degradation of lignin in plant litter. In order to degrade lignin, these organisms
require the presence of an alternative energy source, which functions as the primary growth sub-
strate. As a result, only small portions of the C contained in lignin are actually incorporated into
fungal cells; the majority enters into reactions that eventually form humus.
Dynamics of Decomposition The chemical constituents of plant litter and their use for
microbial biosynthesis directly control the rate at which plant litter decays on and in the soil. The
microbial breakdown of the biochemical constituents in plant litter can be described using the
following first-order rate equation:
At Aoe kt .
In this equation, the amount of substrate remaining (At in μg g−1) at any point during the
decomposition process is proportional to its initial concentration (A0 in μg g−1) and its first-order
rate constant for decomposition (k in days−1, months−1, or years−1). First-order decomposition rate
constants are experimentally derived values obtained by measuring the decline in a substrate
Table 19.7 Rate constants for the decomposition of organic compounds contained in plant litter.
Compound Rate constant for decay k (day−1)

Glucose 0.500–1.000
Protein and simple sugars 0.200
Cellulose 0.036–0.080
Hemicellulose 0.030–0.080
Lignin 0.003–0.010
Values were determined by decomposing purified compounds under laboratory conditions.

Source: Adapted from Alexander (1977), Paul and Clark (1996), and van Veen et al. (1984).
during its decomposition, often under laboratory conditions. As such, they directly reflect how
rapidly a particular compound can be used for microbial biosynthesis. Rapidly decomposed sub-
strates have a high k, whereas substrates that provide little energy for microbial growth and main-
tenance have low values. Decomposition rate (k) constants also are sensitive to temperature and
soil water potential, doubling for a 10 °C increase in temperature and attaining maximum values
near field capacity.
In Table 19.7, we summarize decomposition rate constants for the primary constituents of
plant litter. Notice that glucose, a small, high-energy-yielding molecule has a very-rapid decompo-
sition rate constant, as do other simple carbohydrates and proteins. Contrast these values with the
low-rate constant of lignin. Complex chemical bonds, subunits that provide little energy for micro-
bial growth, and a folded three-dimensional structure that protects the inner portion of the lignin
molecule against enzymatic attack all contribute toward its slow rate of microbial degradation.
Also notice that cellulose and hemicellulose have similar decomposition rate constants, reflecting
the fact that these molecules provide equivalent amounts of energy for microbial growth
(Table 19.7).
The biochemical attributes of plant-derived molecules that influence their use for microbial
growth are clearly reflected in the decomposition rate constants listed in Table 19.7. However,
plant litter is not composed of a single organic compound, but instead contains varying propor-
tions of simple sugars, protein, cellulose, hemicellulose, and lignin. How do different proportions
of these compounds in plant litter influence the rate at which it is metabolized by soil
microorganisms?
Consider the decomposition of an abscised leaf composed of 10% protein and simple carbo-
hydrates, 25% lignin, and 65% cellulose and hemicellulose (Figure 19.13a). The overall decline in
leaf mass seen in Figure 19.11a results from microbial respiration for growth and maintenance,
which returns this photosynthetically fixed CO2 to the atmosphere. Notice that the decomposition
of this leaf is initially rapid (0–20 days), but that rate slows down during the later stages of decay.
During the initial stages of decomposition, the leaching of water-soluble compounds and the
metabolism of proteins and simple carbohydrates (high decomposition rate constants) contribute
to a rapid loss of C from the leaf. In this example, proteins and simple carbohydrates are almost
totally degraded after only 15 days of microbial decay. Metabolism of cellulose and hemicellulose
influences the intermediate stages of decomposition (10–50 days), while the latter stages (50–
100 days) of decomposition are dominated by lignin breakdown. The concentration and decompo-
sition rate constants of these compounds additively yield the overall pattern of leaf decomposition
in Figure 19.13a.
The situation is much different for an abscised leaf containing lower amounts of lignin (5%)
and higher proportions of protein and simple sugars (25%), and cellulose and hemicellulose (70%;
Figure 19.13b). Compared to the leaf in Figure 19.13a, the leaf containing 5% lignin (Figure 19.13b)
exhibits a more rapid overall decline in C than the leaf containing 25% lignin (Figure 19.13a).
(a)
100
90 Leaf
80
C Remaining (%)
70 Cellulose + Hemicelluous
60
50 Proteins + Simple Carbohydrates
40
30 Lignin
20
10
0
0 20 40 60 80 100
Time (days)
(b)
100
90 Leaf
80
C Remaining (%)
70 Cellulose + Hemicelluous
60
50 Proteins + Simple Carbohydrates
40
30 Lignin
20
10
0
0 20 40 60 80 100
Time (days)
F I G U R E 19.13 The decomposition of an abscised leaf in relationship to initial concentrations of protein,

simple carbohydrates, hemicellulose, cellulose, and lignin. The amount of leaf C remaining is controlled by
the concentration and decomposition rate constant of protein and simple sugars, cellulose and hemicellulose,
and lignin. The amounts of leaf C remaining in panels (a) and (b) were calculated using the following series
of first-order decay equations (percent leaf C = A1e−k1t + A2e−k2t + A3e−k3t, where A1 = percent protein and
simple carbohydrates, A2 = percent cellulose and hemicellulose, and A3 = percent lignin). The decomposition
rate constant values (k) are 0.2 days−1 for protein and simple carbohydrates, 0.08 days−1 for cellulose and
hemicellulose, and 0.01 days−1 for lignin. Note that differences in initial concentrations have a profound
influence on leaf decomposition. High lignin concentrations and low concentrations of other constituents
result in relatively slow decomposition rates.
Moreover, a higher proportion of protein and simple carbohydrates (Figure 19.13b) results in a
rapid initial decline in leaf C, much greater than that of the leaf material lower in protein
and simple carbohydrates (Figure 19.13a). Clearly, the relative proportion of easily metabolized,
high-energy-yielding plant compounds, such as simple sugars, and of biochemically complex
molecules that function as poorer substrates (i.e., lignin) for microbial growth has a profound
influence on the overall decay of plant litter entering forest floor and mineral soil.
Decomposition is not solely controlled by differences in the biochemical constituents of
plant litter. It also is controlled by the quantity of N and other nutrients available for microbial bio-
synthesis. In the Alaskan taiga, N and lignin concentrations in leaf litter have an important
influence on its decomposition rate constant and the rate at which it decomposes in the forest floor.
High concentrations of lignin, which reduce the overall energy yield of litter for microbial biosyn-
thesis, result in substantial declines in decomposition rate constants (Figure 19.14a). In contrast to
(a)
0.003 Black Spruce
White Spruce
Trembling Aspen
0.002 Paper Birch
k (days–1)
0.001
0
10 15 20 25
% Lignin
(b)
0.003
0.002
k (days–1)
0.001
0
10 15 20 25
% N per g Organic Matter
F I G U R E 1 9 .1 4 The relationship between decomposition rate constant (k) and (a) the lignin and
(b) nitrogen concentration of leaf litter produced in the Alaskan taiga. Source: After Flanagan and Van Cleve
(1983) / Canadian Science Publishing. Reprinted with permission of National Research Council Canada.
this relationship, decomposition rate constants are rapid in leaf litter with high concentrations of
N (Figure 19.14b). A similar relationship can be observed in the northeastern United States, where
high lignin:N ratios in the leaf litter of deciduous trees are reflected in slow rates of decomposition
(Melillo et al. 1982).
The relationship between decomposition rate and litter N concentration results from the fact
that soil microorganisms, like plants, require N to synthesize the biochemical constituents that
form new cells and maintain existing functions. Fungal biosynthesis generally requires 1 atom of
N for every 5–15 atoms of C assimilated during the decomposition of a particular substrate.
Bacteria, on the other hand, require 1 atom of N for every 3–5 atoms of C. As a result, soil fungi
have a C:N ratio ranging from 15:1 to 5:1, whereas those of soil bacteria are much lower (i.e., 5:1 to
3:1; Paul and Clark 1996). Look again at Figure 19.12 and realize that proteins and amino acids are
the primary N-containing constituents of plant litter. Because protein and amino acid concentra-
tions are low (2–15%) in most plant tissues, newly abscised leaves and dead fine roots are energy-rich
and N-poor substrates for microbial metabolism.
Once litter enters forest floor and mineral soil, its organic constituents can only be used for
microbial growth if there is a sufficient source of N and other nutrients. Leaf litter always has a
C:N ratio wider (e.g., 50:1) than that of microbial cells, and its decomposition requires the presence
of an additional source of N. Otherwise, soil microorganisms are unable to use the organic com-
pounds contained in litter for biosynthesis. The positive relationship between litter N concentration
and decomposition rate constants illustrates this point (Figure 19.14b); low concentrations of N in
plant litter limit the rate at which organic compounds are used for microbial growth and hence
slow its rate of decomposition. The inverse relationship between lignin and decomposition rate
indicates that increasing proportions of poor-energy-yielding compounds in plant litter also slow
the process of decomposition (Figure 19.14a).
In summary, decomposition is controlled by the energy yield of organic compounds (e.g.,
proportion of cellulose vs. lignin) in plant litter and the amount of N and other nutrients available
for microbial growth and maintenance. Herein lies an important ecological link between N-use
efficiency, litter N concentration, and microbial activity during the process of decomposition.
Plants with a high N-use efficiency withdraw a relatively large proportion of N prior to leaf abscis-
sion, thereby lowering the N concentration of leaf litter. The production of leaf litter with a low N
concentration slows decomposition and the subsequent release of inorganic N and other nutrients.
Thus, nutrient-use efficiency and plant litter biochemistry can exert an important feedback on the
rate at which nutrients cycle between plants and soil. In the following paragraphs, we further
explore the microbially mediated release of N during organic matter decomposition—processes
that influence the productivity of many forest ecosystems.
Nitrogen Immobilization and Mineralization Although forest floor and mineral soil rep-
resent the largest pools of N in forest ecosystems, the majority (90%) of N contained within them
resides in organically bound forms that are largely unavailable to plants. Saprotrophic soil micro-
organisms mediate soil N availability within terrestrial ecosystems, because their growth and
maintenance control the amount of inorganic N (i.e., NH4+) that is released during the decay of
litter and soil organic matter. The balance of microbial NH4+ assimilation and release supplies
approximately 90% of the N that is annually taken up and assimilated by plants in unmanaged eco-
systems and is well correlated with NPP in many forest ecosystems.
In the previous section, we learned that litter decomposition was controlled by the types of
organic compounds and the N required to use them for microbial growth. In the following para-
graphs, we further explore the extent to which these factors control microbial requirements for N
and the release of NH4+ from soil organic matter, the rate-limiting step in soil controlling the supply
of N for plant growth in unmanaged terrestrial ecosystems.
When plant litter enters the soil, it is colonized by saprotrophic microorganisms, and
high-energy-yielding constituents (i.e., organic acids, simple sugars, and carbohydrates) are prefer-
entially used for microbial biosynthesis. Because this material is an energy-rich and N-poor sub-
strate for microbial growth, NH4+ must be assimilated from soil solution to form new N-containing
compounds within microbial cells. Nitrogen immobilization is the microbial uptake and assim-
ilation of NH4+ into organic compounds. This process is illustrated by the following reaction in
which an organic acid combines with NH4+ to produce an amino acid:
COO– COO–
C = O + NH4 + H–C–NH2 + H2O

Glutamate dehydrogenase
CH2 CH2
CH2 CH2
COO– COO–
Organic Acid Amino Acid

(α-ketoglutarate) (L-glutamate)
Amino acid synthesis, facilitated by the glutamate dehydrogenase enzyme, is the first step in
the production of proteins needed for the growth and maintenance of microorganisms. Nitrate
(NO3−) also can be used by soil microorganisms, but it first must be reduced to NH4+ before it can
form an amino acid. Nitrogen immobilization is the result of microbial growth on an energy-rich
and N-poor substrate, and it characterizes the initial stages of litter decomposition. The hyphal
networks of fungi, which extend from forest floor into mineral soil, are particularly effective at
transporting NH4+ into freshly decaying litter.
Nitrogen mineralization is the release of NH4+ during the microbial breakdown of pro-
teins, amino acids, and other N-containing organic compounds. Consequently, N mineralization is
the reverse of the N immobilization reaction illustrated earlier in the text, and it is carried out by
all microorganisms involved with organic matter decomposition. Whether an amino acid pro-
duced during protein degradation is used as an energy source or as a building block for microbial
protein synthesis depends on a series of feedback controls. For example, carbohydrates contained
in litter are preferentially used to generate energy in microbial cells, thereby preserving any amino
acids present for microbial protein synthesis. Nonetheless, this relationship can change depending
on the availability and energy yield of organic compounds remaining in plant litter.
During the initial stages of decomposition, carbohydrate availability is relatively high and
amino acids contained in litter are directly assimilated for microbial protein synthesis. However,
carbohydrate supply (i.e., energy) begins to limit microbial growth during the latter stages of litter
decomposition (Figure 19.13). When quantities are insufficient to meet the maintenance energy
requirements of microbial cells, mortality occurs and the constituents of the dead cells begin to
serve as substrates for surviving microorganisms. Under energy-limited conditions, the C-skeletons
(i.e., organic acids) of amino acids are used by surviving microbial cells to generate energy. During
this process, NH4+ is released into soil solution where it can be assimilated by plant roots, partici-
pate in cation exchange reactions (Chapter 9), or enter into other microbially mediated processes.
Because of this, virtually all N molecules in plants have first cycled microbial cells.
Nitrogen mineralization and immobilization occur simultaneously during the process of
litter decomposition, albeit at different rates depending on the organic compounds present and the
amount of N available to use those compounds for microbial metabolism. At what stage during
litter decomposition is there a net release of NH4+ into soil solution where it can be taken up and
assimilated by plant roots? Nitrogen is released from decomposing litter only when the gross rate
of N mineralization exceeds the gross rate at which NH4+ is incorporated into microbial biomass
(i.e., gross N immobilization).
Figure 19.15 illustrates the relationship between the C and N contents of leaf litter and
changes in microbial metabolism that influences the gross immobilization and mineralization of
N. During the initial phases of decomposition, plant litter is relatively rich in energy-yielding
compounds, poor in N, and has a C:N ratio (50:1; Figure 19.13a) wider than that of microbial cells
(approximately <8:1). In order to use litter constituents for growth or maintenance, soil microor-
ganisms need to assimilate NH4+ or NO3− from soil solution. The initial demand for N for microbial
biosynthesis causes high rates of gross N immobilization and an initial accumulation of N in
decomposing litter (compare panels a and c of Figure 19.15). Comparing the decline in leaf C over
time with rates of microbial respiration illustrates that litter mass loss results from the return of
leaf C to the atmosphere as CO2 during microbial respiration (Figure 19.15a,b). The initial increase
in litter N from gross N immobilization, in combination with the loss of C during microbial respi-
ration, causes leaf C:N to decline over time.
Later in the decomposition process, decomposing microorganisms are limited by substrates
that provide little energy (i.e., energy- or C-limited phase), and microbial respiration and biosyn-
thesis subsequently decline (Figure 19.15b). Proteins and other N-containing compounds released
from cell death during this energy-limited phase of decomposition are then metabolized by sur-
viving microbes. These organisms release NH4+ and use the resulting organic compounds for
(a)
% C or N Remaining
50
100 40
Litter C:N
30
50
Carbon C:N Nitrogen 20
0 10
0
(b)
Microbial Respiration
Microbial Activity
Microbial Biosynthesis
(c) Net
Nitrogen Dynamics
Immobilization Gross N
Mineralization
Net
Mineralization
Gross N Immobilization
N-Limited C-Limited Microbial Metabolism

Microbial
Metabolism
Time
F I G U R E 1 9 .1 5 Changes in the C and N contents of leaf litter during the process of decomposition. Leaves
enter the forest floor at time = 0, and (a) the initial decline in C results from (b) microbial respiration and
biosynthesis. (c) Gross rates of immobilization exceed those of mineralization, due to the high-energy–low N
stratus of leaf litter, which creates a microbial demand for N. The amount of N available for microorganisms
to harvest the energy contained in plant litter initially limits decomposition. Following the use of
high-energy-yielding substrates in litter, microbial biosynthesis slows and the demand for N declines. Gross
mineralization exceeds gross immobilization during the period of C limitation, in which soil microorganisms
are releasing NH4+ from N-containing organic compounds in order to harvest the energy contained
within them.
energy production. As a result, gross rates of N immobilization fall below gross rates of N mineral-
ization, and there is a net release of NH4+ from decomposing litter (net N mineralization). These
dynamics illustrate an important point—the microbially mediated processes of gross N minerali-
zation and gross N immobilization control the transfer of N from soil organic matter to soil solu-
tion where it can be taken up by plant roots. Although the biomass of soil microorganisms contains
only 1.5% of the C and 3.0% of the N within forest ecosystems (Wardle 1992), their metabolic activ-
ities truly make them the “gate keepers” of soil N availability. We direct the reader to Staaf and
Berg (1982) for a discussion of P, S, K, Ca, Mg, and Mn mineralization during plant litter
decomposition.
Nitrogen Availability in Forest Ecosystems It is important to realize that net N immobi-

lization only characterizes the initial phases of plant litter decomposition (Figure 19.15c). Soil
organic matter or humus is the end product of litter decomposition; it has a low C:N (10:1) and
contains few compounds that can be used for microbial biosynthesis. It is characterized by the net
release of NH4+, which supplies the majority of N that is taken up and assimilated by plants on an
annual basis. Generally, net N mineralization rates increase with the organic matter content of
mineral soil (Marion and Black 1988). We direct the reader to Binkley and Hart (1989) and Hart
et al. (1994) for a review of current techniques for measuring soil N transformations.
In North American temperate forests, rates of net N mineralization range from 30 to
120 kg N ha−1 y−1 (Pastor et al. 1984; Zak and Pregitzer 1990; Binkley 1995). This process supplies
approximately 90% of the N available for plant uptake on an annual basis in forest ecosystems. For
example, compare rates of net N mineralization with inputs from atmospheric deposition
(Figures 19.2 and 19.3) and free-living N2 fixation (see Biological Fixation of Nitrogen). Notice that
rates of net N mineralization in most forest ecosystems are approximately 10 times greater than the
combined annual inputs from atmospheric deposition and free-living N2 fixation.
Nitrification The NH4+ produced during the process of net N mineralization can be retained by
cation-exchange reactions, taken up by plant roots, or it can be oxidized to NO3− during the process
of nitrification. Nitrification is an ecologically important N-cycling process because, in the
absence of plant uptake or microbial immobilization, NO3− can rapidly move in the downward
flow of water through the soil profile. The high mobility of this ion, in combination with the low
anion-exchange capacity of most temperate soils, makes NO3− the form of N most often lost from
terrestrial ecosystems. Once NO3−, or any other nutrient, has passed below the majority of plant
roots, it is lost from the ecosystem and eventually enters groundwater, streams, or lakes. Conse-
quently, nitrification is of great importance for understanding patterns of N loss from forest
ecosystems.
Nitrification is the two-step oxidation of NH4+ to NO3− by chemoautotrophic bacteria and
archaea. For well over 100 years, NH4+ oxidation was thought to be carried out exclusively by che-
moautotrophic bacteria. With the advent of molecular tools, we now understand that chemoauto-
trophic archaea can be dominant agents of NH4+ oxidation in soil. They were previously unknown,
because they will not grow on laboratory media, as do NH4+ oxidizing bacteria. Both NH4+ oxidizing
bacteria and archaea use NH4+ as an energy source to fix CO2 from the soil atmosphere. Nitrifying
bacteria are sensitive to pH, water potential, and aeration; nitrifying archaea appear to be more
important where NH4+ concentrations in soil are very low. In general, bacterial nitrification dimin-
ishes below pH 6.0, is most rapid at matric potentials of −0.01 to −1.0 MPa, and requires the
presence of O2 (Paul and Clark 1996). However, nitrification has been observed in some acidic
forest soils. There is some evidence to suggest that several heterotrophic bacteria (Arthrobacter
spp.) and fungi (Aspergillus spp.) can produce small amounts of NO3− from NH4+, but their ecolog-
ical significance in forest soils is largely unknown (Schimel et al. 1984).
The first step in the process of nitrification in soil is mediated by chemoautotrophic bacteria
(Nitrosomonas, Nitrosospira, Nitrosococcus, and Nitrosovibrio), as well as chemoautotrophic
archaea (Nitrososphaera and Nitrosopumilus) in the following manner:
NH 4 1.5 O2 NO2 2H H 2O.
The NO2− produced in the above reaction is used by Nitrobacter winogradskyi and Nitrobacter agilis
to obtain energy in the second reaction in the nitrification process:
NO2 H 2O H 2O NO2 NO3 2H

.
2 H 0.5 O2 H 2O
Notice that NH4+ and NO2− oxidation require O2, making nitrification unlikely in very poorly
drained soil. The production of H+ during NH4+ oxidation (first reaction) can lower the pH of soil
with high NH4+ concentrations and rapid nitrification rates.
Nutrient Loss From Forest Ecosystems 523
The potential for NO3− loss following nitrification has drawn the attention of ecologists and
soil microbiologists, who have sought to understand the factors controlling the activity of nitrifying
bacteria and archaea. Competition appears to be an important controlling factor, because plants,
decomposing microorganisms, and nitrifying bacteria all require NH4+ as either a biosynthetic
building block or as an energy source. Most plant roots and decomposing microorganisms have
NH4+ uptake rates more rapid than those of NH4+-oxidizing bacteria, making plant uptake and
microbial immobilization important controls constraining nitrification in many ecosystems (Rob-
ertson and Vitousek 1981; Robertson 1982; Zak et al. 1990). Nitrification rates can greatly increase
in the absence of plant uptake, further suggesting that plant-microbe competition can limit nitrifi-
cation. There is some evidence that plant compounds synthesized to deter herbivory (tannins and
terpenoides) can inhibit nitrifying bacteria (White 1986, 1988).
Nitrification rates dramatically differ among intact forest ecosystems, ranging from 0 to
100% of net N mineralization (Pastor et al. 1984; Zak and Pregitzer 1990; Binkley 1995). The high-
est rates of nitrification often occur in forest soils with rapid rates of net N mineralization, ample
soil water, and neutral pH. Because of its high rate of diffusion and low retention by anion-
exchange reactions in soil, one would expect that NO3− loss following disturbance should vary
according to initial nitrification rate and the extent to which it increases following the removal of
plant uptake. We build on our understanding of nutrient uptake by plants, organic matter decom-
position, and nitrification to understand patterns of N loss prior to and following disturbance
later in this chapter.
NUTRIENT LOSS FROM FOREST ECOSYSTEMS

In the previous sections, we saw that
relatively large quantities of nutrients Atmospheric
Deposition
annually circulate between overstory
trees, forest floor, and mineral soil. In
intact forest ecosystems, a small
proportion of the nutrients that annually
cycle among these ecosystem pools are Translocation
lost through physical export via the Aboveground

Litter
hydrologic cycle or biological export to
Fixation
Production
the atmosphere. Highlighted in the Gaseous

Loss
reduced version of Figure 19.1 (to the
right) are leaching and gaseous loss,
major pathways of nutrient loss.
Nutrients in soil solution that
exceed plant and microbial demands can
be transported below the rooting zone by Organic Matter
Forest Floor and
the downward flux of water through the
Belowground
Litter Mineral Soil
soil profile. Nutrients moving in this way Nutrient Uptake

Production
eventually enter ground or surface and Assimilation Decomposition

and Mineralization
waters where they then become a Nutrients
in
nutrient input for aquatic ecosystems. Soil Solution
Therefore, the nutrient cycles of terres- Mineral

Weathering
trial and aquatic ecosystems are inti- Leaching
mately linked through the flow of water

in the landscape. Nutrient loss from forest ecosystems also influences water quality and the
nutrient dynamics of aquatic ecosystems. In the paragraphs that follow, we discuss the physical
and biological loss of nutrients from intact forest ecosystems, thus completing the journey of
nutrients into, through, and out of forest ecosystems. We later build upon this foundation to
understand the extent to which natural disturbance and human activities influence nutrient loss
from forest ecosystems.
NUTRIENT LEACHING FROM FOREST ECOSYSTEMS

Leaching is the physical process by which nutrients exit terrestrial ecosystems in the downward
flow of water through the soil profile. This process can occur when precipitation exceeds the
amount of water lost through transpiration and evaporation. Water moving downward through the
soil profile has the potential to entrain and transport nutrients dissolved in soil solution. Nutrients
are subject to leaching only when water is moving downward through the soil profile. Therefore,
the hydrologic cycle provides the transport mechanism for the leaching of nutrients from terres-
trial ecosystems. Once below the rooting zone, nutrients are lost from the forest ecosystem and are
headed for groundwater or surface water.
The high demand for growth-limiting nutrients (N or P) by overstory trees typically main-
tains low concentrations in soil solution (also see Nutrient Uptake and Assimilation) and thus
lessens the potential for nutrient leaching in forest ecosystems. The leaching loss of N is often
much less than N inputs via atmospheric deposition and free-living N2 fixation (Vitousek and Mel-
illo 1979). On the other hand, non-limiting nutrients (i.e., Ca or K) can accumulate in soil solution,
because the rate of supply from weathering or the atmosphere exceeds their biological demand.
Under these conditions, leaching can export relatively large quantities of non-limiting nutrients
from terrestrial ecosystems.
It should be remembered, however, that excess water must be present to serve as the trans-
port mechanism for nutrient leaching. In temperate deciduous forests, the greatest potential for
nutrient leaching occurs in early spring prior to canopy development and in late autumn after leaf
fall. During these periods, a low biological demand for nutrients and water, and the resulting
ample amounts of water moving through the soil profile due to low transpiration rates, provide the
potential for leaching.
The potential for nutrient leaching, particularly the loss of limiting nutrients such as N
and P, has drawn the attention of forest ecologists over the past 50 years. In 1963, one of the first
attempts to quantify nutrient leaching from intact forests began at the Hubbard Brook Experi-
mental Forest in New Hampshire (Likens et al. 1977). Using small, bedrock-sealed watersheds
drained by a single stream, a team of ecologists began measuring nutrient leaching from northern
hardwood forests (also see Mineral Weathering). In this unique situation, leached nutrients could
only exit the bedrock-sealed watershed in streamflow. Many studies have been conducted
subsequent to this pioneering research; the reader is directed to Henderson et al. (1978), Vitousek
and Melillo (1979), and Swank and Crossley (1988).
Table 19.8 summarizes leaching losses of limiting (N, P, S), non-limiting (K, Ca), and nones-
sential (Si) plant nutrients from the second-growth northern hardwood forests at the Hubbard
Brook Experimental Forest; also provided are nutrient inputs from atmospheric deposition and
biological N2 fixation. Notice that leaching losses of N, P, and S are much lower than their input,
suggesting that these nutrients are being taken up and assimilated by forest vegetation. Contrast
this pattern to the losses of K, Ca, and Si, which exceed inputs from atmospheric deposition. These
nutrients weather from the K-, Ca-, and Si-rich bedrock of the Hubbard Brook Experimental Forest
at a rate surpassing plant demand, giving rise to their net export from the forest. In Table 19.8,
export or retention is the difference between leaching loss and input.
Similarly, Henderson et al. (1978) compared leaching losses of N, K, and Ca from watersheds
dominated by old-growth Douglas-fir in the Pacific Northwest with those dominated by oaks and
hickories in the southeastern United States. Despite ample amounts of precipitation in these
widely different forests (158–233 cm y−1), leaching losses of N were low (<0.5 kg N ha−1 y−1) and
Nutrient Loss From Forest Ecosystems 525
Table 19.8 Nutrient leaching from an intact northern hardwood-dominated watershed in the Hubbard
Brook Experimental Forest, New Hampshire.
Nutrient Leaching loss Input Export (−) or retention (+)

kg ha y
−1 −1
N 4.0 20.7 16.7

P 0.019 0.036 0.017
S 17.6 18.8 1.2
K 2.4 0.9 −1.5
Ca 13.9 2.2 −11.7
Si 23.8 0 −23.8
Nutrient inputs via atmospheric deposition and biological fixation are also summarized. Nutrients required for plant growth in
large quantities are accumulating within these second-growth forests. Leaching losses are greater than additions for nutrients
in excess of plant demand (i.e., Ca) or those not required for plant growth (i.e., Si).
Source: Likens et al. (1977) / Springer Nature. Reprinted from Biogeochemistry of a Forested Ecosystem, © 1977 by Springer-
Verlag New York, Inc. Reprinted with permission from Springer-Verlag.
were much less than the annual input of N from atmospheric deposition (1–6 kg N ha−1 y−1). These
ecosystems also leached higher amounts of K and Ca than they received in atmospheric deposi-
tion, again reflecting mineral weathering rates in excess of plant demand. In summary, these
examples illustrate that leaching is controlled by the biological demand for nutrients and the avail-
ability of excess water to transport nutrients below the rooting zone.
In the next section, we will consider how these factors change following natural and human
induced disturbances which initiate the process of ecosystem development (i.e., secondary
succession). Overstory harvesting, which also constitutes a loss of nutrients, will be treated in our
discussion of forest harvesting and nutrient loss later in this chapter. In the following section, we
discuss the biological loss of N from forest ecosystems through the process of denitrification.
DENITRIFICATION
Nitrogen is the only plant nutrient that enters terrestrial ecosystems through a biological process.
It is also one of the few plant nutrients that leaves terrestrial ecosystems via a biological process.
Denitrification is the microbially mediated reduction of NO3− to nitrous oxide (N2O) or N2, which
returns N to the atmosphere (Tiedje 1988). Denitrification results in the loss of a limiting nutrient,
potentially influencing the productivity of terrestrial ecosystems. Moreover, N2O produced during
denitrification is a “greenhouse gas” and its concentration in the atmosphere continues to rise
(Elkins and Rosen 1989). Consequently, atmospheric scientists, ecosystem ecologists, and microbial
ecologists have become increasingly interested in understanding the environmental and biological
controls on this process (Davidson and Swank 1987).
The denitrification pathway is used by soil bacteria (facultative anaerobes) as an alternative
respiratory pathway to generate ATP when O2 is absent in the soil atmosphere. Species within 13
genera of bacteria have the ability to use NO3− for respiration. Those common in soil (Pseudo-
monas, Bacillus, and Alcaligenes) carry out the following overall reaction:
5CH 2O 4 H 4 NO3 2N2 5CO2 7H 2O.
The reduction of NO3− is accomplished in several steps (NO3− → NO2− → NO → N2O → N2) that are
mediated by a series of enzymes collectively known as reductases (Knowles 1981). Some organ-
isms lack the enzymatic capacity to carry out the final reductive step (N2O → N2) and produce N2O
NO N2O N2O
Mineralization
+
Biota NH 4 NH–3 N2
Nitrification Denitrification
Immobilization
F I G U R E 1 9 .1 6 Transformation of N by soil microorganisms resulting in the gaseous loss of N from forest

ecosystems. Gaseous loss of N can occur during the process of nitrification as well as during denitrification.
Source: Firestone and Davidson (1989) / John Wiley & Sons. Reprinted from “Microbial basis of NO and N2O
production and consumption in soil” by M. K. Firestone and E. A. Davidson, In M. O. Andreae and D. S.
Schimel (eds.), Exchange of Trace Gasses between Terrestrial Ecosystems and the Atmosphere, © 1989 by John
Wiley & Sons Limited. Reproduced with permission.
as the end product of denitrification. In some situations, N2O can account for one-third of N lost
during denitrification (Robertson and Tiedje 1988). Whether the end product of denitrification is
N2O or N2, the majority of denitrifying bacteria require organic compounds (CH2O) in order to
generate energy. Field techniques for measuring denitrification in terrestrial ecosystems are sum-
marized by Tiedje et al. (1989).
Summarized in Figure 19.16 are the combined sources of gaseous N loss during denitrifica-
tion and nitrification. There is a substantial body of evidence suggesting that gaseous N loss can
also occur during nitrification. Nitric oxide (NO) and N2O are gaseous by-products that are released
by nitrifying bacteria and archaea as they oxidize NH4+ and NO2− (Firestone and Davidson 1989).
Apparently, NO and N2O are chemical intermediates that can dissociate from the NH4+- and
NO3—oxidizing enzyme systems—processes that may be an important source of gaseous N loss in
some forest soils (Robertson and Tiedje 1984).
The physiological requirements for denitrification place considerable restrictions on where
this process occurs in the landscape. Most often, denitrification is associated with poorly drained
soils in which the water table is near or at the soil surface for a considerable portion of the year.
Soils in these landscape positions are often rich in organic matter and experience periodic low
water levels during which nitrifying bacteria can produce NO3− when O2 is present in the soil
atmosphere. Once the soil becomes inundated, O2 diffusion slows and the soil becomes anaerobic.
Denitrification can ensue, provided that organic substrates and NO3− are present.
Groffman and Tiedje (1989) studied denitrification in northern temperate forests occurring
on soils of sand, loam, and clay loam texture. Soils of each textural class were also located in well
drained, moderately well-drained, and poorly drained landscape positions. Denitrification
increased from well-drained to poorly drained landscape positions and was generally higher on
finer-textured soils in which anaerobic conditions are present within soil aggregates. For example,
losses were substantial (40 kg N ha−1 y−1) in lowland forests occurring on poorly drained clay loam
soil, whereas denitrification was minimal in upland forests located on well-drained sandy soils
(<1 kg N ha−1 y−1). The lack of available NO3−, the result of rapid plant uptake, was the primary
factor limiting denitrification during midsummer.
Nitrate uptake by plants also appears to be an important constraint on denitrification in wet
tropical forests. Denitrification is rapid in lowland tropical rainforests that were recently cleared of
vegetation, but only for the first 6 months following harvesting (Robertson and Tiedje 1988). After
6 months, denitrification was substantially lower than that in intact rainforests, apparently the
result of N uptake by rapidly regrowing forest and accelerated decomposition, which lessened
The Cycling and Storage of Nutrients in Forest Ecosystems 527
organic substrate availability to denitrifying bacteria. In combination, these results suggest that
physical factors, such as soil drainage and aeration, directly control the activity of denitrifying
bacteria and archaea. However, denitrification is also controlled by the activities of plants and het-
erotrophic microorganisms that influence NO3− availability in soil, which can dramatically change
following the removal of overstory trees.
THE CYCLING AND STORAGE OF NUTRIENTS IN FOREST ECOSYSTEMS

In the preceding sections, we have discussed the physical, chemical, and biological processes
controlling the flow of nutrients into, within, and out of forest ecosystems. Recall that nutrients
enter forest ecosystems from mineral weathering, atmospheric deposition, and biological N2 fixa-
tion. They cycle within forest ecosystems due to root uptake, aboveground and belowground litter
production, decomposition, and nutrient mineralization. And nutrients are lost from forest ecosys-
tems through leaching and denitrification. Here, we synthesize the specific processes of nutrient
input, internal flow, and loss into overall patterns of nutrient cycling and storage in forest
ecosystems.
NUTRIENT STORAGE IN BOREAL, TEMPERATE, AND TROPICAL FORESTS

The uptake of nutrients by roots, the return of nutrients in aboveground and belowground litter,
and the release of nutrients during decomposition result in the continual movement of nutrients
between plants and soil, albeit at different rates depending on the particular ecosystem. It is a
widely held misconception that overstory trees contain the majority of nutrients within wet
tropical forests, and that the soils beneath them are relatively poor in organic matter and nutrients.
The origin of this misconception is obscure, but our current understanding of tropical forests
clearly refutes this notion. For example, the organic matter and nutrient content of many tropical
forest soils are equivalent to that of temperate forest soils (Klinge and Rodrigues 1968a,b; San-
chez 1976). In some instances, as much as 60% of N and P in wet tropical forests resides in the
organic matter-rich surface and subsurface horizons of Spodosols (Klinge and Rodrigues 1968a,b).
What differs between temperate and tropical forests is the rate at which nutrients cycle from soil to
plants and back to soil again. Recall the rapid mean residence time of organic matter and nutrients
in the forest floor of some tropical forests (Table 19.6). Declines in soil fertility following the har-
vesting of tropical forests are more likely related to a rapid decomposition and loss of nutrients
than to the removal of nutrients in the harvested trees (Cole and Johnson 1980).
In Table 19.9, we summarize the distribution of N within boreal, temperate, and tropical for-
ests. Notice that overstory trees in the wet tropical forest contain 10 times more N than those of the
boreal forest and 3–5 times more N than the overstory of temperate forests. Regardless of this large
difference, comparable quantities of N reside within the soil beneath tropical, temperate, and
boreal forests. For example, mineral soil contains 62% of the N in the wet tropical forest and 69–84%
of the N within the temperate and boreal forests in Table 19.9. When forest floor and mineral soil
are combined, these proportions increase to 70% of total ecosystem N in the wet tropical forest, and
90% of total ecosystem N in the boreal and temperate forests.
From this analysis, it should be clear that forest management practices that erode or remove
(i.e., windrowing or scalping) nutrient-rich O and A soil horizons have the potential to greatly
reduce the nutrient capital of forest ecosystems, regardless of whether they occur in boreal,
temperate, or tropical regions. Recall the rates of nutrient input from atmospheric deposition and
mineral weathering earlier in this chapter (Figure 19.3 and Table 19.1) and realize that substantial
reductions in forest floor and mineral soil nutrient pools can only be replenished over relatively
long periods of time. Also realize that the loss of organic matter-rich surface horizons reduces rates
of nutrient mineralization.
Table 19.9 Distribution of N in tropical temperate, and boreal forest ecosystems.
Boreal Temperate Wet tropical

Location Alaska, USA Washington, New Hampshire, Tennessee, Amazon,
USA USA USA Venezuela
Dominant Black spruce Douglas-fir sugar Oak–hickory Mixed species
species maple–beech
Age (y) 55 42 55 30–80 Mature
Nitrogen kg N ha−1
Overstory 134 316 491 497 1670

Understory 51 21 9 – –
Forest floor 657 233 1100 334 406
Mineral soil 2200 2476 3600 4500 3507
Total
ecosystem 3042 3046 5200 5331 5583
The amount of N in the overstory (leaves, branches, stems, large roots) increases from boreal to tropical forests, but the largest
amount of N resides within the forest floor and mineral soil of these markedly different forest ecosystems.
Source: Adapted from Cole and Rapp (1981), Jordan et al. (1982), and Likens et al. (1977).
THE NITROGEN AND CALCIUM CYCLE OF A TEMPERATE FOREST ECOSYSTEM

The complete N cycle of a second-growth northern hardwood forest is depicted in Figure 19.17a.
In this diagram, boxes represent the storage of N in ecosystem pools (kg N ha−1) and the arrows
depict processes (kg N ha−1 y−1) by which nutrients flow from one ecosystem pool to another. The
N cycle of this forest ecosystem is dominated by storage within soil (3600 kg N ha−1), forest floor
(1100 kg N ha−1), and overstory trees (532 kg N ha−1). Look at Figure 19.17a and compare nutrients
stored in overstory trees, forest floor, and mineral soil with rates of aboveground litter production
(54.2 kg N ha−1 y−1), belowground (6.2 kg N ha−1 y−1) litter production, and net N mineralization
(69.6 kg N ha−1 y−1). Notice that relatively small amounts of N are annually transferred from one
ecosystem pool to another, compared to amounts stored in trees, forest floor, and mineral soil. Also
notice that the total input of N is greater than the total loss of N, indicating an overall retention of
N in this rapidly growing forest ecosystem.
The Ca cycle of the same northern hardwood ecosystem is depicted in Figure 19.17b; it is
similar to and different from the N cycle in several important aspects. The Ca cycle is dominated
by storage in mineral soil, forest floor, and overstory biomass with relatively small amounts of Ca
annually flowing between these pools, similar to the pattern of N cycling and storage. However, the
relatively large amounts of Ca stored in soil occur in the chemical structure of inorganic minerals,
whereas N is stored in soil organic matter, not soil minerals. The large mineral pool of Ca provides
weathering inputs (21.1 kg Ca ha−1 y−1) that are relatively large. Also note that Ca translocation
(arrow from canopy to stem) during leaf senescence is not an important process; in fact, the Ca
increases in leaves prior to abscission. Contrast this with the large amount of N translocated from
leaves prior to senescence. Also contrast patterns of nutrient input and loss for the Ca cycle with
that of the N cycle where losses are much less than inputs. Calcium losses are six times greater
than Ca inputs from atmospheric deposition, suggesting that Ca is not being retained by this eco-
system (compare panels a and b of Figure 19.17).
The Cycling and Storage of Nutrients in Forest Ecosystems 529
(a)
NITROGEN CYCLE OF A
NORTHERN HARDWOOD FOREST ECOSYSTEM
ABOVEGROUND INPUT
LIVING BIOMASS N
251 (4.8) Inorganic OL IMPACTIO
fraction ROS N?
AE IP
THROUGHFALL 6.5 K PREC ITATION
AND STEMFLOW UL 6.5
IO
9.3
LITTER
B FIXAT N 14.2
N2
39.8 FALL
54.2
Inorganic Organic
fraction fraction
6.6 2.7
BIOSPHERE
BELOWGROUND
UPTAKE LIVING BIOMASS N
79.6 181 (4.2) FOREST
FLOOR N
ROOT LITTER 100 (7.7) DI
6.2 NI
TRI 9.0
HY F I CAT I O N
Inorganic Organic ? Organic DR 4. 0
fraction ROOT fraction fraction O LO RT
0.1 GIC E X P O
Available 0.8 EXUDATES 0.1
0.9
Soil N
8 NET
MINERALIZATIONS MINERAL OUTPUT
69.6 SOIL N
3600
Dissolved inorganic fraction
3.9
(b) CALCIUM CYCLE OF A

NORTHERN HARDWOOD FOREST ECOSYSTEM
Inorganic
fraction
ABOVEGROUND 2.2
LIVING BIOMASS Ca INPUT
383 (5.4)
LITTER
FALL
–5.1 40.7
THROUGHFALL
AND STEMFLOW RECIPIT
6.7 LK P 2.2 ATION
BU
UPTAKE BELOWGROUND
62.2 LIVING BIOMASS Ca BIOSPHERE
101 (2.7)
FOREST
FLOOR Ca
370 (1.4) HY
ROOT LITTER DRO RT
LIC EXPO
3.2 13.9
Inorganic ROOT Organic Organic
fraction EXUDATES fraction fraction
Available <0.1
Soil Ca 3.5 3.5 0.1 ? OUTPUT
Inorganic
510 fraction
NET MINERALIZATION (particulate)
42.2 0.2
MINERALSOIL Ca
9600
WEATHERING ROCK
21.1 64,600
Dissolved inorganic fraction
13.7
F I G U R E 1 9 .1 7 (a) The nitrogen and (b) calcium cycles of a second-growth northern hardwood forest at the
Hubbard Brook Experimental Forest. Boxes in each diagram represent ecosystem pools in which N or Ca is
stored, and the arrows represent processes by which nutrients are transferred from one ecosystem pool to
another. Nutrient pools are measured in kg ha−1 and the processes of transfer between pools are presented in
kg ha−1 y−1. Source: (a) Bormann et al. (1977) / American Association for the Advancement of Science.
Reprinted with permission from Science 196, p. 982. © 1977 by American Association for the Advancement
of Science. (b) Likens et al. (1977) / Springer Nature, Biogeochemistry of a Forested Ecosystem, © Springer-
Verlag Berlin, Heidelberg 1977. Reprinted with permission of Springer-Verlag, New York, Inc.
Why is N retained by this forest ecosystem, and why is there a net export of Ca? Recall that
biomass accumulates early in ecosystem development, because net ecosystem productivity (NEP)
is positive (NEP > 0, when GPP > RA + RH; Chapter 18). Growth-limiting nutrients such as N are
rapidly taken up to build new biomass, whereas non-limiting nutrients such as Ca can accumulate
in soil solution—their supply exceeding the biological demand to build new biomass. A positive
NEP and the resulting accumulation of biomass within this forest ecosystem (Gosz et al. 1978),
combined with the fact that N limits the rate of biomass accumulation and Ca does not, cause the
net retention of N and a net export of Ca from this developing forest ecosystem. In the following
section, we build upon these principles to understand patterns of nutrient retention and loss dur-
ing ecosystem development.
ECOSYSTEM C BALANCE AND THE RETENTION AND LOSS OF NUTRIENTS

In 1975, Peter Vitousek and William Reiners proposed and tested a hypothesis integrating patterns
of nutrient loss and retention with the C balance of forest ecosystems (Vitousek and Reiners 1975).
It was based on the idea that patterns of NEP control nutrient loss and retention in forest ecosys-
tems. Notice in Figure 19.18a that net ecosystem production is low or negative during the initial
stages of succession, increases during the intermediate stages of succession, and declines in the
(a)
Primary Approach to Secondary
Succession Steady State Succession
Destructive
+ Event
Net Ecosystem
(Mg•ha–1•y–1)
Productivity
(b)
Input = Loss
Elemental Outputs
Non-Limiting Nonessential
(kg•ha–1•y–1)
Nutrient Element
Input
Rate
Limiting
Nutrient
0
Time
F I G U R E 1 9 .1 8 The conceptual relationship between net ecosystem productivity (NEP) and nutrient loss
from a terrestrial ecosystem during primary and secondary successions. The upper panel depicts changes
in NEP over time and the lower panel illustrates the corresponding changes in nutrient input and loss.
Notice that nutrient inputs are constant over time, whereas the loss of nonessential, essential non-limiting,
and essential limiting nutrients mirrors changes in NEP. Source: Vitousek and Reiners (1975) / Oxford
University Press. Reprinted from BioScience (Vol. 25, No. 6, page 377), © 1975 by the American Institute of
Biological Sciences.
Ecosystem C Balance and the Retention and Loss of Nutrients 531
latter stages of succession (i.e., ecosystem development). In Chapter 18, we saw that patterns of
NEP during ecosystem development result from changes in gross primary productivity (GPP), the
respiration of plants (RA), and heterotrophic organisms (RH). When net ecosystem production is
positive (GPP > RA + RH), biomass and nutrients should accumulate within ecosystems, because
they are needed to build the accumulating biomass.
Vitousek and Reiners (1975) reasoned that the loss of growth-limiting nutrients should be
greatest early or late in succession when NEP is low and there is little or no demand for nutrients.
Examine Figure 19.18a,b and observe the proposed inverse relationship between NEP and the loss
of limiting nutrients. When NEP is less than zero, total ecosystem biomass declines (not shown in
Figure 19.18), and the loss of limiting nutrients should exceed their input. When NEP is zero, total
ecosystem biomass is at an equilibrium, and the loss of limiting nutrients should equal their input.
And, when NEP is greater than zero, total ecosystem biomass increases, and the loss of limiting
nutrients should be less than their input. That is, limiting nutrients should be retained in ecosys-
tems in which NEP is positive and biomass is accumulating. They also predicted that these same
patterns should hold for non-limiting nutrients, but to a much smaller extent (Figure 19.18a,b).
Because nonessential nutrients are not required to build plant biomass, changes in NEP during
ecosystem development should have little influence on their retention or loss (Figure 19.18a,b).
To test this hypothesis, Vitousek and Reiners (1975) located a series of bedrock-sealed water-
sheds that were entirely dominated by mid-(high NEP) or late-successional (low or 0 NEP) forests.
If their hypothesis was correct, differences in NEP between mid- and late-successional forests
should be reflected in the loss of limiting, non-limiting, and nonessential nutrients. These water-
sheds were located in close proximity to one another and received an equivalent input of nutrients
in precipitation. Because each was sealed by underlying bedrock, nutrients that leached below the
rooting zone could only exit in streamwater. Over a 2-year period, they compared the concentration
of NO3− (limiting nutrient), K+, Mg2+, Ca2+ (non-limiting), and Na+ (nonessential) in the streams
draining the watersheds dominated by the mid-and late-successional forests.
By comparing nutrient concentrations in streamwater, it was clear that NO3− loss from late
successional forests was much greater than NO3− loss from mid-successional forests (Table 19.10
and Figure 19.19), supporting the contention that loss of a growth-limiting nutrient is low in eco-
systems where NEP is positive. The export of non-limiting nutrients (K+, Mg2+, Ca2+) was slightly
greater from late-successional forests, whereas the loss of Na+, a nonessential nutrient, did not
differ between mid-and late-successional forests. These results support the idea that the C balance
of terrestrial ecosystems exerts a substantial influence on the cycling and storage of nutrients.
Table 19.10 The mean concentration of limiting, non-limiting, and nonessential plant nutrients in the
streamwater draining watersheds dominated by mid-and late-successional forests in New Hampshire.
Streamwater concentration μeq L−1

Nutrient
Late-successional Mid-successional Late:mid ratio
NO3− 53 8 6.6
K +
13 7 1.8
Mg2+ 40 24 1.7
Ca 2+
56 36 1.6
Na +
29 28 1.0
Source: Vitousek and Reiners (1975) / Oxford University Press. Reprinted from BioScience (Vol. 25, No. 6, page 378), 1975 by
the American Institute of Biological Sciences.
70
60
Nitrate Concentrations (µEq/Liter)
50
40 Late-Successional
Forest
Mid-Successional
30 Forest
20
10
0
J J A S O N D J F M A M J J A S O
1973 1974
F I G U R E 1 9 .1 9 Streamwater nitrate (NO3−) concentrations from watersheds dominated by mid-and

late-successional forests in the White Mountains of New Hampshire. Source: After Vitousek and Reiners
(1975) / Oxford University Press. Reprinted from BioScience (Vol. 25, No. 6, page 379), © 1975 by the
American Institute of Biological Sciences.
FOREST HARVESTING AND NUTRIENT LOSS

Following the removal of overstory trees, there are several fundamental changes that drastically
alter the C balance of forest ecosystems and, consequently, patterns of nutrient loss. Gross primary
productivity immediately declines after overstory removal, because there is little leaf area to con-
vert solar radiation into plant biomass. Greater amounts of solar radiation reaching the soil warm
its surface and increase the rates of organic matter decomposition. Large declines in transpiration
increase soil water contents, which also increases organic matter decomposition and the return of
CO2 to the atmosphere. These changes in ecosystem C balance are illustrated in Figure 19.18 as the
negative values of NEP (i.e., GPP < RA + RH) following a disturbance that initiates secondary
succession. Low rates of nutrient uptake by overstory trees, in combination with a greater nutrient
supply from accelerated rates of organic matter decomposition, may lead to the loss of nutrients
from forest ecosystems.
One of the first and probably most well-known attempts to quantify the influence of forest
harvesting on nutrient loss occurred at the Hubbard Brook Experimental Forest in the White
Mountains of New Hampshire (Likens et al. 1977; Bormann and Likens 1979). To quantify nutrient
loss, scientists located a series of similar watersheds (i.e., size, topography, geology, forest vegeta-
tion, and harvesting history) that were drained by a single stream. Nutrients leached below the
rooting zone encounter shallow bedrock and flow laterally toward the stream draining each water-
shed. For a period of several years (1957–1965), researchers measured the volume and nutrient
concentration of streamwater draining two watersheds dominated by 55-year-old northern hard-
wood forests. Notice in Figure 19.20 that precipitation, streamflow, evapotranspiration, and
nutrient concentrations are nearly identical during this period. In 1965, however, the overstory of
one watershed was clear-cut and herbicide was applied for 3 subsequent years to preclude revege-
tation (Figure 19.20). This treatment resulted in a decrease in evapotranspiration, an increase in
streamflow, and the greater export of NO3−, K+, and Ca2+ (compare control versus clear-cut in
Forest Harvesting and Nutrient Loss 533
Control
(a) (b) Clearcut
Clear-cut Clear-cut
50 Precipitation 6 Calcium
40 4
30 2
0 0
Concentration, mg·L–1
Centimeters of Water
40 Steamflow 3 Potassium
30 2
20 1
10 0
0
50 Nitrate
40 40
30 30
20 20
10 Evapotranspiration (P-S) 10
0 0
1957 1959 1961 1963 1965 1967 1963-64 1965-66 1967-68
Summer Water-Year
F I G U R E 1 9 .2 0 Short-term changes in the flow of water and nutrients through intact and clear-cut
northern hardwood forest ecosystems at the Hubbard Brook Experimental Forest in New Hampshire. The
shaded portion of the figure indicates the period of time during which regeneration was prevented with the
use of herbicide. Notice that harvesting increases streamflow and the export of nutrients from the harvested
watershed. Source: Bormann and Likens (1979) / Springer Nature. Reprinted from Pattern and Process in a
Forested Ecosystem, © 1979 by Springer-Verlag New York Inc. Reprinted with permission of Springer-Verlag,
New York, Inc.
Figure 19.20). Related studies demonstrated that clear-cut harvest also increased denitrification
(Melillo et al. 1983). As plants recolonized the clear-cut site, NEP began to increase, biomass accu-
mulated on the site, and streamwater nutrient concentrations returned to preharvest concentra-
tions (Figure 19.21). This example illustrates the importance of NEP in regulating the loss of
nutrients from clear-cut forest ecosystems.
It is important to point out, however, that not all forest management is conducted in such a
manner nor are these northern hardwood forests typical of all temperate forests. Silvicultural sys-
tems ranging from single-tree selection to whole-tree harvest, soils of broadly different fertility,
and differences of the balance between precipitation and evapotranspiration (i.e., excess water) all
contribute to variation in nutrient loss from managed forest ecosystems. Moreover, not all
temperate forests experience large losses of NO3− following disturbance. In the southeastern
United States, for example, NO3− loss from intensively harvested loblolly pine plantations was
reduced by the addition of logging slash to the forest floor, which increased net N immobilization
by soil microbial communities and fostered N retention (Vitousek and Matson 1984). Temperate
forests prone to NO3− loss are generally those with high rates of net N mineralization and nitrifica-
tion prior to disturbance (Vitousek et al. 1982).
In addition to the export of nutrients to surface waters and the atmosphere, nutrients can
also be lost from forest ecosystems in harvested biomass. One approach for quantifying the impact
for forest harvesting on nutrient cycles uses a “balance sheet” to keep track of all nutrient inputs
and losses (Silkworth and Grigal 1982). Using this technique, one can estimate the influence of
Clear-cut
400
Annual Weighted Concentration
300 Ca2+ Clear-cut

200 Control
100
0
(mEq·L–1)
800
600 NO3–
400
200
0
64 66 68 70 72 74 76 78 80 82 84
Water Year
Long-term trends in NO3− and Ca2+ loss from intact and clear-cut northern hardwood forest
F I G U R E 1 9 .2 1
ecosystems. The shaded portion of the figure indicates the period of time during which regeneration was
prevented with the use of herbicide. The decline in nutrient loss from the clear-cut forest coincides with the
accumulation of biomass (i.e., GPP > RA + RH, therefore NEP > 0) and a renewed demand for nutrients.
Source: Likens et al. (1978) / American Association for the Advancement of Science. Reprinted with
permission from Science 199, p. 493, © 1978 by American Association for the Advancement of Science.
repeated harvesting on nutrient storage in forest ecosystems. If the loss of nutrients during
harvest is not met by nutrient additions during the next rotation, then forest harvesting can cause
an overall decline in nutrient storage. The time it would take to replenish harvest-associated
nutrient losses can be greatly reduced for some nutrients by altering the amount and type of bio-
mass removed from forests. The leaves of trembling aspen are relatively rich in N and P and leav-
ing them on the site dramatically reduces harvest losses. Similarly, trembling aspen has Ca-rich
bark, and leaving this biomass component on site can reduce the net export of Ca from harvest-
ing. This example illustrates the need to assess nutrient inputs and harvest-associated exports to
assess the long-term influence of harvesting on the storage of nutrients within forest ecosystems.
SUGGESTED
READINGS
Anderson, J.M. and Swift, M.J. (1983). Decomposition in Roy, J. and Gamier, E. (1994). A Whole Plant Perspective
tropical forests. In: Tropical Rain Forest: Ecology and on Carbon-Nitrogen Interactions. The Hague, The
Management (ed. S.L. Sutton, T.C. Whitmore and Netherlands: SPB Academic Pub. bv 313 pp.
A.C. Chadwick). Oxford: Blackwell 498 pp. Schlesinger, W.H. (1991). Biogeochemistry: An Analysis
Bormann, F.H. and Likens, G.E. (1979). Pattern and of Global Change. New York: Academic Press 443 pp.
Process in a Forested Ecosystem. New York: Springer- Staaf, H. and Berg, B. (1982). Accumulation and release
Verlag 253 pp. of plant nutrients in decomposing scots pine needle
Hedin, L.O., Armesto, J.J., and Johnson, A.H. (1995). litter: long-term decomposition in a scots pine forest
Patterns of nutrient loss from unpolluted, old-growth II. Can. J. Bot. 60: 1561–1568.
temperate forests: evaluation of biogeochemical the- Vitousek, P.M. and Sanford, R.L. Jr. (1986). Nutrient
ory. Ecology 76: 493–509. cycling in moist tropical forests. Ann. Rev. Ecol. Syst.
Likens, G.E., Bormann, F.G., Pierce, R.S. et al. (1977). 17: 137–167.
Biogeochemistry of a Forested Watershed. New York: Vogt, K.A., Grier, C.C., and Vogt, D.J. (1986). Produc-
Springer-Verlag 146 pp. tion, turnover, and nutrient dynamics of above- and
Paul, E.A. and Clark, F.E. (1996). Soil Microbiology and belowground detritus in world forests. Adv. Ecol. Res.
Biochemistry, 2e. New York: Academic Press 340 pp. 15: 303–377.
Forests of the Future PA R T 6
T he intentional and unintentional influences of humans on forest ecosystems

cannot be denied. These influences only intensify as the human population con-
tinues to increase. Notably, many of the issues faced by forests today are unprecedented in
human history, such as climate change, or the broad scale loss of diversity, or wholesale
reconfiguration of landscapes, or the addition of new species to areas where they become
invasive. Earlier editions of this textbook focused on renewable forest resources and the
dangers of overexploitation or unwise management as the most pressing issues faced by
forests. Over the course of only 60 years since the first edition of this book, the issues,
though still largely anthropogenic in nature, have become far larger and more daunting.
In this part, we explore the ecological concepts related to some of the major human
influences on forests in the twenty-first century. Climate change is perhaps the most
prevailing environmental issue faced by forests and most other ecosystems. We examine
the effects of climate change on forests in Chapter 20 by first summarizing the changes
in temperature and precipitation patterns over the last century, and then exploring the
potential effects of these changes on tree growth, mortality, phenology, and regeneration.
The potential for changes in tree species and forest type distributions is studied, as are the
effects of climate change on disturbances and forest carbon storage. Finally, we address
climate change adaptation for forests.
A second enormous human influence on forests, the introduction of invasive plants
and animals, is addressed in Chapter 21. We first examine the traits characteristic of
invasive species. We then consider the potential impacts of invasive species on forests,
including their effects on competition, fire regimes, and carbon and nutrient cycling.
Finally, a short description of basic invasive species mitigation is presented.
The final two chapters of this book are not focused on specific environmental issues
for forests, but rather review ecological subdisciplines that explore such issues in various
ways. In Chapter 22, we summarize landscape ecology—the study of spatial patterns and
heterogeneity of landscape features—as it pertains to forests. Landscape ecology is rooted
in broad-scale, human-centered environmental issues and has developed its own ecolog-
ical theory and analyses to study them. A selection of landscape ecological concepts is
535
536 PA RT 6 Forests of the Future
presented for forests, including forest fragmentation and connectivity, the interactions
of disturbance and heterogeneity, historical range of variability, and the interactions of
landscape pattern and ecological processes. As a conclusion, the basis for understanding
and addressing sustainability in forests is presented in Chapter 23, including a review of
the concept of ecosystem services. We end the chapter with a perspective for humans and
their relationship to ecosystems, with the hope that this age-old question has become
somewhat clearer (or at least more compelling) throughout the pages of this book.
Climate Change and Forest
CHAPTER 20
Ecosystems
A s we have seen throughout the chapters of this book, climate is the overarching driver of
forest ecosystems on the Earth. Because it acts on specific physiological mechanisms of plants
and other organisms, climate is a major factor determining genetic differentiation and speciation
(Chapter 3), species distributions (Chapters 7 and 15), competition (Chapter 13), disturbance
regimes (Chapters 10 and 16), and growth rates and carbon balance (Chapter 18). The distribution
of vegetation at multiple scales is largely determined by variation in climate with latitude, elevation,
and proximity to large water bodies and mountain ranges (Chapter 2). Characteristic temperature
and precipitation patterns, as they interact with vegetation, parent materials, and physiographic
position, are important in determining soil processes and soil development (Chapter 9). It there-
fore stands to reason that changes in climate, whether natural or anthropogenic, will alter the dis-
tribution of forests and their future productivity, much as they have altered the spatial distribution
and species composition of forest ecosystems in the past (Chapter 15). In the last edition of this
textbook, we posited that we did not know the likely biological consequences of changes in cli-
mate, and instead suggested likely changes in plant distributions and species migrations based on
forest responses to past glaciations. Now that nearly 25 years have passed, we have sufficient data
and improved technology to predict those consequences with more confidence.
Climate change is an enormously broad topic, and the field of climate change biology is both
emerging and constantly evolving. We present in this chapter selected elements of climate change
biology relevant to forest ecosystems. We first provide an overview of climate history and the
natural and anthropogenic causes of variability, and we summarize our current understanding of
climate trends. We then proceed with a summary of climate change impacts on tree physiology,
growth, and phenology; potential tree population responses including range shifts; effects on dis-
turbances and succession; and impacts on carbon storage and its feedback to climate change. It is
challenging to present even an adequate sampling of climate change impacts on forests in a single
chapter; recent textbooks such as those by Hannah (2021), Post (2013), and Newman et al. (2011)
provide a comprehensive background on climate effects on biota and ecosystems. We provide
many examples in the following text from the United States and North America, rather than at
broader spatial scales, as an attempt to distill the extensive data now available.
CLIMATE CHANGE CONCEPTS

In describing climate change we must first distinguish between “climate” and “weather.” Climate
is the long-term trend in weather conditions of a particular region of the Earth. A relatively long
timescale—the rule of thumb is typically a minimum of 30 years—is necessary to reliably deter-
mine a climate trend, because shorter timescales are subject to enough variability that trends are
difficult to ascertain. This is why a single, very dry summer is no more indicative of climate change,
or lack thereof, than a single, very warm winter. Such conditions are more appropriately referred
to as weather conditions rather than climate.
537
538 Chapter 20 Climate Change and Forest Ecosystems
The basis of climate change is abiotic, the physics of which hinge on the prevalence of water
vapor and carbon dioxide (CO2) in the Earth’s atmosphere. A process known popularly as the
“greenhouse effect” occurs because both gases are transparent to visible light from the sun and
also trap heat. As sunlight warms the surface of the Earth, long-wave radiation is emitted back
from the surface, some of which is absorbed by water vapor and CO2. The absorbed radiation is
reemitted by the gas molecules, and some of this reemitted radiation is directed back toward the
Earth, warming the lower atmosphere. The relationship between these “greenhouse gases” and
global temperature is no longer hypothesized but observed in retrospect using ice-core data (see
Chapter 15 for description of ice-core data). Carbon dioxide is the best-studied greenhouse gas and
is our focus in this chapter, although others, such as methane (CH4), also exist and have important
relationships to land use and ecological systems. There has been a consistent and repeated occur-
rence of warm periods of climate over at least the last 800 000 years that correspond with high
levels of atmospheric CO2 (Figure 20.1). Temperature and atmospheric CO2 are intricately related
in a positive feedback in the Earth’s climate system: increasing CO2 causes rising temperatures due
to the greenhouse effect, and increasing temperature releases stored CO2 from the oceans into the
atmosphere.
The long-term variability in temperature and CO2 over the past 800 000 years is certainly not
the result of human activities, but due to natural phenomena such as variations in the Earth’s tilt
and wobble on its axis (Milankovitch cycles), solar intensity, volcanic eruptions, and natural var-
iations in greenhouse gas concentrations, all of which have occurred over very long temporal
scales. Many of these natural phenomena are thought to be the drivers of past glacial and inter-
glacial periods. Virtually all climate scientists today acknowledge that modern atmospheric CO2
concentrations and global temperatures have increased at unprecedented rates since the middle
of the nineteenth century, and that natural causes alone cannot explain these changes. Instead, it
is human activity—primarily the burning of fossil fuels and deforestation, both of which trans-
form sequestered carbon at or beneath the Earth’s surface into atmospheric carbon—that best
explains the rise of surface temperatures in the last 150 years (Figure 20.2). Meehl et al. (2004)
used hindcasting climate models to show that only models that incorporated both natural and
anthropogenic forcings could reproduce the rate and degree of global temperature change over
the twentieth century. The 2021 report of the Intergovernmental Panel on Climate Change has
Carbon Dioxide (parts per million)

300
250
200
800 000 600 000 400 000 200 000

Antarctic Temperature (°C)
5
0
–5
–10
800 000 600 000 400 000 200 000 0
Date (Years Before Present)
F I G U R E 2 0 . 1 Variation in temperature and atmospheric CO2 concentration over the last 800 000 years in
Antarctica. Temperature and CO2 have varied in concert from at least 800 millennia. Data are from the
European Project for Ice Coring in Antarctica (EPICA) Dome C ice core in Antarctica; CO2 was determined
from air bubbles in the ice core. Source: Data from Jouzel et al. (2007) and Lüthi et al. (2008); graphic by
Robert Simmon, NASA Earth Observatory.
Climate Change Concepts 539
(a) Observed
All Natural Influences
2.0
1.5
1.0 All Natural

Drivers
0.5
Volcanic
Solar
0.0
Orbital
–0.5
–1.0
1880 1900 1920 1940 1960 1980 2000 2020
(b)
Temperature Difference from Average (°F)
Observed
2.0 All Human Influences
Greenhouse Gases
1.5
1.0
All Human
0.5 Drivers
Ozone
0.0
Land Cover
–0.5
Aerosols
–1.0
1880 1900 1920 1940 1960 1980 2000 2020
(c) Observed
All Human and Natural Influences
2.0
1.5
1.0 All Drivers
0.5
0.0
–0.5
–1.0
1880 1900 1920 1940 1960 1980 2000 2020
F I G U R E 2 0 . 2 Global temperature changes due to (a) natural forcing, (b) human forcing, and (c) both
natural and human forcings. Using climate models (colored lines) to reproduce observed changes in global
temperatures (black lines), natural phenomena such as volcanic eruptions, solar activity, and variations in
the Earth’s orbit show no long-term trend in global temperatures over the last 140 years. While human
influences such as aerosols and changes in land cover have had a net cooling effect over the last 80 years,
most of the long-term warming trend in global temperatures is explained by increases in greenhouse gases.
Combining both natural and human factors in the climate model matches the warming trend closely since
1950, suggesting the dominant role of human influences on global temperature changes. Source: From
Hayhoe et al. (2018) / U.S. Government Printing Office / Public Domain.
declared that the increase of CO2 and other greenhouse gases since 1850 is the result of human
activities (IPCC 2021). Notably, atmospheric CO2 concentration, at 421 ppm at the time of this
writing, only in 1950 exceeded 300 ppm after being less than that for hundreds of thou-
sands of years.
A discussion of climate change is often distilled to one of temperature change because it is
among the most easily measured and observed factors of the Earth’s climate system. Though the
popular term “global warming” is often used interchangeably with climate change, it oversimplifies
a changing climate in a complex system such as the Earth because it suggests that temperature is
the only factor affected and that it only changes in a specific way (Post 2013). Instead, climate
change is most usefully considered to include all of the abiotic changes that have occurred over
the last 150 years. Obvious indicators of climate change include rising and more variable sur-
face temperatures, changes in precipitation and snow cover, and changes in ice cover on sea and
land. The first two of these indicators, changes in temperature and precipitation and their corol-
laries, are the main emphasis of this chapter.
EFFECTS ON TEMPERATURE
Temperature has many direct effects on trees, and many of these effects were reviewed in Chapter 7.
Thus, changes in climate that affect temperature are extremely relevant for forests. The surface of
the Earth has increased in temperature by 1.09 °C between about 1850 and 2020 (IPCC 2021). This
estimate suggests a rate of increase in temperature higher than even that suggested in the last
edition of this textbook, which reported a warming of 0.5 °C since 1800 and an increase of 0.27–0.39 °C
since 1900 based on estimates by Henry et al. (1994). Remarkably, the period 2010–2019 was likely
the warmest decade over the last 125 000 years, and the period of 2016–2020 was the hottest 5-year
period since 1850. The majority of the warming is attributed to human activities, primarily
greenhouse gas emissions, with only a small fraction (−0.1 °C to +0.1 °C) due to natural forcing
(IPCC 2021). Hansen et al. (2006) reported a temperature increase of 0.2 °C per decade since 1980;
despite a slower period of warming between 1998 and 2012 relative to the periods preceding and
following it, the Earth’s surface temperature appears to be warming at an increasing rate. In the
United States, annual average temperature increased by 0.7 °C between 1901 and 2016. This
warming accelerated between 1976 and 2016, with recent decades the warmest in 1500 years (Hay-
hoe et al. 2018).
The Earth’s warming contains a great deal of variation, among years, decades, centuries,
and millennia (Figure 20.2), as well as spatially. Some parts of the United States have experienced
general warming trends while others have not (Figure 20.3). Between 1901 and 2016, the largest
increases in temperature occurred in the western half of the United States, particularly in Alaska,
the Northwest, the Southwest, and the northern Great Plains, where average temperature
increased by more than 0.8 °C (1.5 °F; Hayhoe et al. 2018). However, much of the southeastern
United States has had a neutral or a slight cooling trend during this period, apparently due to a
winter shift in the position of the polar jet stream around 1950 that allowed cold air masses to
more frequently penetrate the region. It is notable, however, that the Southeast has been warming
at an accelerated rate since the early 1960s. This spatial variability in warming trends presents a
significant challenge to understanding how forests will respond to climate change over broad
regional scales (Post 2013).
In addition to long-term trends in surface temperature, there are several other temperature-
related indicators of climate change that are relevant to forest ecosystems. As the degree of warming
varies across years and decades, it also varies seasonally, with some seasons warming more than
others in some regions. In the United States, minimum temperatures have increased faster than
maximum temperatures, and a trend toward warmer winters has been evident since 1896, particu-
larly at night and in the northern parts of the lower 48 states (USGCRP 2017). Average winter
Change in Temperature (°F)
–1 0 1 2 3 4 5 6 7 8
F I G U R E 2 0 . 3 Observed changes in annual average temperature across the United States for the period
1986–2016 relative to 1901–1960 for the contiguous United States and 1925–1960 for Alaska, Hawaii, Puerto
Rico, and the US Virgin Islands. Significant warming has occurred across much of the western United States
and Alaska, though moderate cooling has occurred in the southeastern United States. Note that scale is in
degrees Fahrenheit; for reference, 1° of change in Fahrenheit degrees = 0.56° of change in Celsius degrees.
Source: From Hayhoe et al. (2018) / U.S. Government Printing Office / Public Domain.
temperatures have increased by nearly 1.7 °C (3 °F) across the contiguous 48 states, spring temper-
atures by 1.1 °C (2 °F), and fall temperatures by 0.8 °C (1.4 °F). These seasonal changes also vary
spatially. The largest winter increases occur in the northern states (the Northeast, Lake States,
Montana, and Dakotas) with additional increases in the Intermountain West; winter increases
were smaller or neutral in the Southeast. Most states warmed in the spring, summer, and fall,
except for a few in the Southeast that had little to no overall change or cooled slightly (NOAA 2021a).
As discussed in the following sections, seasonal changes in temperatures may have important
implications for plants in that they affect water supply and usage by plants, and may affect growing
season length and heat sums, which are important in setting limits of some tree distributions.
An additional factor to consider regarding effects of climate change on temperature is its
effect on temperature extremes. Unusually hot temperatures in the summer can affect water usage
by trees, as well as lead to stressed conditions causing increased susceptibility to attack by insects
and diseases, increased frequency or extent of fire or other disturbances, or death. Extreme or
uncharacteristic hot or cold periods are a natural part of weather patterns in most regions, but the
frequency and intensity of such periods are likely to change as overall surface temperatures warm.
Across the United States, an increasing trend of unusually hot summer days (defined as those
above the 90th percentile) has occurred since about 1980 (NOAA 2021b). Notably, the frequency of
unusually hot summer nights, indicating a lack of nighttime cooling, has increased at an even
faster rate than that for daytime temperatures, and the occurrence of unusually cold temperatures
(especially nighttime extreme lows) has decreased (Figure 20.4). Unusually hot days have increased
(a)
90
Hot daily highs
80
Hot daily highs (smoothed)
Hot daily lows
70
Hot daily lows (smoothed)
60
Percent of land area
50
40
30
20
10
0
1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010 2020
Year
(b)
90
Cold daily highs

80
Cold daily highs (smoothed)
70 Cold daily lows
Cold daily lows (smoothed)
60
Percent of land area
50
40
30
20
10
0
1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010 2020
Year
F I G U R E 2 0 . 4 Percentage of the land area of the contiguous United States with (a) unusually hot daily high
and low temperatures during June, July, and August, and (b) unusually cold daily high and low temperatures
during December, January, and February. Unusually hot summer days and nights have increased since the
1980s, whereas unusually low winter temperatures (particularly at night) have become less common. Source:
Data from NOAA (2021b); graphic from https://www.epa.gov/climate-indicators.
in frequency along the Gulf and Atlantic Coasts and in the Southwest (including southern
California) since 1948 but decreased in frequency in the Upper and Lower Midwest and along the
Mississippi River (NOAA 2021a). Unusually cold days have become less common across most of
the contiguous United States, particularly in the West.
EFFECTS ON PRECIPITATION
Precipitation is an additional abiotic component vastly affected by climate change, but the effects
of climate change on precipitation are somewhat more difficult to interpret from an ecologically
relevant perspective compared to temperature. Precipitation patterns are strongly influenced by
temperature, as well as the interrelationships among air currents, large water bodies, and topogra-
phy. Precipitation alone does not correlate well with tree distributions or the boundaries of most
major forest communities because evaporation and transpiration are heavily influenced by tem-
perature as well as water. Potential evapotranspiration, or the amount of water that could be
evaporated from land, water, and plant surfaces if soil water was limitless, has been shown to be
more closely related to forest distribution than temperature in many regions (Patric and Black
1968). Although we review in the following text the changes and spatial variation that occur in
precipitation with climate change, we caution that such changes are closely linked to temperature.
As with air temperature (Chapter 7), precipitation varies with latitude and altitude. Along
the Pacific Coast, mean annual precipitation increases sharply with altitude, at rates varying from
13–17 mm per 100 m in the coastal range of Washington to 7–8 mm in the Sierra Nevada. Maximum
precipitation occurs at the middle elevations—ranging from perhaps 900 m in the Olympics to
1500 m in northern California and 2400 m in the southern Sierra Nevada. Above these elevations,
air currents become depleted of moisture and precipitation decreases with further altitude. The
effect of elevation on rainfall is much more predictable in the interior mountains of western North
America, being approximately 3–4 mm per 100 m rise in elevation (range 900–1500 m).
Precipitation also varies longitudinally across North America. The mountain ranges become
progressively drier moving eastward from the Pacific Coast. Globally, forests tend to occur where
precipitation exceeds transpiration, and precipitation is generally too low in the major rain shadow
that characterizes the ranges and prairie regions east of the Rocky Mountains. Precipitation grad-
ually increases further east due to the influence of maritime moisture from the Gulf of Mexico.
High precipitation is found in the Southeast, particularly in the southern Appalachian Mountains,
and in the Northeast, especially in the White Mountains.
Compared to temperature, there are far fewer data and thus more uncertainty in the his-
torical levels of global precipitation, but average precipitation over land appears to have increased
since 1950, with a faster rate of increase since the 1980s (IPCC 2021). Across the United States,
average annual precipitation has increased by about 4% since the beginning of the twentieth
century with marked regional differences (Figure 20.5). Much of the increase in precipitation
has occurred in the northern, eastern, and central parts of the country (including the Great
Plains), while the southeastern and western (particularly southwestern) regions of the country
have experienced notable decreases, particularly in southern California and Arizona. Precipita-
tion has increased at an average rate of 0.25 cm (0.1 in.) per decade globally since 1901; precipi-
tation in the contiguous 48 states has increased at a rate of 0.51 (0.2 in.) per decade (Hayhoe
et al. 2018).
As with temperature, trends in precipitation vary seasonally. Comparing the present day
(1986–2016) to the early twentieth century (1901–1960), the largest increase in precipitation
occurred in the fall for the contiguous United States, where the change is greater than 15% in the
northern Great Plains and most of the eastern portion of the country (Easterling et al. 2017).
Increases in precipitation are smaller in the spring months (3.5%), with the northern half of the
country wetter and the southern half drier. Summer precipitation also increased across the country
(3.5%), with patterns including increases in precipitation in the Upper Midwest, Northeast,
southern Great Plains, and southern California, but drying in the Intermountain West, Southwest,
Southeast, and Mid-Atlantic. Winter precipitation increased the least (2%), with drying over much
of the western United States and Southeast but wetter conditions in the southern Great Plains and
parts of the Lower Midwest (Easterling et al. 2017). Winter drying in the western United States
Change in Precipitation (%)
–30 –25 –20 –15 –10 –5 0 5 10 15 20 25 30
F I G U R E 2 0 . 5 Observed changes in annual average precipitation across the United States for the period
1986–2016 relative to 1901–1960 for the contiguous United States and 1925–1960 for Alaska, Hawaii, Puerto
Rico, and the US Virgin Islands. Increased precipitation has occurred in the Great Plains and Northeast, but
the Southwest has experienced a notable decrease. Source: From Hayhoe et al. (2018) / U.S. Government
Printing Office / Public Domain.
suggests declines in snowpack, particularly in middle to high elevations in mountainous areas and
shifts to more winter precipitation falling as rain rather than snow.
Changes in the type of precipitation falling on some regional ecosystems can have important
implications for the water balance of forests and other ecosystems. If climate change causes a
greater proportion of winter-season precipitation to fall as rain rather than snow in such ecosys-
tems, it will affect the extent and depth of snowpack, which is the amount or thickness of accu-
mulated snow on the ground. In mountainous areas, snowpack is essential to the water balance of
forest ecosystems, as it slowly releases stored water as runoff throughout much of the spring and
summer as the snow melts. Snowpack is also important for insulating trees from freezing temper-
atures and winter wind at high elevations. Earlier melting of snowpack due to warmer winters and
springs has the potential to contribute to soil-water deficits in mid and late summer because of the
short duration of summer rainfall events that are ineffective at recharging the soil profile for uptake
by plants. As such, earlier snowmelt also accelerates weather conditions conducive to wildfire,
particularly in the western United States.
The average amount of snowfall across the country has decreased by about 0.2% each year
since 1930, in part because the proportion of precipitation falling as rain instead of snow has also
increased across a large portion of the contiguous United States. The proportion of rainfall has
increased especially in the Pacific Northwest and the Midwest, though some areas of the Lake
States have experienced a higher proportion of snowfall (Kunkel et al. 2009). April snowpack has
decreased across much of the western United States, especially in the Pacific Northwest and
northern Rocky Mountains. Snowpack has peaked earlier in the year across much of the West,
averaging 8 days earlier since 1982. These patterns are particularly evident in states such as
Colorado, New Mexico, and Utah (USDA-NRCS 2020).
Given the importance of precipitation for forest ecosystems, changes in the variability in
precipitation within and among years are an additionally important aspect of climate change.
Changes in the intensity and frequency of precipitation are characteristic of changing climate
during the last 150 years. Higher surface temperatures, particularly those of the oceans, increase
evaporation, creating heavier rain or snow events when the more moisture-laden air converges
into storm systems. Heavy precipitation suggests that precipitation is occurring in more intense
events and may not necessarily mean that the total amount of precipitation at a location has
increased. The relevance to forests of heavy precipitation events is higher runoff and reduced infil-
tration of water into the soil, resulting in lower soil-moisture levels and potentially higher soil-
water stress. Heavy precipitation events also lead to increases in soil erosion and flooding, the
latter of which can be devastating depending on its timing and duration. Unusually heavy snow
events may result in breakage and damage to trees, particularly where heavy snow is uncharacter-
istic of local or regional macroclimate and species lack adaptations to it.
Globally, heavy precipitation events have increased in frequency and intensity over a majority
of land regions since 1950, with expectations that such increases will continue with further surface
warming (IPCC 2021). In the United States, the frequency and intensity of heavy precipitation
events have increased more than average precipitation (Hayhoe et al. 2018), and the proportion of
precipitation that falls as intense, single-day events has increased since the 1980s and especially
since 1996 (Easterling et al. 2017). Since 1901, the Northeast and the Midwest have increased 38%
and 42%, respectively, in the proportion of total annual precipitation falling in the heaviest 1% of
events. The value for the Northeast increases to 55% when measured from 1958 (Hayhoe et al. 2018).
Increases for the Pacific Northwest and the Southeast were 22% and 18%, respectively, since 1901
and 9% and 27% since 1958.
At the other end of the spectrum from heavy precipitation events are droughts, or prolonged
and abnormal periods when precipitation is lacking, creating water deficiencies for ecological sys-
tems. Droughts are often but not always accompanied by unusually warm conditions. Drought
may affect trees directly by reducing growth during the growing season, or by reducing flowering,
seed production, seed germination, and seedling survival. Indirectly, drought may affect trees by
creating dry conditions that facilitate more frequent ignition, higher intensity, and faster spread of
fires. Drought is also associated with insect attack and diseases in trees because it creates an addi-
tional stressor to trees that may weaken their resistance to these agents. Recurrence of adverse
climatic patterns at a given location may increase the probability of disease or insect attack and
would favor more drought-resistant species on sites with high soil-water stress.
The frequency and intensity of droughts have increased in many areas around the world,
particularly in those that are already drought-prone. It remains difficult to characterize the role of
climate change in the occurrence of droughts, however; some regions of the world, such as
southern Europe and West Africa, have experienced longer and more intense droughts, but others,
such as central North America, have experienced less frequent, less intense, or shorter droughts
since the 1950s (IPCC 2013). This pattern holds true in the United States, which has experienced
wetter than average conditions over the last half-century (Wehner et al. 2017). The most consistent
and common droughts over the last 50 years have been observed in southwestern states (California,
Arizona, and New Mexico), but the eastern United States has been generally wetter. Over the last
two decades (2000–2020), as much as 70% of the land area of the United States experienced abnor-
mally dry conditions at some point; more than half of the country experienced a moderate to
severe drought in 2012 (National Drought Mitigation Center 2021).
CLIMATE CHANGE EFFECTS ON THE FOREST TREE

TREE GROWTH AND MORTALITY
Tree growth is closely linked to climate, and the direct effects of climate change on tree growth
depend on several factors, including water and nutrient availability, increasing atmospheric CO2,
the timing of increases in temperature, the adaptability of species to new climatic conditions, the
trees’ thermal optimum for growth, and potential changes in trees’ competitive ability (Way and
Oren 2010). The temperature regime of a given regional or local ecosystem has particularly strong
implications for plant growth (Chapter 7). The simplest evidence of this relationship comes from
tree rings, which widely show increased growth during warmer growing seasons, especially in
northern forests or at high elevations (Bunn et al. 2005; D’Arrigo et al. 2008). Notably, however, the
relationship between warming and growth quickly degrades after 1950 for many species (Briffa
et al. 1998; Lloyd and Fastie 2002; D’Arrigo et al. 2004). This weakening relationship is thought to
be explained by increasing physiological stress and reduced growth that occurs when trees experi-
ence an increase in temperature beyond a threshold without also increasing moisture availability
(i.e., a drought; D’Arrigo et al. 2008). Examining white spruce at tree line in northwestern Canada,
D’Arrigo et al. (2004) found a positive correlation between tree growth and temperature between
1901 and 1965, but a negative correlation after 1965. A threshold summer temperature of 7.8 °C
was identified as repeatedly occurring after 1965, suggesting that increasing temperatures associ-
ated with climate change below a threshold may promote tree growth, but those beyond the thresh-
old may reduce it. Positive effects of increasing temperatures on tree growth are thought to be
limited to species found in temperate and boreal regions with cooler climates and do not appear to
benefit those found in warmer environments such as the tropics (Way and Oren 2010).
In a meta-analysis of studies that examined effects of temperature on tree growth, Way and
Oren (2010) identified the following trends for trees experiencing higher temperatures: (i) increased
leaf biomass and decreased root mass. This shift to a lower root-to-shoot ratio at higher tempera-
tures likely increases susceptibility to drought stress; (ii) increased height growth and shoot growth
affected mainly by daytime temperatures; (iii) moderate increases in stem growth, affected mainly
by nighttime temperatures; (iv) stronger responses by deciduous species compared to conifers.
Coniferous trees, found often on poor-nutrient sites and whose life history trade-offs include
balancing productivity with nutrient retention, may be less able to increase their growth under
favorable conditions; and (v) higher sensitivity of trees from warmer locales to warming tempera-
tures compared to those from cooler locales, probably because trees at warmer locales typically
persist closer to their temperature threshold or optimum.
The general pattern of tree growth described in the earlier text—an increase in growth with
increasing temperatures to a maximum, followed by a decrease—mirrors those of plants’
physiological processes. Photosynthesis and respiration generally increase to an optimum temper-
ature before they decline, due mainly to quicker enzyme function, until lethal high temperatures
are reached. The same processes that lead to increased growth in a plant may also lead to its being
stressed. For example, the rate of photosynthesis theoretically increases with increasing tempera-
ture (Chapter 18), but the greater vapor pressure deficit of warmer air may cause stomata to close
in response to greater transpiration (Mott and Parkhurst 1991) and/or due to soils dried faster by
increased soil-water uptake by roots (Will et al. 2013), reducing photosynthesis. Despite increasing
photosynthetic rates, rising temperatures associated with limited precipitation or soil moisture
may therefore eventually lead to tree mortality (Allen et al. 2010, 2015). Thus, warming alone
without substantial increases in precipitation may induce tree mortality if warming is severe. This
is because drought in combination with warming only requires a moderate temperature increase
to induce mortality (Figure 20.6) because a lack of (or even small increases in) precipitation does
not compensate for increased vapor pressure deficits. Overall, a warmer climate will decrease tree
growth in forests where water is limiting.
Climate Change Effects on the Forest Tree 547
F I G U R E 2 0 . 6 Conceptual diagram Drought

showing potential shifts of current climate Low Frequency / Intensity / Duration High
and results on tree mortality. Only a small
Warmer
portion of the current climate exceeds the Future
upper or lower threshold for temperature climate
and precipitation to support trees, warming Future
respectively, but more extreme tempera- only climate
drought
tures alone or more droughty conditions
+warming
together with warming may produce
Temperature
greater risk of mortality for current tree
populations. Source: From Allen et al. te
ma
(2015) / John Wiley & Sons. t Cli
en
C urr
Tree
Mortality
Threshold
Cooler Mortality
No Mortality
Wetter Drier
Precipitation
At the same time, photosynthesis and ultimately tree growth increase when trees are grown
in an atmosphere enriched with CO2, as is present in most scenarios of climate change (Ainsworth
and Rogers 2007; Norby and Zak 2011; Dusenge et al. 2019). Notably, trees use water more effi-
ciently as atmospheric CO2 increases because they reduce water loss via stomatal conductance
(Keenan et al. 2013). There may therefore be a “fertilization effect” of increasing CO2 that would
increase tree growth in the presence of elevated CO2. Depending on the strength of such an effect
and the ability of trees to acclimate to rising temperature and CO2, the higher photosynthetic rates
and water use efficiency associated with CO2 fertilization could theoretically counteract the
decreased growth that occurs with warming temperatures (Sperry et al. 2019). It appears likely that
tree growth will increase with future increases in CO2, if only because rising atmospheric carbon
appears to be particularly well correlated with increased water-use efficiency (Adams et al. 2020).
The concept of CO2 fertilization is appealing because it suggests an important negative
feedback. Rising concentrations of CO2 in the atmosphere may increase tree growth and sequestra-
tion of carbon in forest biomass, thereby dampening future rates of increase in atmospheric CO2.
However, experimental tests of CO2 fertilization, most of which have been conducted at free-air
CO2 enrichment (FACE) facilities or within open-top experimental chambers, have often raised
more questions about the wider implications of such an effect. Experimental data have shown that
CO2 fertilization is a short-lived phenomenon when soil nutrients (especially nitrogen) are limit-
ing, particularly in older forests where its effect may otherwise be weak (Reich et al. 2006; Norby
et al. 2010). For example, in a FACE experiment in a sweetgum forest in Tennessee beginning in
1997, increases in growth under elevated CO2 declined from 24% in 2001 to 9% in 2008 (Norby
et al. 2010). When plots were fertilized with nitrogen, tree growth immediately responded posi-
tively and in a sustained manner, even as the CO2-fertilized plots (and non-nitrogen–fertilized
plots) were declining (Figure 20.7). These results suggest that nitrogen availability was limiting to
tree growth and declining over time in a maturing forest whose growth was initially enhanced by
elevated CO2. A similar trend was documented globally by Wang et al. (2020), who used satellite
and ground-based data sets to show that the CO2 fertilization effect declined across the world from
1982 to 2015, likely due to changing nutrient concentrations and soil-water availability.
1.8 N-fertilization treatment

initiated
Wood increment (kg m–2 y–1)
1.6
1.4
1.2
1.0
0.8 CO2 treatment

initiated
0.6
0.4
1996 1998 2000 2002 2004 2006 2008
F I G U R E 2 0 . 7 Tree growth response to CO2 enrichment and nitrogen addition in a sweetgum forest in
Tennessee. Elevated CO2 (solid circles) initially increased growth but declined thereafter and did not differ
from non-enriched plots. Nitrogen fertilization (closed squares) caused an immediate and sustained increase
in growth. Source: From Norby et al. (2010) / with permission of PNAS.
An additional issue with CO2 fertilization effects on tree growth is its potential trade-off with
tree longevity. Utilizing 539 permanent plots sampled between 1960 and 2009 in the boreal forest
of Alberta, Canada, Searle and Chen (2018) found an increase in tree mortality over 50 years, as
well as a correlation between high lifetime growth rate and higher probability of mortality com-
pared to slower growing trees. Reduced longevity was explained largely by decreasing water avail-
ability, to which faster growing or larger trees apparently are more sensitive (Hember et al. 2017).
This trend has been corroborated for many tree species and growing environments across North
and South America, Europe, Australia, and Asia (Brienen et al. 2020). Given that the rate of tree
mortality has increased in recent decades (van Mantgem et al. 2009), the potential trade-off bet-
ween rising atmospheric CO2 and tree longevity is compelling for carbon dynamics in a changing
climate. Fast-growing trees contribute to rapid carbon uptake from the atmosphere, but large, old
trees are critical for maintaining the carbon in biomass rather than the atmosphere. Tree mortality
returns stored carbon to the atmosphere as dead trees decompose over decades, offsetting any CO2
fertilization effect (Bugmann and Bigler 2011), and fertilization-related gains in carbon storage
from faster tree growth will therefore be short-lived if the trees die before they are large enough to
store large amounts of carbon (Körner 2017). Forest productivity and carbon storage are discussed
in detail in Chapter 18, and their role in climate change adaptation is discussed further later in
this chapter.
PHENOLOGY
Trees, like all plants, vary in their seasonal rhythms, and important events such as flowering,
spring bud burst, fall bud set, and seed maturation and dispersal have very regular and distinct
cycles during and among years. Seasonal variations in temperature are important cues for the
timing of these events, known as a species’ phenology, and climate change has the potential to
alter this timing (Cleland et al. 2007). Temperature-related factors such as growing season
length, heat sum, and last spring and first fall frosts direct many phenological activities
(Chapter 7). Many of the same ecological factors and relationships that affect tree species
istribution, such as the relationship between physiography and temperature (Chapter 8) or the
d
spatial relationships of landscape ecosystems, have similar, observable effects on phenology. For
example, Fisher et al. (2006) used satellite data to detect leaf flush phenology in deciduous for-
ests in southern New England. Satellite data were able to quantify variations in phenology with
local topography, and in one example detected a gradient of leaf flush over 2 weeks over a 500-m
distance and a 30-m elevational change on the side of a shallow valley (Figure 20.8). As dis-
cussed in Chapters 7 and 8, cold-air drainage into the valley significantly alters its microclimate,
reducing temperatures to the point that phenology is also affected. Moreover, several studies
have documented that vegetation in urban areas experiences earlier springs leading to earlier
tree flowering and leaf flushing, as well as longer growing seasons, compared to adjacent non-
urban areas due to the warmer microclimate characteristic of cities (Jochner and Menzel 2015).
Such long-standing examples of the link between phenology and temperature should provide a
great deal of foresight into the effects of rising global temperatures on tree phenology that are
likely to occur with climate change.
Many physiological events that occur in the spring have been linked to warmer spring tem-
peratures (IPCC 2014), and thus phenological events have begun to serve as important and effec-
tive indicators of climate change. Perhaps the most well-cited example of climate change effects
on phenology is a study by Schwartz et al. (2006), who studied the first bloom and first leaf dates
of lilacs and honeysuckles around the world as a means of detecting changes in spring onset. First
leaf and first bloom (approximating the onset of bud burst in deciduous forest trees) dates in the
contiguous United States, for example, have occurred earlier in the spring in the last few decades
in the northern and western parts of the country where temperature increases have occurred, but
later in the South where cooling trends have been prevalent (see temperature trends discussed in
the earlier text). First leaf dates averaged 1.2 days earlier per decade across the world for the
period 1955–2002, and first bloom dates averaged 1 day earlier per decade. Notably, the winter
chill date—the date at which lilacs and honeysuckles are ready to respond to spring warmth after
satisfying their winter chilling requirement—is occurring 1.5 days earlier per decade in central
and eastern North America, increasing the length of the growing season (Schwartz et al. 2006).
Studies linking altered tree phenology to altered temperature regimes have become increasingly
common, especially for forests in seasonally cold climates (Gunderson et al. 2012; Gill et al. 2015;
Richardson et al. 2018).
The phenology of species and individuals—even those co-occurring—does not necessarily
respond to rising temperatures in the same way. For example, leaf flushing and flowering of
co-occurring species appear to become less synchronous in warm, early springs (Zohner et al. 2018;
Montgomery et al. 2020), probably because species differ in how they respond to early onset of
warm spring temperatures. Co-occurring species with differing heat sums (Chapter 7) exhibit dif-
ferent sensitivities to warm springs and thus their timing of bud burst respond differently to earlier
spring onset. There is also evidence that bud burst in some tree species is particularly cued by their
winter chilling requirement (Zohner et al. 2016; Nanninga et al. 2017) or by photoperiod (Basler
and Körner 2012) even more so than heat sums, and these species may respond more sluggishly to
earlier springs (Montgomery et al. 2020). Cooler, later springs allow most species to meet their
chilling and/or photoperiodic requirements, and such species are therefore more likely to be
synchronized in their bud burst. Synchrony in leaf phenology is important for forest trees because
those that exhibit early bud burst have a higher risk of damage due to late frosts (Kollas et al. 2014;
Vitasse et al. 2014), which could facilitate shifts in species composition in a changing climate. Like-
wise, asynchronous flowering, especially among individuals of the same species, could have nega-
tive effects on pollination success (Zohner et al. 2018).
Phenological events are not restricted to overstory trees, of course, but also occur in the
shrubs, seedlings, and saplings of the understory. Leaf phenology of many plants in the understory
Conifer and
Mixed Forest
Urban and
Low Veg Cover
148
Date of Onset
130
(a) (b) (c) (d)
Optical Estimate: 2.75 Optical Estimate: 2.00 Optical Estimate: 1.50 Optical Estimate: 0.50
Onset DOY = 135 ± 1.4 Onset DOY = 139.1 ± 1.9 Onset DOY = 141.9 ± 1.4 Onset DOY = 145.1 ± 1.4
F I G U R E 2 0 . 8 Varying leaf phenology related to topography and microclimate in Rhode Island. The study
location included a 30-m elevational change over 500 m of horizontal distance. Leaf flush is phenologically
later at lower points in the landscape (the phenological gradient encompassed a 2-week duration) because of
cold-air flow and a colder microclimate. This difference in phenology was detectable with satellite imagery.
Source: From Fisher et al. (2006) / with permission of Elsevier.
is often characterized by leaf flush and expansion days or even weeks before the overstory trees in
order to uptake sufficient carbon prior to canopy closure (Heberling et al. 2019). This shade-
avoidance strategy, known as phenological escape, not only enhances the growth of understory
plants but is critical for their growth and survival (Augspurger 2008; Lee and Ibáñez 2021a) because
it provides the majority of the seedlings’ annual carbon assimilation (Figure 20.9). At issue is that
warming spring temperatures are currently driving earlier canopy leaf flush (Piao et al. 2019), such
that the mismatched phenology between overstory and understory may diminish. However, using
phenology modeling and field experiments of sugar maple and northern red oak seedlings in
southern Michigan, Lee and Ibáñez (2021b) predicted that seedling leaf phenology would be more
sensitive to spring warming relative to the nearby overstory trees that shade them. As a result, the
duration of phenological escape by tree seedlings should increase under warmer springs, thus
(a) Spring Summer Fall (b)

0.15
Accum. CO2 (mol)

0.100
0.10
Seasonal assimilation (mol CO2 ± 95% CI)
A. saccharum
0.05
0.075 Q. rubra
0.00
May 1 Aug 1 Nov 1
0.050 (c)
Assim. rate (cmol CO2/day)

0.008
0.025
0.004
0.000
0.000
A. saccharum Q. rubra May 1 Aug 1 Nov 1
F I G U R E 2 0 . 9 Seedling carbon assimilation for sugar maple and northern red oak resulting from phenolog-
ical escape in southern Michigan. The majority of carbon is assimilated by tree seedlings in early spring,
when light is most available in the understory prior to leaf flushing of the canopy trees. (a) Seasonal foliar
carbon assimilation estimates. (b) Additive carbon assimilation over the growing season. (c) Daily assimila-
tion rates for individual sugar maple and northern red oak seedlings. Source: From Lee and Ibáñez (2021a) /
John Wiley & Sons.
increasing seedling access to early spring light. In the end, however, increased phenological escape
is not likely to increase seedling performance in the understory because of likely increases in
growing season respiration resulting from reduced water availability and warmer temperatures
under climate change (Lee and Ibáñez 2021b). Such increased respiration would largely offset or
exceed additional carbon gained by tree seedlings via earlier leaf flush in warmer springs. Notably,
phenological escape by tree seedlings has also been documented in the fall, where seedlings main-
tain their leaves beyond canopy tree leaf drop (Gill et al. 1998), but late-season carbon uptake via
this phenomenon is minimal (Lee and Ibáñez 2021a).
REGENERATION
Forest ecologists often focus on the response of mature trees to rising temperatures and changes in
moisture availability, but tree regeneration deserves equal attention. Even prior to anthropogenic
climate change, forests are likely to have adjusted to changes in environmental conditions primar-
ily by changes in seedling occurrence and abundance. Regeneration trajectories are also useful in
predicting future forest structure, distribution, and condition. Regeneration is directly affected by
changes in temperature and precipitation, and these effects apply at multiple points of the
r eproductive cycle—during flowering and cone production, seed germination, seedling establish-
ment, and seedling growth and survival. Increased atmospheric CO2 concentrations are also likely
to affect seedling growth, as described in the earlier text, though our focus here is on changes in
temperature and moisture availability.
Importantly, effective tree regeneration depends on success at each stage of reproduction
(Chapter 4), and specific changes in temperature and/or precipitation may be beneficial at one
regeneration stage but detrimental at another. Overall, higher growing season temperatures benefit
the early stages of regeneration, including flowering, pollination, and germination, because these
processes accelerate under warmer spring temperatures (Boucher et al. 2019) and may begin
sooner in early springs (Classen et al. 2010; Prevéy et al. 2018). Moderately higher temperatures
may also benefit seedling growth and survival so long as adequate moisture is available (Fisichelli
et al. 2014a), but increasingly higher temperatures eliminate most benefits of warming (Boucher
et al. 2019). For example, a study of ponderosa pine in the western United States compared histor-
ical regeneration potential (1910–2014) to future potential using a climate–water balance model
(Petrie et al. 2017). The study found that higher temperatures and lower moisture availability
under climate change reduced seedling growth and survival relative to the historical period, but
higher temperatures favored “flowering” (the pollination period for conifers), seed production,
and germination. The models predicted higher future regeneration for 2020–2059, but also that
increasingly higher temperatures after 2059 would reduce seedling growth and survival enough
that the likelihood of regeneration failure in ponderosa pine would approach 60% (Petrie
et al. 2017). Similarly, Ibáñez et al. (2017) found in Michigan that warmer temperatures benefitted
seed production of red and sugar maples, but were detrimental to seedling establishment and
survival, which were highest in cooler years. Moreover, increasing temperatures were more detri-
mental to seedling establishment and survival in southern locations compared to those three
degrees latitude further north.
Of course, temperature changes do not occur in isolation, and moisture availability is impor-
tant at every stage of regeneration. We differentiate changes in “moisture availability” from those in
precipitation, given that flowers and seeds, germination, establishment, growth, and survival all
depend specifically on soil moisture (Vose et al. 2016; Boucher et al. 2019). Soil-moisture availability
may decrease even where precipitation is increasing if warming temperatures increase transpiration
rates and evaporation from the soil surface. Seedlings are particularly susceptible to decreases in soil
moisture because of their relatively small root system that limits their contact with, and thus mois-
ture uptake from, soil water (Will et al. 2013). Seedling mortality and reduced growth associated
with reductions in moisture availability have been documented for many regional ecosystems (Er-
ickson et al. 2015; Rother et al. 2015). In eastern North America, seedlings that persist in the under-
story are typically shade-tolerant, and many shade-tolerant species lack tolerance of drought
(Chapter 13), potentially making seedlings of shade-tolerant species particularly susceptible to
mortality under reduced soil-moisture conditions. Projected soil-moisture deficits under climate
change are particularly concerning where regeneration occurs only episodically in wetter years, as
in western North America. For example, ponderosa pine forests of the western United States typi-
cally have a regeneration niche limited to periods of high moisture availability that are likely to
decrease in frequency with climate change (Petrie et al. 2017; Davis et al. 2019).
Given the likelihood for increasing occurrence of wildfires (see discussion in the following
text), rising temperatures, and increasing drought under climate change, as well as questions about
the resilience of forests under such conditions, much attention has been given to tree regeneration
following wildfires in western North America. Overall, tree regeneration has thus far declined fol-
lowing wildfires in the twenty-first century compared to the end of the twentieth century due to
warmer and drier conditions (Stevens-Rumann et al. 2018; Stevens-Rumann and Morgan 2019;
Rodman et al. 2020), suggesting that either longer recovery periods will be necessary for burned
areas or a conversion of dry forests to non-forest cover types will occur. Using a series of models to
examine tree seedling regeneration of Douglas-fir and ponderosa pine in 21 wildfire areas of the
northern Rocky Mountains, Kemp et al. (2019) found that average summer temperature best
explained post-fire seedling densities for both species. A temperature-tolerance threshold for
Douglas-fir seedlings was determined to be 17 °C, above which Douglas-fir would not regenerate.
Notably, over 80% of the 177 sampled sites were projected to exceed the threshold by the middle of
the twenty-first century, which would sharply decrease Douglas-fir regeneration (Figure 20.10).
(a) PSME
<180 <180 <5%
180–370 180–370 5–50%
370–1110 370–1110 50–100%
>1110 >1110 >100%
0 2040 km 0 2040 km 0 2040 km
(b) PIPO
<180 <180 <5%
180–370 180–370 5–50%

370–1110 370–1110 50–100%
>1110 >1110 >100%
0 2040 km 0 2040 km 0 2040 km
8 10 12 14 16 18 20 22 24
Summer (JJAS) Temperature (°C)
F I G U R E 20 .1 0 Map of current (left), predicted future (center), and change in (right) (a) Douglas-fir
(PSME) and (b) ponderosa pine (PIPO) seedling densities. Map backgrounds are the average summer
temperature of dry mixed-conifer forest currently (1981–2010; left) and predicted for mid-century (2041–
2070; center and right). Seedling densities are considered to have changed if the density increased or
decreased by at least 5% over the predicted period. Pie charts display the percent of sites with increase (blue),
decrease (red), and no change (gray) for each species. Source: Kemp et al. (2019) / John Wiley & Sons.
The threshold for ponderosa pine was slightly higher at 19 °C, but future regeneration potentials
were also projected to decrease. In subalpine forests, post-fire subalpine fir and Engelmann spruce
regeneration declined with drought severity and distance to seed sources but was higher on cooler
and wetter aspects (Harvey et al. 2016b). Post-fire regeneration may also be limited indirectly by
climate change if wildfire severity, frequency, and/or extent are increased by increasing tempera-
tures and decreasing precipitation; these scenarios are discussed in the Disturbance section in the
following text.
CLIMATE CHANGE EFFECTS ON TREE SPECIES DISTRIBUTIONS

Perhaps of longest interest to ecologists concerned about climate change is the potential for
changes in tree species’ distributions. Range shifts and species migrations in relation to climate
were considered well before modern climate change issues were ever examined as part of ecolog-
ical research. Predicting how tree species may shift in their distributions is largely based on the
idea that populations at the margins of a species’ range occur on less favorable sites and in sub-
optimal climates. Such populations are therefore more susceptible to decline or to go locally
extinct, leading to gradual changes in their geographical ranges (Lawton 1993). Given that cli-
mate change is particularly acute at northern latitudes, species in those locales might be expected
to migrate toward more suitable growing conditions. Studies of migrations and shifting tree
distributions may be categorized as those that (i) reconstruct the migration and re-distribution of
tree species following glacial retreat at the end of the last ice age (Chapter 15); (ii) examine how
species’ ranges have relatively recently shifted since the presumed onset of anthropogenic cli-
mate change (about 1850); or (iii) project future distributions using modeling techniques. As
paleoecological reconstructions of immediate post-glacial tree distributions were discussed in
detail in Chapter 15, observed changes over the last 150 years and projected changes are our
focus here.
OBSERVED RANGE SHIFTS

Given the hypothesis that species’ distributions will change mostly at their margins, ecologists
have expected that species would migrate toward the poles (northward in the Northern Hemi-
sphere) and upward in elevation as temperatures warm. Most studies searching for evidence of such
movement have depended on large networks of repeatedly measured forest inventory plots to
examine changes in species occurrence across large areas over time (e.g., Woodall et al. 2009; Zhu
et al. 2014; Fei et al. 2017; Sittaro et al. 2017; Boisvert-Marsh et al. 2019). In general, tree species in the
Northern Hemisphere have been found to be moving higher in elevation (Holzinger et al. 2008;
Lenoir et al. 2008), but northward migration has been somewhat less consistent. In one of the first
studies to examine climate-related tree species migration across broad areas of the eastern United
States, Woodall et al. (2009) compared the average latitude of large adult trees to the average lati-
tude of seedling density for 40 species. In comparing seedlings to trees, those species having seed-
lings both further to the north and south of adult trees at the extent of the species’ range would be
expected to be expanding their range; those with seedlings further to the north but not to the south
to be expanding northward; those with seedlings further to the south but not the north to be
expanding southward; and those with adults further to the north and south than seedlings to be
contracting in their range (Zhu et al. 2012). The average latitude for seedlings was more than 20 km
further north than that of tree biomass for northern species (Figure 20.11), revealing tree regener-
ation in the northern portion of these species’ ranges and suggesting northward tree migration by
as much as 100 km per century for some species. Notably, migration in the southern portion of tree
species distributions did not appear to occur (i.e., ranges expanded to the north but did not contract
from the south; Woodall et al. 2009). Additional studies have corroborated similar northward shifts
Yellow Birch Sugar Maple
6 30 20 2500
18
5 25
16 2000
4 20 14
12 1500
3 15 10
8 1000
2 10
6
Seedling Density (stems/ha)

5 4 500
1
Tree Biomass (tonnes/ha)
2
0 0 0 0
<34 34–36 36–38 38–40 40–42 42–44 44–46 46–48 48+ <34 34–36 36–38 38–40 40–42 42–44 44–46 46–48 48+
Northern Red Oak American Basswood

16 160 2.5 50
14 140 45
2 40
12 120
35
10 100 1.5 30
8 80 25
1 20
6 60
15
4 40
0.5 10
2 20 5
0 0 0 0
<34 34–36 36–38 38–40 40–42 42–44 44–46 46–48 48+ <34 34–36 36–38 38–40 40–42 42–44 44–46 46–48 48+
Latitude Classes (degrees)
FIGURE 20.11 Comparison of tree biomass (solid lines) and seedling densities (dotted lines) across 2° latitude classes for four northern species
in the eastern United States. Average and peak seedling densities are found further north than average and peak tree biomass for each species,
suggesting a northward expansion of each species’ range. Source: Woodall et al. (2009).
for portions of species’ ranges in Canada (Boisvert-Marsh et al. 2014, 2019; Sittaro et al. 2017). For
example, rising temperatures increased seedling recruitment for sugar maple, American beech,
and red maple in the northern portions of their range in Quebec, Canada, but decreased it in the
south, driving range shifts northward (Boisvert-Marsh et al. 2019).
Subsequent studies have found less evidence for northward migration in North America.
Using the same data set as Woodall et al. (2009) to examine potential range shifts for 92 species,
Zhu et al. (2012) emphasized that such analyses must be conducted at the boundaries of species’
distributions rather than simply comparing average latitudes for seedlings and adult trees, as geo-
graphic range is realistically defined by boundaries rather than central tendencies. In examining
adult trees and seedlings only within the 95th percentile of latitude for each species, they found no
evidence that migration was greatest where climate change is the strongest. Instead, they found
evidence for range contraction rather than northward expansion in 59% of the species studied; only
21% of the species showed a northward shift, and 16% showed a southward range shift. Additional
studies using this methodology have suggested that most species’ ranges in eastern North America
are actually stable or possibly contracting rather than expanding northward (Woodall
et al. 2013, 2018; Zhu et al. 2014). The concern with this “failure” to migrate northward of many
species is the potential that tree species are unable to migrate at a pace similar to that of climate
change occurring at northern latitudes (Neilson et al. 2005; Sittaro et al. 2017).
An important consideration is whether range shifts could occur in an eastward–westward
shift in addition to (or instead of) a northward–southward shift. Tree species ranges in the eastern
United States have been shown to have shifted to the west (median rate 15.4 km per decade) more
than to the north (11 km per decade) in response to recent increased precipitation and moisture
availability (Fei et al. 2017). For 86 tree species distributions (defined by abundance) over 35 years,
important regional differences also exist in species migrations. Most species in the northern hard-
wood region shifted north at 20.1 km per decade, those in the central hardwood region shifted west
at 18.9 km per decade, those in the southern pine–hardwood region shifted west at 24.7 km per
decade, and those in the forest–prairie transition moved west at 30 km per decade. The faster west-
ern shift relative to the northern shift suggests that tree species distributions were more sensitive to
precipitation compared to temperature between 1980 and 2015 (Fei et al. 2017), and emphasizes
that change in precipitation is an equally important factor to consider in climate change effects on
range shifts.
PROJECTED CHANGES IN TREE SPECIES DISTRIBUTIONS

Projecting future tree distributions requires some sort of species distribution modeling or forecast-
ing of both climate and climate–species relationships. Models are often criticized by scientists and
laypersons because they are by definition a simplification of what happens in the natural world
and depend on a modeler’s understanding of how the natural world works. Nevertheless, predict-
ing something that has not yet happened requires such models, and projecting future tree
distributions requires accepting some uncertainties in the data. A recent review of forest modeling
challenges under climate change is presented by Scheller (2018). We also emphasize that projections
of tree species are almost continuously changing with improvements in climate model projections
and tree sampling data, such that those presented in the following text are for illustrative purposes
rather than precise predictions of where a given tree species may be distributed under various cli-
mate change scenarios.
One method of predicting future distributions is the use of climate envelopes, which iden-
tify the values of a series of climate variables at each geographic location where the species
occurs, then define the species’ range using the maximum and minimum values of those vari-
ables (McKenney et al. 2007a). Using projections of future climate variables from one of several
climate models with scenarios of high or low CO2 emissions, a climate envelope may then be
Climate Change Effects on Tree Species Distributions 557
established for future climate scenarios and compared to the current envelope (Figure 20.12).
Such a method thereby predicts future potential tree species habitat—in other words, where a
given tree species could grow, ignoring the influence of biotic interactions such as competition,
dispersal, or genetic adaptation to growing conditions (McKenney et al. 2007b). A second method
uses statistical methodology called regression tree analysis (RTA) to project future tree species
distribution from predictor variables that include climate variables as well as soil, elevation, land
use, and indices of landscape fragmentation (Iverson et al. 2008, 2019a). Predictor variables are
used to estimate the importance value (an index integrating the frequency, size, and density of a
given species) of tree species under current and future climate scenarios. The importance values
are mapped in contiguous 400-m and 100-m plots across the landscape, such that current and
projected distributions can be compared (Figure 20.12). Notably, the RTA method also predicts
only potential tree species habitat in the absence of biotic interactions. Neither method necessar-
ily predicts species migration or even the potential distribution of tree species, but instead
Legend
Future climate range
Current climate range
Importance
Value
0
1–3
4–6
7–10
11–20
21–30
31–50
51–100
F I G U R E 20 .1 2 Future projections of the distribution of sugar maple in the eastern United States using
climate envelopes (left) and regression tree analysis or RTA (right). Climate envelopes show a northward
movement of potential habitat under future climate for the period 2071–2100, suggesting that much of the
southern portion of the current climate envelope will become unsuitable due to warming temperatures. The
RTA analysis suggests that sugar maple will become a less important species on the landscape by 2100
(bottom right) compared to its current distribution (top right). Future climate projections are based on
different climate models for the two methods. Source: Climate envelopes map from McKenney et al. (2007b)
/ with permission of Oxford University Press; RTA map from Peters et al. (2020).
resents the area that would be suitable to colonize under climate change. Other modeling tech-
p
niques have also been used to project future tree distributions (e.g., He et al. 2017; Wang
et al. 2017).
Results of species distribution modeling suggest rather drastic changes for many tree species
by the year 2100. Using a model that assumed any of the 130 species studied could disperse beyond
their current climate envelope, McKenney et al. (2007a) found an overall average of a 12% decrease
in envelope size and a northward shift of about 700 km. Notably, 72 of the 130 species exhibited a
shrinking envelope, and 11 of the 72 species, most currently located in the southeastern United
States, had envelopes that decreased in size by over 60%. Compared to current climate envelopes,
future envelopes of 58 species were projected to increase in size up to 43%, mostly those located in
the eastern United States and along the West Coast. These results suggest two important points.
First, climate envelopes can shift rather large distances quite quickly, and it is unlikely that trees
will migrate at a similar rate. Relatively slow migration rates have been corroborated by Iverson
et al. (2004, 2019b) in a modeling context. In such a case, it is likely that many tree species will be
mismatched with climate by the end of the twenty-first century. Second, species with currently
limited distributions, such as many in the southeastern United States, are likely to be in danger of
extinction because climate change drastically reduces the size of their climate envelope.
Results were even more drastic when the models were run under the assumption that tree
species would only persist within current envelopes (i.e., they were unable to disperse beyond their
current envelopes). In this scenario, future envelopes decreased by 58% and shifted northward only
330 km—probably decreasing more in size due to a more limited northward shift. Again, species
with the largest decreases in envelope size were coastal with limited current distributions, partic-
ularly in the southeastern United States. A compelling prediction to emerge from this work was
that by 2100 global change will have altered the climate of the southern United States beyond what
any of the 130 species studied can currently tolerate, and that the species whose climate envelope
showed the greatest latitudinal shift northward were from the southeastern portion of North
America (McKenney et al. 2007a). When model runs were completed again using updated climate
models (McKenney et al. 2011), climate envelopes were found to be as much as 10% larger in size
but shifted up to 2.4° latitude further (Figure 20.13).
Similar results surface using RTA modeling. Iverson et al. (2019a) examined 125 tree species
in the eastern United States and found that 72% of the species would gain suitable habitat under
climate change, 21% of the species would lose suitable habitat, and 8% would remain unchanged,
although model reliability varies widely among species. Those species gaining the most habitat
were either (i) currently found in a warm and dry region (i.e., the southwestern portion of the
United States). These species’ distributions are heavily driven by temperature and expand greatly
when provided much warmer temperatures; (ii) relatively rare and able to expand their range
under climate change; or (iii) currently distributed in a southern location and expected to expand
northward by 2100. About three to four times as many species moved northward instead of south-
ward, with further distances occurring under more severe climate change scenarios (Figure 20.14).
In recent RTA work that combined forest tree data from the United States and Canada for 25
species, Prasad et al. (2020) found that species projected to lose habitat did so mainly along the
southern limits of their range, while habitat gains occurred mainly along the northern limits of
their range.
The use of importance values in the RTA methodology provides the added advantage of
examining the changes in species occurrence that might take place with climate change. Iverson
et al. (2019a) identified the top three species for each state in the eastern United States as deter-
mined by their importance value and found that most of these species decreased in suitable habitat
under climate change. For example, in Michigan, the species with the top three importance values
are red maple, sugar maple, and trembling aspen, which decline to 98%, 82%, and 89% of their
Climate Change Effects on Tree Species Distributions 559
–119 to –85 –38 to –21 11–26
–84 to –61 –20 to –6 27–45
–60 to –39 –5 to 10 46–84
F I G U R E 20 .1 3 Climate envelope richness (i.e., number of tree species) differences between current
(1971–2000) and future (2071–2100) periods for North America using two different climate models. Larger
numbers (warmer colors) represent decreases in the number of tree species. Northern gains with southern
losses (particularly from the southeast quarter of the continent) suggest northward migrations of many tree
species. Left panel utilized the Parallel Climate Model (PCM); right panel utilized CCSM3.0 climate model.
Source: McKenney et al. (2011).
current importance values under a low carbon emissions scenario and 87%, 74%, and 80% under a
high emissions scenario. Over all states and regions in the eastern United States, importance values
of the first (loblolly pine) and third (sweetgum) most common species increased under climate
change by 21% and 48% of current importance values under a low emission scenario and by 32%
and 73% under a high emissions scenario. The second most common species, red maple, decreased
to 99% and 92% of its current importance values under low and high emissions scenarios, respec-
tively (Iverson et al. 2019a).
PROJECTED CHANGES IN FOREST TYPE DISTRIBUTIONS

Notably, efforts to predict future tree distributions are focused primarily on species shifts rather
than shifts in communities or forest types. Evidence from the pollen record and post-glacial recon-
structions of tree species distributions clearly shows that post-glacial northward migration of
species occurred independent of each other and not as forest types (Chapter 15). Thus, many forest
types that existed during the period of glacial retreat have no modern analog, and likewise many
modern types did not exist 12 000–15 000 years ago. Most ecologists therefore assume that such
independent migration is likely to again occur as anthropogenic climate changes take place, and
that tree species’ range shifts are most likely to result in community disassembly and re-assembly,
(a) N F I G U R E 20. 14 Modeled changes in

NNW NNE distance and direction of mean centers
of suitable habitat for 125 species
NW NE utilizing different climate models.
(a) Average of three climate models
using low-emission scenario. (b) Average
of three climate models using current
WNW ENE
emission scenarios. Both models suggest
a northeasterly migration for most
species, with further distances predicted
W E using the high-emission scenarios.
Source: Iverson et al. (2019a).
WSW ESE
SW SE
SSW SSE
S
(b) N
NNW NNE
NW NE
WNW ENE
W E
WSW ESE
SW SE
SSW SSE
S
creating new forest types over the long term. Three important differences exist between current
species shifts and historical post-glacial migrations that create uncertainties, however. First, the
rate of anthropogenic climate change is higher than what occurred during glacial retreat, and it is
unclear how a higher rate of change will affect tree migrations relative to those that occurred his-
torically. Second, human influences that may affect migration rates via their impacts on dispersal—
such as landscape and forest fragmentation, alteration of disturbance regimes that may encourage
or inhibit tree regeneration, and invasive species that may compete with tree species—did not exist
Climate Change Effects on Forest Disturbances 561
during post-glacial migration, and the presence of such barriers to migration on the landscape may
affect tree species associations. Finally, current range shifts are occurring in the presence of well-
established communities rather than recently de-glaciated terrain. These “occupied landscapes”
may affect dispersal and competitive outcomes, and thus how future associations are assembled.
Despite these caveats, several researchers have attempted to model how forest types may
be affected by climate changes, at least in the short term. A study by Iverson et al. (2008) using
previous climate models that have since been updated suggested that changes in forest type dis-
tribution in the United States are most likely at northern latitudes. Spruce-fir forests, already
limited to the far northern areas of the eastern United States, are likely to be essentially elimi-
nated south of the Canadian border (Figure 20.15), as are the white-red-jack pine forests cur-
rently common in the Lake States. Refugia for some of these boreal species may include high
elevations in the northern Appalachian Mountains or north-facing slopes where microclimates
mitigate rising temperatures. Aspen–birch forests may be retained only in northern Minnesota
under low-emissions scenarios and are eliminated under high-emissions scenarios. By contrast,
oak–hickory and oak–pine forests were projected to increase to the north; maple–beech–birch
forests were stable under low-emissions scenarios but reduced under high emissions, often
replaced by oak–hickory forests. There is some evidence that red maple has already moved into
the southern edge of the boreal forest due to warming temperatures at its northern extent (Fisi-
chelli et al. 2014b). Other forest types, such as elm–ash–cottonwood and the southern pines,
were stable in the analysis. The northern forests lost under climate change currently provide
significant economic values to the region in the form of commercial products and tourism
(Iverson et al. 2008).
In considering the potential migration of forest types, it is important to reiterate that com-
munities are grounded within sites, and sites do not migrate with climate change. Important site
factors other than climate also shape the composition of species within ecosystems, and enor-
mous variation in these factors exists within a given species’ climate envelope or suitable habitat
that is not and cannot be easily modeled. For example, in the northeastern United States, upland
oak species, with or without warming temperatures, will require periodic fires to open the canopy
and remove competitors if they are to persist. Warming temperatures and reduced precipitation
in the eastern United States would facilitate such disturbances and perhaps will accompany the
northward migration of oak-dominated forests, but oaks probably will still be limited to drier sites
with droughty soils, at least in the short term. Floodplain forests dominated by silver maple are
likely to be favored by increased spring precipitation and flooding regardless of warming temper-
atures. Mesic forests dominated by species such as sugar maple, basswood, and American beech,
which prefer moist but not wet sites, may find refugia from warmer temperatures and decreasing
soil moisture in cool, low wetlands commonly nearby their current sites as they dry out. There
will likely come a time when the severity of climate change overrides current site–species rela-
tionships, but short-term changes in forest type distributions are likely to remain closely tied to
physical site factors over the next several decades.
CLIMATE CHANGE EFFECTS ON FOREST DISTURBANCES

The effects of climate change on forests involve much more than the simple ability of tree species
to tolerate changes in temperature and precipitation. We have discussed throughout this book the
central role of disturbances in shaping forest structure and function (Chapters 10 and 16) as well
as landscapes (Chapter 22). Many disturbance regimes are strongly driven by climate, so naturally
climate change is likely to have significant impacts upon where disturbances occur, how often they
occur, their severity, and their extent. Changes in disturbance regimes, in turn, impact how well
FIA-Current RF-Current
Whte/Red/Jck
Sprc/Fir
Lnglf/Sish
Lobly/Shrtif
Oak/Pine
Oak/Hikry
Oak/Gum/Cypr
Elm/Ash/Ctnw
Map/Bch/Brch
Aspn/Brch
NoDat/NoFor
GCM3Avg Lo PCM Lo
GCM3Avg Hi HADLEY Hi
F I G U R E 2 0 .1 5 Maps of current and potential future suitable habitat for current forest types of the eastern
United States. Maps include current inventory estimates of current distribution of abundance (FIA current),
the modeled current map (RF current), two low-emission model scenarios, and two high-emission model
scenarios. Source: From Iverson et al. (2008) / Elsevier.
forests are able to provide ecosystem services (Chapter 23) such as carbon storage, biodiversity, and
water resources. The ecological resilience of forests is also at issue, as changing disturbance regimes
have the capacity to move ecological systems beyond tipping points that would result in the loss of
forests (Reyer et al. 2015; Johnstone et al. 2016). In sum, changes in forests due to climate change
are most likely to be expressed over the next century as extensive changes in forest composition
and structure due to altered disturbance regimes rather than slow-paced species migrations that
reflect physiological responses to changes in temperature, precipitation, and CO2 (Dale et al. 2001).
This is especially the case because disturbances create rapid changes, whereas long-lived organ-
isms such as trees would express changes more slowly—especially if changes are expressed through
regeneration (Johnstone et al. 2010).
The onset of climate change and observable changes in disturbances have greatly increased
attention toward forest disturbances in the last few decades, and general trends across all distur-
bance types may be gleaned from the literature. In a review of 647 studies investigating climate
change effects on disturbances, Seidl et al. (2017) found that 42% of the studies found temperature-
related factors to be the most important climatic driver of changes in disturbance regimes, partic-
ularly in northern latitudes and especially in the boreal forest. About 38% focused on precipitation,
primarily water availability, mostly toward southern latitudes and especially in the tropics. These
two broad climatic factors are as impactful in their effects on disturbance regimes as they are on
growth, regeneration, phenology, and species distributions.
Climate change is likely to affect the frequency and severity of many kinds of disturbances,
such as fire, drought, windstorms, ice storms, hurricanes, invasive species, insect and pathogen
outbreaks, and landslides (Dale et al. 2001). In the following sections we focus on fires, insect and
pathogen activity, and wind. In their review, Seidl et al. (2017) noted that where climate change is
projected to produce warmer and drier conditions, over 82% of the studies identified an increase in
fires and 78% in insect activity. Where climate was projected to be warmer but wetter, only 55%
identified an increase in fires and 65% in insect activity. Wetter climates also translated to an
increase in wind disturbance (89%) and diseases (69%) (Seidl et al. 2017). The importance of these
disturbances for forests is described in Chapters 10, 12, and 16; in the following text we focus
mainly upon how they may be altered by climate change.
FIRES
With much of the climate across the globe predicted to become warmer and drier and the dramatic
effects of fires on forests, ecologists have paid great attention to the effects of climate change on
fire regimes across the world. Fire frequency, extent, intensity, and seasonality are closely linked
to weather and climate as well as to forest species composition, stand structure, and site
characteristics (Chapter 10). Many decades of fire suppression in many forests have altered forest
structure to such an extent that it can sometimes be difficult to differentiate the effects of climate
change on fire behavior and occurrence from the effects of fire suppression. Climate change
directly affects fire regimes by influencing the moisture content of fuel, which in turn affects fire
initiation and spread (Williams and Abatzoglou 2016), ignition sources such as lightning (associ-
ated with storms), and the way fire spreads (via influences on wind speed and other weather
factors). Williams and Abatzoglou (2016) estimated that human-caused climate change in the
western United States resulted in extremely dry fuels in 75% more forested area for an additional
9 days during the fire season. They also attributed 4.2 million ha more burned area in the West to
climate change between 1984 and 2015, which was almost twice the area expected otherwise.
Climate change can also influence fire regimes indirectly by affecting the productivity of vegeta-
tion and the decomposition of dead vegetation, both of which impact the availability of fuels
(Pausas and Ribeiro 2013), by driving forest composition thereby increasing the flammability of
forests, and by altering forest structure which influences fuel continuity. The effects of fire on for-
ests are discussed in detail in Chapter 10.
Changes in climate have already affected fire regimes in many forests in observable ways. In
dry, fire-prone forests such as those of the boreal region in Canada and those in the western United
States, warmer temperatures and lower precipitation over the last 20 years have increased the area
burned by fires (Gillett et al. 2004; Abatzoglou and Kolden 2013). Moreover, the likelihood of fires
>5000 ha in size is projected to increase across much of the United States by the mid-twenty-first
century, especially in the northern Rocky Mountains, the northern Lake States, and the southeast-
ern Coastal Plain (Figure 20.16; Barbero et al. 2015). Some researchers have estimated that the area
burned each year in the western United States could increase as much as sixfold by the middle of
the twenty-first century with continued climate change (Litschert et al. 2012). The western United
States has also experienced an increase in how often fires burn, due to earlier snowmelt and
warmer spring and summer temperatures that lead to longer fire seasons (Westerling et al. 2006).
The extraordinary fire season of 2020 in the western United States alone effectively doubled the
area burned in the Rocky Mountains since 1984, and as a result the rate of burning in subalpine
forests is now double that of the last 2000 years (Higuera et al. 2021).
The potential for significant effects of climate change on fire regimes in the future has
been predicted for several decades (Dale et al. 2001). One of the most compelling examples of
potential changes in fire regimes due to climate change is found in the Greater Yellowstone Eco-
system in the central Rocky Mountains, where the climate is predicted to increase temperatures
and reduce precipitation to an extent that the disturbance regime will move well out of its his-
torical range of variability (Chapter 22). Fires larger than 200 ha have burned every 100–300 years
in the region for as many as 10 000 years, but the fire rotation is projected to shorten to less than
30 years by the mid-twenty-first century with continued warming, and the frequency of years
with synchronized large fires across the region is likely to increase (Westerling et al. 2011).
Altered fire frequency and extent to this degree are incompatible with the tree species that cur-
rently occupy the landscape and will almost certainly have consequences for the persistence of
forests and c arbon storage.
The important interactions between climate change and fire are not limited to how fire
regimes are altered. An important effect only recently documented is how well forests recover
following fires in an altered climate—that is, how might climate change affect the resilience of
forests when the fires that burn them are more frequent, larger, or more intense? The important
effect in such a scenario is the change to the fire regime as well as to the post-fire climate under
which the forest must recover. One of the easiest ways to envision such a scenario is to consider
post-fire regeneration. When fire burns through a forest, the severity of the fire determines how
the forest recovers (Chapter 10). A low-severity fire that does relatively little harm to mature trees,
such as historical fires in ponderosa pine forests, may recover rather easily and leave the forest
appearing relatively unchanged. However, recovery of forests following high-severity fires that
kill most of the mature trees, such as those that occur in boreal and subalpine forests, depends
heavily on post-fire regeneration that will eventually replace the killed mature trees. Such regen-
eration is likely to gain more attention from ecologists because (i) climate change is causing
larger, more frequent, and/or more severe fires, such that more burned area will probably rely on
post-fire regeneration for recovery, and (ii) the tree seedlings that regenerate following fires will
be subject to novel and in many cases harsher growing conditions under which they have to sur-
vive and grow. Reduced or failed regeneration, even when numerous fire adaptations and recov-
ery mechanisms are present, strongly affects the ability of a forest to be resilient to fires. Forests
that successfully regenerate will be able to recover from fires as they have in the past, but those
that do not may be converted to a non-forested state (Coop et al. 2020). Thus, changing fire
regimes and/or decreasing post-fire regeneration have the ability to reduce the amount of forest
cover across the world (McDowell et al. 2020).
Mean no. of VLF weeks expected (1971–2000)
0.5
0.2
0.15
0.10
0.05
Mean no. of VLF weeks expected (2041–2070) 0.04
0.03
0.02
0.01
No models
Relative changes in % (future–present)

600
500
400
300
200
150
100
75
50
25
0
No models
F I G U R E 20 .1 6 Mean annual number of very large fire (VLF; >5000 ha) weeks per surface unit for
1971–2000, 2041–2070, and the relative change between the two time periods. These large disturbances are
projected to increase by mid-century, particularly in the northern Rocky Mountains, the Lake States, and the
southeastern Coastal Plain. Source: From Barbero et al. (2015) / CSIRO Publishing.
As climate change progresses with increasing fire activity, rising temperatures, and reduced
soil moisture, forest conversion has become increasingly common, especially in western North
America. In general, the mechanisms behind such conversion can be summarized as (Coop
et al. 2020):
Increased fire activity that kills all or most trees: Many current and historical fire regimes are
stand-replacing and kill many or most trees, such as in many high-elevation forests dominated
by lodgepole pine, spruce, and fir. However, increasing area burned, fire size, number of fires,
and proportion of area burned at high severity are likely to reduce the resilience of such forests
historically burned by large, infrequent crown fires as well as those burned by smaller, frequent
surface fires (e.g., ponderosa pine forests) or mixed-severity fires (e.g., mixed conifer forests).
Such changes in fire regimes have several important effects. First, a growing number of large fires
that burn at high severity will raise the number of large openings on the landscape (Figure 20.17a).
Unless the forests are dominated by serotinous species (e.g., lodgepole pine), reseeding of the
burned area would need to occur via live tree seed sources whose distance from many portions of
a large patch may be beyond the trees’ dispersal limit (Figure 20.17b; Johnstone et al. 2016).
Moreover, many such landscapes become extremely harsh environments for seedling survival
after they are burned, which may lead to forest conversion to non-forest or, at the very least, delay
its recovery (Harvey et al. 2016a,b). For example, almost 40% of ponderosa pine forests burned by
(a) (b)
Elimination of mature tree cover Inhibition of recovery processes
1) Increased area burned & fire severity 3) Fewer & more distant trees reduce seed availability
create more, larger openings
4) Short-interval burning eliminates
Conversion under regeneration Conversion under
altered conditions altered conditions
5) Heat & drought constrain establishment
Spatial Extent
Spatial Extent
2) Heat & drought may elevate tree

mortality before and during fire
High-severity patches High-severity patches

under prior conditions under prior conditions
Temporal Duration Temporal Duration
(c)
Feedbacks
+
6) Positive: increased fuels &
flammability promote repeated
burning, reinforcing or
amplifying conversion Conversion under
altered conditions
Spatial Extent
7) Negative: decreased
fuels & flammability of
post-fire vegetation
permit recovery
High-severity patches
under prior conditions
Temporal Duration
F I G U R E 2 0 .1 7 Processes directing forest conversion due to fire. (a) Conversion is initiated by fire (red
arrow) or climate (yellow arrow) that kills extensive forested areas. (b) Conversion is maintained by failed or
reduced post-fire regeneration of pre-fire tree species, prolonging forest recovery time. (c) Repeated or
short-interval burning may further reinforce conversion by further limiting regeneration and recovery.
Source: From Coop et al. (2020) / with permission of Oxford University Press.
high-severity fire as part of the 2002 Hayman Fire near Denver, Colorado still lacks significant
tree regeneration nearly two decades after they were burned, due to a lack of seed sources and
harsh post-fire environment (Rodman et al. 2020).
Increasing the area burned each year also increases the chances that an area recently burned
will be burned again with relatively little time for the forest to recover between the two burns. Such
short-interval fires may remove remaining live tree seed sources, reduce or eliminate any established
tree regeneration, or shift vegetation composition toward species with superior post-fire resprout-
ing ability (Coop et al. 2016). For example, short-interval fires in the Greater Yellowstone Ecosys-
tem, where historical high-severity fires burned lodgepole pine forests on a 100–300-year rotation,
combusted almost all pre-fire biomass in some reburns, sharply reduced post-fire regeneration
relative to the previous (long-interval) fire, and converted many high-density stands on the
landscape to low-density stands (Turner et al. 2019). Notably, forest resilience in this case was
reduced by short-interval fires, even for a highly fire-adapted tree species with serotinous cones,
because a second fire occurred before the regenerating trees from the first fire were able to begin
cone production. Similar studies of black spruce, a semi-serotinous species in the boreal forest,
have documented a switch between forest types (e.g., Hart et al. 2019) or a conversion of forest to
grassland (Brown and Johnstone 2012) due to short-interval fires when trees are burned before
they are able to produce cones.
Warmer and drier climate making survival of post-fire regeneration more difficult: Climate change
may directly influence regeneration failure and potential forest conversion via the impacts of
warming temperatures and reduced precipitation on tree and seedling survival and growth. As
discussed in the earlier text, these changes in climate have resulted in significant tree mortality
in dry forests across North America. Such mortality is unrelated to fires but may contribute to
reduced seed sources critical to revegetating burned areas and may increase fuels for future fires
(Figure 20.17a,b).
Perhaps more critical to potential forest conversion is the potential inability of trees to regen-
erate in a warmer and drier climate following fire, even when seed sources are available. Such
post-fire regeneration failure attributed to climate change and potentially leading to forest
conversion has already been observed in the western United States. For example, regeneration of
ponderosa pine and Douglas-fir was low 8–15 years after fire at lower elevations in the Colorado
Front Range, especially on dry sites and at greater distances from post-fire seed sources (Rother
et al. 2015). Similarly, at lower tree line in Yellowstone National Park, Douglas-fir regeneration and
forest replacement were low on dry sites, far from seed sources, and at the lowest elevations in
areas of high-severity burning (Donato et al. 2016). These trends contrasted with rigorous Douglas-
fir regeneration on mesic sites, possibly signaling effects of warming temperatures and reduced
precipitation. At high elevations in Yellowstone, Rammer et al. (2021) used simulation models to
predict widespread regeneration failure (between 28% and 59% of the forested area) in highly fire-
adapted lodgepole pine forests by 2100 and particularly in non-fire-adapted forests dominated by
Engelmann spruce and subalpine fir. Finally, Stevens-Rumann et al. (2018) attributed sharp
decreases in twenty-first century post-fire tree regeneration at 52 wildfires in the Rocky Mountains
to greater annual moisture deficits that occurred from 2000 to 2015. Importantly, it is not known
whether such regeneration failures will truly lead to permanent forest conversion, or whether
recovery to the pre-fire forest will simply be delayed, because post-fire regeneration in dry forests
of the region occurs in pulses for many forest types such as ponderosa pine and mixed-conifer for-
ests (Coop et al. 2020).
Disruption of fire–vegetation feedbacks: Fire regimes generally include feedbacks that promote the
persistence of that regime, and alteration of fire regimes by climate change may change these
feedbacks (Figure 20.17c). For example, fire managers have long known that young forests are
less susceptible to fire than older forests because fuels are consumed by the previous fire;
flammability increases as fuels re-accumulate. This negative feedback between fire and vegetation
limits or delays future burning and thus typically provides an extended window during which
post-fire regeneration and forest recovery may occur (Coop et al. 2020). However, such a feedback
may be overridden by weather that promotes severe burning, and such weather is likely to
increase under current climate trends. For example, windy conditions and dramatic declines in
precipitation preceding and during the 1988 fires in Yellowstone National Park caused the fires to
spread irrespective of forest age or fuel type (Christensen et al. 1989), essentially negating any
negative feedback. Fire regimes can also be bolstered by positive feedbacks, whereby fires promote
future fires, particularly when they are severe. In some systems, severe burning quickly creates
high levels of fuels from killed vegetation and dense re-vegetation through sprouting that pro-
mote additional burning. Although such positive feedbacks perpetuate a fire regime under a
given climate, a changing climate that results in even small changes in fire activity may result in
relatively abrupt forest conversion (Figure 20.17c). For example, Tepley et al. (2017) documented
conversion of dry montane conifer forests in California and Oregon to highly flammable shru-
bland under drier post-fire conditions and at greater distances from seed sources. In this system,
recruitment of Douglas-fir occurs only in the first few years after severe fires, facilitating rapid
forest recovery, because competing broadleaf trees and shrubs prevent a longer duration of seed-
ling establishment. Warmer and drier temperatures that limit or reduce this immediate seedling
recruitment are likely to shift the system toward a prolonged recovery time, and the onset of
shorter-interval fires in a highly flammable landscape is likely to perpetuate shrub cover rather
than forest cover (Tepley et al. 2017).
The potential for forest conversion due to changing fire regimes is at the forefront of forest
ecology in an era of rapid climate change. Excellent reviews of this potential are provided by Buma
et al. (2020), Coop et al. (2020), Hart et al. (2019), Hessburg et al. (2019), Stevens-Rumann et al.
(2018), Tepley et al. (2017), Johnstone et al. (2016), and Anderson-Teixeira et al. (2013).
INSECTS AND PATHOGENS

Climate change, particularly the effect of warming temperatures, is likely to influence insect and
pathogen effects on forests nearly as much as it does for fires. Insect outbreaks are already a major
source of tree mortality across the United States; between 1984 and 2012, bark beetles alone killed
more forest area than wildfires (Hicke et al. 2016). Many current and recent bark beetle outbreaks
have been associated with rising temperatures and decreasing precipitation (Bentz et al. 2010),
leading many ecologists to speculate that climate change is likely to further increase such out-
breaks in the future. Pathogens are similarly expected to have large effects on forests because of
climate change (Sturrock et al. 2011; Kolb et al. 2016). Climate change may facilitate outbreaks of
insects and disease by (i) affecting the development and survival of insects and pathogens, and (ii)
changing the susceptibility of the trees that they attack.
Insect and pathogen activity is strongly controlled by effects of climate. For insects, the most
obvious effect is that warmer temperatures accelerate the rates at which they proceed through
their life cycles. In some cases, warming temperatures have increased developmental rates to the
extent that species with a 2- or 3-year life cycle have been able to complete their development
within a single year (Figure 20.18; Bentz et al. 2010), which would create quicker population
growth that potentially leads to outbreaks. Moreover, increased developmental rates may also lead
to higher synchrony among populations, allowing mass attack that overwhelms tree defenses and
leads to outbreaks (Logan and Powell 2001). Some insects have also been documented to increase
their reproductive output by producing multiple generations per year instead of one. For example,
Mitton and Ferrenberg (2012) observed that 20 years of rising temperatures in the Colorado Front
(a) (d)
(b) (e)
(c) (f)
High
Low
F I G U R E 20 .1 8 (a)–(c) Predicted probability of spruce beetle offspring developing in a single year in spruce
forests across the range of this insect in North America during 1961–1990, 2001–2030, and 2071–2100. Forests
with a higher probability of 1-year life-cycle duration have a higher probability of population outbreak and
increased levels of spruce-beetle-caused tree mortality. (d)–(f) Predicted probability of mountain pine beetle
cold survival across the range of pine species in the United States and Canada during the same three climate
periods. Based on the response of mountain pine beetle to temperature, results suggest a low to moderate
probability of range expansion across Canada and into central and eastern US forests by the end of the
twenty-first century. Source: From Bentz et al. (2010) / with permission of Oxford University Press.
Range resulted in an earlier and longer dispersal period for mountain pine beetle and increased the
life cycle from one to two generations per year. The potential for two generations is probably
limited to warmer forests in North America until at least the middle of the twenty-first century
(Bentz et al. 2019), but continued rising temperatures may eventually expand the potential to other
forests as well. The effect of warming temperatures on insect development is not always straight-
forward, because some species have a winter-dormancy requirement that may become less likely
to be met, which could upset insect development (Ayres and Lombardero 2000). Thus, climate
change is not likely to affect all insect species in the same way. Pathogens tend to be more affected
by moisture than by temperature because moisture is required for reproduction and dispersal
(Desprez-Loustau et al. 2006), although the effects vary by species.
Climate change is also important in affecting the survival of insects and pathogens because
it affects winter temperatures that may influence overwinter survival. Winter temperature is
known to be a strong determinant of many insect species’ ranges, such that small rises in winter
temperature could permit such species to expand into previously unsuitable forests (Figure 20.18).
Northward expansion of several insect species due to increased winter temperatures has been doc-
umented. For example, the southern pine beetle was historically limited to pine forests of the
southeastern United States, but has moved northward into New Jersey, New York, and New
England following a rise of winter temperatures there of 3–4 °C (Dodds et al. 2018). Similarly, an
enormous mountain pine beetle outbreak in British Columbia between the mid-1990s and 2017
(Chapter 16) is likely related to consistently warmer temperatures there that began in 2003, and
even included tree species that are not normally susceptible to this insect (Logan et al. 2010).
Finally, the hemlock woolly adelgid, an invasive insect that attacks eastern hemlock and was
limited to the mid-Atlantic in the 1970s has expanded northward into central New England, where
its further spread has been significantly slowed by winter temperatures (Paradis et al. 2008). Nota-
bly, the short life cycle, sensitivity to temperatures, reproductive potential, and mobility of invasive
insects and pathogens will likely result in rapid adaptation to climate change relative to native
species (Newman et al. 2011).
Many insects and pathogens most successfully attack weakened trees, and a warmer climate
is likely to have marked effects on tree vigor. Drought, characterized by reduced precipitation and
increased temperatures, is the climate effect most commonly linked to insect and pathogen attack
of stressed or low-vigor trees because it reduces the efficacy of tree defense (Raffa et al. 2008; Bentz
et al. 2010). In turn, reduced tree defense across broad areas may result in outbreaks. For example,
a study by Gaylord et al. (2013) examined the relationship between drought, tree defense, and
insect attack in forests dominated by pinyon pine and one-seed juniper in New Mexico by experi-
mentally manipulating precipitation. They found pinyon pine mortality to be the highest in the
treatment that simulated drought, with pinyon ips bark beetle, which generally attacks stressed or
recently dead trees, present in 92% of dead trees. Juniper was not susceptible to beetle attack but
suffered canopy loss after 3 years of drought. Notably, the increased incidence of extreme weather
events related to climate change also has the potential of reducing tree vigor and increasing sus-
ceptibility to insect and pathogen attack, for example, via damage by heavy snow, ice storms, or
lightning (Ayres and Lombardero 2000).
WIND
Hurricanes drive the structure and composition of forests of Atlantic and Gulf coastal areas, but
smaller-scale wind events such as downbursts, derechos, and tornadoes are a fundamental
component of the disturbance regime of most deciduous forests in eastern North America
(Peterson 2000). As discussed in Chapter 16, wind events of varying severity kill trees and disturb
forest canopies, thereby initiating gap dynamics and altering stand structure and greatly impacting
tree regeneration, establishment, and growth. It remains difficult to predict the general response of
forests to wind events because forest types and tree species vary widely in their susceptibility to
Climate Change Effects on Forest Carbon 571
wind damage, and susceptibility also changes with forest age and density. However, it is certainly
sound speculation that increasing intensity, duration, or frequency of wind events would increase
disturbance to forests. Heavy precipitation events, often associated with thunderstorms, have
become more frequent and intense in the United States since the 1990s (Easterling et al. 2017),
particularly in eastern deciduous forests in the Midwest and Northeast. Given that downbursts and
derechos arise from thunderstorms and thunderstorm conditions contribute to tornado formation,
it is likely that such events will continue to increase in frequency and severity. Gregow et al. (2017)
found increased windstorm damage to European forests between 1951 and 2010, with a threefold
increase in average intensity of windstorms after 1990 attributed to climate change.
Perhaps equally important as direct effects, however, are indirect effects of climate
change-induced increases in wind damage to forests, such as the effects of changing temperature
and precipitation on the resistance of individual trees and forests to wind. Where winter tempera-
tures are increasing, normally frozen soils that provide stable anchorage for trees during the winter
months may thaw, leaving them susceptible to windthrow. Such effects have been a strong empha-
sis of forest research in Europe. For example, Usbeck et al. (2010) found that windstorm damage to
forests increased by 22 times since 1858 in Switzerland, and that the most severe wind damage
occurred when soil conditions prior to storms were thawed and wet. In this region, weather condi-
tions attributed to climate change included much warmer and wetter winter conditions and higher
maximum gust wind speeds. Blennow et al. (2010) modeled wind effects on forests in Sweden with
particular attention to changes in frozen soils, finding that a changed wind climate and soil thaw-
ing increased the probability of wind damage especially in southern compared to northern forests.
Climate change may also indirectly influence wind damage in forests by changing tree growth and
forest dynamics. A modeling study of radiata pine plantations in New Zealand found that increased
tree growth under climate change scenarios increased susceptibility to wind damage (Moore and
Watt 2015). In addition to experiencing higher wind speeds, trees grown under increased CO2
emissions grew taller but not larger in diameter, leaving them more susceptible to breakage by
wind, particularly when grown at high density.
An important point about climate change effects on forest disturbances is that disturbances
interact, such that changes in the frequency, intensity, duration, extent, or seasonality of one
disturbance may affect the same characteristics of the next (Seidl et al. 2017, Chapter 16). For
example, increases in wind damage or insect outbreaks brought about by climate change may
increase fuel availability, leading to larger and more severe fires. Likewise, increases in pathogens
may reduce root or stem stability, heightening the risk of wind damage. Climate change will affect
all disturbances within any disturbance regime in both obvious and unforeseen ways, such that
precisely predicting future forests will remain very difficult.
CLIMATE CHANGE EFFECTS ON FOREST CARBON

Climate change is often discussed today in terms of its impacts on ecosystem services provided by
forests. Ecosystem services affected by climate change span a wide range of ecological processes,
such as diversity (Chapter 14), forest resilience to disturbances (this chapter), and carbon storage.
Forests are an enormous pool of carbon in the global carbon cycle, which is important because
carbon stored in biomass is carbon not found in the atmosphere acting as a greenhouse gas (Pan
et al. 2011). In the context of climate change, there is therefore great interest among ecologists in
carbon sequestration in forests, or their ability to uptake carbon from the atmosphere, fix it into
biomass, and thereby store it within an ecosystem. Greatly simplified, the amount of carbon stored
in forest ecosystems is largely the result of the balance between the amount of carbon taken up
into ecosystems via photosynthesis (net primary productivity or NPP) and the amount lost from
ecosystems via decomposition or heterotrophic respiration. When NPP is greater than decomposi-
tion, an ecosystem acts as a sink of carbon, but it acts as a source of carbon to the atmosphere when
decomposition is greater than NPP. In determining the location of carbon, this carbon balance is
extremely important for the global carbon cycle, and anything that changes the balance, such as
disturbance, human activities, climate change, or their interactions, can alter atmospheric CO2.
The carbon balance of forest ecosystems is discussed in detail in Chapter 18, and understanding
how climate change may affect the components of carbon balance in forests is the brief focus here.
CLIMATE CHANGE EFFECTS ON CARBON GAIN: PRIMARY

PRODUCTIVITY
In the context of climate change, we discussed in the earlier text that tree growth generally
decreases with warming temperatures but increases in the presence of elevated CO2. Individual
tree growth generally decreases with warming temperatures where water is limiting, but rising
CO2 and the increased water use efficiency associated with it may outweigh the decrease. These
trends in tree growth are scalable to ecosystems, in that NPP increases with rising temperatures
when water is available but decreases with drought or reduced precipitation. Likewise, the rate of
nutrient cycling and availability may potentially increase with warming temperatures depending
on moisture availability, sustaining positive growth responses to elevated CO2 (Norby et al. 2010).
These physiological mechanisms are coupled to phenology, in that lengthened growing seasons
resulting from the onset of earlier springs and later falls are likely to increase forest productivity
(Myneni et al. 1997; Keenan et al. 2014). Productivity responses to climate change are notoriously
variable among regions, however. In general, forest prod‑‑uctivity in cooler and colder climates
tends to respond favorably to warming (Rustad et al. 2001), but that in warmer and drier climates
often does not (Peñuelas et al. 2007). However, Charney et al. (2016) used a modeling study to proj-
ect changes in forest productivity with rising CO2 in North America, finding negative impacts in
the interior West but positive impacts on coastal forests in the West, Southeast, and Northeast.
Studies of tree rings at the end of the twentieth century had also shown variable importance of
warming temperatures, rising CO2, and reduced precipitation among regions in the United States,
with some factors more important in some regions than others (e.g., Graumlich et al. 1989; Peter-
son et al. 1990). Overall, primary productivity will probably respond favorably to climate change in
the near future.
An important point is that productivity is not affected independently by warming tempera-
tures, rising CO2, or reduced precipitation. In fact, each of these climate change effects occur con-
currently, and their interactions may have different implications for forest productivity (Newman
et al. 2011). Warming temperatures, for example, might increase rates of nutrient cycling in cool
and cold climates that are nutrient-limited, but they may simultaneously reduce soil moisture via
evaporation. At the same time, even while warming temperatures may reduce soil-moisture avail-
ability, rising CO2 increases water-use efficiency such that moisture stress is less acute or unimpor-
tant to growth. Site conditions, of course, will also play an important role in providing the context
for potential changes in productivity with climate change. Cool, moist ecosystems such as those in
coves or northern and eastern aspects are less likely to experience soil-moisture deficits in the
presence of warming temperatures, and warmer, drier ecosystems or those on droughty soils are
more vulnerable to rising temperatures, heavily influencing their primary productivity.
CLIMATE CHANGE EFFECTS ON CARBON LOSS: HETEROTROPHIC

RESPIRATION
Primary productivity is only part of the carbon cycle, and the impacts of climate change on hetero-
trophic respiration are equally important for forest carbon storage. Enormous carbon pools are
found in dead organic matter in soils, litter, and deadwood, and decomposition of this organic
matter constitutes a release of carbon from ecosystems back into the atmosphere as CO2. Increases
in decomposition due to climate change could therefore create a positive feedback, as it would
c ontribute greenhouse gases that would further exacerbate the process (Davidson and Janssens 2006).
Decomposition is critical in making nutrients available for tree growth and thus for increasing pri-
mary productivity, but such increases in productivity are meaningless for carbon sequestration if
the amount of carbon lost through decomposition equals or exceeds that gained through primary
productivity. The process of decomposition and its role in nutrient cycling are detailed in
Chapter 19. Here, we briefly review potential effects of climate change on decomposition, as it
affects the carbon cycle.
Carbon will accumulate in the soil if the rate of decomposition is slower than the rate of car-
bon inputs from litter and dead vegetation. If primary productivity increases as a result of climate
change, the amount of carbon available for decomposition will increase. Thus, the way climate
change affects decomposition—through changes in temperature and soil-moisture availability—
will greatly affect how forests store and cycle carbon. Microbial activity is clearly affected by temper-
ature, and many ecologists have therefore assumed that warming temperatures will, in turn, increase
decomposition rates. Microbial activity is also sensitive to soil moisture, however, and warming
temperatures that lead to higher evaporation and thus drier soils may slow decomposition rates,
depending on the soil-moisture conditions that result from the drying (i.e., decomposition in poorly
drained soils may increase as they dry). Litter decomposition rates are usually low when mean
annual temperatures are lower than about 10 °C, but much more variable when warmer than 10 °C
(Figure 20.19), suggesting that factors in addition to temperature and precipitation are also impor-
(a)
5
3
k value
0
–10 0 10 20 30
MAT (°C)
(b)
5
3
k value
0
0 2000 4000 6000
MAP (mm)
F I G U R E 2 0 .1 9 Leaf litter decomposition rates (k value) with (a) mean annual temperature, and (b) mean
annual precipitation for an analysis of 70 publications of litter decomposition. Source: From Zhang et al.
(2008) / with permission of Oxford University Press.
tant for decomposition (Zhang et al. 2008). Overall, much like primary productivity, the rate of
decomposition is likely to rise with warming temperatures, but only if soil moisture remains avail-
able (Aerts 2006), and in concert with other factors such as litter quality (see the following text).
Notably, several soil-warming experiments in forests have shown that increased decomposition
rates may be limited in duration if litter input rates do not also increase, because available soil car-
bon will be effectively respired away (e.g., Bradford et al. 2008b; Bronson et al. 2008).
The effect of warming temperatures on decomposition is likely to be especially critical near
the poles in permafrost areas—both forested and non-forested—where soils remain frozen
year-round. The world’s permafrost is extremely rich in soil carbon, estimated to hold twice as
much carbon as is currently present in the atmosphere (Schuur et al. 2015), but most of it is
unavailable to microbial activity and decomposition when frozen. As warming temperatures cause
permafrost to thaw, decomposition of formerly unavailable carbon in the soil and its release into
the atmosphere are likely to create an enormous positive feedback to climate change (Turetsky
et al. 2019). Similarly, peatlands, many of which are forested and also in boreal and polar regions,
contain an enormous amount of soil carbon that decomposes very slowly because of the anaerobic
conditions found in waterlogged soils. If warming temperatures are accompanied by drier condi-
tions, drying of peatlands could aerate soils and make carbon available for decomposition. In the
case of both permafrost and peatlands, it is unknown how much of the carbon could be released to
the atmosphere through decomposition, how quickly the release may occur, or how much primary
productivity might compensate for the release (Turetsky et al. 2019), but these examples of
temperature-induced changes on decomposition may demand the most attention from ecologists
in the coming decades.
In considering leaf litter, the rate of decomposition also depends on the compound being
decomposed, and thus substrate quality is likely to influence climate change effects on decomposi-
tion (Davidson and Janssens 2006). Global rates of litter decomposition are primarily driven by
litter quality (Cornwell et al. 2008); plant species producing litter with high nutrient concentra-
tions and lower concentrations of compounds more difficult to decompose (e.g., lignin) tend to
decompose more rapidly (Chapter 19). Therefore, the potential for climate change to alter the
quality of litter produced by tree species could be important for decomposition rates and loss of
carbon from ecosystems. Relatively few studies to date have examined this relationship experimen-
tally. In central Spain, experimentally increased temperature, reduced precipitation, and a
combination of the two factors each reduced decomposition of a dry shrub species, attributed to
reduced leaf quality (Prieto et al. 2019). Where temperature was increased, leaf litter had reduced
concentrations of phosphorous and iron and higher C:P and C:N ratios—the latter of which stim-
ulated nitrogen immobilization by microbes, thus decreasing decomposition rates. If such trends
are consistent among other woody species across regional ecosystems, such reduction in litter
quality could significantly reduce carbon losses from soil in a changing climate.
Despite the possibility that climate could alter the quality of litter in a given species, the
changes in forest species composition brought about by climate change are most likely to have the
strongest effects on litter quality within an ecosystem (Newman et al. 2011). It is well known, for
example, that deciduous tree litter (such as that of maples) is more rapidly decomposed than conif-
erous litter (such as that of pines), at least in the ecosystems where each type of tree is found
(Gholz et al. 2000). Likewise, deciduous litter decomposition rates vary by species (e.g., sugar
maple leaves decompose faster than northern red oak leaves). Therefore, any changes in tree
species composition on a given site might be expected to bring about changes in decomposition.
For example, Alexander and Arthur (2014) examined changes in decomposition rates within
Kentucky oak forests invaded by red maple due to fire suppression by manipulating leaf litter in
field plots. Over only 18 months, red maple leaf litter initially decomposed much faster than litter
from scarlet or chestnut oak, decreased nitrogen immobilization, and slowed mineralization.
Although this experiment did not directly examine the effects of climate change and was conducted
under current climate conditions, it reiterates a tight link between decomposition rates and species
composition and suggests that decomposition rates are likely to respond quickly to climate-induced
changes in species composition.
Finally, in addition to leaf litter, coarse woody debris is a major carbon pool representing
10–20% of heterotrophic respiration in mature forests (Bond-Lamberty et al. 2004), and the carbon
it contains has a longer residence time before it is lost to the atmosphere. There is relatively little
information about wood decomposition compared to that for leaf litter in forests (Harmon
et al. 2011), and thus even less evidence for an influence of climate on the decomposition of wood.
Overall, most studies find that the specific characteristics of the wood being decomposed—such as
nutrient content, size, and tree species—are more important in explaining variations in the rate of
decomposition than are variations in climate. In general, only 20–30% of the variation in wood
decomposition is explained by climate both regionally (Bradford et al. 2014) and globally (Hu
et al. 2018), and the majority of the variation is explained by wood characteristics. Understanding
climate-induced changes in wood decomposition as it affects carbon loss from ecosystems is clearly
an important area of additional research.
FEEDBACKS AMONG DISTURBANCE, CLIMATE CHANGE, AND CARBON

IN FORESTS
Disturbances in forests kill trees and, in most cases (except for harvesting), do not remove the dead
biomass, such that they are likely to alter the balance between primary productivity and decompo-
sition and alter an ecosystem’s carbon storage. Depending on the disturbance type, carbon may be
immediately lost from the ecosystem; for example, fire oxidizes biomass and immediately re-
moves carbon from the system, but insect outbreaks more slowly kill trees and do not create
an immediate loss. Regardless of disturbance type, however, an ecosystem’s carbon balance
(net ecosystem productivity or NEP; Chapter 18) will shift from sink to source and back to sink
as the forest recovers as a result of the balance between primary productivity and decomposition
(Figure 18.18, Chapter 18). For a simple example, a stand-replacing fire in a mature conifer forest
will shift the ecosystem from a carbon sink prior to the fire to a source after the fire (Kashian
et al. 2006). Immediately following the fire, live vegetation is lacking such that primary production
is low, but dead vegetation is prominent such that decomposition is high, and NEP is negative (i.e.,
the system is a carbon source; Figure 20.20). As post-fire succession proceeds, carbon sequestered
by young, vigorous vegetation eventually exceeds that lost from decomposing dead vegetation, and
NEP is positive (i.e., a carbon sink). Primary production decreases over decades as the forest ages,
most or all of the post-fire dead vegetation decomposes away, and NEP approaches zero until the
next disturbance. Thus, the landscape acts as a small carbon sink over many years, but as a strong
carbon source for a few decades after it has burned. Provided that there is enough time for forests
to recover between disturbances, most forests recover most of the carbon loss resulting from a
given disturbance (Kashian et al. 2013). Post-disturbance forest carbon dynamics such as these are
discussed in detail in Chapter 18.
Changes in forest carbon storage with succession and associated shifts between carbon sources and
sinks have probably occurred at least since the Pleistocene. In the context of anthropogenic climate
change, however, carbon storage is a larger concern because of its implications for atmospheric
CO2 and because of potential effects of climate change on disturbance regimes. This is especially
true when the effects of changing disturbance regimes are considered across large, forested
landscapes. In the example cited in the earlier text, more frequent fires would cause more of
the landscape to act as a carbon source, which cumulatively would release more carbon into the
atmosphere at a given point in time. Perhaps more concerning regarding carbon storage are the
effects of climate change on post-disturbance recovery of forests. If, for example, more frequent or
severe fires converted many high-density forests on a landscape to low-density forests, as has been
ΔC Vegetation
Total
NEP
(g C yr-1)
ΔC
ΔC Dead wood
Time since stand-replacing fire (years)
F I G U R E 2 0 .2 0 Hypothetical trajectory for forest carbon storage (net ecosystem production; NEP) after a
stand-replacing fire. NEP is negative immediately after the fire because sparse vegetation has little carbon
uptake while large amounts of dead vegetation have high carbon loss (ΔC is strongly negative) from
decomposition. NEP soon increases, reflecting high carbon uptake in live vegetation (ΔC is strongly positive)
as trees and other vegetation re-establish. NEP eventually decreases toward zero as low carbon uptake by
vegetation is balanced by low carbon loss from dead wood. Source: From Kashian et al. (2006) / with
permission of Oxford University Press.
documented for lodgepole pine in the Greater Yellowstone Ecosystem (Turner et al. 2019), carbon
loss will occur for the landscape because the fires will create much dead vegetation to decompose
but reduce the amount of vegetation available to uptake carbon. An even more extreme example is
the potential for conversion of forest to non-forest, due to increased fire size or severity, short-
interval fires, or less-suitable post-fire growing conditions. Although forests eventually recover all
of their lost carbon after a disturbance, such recovery requires that forests remain forests;
conversion to other cover types would result in a net loss of carbon from the landscape. Such
dramatic cover type conversions were once considered only remotely possible (Kashian et al. 2006),
but have become considerably more common as climate change progresses.
Because of their ability to affect carbon storage, insect outbreaks are also an important dis-
turbance affected by climate change. Insect outbreaks are second only to wildfires as the largest
source of tree mortality in western North America (Samman and Logan 2000). We have described
in the earlier text that insect outbreaks are likely to increase in frequency, extent, and severity for
some insect species as development accelerates with rising temperatures, overwinter survival
increases with rising winter temperatures, and host trees become stressed by moisture deficits. The
response of forest carbon storage to insects is not dissimilar to its response to fire, although there
exists tremendous variability in carbon storage response across different types of insects (Hicke
et al. 2012). In general, much like fire, insect outbreaks reduce primary productivity and shift car-
bon pools from live vegetation to deadwood as they kill trees, in some cases switching the ecosys-
tem to a carbon source. Depending on the severity of the outbreak (proportion of trees killed),
forest recovery following insect outbreaks is typically much quicker, and thus carbon sources are
generally much shorter-lived. For example, Romme et al. (1986) found that lodgepole pine forests
returned to pre-disturbance levels of primary productivity after only 5–15 years following moun-
tain pine beetle outbreaks in Yellowstone National Park. Similarly, Kashian et al. (2013) were
unable to detect differences in ecosystem carbon between mature lodgepole pine stands attacked
by mountain pine beetle and unattacked stands by 25–30 years after the outbreak. For comparison,
they found that the same ecosystems required about 250 years to recover most of the carbon follow-
ing stand-replacing fires. The rate of recovery of carbon reflects the severity of the outbreak, with
stronger sources and longer recovery times associated with more severe outbreaks. Even severe
outbreaks rarely kill every tree or even most trees. The surviving trees are critical as seed sources
for post-disturbance regeneration, and to partially compensate for carbon losses with rapid growth
following release from competition with trees killed by the outbreak (Hicke et al. 2012; Han-
sen 2014). Compensatory contributions of surviving trees to post-outbreak primary productivity
are a main reason that outbreaks have a reduced impact on carbon storage compared to fires.
Despite reduced impact compared to fires, insect outbreak-induced carbon losses from for-
ests are certainly not trivial. For example, in its worst year, the enormous bark beetle outbreak that
occurred in British Columbia in the early 2000s produced carbon emissions equal to about 75% of
those from forest fires across all of Canada over a 40-year period and reduced primary production
by the same amount that it increased in the 1980s and 1990s due to climate change (Kurz
et al. 2008a). Notably, Thom et al. (2017) used a simulation model to find that bark beetle activity
in an Austrian national park increased with climate change in the twenty-first century, but that
changing tree species composition reduced host tree availability and decreased insect activity in
the long run. Despite these results and the bevy of unknown factors that may shape future biotic
disturbances, it is likely that continued and increasing insect outbreaks are likely to decrease forest
carbon for at least the next several decades (Williams et al. 2016).
Fire, Carbon, and Climate Change in Forests of Yellowstone National Park The
lodgepole pine forests of Yellowstone National Park in northwest Wyoming are an excellent
example of how climate change may interact with disturbance regimes and carbon storage. Large,
stand-replacing fires, comparable to those that burned 3200 km2 of the park in 1988, have infre-
quently burned this landscape on a rotation of 100–300 years. Kashian et al. (2013) used a repli-
cated chronosequence approach (Chapter 17) to measure all the carbon pools in 77 lodgepole pine
forests aged 12 to >350 years having varied structures (tree densities) to characterize changes in
carbon storage following large fires. They were unable to detect a carbon source (negative NEP),
although the youngest stands in their analysis were 12 years old, suggesting that any carbon source
is relatively short-lived on this landscape. This carbon source, if it exists, is probably short-lived
because of the rapid regeneration of the burned area by lodgepole pine via serotinous cones and
slow decomposition of deadwood in the cold, dry subalpine landscape. The duration of a carbon
source would likely be longer on a landscape dominated by non-serotinous species or where
decomposition was more rapid. The recovery of ecosystem carbon was driven mainly by primary
productivity, as carbon gain in live vegetation (7780 g C/m2) greatly exceeded carbon lost through
decomposition (1960 g C/m2) over the first 100 years. Most carbon accumulation occurred early;
NEP was not statistically different from zero for forests older than 70 years (Kashian et al. 2013).
As discussed in the earlier text, most forests are able to recover carbon lost as a result of dis-
turbance so long as forests regenerate rather than undergo a significant state change to another
cover type (Kashian et al. 2006). However, the ability of a forest to recover after a disturbance
requires a disturbance-free interval to which a particular forest type is adapted. In Yellowstone,
Kashian et al. (2013) found that 90% of total ecosystem carbon in lodgepole pine forests was recov-
ered in the first 100 years after fire, making landscape carbon well buffered for fires that recur every
100–300 years (Figure 20.21). Likewise, coniferous subalpine forests in Colorado were found to
recover most of the carbon lost to disturbance after 100 years for fires that recur every 150–400 years
(Bradford et al. 2008a), and jack pine forests in Michigan after about 20 years for fires that recur
every 30–80 years (Rothstein et al. 2004). In these forests, carbon lost to disturbance is recovered
prior to the next disturbance, and there is no net carbon lost to the atmosphere. Of issue, however,
300
25 000
C accumulation rate (g C·m–2·yr–1)

250
Total ecosystem C (g/m2)
20 000
200
15 000
150
10 000
100
5000
50
0 0
0 100 200 300 0 100 200 300
Stand age (yr) Stand age (yr)
F I G U R E 2 0 .2 1 Total ecosystem carbon (left) and carbon accumulation rate (right) for 77 lodgepole pine
stands in Yellowstone National Park. Total ecosystem carbon increases with stand age, but 90% of the
increase occurs before age 100 years (blue dashed line). Carbon in the ecosystem was resilient to fire because
fire frequency was historically 100–300 years (shaded box). However, fire rotation is expected to increase to
30 years (red dashed line) by 2050, which would prevent the system from recovering more than 75% of its
carbon after a large stand-replacing fire. Source: From Kashian et al. (2013) / John Wiley & Sons.
is the possibility that climate change could shorten the fire rotation (or the rotation of non-fire dis-
turbances on other landscapes), such that subsequent disturbances occur before carbon is able to
recover. In Yellowstone, current projections allow for fire rotations as short as 30 years in the next
several decades (Westerling et al. 2011), which would significantly reduce the amount of forest
carbon on the landscape. Such short-interval fires would almost certainly lead to forest conversion
to non-forest cover types (e.g., Harvey et al. 2016b; Turner et al. 2019), a significant net loss of car-
bon from the landscape, and potential feedback to further climate change.
Although our discussion of climate-disturbance–carbon interactions here is limited to fire
and insect outbreaks, we emphasize that other disturbance types could also produce significant
carbon losses from forested landscapes if altered from their current regimes by climate change.
Wind and drought are two such examples. A third disturbance type, timber harvest, is extremely
important in its potential to reduce carbon storage in forests despite its anthropogenic nature, and
it leaves a long legacy of effects. We discuss timber harvesting in the following text; more detailed
treatments of climate change and disturbance interactions for carbon storage are presented in
Williams et al. (2012, 2016), Seidl et al. (2014), Vanderwel et al. (2013), and Kurz et al. (2008b),
among others.
ADAPTING TO CLIMATE CHANGE EFFECTS ON FORESTS

There are several reasons to expect human responses to climate change. First, modern climate
change is largely anthropogenic, and it is human nature to want to correct past mistakes. Second,
forests provide a wide variety of values (such as timber resources but also clean water, wildlife hab-
itat, recreation, scenic beauty, cultural values, etc.) that many ecologists and managers feel strongly
about preserving for future generations. Responses to climate change might be categorized as
either climate change mitigation, where emphasis is placed upon reducing carbon emissions from
fossil fuels and attempting to stabilize or reduce the concentrations of greenhouse gases in the
Adapting to Climate Change Effects on Forests 579
atmosphere, or adaptation, where ecologists and land managers develop strategies that assist eco-
systems in coping with current and future climate change impacts. These responses have impor-
tantly different goals. Mitigation attempts to reduce or avoid human interference with climate to
lessen the rate of climate change so that ecosystems may naturally adapt while maintaining food
production and economic development (IPCC 2014). By contrast, adaptation has as its goal
reducing vulnerability to the harmful effects of climate change. Thus, adaptation involves adjust-
ments by humans to current or expected changes in climate and is our major emphasis in this
section. As we have discussed, forests are a major player in the global carbon cycle, such that the
ultimate and most obvious adaptation to climate change in forests is to avoid deforestation at a
global scale. Our emphasis is not upon the obvious, but instead on a few major strategies of adap-
tion to climate change effects in forests: assisted migration, the use of climate refugia, and carbon
management. There include a host of other equally important adaptation strategies that are not
discussed in this chapter. Detailed treatments of such strategies are presented in Halofsky et al.
(2018), Swanston et al. (2016), Millar et al. (2014), and Janowiak et al. (2014), among many other
resources.
ASSISTED MIGRATION
The process of intentional, human-assisted movement of species or genotypes in response to cli-
mate change is called assisted migration, also called assisted colonization or managed relocation
(Williams and Dumroese 2013). The ecological basis for assisted migration is the notion that the
rate of climate change is likely to exceed the rate and extent of tree dispersal to newly suitable hab-
itat, and the speed and distance at which climate zones will change will not allow species the time
to relocate to suitable environments. As a result, many tree species are likely to be overtaken by an
unsuitable climate and potentially be outcompeted by more quickly dispersing species (Minteer
and Collins 2010, Vitt et al. 2010). Impediments to movement of tree species may include gradual
rates of adaptation, but also may include physical barriers, such as fragmentation or geographic
features, within a species’ current range. Because it may lifeboat (Chapter 22) at-risk species or
populations at risk due to climate change, assisted migration is considered by many to be an impor-
tant tool for endangered species conservation (Hoegh-Guldberg et al. 2008) and building resilience
when restoring forests (Dumroese et al. 2015).
In practice, assisted migration has been efficiently categorized into three forms, depending
on the distance over which movement occurs (Figure 20.22; Handler et al. 2018). Assisted
population migration, also called assisted genetic migration or assisted gene flow, involves within-
range movement of seed sources or populations to new locations. Assisted population migration is
an important tool for helping species to overcome physical barriers to movement within their
current range. Assisted range expansion is defined as the movement of seed sources or populations
from their current range to suitable areas just beyond the historical species range. Assisted range
expansion has the goal of facilitating or mimicking natural dispersal, particularly when human
activities have created barriers to such dispersal. Assisted species migration, also called species
rescue, managed relocation, or assisted long-distance migration, moves seed sources or popula-
tions to locations far outside the historical species range, beyond locations accessible by natural
dispersal. All three forms of assisted migration are meant to ensure that a species occurs in many
redundant locations or across a range of conditions such that risks of unforeseen climate impacts
are reduced (Handler et al. 2018).
The idea of assisted migration—particularly assisted species migration—has not been with-
out controversy. Some ecologists have argued that anthropogenic-based migration of species may
facilitate the effects of invasive species on other plants and ecosystems when the assisted species
become established in new areas (Ricciardi and Simberloff 2009). Other potential challenges to
Natural
Migration
Boundary
Assisted Population Assisted Range Assisted Species

Migration Expansion Migration
F I G U R E 2 0 .2 2 The three forms of assisted migration. Assisted population migration transfers seed sources
of populations of a species to new locations within its current range. Assisted range expansion moves seed
sources or populations from their current range to suitable areas just beyond the historical species range to
mimic natural dispersal. Assisted species migration transfers seed sources or populations to locations far
outside the historical species range, inaccessible by natural dispersal. Source: From Handler et al. (2018) /
U.S. Department of Agriculture / Public Domain.
assisted migrations are also notable. For example, species established in new areas may hybridize
with local species if they are capable of doing so, as in the spruces, pines, aspens, and oaks (Aitken
et al. 2008), increasing the influence of humans on genetic diversity. The introductions of new
species into an area may introduce or attract previously absent insects or pathogens. Pragmatically,
moving species long distances outside of their current range, even when climate projections predict
that such locations will provide suitable growing conditions by the end of the century, may not be
successful if current growing conditions at those locations are not yet suitable (Handler et al. 2018).
Assisted migration can also be designed to maintain high levels of productivity and
diversity in commercial forests (termed forestry-assisted migration; Pedlar et al. 2012). Forestry-
assisted migration emphasizes moving populations (usually seed sources) within the current
range of the species (assisted population migration) or expanding ranges along the margins of
current geographical limits (assisted range expansion; Gray et al. 2011; Williams and Dumro-
ese 2013). Assisted population migration has been incorporated into forestry practice in parts
of Canada (Pedlar et al. 2011), led by British Columbia, and in the United Kingdom (Whittet
et al. 2016).
Perhaps the best example of the complexities of assisted migration is that of whitebark pine,
whose seeds were tested for germination and establishment far north of their current native range
(McLane and Aitken 2012). Whitebark pine is a high-elevation conifer of the Rocky Mountains
that is an important food source for wildlife and is currently threatened by both warming
t emperatures and an exotic pathogen (white pine blister rust). Seeds were planted from 7 popula-
tions in 8 locations bracketing the current northern (800 km north of the current boundary) and
southern (600 km south of the current boundary) boundaries, and data were collected after
10 years. Germination occurred at all locations, but germination, establishment, and seedling
growth were affected by seed size, winter snow cover, and the timing of snowmelt much more than
latitude. In fact, the seedlings at the northernmost site were twice as tall on average as those on any
other site in the study. Seedlings on sites with snowpacks at one of the extremes (i.e., either little
or no protective snow cover in winter, or deep snowpacks that melted late in the year) had poorer
rates of survival than on sites where snowpacks were continuous in winter and melted in April or
May. Notably, the northernmost site with the tallest seedlings was also near the Pacific Coast and
had relatively mild mean annual temperatures, the warmest average summer temperature of all
other sites, and had intermediate snowfall (McLane and Aitken 2012). These results suggest that
climate is not likely to be the sole driver of successful assisted migration, and that local site condi-
tions are clearly important.
The information required for high confidence of success with assisted migration includes
much more than simple changes in climate envelopes. Climatic extremes outside of the historical
range of variability may be more important to assisted migration than averages; phenological rela-
tionships with temperature, photoperiod, and precipitation are likely to be extremely important in
determining the establishment and survival of relocated species; and local effects of microclimate,
herbivory, and competition of a species in a new environment cannot be ignored (Park and Tal-
bot 2018). If eventual adoption of the approach is widespread, assisted migration will likely be
complemented with other approaches designed to increase forest resilience in the face of climate
change (see Janowiak et al. 2014) rather than serving as a stand-alone strategy. Detailed discus-
sions of assisted migration are presented in Park and Talbot (2018), Handler et al. (2018), Park
et al. (2014), Williams and Dumroese (2013), and Pedlar et al. (2012). A review of available assisted
migration experiments is provided by Sáenz-Romero et al. (2021).
REFUGIA
Glacial refugia have long been discussed as pockets of species distinct from their surroundings
that resulted from some sort of protection from the intense cold and fluctuations of glacial climates
(Chapter 15). During glaciations, these areas harbored populations during periods of unfavorable
climate in regions where the ice displaced suitable climates to the south or to lower elevations.
Glacial refugia also allow the species they contained to re-colonize the areas around them as the
climate warmed and the ice retreated. Species that thrive in warmer climates shrink into refugia
during glacial periods, but those that thrive in cooler climates are often restricted to refugia bet-
ween glaciations (Stewart et al. 2010). Today, as the climate continues to change, the latter refugia
concept has been applied to small areas that are relatively buffered from those changes, and thus
refugia may be effective in protecting especially sensitive species and ecosystems from climate
change over the short term (Morelli et al. 2016). These refugia are considered to be potentially
important foci of biodiversity conservation in an altered climate. They are not completely analo-
gous to the interglacial refugia described in the earlier text because the climate will be warmer and
the landscape will be fragmented, limiting dispersal (Ashcroft et al. 2012). Nevertheless, climate
change refugia have been described as a “slow lane” for species diversity because they safeguard
species for longer periods of time than those existing within the faster climatic changes of the
broader region (Morelli et al. 2020), although these relative differences may often be short-lived
(McLaughlin et al. 2017). However, even transient delays of climate change effects may be impor-
tant in allowing species additional time to adapt to the new climate, or, perhaps more importantly
and likely, to disperse.
Modern refugia often occur because of specific physical site conditions that locally influence
resources available to species. Climate change refugia generally have locally stable and persistent
climatic conditions despite changes that may occur around them (Ashcroft et al. 2012). Species
within the refugia do not experience the climate conditions in the broader area, at least in the same
way as the species outside of the refugia. The classic example of refugia are low areas in the
landscape such as depressions and valleys that pool cold air (see Chapter 8; Figure 7.2 (Chapter 7)
and Figure 20.8), particularly at night, resulting in a markedly different microclimate and often
markedly different species composition and phenology. Other climate change refugia may buffer
against moisture deficits, such as wetlands, riparian areas, or groundwater-fed springs. In moun-
tainous areas of the Northern Hemisphere, north- and east-facing slopes and draws are typically
shaded and are thereby cooler and wetter than their surroundings. Additional examples of climate
change refugia are presented in Figure 20.23. These refugia are often considered microrefugia;
macrorefugia have also been identified as broader areas occurring due to elevation and proximity
to coastal areas (Stralberg et al. 2018).
In addition to protection from rising temperatures and reduced moisture resulting from cli-
mate change, other factors such as extreme events, streamflow timing and rate, and disturbances
have also been associated with refugia (Figure 20.23; Krawchuk et al. 2020). In particular, distur-
bance refugia may include areas protected from disturbances that may increase in frequency or
severity under climate change. For example, in a landscape study of mixed-conifer forests burned
frequently by fires in the Sierra Nevada of California, Wilkin et al. (2016) found that cold-air pools
were not only important microrefugia from rising temperatures, but they burned less frequently
and at lower severity than the greater region. These reduced effects of fire are likely to have a
positive effect on the protective capacity of the refugia and suggest that potential disturbances to
refugia may override their buffering of rising temperatures or reduced moisture availability.
Much like assessing assisted migration, potential climate change refugia may be identified
but are truly only working hypotheses as climate change is ongoing. Like most other projections
Topographically complex terrain creates Deep snow drifts provide insulation

varied microclimates and increases to the surface below and provide water Valleys that harbor cold air pools
the likelihood that current climates later in the season. and inversions can decouple local climatic
will continue to exist nearby. conditions from regional circulation
patterns.
Canopy cover can buffer

local temperature maximums
and minimums throughout
the year.
Cold groundwater inputs

produce local cold-water
Poleward-facing slopes refuges in which stream
and aspects result in temperature is decoupled
shaded areas that buffer from air temperature.
solar heating, particularly
during the low solar Areas near or in large deep lakes
angles of winter and or oceans will warm more slowly
early spring. due to the high heat capacity of
water.
F I G U R E 2 0 .2 3 Examples of potential climate change refugia that may result from physical site characteris-
tics on a landscape. Source: From Morelli et al. (2016) / Morelli TL et al / Public Domain CC BY 4.0.
associated with climate change, identification of potential refugia has had to rely on simulation
modeling of climate and species distributions. Michalak et al. (2018) identified potential future
climate change refugia across North America using projected changes in the size and distributions
of current climate envelopes over time and different dispersal differences. They found that only
12% of North America could function as climate refugia, mostly at high elevations and in moun-
tainous areas. Existing protected areas on the continent disproportionately included potential refu-
gia, and refugia at lower latitudes and lower elevations were less likely to be included within pro-
tected areas. The refugia were also projected to shrink in size over time as climate change progresses,
reiterating their vulnerability to climate change themselves.
The maintenance or creation of refugia is increasingly included as a strategy in management
portfolios for climate change adaptation (e.g., Swanston et al. 2016). Approaches to such a strategy
may include: (i) identifying unique sites and prioritizing them for protection. Such sites might
include those with unique or diverse geology, physiography, soils, or biodiversity, particularly
those relatively undisturbed by humans. They might also be identified as potential sites for assisted
migration species that may be vulnerable elsewhere on the landscape; (ii) identifying and protect-
ing sensitive or threatened species and ecosystems, such as those near the southern boundaries of
their current range or restricted to a narrow range of site conditions. Methods to protect such
species are not unlike those typical of species conservation but may also include protection of
species and ecosystems across a range of environmental conditions and monitoring for potential
evidence of migration; and (iii) establishing artificial refugia. The development of highly con-
trolled environments such as nurseries, arboretums, or botanical gardens may be used to tempo-
rarily harbor rare or at-risk species until they can be moved to a new suitable habitat. Such an
approach mirrors that of assisted migration but uses “refugia” to maintain species before they are
relocated. Tactics include seed collection, seedling cultivation, and planting in a natural setting
where protection from climate change can be assured. Much like establishing general protocols for
conservation planning, effectively incorporating refugia into management, especially quickly
enough to keep pace with a changing climate, remains an important challenge.
FOREST CARBON MANAGEMENT

The largest driver of climate change today is the combustion of fossil fuels, but forests have
continuously been cited as important buffers in the global carbon cycle as it relates to atmospheric
CO2 (Harris et al. 2021). Management of forests has a strong influence on how they store carbon
because it involves manipulation of tree species composition, competition within and between
species, and stand structure in ways that shape primary productivity and, to a lesser extent, decom-
position. Centuries-old forest management decisions—mainly past timber-harvesting activities—
have left important legacies for carbon storage today. In the United States, forests currently act as a
major sink of carbon and have done so since about the 1950s (Figure 20.24). By 1915, forests in the
United States, most of them harvested mostly in the 1800s, were losing about 760 Tg of carbon each
year, but by 2010 were gaining about 200 Tg each year (Birdsey et al. 2006). The nature of this
increased forest growth, and the switch between carbon source and sink, is critical for future carbon
storage and thus for climate change. Factors such as nitrogen deposition and CO2 fertilization could
explain the increased forest growth and would probably result in a stronger carbon sink in the short
term. More likely, however, is that forests in the United States are responding to past and current
land use such as reforestation of old agricultural land or regrowth of harvested forests (Ryan
et al. 2010). In that case, the carbon sink will likely weaken as regrowth concludes, and the ability
of forests to buffer fossil fuel combustion will be reduced. Current projections of forest carbon
storage appear to show a weakening sink but suggest that effective forest carbon management may
be able to reduce the weakening (Figure 20.24). Thus, the size of the carbon sink is only one part of
the story, because the potential saturation of the sink is also fundamental to future c arbon storage.
1000
800
600
400
Tg C/yr
200
0
–200
–400
1700 1800 1900 2000 2100
F I G U R E 2 0 .2 4 Carbon emissions for forests in the United States for 1700–2100. Values greater than zero
represent a carbon source, whereas negative values represent a carbon sink. Projections beyond 2000 suggest
a weakening sink but could be stabilized with forest management for carbon (dashed line). Source: From
Birdsey et al. (2006).
Over the last two decades, forest managers have developed several silvicultural strategies
to increase carbon stores. A detailed discussion of these strategies is beyond our scope; examples
of such strategies as summarized by Ontl et al. (2020), Ryan et al. (2010), and Canadell and
Raupach (2008) include:
Avoid the loss of forests: As we have described in the earlier text, forests are one of the largest pools
of carbon in the terrestrial biosphere, such that deforestation results in an obvious release of car-
bon to the atmosphere. Of particular concern is conversion of forests to non-forest cover types,
which not only results in the loss of carbon storage in trees but also losses through increased
decomposition in the forest floor and soil. Avoiding deforestation may be achieved with
conservation easements or best management practices on forested land.
Increase the extent of forested land: Just as avoiding forest loss will avoid loss of carbon to the
atmosphere, adding new forest to unforested lands will increase carbon uptake and storage and
remove significant amounts of carbon from the atmosphere. Such afforestation might occur
where partial regeneration failure has resulted in very sparse forests, on marginal sites for agri-
culture or where forests have been cleared for crop land, or as riparian buffers.
Manage forests to reduce disturbances: Disturbances inevitably cause carbon losses from forests,
and enhanced disturbances caused by climate change will exacerbate this process. Fire is the
most influential disturbance in this regard, and several approaches may be used to reduce
the incidence of fire in forests. These include using prescribed fire to reduce fuel loads
and tree density or establishing an effective system of fuel breaks and fire lines to slow fire
spread or reduce its intensity. Altering forest structure by thinning is also an important approach
to reduce the threat of fire, lower the density of host species for a given insect or pathogen, reduce
the risk of drought-induced mortality, and favoring species less susceptible to wind or ice damage.
Thinning with an objective of increasing carbon storage should be conducted thoughtfully, as
thinning nearly always reduces forest carbon at landscape scales (Ryan et al. 2010). Note that
reducing disturbances in forests may be a competing objective with managing to mimic or incor-
porate the natural disturbance regime (see Chapters 10 and 16).
Facilitate post-disturbance recovery: Regeneration failure following disturbance, especially fires, is
likely to be a major driver of forest conversion and carbon loss in warmer and drier climates.
Examples of approaches that enhance rapid post-disturbance carbon uptake include quickly
r evegetating disturbed areas with seeding or planting, using native species likely to be well adapt-
ed to an altered climate, and protecting established seedlings and sapling from ungulate browsing
or tree harvesting activities.
Manage forests to increase forest growth: Increasing forest productivity has always been an objective
of forest managers, but in many cases current forest structure may require significant alterations
to increase carbon uptake and storage. In doing so, managers might plant or favor with silvicul-
ture species that are better adapted to future conditions, creating large openings to encourage
regeneration of shade-intolerant species that grow quickly, thinning to enhance age-or size-class
diversity, or promoting species with higher wood density (and thus higher carbon density), se-
lecting species with high growth rates, and controlling competing vegetation.
Many specific silvicultural techniques are available in managing for forest carbon storage. Some of
these are discussed by Franklin et al. (2018) and Ashton et al. (2012). D’Amato and Palik (2021)
place some of these strategies in the context of ecological forestry (Chapter 14). As the climate
continues to change, management techniques will likely need to continually adapt to additional
changes and conditions.
SUGGESTED
R E A D I N G S
Coop, J.D., Parks, S.A., Stevens-Rumann, C.S. et al. Ontl, T.A., Janowiak, M.K., Swanston, C.W. et al. (2020).
(2020). Wildfire-driven forest conversion in western Forest management for carbon sequestration and
North American landscapes. Bioscience 70: 659–673. climate adaptation. J. For. 118: 86–101.
Hicke, J.A., Allen, C.D., Desai, A.R. et al. (2012). Effects Seidl, R., Thom, D., Kautz, M. et al. (2017). Forest dis-
of biotic disturbances on forest carbon cycling in the turbances under climate change. Nat. Clim. Chang.
United States and Canada. Glob. Chang. Biol. 18: 7–34. 7: 395–402.
Iverson, L.R., Prasad, A.M., Matthews, S.N., and Peters, Westerling, A.L., Hidalgo, H.G., Cayan, D.R., and
M. (2008). Estimating potential habitat for 134 Swetnam, T.W. (2006). Warming and earlier spring
eastern US tree species under six climate scenarios. increase western US forest wildfire activity. Science
For. Ecol. Manag. 254: 390–406. 313: 940–943.
McKenney, D.W., Pedlar, J.H., Lawrence, K. et al. (2007). Williams, M.I. and Dumroese, R.K. (2013). Preparing
Potential impacts of climate change on the distribu- for climate change: forestry and assisted migration.
tion of North American trees. Bioscience 57: 939–948. J. For. 111: 287–297.
Invasive Species in Forest
CHAPTER 21
Ecosystems
T hroughout this book we have emphasized the relationship between plant species and physical
site factors, and that plants occur in ecosystems where they best compete and are adapted to
conditions of climate, physiography, soil moisture and nutrients, and microclimate. However, the
species occupying a given site are not necessarily those best adapted to grow on that site, but merely
those that have access to that site at the time of its availability and are superior competitors once
established. Invasion of a site by better competitors often results in substantial changes following
natural or human disturbances. As human effects on the ecosphere increase and become more
widespread, many of these better competitors are those introduced by humans, whose ecological
adaptations are to sites and ecosystems from other parts of the continent or world.
The distribution of biota of the ecosphere has been restricted by oceans and other natural
barriers for millions of years. During the last 125 years, human activities, especially international
travel and trade, have broken these barriers, allowing introduced or “exotic” species to invade new
continents at an increasing rate. Such invasive species—those species that arrive with human
assistance, establish populations, and spread (Simberloff 2013)—have dramatically altered forest
ecosystem diversity, function, and productivity. Many ecologists agree that the spread of invasive
species is now one of the most serious ecological threats to the sustainability and productivity of
native ecosystems (Mack et al. 2000; Pyšek et al. 2012). In a study of stressors to 1055 threatened
or endangered plants in the United States conducted almost 25 years ago, Wilcove et al. (1998)
found that 57% were impacted by invasive species; this proportion has almost certainly risen over
the last three decades. Economic losses to society due to invasive species in the United States are
estimated to be $120 billion each year, and losses of forest products alone are at least $2 billion
annually (Pimentel et al. 2005). Together with habitat loss and fragmentation (Chapter 22), reduc-
tions in biodiversity (Chapter 14), and climate change (Chapter 20), the threat of invasive species
is one of the most pressing issues faced by forest ecologists and managers in the twenty-first century.
In this chapter, we present multiple examples of the effects of invasive species on forest eco-
systems. We first examine the characteristics that make a plant species particularly invasive and
those of the ecosystems that appear to be particularly subject to invasion. We then discuss the
impacts of invasive species on forests, with special emphasis on the effects of invasive plants,
insects and pathogens, and animals on biota, soils, and disturbance regimes. The general threat to
forest ecosystems by invasive plants is summarized by Poland et al. (2021) and Liebhold et al.
(1995, 2017).
CONCEPTS OF INVASIVE SPECIES

DEFINITION OF INVASIVE SPECIES
Humans have caused many species to be introduced to new environments, but not all introductions
are deliberate, and not all introduced species become invasive. Many or even most new plant
species are introduced for horticultural or landscape uses, while many (not all) new insects are
587
588 Chapter 21 Invasive Species in Forest Ecosystems
introduced accidentally through cargo, packing materials, or even through transported plants or
soil. Invasive species occur along a general three-stage continuum: introduction, naturalization,
and invasion and spread. An introduced species is simply one that has arrived with either purposeful
or accidental human assistance, whether or not it becomes established and develops populations
in a new environment (Simberloff 2013). To become invasive, an introduced species must
(i) establish new populations, and (ii) spread widely from that point of establishment. Introduced
species that establish but do not spread widely are naturalized species. Naturalized species are per-
haps most common in highly disturbed areas or in heavily human-modified landscapes. Relatively
few woody species are considered naturalized but not invasive; one example is the common lilac,
which is hardy enough to survive for centuries once established but fails to spread widely because
it lacks viable seeds for dispersal.
Williamson (1996) suggested that the development of invasive species follows the “rule of
tens,” which predicts that roughly 10% of introduced species become established (naturalized),
and 10% of those established then spread and become invasive (meaning that, ultimately, approxi-
mately 1% of introduced species become invasive). A general lack of data makes it difficult to truly
quantify the proportion of introduced species that follow the rule of tens. In California, Florida,
and Tennessee, roughly 6, 10, and 13% of introduced plant species, respectively, have become inva-
sive (Simberloff 2013).
Invasive species are most often considered to be non-native, suggesting that they were moved
with human assistance to a new environment where they did not evolve. These non-native species
may also be called alien, exotic, or non-indigenous, although the first two terms have fallen out of
favor for their value-laden and somewhat judgmental use. Importantly, again, such species are not
always invasive. It is also noteworthy that even native species may behave like invasive species if
human disturbance allows them to colonize, establish, and spread into ecosystems normally inac-
cessible to them. For example, in the eastern United States, oak-dominated forests maintained by
relatively frequent surface fires prior to European colonization have been invaded by native red
maples and other mesophytic species due to fire suppression (Nowacki and Abrams 2008). With
fire suppression or exclusion in these ecosystems, oak forests developed into closed-canopy forests
that reduced the ability of oak to regenerate under the diminished light conditions and allowed
shade-tolerant species such as red maple to establish in the understory. As the abundance of
shade-tolerant species increased, the probability of fire decreased with the development of moist
and cool microclimates and the accumulation of less-flammable fuels, reducing oak dominance.
This process of “mesophication”—facilitated by the invasion of dry and dry-mesic sites by native,
shade-tolerant, mesic species—has threatened to convert many oak forests to those dominated by
maple and other mesic species (Nowacki and Abrams 2008).
A final point about invasive species is that their significance is heavily value-laden and varies
from location to location (Coates 2007). That is, a species invasive and perceived as an ecological
nuisance in one location is often heavily valued within its native range. There are numerous exam-
ples of tree species in this regard (Table 21.1). For example, northern red oak is a late-successional
canopy dominant native to eastern North America that is valued for its wood quality, whose persis-
tence is a major emphasis of forest management in the region. In forests of western and central
Europe, however, introduced northern red oak is an invasive species that establishes and spreads
in native woodlands, limiting regeneration of native tree species and severely reducing native
diversity. Likewise, Douglas-fir is an iconic, long-lived canopy dominant of the old-growth forests
of the Pacific Northwest of the United States and Canada whose protection from logging has been
a major environmental issue in North America over the last four decades. In the Southern Hemi-
sphere, however, particularly in New Zealand and in Patagonia, introduced Douglas-fir aggres-
sively colonizes grasslands and forest understories, altering forest composition and reducing
diversity. The perception of these two tree species valued in North America but abhorred else-
where does not differ from that of the tree-of-heaven, one of the most notoriously invasive trees in
Concepts of Invasive Species 589
Table 21.1 Selected list of trees invasive in North America and those native to North America but inva-
sive elsewhere.
Habit in e cosystems Habit in invaded

Species Origin Invaded
of origin ecosystems
Tree-of-heaven China Widespread in Widespread, Dense, clonal thickets
the United probably as an on forest edges even
States except early- on poor sites;
extreme north successional species displaces natives
Norway maple Europe, northern Northeastern Generally, a Forms closed-canopy
Turkey and Iran and northwestern scattered understory forests and casts
United States tree; rarely deep shade
a dominant
Russian olive Europe and Asia Mostly western Shrublands and Grows in riparian
United States woodlands in riparian areas and outcom-
and moist areas petes native
vegetation there
Common Europe and Asia Northern United Dense thickets to Invades forests,
buckthorn States and scattered plants on prairies, and
California open sites savannas and forms
dense thickets
Box elder North America Central Europe, Scattered short-lived Dominant in lowlands,
Australia tree in riparian areas riparian areas, and
and colonizes disturbed areas
uplands
Douglas-fir Pacific Northwest New Zealand, Long-lived canopy Invades open
and Intermountain Chile, dominant in pure and grasslands and forest
West of and Argentina mixed stands understories
North America
Black cherry Eastern Western and Early-successional Invades understories
North America central Europe and edge species of oak and pine
dominant; common forests, impedes
in understory regeneration of other
tree species
Northern red Eastern Western and Late-successional Establishes and
oak North America central Europe canopy dominant competes with native
oaks, impedes
regeneration, and
reduces diversity
North America but highly valued in China for its medicinal, ornamental, and commercial value as
a host plant for silk moths. Much like their invasiveness itself (see the following text), the percep-
tion of such species is highly context-dependent.
CHARACTERISTIC TRAITS OF INVASIVE PLANT SPECIES

What is it about invasive species that makes them particularly able to establish and spread beyond
their point of introduction? A classic explanation is that species become successful in new locations
in the absence of their coevolved natural enemies. Commonly called the enemy release hypothesis,
the idea is that some species become invasive when introduced to a new environment that does not
contain herbivores or pathogens that fully replace the natural enemies present in the native
environment. As a result, an increase in abundance and distribution of the invasive plant may
proceed unchecked in the new environment. Although criticized for its oversimplicity in explain-
ing why some plants become invasive (Colautti et al. 2004), there indeed exists evidence for the
enemy release hypothesis when the plant species is strongly regulated by pathogens in its native
environment. For example, Norway maple is a popular street tree introduced from Europe in
the eighteenth century that has invaded both urban woodlots and undisturbed forests of the
northeastern United States and Canada (Martin 1999). Tar spot disease is a pathogen that affects
Norway maple in Europe, and it was introduced to North America near the end of the nineteenth
century. Webster et al. (2005) attributed a decline in stem growth of overstory Norway maples in
Michigan to an outbreak of the disease. Moreover, studies near Montreal, Quebec have shown that
the growth of Norway maple saplings and trees declined sharply and mortality increased as a result
of an outbreak of tar spot disease, while the disease’s effects on native sugar maple remained far
less detrimental (Lapointe and Brisson 2011). Although the interactions between introduced
herbivores/pathogens and invasive plants may differ in locations beyond those of the ecosystems
in which they evolved, these “reunion” studies suggest that in at least some cases plants may
become invasive when freed from natural enemies.
Despite the potential importance of natural enemies, humans have preconceptions about
what allows invasive plants to establish and spread in new environments. One of the most common
traits attributed to invasive plants is an extremely high growth rate that allows them to grow faster
than native plants, thus outcompeting them or overgrowing them, often to the point of local
extinction. Perhaps the best example of a fast-growing invasive plant is kudzu, a semi-woody vine
native to Asia and introduced to the United States in the mid-1870s. Kudzu may grow as fast as
a foot per day once established, may develop vines as long as 100 feet, and is well known for over-
taking and overgrowing anything in its path, particularly in the southeastern United States
(Figure 21.1). The other major characteristic of invasive plants, including many invasive trees, is
their ability to form dense, monotypic (single-species) stands that shade out native plants. For
example, in eastern North America, Norway maple forms dense stands of overstory trees that cast
Photo by John D. Byrd, Mississippi State University, Bugwood.org
F I G U R E 2 1 .1 Aggressive invasion of kudzu overgrowing trees, shrubs, and virtually all other features on a
landscape in the southeastern United States.
deep shade, limiting or eliminating the growth of native herbaceous plants and the regeneration of
native trees. Norway maple regeneration is much more successful beneath the dense shade cast by
its own overstory than is regeneration of native trees, and its saplings also appear to outcompete
native saplings in the understory (Galbraith-Kent and Handel 2008).
The traits that make some plant species invasive are more complex than simply a fast growth
rate and an ability to exclude other species with shade, however. In general, invasive species are
superior to native species in some functional traits that allow them to opportunistically acquire
resources that help them to excel at growth, reproduction, and spread (Blumenthal 2005). Such
traits include high relative growth rates, but also short generation time; light, small, and widely
disseminated seeds; and fruits attractive to vertebrate dispersers (Rejmánek and Richardson 1996).
Notably, a wide variety of fast-spreading plant species fit this description, including many of those
classically considered to be pioneer or early successional species that are not invasive in their
native environments.
Traits most likely related to invasiveness are often determined from meta-analyses of many
studies that compare co-occurring native and naturalized species to those that are invasive, and
such studies often indicate that many presumptions about what makes a species invasive are not
necessarily correct. For example, Daehler (2003) examined 79 comparisons of co-occurring native
and invasive plants and found that invaders were more likely to have higher leaf area and lower
cost of construction for leaf tissue. Higher leaf area and lower tissue construction costs are
together likely to provide a growth advantage to invasive species. However, most other traits did
not differ between native and invasive plants (Figure 21.2); invasive plants did not necessarily
grow faster, were not better competitors, nor were they better dispersers. In a study of 146 com-
parisons of naturalized and invasive plants of the same genus in Australia, Gallagher et al. (2015)
found that invasive plant species had larger specific leaf area (the amount of leaf area per unit
biomass), longer flowering periods, and were taller than naturalized species. Specific leaf area
was especially important in distinguishing invasive from naturalized species and was associated
with faster growth rates, higher leaf turnover, and shorter life spans. Invaders were also more tol-
erant of a wider range of rainfall and temperature conditions in the native range. Finally, invasive
plants tend to show greater phenotypic plasticity than non-invasive plants (Davidson et al. 2011;
Daehler 2003). Theoretically, higher plasticity should allow an invading species to better adapt to
novel environmental conditions to which it is introduced, and to take advantage of environmental
fluctuations common to disturbed ecosystems where invasive species commonly establish
and spread.
In a comprehensive study of 117 trait comparisons between invasive and non-invasive plant
species, Van Kleunen et al. (2010) identified multiple trends attributable to species’ invasiveness.
Invasive species had higher values for traits such as photosynthetic rate, transpiration, tissue
nitrogen content, and lower leaf construction costs; allocation to leaf area and shoots, growth rate,
and size (biomass of roots, shoots, and whole plants, and plant height); reproduction (number of
flowers or seeds, germination); and survival. In other words, invasive species had higher values for
traits related to high performance and faster growth, and ultimately higher fitness. Importantly,
traits differed more between invasive and native plants than between invasive and naturalized
non-native plants, and the traits of invasive plants generally did not differ from native plants when
the native plants were invasive in other environments (Van Kleunen et al. 2010).
The traits associated specifically with invasive woody plants include a subset of those typi-
cally mentioned as being characteristic of invasive plants in general. In a study of 28 woody species
known to be invasive in California, Grotkopp and Rejmánek (2007) compared the traits of invasive
and naturalized (non-invasive) species within the same genus or family. They found that high
relative growth rate in the seedlings of trees and shrubs and high specific leaf area were the best
indicators of woody plant invasiveness, probably because these traits are related to a higher ability
to opportunistically capture available resources. High growth rate was also a major predictor of
r l r
tte ua te
et
be eq b
e e er
tiv tiv va
d
Na Na In n
growth rate 41
photosynthesis 21
construction 8
Growth-related
leaf area 16
roots 7
herbivore damage 9
grow season length 8
water use 5
salt tolerance 5
fecundity 13
Spread-related
germination 18
dispersal 7
survival 15
competition
15
Composite
standing biomass
9
plasticity
12
0 25 50 75 0 25 50 75 0 25 50 75 100
% of comparisons
F I G U R E 2 1 .2 Performance of native versus invasive plants using 16 growth-related, spread-related, and

composite traits. The first column shows the number of comparisons between co-occurring native and
invasive plants in which the native performed better; the center column shows equal performance between
native and invasive plants; and the third column shows better performance by the invasive plant in the
comparison. Differences between invasive and native plants were relatively few; invaders had higher leaf
area, lower construction costs, and higher phenotypic plasticity, and natives had higher survival.
Source: Daehler (2003) / Annual Reviews Inc.
invasiveness for trees in a literature review by Lamarque et al. (2011), who noted that several
mechanisms related to invasiveness were useful in explaining tree invasions, and that no single
factor was sufficient to predict whether an introduced tree would become invasive. This lack of a
“silver bullet” in explaining tree invasions is well in line with invasive plants in general.
The physical site factors of the ecosystems being invaded seem to also heavily influence the
incidence of plant invasions. For example, invasive species in regional ecosystems having a
Mediterranean climate (long, dry growing seasons and mild, wet winters) tend to allocate more
biomass to roots than non-invasive plants, allowing them to better survive the characteristic dry
summers (Grotkopp and Rejmánek 2007). These patterns suggest that invasive species must not
only be able to grow quickly and excel at acquiring resources, but also must be able to tolerate
certain growing conditions or environmental filters that determine where they may be effective
invaders. Such environmental filters appear to be important regardless of the traits possessed by
the invader. Apparently, the site factors of the ecosystems being invaded select for similar traits as
those found in native species, such that introduced plants that become naturalized typically have
similar traits as native plants (Divíšek et al. 2018; Kraft et al. 2015). In some cases, these similarities
occur because the introduced plants are closely related to the native plants (Strauss et al. 2006). To
proceed from naturalized to invasive species, however, plants must also have traits that are differ-
ent enough from those found in the community in which they have been introduced to give them
an advantage over native species (Ordonez 2014).
Adding to the complexity of invasiveness are the growing conditions of the new environ-
ment (Daehler 2003). An advantage of invasives over native species depends on the availability of
resources in a given ecosystem. In general, invasive plant performance increases under conditions
of high resource availability (light, water, or nutrients), but native plants outperform invasives
when resources are low or when natives are favored by a specific disturbance regime (Figure 21.3).
As a result, very few species are able to be invasive everywhere they are able to establish (Dae-
hler 2003). At the same time, these trends explain why ecosystems with high resource availability
(such as open, urban, or other ecosystems highly disturbed by humans) or altered disturbance
regimes (such as ecosystems affected by fire exclusion) are particularly susceptible to plant inva-
sions. Exceptions to this trend, whereby invasive plants perform well when resources are limited,
are not uncommon and are often related to the high resource-use efficiency of some invasive
species (Funk and Vitousek 2007; Heberling and Fridley 2013; Jo et al. 2015, 2017).
100
90 13
12
80 9
% of independent comparisons
10
70 12 6
60
8
50
40
30
20 79
10
0
r
ht
rs
n
ve
te
te
t
ur
io
en
ze
lig
wa
wa
ne
tit
t
is
ra
tri
pe
ld
w
h
nu
tg
lo
m
g
lo
ia
hi
ou
co
w
ec
lo
or
ith
sp
w
w
lo
Conditions when performance of native > invader
F I G U R E 21 .3 Growing conditions under which the performance of the native species was equal to or
better than that of the invasive species in a pairwise comparison. Numbers over each bar are the total
number of studies that manipulated each condition. Source: Daehler (2003) / Annual Reviews Inc.
NON-P LANT INVASIVE SPECIES IN FORESTS

Thus far we have described invasive species in forests as if they only included plants, but exotic
insects, pathogens, and other animals are also critical in affecting forest ecosystems and deserve
mention here. Phytophagous (plant-feeding) insects and pathogens are among the most widespread
and impactful agents to forests in North America. Like invasive plants, insects and pathogens are
often introduced unintentionally with their host plants during shipping of nursery stock, produce,
or related products, or simply when the species hitchhike on any cargo transported from their
native range to a new environment (McCullough et al. 2006; Westphal et al. 2008). Also, much like
invasive plants, only a fraction of new introductions become established, and only a fraction of
established populations begin to spread to become invasive.
More than 450 non-indigenous insect species and at least 16 pathogens, introduced between
1635 and 2006, are documented to attack one or more tree species in the United States (Aukema
et al. 2010). The majority of these introductions have occurred since 1860 (Figure 21.4a). About
14% of these species are high-impact insects and pathogens, in that they kill trees at a rate above
background levels; these high-impact species may cause tree mortality over broad areas. A selected
list of some of these high-impact insects and pathogens is presented in Table 21.2. Notably, of the
high-impact species, only two insects and one pathogen were introduced before 1860; between
1860 and 2006, a new damaging insect or pathogen was detected in the United States about every
two years. Accumulation of non-indigenous species in the United States has been dominated by
those that feed on sap (e.g., adelgids, aphids, and scales) or on foliage, although a significant
increase in the number of introductions that feed on phloem and wood has occurred since about
1970 (Figure 21.4b).
Much like invasive plants, many, if not most, invasive insects are less ecologically trouble-
some in their native range. For example, the emerald ash borer feeds on stressed and dying ash
trees in its native range in east Asia and is fairly uncommon at endemic levels. In North America,
however, the borer was able to colonize and kill healthy ash trees, reproduce, and spread quickly
across the continent (Herms and McCullough 2014), and has emerged as one of the most destructive
invasive insects ever introduced there. In fact, North American ash used as ornamental trees in
(a) (b)
150 Phloem and wood borers
Foliage feeders
Sap feeders
Cumulative pest detections
Cumulative pest detections
300 Other
100
200
50
100
0 0
1700 1800 1900 2000 1750 1800 1850 1900 1950 2000
Year Year
F I G U R E 2 1 .4 Cumulative detections of established forest pests over time for (a) total non-indigenous
insects with line fitted for the years 1860–2006 (blue) and high-impact insects and pathogens for the years
1860–2006 (red); and (b) cumulative detections of non-indigenous forest insects by guild over time.
Source: Aukema et al. (2010) / with permission of Oxford University Press.
Table 21.2 Selected list of important forest insects and diseases introduced into North America that have severely disrupted native ecosystems and decimated
or displaced native species.
Agent Latin name Origin Detection year Hosts Tissue attacked

Insects
Spongy (formerly gypsy) moth Lymantria dispar Europe, Asia 1869 Hardwoods Foliage
Winter moth Operophtera brumata Europe 1930 Hardwoods Foliage
European pine sawfly Neodiprion sertifer Europe 1925 Pines Foliage
Hemlock woolly adelgid Adelges tsugae Asia 1952 Hemlocks Foliage
Balsam woolly adelgid Adelges piceae Europe 1908 True firs Phloem
Asian longhorned beetle Anoplophora glabripennis Asia 1996 Hardwoods Cambium, phloem
Emerald ash borer Agrilus planipennis Asia 2002 Ashes Phloem
Beech scalea Cryptococcus fagisuga Europe 1890 Beech Bark, cambium
Pathogens
Chestnut blight Cryphonectria parasitica Asia 1876 Chestnuts Cambium, phloem
White pine blister rust Cronartium ribicola Europe, Asia 1900 White pines Needles, stems
Dutch elm disease Ophiostoma ulmi Europe 1927 Elms Xylem, phloem
Dutch elm diseaseb Ophiostoma novo-ulmi Europe 1945 Elms Xylem, phloem
Oak wilt Bretziella fagacearum Central or 1944 Oaks Xylem
South America
Laurel wilt Raffaelea lauricola Asia 2002 Redbay, Xylem
sassafras
a
Beech scale is associated with the spread of the pathogen that causes beech bark disease (Nectria coccinea var. faginata). Trees remain free of the disease until they are infested with
beech scale.
b
Dutch elm disease is caused by two distinct species of a pathogen successively introduced into North America.
Source: Data from Aukema et al. (2010).
0005430862.INDD 595 12-21-2022 15:08:55

Asia have allowed local emerald ash borer populations to approach outbreak levels (Liu et al. 2003).
Lack of resistance in North American ash relative to Asian ash species is a likely explanation for
the rapid increase in emerald ash borer population in a new environment and its consideration as
a highly destructive invasive species there (Liebhold et al. 2017).
The presence of invasive insects appears to vary widely across North America. There are far
more non-indigenous, high-impact forest insects and pathogens present in eastern North America
compared to western North America (Figure 21.5). There are likely multiple mechanisms driving
this spatial pattern. First, the mostly deciduous forests of eastern North America contain a higher
number of tree genera that could serve as hosts for newly established non-indigenous insects com-
pared to the mostly coniferous forests of the West (Liebhold et al. 2013). Of the 450 non-indigenous
insect species identified by Aukema et al. (2010), about half of the species feed on host species
within a single genus, about 18% attack more than one genus within a single plant family, and
about a third have a broader range of hosts that may span multiple plant families. As such, it makes
sense that most insects are more likely to invade and become established where genera are more
diverse, which might also explain why invasive insects from Europe are much more pervasive in
North America than North American insects are in Europe (Mattson et al. 2007). In addition, given
that international trade is a main vector of new introductions (Westphal et al. 2008), the longer
history of such trade in eastern North America compared to the West is also likely an important
factor driving the pattern of insect invasions across the continent.
Invasive tree pathogens in many ways have similar effects as invasive insects. Invasive path-
ogens may also cause widespread tree mortality. Classic examples include chestnut blight, which
was introduced from Asia into North America shortly after the Civil War and killed the majority of
American chestnut trees in eastern forests by the 1940s (Paillet 2002); and Dutch elm disease,
which was first introduced in the 1920s from Europe, killing most large elms across the eastern
United States in natural forests as well as those widely planted and prized as street trees
(Hubbes 1999). Important differences exist between insects and pathogens, however. First, inva-
sive pathogens are often extremely difficult to identify, such that pathogenic establishment and
spread and resulting tree mortality may proceed for many years unchecked. In addition, pathogens
Alaska
250 0 1,000 km
Number of pests per country

250 0 1,000 km
3 22 45
F I G U R E 2 1 .5 Geographical variation of the numbers of invasive, high-impact forest insects and pathogens
across the United States. There are far more invasive insects and pathogens in the eastern United States with
a clear area of concentration in the Northeast. Source: Liebhold et al. (2013) / John Wiley & Sons.
Impacts of Invasive Species on Forests 597
often exist in several strains or even species that may be introduced in succession and, as a result,
increase the severity of a particular epidemic. For example, Dutch elm disease was caused by two
distinct pathogen species that were introduced to North America in succession (Table 21.2), the
second two decades after the first, resulting in prolonged and more severe effects of the disease
(Brasier 2001).
Tree pathogens—including those invasive to a particular region—often work in tandem
with either invasive or native insects that serve as vectors for infection or as dispersal agents (Wing-
field et al. 2016). There are many examples of such mutualisms that have developed between inva-
sive species. Dutch elm disease results from a pathogen introduced with timber from Asia; this
timber was infested by both native and introduced bark beetles that facilitated the spread of the
pathogen to native North American elms. More recently, beech bark disease has begun to cause
extensive tree mortality of American beech across the eastern United States. The pathogen that
causes beech bark disease does not attack trees until the trees have been heavily infested by the
beech scale, an invasive insect introduced from Europe around 1890. The pathogen presumably
enters the tree through the puncture wounds caused by the extensive feeding by the beech scale.
The association between the beech bark disease pathogen and the beech scale is also present in the
native range of each agent in Europe. Although the pathogen in North America was once thought
to be the same as that causing beech mortality in Europe, recent evidence suggests that closely
related pathogens might cause the disease on both continents, and the pathogen causing disease in
North America is actually native (Wingfield et al. 2016).
Non-plant invasive species in forests are not limited to insects and pathogens, but also
include soil invertebrates and large vertebrates (Liebhold et al. 2017). Invasive soil invertebrates
include many that affect decomposition, such as some beetles and particularly earthworms.
Earthworms have been documented to have a large effect on microbial and invertebrate soil
communities and the cycling and storage of soil nutrients (Frelich et al. 2006). Introduced large
vertebrates, such as feral swine, also have large impacts on forests. Extensive rooting and digging
by feral swine have been documented to destroy tree regeneration and damage the litter layer in
the southeastern United States (Campbell and Long 2009). Impacts to forest ecosystems of these
and other invasive species are discussed in the following sections.
IMPACTS OF INVASIVE SPECIES ON FORESTS

Separate from understanding the invasiveness of a species—that is, the traits that allow an intro-
duced species to be effective at establishing and spreading from its point of introduction—is
understanding the impact of an invasive species, or its effect on native plants or ecosystems. The
impact of an invasive species varies depending on the type of taxa (e.g., plant, insect, pathogen, or
vertebrate) and ranges from changing abundance of an individual species to altering ecosystem
function (Parker et al. 1999). Impacts of an invasive species depend largely on its distribution and
abundance, which, in turn, depend on the stage of invasion (Mayfield et al. 2021). In this section,
we focus on the ecological impacts of the major taxonomic groups of invasive species in forests and
provide examples therein. A full treatment of invasive species impacts is given by Mayfield et al.
(2021), Miniat et al. (2021), and Liebhold et al. (2017).
IMPACTS OF INVASIVE PLANTS ON FORESTS

Invasive plants may impact forests in several different ways, including changing the abundance
and distribution of native species (for native plants, through competition; for other native taxa, by
altering habitat structure and resource availability), altering fire regimes, and affecting processes
that govern nutrient cycling and storage. Although plants are but one taxonomic group of invasive
species in forests, they have a disproportionate effect on ecosystems for their ability to shape
important factors such as vegetation structure, successional trajectories, primary productivity,

nutrient cycling, and species habitats.
Competition One of the most obvious ways invasive plants affect forests is by outcompeting
native plants for resources such as light, water, and nutrients. As we have seen, invaders tend to be
favored where such resources are abundant, but natives are favored where resources are limited
(Daehler 2003). Once plants invade a forest, they can drastically change its composition by compet-
ing with native tree regeneration and, in the case of invasive trees, by shading out (e.g., Norway
maple) or crowding out (e.g., tree-of-heaven) native woody plants. In the case of invasive woody
vines, such as kudzu or Asian bittersweet, the invasive plants may climb, overgrow, or smother
native trees and other plants (Figure 21.1). For example, infestation by Asian bittersweet sharply
reduced the growth of red oak and trembling aspen in a Massachusetts forest 14 years after
establishment (Delisle and Parshall 2018). Although local changes in species abundance and com-
positions are well documented to be associated with invasive plants, invasive plants rarely push
native plants fully to extinction (Gurevitch and Padilla 2004). By affecting native plants, however,
invasives may also affect native invertebrate and vertebrate species that utilize or specialize on the
native plants (Gandhi and Herms 2010).
Altered Fire Regimes In addition to altering the vegetation structure or the abundance of
species in forests, some invasive plants cause system-level effects that create new disturbance
regimes under which both native and invasive plants must persist (Asner and Vitousek 2005).
Perhaps the best examples of such system-level effects include invasive plants that alter the fire
regimes of forests—recognized as some of the most important ecosystem-altering species on Earth
(Brooks et al. 2004; Mayfield et al. 2021). The best-studied of such plants are grasses. When an
invasive grass establishes in a forest of a dry regional ecosystem, it provides the fine fuel needed to
initiate and carry a fire, resulting in increased fire frequency and intensity. Following such fires,
the invasive grasses recover more quickly than native grasses and plants, further increasing the
susceptibility of the forest to additional fires (D’Antonio and Vitousek 1992).
The most notorious example of the invasive grass–fire cycle is that of cheatgrass in the Inter-
mountain West. Cheatgrass was introduced from Europe or Asia prior to the Civil War and is now
widespread across the United States, though it is most prominent and problematic in the West.
Though its major effects are on rangelands and sagebrush steppe ecosystems, cheatgrass is com-
monly found in pinyon–juniper forests of the Great Basin, ponderosa pine forests across the West
(Figure 21.6), and interior Douglas-fir forests of British Columbia. It can outcompete native plant
species through soil water depletion (Melgoza et al. 1990) and may displace them as fire frequency
increases. Repeated and frequent burning due to cheatgrass invasion has converted millions of
acres of sagebrush steppe to annual grasslands (Monsen and Shaw 2000), and it has had similar
effects on the forest types described in the earlier text. Cheatgrass is very difficult to control once it
achieves dominance; those areas where perennial plants have been nearly completely displaced by
cheatgrass and fires occur every five or fewer years are likely to become permanently dominated by
cheatgrass over the long term (Mosely et al. 1999).
A similar impact of invasive grasses on forests is found in the southeastern United States,
where cogongrass was introduced from Japan in 1912. Cogongrass is a tall, dense, and aggressive
grass that outcompetes native plants for nutrients, crowds out native species, reduces light levels
at the forest floor, and alters decomposition rates in southern pine forests (Brewer 2008; Holly
et al. 2009). In a study of longleaf pine forests on the Gulf Coastal Plain in Mississippi, Brewer
(2008) found that cogongrass invasion over 3 years displaced shorter plants that were endemic to
longleaf pine forests and sharply reduced species richness. Tall saplings, shrubs, and vines, most of
which were not endemic, were not displaced. In addition to displacing native plants, cogongrass
also alters fire regimes by increasing fine fuel loads and fire intensity, resulting in higher tree
Left photo by Becky Kerns, USDA Forest Service; right photo by John
M. Randall, The Nature Conservancy, Bugwood.org.
F I G U R E 21 .6 Cheatgrass invasion into ponderosa pine forest in the Blue Mountains, Oregon (left) and
cogongrass invasion into slash pine forest in the southeastern United States (right).
mortality relative to historical fire regimes. For example, Platt and Gottschalk (2001) found higher
aboveground biomass of fine fuels and litter biomass in slash pine forests where cogongrass was
present than in nearby areas without it. Likewise, Lippincott (2000) found higher fuel continuity
in longleaf pine forests invaded by cogongrass that led to more intense fires and higher young tree
mortality, creating the potential that such forests could convert to non-indigenous grasslands.
Carbon and Nutrient Cycling Invasive plants may create system-level changes beyond
a ltering fire regimes. On the whole, invasive plants appear to increase ecosystem productivity
because they affect carbon uptake, storage, and cycling. For example, in a meta-analysis of
94 experimental studies, Liao et al. (2008) found that aboveground net primary production
increased by 50–120% in ecosystems with significant plant invasions—particularly by woody
plants—compared with uninvaded ecosystems. Such effects are greatest when invasive plants have
significantly higher light and nutrient use efficiencies and growth rates than the native plants they
are replacing. Effects on carbon cycling, in general, increase with the dominance of such plants.
Interestingly, woody plants tend to increase carbon stocks in aboveground and belowground bio-
mass and soil organic matter when they invade grasslands, but annual grasses tend to decrease soil
carbon stocks when they invade forests (Miniat et al. 2021). Given that invasive plants are gener-
ally more competitive when limiting resources are more abundant (Daehler 2003), the strength of
the effects of invasive plants on carbon cycling is likely to vary widely across ecosystems differing
in resource availability and is difficult to generalize.
The differences in growth rates and tissue chemistry of invasive plants often increase their
litter production and decomposition rates (Wardle and Peltzer 2017), respectively, and these
differences can facilitate changes in fire regimes. In turn, altered fire regimes, particularly when
fire frequency or intensity is sharply changed, will affect carbon and nutrient cycling in invaded
ecosystems. For example, cover-type conversion of forests to grasslands due to invasion by cheat-
grass or cogongrass, as described in the earlier sections, will quickly result in a loss of carbon
storage (see Chapter 20). Studies of cheatgrass invasion in the Great Basin of the western United
States have shown that an estimated 8 Tg of carbon has been released to the atmosphere due to
increased fire frequency and intensity and land cover change from forest to grassland (Bradley
et al. 2006). Continued invasion by cheatgrass will likely release another 50 Tg of carbon in the
coming decades, converting large portions of the landscape from carbon sinks to carbon sources.
Invasive plants also appear to have a strong effect on the cycling of nutrients in forests. Plant
invasions generally increase the amount of nitrogen found in aboveground plant tissues (Liao
et al. 2008), increase nitrogen availability in soils, and stimulate microbial communities. In com-
paring five pairs of native and non-native buckthorn, honeysuckles, and bittersweets in deciduous
forests of the eastern United States, Jo et al. (2017) found that invaders had greater litter produc-
tion and litter nitrogen content that enhanced the flow of nitrogen from litter to the soil after
3 years. Moreover, available soil nitrogen uptake was increased in invaded forests because invaders
had higher fine root production and longer roots. Increased rates of soil nitrogen mineralization
and uptake by invasive plants therefore accelerate soil nitrogen cycling, which, in turn, is likely to
support further increased productivity of invaders.
An additional, and perhaps most pronounced, mechanism of impacting nutrient cycling by
invaders is by nitrogen fixation. Many invasive plants can house bacterial symbionts that are able
to fix nitrogen and therefore provide the plant a significant competitive advantage on sites where
nutrients are limiting. For example, black locust is a nitrogen-fixing tree native to the Ozarks and
southern Appalachian Mountains, but widely introduced and naturalized across much of the
United States. Black locust is considered invasive in Europe as well as in the western United States,
the Midwest, New England, and northern California, even though the species is native to North
America. Rice et al. (2004) found that black locust invasion in pitch pine-scrub oak forests of
New York increased soil nitrogen concentrations and net nitrification rates relative to non-invaded
sites. Net mineralization rates also greatly increased relative to non-invaded sites and were attrib-
uted to leaf litter with high nitrogen content and high turnover. The longevity of invasive effects on
nutrient cycling remains an area of active research, and the magnitude of impacts is likely to be
highly dependent on the ecological characteristics of a given ecosystem (Kumschick et al. 2015).
IMPACTS OF INVASIVE INSECTS AND PATHOGENS ON FORESTS

Most invasive insects and pathogens have relatively little effect on forests, but a small fraction have
gained notoriety for killing or damaging millions of trees across very wide areas. The type of
damage caused by invasive insects depends largely on the type of tissue they feed upon. For
example, insects that feed on seeds and flowers reduce the ability of host trees to reproduce,
whereas those that feed on foliage, sap, or roots generally reduce the vigor of trees. The spongy
moth (formerly known as gypsy moth) is a defoliator of a wide variety of deciduous tree species in
eastern North America. Severe defoliation by larval spongy moths creates a deficit of carbohydrate
allocated to roots and shoots, thereby reducing nitrogen uptake and resulting in loss of growth and
dieback in the crown (Kosola et al. 2001). The hemlock woolly adelgid is an aphid-like insect that
feeds on the storage cells of hemlocks, mainly in eastern North America. Such feeding not only
reduces carbohydrates in the tree but may cause xylem to form abnormalities that affect its ability
to uptake water (Domec et al. 2013) and may slowly kill the tree. Widespread mortality of hem-
locks, considered to be foundational species whose loss is likely to cause significant ecological
shifts in habitat of both aquatic and terrestrial taxa, has occurred across much of the southern
Appalachians.
Extensive tree mortality is most often caused by invasive bark and wood-boring insects.
Emerald ash borer is a wood borer introduced in Michigan from China in the early 1990s that has
killed millions of ash trees in eastern North America and is considered the most destructive invasive
insect ever to be introduced in the United States (Aukema et al. 2011). Larval emerald ash borers
feed at the nexus of phloem and xylem and rapidly girdle and kill the tree. Generally, invasive insects
having the most severe and long-term impacts are those most likely to spread quickly, are specific to
one or a few host species, and kill hosts that are abundant, dominant, and ecologically unique
(Lovett et al. 2006). As such, emerald ash borer and hemlock woolly adelgid are likely to be more
impactful than spongy moth, as are pathogens such as chestnut blight and Dutch elm disease.
By killing trees, invasive insects or pathogens create important changes in vegetation struc-
ture or species composition of a forest that may affect the rest of an ecosystem. When invasive
insects or pathogens kill trees, they reduce canopy coverage or create canopy gaps that increase
light availability and alter temperature and moisture regimes at the forest floor. In the short term,
such disturbances are likely to affect tree regeneration, density, and basal area as well as the estab-
lishment and growth of both native and invasive plants. In killing or damaging trees, most invasive
insects and pathogens will also increase organic inputs to the forest floor as leaves, branches, insect
excrement and biomass, and coarse woody debris (Figure 21.7; Gandhi and Herms 2010). Over
longer time periods, tree mortality caused by invasive insects and pathogens has the potential to
alter tree species composition and forest structure, especially if a large proportion of the dominant
overstory trees are killed by host-specific invasives (Figure 21.7), and thus sharply influence
succession (e.g., Morin and Liebhold 2016). When a host species largely defines forest structure
and/or stabilizes the functioning of an ecosystem as a foundational species (Chapter 14), removal
of such a host may have dramatic effects (Ellison et al. 2005; Wilson et al. 2018).
Chestnut Blight, Dutch Elm Disease, and Forest Succession A host species that is
eliminated or greatly reduced in abundance in forests is eventually replaced by other tree species.
An outstanding example is the virtual elimination of American chestnut from eastern deciduous
forests by chestnut blight. This invasive fungus was introduced from Asia about 1904, killing most
of the mature chestnuts in New England within 20 years, and most in the southern Appalachians
by the 1940s. It is unlikely in recorded history that a major forest tree has been so nearly eradicated.
American chestnut was a major dominant of many dry and dry-mesic ecosystems in the southern
Appalachian Mountains and in deciduous forests of eastern North America (Braun 1950), growing
to enormous sizes in old-growth forests of the southern Appalachians (Figure 21.8). Surviving
chestnuts infected with the blight typically produce non-viable fruit. Although we still observe
sprouts from root systems (Griffin 1992), few trees are vigorous or survive long enough to reach the
overstory.
Succession following the elimination of American chestnut often resulted in the simple
replacement of that species by its former associates (Illick 1914, 1921; Korstian and Stickel 1927;
Augenbaugh 1935; Keever 1953; Woods 1953; Woods and Shanks 1959; Day and Monk 1974;
Stephenson 1974). Succeeding trees in the southern Appalachians are mainly oaks (especially
chestnut, red, and white oaks), along with various hickories, and tuliptree on more mesic sites.
Thus, eradication of chestnut in this region has caused the replacement of the former oak–chestnut
type with an oak or oak–hickory type. Similar changes, but involving some different species, previ-
ously occurred in the middle Atlantic states and southern New England (Good 1968).
In the Allegheny Mountains of western Pennsylvania, logging and fire following the death of
chestnut trees created open sites for invasion by early and mid-successional species (black cherry,
black birch, black oak, sassafras, and black gum) together with species already present on the site
such as sugar maple, red maple, white ash, and beech (Mackey and Sivec 1973). Without fire or
other disturbances in the former chestnut forest, succession gradually proceeds to more tolerant
and mesophytic species such as hemlock, sugar maple, and beech.

Photo by Kayla Perry, Kent State University
F I G U R E 2 1 .7 Impacts of emerald ash borer on forests in southeastern Michigan. Top: Extensive mortality
of overstory ash trees. Note host-specific mortality in this mixed deciduous forest. Bottom: Dead overstory
ash trees left as coarse woody debris on the forest floor following emerald ash borer infestation. Note the
amount of sunlight now reaching the forest floor through canopy gaps created by killed ash trees.
Later in the twentieth century, Dutch elm disease, caused by a fungus first introduced into
Ohio from Europe around 1927, virtually eliminated mature American elm trees from swamp and
river floodplain forests of the eastern and midwestern United States within about 50 years. Studies
of succession following the elimination of overstory elms in the Midwest indicate that elms are
F I G U R E 21 .8 American chestnut trees in an old-growth forest of the Great Smoky Mountains, North
Carolina. On favorable sites, American chestnut trees could grow over 2 m in diameter and 36 m in height.
Source: Photo courtesy of the American History Society.
replaced by a number of different species, depending on regional and local site conditions. In Illi-
nois woodlands, sugar maple is the species most likely to increase in dominance where soils are not
too poorly drained (Boggess 1964; Boggess and Bailey 1964; Boggess and Geis 1966). In south-
eastern Iowa, hackberry and box elder were the most frequent trees replacing elm (McBride 1973).
In southeastern Michigan, American elm was a late-successional dominant in deciduous
swamp forests with black ash, red maple, and yellow birch. The latter two species now dominate
the swamps in different proportions depending on site conditions (Barnes 1976). In contrast with
chestnut in dry-mesic forests, American elm today remains in deciduous swamps, making up
about 10–15% of the understory. However, old fields and other open upland areas are now much
more important sites for elm regeneration than swamps. Unlike American chestnut, American
elm typically reaches reproductive maturity before it succumbs to the invasive pathogen, such that
it will likely be perpetuated for generations by seeds from young elm trees but the average life span
will be drastically reduced. Resistant elm clones continue to hold promise for horticultural plant-
ing (Townsend et al. 1995).
The death of host trees due to invasive insects and pathogens has obvious effects on carbon
and nutrient cycling in the forests that they affect. In general, invasive insects and pathogens
reduce the productivity of forests, at least in the short term (Miniat et al. 2021), and the input of
organic material to the forest floor and to dead wood carbon pools affects decomposition, soil
respiration, and annual carbon budgets. Invasive species that extensively damage but do not
necessarily kill their host species, such as spongy moth, are still likely to reduce tree growth and
thus net ecosystem production (Clark et al. 2010). Invasive insects or pathogens that kill trees may
also have indirect effects on carbon cycling when tree mortality results in replacement of the host
species by new species with different rates of nutrient uptake, growth, or litter quality. Such
replacements in some cases may have long-term implications for carbon cycling in affected forests
(Lovett et al. 2006), depending on the specific successional dynamics. For example, the loss of ash
in the United States due to emerald ash borer is likely to result in a significant loss of aboveground
carbon in the short term, but much of this carbon is likely to be recovered by the increased growth
of other tree species already present that will succeed ash (Flower et al. 2013).
Invasive insects and pathogens can also cause changes in nutrient cycling, often dramati-
cally (Crowley et al. 2016). Again, the pulse of organic matter to the forest floor following invasive
insect outbreaks or pathogen infestations, as well as the decrease in nitrogen uptake by trees
immediately following outbreaks or infestations, can result in short-term increases in soil nitrogen
availability (e.g., Orwig et al. 2008). Moreover, leaching and loss of nitrogen from soil may also
result. For example, Eshleman et al. (1998) found that nitrogen “leakage” from forested water-
sheds in the Chesapeake Bay region occurred several months to a year after spongy moth defolia-
tion, probably due to reduced nitrogen uptake from recovering trees and increased water discharge
from soil resulting from reduced transpiration following defoliation. Nutrient cycling may also be
affected over the long term when invasive-caused tree mortality results in tree species replacement.
For example, hemlock woolly adelgid infestations have shifted dominance from eastern hemlock
toward black birch in New England, increasing aboveground productivity and nitrogen uptake
(Finzi et al. 2014). Likewise, beech bark disease in beech–sugar maple forests of the eastern United
States has resulted in a shift toward sugar maple-dominated forests, in turn increasing litter decom-
position, nitrogen cycling, and nitrogen leaching from soils in the region (Lovett et al. 2010).
IMPACTS OF INVASIVE ANIMALS ON FORESTS

A large number of animals other than insects have been introduced and become invasive in forests
across the world. In the United States, over 260 vertebrates, including mammals, birds, amphib-
ians, and reptiles, could be considered invasive (Fall et al. 2011). The effects of many of these
species are local or regional in scale, such as those of goats or burros whose selective grazing
impacts native vegetation; or are difficult to isolate, such as various bird species that aid in the dis-
persal of invasive plants. In this section, we describe the notorious impacts of one invertebrate
(earthworms) and one vertebrate (feral swine) on forests in North America.
In northern regions of the United States that were glaciated during the last ice age, non-native
earthworms have invaded many forests that had been free of such fauna for more than 12 000 years.
North American earthworms found in unglaciated areas in the southeastern United States and the
Pacific Northwest have not expanded their range into this region, either because of slow dispersal
or because they do not tolerate relatively harsh winters. Where forests evolved without earth-
worms, litter decomposes more slowly on the forest floor, depending on site conditions. When
earthworms are introduced to these areas, typically from non-native plants and soils and discarded
fishing bait, the invasive worms voraciously eat the litter layer and may eliminate it completely. In
doing so, earthworms may redistribute carbon throughout the soil profile, dramatically reducing
carbon stocks on the forest floor (Bohlen et al. 2004). This redistribution may alter the distribution
and composition of microbial and microarthropod communities in the soil, as well as the seedbed
A Primer of Invasive Species Management in Forests 605
conditions for forest herbs and woody plant regeneration. In some forests, invasive earthworms
also contribute to soil erosion, nutrient leaching, and acceleration of nutrient cycling to the extent
that nutrient uptake by native plants cannot keep pace with available nutrients (Frelich et al. 2006).
These impacts are likely to alter the abundance and diversity of tree regeneration and could create
large, long-term changes in forest composition (Bohlen et al. 2004).
The presence of earthworms in agricultural settings was once universally considered to be
beneficial (for example, see the third edition of this textbook), but the potential negative impacts
of invasive earthworms on forests have only recently been recognized by ecologists. The leading
edge of earthworm invasion was examined in a series of studies of four forests dominated by sugar
maple in north-central Minnesota. Hale et al. (2006) found that high earthworm biomass was asso-
ciated with lower herbaceous plant abundance and diversity as well as lower abundance and
density of tree seedlings. High earthworm biomass was also associated with rapid disappearance
of organic horizons, and increases in the thickness, density, and total soil organic matter content
of the A horizon. Moreover, availability of nitrogen and phosphorus was lower where earthworm
biomass was high (Hale et al. 2005a). Hale et al. (2005b) found a clear succession of invasive earth-
worm species, where species that live and feed exclusively in the litter layer may facilitate the
establishment of species of earthworms that live and/or feed in the mineral soil, the latter of which
prevent recovery of the forest floor. Thus, not only do earthworms invade forest ecosystems, but
they also develop entire communities with species that compete and succeed one another as if they
were native.
The first swine in the United States was introduced to North America from the domestic
stock of early European explorers and colonists. Eurasian wild boars were later introduced for
hunting purposes, and they interbred with domestic pigs where their occurrences overlapped.
Today, feral swine cause extensive damage to forests and other ecosystems due to their digging and
rooting activities, which has been shown to influence plant succession and species composition
(Engeman et al. 2007) and decrease tree diversity and regeneration (Siemann et al. 2009). In
altering species composition, feral swine indirectly impact the species by disrupting the ecosys-
tems they inhabit. Feral swine also have direct effects on other animal species, in that they prey on
many native invertebrates, birds, amphibians, and reptiles (Jolley et al. 2010) and compete with
native fauna for food resources. In addition, feral swine consume the acorns of oaks and the nuts
of hickories, representing an important source of competition with native wildlife for a critical
food source as well as important seed predation for native tree species (Elston and Hewitt 2010).
The degree of soil disturbance provided by feral swine makes it likely that they substantially
affect forest ecosystem nutrient cycles, although empirical data are not definitive in this regard
(Barrios-Garcia and Ballari 2012).
A PRIMER OF INVASIVE SPECIES MANAGEMENT IN FORESTS

Prevention, management, and control of invasive species are within the enormous and complex
fields of applied and restoration ecology with textbooks dedicated to those subjects alone. There
are literally hundreds of strategies available to control and manage invasive species, and a detailed
treatment of such strategies is well beyond our scope. However, a series of common concepts,
caveats, and methods characterize the management and control of all invasive species regardless
of origin, taxonomy, or ecosystem, and we attempt in this section to provide a brief summary in the
context of forest ecosystems. Recent exhaustive and comprehensive reviews of invasive species
management are provided by Venette et al. (2021), Poland et al. (2021), Koch et al. (2021), Green
and Grosholz (2021), Simberloff (2013), and Clout and Williams (2009), among many others.
A discussion of invasive species management first requires an understanding of the stage of
the invasive process being addressed for a particular species. Regardless of the invasive species in
question, management needs and approaches vary widely at different points along this invasion
Proactive Active Reactive

Eradication
would require
intense effort
Local control and

management only
Eradication
simple or
feasible
Area infested
Control costs
Public awareness
typically begins
Introduction
Detection
Species Scattered Numerous At or near

absent locations locations biological potential
Time
F I G U R E 2 1 .9 Population growth of a hypothetical invasive species and the timeline of associated

management activities to control it. Eradication is most likely at the first sign of detection, which would
require proactive management. Once the species has invaded numerous locations, eradication is more
difficult, and reactive management when the invasive species approaches its biological potential is usually
successful only locally. Source: Adapted from Hobbs and Humphries (1995).
process. The timeline for invasion by a hypothetical invasive species is illustrated in Figure 21.9.
If we assume that the area infested by an invasive species can be represented by a logistic growth
curve, with low initial area infested followed by rapid infestation of a landscape and finally a long-
term stabilization of the area infested, then management effort (and thus activities) clearly must
differ over time. Eradication of an invasive species is fairly simple or feasible, as well as inexpen-
sive, when the species is present only in scattered locations across limited areas. Management at
this stage would be considered proactive and would depend heavily on predicting which species
might become invasive, preventing their further establishment and spread, and early detection if
they do begin to spread (Epanchin-Niell and Liebhold 2015; Rout et al. 2014).
Unfortunately, most invasive species are not detected on a landscape until they have already
begun to increase their populations, typically just before they experience the fastest increase in
the amount of area infested and are detected at many locations (represented by the steep portion
of the logistic growth curve; Figure 21.9). Because invasives exhibit rapid increases during this
period, management is necessarily active and typically requires intense effort, which is accompa-
nied by rapidly increasing cost of control. Even more unfortunate is that public awareness of the
issues surrounding an invasive species often is not substantial until the invasive species is already
rapidly increasing. When invasives infest all or most of what they are able on a landscape, their
increasing distribution slows. Management at this point is reactive and is too late for eradication,
resorting to control of the invasive in small, local areas at high cost. Among the important lessons
illustrated in Figure 21.9 are that (i) early intervention is crucial in invasive species management,
and (ii) a suite of intensive management strategies is necessary for active management of
invasive species.
EARLY INTERVENTION STRATEGIES

Early intervention includes prediction, prevention, early detection, eradication, and other rapid
responses (Venette et al. 2021). Prediction occurs by forecasting the likelihood that a species will
become invasive and its potential consequences, a process called risk assessment. Prediction also
depends heavily on pathway analysis, which is a standard process used to consider the potential
avenues for a species to enter a given ecosystem. Prevention typically involves regulations and
quarantines, which, together with pathway analysis, have been considered extremely valuable and
inexpensive in their ability to avoid new invasions (Hulme 2009). Early detection depends heavily
on surveillance, inventory, and monitoring of potential invasives in a new area of concern; specific
methodologies are described in Venette et al. (2021) and Oswalt et al. (2021). Detection of a new
invasive species or the spread of an invasive species to a new area then requires rapid responses
such as eradication, containment, or suppression (Beric and MacIsaac 2015). Management at the
early intervention stage may occur for recently introduced species that are not yet spreading,
species present at low levels but not yet considered to be threatening, or invasive species threat-
ening an area of interest (Venette et al. 2021).
MANAGEMENT APPROACHES FOR ESTABLISHED INVASIVE SPECIES

The priorities and strategies of invasive species management change quickly once an invasive
species becomes established and has spread. These strategies turn from goals of eradication and
containment to management strategies ranging in spatial scale from local approaches to land-
scape-or region-wide programs. Important principles of effective invasive species management at
this stage, as reviewed by Poland et al. (2021), include the following: (i) it is necessary to under-
stand the potential impacts of an invasive species as well as the impacts of its management;
(ii) objectives should be specific to a given ecosystem or landscape; (iii) knowledge of thresholds is
necessary for decision-making; and (iv) resources are inherently limited for managing invasive
species and must be prioritized. Prioritization of resources is generally based on the degree of neg-
ative impact by the invasive species, the value or rarity of the ecosystems at risk for invasion, the
size of the area potentially invaded and the stage of invasion, expected outcomes of management,
and the availability of management tools.
Management approaches for invasive species have been categorized by Poland et al. (2021)
to include the following specific to forests:
Regulatory control consists of measures to monitor or eliminate pathways that might allow intro-
ductions of invasives into a given ecosystem, such as seed certification programs, quarantines,
and best management practices. Regulatory control is most common for recently introduced
invasive species rather than long-established, widespread species. It is typically used for invasive
insects, pathogens, and terrestrial wildlife, but not for plants.
Physical control involves the physical removal of invasive species or the construction of physical
barriers to their invasion or spread. Examples include the pulling of invasive plants, trapping and
removing invasive insects and other animals, or fencing to exclude large invasive animals.
Physical control is expensive and generally limited to local areas. It is used in the control of all
invasive taxa.
Cultural control includes activities that humans conduct in culturing or managing resources that
make the establishment or spread of invasive species less likely. Avoiding the use of a host species
susceptible to an invasive species in urban areas or using silvicultural techniques that increase
health and vigor such as thinning or increasing diversity in forest stands are examples of cultural
control. Cultural control is especially used for the control of invasive insects and pathogens.
Chemical control employs natural or human-made chemicals or microbial agents to directly affect
invasive species in an effort to prevent infestation, eradicate populations, reduce impacts, or slow
down spread by reducing the population. Chemical control is very commonly used for invasive
insects and pathogens but is generally used for all invasive taxa.
Biological control uses living organisms to reduce the numbers of targeted invasive species such
that the invasive species can be better managed with other strategies. The most common biological
control agents are insects, usually predators or parasitoids, but may also include fungi, bacteria,
and viruses. Biological control is generally highly successful but requires a long period of
development to ensure its effectiveness against the invasive species but its harmlessness against
non-target species. Biological control has been most successful with invasive insects but also
holds much promise for the control of invasive plants.
Host resistance involves the propagation of host species populations that have genetic traits that
leave them protected against (or less susceptible to) attack by invasive species. Host resistance is
largely used as a strategy to reduce the impacts of invasive insects and pathogens, particularly
those long established, on native plants. Identifying trees resistant to invasive pathogens is gen-
erally most successful in genetically diverse populations.
Reproductive control refers to the prevention of mating of invasive species or the development of
their offspring using genetic or chemical manipulation. Classic examples include using phero-
mones to disrupt the mating of invasive insects or controlling fertility in other invasive animals.
Reproductive fertility is effective but is currently restricted to a relatively small number of inva-
sive species.
As suggested in the earlier text, a given invasive species will demand its own set of
management approaches; a list of the management approaches used for the invasive species
discussed in this chapter is presented in Table 21.3. The simultaneous application of several con-
trol methods in a program emphasizing sound ecology and economics is called integrated pest
management (IPM). IPM programs have been developed for several of the invasive species
discussed in this chapter. For example, such a program has been developed to control the spread
of spongy moth, integrating all seven of the management approaches discussed in the earlier
text (Sharov et al. 2002). Other examples of IPM programs include those for hemlock woolly
adelgid and emerald ash borer, as well as several invasive plants and tree pathogens (Poland
et al. 2021).
NOVEL ECOSYSTEMS AND INVASIVE SPECIES

In many ecosystems, humans have altered species composition, disturbance regimes, and ecosys-
tems processes well beyond their natural range of variability (Chapter 22), and together with cli-
mate change, have created an entirely new ecological context in which ecosystems occur. In such
cases, non-indigenous, introduced species may become better adapted to the new conditions
compared to native species, resulting in new combinations of species. Moreover, the new ecolog-
ical situation may be extremely difficult to reverse despite intensive management efforts. Such
ecosystems are termed novel ecosystems (Hobbs et al. 2006), and may include many introduced
species that represent the “new normal” in that they develop as unique, long-term stable states
that feature the dominance or abundance of one or more invasive species. Because they are
unlikely to be restored to a condition resembling that prior to human alteration, management of
novel ecosystems may shift toward those aspects that best meet the needs of humans—such as for
pollination, recreation, or carbon uptake—rather than toward the most natural condition possible
(Hobbs 2007).
Table 21.3 Management approaches employed for the invasive species discussed in this chapter.
Control method for established species

Species
Regulatory Physical Cultural Chemical Biological Resistance Reproductive
Plants
Tree-of-heaven X X
Norway maple X X
Common buckthorn X X
Amur honeysuckle X X
Kudzu X X X
Asian bittersweet X X
Cheatgrass X X
Cogongrass X X
Insects
Emerald ash borer X X X X X X
Beech scale X X X
Spongy (gypsy) moth X X X X X X X
Hemlock woolly adelgid X X X X X X
Pathogens
Dutch elm disease X X X X
Chestnut blight X X
Animals
Earthworms X X X X
Feral swine X X X
Source: Data from Poland et al. (2021).
Several important examples of invasive species already described in this chapter fit the defi-
nition of novel ecosystems outlined in the earlier text. For example, ponderosa pine forests heavily
invaded by cheatgrass in the western United States are systems that include new combinations of
species that have altered fire regimes. The altered fire regimes are likely to perpetuate the plant
communities that now include invasive plants at the expense of native species, including pon-
derosa pine itself. Initiated by human activities, the new ecological conditions are likely to be irre-
versible, and more frequent and intense fires that accompany cheatgrass invasion are likely to
maintain a new stable state of cheatgrass dominance. As a result, management is likely to empha-
size land use options viable under the new conditions rather than maintaining some historical
benchmark of ecological integrity; in other words, future management may emphasize the novel
conditions rather than historical conditions.
Related to the concept of novel ecosystems is the opinion that the impact, rather than
whether a species is native or non-indigenous, should shape how humans value species. For
example, Davis et al. (2011) present the example of tamarisk, a shrub genus from Europe, Asia, and
Africa, that has invaded riparian areas of the dry American Southwest. Tamarisk has a reputation
for high levels of evapotranspiration and thus its ability to alter hydrologic regimes in ecosystems
where water is limiting, and has been targeted for removal or suppression since the 1930s. How-
ever, tamarisks have recently been documented to have similar rates of water use compared to
native shrubs of the region (Stromberg et al. 2009), and today provide the preferred habitat for at
least one endangered bird. Thus, despite it being a heavily human-modified riparian system of the
arid Southwest, this novel ecosystem, together with its dominant, non-indigenous species, may
actually be a desirable element on this landscape that performs valuable ecosystem functions
(Aukema et al. 2010).
The concept of novel ecosystems as it pertains to invasive species has not been without con-
troversy. Several ecologists have argued, often confrontationally, that the concept of novel ecosys-
tems is seriously flawed because it enables conservationists and managers to ignore the enormous
ecological effects of invasive species that we have discussed in this chapter (Simberloff et al. 2015).
Similarly, they argue that equating the ecological value of invasive species to that of native species
ignores the impacts of invasives (Simberloff 2015). Moreover, the central concept of novel ecosys-
tems relating to non-indigenous species—that it might not be possible to remove all invasive or
non-native species from an ecosystem—is argued by some ecologists to be a dangerous avenue for
managers to forgo attempts at invasive species eradication and management when such attempts
may be completely feasible (Aronson et al. 2014).
As with all controversies, the most useful answer is probably the one lying squarely in the
middle. Most forest ecologists would probably rather see a near-natural forest ecosystem domi-
nated by native species than a highly altered forest dominated by non-native and invasive trees and
shrubs. At the same time, many ecologists would rather see a highly altered forest ecosystem on a
landscape than to see no wooded areas at all. Perhaps the value of the novel ecosystem concept is
that it labels non-natural ecosystems as something other than “degraded,” thus providing managers
and ecologists a new framework in which to define new objectives or paradigms for a seemingly
different, “novel” system that is not likely to be restored to its original composition, structure, or
function (Miller and Bestelmeyer 2016). If there is ecological value in such forest systems—and
there may not always be—then it is probably beneficial to emphasize that value rather than
abandon any concern because of its dissimilarity to pre-European colonization conditions.
SUGGESTED
READINGS
Aukema, J.E., McCullough, D.G., Von Holle, B. et al. Liebhold, A.M., Brockerhoff, E.G., Kalisz, S. et al. (2017).
(2010). Historical accumulation of nonindige- Biological invasions in forest ecosystems. Biol. Inva-
nous forest pests in the continental United States. sions 1911: 3437–3458.
Bioscience 60: 886–897. Poland, T.M., Patel-Weynand, T., Finch, D.M. et al. (2021).
Daehler, C.C. (2003). Performance comparisons of Invasive Species in Forests and Rangelands of the Unit-
co-occurring native and alien invasive plants: ed States: A Comprehensive Science Synthesis for the
implications for conservation and restoration. Ann. United States Forest Sector. New York: Springer 500 pp.
Rev. Ecol. Evol. Syst. 34: 183–211. Simberloff, D. (2013). Invasive Species: What Every-
Lamarque, L.J., Delzon, S., and Lortie, C.J. (2011). one Needs to Know. Oxford, UK: Oxford University
Tree invasions: a comparative test of the dominant Press 352 pp.
hypotheses and functional traits. Biol. Invasions 13: Williamson, M. (1996). Biological Invasions. New York:
1969–1989. Springer 256 pp.
Forest Landscape
CHAPTER 22
Ecology
L andscape ecology is a subdiscipline of ecology that, above all else, emphasizes spatial
heterogeneity (variation from place to place) and patterns in ecological systems. In most (not
all) cases, the heterogeneity of concern is that of vegetation, in part because it is most easily observed.
For a long time, much of the field of ecology (including much of forest ecology) ignored heteroge-
neity and assumed, for the sake of simplicity, everything to be non-spatial. However, we casually see
spatial heterogeneity in the natural world every day, by observing that one location is a better place
to see wildflowers than another, or perhaps that some tree species are likely to be more productive on
one site compared to another. Ecology is concerned with interactions, and it is clear that the strength
of interactions changes with distance (i.e., objects closer to each other interact more strongly than
those further apart). This is because distance implies spatial location, and therefore the patterns we
see across landscapes and within ecosystems are likely to be important for ecological processes.
The discipline of landscape ecology is fairly new, having been developed only in the early
1980s, though its broad conception arose much earlier from a combination of geography and plant
ecology, especially through the work of the German geographer Carl Troll (1939). Landscape ecology
is well known for having two well-defined “schools” depending on where the discipline is centered.
In Europe, the field is strongly human-centered (Wu and Hobbs 2007; Brandt et al. 2009) and deals
mainly with the “built” environments of cities and densely populated areas of the region. There,
landscape ecology focuses on the interrelationship between humans and their open and built-up land-
scapes (Naveh and Lieberman 1994). Together with biology and ecology, it emphasizes human-centered
fields, such as the social, economic, geographic, and cultural sciences connected with modern land
uses, and is used as the scientific basis for landscape planning, management, conservation, and res-
toration. In North America, the concepts of the European school of landscape ecology are often
emphasized in the fields of landscape architecture, urban planning, and landscape design.
In North America, the framework for a new discipline in landscape ecology was defined by
a small group of ecologists and geographers in 1983. Specifically, the field was defined as the study
of relationships between spatial patterns and ecological processes unrestricted by a particular scale
(Risser et al. 1984). This definition deviated from that of the European school, probably because of
the relatively small number of participants and their particular interests and expertise (With 2019).
Its emphasis on spatial patterns and ecological processes centered its stronger focus on natural sys-
tems, quantitative analyses using statistics and modeling, and the development of ecological the-
ory. In contrast to a focus in Europe on cultural landscapes, the North American school was also
more able to stress management of its still-remaining natural resources and large tracts of unbuilt
land. Forest and other natural resource management has played a major role in how landscape
ecology has developed as a field in North America over the last four decades (With 2019). As such,
landscape ecology is much more commonly taught and studied in biology, ecology, and natural
resource departments in North American colleges and universities.
Landscape ecology is commonly considered to be the study of features called patches or
elements situated in a landscape, where a “landscape” occurs at spatial scales broader than
611
612 Chapter 22 Forest Landscape Ecology
individual organisms or ecosystem types. It considers the spatial pattern and heterogeneity of
these features, their development and dynamics, exchanges between them, and their management
(Turner and Gardner 1991). The focus on broad spatial scales is an artifact of the human-centered
focus of the European school, but landscape ecology does not need to concentrate only on large
spatial extents. Indeed, technological advances such as satellite and other remotely sensed imag-
ery, the maturity of geographical information systems, and rapid improvements in the speed and
capacity of computer processing have allowed ecologists to more easily make observations and
collect data over large areas of land (Turner and Gardner 1991). However, landscape ecology does
not dictate specific scales that should be studied, instead emphasizing the identification of the
scales most appropriate for characterizing the relationships of spatial patterns and the ecological
process of interest (Turner and Gardner 2015). For example, With et al. (1999) examined the
response of crickets to habitat abundance and fragmentation using varying spatial patterns mowed
into 25 m2 plots in a field, because this was the appropriate “landscape” perceived by the crickets.
Landscape ecology encompasses an enormous diversity of possible topics that cannot pos-
sibly be covered well in a single chapter. Therefore, we emphasize some of the most important
topics with particular importance to forests, including the ecological importance of landscape
fragmentation, disturbances on landscapes, and the effects of landscape pattern on ecological
processes. A full treatment of landscape ecology is provided in the definitive texts by Turner and
Gardner (2015) and With (2019), as well as works by Wu and Hobbs (2007), Wiens and Moss
(2005), Forman (1995), Naveh and Lieberman (1994), and Forman and Godron (1986).
CONCEPTS OF LANDSCAPE ECOLOGY

A landscape is something with which most of us have some level of familiarity. As described by
Turner and Gardner (2015), we intuitively think of a landscape as a large area that we can observe
from a single place at a single point in time. It could be a forested landscape, or a prairie landscape,
or an agricultural or urban landscape, all of which contain patches of different types. For example,
a forested landscape might include lakes and ponds as well as forests, and the forested areas might
include patches of different forest types (e.g., oak forests and beech–sugar maple forests) or forests
of different ages. An urban landscape might include patches of industrial areas, residential areas,
and green areas such as parks or golf courses. As described in the earlier text, the concept of a
landscape could be broadened to include something more than the human perspective, to include
patches of things important to a particular organism of interest. Thus, a cricket’s landscape might
include patches of water, thick vegetation, and open areas, all in a small area as perceived by hu-
mans. Therefore, important is the presence of a mosaic of different patch types that may interact
with each other—the presence of heterogeneity. Although many definitions exist for a landscape
(Wiens and Moss 2005), simply defining it as an area exhibiting spatial heterogeneity in at least one
factor of interest (Turner and Gardner 2015) seems most appropriate.
All landscapes may be broadly described with a few simple attributes. Landscape structure
describes both the composition (different kinds or diversity of patches) and the configuration
(spatial arrangement of the patches) of the landscape. Composition might include metrics such as
the proportion of the landscape represented by each patch type, or the number of different patch
types (patch richness). Configuration metrics might include the distribution of sizes of patches of
a certain type, the isolation or connectedness of patches on a landscape, or the complexity of patch
shapes. Landscape structure represents the patterns that have been quantified by landscape ecolo-
gists since the birth of the discipline. For example, Kashian et al. (2017) quantified the spatial
pattern of jack pine seedling densities across a recently burned landscape in northern Michigan as
an attempt to create a template for managers to mimic post-fire jack pine regeneration in planta-
tions (Figure 22.1). Within the burned area, seedlings varied in density from fewer than
100 seedlings per hectare to more than 5000 per hectare, and were classified into five patch types
Concepts of Landscape Ecology 613
Jack pine stems/ha
< 100 100–500 500–1000 1000–5000

(a) (b)
F I G U R E 2 2 . 1 Characterization of post-fire landscape structure using jack pine seedling densities across a
burned landscape in northern Michigan. Seedling densities were classified into five density classes on the
burned landscape and the pattern quantified using a few indices of landscape structure. These indices were
then used to simulate potential planting configurations that would best emulate the post-fire landscape.
(a) Seedling densities on the actual landscape; (b) simulated pattern of seedling densities. Areas shown
represent 164-ha regions of the burned area. Source: Kashian et al. (2017).
based on their density. Landscape structure was quantified using the proportion of landscape,
mean patch area, patch density (number of patches per 100 ha), and distance to the closest patch
of the same type for each patch type in the burned area. These indices were then used to simulate
similar landscape patterns that could be used as planting configurations that would emulate a nat-
urally regenerated landscape (Figure 22.1).
Landscape ecologists are concerned not just about the patterns present on a landscape, but
also about what the patterns mean for ecological processes. Landscape function describes how
the patches on a landscape interact with each other, often by examining flows of energy, mate-
rials, and species within and among them. A familiar example of interacting patches is the con-
cept of riparian buffers, which are regions of intact vegetation (often forests) around streams and
wetlands that act to increase water quality despite adjacent land uses. On agricultural land-
scapes, riparian buffers intercept sediment, nutrients, pesticides, and other materials in surface
runoff from fields into the stream or wetland, as well as reduce nutrients and other pollutants
moving with water flowing just below the soil surface. Riparian buffers are particularly impor-
tant in removing nitrogen, the excess of which may originate from fertilizers or animal waste on
agricultural landscapes and is a pollutant in surface and ground water (Carpenter et al. 1998).
Nitrogen removal from water in riparian buffers may occur via uptake by vegetation or microbial
activity, storage in the soil, or denitrification, the latter of which represents a loss of nitrogen
from the system in gaseous form. Patch size—in this case buffer width—plays an important role
in the effectiveness of buffers in removing nitrogen, with wider buffers more effective (Mayer
et al. 2007). The spatial arrangement of riparian buffers on a landscape is also important, as
maintaining buffers around stream headwaters appears to be most effective in sustaining water
quality (Peterson et al. 2001). Thus, the structure of the landscape (the abundance and
a rrangement of patches of vegetation adjacent to streams) has a clear effect on the ecological
process of nitrogen movement across the landscape.
A third major attribute of all landscapes is the process of landscape change, which exam-
ines how landscape structure and function change over time. Landscape change represents a major
component of landscape ecology in forests because it incorporates the temporal effects of distur-
bances, succession, land use, and management. If the relationship between landscape structure
and function is well understood, comparisons of how landscape patterns change over time allow
explanations of functional changes as well as predictions of how those functions may change in
the future. In a study in the midwestern United States, Radeloff et al. (2005) found that the number
of housing units grew by 146% between 1940 and 2000, particularly around the suburbs of major
metropolitan areas in the 1950s and 1960s and in rural forested areas separated from metropolitan
areas in the 1970s and 1980s (Figure 22.2). Together these trends of landscape change have
conservation concern for forests. Suburban sprawl intensively affects smaller forested areas with
higher housing density, but rural sprawl probably has a stronger negative impact because it affects
larger, less-altered forested areas, albeit at lower intensity. If continued, these changes have
potential to impact breeding habitat for bird populations, as well as large mammals. Thus, under-
standing landscape change in this region provides important opportunities for using ecological
principles in land use planning and policy development for growth management.
Landscape ecology can thus be summarized as being concerned with spatial heterogeneity at
a variety of scales. The nature of this heterogeneity has specific implications for the way ecological
flows of energy, nutrients, and species occur across the landscape, and those flows rely somewhat
on how well patches are connected to one another. Finally, landscapes are dynamic, and the het-
erogeneity they contain is likely to change over time (Wiens 2005). We described the importance
of ecological scale in Chapter 2; many of the other aspects of landscape ecology we cover in the
following text with a series of examples specific to forests.
1940 2000
Housing density (no./km2)

Water 4–8
0 8–32
0–2 32–128
2–4 > 128
F I G U R E 2 2 . 2 Housing density in 1940 and 2000 directing landscape change in the midwestern United
States. Housing growth occurred both as suburban sprawl along the fringes of metropolitan areas and rural
sprawl in previously unaltered forested areas. Source: Modified from Radeloff et al. (2005).
Forest Fragmentation and Connectivity 615
FOREST FRAGMENTATION AND CONNECTIVITY

Among the first concerns of landscape ecology was the idea of landscapes changing as a result of
human activities. The most concerning of these changes was the idea that human land use such as
forest clearing for agriculture, colonization, or forest harvesting was acting to break up large, con-
tiguous tracts of forests with deleterious ecological effects (Harris 1984). Given its occurrence at
broad spatial scales and its emphasis on spatial heterogeneity, forest fragmentation was a natural
focus for landscape ecologists. The classic example of forest fragmentation was shown in a town-
ship in southern Wisconsin where land clearing and farming created an increasingly fragmented
forest from 1882 (when only 30% of the original forest remained) to 1950 (Figure 22.3). By 1954,
1831 1882
1902 1950
F I G U R E 2 2 . 3 Fragmentation of a forested area of Cadiz Township, Green County, Wisconsin, during the
period of European colonization. The township is 6 miles on a side. The shaded areas represent land
remaining in, or reverting to, forest in 1882, 1902, and 1950. Source: Curtis (1956) / with permission
University of Chicago Press.
Curtis (1956) found only 3.6% of the original forest remained, and 77% of what remained was so
heavily grazed that tree regeneration was absent. The forest fragments are therefore functionally
isolated from other and larger areas of forest. Much of our understanding of forest fragmentation
and its ecological effects is based in patch-based ecological theory, as described in the following text.
PATCHES IN FOREST ECOLOGY

Many aspects of landscape ecology today are based on ecologists’ conceptualization of landscapes
as a mosaic of patches of habitats, ecosystems, or cover types. The patch–corridor–matrix model,
as described by Forman (1995), suggests that the extent and configuration of patches on a landscape
is what defines landscape pattern. As described in the earlier text, patches are to some extent
human constructs and are defined based on the phenomenon under consideration, such as forest
ages, forest types, or vegetation type (e.g., forest vs. field). Patches are considered to be distinct
areas at a given point in time, with boundaries different enough from what surrounds them that
they can be identified as a relatively homogenous area at a given spatial scale. Corridors are
narrow, linear patches of a patch type that differs from those on either side (Forman 1995) and
typically connect similar patch types on a landscape. Corridors may function as breeding habitat
for organisms that enables gene flow across a landscape, may allow dispersal or migration across a
landscape, or may act as a barrier or a filter to the movement of organisms. The most common or
connected landscape element is known as the matrix, which is often considered to be an unusable
space to the organisms or phenomenon of interest. Thus, the matrix in Figure 22.3 is non-forested
area beginning in 1882.
Patch-based ecological theory is heavily influenced by the theories of island biogeography
(MacArthur and Wilson 1967) and metapopulations (Levins 1969, 1970). Island biogeography
theory was originally developed to explain community dynamics on oceanic islands and suggests
that the number of species on island communities is affected by the size of the island and its dis-
tance from the mainland from which organisms disperse. Larger islands support more species
because they have more complex habitat and large populations less threatened by extinction. In
addition, islands further from the mainland support fewer species because they are colonized less
frequently. Therefore, large islands closest to the mainland are likely to have the most species, while
small islands far from the mainland are likely to have the fewest species. Island biogeography theory
is rather easy to transpose onto fragmented landscapes, allowing assumptions about biodiversity in
small, isolated patches of forest compared to larger, well-connected fragments.
Metapopulation theory adds an additional layer of complexity to patch-based theory.
Metapopulations include groups of populations that are spatially separated, as they would be when
occupying fragments of habitat on an altered landscape. Though separated in space, the popula-
tions interact through dispersal among the patches, forming a “population of populations”
(Levins 1969). Metapopulation theory predicts that a specific population in the metapopulation
may go extinct, leaving some patches on the landscape suitable but unoccupied, but the patch will
eventually be re-colonized by immigration from an occupied patch. Thus, the metapopulation as a
whole remains stable even if some of the populations within it do not.
Patch-based theory is one of the foundational aspects of forest landscape ecology and still
permeates the field today. In general, most forest ecologists have realized that the natural world is
not simply a series of patches surrounded by a featureless, unhabitable matrix. In such a scenario,
every species—including trees and other plants—would “perceive” a very different landscape,
making management or conservation nearly impossible. Instead, the matrix itself is made up of a
mosaic of different patch types that may be suitable for dispersal or other activities of life history
but not for breeding (Figure 22.4). As such, the matrix surrounding patches may not be completely
analogous to an inhospitable ocean surrounding islands. Forest fragmentation such as that shown
in Figure 22.3 may indeed create functionally isolated forests and woodlots for trees because the
F I G U R E 2 2 . 4 Distribution of eastern white pine on a portion of Harvard Forest, Massachusetts. White pine
forest is highly fragmented when only the white pine patch type and the matrix are considered (left), but in
reality the matrix forms a complex mosaic (right) of patches of other conifers (pink), mixed conifer stands
(lighter greens), open areas (yellow), hardwood stands (orange), mixed hardwood and conifer (brown), and
wetlands (blue).
nature of the matrix may prevent dispersal, regeneration, and/or establishment. We cannot
assume, however, that some organisms, particularly animals, may not be able to traverse the matrix
even as the patches within it become smaller and more isolated. This landscape-mosaic view is
probably a more accurate representation of the natural world than that of the patch–corridor–
matrix view, and its use for understanding and characterizing landscapes is at the forefront of
forest landscape ecology.
FOREST FRAGMENTATION
As described in the earlier text, human activities have reduced total forest area and have frag-
mented remaining forests into progressively smaller patches isolated by adjacent plantations,
roads, or agricultural and urban development (Harris 1984; Saunders et al. 1991). Forested land in
the 48 contiguous United States is estimated to have occupied 400 million ha in the 1500s
(Harrington 1991), and it was reduced by about 53% to approximately 188 million ha by the 1920s.
By the early 1990s only 3–5% of pristine old-growth forest remained, principally in the Pacific
Northwest (Miller 1992). Forest fragmentation is widespread in the eastern United States, where
remaining forests date only from the 1920s and 1930s (Rudis 1995; Vogelmann 1995) and small,
isolated woodlots interspersed in farmlands and suburbia comprise about 40% of the deciduous
forest (Terborgh 1989). We discuss in the following text the ecological effects of fragmentation (and
associated forest loss).
Ecological Effects Forests have always been naturally fragmented by large disturbances even
prior to European colonization (Chapter 16). Disturbances such as fires, windstorms, floods, and
avalanches characteristic of regional and local ecosystems created openings of various sizes on the
landscape, providing the heterogeneity necessary for a diverse array of animals and plants. Pattern
on modern landscapes differs from that prior to European colonization, however, at least in part
because changes in land use, conversion of cover types, or other drivers of current fragmentation
often prevent or reduce regeneration in disturbed patches.
An important consideration regarding forest fragmentation is that many of its effects are actu-
ally due to the loss of forest cover rather than simply its subdivision into relatively small, isolated
patches (Fahrig 2003). Most landscape ecologists agree that fragmentation affects forests on a
landscape by (i) reducing the amount of forest, (ii) increasing the number of forest patches,
(iii) reducing the size of the forest patches, and (iv) increasing the isolation of the forest patches.
However, these aspects are not independent of each other. For example, removal of some patches via
forest loss would reduce the number of patches but increase the isolation of those that remain
(Figure 22.5a), but subdivision of patches would increase the number of patches and reduce their
size while decreasing their isolation (Figure 22.5b). Thus, forest fragmentation and loss are closely
related to each other but are not equivalent. An increase in the number of patches on a landscape is
clearly caused by fragmentation rather than loss (With 2019), but both patch size and isolation may
be at least as well explained by loss as by fragmentation (Bender et al. 2003; Fahrig 2003). The impor-
tant assertion, therefore, is that forest fragmentation changes the function of the patches of forest
that remain as well as decreasing their size and increasing their isolation (van den Berg et al. 2001).
Fragmentation effects may be classified into those resulting from patch size and isolation, and
those resulting from edge effects. Predictions about the effects of decreasing patch size on the diver-
sity and abundance of species (most often animals) lean heavily on island biogeography theory:
larger patches are expected to favor larger populations compared to smaller patches, and thus
(a) (b)
Loss of patches: Subdivision of patches:

# patches reduced, # patches increased,
patch isolation increased, patch isolation decreased,
patch size unchanged patch size decreased
F I G U R E 2 2 . 5 Potential effects of forest loss on landscape pattern. Fragmentation is expected to increase

the number of small, isolated patches as it proceeds on a landscape, but these effects are not always
consistent and some may be achieved by forest loss alone. (a) Forest loss may increase patch isolation yet
reduce the number of patches, and (b) subdivision may increase the number of patches and patch size yet
decrease patch isolation.
extinction rates are expected to be lower in larger patches. Most species have a minimum patch size
requirement (Diaz et al. 2000), and thus fragmentation of forests may result in patches too small to
provide adequate heterogeneity for territory size, food supply (Whitcomb et al. 1981), or other
required features such as streams or wetlands. As described in the earlier text, patch isolation reflects
the amount of forest in the landscape more so than its fragmentation, but isolated patches are less
likely to be recolonized if a population within them goes extinct, depending on how well dispersers
are able to move through the matrix among the patches (Ricketts 2001). For example, small mam-
mals are more likely to move among forest patches through open habitat if the patches are close
together and exposure to predation in the open areas will be limited.
The most unambiguous indicator of fragmentation is the amount of edge that is increased as
the process takes place. The amount of edge increases proportionately as patch size decreases,
often represented as a ratio of the length of the patch perimeter to the patch area. Edge effects are
the ecological consequences of sharp boundaries between two patch types, and they tend to be
stronger than effects of patch area (With 2019). In forests, edge effects are often represented by
adjacent forests and disturbed areas or openings such as those created by harvesting; edges thus
ring the perimeter of each forest patch. The forest edge is characterized by relatively sharp bound-
aries with microclimatic, vegetational, and biotic conditions markedly different than in the interior
of the forest. In a series of studies of clear-cut harvesting in old-growth Douglas-fir forests of the
Pacific Northwest, Chen et al. (1992, 1995) found warmer air and soil temperatures during the day
and cooler temperatures at night at the edge compared to the interior forest, as well as lower
relative humidity, higher short-wave radiation, and exponentially higher wind speed. These effects
extended 30 to over 240 m into the forest from the edge. Although not tested explicitly for an edge
effect, canopy cover, basal area, and trees per unit area decreased in the forest nearest the edge,
dominant tree growth increased, tree mortality increased, and tree regeneration was impacted in a
species-specific way.
The abrupt, human-created sharp boundaries along forest edges differ from those to which
many forest species have adapted. Edges favor species adapted to them and select against those
requiring interior conditions (Whitcomb et al. 1981; Yahner 1988, 1995; Saunders et al. 1991),
potentially changing species richness. Such edge effects are evident with many forest animals but
are particularly obvious with birds. Edge species of birds that forage and breed near forest edges
are relatively independent of forest patch size and may be unaffected or even positively affected by
fragmentation. By contrast, forest interior species—those that depend on large forest tracts and
avoid edges and their effects—may decline in numbers with fragmentation. For example, Robbins
et al. (1989) found that encountering a breeding pair of an area-dependent bird species such as the
scarlet tanager is much more likely (>70% probability) in a 100-ha forest compared to a tract
smaller than 10 ha (<50%).
Edge effects also favor certain edge-dwelling game species (white-tailed deer, red fox), raptors
(great horned owl, red-tailed hawk), and songbirds (gray catbird, brown-headed cowbird; Yahn-
er 1995). However, edges clearly have a negative effect on many other fauna and flora such as neo-
tropical migrant birds, often because of higher predation and parasitism in edge habitat (Wilcove
et al. 1986; Robinson et al. 1995; Yahner 1995). For example, nest predation and parasitism by the
brown-headed cowbird increased with forest fragmentation in nine landscapes in Missouri, Indi-
ana, Illinois, and Wisconsin. Cowbirds are nest parasites that lay their eggs in the nests of other
“host” species, which then raise cowbirds at the expense of their own. Cowbird parasitism signifi-
cantly increases with decreasing amount of forest cover. In some cases, nest parasitism so decreased
reproductive success for some host species that they required immigration from source populations
in heavily forested, less-fragmented landscapes for persistence (Gustafson and Crow 1994).
A classic and severe example of forest fragmentation occurs in Florida, in large part due to the
influence of highways. Florida’s native species evolved in a setting surrounded by the sea and not
in the presence of high-density, high-speed traffic (Harris and Silva-Lopez 1992). Hard-surface
highways in Florida were built at a rate of over 6 km/day between the 1940s and 1990s to accommo-
date human population growth rate and tourism. In this case, highways act as barrier corridors to
animal movement, and automobile–animal collisions are the primary source of mortality for most
of Florida’s endangered large wildlife species, including the panther, black bear, key deer, and
American crocodile. Harris and Silva-Lopez (1992) conclude that roads are perhaps the most impor-
tant driver of forest fragmentation in the United States.
CONNECTIVITY
An important aspect of fragmentation is connectivity, which measures how well the structure of
a landscape enables or hinders movement or flows among its patches. Connectivity provides an
explicit measure of landscape function, in that it describes how pattern influences ecological
processes on a landscape (With 2019). As it relates to fragmentation, connectivity is generally con-
sidered to mitigate patch isolation, even in a subdivided landscape, if movement among the patches
is possible. The connections among patches on a landscape are typically functional, and as with
many aspects of landscape ecology, what constitutes a functional connection between patches
depends on the application or process of interest. Patches that are connected for birds, for example,
may not be connected for hydrologic flow. Connectivity might be defined for one species or process
as strict adjacency (i.e., patches that are touching or physically connected by a corridor), but by a
threshold distance (such as a maximum dispersal distance) for another. For example, in a study of
seed dispersal of hardwood species among forest fragments in southern Ontario, breaks in forest
cover of 15–30 m, composed of open fields or roads, were not sufficient enough to limit seed dis-
persal among forest patches (Hewitt and Kellman 2002). Therefore, the landscape remained
connected even when fragmented, presumably because the movements of birds and rodents were
not restricted. Small patches of habitat that are physically isolated but within the threshold distance
may exhibit connectivity by acting as stepping-stones for species’ movement across a landscape that
appears highly fragmented. Simulation modeling has shown that a landscape generally must be
about 30% occupied by a given patch type for that type to be functionally connected (Fahrig 2003).
Examples of connectivity affecting the movement of organisms are many. A study examining
the colonization of planted windbreaks in agricultural landscapes in Costa Rica found that forest
connectivity increased tree recruitment (Harvey 2000). When windbreaks were connected to intact
forests, tree seedlings were greater per unit area and tree species richness was higher compared to
windbreaks not connected to forests. In particular, bird-dispersed tree species were more common
and abundant in connected windbreaks because bird species were more likely to move through
forested areas than open areas, and dispersal was therefore more likely in the connected wind-
breaks (Harvey 2000). D’Eon et al. (2002) examined connectivity between patches of old-growth
forest, recent harvest, and recent wildfire in British Columbia as it varies among organisms’ ability
to move freely across its range. As described earlier in this chapter, old-growth patches were
increasingly connected as the amount of total old-growth increased, but decreased where harvest-
ing was more common. Carnivorous bird species, which were most able to move freely, perceived
the landscape as connected and were able to access all patches, but smaller, less mobile species
appeared to perceive a lack of connectivity. The degree of connectivity on landscapes has also been
proposed to have importance for the ability of species to shift their ranges under changing cli-
mates. Fragmented landscapes are likely to slow but not prevent range shifts unless connectivity is
reduced below the level at which metapopulations may persist (Opdam and Wascher 2004),
although some empirical studies have suggested that connectivity is only one of several factors that
may affect changes in species’ distributions (e.g., Melles et al. 2011).
Although most often considered for its effects on organisms’ ability to move across a
landscape, connectivity also has important implications for ecological flows and processes. Wildfire
spread and behavior are perhaps the most obvious examples of how ecological processes may be
influenced by forest connectivity. The spatial arrangement of fuels on a landscape is a major driver
of how fires spread, at least when weather conditions are not severe enough to override its impor-
tance (Turner and Romme 1994; Turner et al. 1994). Miller and Urban (2000) used a simulation
model to examine the interaction between surface fire frequency and the connectivity of burnable
area for forests in the Sierra Nevada in California. Connectivity of fuels was higher where fire was
less frequent, probably because fuel is able to accumulate during longer fire intervals. However,
fuel moisture also plays a critical role in fuel connectivity. Surface fuels become very connected
when fuel moisture is very low and very disconnected when it is very high (Figure 22.6). Thus, the
1% fuel moisture 5% fuel moisture

p = 0.86 p = 0.59
CL = 136m CL = 135m

p = 0.34 p = 0.21
CL = 69m CL = 40m

p = 0.10 p = 0.02
CL = 21 CL = 14m
F I G U R E 2 2 . 6 Simulated effects of fuel moisture on surface fuel connectivity in the Sierra Nevada of
California. The proportion of the map that is burnable (p) and the correlation length (CL) are shown for each
map. Correlation length is an index of connectivity that reports the average distance a fire can spread
without leaving a patch of burnable area (higher correlation length suggests higher connectivity). Surface
fuels become far more connected as fuel moisture decreases. Source: Miller and Urban (2000) Miller and
Urban, 2000 / with permission Springer Nature.
spatial arrangement of fuels probably determines fire behavior and spread prior to fire suppression
at intermediate fuel moisture levels. Fuel connectivity likely has no effect during wet periods or
during droughts (Miller and Urban 2000). Fire suppression likely increases fuel connectivity by
allowing fuel loads to accumulate and may result in more widespread fires.
Forest fragmentation is one result of a suite of human activities that have intensified con-
cerns of ecologists and the public at large over diminishing forest lands, declining native plant and
animal diversity, concomitant increases in exotic and invasive species, and sustaining natural eco-
logical processes. Our understanding of forest fragmentation effects has led to an examination of
forest and resource management practices and basic changes in policies and practices to sustain
ecosystems. Detailed treatments of fragmentation are provided by Lindenmayer and Fischer
(2013), Collinge (2009), and Bradshaw and Marquet (2002).
DISTURBANCES ON LANDSCAPES
Disturbances have been significant in shaping the physical features of landscape ecosystems and
are instrumental in the evolution of biota. Heterogeneity on landscapes may affect the susceptibil-
ity and spread of disturbances, but disturbances may also act as an important driver of heterogene-
ity. The fact that forest disturbances both respond to and create landscape pattern makes them
conventional objects of study in landscape ecology (Turner and Gardner 2015). We examine both
of these interactions in the following text using examples of major disturbance events.
EFFECTS OF HETEROGENEITY ON DISTURBANCES

We discussed in Chapter 16 that specific regional ecosystems and certain local ecosystems are
more vulnerable to some types of disturbances because of the unique combinations of characteristic
physical site factors and biota. The susceptibility of certain ecosystems to disturbances can be as-
sessed by examining the likelihood that the disturbance will occur, typically using historical occur-
rences of the disturbance to calculate a probability. When the historical record of disturbances is
long, it is possible to identify the regional ecosystems most likely to experience a given disturbance
as well as the physiographic sites within the regional ecosystem where disturbance is most
common. Long-term records of disturbances in forests in New England have shown differential
disturbance effects that correspond to these sites.
Hurricanes in New England The forests of central New England have been repeatedly affected
by many types of disturbances, but the ecological effects of hurricanes have been most influential
and studied at several spatial scales (Foster 1988a, b; Boose et al. 1994; Foster and Boose 1992, 1995).
Hurricanes have historically struck southern and central New England every 20–40 years, with the
most destructive storms in 1625, 1788, 1815, and 1938 moving over fairly generalized pathways
(Figure 22.7). The 1938 hurricane included winds exceeding 200 km h−1 and affected forests along a
100-km2 path, leaving a mosaic of differentially damaged forest controlled by physiography, wind
direction, and the nature of the pre-hurricane vegetation. Southeastern winds were the strongest,
and the most severe damage to forests occurred on south-to east-facing slopes and the northwestern
shores of lakes. By contrast, damage was far less to forests in the lee of broad hills (northwest
exposure), on the southeast and east shores of lakes, or in valleys. The severity of vegetation damage
was related to physiography and forest type. For example, damage was most severe where exposure
was greatest, such as immediately adjacent to boundaries of a pond and a large field (Figure 22.8).
Regionally, damage severity was also related to historical factors, such as the time since the
last major storm and the previous pattern of human clearing and agriculture which, in turn, affected
the heterogeneity of the pre-hurricane vegetation. Landscapes of central New England were espe-
cially prone to disturbance in 1938, because 30-to 100-year-old white pines dominated old fields and
Disturbances on Landscapes 623
N 1815 CANADA
1788
1938
e
Main
ni
Vermo
Frequency of
occurrence of
severe damage ire
ampsh
to forests due New H
to hurricanes,
1620–1950 tts
chuse
Massa H.F
1
2
c ticut
Conne
3
4
5–6
D
B
ATLANTIC OCEAN
A
F I G U R E 2 2 . 7 Four generalized pathways (A–D) of hurricanes in the New England region superimposed over
a map of the frequency of occurrence of the major hurricanes from 1620 to 1950. In addition, the historical
tracks of the hurricanes of 1788, 1815, and 1938 are shown. Hurricane damage ratings range from: 1 = light to
5–6 = severe. HF = location of the Harvard Forest, Petersham, MA. Source: Foster and Boose (1992) and
Whitney (1994). Reprinted from the Journal of Ecology with permission of the British Ecological Society.
were highly susceptible to wind damage. A similar hurricane occurred in 1815 but caused much less
damage because it occurred at the height of the agricultural period prior to reforestation. A legacy
of the 1938 storm was that it destroyed older conifers and tall pioneer species, leaving more storm-
resistant hardwoods and young understory conifers (Figure 22.9).
Despite an expectation that the 1938 hurricane would create large openings across a region,
whereas smaller storms create small gaps, the hurricane produced a continuum of damage from
individual trees to uprooting of entire stands. Entire windthrown stands are spectacular, but the
hurricane actually created a fine-grained heterogeneous mosaic of damage, with most patches
smaller than 2 ha (Foster and Boose 1992). This heterogeneity was due in part to broadscale physi-
ography, land-use history, and the interrelated factors of stand structure and species composition.
Stand-scale damage was closely related to topography and increasing stand age and height. Higher
Medium H. Canopy of Young H

pioneer
tree species Severe
Ridge Gap
damage
H.
g
un
Old H.
H.
nd
Yo
Low or
Po
ng
ld
Fie
u
moderate
Yo
damage
Medium H
Escapes
Old C. damage
Swamp
Young C.
Medium C. C = conifer
gC
H = hardwoods
n
You
Medium H.
Hurricane
N winds
Medium H. Pond
0 meters 500 Hilltop SE
0 1500 slope
feet Medium H.
F I G U R E 2 2 . 8 Diagram of post-hurricane damage classes from a characteristic landscape in southern New

England. Young, medium-age, and old stands of conifers (C) or hardwoods (H) represent a hypothetical
pre-hurricane vegetation mosaic. Conifer = white pine; hardwoods = oaks, red maple, birch, beech,
hickories (and hemlock); pioneers = pines, aspens, white birch. Predicted damage distribution is primarily
based on studies of Foster (1988a,b) and Foster and Boose (1992). Source: After Forman (1995) / Cambridge
University Press.
Photo by Stephen H. Spurr. Reprinted courtesy of the
Harvard Forest, Harvard University
FIGURE 22.9 The Harvard Tract of the Pisgah National Forest in 1942, 4 years after the 1938 hurricane.
stand density also reduced damage, apparently because stand opening increased air turbulence dur-
ing the storm. White-pine-dominated conifer forests were significantly more susceptible than hard-
wood forests, as were fast-growing species that tended to occupy dominant canopy positions such as
red pine, aspens, and white birch. Slower-growing oaks, hickories, red maple, and hemlock occu-
pied subordinate canopy positions and were more protected from wind. The 1938 hurricane was
also notable for the catastrophic uprooting of trees, especially white pine, attributed to soil satura-
tion and reduced rooting strength resulting from the 15–35 cm of rain that preceded the storm.
Although the damage to forests caused by hurricanes in New England is intimately related
to exposure as mediated by physiography and associated ecosystem characters, not all disturbances
are influenced by landscape pattern. For example, in the Upper Peninsula of Michigan, topo-
graphic relief is far more subtle than in New England, and most widespread wind damage in this
region is caused by vertical downbursts from thunderstorms rather than horizontal winds from
hurricanes. In studying this system, Frelich and Lorimer (1991) were unable to find any effect of
physiography (including slope position, aspect, or slope inclination) on patterns of wind distur-
bance, reflecting stark differences in regional weather patterns and broadscale physiography com-
pared to hurricane-prone areas. Likewise, Hurricane Katrina struck the forests of the central Gulf
Coast in southern Mississippi in 2005, an area also lacking appreciable topography. Stand age
rather than physiographic characteristics best predicted damage severity in forests affected by this
storm; aspect reflects gently sloping topography and was only marginally important in predicting
damage (Kupfer et al. 2008). These studies reiterate that although landscape pattern may affect
disturbances, these ecological effects are firmly rooted in the regional and local landscape ecosys-
tems in which they occur.
Landscape Pattern Effects on Disturbance Spread The spread of a disturbance across the
landscape may also be affected by heterogeneity, although studies are somewhat inconsistent in
showing this relationship. Generally, disturbances spreading within a cover type are dampened when
the landscape is more heterogeneous, and disturbances spreading among cover types are enhanced
(Turner et al. 1989). For example, many coniferous forests in western North America were impacted
by widespread and severe outbreaks of bark beetles in the early 2000s. Given that most bark beetles
are host-tree-specific, outbreak spread generally occurs within a cover type. Though many factors
govern the occurrence of outbreaks, a lack of heterogeneity in forest type and age likely enhanced the
spread of outbreaks, especially where forests were widespread and well connected (Raffa et al. 2008).
A study of the astonishing mountain pine beetle outbreaks in western Canada in the early 2000s
found that almost 70% of the forest area dominated by the beetle’s primary host, lodgepole pine, had
simultaneously reached the preferred age and size for beetle attack (Taylor and Carroll 2004). This
extremely low landscape heterogeneity clearly translated to high susceptibility to this disturbance.
By contrast, disturbances that spread among multiple cover types may be enhanced by
heterogeneity. In studying outbreaks of the forest tent caterpillar in northern Ontario, Roland
(1993) found that increasing amounts of edge on a landscape had strong effects on outbreak dura-
tion, and fragmented landscapes in general also experienced longer outbreak duration.
Fragmentation may favor the caterpillar by restricting dispersal of the pathogens that limit its
populations, and the warmer microclimate of edge habitat may favor caterpillar development
(Roland 1993). In this case, heterogeneity represented by fragmentation enhanced the spread of
the outbreak as it moved between forested and non-forested patches.
EFFECTS OF DISTURBANCES ON HETEROGENEITY

In describing how landscape heterogeneity affected disturbances in New England, we referenced
the idea that the 1938 hurricane created a fine-grained mosaic of damage severities to forests. This
mosaic of disturbance severities imparts heterogeneity in vegetation structure and composition as
the forests recover. This heterogeneity is most obvious following large disturbances, and Foster
et al. (1998) compared the patterns left by a diverse series of five large disturbance events in the
1980s and 1990s (Figure 22.10). Hurricane events in New England, as described in the earlier text,
create a fine-grained mosaic of damage severities that feature uprooted or broken trees, damaged
trees left alive, and undamaged forests, all varying in structural characteristics. In contrast to hur-
ricanes, wind-related damage patterns to forests due to tornadoes are very complex due to their
extreme variability and the potential for multiple touchdowns. An F4 tornado in the Tionesta
Scenic Area in northwestern Pennsylvania in 1985 created a path of damage 19 km long and 1 km
wide, with extremely sharp boundaries between damaged and undamaged areas but little influence
of physiography on the patterns of damage severity (Figure 22.10).
Flooding of riparian areas occurs regularly, but extreme flooding such as the 1993 Missis-
sippi River flood was notable for its effects on landscape pattern. Long periods of inundation in
lower-lying areas of the floodplain, as well as flooding of areas of floodplain occupied by less
flood-tolerant species, caused widespread tree mortality across the region. Landscape patterns
linked to exceptional flooding events are therefore strongly influenced by the topography of river
floodplains. Landscape patterns created by volcanic eruptions, such as that of Mount St. Helens in
1980, are often a combination of broad area of damage with an epicenter in the blast zone and
“fingers” of damaged areas resulting from lava or pyroclastic flows (Figure 22.10). Forests within
the blast zone at Mount St. Helens were subject to an explosive blast, a thermal wave, and the
deposition of rock and ash. Landscape patterns were very complex within the blast zone due to
influences of pre-eruption vegetation patterns, physiography, and multiple disturbances, and the
degree of heterogeneity is likely to differ depending on the season of eruptions (Foster et al. 1998).
Finally, large wildfires such as those that burned in Yellowstone National Park in 1988 very clearly
create heterogeneity across the landscape as variability in forest age, structure, or species
5 km
0 100 km 5 km
Atlantic Ocean
Hurricane (1938 New England) Fire (Yellowstone National Park) Volcano (Mount St Helens)
5 km
Tornado (Tionesta Scenic Area)

5 km
Flood (Mississippi River)
F I G U R E 2 2 .1 0 Comparison of landscape patterns of five contrasting large disturbances in forests across

North America. For each disturbance, high-severity damage is shown in black, and lesser damage in gray.
Disturbances include the 1938 hurricane in New England, the 1988 fires in Yellowstone National Park, the
1980 eruption of Mount St. Helens, an F4 tornado at the Tionesta Scenic Area in Pennsylvania in 1985, and
the 1993 Mississippi River floods. Source: From Foster et al. (1998) / Springer Nature.
c omposition. Next, we use Yellowstone as an ideal example to illustrate the role of wildfires in
creating significant landscape pattern.
Stand-Replacing Wildfires in Yellowstone National Park Yellowstone National Park is

a large wilderness area in the central Rocky Mountains ideal for examining landscape changes
that occurred for millennia prior to European colonization. Romme (1982) and Romme and
Knight (1982) studied the fire history of lodgepole-pine-dominated forests in the 73-km2 Little
Firehole River watershed prior to the extraordinary Yellowstone fires of 1988. Fire history during
a 350-year period was determined using tree-ring and fire-scar analyses (Chapter 10). Major
wildfires in 1739, 1755, and 1795 burned more than half of the watershed. Large, stand-replacing
fires (>1000 ha) were generally absent until 1988; only three fires >4 ha burned from 1795 to
1978, all of which were <100 ha. Fire suppression was not effective until the late 1940s, so the
absence of large fires was largely due to lack of suitable fuel conditions over most of the water-
shed. Large fires occur infrequently in Yellowstone (every 100–300 years) because of very slow
forest regrowth and fuel accumulation after fire (Romme 1982). Subalpine plateaus feature such
large, infrequent fires in part because physical site factors (geology, physiography, and soils) and
vegetation are relatively homogeneous over much of the area; large, contiguous forests over
broad areas grow and accumulate fuel at approximately at the same rates; and fire spreads easily
because the plateau topography has low relief and few natural barriers.
Romme and Knight (1982) identified three main stages of succession following fire on
upland sites (early, 0–40 years; middle, 40–250 years; late, >250 years). The even-aged pine canopy
that developed after fire deteriorates as the stands age and is replaced by trees able to survive in the
understory and seed into gaps. Most of the subalpine plateaus in Yellowstone are dry and domi-
nated by lodgepole pine, but subalpine fir and Engelmann spruce are common and may co-dominate
with pines on occasional moist sites, particularly in older forests. This relatively homogeneous
site-vegetation pattern contrasts with the diverse mosaic of physiography, human-use history, and
disturbance regime of central New England. As a result, the forest landscape mosaic in Yellow-
stone is largely based on forest age and structure rather than species composition. Large fires in
1739, 1755, and 1795 reduced the high proportion of old-growth forest present in the early 1700s
and replaced it with early-successional forests (Figure 22.11). As the forests burned by these fires
aged, middle-successional stages became most abundant around 1800 and gradually transitioned
to late-successional forest.
After creating a map of the mosaic resulting from multiple fires over 200 years, Romme
(1982) applied species diversity indices (Chapter 14) to the mosaic to characterize landscape diver-
sity from 1778 to 1978. Changes in landscape diversity were quantified using (i) a weighted average
of richness (number of stand ages), evenness (proportion of the landscape occupied by each stand
age), and patchiness; and (ii) the Shannon–Wiener index. Both measures show that the large fires
of 1739, 1755, and 1795 created high diversity in the late 1700s and early 1800s, and a lack of large
fires in the late 1800s during a 70-year period with no major fires resulted in low diversity
(Figure 22.12). Two small fires and a series of mountain pine beetle outbreaks in the twentieth
century led to differential rates of succession that increased landscape diversity (Romme 1982).
These differences in landscape diversity over time in turn had important effects on bird and ungulate
populations, which are described in detail by Romme and Knight (1982).
To this point, we have discussed heterogeneity in Yellowstone as having resulted from a se-
ries of stand-replacing fires over several centuries. As described in the earlier text, single distur-
bance events also may create complex landscape patterns, particularly if they affect broad spatial
extents (Foster et al. 1998), as illustrated by the extensive Yellowstone fires of 1988. As described in
the earlier text, very large, stand-replacing fires were a dominant factor on the Yellowstone
landscape prior to European colonization, with such fires occurring at intervals of 100–300 years
(Romme and Despain 1989). Due to prolonged, extreme drought and high winds that occurred
Middle
60
50
% Watershed Area
40
30 Late
20
10
Early
40 % Watershed Area Burned

1739 1795
20
1755 1834 1905 1949
1738 ‘58 ‘78 ‘98 1818 ‘38 ‘58 ‘78 ‘98 1918 ‘38 ‘58 ‘78
Year
F I G U R E 2 2 .1 1 Percentage of the 73 km2 Little Firehole River watershed occupied by early, middle, and late
stages of forest succession from 1738 to 1878 in Yellowstone National Park, northwestern Wyoming.
Source: Romme and Knight (1982) / with permission of Oxford University Press.
(a) Average of Richness, Evenness,

100 and patchiness Indices
90
80
(b)
Shannon Index
.90
.80
.70
1778 ‘98 1818 ‘38 ‘58 ‘78 ‘98 1918 ‘38 ‘58 ‘78
Year
Time of effective
fire control
F I G U R E 2 2 .1 2 Changes in landscape diversity in the Little Firehole River watershed from 1778 to 1987 in
Yellowstone National Park. (a) Average of richness, evenness, and patchiness indices; Y axis = percentage of
maximum possible average of richness, evenness, and patchiness indices. (b) Shannon–Wiener diversity
index; Y axis = percentage of maximum possible Shannon–Wiener diversity index. Maximum possible
Shannon–Wiener index would occur where all stand ages have equal coverage in the landscape.
Source: Romme and Knight (1982) / with permission of Oxford University Press.
during the summer of 1988, fires burned approximately 250 000 ha of lodgepole pine forest. In con-
trast to expectations at the time, what resulted was not a homogenous landscape, but a markedly
heterogeneous one, including a mosaic of burned and unburned patches with varying burn sever-
ities (Turner et al. 1994).
Sooner than 5 years after the 1988 Yellowstone fires, lodgepole pine regeneration varied sub-
stantially across the landscape, from areas with no seedlings to areas with >500 000 seedlings per
hectare (Turner et al. 2004). Seedling densities truly varied in a fine-grained mosaic, with mean
patch size 1.5 ha, 68 patches occurring per 100 ha, and similar patches separated on average by only
150 m (Kashian et al. 2004). Overall, small, dense patches of seedlings occurred within a matrix of
large, sparser patches across the landscape (Figure 22.13). Variation in burn severity was impor-
tant in determining post-fire seedling density; areas burned by severe surface fires had higher seed-
ling densities than areas burned by crown fire or light surface fires (Turner et al. 1994). Lodgepole
pine has serotinous cones (Chapter 10) that require heat to open, but may be incinerated by high-
intensity fires. While crown fires may have incinerated many cones and light surface fires may
have been insufficient to open them, severe surface fires likely provided the appropriate burn tem-
peratures to maximize post-fire regeneration (Turner et al. 1994). The expression of the serotinous
trait by lodgepole pine also varies across the landscape and is not well understood, but higher
proportions of trees producing serotinous cones in a stand are correlated with shorter fire intervals
and moderately aged stands (Tinker et al. 1994; Schoennagel et al. 2003).
The heterogeneity created by the 1988 fires in Yellowstone has many implications for the
future landscape. Although succession and stand dynamics will proceed as the post-fire forests age,
Legend
Sapling density (stems/ha)

0–1000
1001–5000
5001–10 000
10 001–20 000
20 001–50 000
> 50 000
Lakes
Roads
Kilometres
0 5 10 20
F I G U R E 22 .1 3 Heterogeneity of lodgepole pine regeneration density across the area burned in Yellowstone
National Park in 1988. Regeneration density varied over six orders of magnitude in a fine-grained, complex
spatial mosaic. Source: Kashian et al. (2004) / with permission of Canadian Science Publishing.
the landscape pattern created in 1988 is likely to leave a long-lasting legacy of the fires for decades
to centuries. Studies of stand structural dynamics on the unburned portion of the Yellowstone
landscape found high variability in stand density that decreased over time (Kashian et al. 2005b).
Most stands on the landscape exhibited classic dynamics of a fire-regenerated, shade-intolerant
species, showing evidence that they had initially been relatively dense and their density had
decreased over time due to self-thinning (Chapter 17). However, several stands 50–150 years old
had clearly regenerated sparsely after the fire that last burned them, allowing gradual and contin-
uous recruitment to occur and density to remain stable or increase over time. Younger stands were
not only more dense, but their density varied more across the landscape; stand density was gener-
ally similar among stands 200 years and older (Kashian et al. 2005b). The initial patterns of regen-
eration density produced by the 1988 fires largely remained in 2012 such that convergence had not
yet occurred, but the mechanisms that would likely cause convergence across the landscape were
observed (Turner et al. 2016). Notably, smaller fires (<5000 ha) that burn during the interval bet-
ween large fires (Despain 1990), such as those examined by Romme (1982), become the main
source of heterogeneity once stand structures converge following larger fires.
HISTORICAL RANGE OF VARIABILITY

Related to the concept of disturbance and the successional processes that follow is the idea that
landscape patterns are not static, as illustrated in the earlier text by the studies of convergence in
stand structure and function in Yellowstone National Park. The fact that landscape patterns change
over time suggests that a single ecological description of a landscape would serve as a poor bench-
mark because it is not likely to capture the full range of natural conditions that occur over time. The
historical range of variability (HRV; also called natural range of variability) is the variation in
ecological conditions that occurs over time and space within a specified time frame and area with-
out human intervention (Landres et al. 1999). The benefit of identifying an HRV is that ecologists
and managers may assess the degree of alteration of current landscapes using a range of historical
conditions rather than a specific condition at a single point in time. The HRV concept is useful in
restoring current landscapes to historical conditions as well as designing management, such as
forest harvesting, in such a way that it mimics historical conditions.
Two of the most important disturbance types that affect forests are fire and harvesting, par-
ticularly because humans have heavily influenced the extent and severity of each. As such, many
(but not all) HRV studies have occurred in forests of western North America where these distur-
bances are particularly prevalent. A comparison of the landscape pattern of the wilderness area of
Yellowstone National Park to that of the adjacent Targhee National Forest in Wyoming is an illus-
trative example of the use of HRV. Using a map of historical fires in Yellowstone as a reference
landscape, Tinker et al. (2003) calculated a series of landscape metrics at 20-year intervals for the
period 1705–1995 (Figure 22.14). Metrics included the total number of patches, average size of
patches, the variation in patch size, and the amount of edge. The HRV for each metric was esti-
mated using the range of values for each metric during the study period. The same metrics were
then calculated for the Targhee National Forest, which had undergone clear-cut harvesting since
the 1970s, prior to and after the onset of harvesting. The pre-harvest landscape of the National
Forest fell within the HRV for each of the four metrics, but the post-harvest landscape did not
(Figure 22.14). The post-harvest landscape featured more patches that were smaller and less vari-
able in size compared to the HRV, indicating a heavily altered landscape pattern due to harvesting.
An example of the use of HRV in forests of eastern North America is found in jack pine for-
ests of northern Michigan. Jack pine forests of this region are the breeding grounds for the rare
Kirtland’s warbler, a once-endangered songbird that nests beneath stands approximately 5–25 years
old. Like lodgepole pine, jack pine is a species with serotinous cones (Chapter 4) and is dependent
on fire for its regeneration, and thus the warbler’s need for young jack pine stands leaves it also
(a) Number of Patches (b) Mean Patch Size
1600 400
1400 Targhee NF, 350 1785 1945 1965 1985

1825 1865 1905
Post-harvest
1200 300 1705 1745
Targhee NF,
Pre-harvest HRV
1000 Patch Size (ha) 250
# Patches
800 1995 200
Targhee NF - 1995
600 150
Pre-harvest HRV
1705 1745 1905
1785 1825 1865 1945 1985
400 100
Targhee NF,
200 50 Post-harvest
0 0
(c) Patch Size Standard Deviation (d) Edge Density

7000 30
1985 Targhee NF,
Post-harvest
6000
1945 1965 25
1825 1865 1905
1785 Targhee NF,
5000
Edge Density (m/ha)
Pre-harvest 20
1995
1745 1885 HRV
PSSD (ha)
4000
1705 15 1705
1745
HRV
3000 1825 1865 1905
1785 1945 1965 1985
1995 10
2000 Targhee NF,
Targhee NF, Pre-harvest
Post-harvest 5
1000
0 0
F I G U R E 22.14 (a)–(d) Landscape metrics illustrating the deviation of the harvested landscape of the Targhee National Forest from the historical
range of variability of the reference landscape of Yellowstone National Park. Source: Tinker et al. (2003) / with permission of Springer Nature.
Mean Landscape Age
KW
Historical range
of variability
Range of variability
since KW
management
Time
F I G U R E 2 2 .1 5 Conceptual representation of the historical range of variability of jack pine stand ages
across the landscape in northern Lower Michigan. Prior to European colonization, the landscape was likely
to be both younger and older than the 2015 landscape at various points in time, but current Kirtland’s
warbler habitat management has restricted the variability compared to its historical range. Source: Tucker
et al. (2016) / with permission of Springer Nature.
heavily dependent upon fire. Fire suppression in the region in the twentieth century left the war-
bler reliant on forest management that provided young jack pine habitat using extensive planta-
tions, typically at the expense of mature jack pine forests on the landscape. Historical fire return
intervals in jack pine forests of the region were 30–60 years, and the conversion of older forests to
young plantations was thus assumed to be an accurate mimicry of historical landscape conditions.
Using original General Land Office survey notes collected in 1858, Tucker et al. (2016) reconstruct-
ed the distribution of historical age classes of jack pine forests across northern Lower Michigan.
They found that the landscape under warbler management in 2015 was much younger than that
in 1858; historical coverage of stands <20 years old was only 5% compared to 31% in 2015 and
stands >50 years old occupied 76% of the landscape compared to 30% in 2015. This study did not
specifically estimate an HRV (as did Tinker et al. 2003), and in fact the 2015 landscape likely fell
within the HRV when one considers the historical tendency for this landscape to be burned by
large fires. However, the marked difference in the age structure of the landscape prior to European
colonization in 1858 suggests that the historical landscape was at least occasionally older during
fire-free periods. Thus, forest management for the Kirtland’s warbler, which maintained a much
higher proportion of the forests on the landscape at a young age, had drastically reduced the vari-
ability of landscape age over time and may eventually eliminate it (Figure 22.15).
INTERACTIONS OF LANDSCAPE PATTERNS AND ECOLOGICAL PROCESSES

At the beginning of this chapter, we defined landscape ecology as the study of the interaction bet-
ween pattern and process, and we might consider this to be a bi-directional relationship (Turner
and Gardner 2015). For example, in our references to landscape ecosystems throughout this book,
we have illustrated how ecological processes influence vegetation patterns across the landscape. In
forming ecosystems, physical site factors including climate, physiography, soil, and hydrology
influence ecological processes such as nutrient cycling (Chapter 19) or ecosystem productivity
(Chapter 18), and patterns of species composition and cover type in turn reflect this heterogeneity.
The opposite is also true, in that heterogeneity affects ecosystem processes and how they occur at
broad spatial scales.
Much of our understanding of ecosystem processes in forests is articulated at a stand or site
level from studies whose design ensured homogeneity of study sites to the extent possible. When
Interactions of Landscape Patterns and Ecological Processes 633
ecosystem processes are considered at broad scales, heterogeneity creates an additional layer of
complexity that must be accounted for, but it also provides additional insights (and potentially dif-
ferent answers) about how ecosystem function occurs. Turner and Chapin (2005) developed a
framework distinguishing point processes, which generally include ecological rates, from lateral
transfers, which include transport processes of matter among patches on a landscape. Point
processes are typically measured at a specific place on the landscape, such as net primary produc-
tion, carbon storage, or nutrient cycling, and highlight variability in the rates and the potential
causes of that variability. Lateral transfers focus upon the flow of materials from one place to
another, such as the flow of nitrogen across an agricultural landscape in the context of riparian
buffers. We focus mostly on point processes in this section to briefly illustrate the interaction bet-
ween heterogeneity and ecological processes. Detailed treatments on the effects of heterogeneity
on ecological processes are presented by Turner and Gardner (2015), Lovett et al. (2005), and
Hutchings et al. (2000).
LEAF AREA AND PRODUCTIVITY

The prevalence of climate change has spawned many recent studies that have examined carbon
storage and ecosystem productivity over large extents. One of the earliest attempts at understand-
ing how carbon dynamics vary at broad spatial scales examined evapotranspiration (ET) and pho-
tosynthesis for a 1540 km2 coniferous forest landscape in northwestern Montana (Running
et al. 1989). These processes represent important interactions between vegetation at the Earth’s
surface and the atmosphere and are closely related to net primary productivity (Chapter 18).
Within a geographic information system, Running et al. (1989) determined leaf area index (LAI)
from satellite imagery for the forest in 1200 grid cells, and ran an ecosystem model in each cell
using LAI as an input variable. The model estimated annual ET and net photosynthesis for the
landscape and demonstrated the enormous degree of spatial variability in these ecosystem
processes and its association with physiography. For example, leaf area was highest where precip-
itation was the highest, which happened to be on the western slopes of the study area. Soil water
holding capacity was also an important determinant of LAI and was lower on steep slopes or shal-
low soils often in spite of higher precipitation (Figure 22.16). Variation in ET was in part due to
land use, where low ET was associated with young post-logging regeneration. ET dried out soils
earlier in the season where snowmelt occurred earlier, and those regions of the landscape experi-
enced a longer period of soil moisture stress. Photosynthesis reflected LAI but varied with temper-
ature, decreasing at low elevation due to warmer temperatures and at high elevations due to cooler
temperatures, and varied over the study area with soil moisture (Figure 22.16). Overall, this work
showed the importance of quantifying the variability of ecosystem processes in attempting to
understand and predict carbon storage and flux over large landscapes.
Heterogeneity in ecological processes that affect carbon dynamics may result from distur-
bances as well as physical site factors. Carbon dynamics, particularly when heavily influenced by
aboveground productivity, are closely tied to disturbances and the recovery and succession of for-
ests that follow them (Bradford et al. 2008a; Kurz et al. 2008b). For example, the heterogeneity of
lodgepole pine regeneration following the 1988 Yellowstone fires (Figure 22.13) had important
implications for leaf area and aboveground net primary productivity (ANPP) approximately
10 years after the disturbance (Turner et al. 2004). In fact, ANPP was closely related to seedling
density and varied in a fine-grained mosaic much like post-fire regeneration. About 68% of the
burned landscape had low ANPP (<2 Mg ha−1 year−1), reflecting a matrix of lower-density patches
of seedlings, and 10% had high ANPP (>4 Mg ha−1 year−1). Similar to seedling density, Kashian
et al. (2005a) predicted that the variation in productivity would decrease over decades or centuries
during the fire-free interval as contrasting stand structures began to converge, and the initiation of
this convergence has already been observed 24 years after the fire (Turner et al. 2016).
4–6
6–8
8–10
10–12
12–15
LEAF AREA INDEX
25–34
35–39
40–44
45–49
50–54
> 55
EVAPOTRANSPIRATION (cm/yr)
9–11
12
13
14
15
16
17
18
> 19
PHOTOSYNTHESIS (Mg · ha–1· yr–1)
F I G U R E 2 2 .1 6 Heterogeneity of evapotranspiration and net photosynthesis as determined by leaf area

index and an ecosystem model for a 1540 km2 landscape in northwestern Montana. Heterogeneity of the
ecosystem process closely shows physiographic differences of the forested landscape. Source: Running et al.
FOREST CARBON DYNAMICS

Incorporating heterogeneity is particularly important for creating carbon budgets and for under-
standing the carbon balance on landscapes because of the variability of forest species composition,
age, and other factors. A carbon balance may be determined from leaf to global scales, but net eco-
system productivity (NEP; also termed net ecosystem carbon balance, or NECB) in a forest essen-
tially represents the balance between carbon fixed into live biomass via photosynthesis and that
lost from dead biomass as it decomposes (Chapter 18). The importance of NEP for large landscapes
relates to whether a region acts as a carbon source (losing more carbon to the atmosphere than it
gains) or a carbon sink (gaining more carbon than it is loses to the atmosphere). Because forests are
a major pool of carbon on the earth’s surface, and they act to remove carbon from the atmosphere,
disturbances that impact forests are especially important in determining whether a landscape is a
source or sink of carbon and therefore how it contributes to climate change (Chapter 20).
As discussed in Chapter 18, forests are typically carbon sources immediately after a distur-
bance when small, live trees with relatively low ANPP do not compensate for carbon lost from high
levels of dead, decomposing biomass killed by the disturbance. Depending on the forest type, NEP
reaches its maximum several years or decades after the disturbance when tree growth is most vig-
orous (primary productivity is highest) and most dead biomass has decayed away, and forests then
act as strong carbon sinks. By contrast, older forests have accumulated large stocks of carbon even
as primary productivity declines and dead biomass begins to re-accumulate, and NEP approaches
zero. Many studies employ a chronosequence approach to determine how carbon storage changes
after a disturbance, whereby a series of stands are sampled that vary in age, and all differences
among the stands are assumed to occur due to the effects of succession (Chapter 17). For example,
Rothstein et al. (2004) sampled 11 jack pine stands ranging in age between 1 and 72 years in
northern Michigan. They found that jack pine-dominated forests are initially a carbon source for 6
years after wildfires because decomposition exceeds primary productivity, then shift to a strong
carbon sink until stands matured and primary productivity declined.
The patterns described in the earlier text are very common for systems where disturbances
are stand-replacing and may also vary with stand structure (Figure 22.17; Kashian et al. 2006).
Stand density is important because it affects the rate of carbon uptake and storage by vegetation.
Because they have higher leaf area and photosynthate accumulates more quickly, dense stands
should accumulate more carbon quicker to form a strong sink after the disturbance. Sparse stands
should take longer to shift to positive NEP and achieve a weaker sink, but the duration of the sink
is likely to be prolonged compared to dense stands (Kashian et al. 2006). Notably, Kashian et al.
(2013) found that stand density was important for carbon storage in forests in Yellowstone National
Park that had not yet reached canopy closure, but that tree size and age were most important after
this point. Canopy closure represents the point at which stands reach their maximum leaf area,
and thus changes in leaf area may be a critical driver of NEP after disturbances.
Given the heterogeneity of both stand age and structure across landscapes, estimates of NEP
at broad spatial scales are extremely useful because they account for the variability that directly
affects carbon storage. Landscapes dominated by young or sparse forests may be carbon sources
while those dominated by older or dense forests may be carbon sinks, or any combination therein.
In addition to the influence of physiography as described by Running et al. (1989), species compo-
sition also imposes spatial heterogeneity in carbon dynamics across landscapes. For example,
boreal forests with mixed conifer and hardwood composition were found to store 47% more carbon
in the aboveground vegetation compared to black spruce forests and 44% more than jack pine for-
ests (Martin et al. 2005). In part, mixed stands were able to accumulate carbon more quickly
because they contain species with contrasting shade tolerances and thus are able to form multilay-
ered canopies, supporting a greater leaf area. Landscape carbon studies able to incorporate all
factors that are spatially heterogeneous are very difficult and remain elusive, but clearly accounting
for such heterogeneity is critical in understanding carbon storage at broad spatial scales.
150
Dense
100
ΔC Vegetation
(g C m-2 yr-1)
50 Sparse
0
100
Dense
50
(g C m-2 yr-1)
Sparse
Total NEP
–50
–100
0 50 100 150 200 250

Age since fire
F I G U R E 2 2 .1 7 Carbon balance following disturbances for lodgepole pine forests of contrasting structure
in the central Rocky Mountains. Carbon accumulates quickly and to a higher maximum for a short duration
in dense stands compared to sparse stands, which are slower to accumulate carbon. These differences in
vegetation drive similar changes in net ecosystem productivity (NEP). Source: Kashian et al. (2006).
Incorporating the heterogeneity present on landscapes into our understanding of carbon

dynamics has implications for predicting how forested regions may contribute to atmospheric car-
bon and climate change. Any disturbance to a forest will release carbon to the atmosphere if it kills
trees and will therefore temporarily shift the landscape to a carbon source. The amount of carbon
released, and the strength of the source, will vary with the type and severity of the disturbance and
the heterogeneity of these factors that may occur across a landscape. For example, stand-replacing
wildfires will create a strong and immediate carbon source due to combustion and widespread tree
mortality and subsequent decomposition, but less-severe surface fires may create only a weak
source if relatively few trees are killed. Likewise, moderate insect outbreaks or windstorms will
create a weaker, more prolonged carbon source because a higher proportion of trees survive the
disturbance, and variation in their severity may affect the strength or duration of the source.
NUTRIENT DYNAMICS
Although it is clear that landscape heterogeneity in nutrient pools and fluxes is the rule rather than
the exception for most systems, fewer researchers have been successful in both quantifying and
explaining such patterns (Turner and Gardner 2015). Much of this work in temperate forests has
focused on the nitrogen cycle, given that nitrogen is often the limiting nutrient for forest produc-
tivity (Chapter 19). A series of studies by Zak et al. (1986, 1989) and Zak and Pregitzer (1990)
related the spatial variation in nitrogen-related processes to the patterning of landscape ecosys-
tems in northern Lower Michigan. In comparing nine ecosystems located on various landforms
and soil textures, net mineralization was lowest in a dry oak ecosystem and highest in a mesic
northern hardwood ecosystem (Table 22.1). Nitrification was also highest in the northern hard-
wood ecosystem and far higher than in any other ecosystem sampled. Overstory biomass was
asymptotically related to differences in mineralization, but was unrelated to nitrification rates.
Thus, the spatial heterogeneity in at least some nutrient dynamics closely tracks the spatial distri-
bution of ecosystems across the landscape (Zak and Pregitzer 1990).
Table 22.1 Association of mineralization and nitrification rates with landscape ecosystems
in northern Michigan.
Percent coarse Mineralization Nitrification

Dominant overstory species Landform
and medium sand (μg N/g) (μg N/g)
Black and white oak Outwash plain 72 52.0 0.9
Black and white oak Outwash plain 70 49.6 0.3
Black, white, and red oak Kame 67 47.2 0.2
and moraine
Red, white, and black oak Kame 63 60.2 0.2
and moraine
Red and white oak Kame 55 63.0 0.6
and moraine
Red oak and red maple Kame 41 81.8 0.2
and moraine
Sugar maple and red oak Moraine 61 75.5 4.1
Sugar maple and red oak Moraine 56 93.4 1.2
Sugar maple, basswood, Moraine 55 127.8 45.5
and white ash
Source: Zak et al. (1989) / with permission of Canadian Science Publishing.
At broad spatial scales, spatial variability in some nitrogen transformations may be explained
by the heterogeneity of soil texture. For example, Reich et al. (1997) found a strong influence of soils
on mineralization for forests in Wisconsin and Minnesota, with higher rates of mineralization on
finer-textured soils. Likewise, Groffman and Tiedje (1989) and Groffman et al. (1992) examined
rates of denitrification, which tends to occur on wet sites in the absence of oxygen, in forest soils of
southern Michigan. Rates of denitrification across the landscape varied with soil texture and associ-
ated drainage, and were differentially responsible for regional estimates of denitrification. Loamy
soils represented 47% of the forest but 73% of the denitrification, and clay soils represented 9% of the
forest but 22% of the denitrification. Meanwhile, sandy soils represented 44% of the forest but only
5% of the denitrification. Together, these studies are examples of ecosystem processes whose hetero-
geneity may be predicted by the heterogeneity of physical site factors across the landscape.
Heterogeneity of nutrient dynamics at broad spatial scales appears to be most easily related
to physical site factors, but a general framework that associates nutrient dynamics with more gen-
eral characteristics of landscape pattern has been more fleeting. However, an experimental study
of forest fragmentation in Kansas may have begun to uncover a relationship between patch size
and soil nitrogen dynamics (Billings and Gaydess 2008). The study was designed with the assump-
tion that edge effects in fragmented forests result in changes in dominant tree species that in turn
may affect soil nutrient cycling. The experimental site contained small (4 × 8 m2) and large
(50 × 100 m2) patches of successional forest that were established in 1984. Large patches had greater
rates of woody plant colonization, higher plant species richness, and likely higher aboveground
net primary productivity (ANPP) per unit area than smaller patches, all of which became evident
after about 15 years of succession in the experimental fragments. These differences in vegetation
apparently resulted in greater rates of mineralization and nitrification in soils of the small patches
compared to the large patches, associated with greater fine root biomass and root nitrogen
concentration. Soils from small patches also had higher potential denitrification rates compared to
large patches, which is often associated with faster nitrogen cycling rates (Venterea et al. 2003).
Although these patterns have yet to be rigorously tested in other landscapes, the potential that
forest patch size—or landscape configuration in general—may govern the spatial heterogeneity in
nutrient cycling is intriguing.
Much like the landscape ecosystem approach we reference throughout this book, landscape
ecology focuses on the importance of scale, is concerned with the effects on a specific place on a
landscape by the ecosystems that surround it, and emphasizes physical site factors to some degree.
The landscape ecosystem approach involves mostly a focus on natural systems, with perhaps
some acknowledgement that humans are a significant disturbance to those systems. Additively,
landscape ecology explicitly includes humans as an important variable in understanding spatial
heterogeneity, including how their perceptions, values, and decisions shape landscapes. We have
not addressed this human dimension in our brief overview of forest landscape ecology, but we do
so in detail in our treatment of sustainability in the chapter that follows.
SUGGESTED
READINGS
Fahrig, L. (2003). Effects of habitat fragmentation on Romme, W.H. (1982). Fire and landscape diversity in
biodiversity. Annu. Rev. Ecol. Evol. Syst. 34: 487–515. subalpine forests of Yellowstone National Park. Ecol.
Forman, R.T.T. (1995). Land Mosaics, the Ecology of Monogr. 52: 199–221.
Landscapes and Regions. Cambridge University Turner, M.G. (1989). Landscape ecology: the effect
Press 632 pp. of pattern on process. Annu. Rev. Ecol. Evol. Syst.
Foster, D.R. and Boose, E.R. (1992). Patterns of forest 20: 171–197.
damage resulting from catastrophic wind in central Turner, M.G. and Gardner, R.H. (2015). Landscape
New England, USA. J. Ecol. 80: 79–98. Ecology in Theory and Practice, 2e. New York:
Landres, P.B., Morgan, P., and Swanson, F.J. (1999). Springer 502 pp.
Overview of the use of natural variability con- With, K.A. (2019). Essentials of Landscape Ecology.
cepts in managing ecological systems. Ecol. Appl. 9: Oxford, UK: Oxford University Press 656 pp.
1179–1188.
Sustainability of Forest
CHAPTER 23
Ecosystems
A t many instances in this book, we have suggested that humans have had far-reaching
influences on forests, both directly and indirectly. Their influences have been discussed
throughout the first 22 chapters of this book, including impacts on the overriding effects of
climate (Chapter 20); extensive alteration of disturbance regimes (Chapter 16); intentional and
accidental introduction of non-native plants, animals, and pathogens that eventually became
invasive (Chapter 21); widespread reductions in biological and ecological diversity (Chapter 14);
and fragmentation of forest landscapes (Chapter 22), among many others. Forest ecology does
not exist in a vacuum, in that every forest ecosystem on Earth has in some way been influenced
by humans. Although we struggle to understand how the natural world works, we cannot com-
pletely study forest ecosystems independent of human influence, especially when we attempt to
consider the future. This final chapter grapples with the concept of sustainability: its meaning, its
role in ecology, and its applications to forest ecology.
CONCEPTS OF SUSTAINABILITY
Without question, humans are part of ecosystems and are not external to them. Although some-
times difficult to imagine because of their wide-ranging influence, humans are supported by the
physical site factors of landscape ecosystems as much as any other organism. At the same time,
however, humans have a disproportionate influence on the ecosphere and have profoundly
changed regional and local ecosystems in both dramatic and subtle ways. Perlin (1989) provides a
remarkably powerful historical record of human destruction of forests that accompanied the
development of civilizations from the Bronze Age in Mesopotamia to nineteenth-century
America. In several papers, Kay (1994) speculated that the role of indigenous people is probably
underestimated in its importance of changing the biota in western North America. He concluded
that the modern concept of wilderness as areas without human influence is a myth because indig-
enous people changed the composition of entire plant and animal communities by limiting ungu-
late numbers and purposefully modifying the vegetation with fire. Of course, the effects of
European colonizers in North America have included forests cleared for agriculture, logging, live-
stock grazing, long-term changes in dominant tree species, forest fragmentation, climate change,
introduction of invasive species, and changes in disturbance regimes. The piercing story of human
interventions in the complexity of ecosystems is compellingly described for the Blue Mountains
of Oregon by Langston (1995).
THE PREVALENCE OF HUMAN VALUES IN FOREST ECOLOGY

The widespread decline of ecosystems has led to a great interest in the idea of sustainability since
the 1970s (Du Pisani 2006). A general notion of sustainability is that current desirable conditions of
an object or process—be they species composition, structure, recreational opportunities, a perception
639
640 Chapter 23 Sustainability of Forest Ecosystems
of wildness, etc.—continue to exist for the benefit of future generations. In most cases, sustainability
considers the maintenance into the future of something currently present. For example, Chapin
et al. (1996) defined a sustainable ecosystem as one that “over the normal cycle of disturbance
events maintains its characteristic diversity of major functional groups, productivity, soil fertility,
and rates of biogeochemical cycling.” Similarly, Turner et al. (2013) defined sustainability as “use of
the environment and resources to meet current needs without compromising the ability of a system
to provide for future generations.” One definition is strictly ecological and the other incorporates
social and economic structures, but both definitions are sufficiently broad and ambiguous to place
human values on real units of nature. Because humans differ in the way they perceive the world,
placing their values onto specific components of landscape ecosystems inevitably results in differ-
ences of opinion.
In the first edition of this textbook, Stephen H. Spurr elegantly described how the ecosystems
we see today are but a snapshot in time and space, and the values humans place on them may also
be temporary. In scrutinizing the concepts of native and introduced species, he reasoned
(Spurr 1964, p. 157):
. . . there is no meaning to the concepts of native and introduced species from a biocentric viewpoint. Char-
acterizing a plant or animal as being exotic or endemic characterizes it only from the standpoint of man’s
relationship to it. Actually, all plants and all animals are introduced or exotic from a biocentric standpoint
except at the very point in space where the particular gene combination was first put together. Whether the
subsequent migratory pattern of that organism took place independent of man or with the help of man is
important to man but not to the wilderness ecosystem itself.
It is an interest of man to know whether he carried the coconut to a given tropical island or not. To the
coconut it is of little importance as to whether it floated by itself in an ocean current or was lodged in the
hull of a native dugout canoe which in turn floated on the ocean current. To a maple growing in given spot
it is immaterial whether its seed flew there on its own wings or whether it was aided and abetted by the
wings of an airplane. A wild cherry is unaffected by concern as to whether its seed was deposited by a sea-
gull who spotted a target below or by a human who brought it thither in a paper bag . . . In short, from the
viewpoint of the forest, there is no distinction between native and introduced species . . . All were
migrants there.
Given what is now known about invasive species (Chapter 21), imagine how most natural-
ists and ecologists would react if such a sentiment was written as fact in a modern textbook!
Indeed, the passage was heavily edited for the second edition (Spurr and Barnes 1973) and elimi-
nated by the third (Spurr and Barnes 1980), primarily because the values of ecologists (and soci-
ety) began to change. Ecological data from the last 58 years suggest that introduced species can be
bad because those that become invasive may outcompete native species, alter nutrient dynamics
and/or soil communities, have negative economic implications, or may even lead to wholesale
changes in a system. And yet relatively few of these data directly contradict Spurr’s initial senti-
ments; instead, it is implicit in modern ecological studies that humans value native species over
introduced species, find the nutrient dynamics of a near-natural ecosystem to be more important
than of a heavily altered ecosystem, and abhor disruptions to a well-functioning economy. Data
over the last six decades have also gradually moved toward acknowledging the importance and
inevitability of change in ecological systems (Chapter 17), the major point that Spurr was at-
tempting to make. Spurr might be amused today to see that we have finally accepted ecological
change—as long as systems change as we think they should. Human values are typically not
scientific, yet they are often overlain onto scientific inquiry, including that undertaken by forest
ecologists. These values are what challenge scientists in incorporating notions of sustainability
into scientific understanding.
Concepts of Sustainability 641
HISTORICAL PERSPECTIVE OF SUSTAINABILITY IN FORESTS

Ours is not a textbook about the management of forests, but a brief discussion here is of use to
illustrate how human values are an intrinsic part of sustainability. In fact, if sustainability’s root
concern is providing for future generations, then those concerned with forests may be among the
first who explicitly integrated sustainability into their operating models (Westoby 1989), even as
those models evolved and changed over decades and centuries. First and foremost, the concept of
sustaining natural resources over the long term was present in the traditions of many ancient cul-
tures, such as those of the indigenous peoples of New Zealand, Indonesia, Mali, India, and British
Columbia (Gadgil and Berkes 1991). Regarding modern times, a review by Wiersum (1995) notes
that sustainability was well expressed in forest management systems in Germany in the eighteenth
century, and the philosophy of wise use consistent with future availability of forest resources was
well articulated in the 1800s. Notably, the creation of the forestry profession in the United States
resulted from concerns about sustainable use of forests in the 1900s. Sustainability in some form
therefore has a very long association with forests and forest ecology (Franklin et al. 2018).
In North America, one aspect of timber management has been deeply associated with at least
one form of sustainability for over a century: the idea of sustained yield. Sustained yield is a Western
concept that because harvested trees can be regrown, forests are a renewable resource and can be
managed for a continuous flow of resources for future generations. Notably, the term “resource” itself
automatically invokes human values, both social and economic. Two general problems, among many,
are inherent with the idea of sustained yield. First, one cannot assume that all forests are equally
renewable, and as such their yields are not always sustainable. For example, aspen in the Lake States
is typically clear-cut for pulpwood, sometimes as often as every 20–30 years, because the trees regener-
ate quickly from root sprouts after cutting and exhibit growth rapid enough to replace themselves
within the time frame of our current social and economic systems. Likewise, Douglas-fir forests of the
Pacific Northwest are also considered quite renewable after harvesting because they may produce saw-
logs within a 50-year time frame on suitable sites, again within an appropriate social and economic
time frame. These models place the emphasis on the potential product (in this case timber) as the
resource of interest, such that clear-cutting a 500-year-old (“old-growth”) Douglas-fir forest and plac-
ing it into a rotation of harvesting every 50 years are appropriate to achieve a sustained yield.
A second problem with sustained yield is its focus on a single resource provided by the forest.
Using the abovementioned example, outside of the timber resource, transitioning an old-growth
Douglas-fir forest to one that produces 50-year-old sawlogs does not consider the potential value of
the old growth itself, which would not be considered to be renewable. The old-growth forest may
provide a limited habitat for certain species (both plants and animals), or it may provide recreational
opportunities or aesthetic values, or it may create an important protection for water resources not
found in younger forests. Forest management in North America broadened to multiple-use forest
management in the 1960s and 1970s in recognition of the idea that sustainability might also target
resources other than timber (Clawson 1978). It also more firmly emphasized the importance of
human values in its treatment of sustainability. Multiple-use management continued to include
sustained yield as one of its multiple uses, but transitioned again in the 1990s to ecosystem
management (Grumbine 1994; Brussard et al. 1998). Ecosystem management suffered somewhat
from varying definitions and a somewhat ambiguous framework, but generally considered the
balance between maintenance of biodiversity and ecosystem processes (including disturbances
and the changes they invoked in ecosystems) and appropriate human uses. Its emphasis on “whole
systems” included social and economic factors as well as ecological ones and hoped to integrate
ecological conservation into socioeconomic models. Social factors, in particular, included concepts
crucial to sustainability, such as freedom, justice, and equity. Ecosystem management therefore
represented a further emphasis of the human factor into sustainability, and many of its basic tenets
are recognizable today in concepts such as ecological forestry (Chapter 14).
ECOSYSTEM SERVICES
Central to the concept of sustainability is a clear definition of what exactly is being sustained.
Simply put, ecosystem services are the objects and processes of natural ecosystems that are
beneficial to humans, and they are typically the implicit or explicit targets of sustainability. The
idea of ecosystem services was described by Daily et al. (2000), codified by the United Nations’
Millennium Ecosystem Assessment (MEA 2005), and has since been discussed by many authors
(e.g., Nesbitt et al. 2017; Alix-Garcia and Wolff 2014; Levin 2012; Perrings 2007; and many others).
Notably, the focus of ecosystem services by design is human well-being; the ecosystems supplying
the services benefit from protection or conservation only because of what they provide to humans
rather than from any intrinsic value they have. It is often argued, however, that protecting ecosys-
tem services contributes to the overall function of the ecosphere and in turn the regional and local
ecosystems within it. Indeed, it has been recognized that many ecosystem services are irreplace-
able, and their loss would somehow harm the integrity of ecosystems, regardless of direct human
benefit (Ekins et al. 2003; Farley 2012; Neumayer 2013). Thus, ecosystem services are considered
to be a vital bridge between ecosystems and society, and they justify the actions needed to undergo
sustainability (Wu 2013, and references therein).
Ecosystem services are often grouped into four broad categories (MEA 2005; Carpenter
et al. 2009): provisioning, regulating, supporting, and cultural. Provisioning services, originally
defined as “ecosystem goods” by Daily (1997), include the potential products extractable from an
ecosystem such as timber, food, water, or minerals, but also genetic and medicinal resources. Regu-
lating services are less tangible than provisioning services because they describe more abstract
benefits humans obtain when they maintain important ecosystems processes. Examples of regulat-
ing services might include pollination, which is important for fruit production and the persistence
of plant species, or trophic dynamics, which are important for control of pests and diseases. Sup-
porting services are those that are vital for the functioning of ecosystems and are necessary for most
other ecosystem services to exist, such as soil formation, nutrient cycling, or oxygen production.
Turner et al. (2013) also treat primary productivity and post-disturbance forest regeneration as
major supporting services on specific landscapes. Supporting services are sometimes difficult to dis-
tinguish from regulating services because many regulating services may be supportive at large
spatial or temporal scales. Finally, cultural services are non-material services obtained from ecosys-
tems such as spiritual enrichment, aesthetic encounters, reflection, or recreation. A detailed
treatment of ecosystem services in forests is provided by Solórzano and Páez-Acosta (2009); a trun-
cated list of examples is shown in Table 23.1.
If ecosystem services are a main target of sustainability, then maximizing those services is
typically its goal. However, not all ecosystem services of a forest are able to be maximized at the
same time, and trade-offs ensue. Perhaps the best example of trade-offs in ecosystem services
would be plantations or tree farms under a sustained yield scenario, where the provisioning service
of timber would be maximized at the expense of biodiversity, wildlife habitat, or aesthetics. Like-
wise, wilderness areas might maximize cultural services such as recreation, spiritual uses, or aes-
thetics at the expense of provisioning services when potential resources are not extracted
(Kinzig 2009). The point is that most services require their own unique set of ecological attributes
such that it is difficult to maximize all or even most simultaneously.
Interestingly, many ecologists have argued that biodiversity preservation is the keystone of
sustainability, because it augments the maintenance of most other ecosystem services through its
heavy influence on ecosystem function (e.g., Cardinale et al. 2012). For example, in one of the ear-
lier explorations of sustainability in forests, Lindenmayer and Franklin (2003) edited a book with
chapters written by 13 forest ecologists and managers from around the world, many of whom
defined sustainability as the preservation of forest biodiversity and one who defined it as “the
Table 23.1 Non-exhaustive list of examples from forest ecosystems of the four categories of ecosystem
services.
Type of ecosystem
Ecosystem good or function Examples from forest ecosystems
service
Provisioning Forest products Timber, fiber
Food Fish, fruits, graze or browse for
livestock, wildlife
Genetic resources Test and assay organisms, bioprospecting
Biochemical/ Drugs and pharmaceuticals, dyes, latex, and
medicinal resources other resins
Ornamental resources Diversity of organisms with ornamental potential
Regulating Climate regulation Altering temperature, albedo, evapotranspiration
Hydrological regulation Regulating volume, quality, and timing of
water flow
Disturbance moderation Dampening of wind disturbance, flood control
Carbon sequestration Cycling and storage of carbon
and storage
Pollination Pollination of timber/fiber species, fruit
production
Trophic dynamics Pest or disease control, reduction of
herbivory on crops
Supporting Soil formation and support Physical support for plants, nutrient
cycling, water
Post-fire forest regeneration Maintenance of cover types and habitats
Site productivity Framework for species composition,
soil processes
Cultural Recreation activities Tourism, camping, hunting, hiking
Spiritual inspiration Use of forests for religious or individual
experiences
Aesthetics Appreciation of scenery
Cultural uses Background for books, music, film, folklore
grand attempt at forest biodiversity rehabilitation.” Although biodiversity is certainly an important

attribute of a sustainable forest, its increase may necessarily be accompanied by trade-offs with
some other services such as food or timber production and may have less or nothing to do with
other services such as carbon storage or erosion control (Kinzig 2009; Reyers et al. 2012). Biodiver-
sity appears to typically be most at odds with provisioning services. Naeem (2009) suggested that
the ecosystem functions most closely related to biodiversity are generally supporting services, such
as net primary production, nutrient cycling, and other biogeochemical processes, but that these
services are not consumed and are least understood by people. Others have argued that biodiver-
sity contributes more to regulating services such as invasion resistance and ecosystem resilience,
as well as cultural services (Mace et al. 2012). In any case, some have argued that biodiversity is
best conserved when ecosystem services are the target rather than biodiversity itself, because eco-
system services provide the economic and social justification that biodiversity has an instrumental
value (Skroch and López-Hoffman 2009; Mace et al. 2012; Balvanera et al. 2014).
TOWARD A DEFINITION OF SUSTAINABILITY

If sustainability is built upon a balance between ecological factors and human well-being, and
ecosystems services are its target, then how exactly might we define the term? Unfortunately,
inconsistencies in its definition have plagued an already complex concept while its study has
increased dramatically over the last three decades. Marshall and Toffel (2005) tabulated at least 100
different definitions of the term. As a matter of comparison, we present three such definitions as
the most common and familiar to those who think about sustainability.
Sustainability has often been described as a “three-legged stool,” with equal-sized environ-
mental (integrity of ecosystems), social (justice, equity, and freedom), and economic (material
goods and services) legs. In this model, also called the “triple bottom line” or the “people–planet–
profit” model (Elkington 2004), all three aspects must be met simultaneously; an emphasis on any
one more so than the others will not achieve sustainability. These three components have been
described as separate but intersecting, best depicted with a Venn diagram where sustainability
represents the limited area where all three components intersect, representing their achievement
at the same time (Figure 23.1a). In natural resource management, the idea is to properly account
for all three components and the interactions among them. The triple bottom line is increasingly
used as a way to plan for sustainability and sustainable development in the business world, with
the model often used as a metric of a corporation’s commitment to sustainability.
If we accept that the triple bottom line describes a basic model of sustainability, then the
likelihood that all three components are not equally emphasized must be considered. If the focus is
on the entire system rather than any of its separate components, achievable sustainability is consid-
ered to be weak. Weak sustainability allows one component to substitute for another in the
triple-bottom-line model, as long as the entire system increases or sustains its total capital
(Figure 23.1b). Further, the model assumes there is no difference between natural capital (ecosys-
tems or environment) and human-made capital when providing for the well-being of humans, and
that natural capital lacks unique properties that cannot be substituted by human-made capital.
Thus, weak sustainability could be achieved on a heavily logged forest landscape if the harvesting
supported economic growth of the region and an appropriate standard of living for its people,
because the social and economic spheres of the model would substitute for the environment. Wu
(2013) contrasts weak sustainability with absurdly strong sustainability, where the environmental
(a) Triple Bottom Line (b) Weak Sustainability (c) Strong Sustainability
Economy
Economy
So
y
Sustain-
om
cie
ability Society
on
ty
Society
Ec
Environment
Environment
Environment
F I G U R E 2 3 .1 Three models relevant to sustainability: (a) the triple-bottom-line model outlines the basis of
sustainability as the intersections of the three spheres of environment, economy, and society, assuming that
all three must be met simultaneously to achieve sustainability; (b) weak sustainability assumes that any of
the three components is substitutable for another, so long as the entire system increases or maintains its
capital; (c) strong sustainability assumes that the three components are nested within each other such that
environment constrains society and economy, and society constrains economy. Source: Wu (2013) /
Springer Nature.
sphere is never substituted even when it could be or perhaps should be. Under such a model, species
and ecosystems are protected above all else, even at the cost of human well-being (e.g., never clear-
ing another forest for agriculture or cutting trees for fuel even if it means people would starve or go
without heat in a local village). Thus, weak sustainability assumes that ecosystems and economic
products or resources are equivalent in value and therefore may substitute for one another, while
absurdly strong sustainability assumes they are polar opposites. A model of sustainability consid-
ered strong therefore lies somewhere between these two ends of the sustainability spectrum. Rather
than assuming any of the three components are substitutable by the others as in weak sustainability,
strong sustainability assumes that each of the three components compliments the others, because
natural capital may not necessarily have an analogous human-made substitute. Rather than viewed
separate from each other, strong sustainability assumes that the three components are viewed hier-
archically, with economy nested within society and society nested within the environment
(Figure 23.1c). In this model, social and economic activities exist within the context of fundamental
ecological processes such as nutrient cycling, climate regulation, and water purification, and
economy functions within the context of a specific sociological regime (Franklin et al. 2018). Strong
sustainability thus firmly places the ecological/environmental aspect at the center of the paradigm.
An example of the differences between weak, strong, and absurdly strong sustainability can
be illustrated using the susceptibility of western landscapes to fire. The Colorado Front Range is an
area undergoing rapid economic development and extensive exurban sprawl into forested,
fire-prone areas. This expansion of the wildland–urban interface (WUI) poses great problems for
federal agencies responsible for protecting private landowners from wildfires burning on adjacent
National Forests. Moreover, fire suppression in the twentieth century in the area has made the
ponderosa pine and mixed-conifer forests overly dense and particularly susceptible to large, severe
fires with the potential to heavily impact lives, property, and the landscape itself. This context has
led to a great many discussions about the sustainability of continued development and the expan-
sion of the WUI, and arguments have varied precisely based on the type of sustainability model
under consideration. Using a weak sustainability model, the continued economic growth and pro-
vision of property might substitute for the ecological integrity of the landscape, such that appropri-
ate management would include substantial fire suppression in the wake of further expansion of
the WUI. On the other hand, an absurdly strong sustainability model might argue that the environ-
ment is not substitutable regardless of the consequences for humans, such that any fires that burn
on the landscape should be allowed to burn without care for lives or property as a means of allow-
ing altered forests to heal themselves. A strong sustainability model would probably propose a
carefully planned development strategy that minimized environmental impact and mitigated risk
to private landowners while allowing the natural disturbance regime to persist. Although it would
take region-wide cooperation and collaboration, extensive regulation and legislation, and corpo-
rate support to achieve strong sustainable development in this manner, it is clear that a strong
sustainability model is the most appropriate.
Different sustainability models will often lead to different landscapes that vary in the ecosys-
tem services they provide (Wu 2013; Figure 23.2). Wilderness areas, for example, that emphasize
the preservation of biodiversity and natural ecosystems provide high natural capital and are most
appropriately viewed through an absurdly strong sustainability model. Regulating ecosystem ser-
vices would dominate in such a landscape, and provisioning services would effectively be absent.
At the other end of the spectrum, urban landscapes would constitute a weak sustainability model
most closely, having low natural capital and providing mostly cultural and regulating services.
Agricultural and semi-natural landscapes might follow more of a strong sustainability model
because they clearly treat social and economic factors within the context of ecological factors.
Agricultural landscapes are generally more altered and thus have less natural capital than
semi-natural landscapes with a higher proportion of provisioning services, but neither landscape
fully substitutes natural capital with human-made capital.
Relative Importance or Dominance

Sustainability Absurdly strong sustainability Strong sustainability Weak sustainability
View
Environmental Nature conservation/preservation Ecological restoration Ecological engineering
Emphasis
Ecosystem
Services
Supporting Regulating Provisioning Cultural
Capital Form Built, Human & Social Capitals

Natural Capital
Ecosystem or
Landscape Type
Natural Semi-natural Agricultural Urban
F I G U R E 2 3 .2 Varying landscapes and their placement on spectra of sustainability views/models,

proportion of ecosystem services categories, and relative amount of natural capital. Source: Wu (2013) /
Springer Nature.
The idea that different landscapes follow different sustainability models and provide different
ecosystem services suggests that scale and spatial heterogeneity have strong implications for this con-
cept. A mosaic of land use types is always present on modern landscapes, such that achieving strong
sustainability at broad scales may require targets of weak sustainability (for example, in urban areas)
or absurdly strong sustainability (for example, in wilderness areas) at smaller spatial scales (Wu 2013).
Moreover, if landscape pattern influences ecological processes (Chapter 22) and biodiversity, then it
likely also influences the provision of ecosystem services, making them patchy across space (and
potentially across time). If true, then some patterns are more likely to favor the maintenance of eco-
system services, and thus some patterns are more likely to lead to sustainable landscapes (Wu 2013).
Notably, the patchiness, or spatial heterogeneity, of some ecosystem services may be impor-
tant for sustaining them. In a comparison of two forested landscapes in the western United States
(the Greater Yellowstone Ecosystem, a subalpine landscape of the central Rocky Mountains; and
the coastal temperate rainforest of the Pacific Northwest), Turner et al. (2013) found that
post-disturbance spatial heterogeneity was important for sustaining the ecosystem services of forest
regeneration, primary production, carbon storage, natural hazard regulation, insect and pathogen
regulation, timber production, and wildlife habitat. Spatial heterogeneity helps to sustain the sup-
porting service of forest regeneration by determining the location of surviving pre-disturbance trees
that will provide a seed source for regeneration, and this supporting service will in turn also affect
other services such as primary production and carbon storage. Regulation of forest insects and path-
ogens is influenced by heterogeneity because it determines the spatial patterns of stands of suscep-
tible age and size for insect or pathogen attack, as well as how their placement on a landscape might
dampen or enhance an outbreak. The provisioning service of timber is clearly affected by site
heterogeneity as determined by physiography, soils, and microclimate (Turner et al. 2013).
WHERE DO WE GO FROM HERE?

It is not arguable that we are in an age of ecological crisis. Nearly 20 years ago, Franklin (2003)
noted that humans have extensively altered the physical and ecological contexts of forests to the
point that they no longer resemble those from which the forests evolved. These include altered
Where Do We Go from Here? 647
isturbance regimes, especially fire; changes in global and regional climate and chemical regimes
d
(i.e., climate change and acid rain); introduction of invasive species; and fragmentation of land-
scapes that have changed the amount and spatial arrangement of various forest conditions. Each
of these topics has been covered in detail in other chapters of this book (Chapters 10 and 20–22)
and will not be revisited here, but their increased pervasiveness in modern forests has warranted
each a detailed treatment that was not presented in this book’s last edition. Franklin (2003) pre-
dicted that these four problems would intensify in the twenty-first century, and it seems as though
he was correct.
So how do we ensure sustainability in forests? In as much as our living within an ecologi-
cal crisis, this is a question still of great debate whose answer is not an easy or unambiguous one,
but one that is clearly related to many of the concepts in this book. In a sense, sustainability of
forests is subject to the “land sparing versus land sharing” debate (Phalan et al. 2011). One could
simply keep humans out of at least some forests and allow the forests to function without human
interference. This land-sparing argument suggests that intensively managed planation forests
(sometimes called “fiber farms,” usually in temperate regions) be used to supply necessary wood
products, such that native forests could be left unaltered by humans and thus sustained (e.g.,
Bremer and Farley 2010; Paquette and Messier 2010). The adoption of fiber farms serves as a
kind of ecological offset that allows larger reserve areas elsewhere on a landscape (Green
et al. 2005; Fischer et al. 2008). This type of forest allocation is viewed by some as the easiest and
most efficient movement toward forest sustainability because often-opposing objectives of stake-
holders will conflict less often, such that substitution of human-made capital for natural capital
will be less common. However, many forest managers counter that such a physical separation of
forests into provisioning areas and preserves is likely to result in a failure of forests to fulfill the
expectations and needs of society as well as undesirable outcomes for ecosystem function and
biodiversity (Franklin 2003). These failures are likely because many forest landscapes have
already been heavily altered by humans if only indirectly, and thus there are few or no candidate
remnants suitable for preservation. Moreover, the lofty expectations humans have for forests in
addition to wood products—such as biodiversity preservation and watershed protection—are
probably not attainable without active management. Critics of land-sparing often follow the
land-sharing model, which integrates different goals (such as biodiversity and/or ecosystem ser-
vices and resource extraction) in the same area. Proponents of land-sharing in forests are pri-
marily those who favor ecological forestry (Puettmann et al. 2008; Lindenmayer et al. 2012;
Franklin et al. 2018), discussed in Chapter 14, where complex silvicultural systems (harvesting
and regeneration) are developed that address more than simply the need for wood products. Of
course, critics of land sharing counter that a focus on multiple goals simultaneously is less effec-
tive than a focus on either goal individually. In addition, decreased resource extraction that
would accompany ecological forestry would necessitate increased extraction elsewhere (Gabriel
et al. 2010).
There are certainly no easy answers for forest sustainability today, in part because sustain-
ability is achieved by following very different pathways depending on the landscape. Linden-
mayer and Cunningham (2013) summarized key principles for sustainable forests. First, there is
an important need for improved planning and monitoring to avoid overcommitment of natural
resources. To do so, managers and scientists must be able to identify the critical thresholds beyond
which resource use will begin to degrade ecosystem services and biodiversity (Carpenter
et al. 2011). Second, landscapes vary in their importance to ecosystem services and biodiversity,
and thus it is critical to identify those places with disproportionate contribution to species persis-
tence and ecological processes, such as breeding areas, refugia, or biodiversity hotspots. Research
begun immediately after major natural disturbance events is often useful in this regard, particu-
larly in understanding how refugia contribute to post-disturbance recovery (e.g., Turner et al.
2003a). Third, it is important to maintain connectivity, but only when connectivity is desirable for
specific processes. Given that connectivity could refer to connected habitat for a species or group
of species of interest or to connected ecological processes, and that some connectivity could lead
to the spread of undesirable disturbances or species (Lindenmayer and Fischer 2013), identifying
the type of connectivity important for sustainable forest management leaves much room for
future research. Fourth, land-use practices may accumulate over space and time in such a way
that threatens forest sustainability, and such land uses should be avoided. A classic example
would be the great eastern white pine forests logged at the end of the nineteenth century in
Michigan and other areas of the Lake States. Wholesale forest clearing and subsequent, repeated
slash fires converted a widespread dry-mesic forest landscape to one dominated by xeric species
such as jack pine and xeric oaks much more likely to burn. Finally, land uses typically fall upon a
gradient between land sharing and land sparing rather than either end of the spectrum and are
extremely context-dependent.
A final point about the prospects of sustainability draws on our original treatment of forest
ecology as a study of landscape ecosystems, which integrates the biotic and physical aspects of
whole, volumetric ecosystems. In his treatment of sustainable planning and design for large
landscapes, Bailey (2002) reasons that sustaining ecosystems must occur at large spatial scales,
and that local management must occur within a strong contextual knowledge of regional ecosys-
tems. The ecological crises we described in the earlier text have at least partially resulted from a
mismatch of human development and the limits of the regional ecosystems in which they occur.
The sense of place provided by regional ecosystems provides patterns and processes that place
bounds or limits on the social and economic components of the triple bottom line of that
particular place, and thus recognizing the sense of place is necessary to conserve resources
(Bailey 2002). Returning to our retrospective example of eastern white pine harvesting in
Michigan, a lack of understanding and recognition of the cold winter climate, relatively flat
topography, and dry, sandy soils that promote fires limited the level of post-harvesting forest
regeneration and recovery that might have been expected in that regional ecosystem. Unfortu-
nately, human activities exceeded the regional ecosystem’s ecological limits in a way that has
threatened its long-term sustainability.
EPILOG: EARTH AS A METAPHOR FOR LIFE

Thus far our discussion of sustainability has detailed the problems without suggesting solutions
(an action Theodore Roosevelt called “whining”). We have been able to discuss in detail what sus-
tainability is, but have been less able to describe how to achieve it in contemporary forests without
a detailed discussion of forest management. Again, the fact that sustainability is defined in part by
human values makes it extremely difficult—and undesirable—to simply stick to the science. As
such, we take the opportunity to end this chapter (and this book) with a final abstract, rather
philosophical consideration.
In thinking about sustainability, perhaps absurdly strong sustainability in particular, certain
things such as biodiversity or ecological integrity have intrinsic value, independent of their value
or contribution to human well-being. Rare or endangered species should be protected because they
have value, regardless of how humans may benefit from them. Wilderness areas are not just
valuable for the recreation, aesthetic, or spiritual values they provide humans, but also because
they are an end in themselves. Ecosystem services, the major target of sustainability, contrast with
this notion of intrinsic value because they are human-centered (Colyvan et al. 2009) and have
value only based on what they provide for humans (Reyers et al. 2012). Focusing on the ultimate
service that ecosystems provide humans is likely to oversimplify how ecosystems function
(Ghazoul 2007). In addition, humans need to adapt to ecosystem function because ecosystems lack
Where Do We Go from Here? 649
intentionality in serving humans (Salles 2011). Sustainability by definition is therefore a

human-centered construct, illustrated by the simple fact that we need to label ecosystems with
social or economic value to justify protecting them.
As described in the earlier text, humans have had an enormous, disproportionate
influence on the ecosphere and on ecosystems at every scale. We could easily venture into a
discussion of understanding humans’ place in the world. Are we part of ecosystems, or
external to them? Should we be stewards of ecosystems, regardless of whether or not we con-
sider ourselves as members? Many (not all) people interested in forest ecology are likely to
answer “yes” to both of these questions, but do humans deserve to be the center of focus?
Rowe (1998) described “Earth” rather than “organism” as a metaphor for “life”. A majority of
ecologists and non-scientists alike consider organisms to represent living things, but Rowe
argues that this is a reductionist view that turns the vantage point of humans inward toward
themselves instead of outward toward their surroundings. Throughout this book, we have
emphasized that organisms are but one part of ecosystems, and they cannot realistically be
separated from physical factors of the Earth (air, water, soil, and energy). Organisms are parts
of wholes, and it is the whole ecosystem that provides life to the organisms. In the past (and to
some degree the present), ecological research often tended to gravitate toward the charismatic
biota, either as the survival physiology of organisms or as the interactions between them. An
ecological viewpoint that places organisms at its center, only later turning to what is around
them, lends itself to a human-centered focus. When the whole ecosphere, rather than just the
biota, is considered to be life, we put spaces at the center of our ecological viewpoint and look
at its contents (Rowe 1998). Thus, humans become an integral part of ecosystems, but are not
central to them.
A viewpoint alternative to one that is human-centered is not new; many have articulated
the dangers of placing humans at the center of the world. In fact, many of the world’s ecological
problems have resulted almost exclusively from an anthropogenic view of the world. Aldo Leop-
old’s land ethic famously moved the focus from humans to the land, with the idea that land was
not a commodity owned by humans but a community that included them (Leopold 1949). Sig-
mund Freud (not an ecologist) suggested that humans make their biggest scientific strides when
they realize they are not the center of the universe. For example, physical science leapt forward
when Copernicus speculated that the Earth was not the center of the universe; biology advanced
greatly when Darwin suggested that humans actually descended from other animals; and psy-
chology developed quickly from Freud’s theory that a personality is driven in part by the uncon-
scious mind, out of a person’s control (Freud 1917). It is certainly true that an additional scientific
leap forward will be needed to find new solutions in this age of ecological crises, perhaps one
facilitated by encouraging humans to consider that they are a tenant rather than the steward of
ecosystems.
How does this philosophy translate to sustainability? It suggests that ecosystems have organ-
isms, not the reverse; thus, ecosystems have humans, not the reverse. Rowe (2000) uses an analogy
of the heart in a human body to describe the relationship of humans to ecosystems. The role of the
heart is to keep the body healthy, not to oversee the body. Likewise, perhaps the role of humans is to
keep ecosystems healthy, but as parts of those wholes rather than as their stewards. Transposing this
to ecosystem services, ecosystems do not serve humans; in fact, the reverse should be true. We have
no illusions to the fact that this view of the world is awkward and would be difficult if not impossi-
ble to instill within our current models of sustainability in forests. At the very least, however, we
return to the notion of understanding forest ecology, and of ecology in general, as the study of whole
ecosystems that include more than organisms (including humans) alone. If nothing else, that
resembles a good start.
SUGGESTED
READINGS
Bailey, R.G. (2002). Ecoregion-Based Design for Sustain- Mace, G.M., Norris, K., and Fitter, A.H. (2012). Biodiver-
ability. New York: Springer 223 pp. sity and ecosystem services: a multilayered relation-
Balvanera, P., Siddique, I., Dee, L. et al. (2014). Linking ship. Trends Ecol. Evol. 27: 19–26.
biodiversity and ecosystem services: current uncer- Turner, M.G., Donato, D.C., and Romme, W.H. (2013).
tainties and the necessary next steps. Bioscience Consequences of spatial heterogeneity for ecosystem
64: 49–57. services in changing forest landscapes: priorities for
Daily, G.C. (ed.) (1997). Nature’s Services: Societal future research. Landsc. Ecol. 28: 1081–1097.
Dependence on Natural Ecosystems. Washington, DC: Wu, J. (2013). Landscape sustainability science: eco-
Island Press 412 pp. system services and human well-being in changing
Lindenmayer, D.B. and Franklin, J.F. (ed.) (2003). landscapes. Landsc. Ecol. 28: 999–1023.
Towards Forest Sustainability. Washington, D.C:
Island Press 212 pp.
References
Abatzoglou, J.T. and Kolden, C.A. (2013). Relationships paleoecology and tree census data for eastern North
between climate and macroscale area burned in America. Ann. For. Sci. 76: 8.
the western United States. Int. J. Wildland Fire 22: Abrams, M.D. and Scott, M.L. (1989). Disturbance-mediated
1003–1020. accelerated succession in two Michigan forest
Abbey, E.P. (1968). Desert Solitaire. New York: McGraw-Hill types. For. Sci. 35: 42–49.
336 pp. Abrams, M.D., Sprugel, D.G., and Dickmann, D.I.
Abella, S.R. and Covington, W.W. (2006a). Vegetation– (1985). Multiple successional pathways on recently
environment relationships and ecological species disturbed jack pine sites in Michigan. For. Ecol.
groups of an Arizona Pinus ponderosa landscape, Manag. 10: 31–48.
USA. Plant Ecol. 185: 255–268. Abrams, M.D., Kubiske, M.E., and Mostoller, S.A.
Abella, S.R. and Covington, W.W. (2006b). Forest eco- (1994). Relating wet and dry year ecophysiology to
systems of an Arizona Pinus ponderosa landscape: leaf structure in contrasting temperate tree species.
multifactor classification and implications for eco- Ecology 75: 123–133.
logical restoration. J. Biogeogr. 33: 1368–1383. Adams, W.T. (1992). Gene dispersal within forest tree
Abella, S. and Shelburne, V.B. (2004). Ecological populations. In: Population Genetics of Forest Trees
species groups of South Carolina’s Jocassee Gorges, (ed. W.T. Adams, S.H. Strauss, D.L. Copes and A.R.
southern Appalachian Mountains. J. Torry Bot. Soc. Griffin), 217–240. Boston, MA: Kluwer.
131: 220–231. Adams J.M. (1997). Global land environments since the
Abella, S.R., Shelburne, V.B., and Macdonald, N.W. last interglacial. Oak Ridge National Laboratory,
(2003). Multifactor classification of forest TN, USA. https://www.esd.ornl.gov/projects/qen/
landscape ecosystems of Jocassee Gorges, adams1.html. Accessed 02-17-2021.
southern Appalachian Mountains, South Carolina. Adams, A.B., Dale, V.H., Kruckeberg, A.R., and Smith,
Can. J. For. Res. 33: 1933–1946. E. (1987). Plant survival, growth form and regen-
Aber, J.D. and Melillo, J.M. (1991). Terrestrial Ecosys- eration following the May 18, 1980, eruption of
tems. Philadelphia: Saunders 429 pp. Mount St. Helens, Washington. Northwest Sci. 61:
Abrams, M.D. (1990). Adaptations and responses to 160–170.
drought in Quercus species of North America. Tree Adams, M.A., Buckley, T.N., and Turnbull, T.L. (2020).
Physiol. 7: 227–238. Diminishing CO2-driven gains in water-use
Abrams, M.D. (1992). Fire and the development of oak efficiency of global forests. Nat. Clim. Chang. 10:
forests. Bioscience 42: 346–353. 466–471.
Abrams, M.D. (2001). Eastern white pine versatility in Adkins, J.K. and Rieske, L.K. (2015). A terrestrial
the presettlement forest. Bioscience 51: 967–979. invader threatens a benthic community: potential
Abrams, M.D. (2003). Where has all the white oak effects of hemlock woolly adelgid-induced loss of
gone? Bioscience 53: 927–939. eastern hemlock on invertebrate shredders in head-
Abrams, M.D. and Downs, J.A. (1990). Successional water streams. Biol. Invasions 17: 1163–1179.
replacement of old-growth white oak by Aerts, R. (2006). The freezer defrosting: global warming
mixed-mesophytic hardwoods in southwest and litter decomposition rates in cold biomes.
Pennsylvania. Can. J. For. Res. 20: 1864–1870. J. Ecol. 94: 713–724.
Abrams, M.D. and Nowacki, G.J. (2015). Exploring Agee, J.K. (1973). Prescribed fire effects on physical and
the early Anthropocene burning hypothesis and hydrologic properties of mixed-conifer forest floor
climate-fire anomalies for the eastern US. J. Sus- and soil. Report 143, Univ. California Resources
tain. For. 34: 30–48. Center, Davis, CA.
Abrams, M.D. and Nowacki, G.J. (2019). Global Agee, J.K. (1993). Fire Ecology of Pacific Northwest
change impacts on forest and fire dynamics using Forests. Washington D.C: Island Press 493 pp.
651
652 References
Agee, J.K. (1998). The landscape ecology of western Alexander, R.R. (1986). Classification of the forest veg-
forest fire regimes. Northwest Sci. 72: 24. etation of Wyoming. USDA. For. Serv. Res. Note
Agren, G.I., Axelsson, B., Flower-Ellis, J.G.K. et al. RM-466. Rocky Mountain For. and Rge. Exp. Sta.,
(1980). Annual carbon budget for a young Scots Fort Collins, CO. 10 pp.
pine. Ecological Bulletins 32: 307–313. Alexander, H.D. and Arthur, M.A. (2014). Increasing
Ahlgren, C. (1974a). Effects of fires on temperate red maple leaf litter alters decomposition rates and
forests: north central United States. In: Fire and nitrogen cycling in historically oak-dominated for-
Ecosystems (ed. T.T. Kozlowski and C.E. Ahlgren), ests of the eastern U.S. Ecosystems 17: 1371–1383.
46–72. New York: Academic Press. Alexander, R.R., Tackle, D., and Dahms, W. (1967). Site
Ahlgren, I.F. (1974b). The effect of fire on soil organ- indexes for lodgepole pine, with corrections for
isms. In: Fire and Ecosystems (ed. T.T. Kozlowski stand density; methodology. USDA For. Serv. Res.
and C.E. Ahlgren). New York: Academic Press. Paper RM-29. Rocky Mountain For. and Rge. Exp.
Ahuja, M.R. (2005). Polyploidy in gymnosperms: revis- Sta., Fort Collins, CO. 18 pp.
ited. Silvae Genet. 54: 59–69. Alexander, N.L., Flint, H.L., and Hammer, P.A. (1984).
Ahuja, M.R. and Libby, W.J. (ed.) (1993). Clonal For- Variation in cold hardiness of Fraxinus americana
estry. Vol. I, Genetics and Biotechnology, 277 pp. stem tissue according to geographic origin. Ecology
Vol II, Conservation and Application, 240 pp. 65: 1087–1092.
New York: Springer-Verlag. Alexander, R.R., Hoffman, G.R., and Wirsing, J.M.
Ainsworth, E.A. and Rogers, A. (2007). The response of (1986). Forest vegetation of the Medicine Bow
photosynthesis and stomatal conductance to rising National Forest in southeastern Wyoming: a hab-
[CO2]: mechanisms and environmental interac- itat type classification. USDA For. Serv. Res. Paper
tions. Plant Cell Environ. 30: 258–270. RM-271. Rocky Mountain For. and Rge. Res. Sta.
Aitken, S.N., Yeaman, S., Holliday, J.A. et al. (2008). Fort Collins, CO. 39 pp.
Adaptation, migration or extirpation: climate Alfaro, R., Campbell, E., and Hawkes, B. (2010). Histor-
change outcomes for tree populations. Evol. Appl. ical Frequency, Intensity, and Extent of Mountain
1: 95–111. Pine Beetle Disturbance in Landscapes of British
Albert, D.A. (1995). Regional landscape ecosystems of Columbia. Victoria, B.C.: Canadian Forest Service,
Michigan, Minnesota, and Wisconsin: a working Pacific Forestry Centre 52 pp.
map and classification. USDA For. Serv. Gen. Tech. Alho, C.J.R. (2008). The value of biodiversity. Brazilian
Report NC-178. North Central For. Exp. Sta., St. Journal of Biology 68: 1115–1118.
Paul, MN. 250 pp. + map. Alix-Garcia, J. and Wolff, H. (2014). Payment for eco-
Albert, D.A. and Barnes, B.V. (1987). Effects of system services from forests. Annu. Rev. Resour.
clearcutting on the vegetation and soil of a sugar Econ. 6: 361–380.
maple-dominated ecosystem, western Upper Allen, G.S. and Owens, J.N. (1972). The Life History of
Michigan. For. Ecol. Manag. 18: 283–298. Douglas-fir. Ottawa: Environment Canada, For.
Albert, D.A., Denton, S.R., and Barnes, B.V. (1986). Serv 139 pp.
Regional Landscape Ecosystems of Michigan. Ann Allen, R.B. and Platt, K.H. (1990). Annual seedfall vari-
Arbor: School of Natural Resources, University of ation in Nothofagus solandri (Fagaceae), Canter-
Michigan 32 pp. bury, New Zealand. Oikos 57: 199–206.
Aldrich, P.R., Parker, G.R., Severson, J.R., and Michler, Allen, A.P., Brown, J.H., and Gillooly, J.F. (2002).
C.H. (2005). Confirmation of oak recruitment Global biodiversity, biochemical kinetics, and the
failure in Indiana old growth forest: 75 years of energetic-equivalence rule. Science 297: 1545–1548.
data. For. Sci. 51: 406–416. Allen, C.D., Macalady, A.K., Chenchouni, H. et al.
Alexander, R.R. (1964). Minimizing windfall around (2010). A global overview of drought and heat-
clear cuttings in spruce-fir forests. For. Sci. 10: induced tree mortality reveals emerging climate
130–142. change risks for forests. For. Ecol. Manag. 259:
Alexander, R.R. (1966). Site indexes for lodgepole pine, 660–684.
with corrections for stand density; instructions for Allen, C.D., Breshears, D.D., and McDowell, N.G.
field use. USDA For. Serv. Res. Paper RM-24. Rocky (2015). On underestimation of global vulnerability
Mountain For. and Rge. Exp. Sta., Fort Collins, CO. to tree mortality and forest die-off from hotter
7 pp. drought in the Anthropocene. Ecosphere 6: 1–55.
Alexander, M. (1977). Introduction to Soil Microbiology, Allen, R.B., Millard, P., and Richardson, S.J. (2017). A
2e. New York: Wiley 467 pp. resource centric view of climate and mast seeding
Alexander, R.R. (1985). Major habitat types, community in trees. Prog. Bot. 79: 233–268.
types, and plant communities in the Rocky Moun- Allendorf, F.W., Luikart, G.H., and Aitkan, S.N. (2012).
tains. USDA For. Serv. Gen. Tech. Report RM-123. Conservation and the Genetics of Populations, 2e.
Fort Collins, CO. 105 pp. Hoboken, NJ: Wiley Blackwell 624 pp.
References 653
Alvarez-Uria, P. and Körner, C. (2007). Low temper- Angelini, C. and Silliman, B.R. (2014). Secondary
ature limits of root growth in deciduous and foundation species as drivers of trophic and
evergreen temperate tree species. Funct. Ecol. functional diversity: evidence from a tree–epiphyte
21: 211–218. system. Ecology 95: 185–196.
Amaral, J., Ribeyre, Z., Vigneaud, J. et al. (2020). Angelini, C., Altieri, A.H., Silliman, B.R., and Bert-
Advances and promises of epigenetics for forest ness, M.D. (2011). Interactions among foundation
trees. Forests 11: 976. species and their consequences for community
Amaranthus, M.P., Trappe, J.M., Bednar, L., and organization, biodiversity, and conservation. Biosci-
Arthur, D. (1994). Hypogeous fungal production in ence 61: 782–789.
mature Douglas-fir forest fragments and surround- Antibus, R.K., Croxdale, J.G., Miller, O.K., and Linkins,
ing plantations and its relation to coarse woody A.E. (1981). Ectomycorrhizal fungi of Salix rotun-
debris and animal mycophagy. Can. J. For. Res. 24: difolia. III. Resynthesized mycorrhizal complexes
2157–2165. and their surface phosphatase activities. Can. J. Bot.
Ambroise, V., Legay, S., Guerriero, G. et al. (2020). The 59: 2458–2465.
roots of plant frost hardiness and tolerance. Plant Antonovics, J. (1971). The effects of a heterogeneous
Cell Physiol. 61: 3–20. environment on the genetics of natural popula-
Amoroso, M.M., Daniels, L.D., Baker, P.J., and tions. Am. Sci. 59: 593–599.
Camarero, J.J. (ed.) (2017). Dendroecology: Tree- Arbeitsgruppe Biözonosenkunde (1995). Ansätze für
Ring Analyses Applied to Ecological Studies. Cham, eine Regionale Biotop-und Biozönosenkunde von
Switzerland: Springer 420 pp. Baden-Wurttemberg. Baden-Wurttemberg. Freiburg,
Amthor, J.S. (1984). The role of maintenance respira- Germany: Mitt. Forst. Versuchs u. Forschungsanst
tion in plant growth. Plant Cell Environ. 7: 561–569. 166 pp.
An, M., Deng, M., Zheng, S.S. et al. (2017). Introgres- Archambault, L., Barnes, B.V., and Witter, J.A. (1989).
sion threatens the genetic diversity of Quercus Ecological species groups of oak ecosystems
austrocochinchinensis (Fagaceae), an endangered of southeastern Michigan, USA. For. Sci. 35:
oak: a case inferred by molecular markers. Front. 1058–1074.
Plant Sci. 8: 229. Archambault, L., Barnes, B.V., and Witter, J.A. (1990).
Anderson, E. (1948). Hybridization of the habitat. Landscape ecosystems of disturbed oak forests of
Evolution 2: 1–9. southeastern Michigan, USA. Can. J. For. Res. 20:
Anderson, M.G. and Ferree, C.E. (2010). Conserving the 1570–1582.
stage: Climate change and the geophysical under- Arduini, G., Chemel, C., and Staquet, C. (2020). Local
pinnings of species diversity. PLoS One 5: e11554. and non-local controls on a persistent cold-air pool
Anderson, J.M. and Swift, M.J. (1983). Decomposi- in the Arve River Valley. Q. J. R. Meteorol. Soc. 146:
tion in tropical forests. In: Tropical Rain Forest: 2497–2521.
Ecology and Management (ed. S.L. Sutton, T.C. Armand, A.D. (1992). Sharp and gradual mountain
Whitmore and A.C. Chadwick), 287–308. Oxford: timberlines as a result of species interaction. In:
Blackwell. Landscape Boundaries: Consequences for Biotic
Anderson, M.G., Clark, M., and Sheldon, A.O. (2014). Diversity and Ecological Flows (ed. A.J. Hansen and
Estimating climate resilience for conservation F. di Castri), 360–379. New York: Springer.
across geophysical settings. Conserv. Biol. 28: Arno, S.F. (1976). The historical role of fire on the Bit-
959–970. terroot National Forest. USDA For. Serv. Res. Paper
Anderson, M. G., Barnett, A., Clark, M. (2016). Resil- INT-187. Intermountain For. and Rge. Exp. Sta.,
ient and connected landscapes for terrestrial Ogden, UT. 29 pp.
conservation. http://easterndivision.s3.amazonaws. Arno, S.F. (1980). Forest fire history of the northern
com/ResilientandConnectedLandscapesForTerres- Rockies. J. For. 78: 460–465.
tialConservation.pdf. Accessed 01-10-2022. Arno, S.F. and Fischer, W.C. (1989). Using vegetation
Anderson-Teixeira, K.J., Miller, A.D., Mohan, J.E. et al. classifications to guide fire management. In D. E.
(2013). Altered dynamics of forest recovery under a Ferguson, P. Morgan, and F. D. Johnson (comps.),
changing climate. Glob. Chang. Biol. 19: 2001–2021. Proceedings—Land Classifications Based on Vegeta-
Andersson, E. (1963). Seed stands and seed orchards tion: Applications for Resource Management. USDA
in the breeding of conifers, World Consult. For. For. Serv. Gen. Tech. Report INT-257, Intermoun-
Gen. and For. Tree Imp. Proc. II FAO-FORGEN tain Res. Sta, Ogden, UT.
63– 8/1:1–18. Arno, S.E, Scott, J.H., and Hartwell, M.G. (1995).
Andrus, R.A., Hart, S.J., Tutland, N., and Veblen, Age-class structure of old growth ponderosa pine/
T.T. (2021). Future dominance by quaking aspen Douglas-fir stands and its relationship to fire
expected following short-interval, compounded history. USDA, For. Serv. Res. Paper INT-RP-481,
disturbance interaction. Ecosphere 12: e03345. Intermountain Res. Sta., Odgen, UT. 25 pp.
654 References
Arno, S.F., Parsons, D.J., and Keane, R.E. (2000). Augspurger, C.K. (2008). Early spring leaf out enhances
Mixed-severity fire regimes in the northern Rocky growth and survival of saplings in a temperate
Mountains: consequences of fire exclusion and deciduous forest. Oecologia 156: 281–286.
options for the future. In: Wilderness Science Augspurger, C.K. (2009). Spring 2007 warmth and frost:
in a Time of Change Conference (ed. D.N. Cole, phenology, damage and refoliation in a temperate
S.F. McCool, W.A. Freimund and J. O’Loughlin) deciduous forest. Funct. Ecol. 23: 1031–1039.
(comps.). 1999 May 23–27; Missoula, MT. USDA Aukema, J.E., McCullough, D.G., Von Holle, B. et al.
For. Serv. Rocky Mountain Res. Sta. Proceedings (2010). Historical accumulation of nonindigenous
RMRS-P-15-VOL-1. Ogden, UT. forest pests in the continental United States. Biosci-
Arnold, M.L. (1994). Natural hybridization and Louisi- ence 60: 886–897.
ana irises. Bioscience 44: 141–147. Aukema, J.E., Leung, B., Kovacs, K. et al. (2011).
Arnold, M.L. (2015). Divergence with Genetic Exchange. Economic impacts of non-native forest insects in
Oxford, UK: Oxford University Press 272 pp. the continental United States. PLoS One 6: e24587.
Aronson, J., Murcia, C., Kattan, G.H. et al. (2014). The Austerlitz, F., Mariette, S., Machon, N. et al. (2000).
road to confusion is paved with novel ecosystem Effects of colonization processes on genetic diversity:
labels: a reply to Hobbs et al. Trends Ecol. Evol. 29: differences between annual plants and tree species.
646–647. Genetics 154: 1309–1321.
Ashcroft, M.B., Gollan, J.R., Warton, D.I., and Ramp, Austin, M.P. (1985). Continuum concept, ordination
D. (2012). A novel approach to quantify and locate methods, and niche theory. Annu. Rev. Ecol. Syst.
potential microrefugia using topoclimate, climate 16: 39–61.
stability, and isolation from the matrix. Glob. Austin, M.P. (1990). Community theory and competi-
Chang. Biol. 18: 1866–1879. tion in vegetation. In: Perspectives on Plant Com-
Ashton, P.S. (1992). Species richness in plant commu- petition (ed. J.B. Grace and D. Tilman), 215–238.
nities. In: Conservation Biology (ed. P.L. Fiedler and New York: Academic Press.
K.S. Jain), 3–22. London: Chapman and Hall. Austin, M.P. and Smith, T.M. (1989). A new model for
Ashton, P.M.S. and Berlyn, G.P. (1994). A comparison of the continuum concept. Vegetatio 83: 35–47.
leaf physiology and anatomy of Quercus (section of Avers, P.E. and Schlatterer, E.F. (1991). Ecosystem
Erythrobalanus-Fagaceae) species in different light classification and management on National Forests.
environments. Am. J. Bot. 81: 589–597. In: Proceedings of the 1991 Symposium on Systems
Ashton, M.S. and Kelty, M.J. (2018). The Practice of Analysis in Forest Resources. March 3–6, 1991,
Silviculture: Applied Forest Ecology, 10e. Hoboken, Charleston South Carolina. USDA For. Serv. South-
New Jersey: Wiley 776 pp. eastern Exp. Sta, Asheville, NC. 423 pp.
Ashton, M.S., Tyrrell, M.L., Spalding, D., and Gentry, B. Ayres, M.P. and Lombardero, M.J. (2000). Assessing the
(ed.) (2012). Managing Forest Carbon in a Changing consequences of global change for forest distur-
Climate. New York: Springer 424 pp. bance from herbivores and pathogens. Sci. Total
Asner, G.P. and Vitousek, P.M. (2005). Remote analysis Environ. 262: 263–286.
of biological invasion and biogeochemical change. Azizi Jalilian, M., Shayesteh, K., Danehkar, A., and
Proc. Natl. Acad. Sci. 102: 4383–4386. Salmanmahiny, A. (2020). A new ecosystem-based
Assmann, E. (1970). The Principles of Forest Yield land classification of Iran for conservation goals.
Studies. Oxford: Pergamon Press 506 pp. Environ. Monit. Assess. 192: 1–17.
Aston, J.L. and Bradshaw, A.D. (1966). Evolution in Baack, E., Melo, M.C., Rieseberg, L.H., and Ortiz-Barrientos,
closely adjacent plant populations. II. Agrostis sto- D. (2015). The origins of reproductive isolation in
lonifera in maritime habitats. Heredity 21: 649–664. plants. New Phytol. 207: 968–984.
Atchley, A.L., Kinoshita, A.M., Lopez, S.R. et al. (2018). Bacles, C.F.E., Lowe, A.J., and Ennos, R.A. (2006).
Simulating surface and subsurface water balance Effective seed dispersal across a fragmented
changes due to burn severity. Vadose Zone J. 17: landscape. Science 311: 628.
180099. Bailey, R.G. (1983). Delineation of ecosystem regions.
Atkins, J.W., Bohrer, G., Fahey, R.T. et al. (2018a). Environ. Manag. 7: 365–373.
Quantifying vegetation and canopy structural com- Bailey, R.G. (1988). Ecogeographic analysis: a guide
plexity from terrestrial LiDAR data using the FOR- to the ecological division of land for resource
ESTR R package. Methods Ecol. Evol. 9: 2057–2066. management. USDA For. Serv. Misc. Publ. No.
Atkins, J.W., Fahey, R.T., Hardiman, B.S., and Gough, 1465. Washington, D.C. 18 pp.
C.M. (2018b). Forest canopy structural complexity Bailey, R.G. (1994). Ecoregions of the United States
and light absorption relationships at the sub- (map). USDA For. Serv., Washington, D.C.
continental scale. J. Geophys. Res. Biogeosci. 123: Bailey, R.G. (1995). Description of the ecoregions of the
1387–1405. United States. USDA For. Serv. Misc. Publ. 1391.
Augenbaugh, J.E. (1935). Replacement of chestnut in Washington, D.C. 108 pp. + map
Pennsylvania. Pa. Dept. For. Waters. Water Resour. Bailey, R.G. (2002). Ecoregion-Based Design for Sustain-
Bull. 54: 1–38. ability, 2e. New York: Springer 236 pp.
References 655
Bailey, R.G. (2009). Ecosystem Geography, 2e. New York: Colonizing Species (ed. H.G. Baker and G.L. Steb-
Springer-Verlag 251 pp + 1 map. bins), 353–372. New York: Academic Press.
Bailey, R.G. and Hogg, H.C. (1986). A world ecoregions Bao, Y. and Nilsen, E.T. (1988). The ecophysiological
map for resource reporting. Environ. Conserv. 13: significance of thermotropic leaf movements in
195–202. Rhododendron maximum. Ecology 69: 1578–1587.
Bailey, R.G., Zoltai, S.C., and Wiken, E.B. (1985). Eco- Barber, H.N. (1955). Adaptive gene substitutions in
logical regionalization in Canada and the United Tasmanian eucalyptus: I. genes controlling the
States. Geoforum 16: 265–275. development of glaucousness. Evolution 9: 1–14.
Bailey, J.J., Boyd, D.S., Hjort, J. et al. (2017). Modelling Barber, H.N. and Jackson, W.D. (1957). Natural
native and alien vascular plant species richness: At selection in action in Eucalyptus. Nature 179:
which scales is geodiversity most relevant? Glob. 1267–1269.
Ecol. Biogeogr. 26: 763–776. Barbero, R., Abatzoglou, J.T., Larkin, N.K. et al. (2015).
Baker, F.S. (1949). A revised tolerance table. J. For. 47: Climate change presents increased potential for
179–181. very large fires in the contiguous United States. Int.
Baker, W.L. (2009). Fire Ecology in Rocky Mountain J. Wildland Fire 24: 892–899.
Landscapes. Washington, DC: Island Press 632 pp. Barbour, M.G. and Billings, W.D. (1988). North
Baker, M.E. and Barnes, B.V. (1998). Landscape eco- American Terrestrial Vegetation. Cambridge:
system diversity of river floodplains in north- Cambridge Univ. Press 434 pp.
western Lower Michigan, USA. Can. J. For. Res. 28: Barbour, M.G. and Christensen, N.L. (1993). Vegetation.
1405–1418. In: Flora of North America, North of Mexico, vol.
Baker, W.L. and Ehle, D. (2001). Uncertainty in 1 (ed. R.R. Morin) (conv. ed.), 97–131. New York:
surface-fire history: the case of ponderosa pine Oxford Univ. Press.
forests in the western United States. Can. J. For. Barden, L.S. (1981). Forest development in canopy
Res. 31: 1205–1226. gaps of a diverse hardwood forest of the southern
Baker, H.G., Bawa, K.S., Frankie, G.W., and Opler, P.A. Appalachian Mountains. Oikos 37: 205–209.
(1983). Reproductive biology of plants in tropical Barnes, B.V. (1967). The clonal growth habit of
forests. In: Tropical Rainforest Ecosystems (ed. F.B. American aspens. Ecology 47: 439–447.
Golley), 183–215. New York: Elsevier. Barnes, B.V. (1969). Natural variation and delineation
Baker, W.L., Monroe, J.A., and Hessl, A.E. (1997). The of clones of Populus tremuloides and P. grandiden-
effects of elk on aspen in the winter range in Rocky tata in northern Lower Michigan. Silvae Genet. 18:
Mountain National Park. Ecography 20: 155–165. 130–142.
Baldwin, D.J.B., Desloges, J.R., and Band, L.E. (2000). Barnes, B.V. (1975). Phenotypic variation of trembling
Physical geography of Ontario. In: Ecology of aspen in western North America. For. Sci. 21: 319–328.
a Managed Terrestrial Landscape: Patterns and Barnes, B.V. (1976). Succession in deciduous swamp
Processes of Forest Landscapes in Ontario (ed. A.H. communities of southeastern Michigan, formerly
Perera, D.L. Euler and I.D. Thompson), 12–29. dominated by American elm. Can. J. Bot. 54: 19–24.
Vancouver, British Columbia: UBC Press. Barnes, B.V. (1977). The International Larch Prove-
Balisky, A.C. and Burton, P.J. (1995). Root-zone soil nance Test in southeastern Michigan, USA. Silvae
temperature variation associated with microsite Genet. 26: 145–148.
characteristics in high-elevation forest openings in Barnes, B.V. (1984). Forest ecosystem classification and
the interior of British Columbia. Agric. For. Meteo- mapping in Baden-Württemberg, Germany. In: Symp.
rol. 77: 31–54. Proc. Forest Land Classification: Experience, Problems,
Baltzer, J.L. and Thomas, S.C. (2007). Determinants of Perspectives (ed. J.G. Bockheim). NCR-102 North
whole-plant light requirements in Bornean rain Central For. Soils Com., Soc. Am. For., USDA For.
forest tree saplings. J. Ecol. 95: 1208–1221. Serv., and USDA Soil Cons. Serv. Madison, WI.
Baluška, F., Čiamporová, M., Gašparíková, O., and Barnes, B.V. (1991). Deciduous forests of North
Barlow, P.W. (1995). Structure and Function of America. In: Ecosystems of the World, Temperate
Roots. Norwell, MA: Kluwer Academic Publishers Deciduous Forests, vol. 7 (ed. E. Röhrig and B.
364 pp. Ulrich), 219–344. New York: Elsevier.
Balvanera, P., Pfisterer, A.B., Buchmann, N. et al. Barnes, B.V. (1993). The landscape ecosystem approach
(2006). Quantifying the evidence for biodiversity and conservation of endangered spaces. Endanger.
effects on ecosystem functioning and services. Ecol. Species Update 10: 13–19.
Lett. 9: 1146–1156. Barnes, B.V. (1996). Silviculture, landscape ecosys-
Balvanera, P., Siddique, I., Dee, L. et al. (2014). Linking tems, and the iron law of the site. Forstarchiv 67:
biodiversity and ecosystem services: current uncer- 226–235.
tainties and the necessary next steps. Bioscience 64: Barnes, B.V. and Dancik, B.P. (1985). Characteristics
49–57. and origin of a new birch species, Betula mur-
Bannister, M.H. (1965). Variation in the breeding rayana, from southeastern Michigan. Can. J. Bot.
system of Pinus radiata. In: The Genetics of 63: 223–226.
656 References
Barnes, B.V. and Wagner, W.H. (2004). Michigan Beck, D.E. (1971). Polymorphic site index curves for
Trees. Ann Arbor, MI: University of Michigan white pine in the southern Appalachians. USDA
Press 456 pp. For. Serv. Res. Note SE-80. Southeastern For. Exp.
Barnes, B.V., Pregitzer, K.S., Spies, T.A., and Spooner, Sta., Asheville, NC. 8 pp.
V. (1982). Ecological forest site classification. J. For. Beck, D.E. (1977). Twelve-year acorn yield in southern
80: 493–498. Appalachian oaks. USDA For. Serv. Res. Note SE-244,
Barnes, B.V., Xü, Z., and Zhao, S. (1992). Forest eco- Southeastern For. Exp. Sta., Asheville, NC. 8 pp.
systems in an old-growth pine—mixed hardwood Beck, E.H., Heim, R., and Hansen, J. (2004). Plant
forest of the Changbai Shan Preserve in north- resistance to cold stress: mechanisms and envi-
eastern China. Can. J. For. Res. 22: 144–160. ronmental signals triggering frost hardening and
Barnes, B.V., Zak, D.R., Denton, S.R., and Spurr, S.H. dehardening. J. Biosci. 29: 449–459.
(1998). Forest Ecology, 4e. New York: Wiley 774 pp. Becking, R.W. (1957). The Zürich-Montpellier school of
Barnosky, C.W. (1987). Response of vegetation to phytosociology. Bot. Rev. 23: 411–488.
climatic changes of different duration in the Late Beese, W.J., Deal, J., Dunsworth, B.G. et al. (2019a). Two
Neocene. Trends Ecol. Evol. 2: 247–250. decades of variable retention in British Columbia: a
Barrios-Garcia, M.N. and Ballari, S.A. (2012). Impact of review of its implementation and effectiveness for
wild boar (Sus scrofa) in its introduced and native biodiversity conservation. Ecol. Process. 8: 1–22.
range: a review. Biol. Invasions 14: 2283–2300. Beese, W.J., Rollerson, T.P., and Peters, C.M. (2019b).
Barry, R.C. (1967). Seasonal location of the Arctic Front Quantifying wind damage associated with variable
over North America. Geogr. Bull. 9: 79–95. retention harvesting in coastal British Columbia.
Barthelemy, D., Edelin, C., and Hallé, F. (1991). Canopy For. Ecol. Manag. 443: 117–131.
architecture. In: Physiology of Trees (ed. A.S. Begon, M., Harper, J.L., and Townsend, C.R. (1990).
Raghavendra), 1–20. New York: Wiley. Ecology, Individuals, Populations, and Communities,
Barton, N.H. (2001). The role of hybridization in evolu- 945. Boston: Blackwell Sci. Publ.
tion. Mol. Ecol. 10: 551–568. Beisner, B.E., Haydon, D.T., and Cuddington, K. (2003).
Barton, K.E. and Koricheva, J. (2010). The ontogeny Alternative stable states in ecology. Front. Ecol.
of plant defense and herbivory: characterizing Environ. 1: 376–382.
general patterns using meta-analysis. Am. Nat. Belahbib, N., Pemonge, M.H., Ouassou, A. et al. (2001).
175: 481–493. Frequent cytoplasmic exchanges between oak
Basey, J.M., Jenkins, S.H., and Busher, P.E. (1988). species that are not closely related: Quercus suber
Optimal central-place foraging by beavers: tree-size and Q. ilex in Morocco. Mol. Ecol. 10: 2003–2012.
selection in relation to defensive chemicals of Bell, A.D. (1991). Plant Form, An Illustrated Guide to
quaking aspen. Oecologia 76: 278–282. Flowering Plant Morphology. New York: Oxford
Basey, J.M., Jenkins, S.H., and Miller, G.C. (1990). Univ. Press 341 pp.
Food selection by beavers in relation to inducible Belleau, A., Leduc, A., Lecomte, N., and Bergeron, Y.
defenses of Populus tremuloides. Oikos 59: 57–62. (2011). Forest succession rate and pathways on dif-
Baskin, Y. (2002). A Plague of Rats and Rubber Vines: ferent surface deposit types in the boreal forest of
The Growing Threat of Species Invasions. Washing- northwestern Quebec. Ecoscience 18: 329–340.
ton, D.C.: Island Press 330 pp. Belsky, A.J. and Blumenthal, D.M. (1997). Effects of
Basler, D. and Körner, C. (2012). Photoperiod sensitivity livestock grazing on stand dynamics and soils in
of bud burst in 14 temperate forest tree species. upland forests of the interior West. Conserv. Biol.
Agric. Meteorol. 165: 73–81. 11: 315–327.
Battaglia, L.L. and Shari, R.R. (2006). Responses of Bemmels, J.B., Knowles, L.L., and Dick, C.W. (2019).
floodplain forest species to spatially condensed gra- Genomic evidence of survival near ice sheet mar-
dients: a test of the flood–shade tolerance trade-off gins for some, but not all, North American trees.
hypothesis. Oecologia 147: 108–118. Proc. Natl. Acad. Sci. 116: 8431–8436.
Bauerle, W.L., Bowden, J.D., Wang, G.G., and Shahba, M.A. Bender, E.A., Case, T.J., and Gilpin, M.E. (1984). Pertur-
(2009). Exploring the importance of within-canopy bation experiments in ecology: theory and practice.
spatial temperature variation on transpiration Ecology 65: 1–13.
predictions. J. Exp. Bot. 60: 3665–3676. Bender, M.H., Baskin, J.M., and Baskin, C.C. (2000).
Baumeister, D. and Callaway, R.M. (2006). Facilitation Age of maturity and life span in herbaceous poly-
by Pinus flexilis during succession: a hierarchy of carpic perennials. Bot. Rev. 66: 311–349.
mechanisms benefits other plant species. Ecology Bender, D.J., Tischendorf, L., and Fahrig, L. (2003).
87: 1816–1830. Evaluation of patch isolation metrics for predicting
Bazzaz, F.A. (1979). The physiological ecology of plant animal movement in binary landscapes. Landsc.
succession. Annu. Rev. Ecol. Syst. 10: 351–372. Ecol. 18: 17–39.
Beaufait, W.R. (1956). Influence of soil and topography Bendix, J., Wiley, J.J. Jr., and Commons, M.G. (2017).
on willow oak sites. U.S. For. Sen., Southern For. Intermediate disturbance and patterns of species
Exp. Sta. Occ. Paper 148. 12 pp. richness. Phys. Geogr. 38: 393–403.
References 657
Bentz, B.J., Régnière, J., Fettig, C.J. et al. (2010). Climate Interior, Nat. Park Serv., Nat. Res. Publ. Office,
change and bark beetles of the western United Denver. CO.
States and Canada: Direct and indirect effects. Bio- Billings, S.A. and Gaydess, E.A. (2008). Soil nitrogen
science 60: 602–613. and carbon dynamics in a fragmented landscape
Bentz, B.J., Jönsson, A.M., Schroeder, M. et al. (2019). experiencing forest succession. Landsc. Ecol. 23:
Ips typographus and Dendroctonus ponderosae 581–593.
models project thermal suitability for intra-and Binkley, D. (1995). The influence of tree species on
inter-continental establishment in a changing cli- forest soils: processes and patterns. In Mead, D.
mate. Front. For. Glob. Change 2: 1. J., and I. S. Cornforth (eds.), Proc. Trees and Soils
Berdanier, A.B. and Clark, J.S. (2016). Divergent repro- Workshop. Agronomy Society of New Zealand
ductive allocation trade-offs with canopy exposure Special Publication No. 10. Lincoln Univ. Press,
across tree species in temperate forests. Ecosphere Canterbury.
7: e01313. Binkley, D. and Hart, S. (1989). The components of
Beric, B. and MacIsaac, H.J. (2015). Determinants of nitrogen availability assessments in forest soils.
rapid response success for alien invasive species in Adv. Soil Sci. 10: 57–115.
aquatic ecosystems. Biol. Invasions 17: 3327–3335. Binkley, D., Moore, M.M., Romme, W.H., and Brown,
Berntson, G.M. and Woodward, F.I. (1992). The root P.M. (2006). Was Aldo Leopold right about the
system architecture and development of Senecio Kaibab deer herd? Ecosystems 9: 227–241.
vulgaris in elevated CO2 and drought. Funct. Ecol. Bird, A. (2007). Perceptions of epigenetics. Nature 447:
6: 324–333. 396–398.
Berry, J. and Bjorkman, O. (1980). Photosynthetic Birdsey, R.A., Pregitzer, K.S., and Lucier, A. (2006).
response and adaptation to temperature in higher Forest carbon management in the United States:
plants. Ann. Rev. Plant Physiol. 31: 491–543. 1600-2100. J. Environ. Qual. 35: 1461–1469.
Berry, J. and Downton, W.J.S. (1982). Environmental Birkeland, P.W. (1974). Pedology, Weathering and Geo-
regulation of photosynthesis. In: Photosynthesis, morphological Research. New York: Oxford Univ.
Development, Carbon Metabolism, and Plant Press 285 pp.
Productivity, vol. 2 (ed. R. Govindjee). New York: Biswell, H.H. (1961). The big trees and fires. Natl. Parks
Academic Press 580 pp. Mag. 35: 11–14.
Berry, P.M., Fabók, V., Blicharska, M. et al. (2018). Why Biswell, H.H. (1974). Effects of fire on chaparral. In:
conserve biodiversity? A multi-national explora- Fire and Ecosystems (ed. T.T. Kozlowski and C.E.
tion of stakeholders’ views on the arguments for Ahlgren), 321–364. New York: Academic Press.
biodiversity conservation. Biodivers. Conserv. 27: Björklund, J., Seftigen, K., Schweingruber, F. et al.
1741–1762. (2017). Cell size and wall dimensions drive distinct
Beschta, R.L. and Ripple, W.J. (2016). Riparian vegeta- variability of earlywood and latewood density in
tion recovery in Yellowstone: the first two decades Northern Hemisphere conifers. New Phytol. 216:
after wolf reintroduction. Biol. Conserv. 198: 728–740.
93–103. Blake, E.S. and Gibney, E.J. (2011). The Deadliest,
Besnard, G., Acheré, V., Jeandroz, S. et al. (2008). Does Costliest, and Most Intense United States Tropical
maternal environmental condition during repro- Cyclones from 1851 to 2010 (and other Frequently
ductive development induce genotypic selection in Requested Hurricane Facts). NOAA Technical Mem-
Picea abies? Ann. For. Sci. 65: 109. orandum NWS NHC – 6. 47 pp.
Bidwell, R.G.S. (1974). Plant Physiology. New York: Blasi, C., Carranza, M.L., Frondoni, R., and Rosati, L.
Macmillan 643 pp. (2000). Ecosystem classification and mapping: a
Bigler, C., Kulakowski, D., and Veblen, T.T. (2005). proposal for Italian landscapes. Appl. Veg. Sci. 3:
Multiple disturbance interactions and drought 233–242.
influence fire severity in Rocky Mountain subal- Blennow, K., Andersson, M., Sallnäs, O., and Olofsson,
pine forests. Ecology 86: 3018–3029. E. (2010). Climate change and the probability of
Billings, W.D. (1973). Arctic and alpine vegetations: wind damage in two Swedish forests. For. Ecol.
Similarities, differences, and susceptibility to dis- Manag. 259: 818–830.
turbance. Bioscience 23: 697–704. Bliss, L.C. and Cantlon, J.E. (1957). Succession on river
Billings, W.D. (1987). Constraints to plant growth, alluvium in northern Alaska. Am. Midi. Nat. 58:
reproduction, and establishment in Arctic environ- 452–469.
ments. Arctic and Alpine Res. 19: 357–365. Blonder, B., Graae, B.J., Greer, B. et al. (2020). Remote
Billings, W.D. (1995). The effects of global and regional sensing of ploidy level in quaking aspen (Populus
environmental changes on mountain ecosystems. tremuloides Michx.). J. Ecol. 108: 175–188.
In D. G. Despain (ed.), Plants and their Environ- Bloomberg, W.J. (1950). Fire and spruce. For. Chron. 26:
ment: Proceedings of the First Biennial Scientific 157–161.
Conference on the Greater Yellowstone Ecosystem. Blumenthal, D. (2005). Interrelated causes of plant
Tech. Report NPS/NRYELL/NRTR-93XX. US Dept. invasion. Science 310: 243–244.
658 References
Boege, K. and Marquis, R.J. (2005). Facing herbivory as Boose, E.R., Foster, D.R., and Fluet, M. (1994).
you grow up: the ontogeny of resistance in plants. Hurricane impacts to tropical and temperate forest
Trends Ecol. Evol. 20: 441–448. landscapes. Ecol. Monogr. 64: 369–400.
Boettcher, S.E. and Kalisz, P.J. (1990). Single-tree Borchert, R. (1991). Growth periodicity and dormancy.
influence on soil properties in the mountains of In: Physiology of Trees (ed. A.S. Raghavendra),
eastern Kentucky. Ecology 71: 1365–1372. 221–245. New York: Wiley.
Boggess, W.R. (1964). Trelease Woods, Champaign Boring, L.R., Swank, W.T., Waide, J.B., and Henderson,
County, Illinois: woody vegetation and stand com- G.S. (1988). Sources, fates, and impacts of nitrogen
position. Trans. Ill. Acad. Sci. 57: 261–271. inputs to terrestrial ecosystems: review and syn-
Boggess, W.R. and Bailey, L.W. (1964). Brownfield thesis. Biogeochemistry 6: 119–159.
Woods, Illinois: woody vegetation and changes Bormann, F.H. and Likens, G.E. (1979). Pattern
since 1925. Am. Midi. Nat. 71: 392–401. and Process in a Forested Ecosystem. New York:
Boggess, W.R. and Geis, J.W. (1966). The Funk Forest Springer-Verlag 253 pp.
Natural Area, McLean County, Illinois: woody veg- Bormann, F.H., Likens, G.E., and Melillo, J.M. (1977).
etation and ecological trends. Trans. Ill. Acad. Sci. Nitrogen budget for an aggrading northern hard-
59: 123–133. wood forest ecosystem. Science 196: 981–982.
Bognounou, F., De Grandpré, L., Pureswaran, D.S., and Boshier, D., Broadhurst, L., Cornelius, J. et al. (2015).
Kneeshaw, D. (2017). Temporal variation in plant Is local best? Examining the evidence for local
neighborhood effects on the defoliation of primary adaptation in trees and its scale. Environ. Evid. 4:
and secondary hosts by an insect pest. Ecosphere 1–10.
8: 1–15. Bossdorf, O., Richards, C.L., and Pigliucci, M. (2008).
Bohlen, P.J., Scheu, S., Hale, C.M. et al. (2004). Non-native Epigenetics for ecologists. Ecol. Lett. 11: 106–115.
invasive earthworms as agents of change in Botkin, D.B. (1993). Forest Dynamics: An Ecological
northern temperate forests. Front. Ecol. Environ. 2: Model. New York: Oxford Univ. Press 309 pp.
427–435. Botkin, D.B., Woodwell, G.M., and Tempel, N. (1970).
Boisvert-Marsh, L., Périé, C., and de Blois, S. (2014). Forest productivity estimated from carbon dioxide
Shifting with climate? Evidence for recent changes uptake. Ecology 51: 1057–1060.
in tree species distribution at high latitudes. Eco- Botting, D. (1973). Humboldt and the Cosmos. London:
sphere 5: 1–33. Harper and Row 295 pp.
Boisvert-Marsh, L., Périé, C., and de Blois, S. (2019). Boucher, D.H. (1985). The Biology of Mutualism. Lon-
Divergent responses to climate change and distur- don: Croom Helm 388 pp.
bance drive recruitment patterns underlying latitu- Boucher, D.H., James, S., and Keeler, K.H. (1982). The
dinal shifts of tree species. J. Ecol. 107: 1956–1969. ecology of mutualism. Annu. Rev. Ecol. Syst. 13:
Bolan, N.S., Robson, A.D., Barrow, N.J., and Aylmore, 315–347.
L.A.G. (1984). Specific activity of phosphorus in Boucher, D., Gauthier, S., Thiffault, N. et al. (2019).
mycorrhizal and non-mycorrhizal plants in relation How climate change might affect tree regeneration
to the availability of phosphorus to plants. Soil Biol. following fire at northern latitudes: a review. New
Biochem. 6: 299–304. For. 51: 543–571.
Bolstad, P.V., Swank, W., and Vose, J. (1998). Predicting Boufford, D.E. and Spongberg, S.A. (1983). Eastern
southern Appalachian overstory vegetation with Asian–eastern North American phytogeographical
digital terrain data. Landsc. Ecol. 13: 271–283. relationships—a history from the time of Linnaeus
Bonan, G.B. and Shugart, H.H. (1989). Environmental to the twentieth century. Ann. Missouri Bot. Gard.
factors and ecological processes in boreal forests. 70: 423–439.
Annu. Rev. Ecol. Syst. 20: 1–28. Bowen, G.D. and Smith, S.E. (1981). The effects of
Bond, W.J. (1993). Keystone species. In: Ecosystem mycorrhizas on nitrogen uptake by plants. Ecol.
Function and Biodiversity (ed. E.D. Schulze Bull. 33: 237–247. Swedish Natural Sci. Res.
and H.A. Mooney), 237–253. New York: Council, Stockholm.
Springer-Verlag. Boyce, S.G. and Kaeiser, M. (1961). Why yellow-poplar
Bond, W.J. and van Wilgen, B.W. (1996). Fire and Plants. seeds have low viability. USDA For. Serv. Central
London: Chapman & Hall 263 pp. States For. Exp. Sta., Tech. Paper 186. 16 pp.
Bond-Lamberty, B., Wang, C., and Gower, S.T. (2004). Boyle, T.J.B. (1992). Biodiversity of Canadian forests:
Net primary production and net ecosystem produc- current status and future challenges. For. Chron.
tion of a boreal black spruce fire chronosequence. 68: 444–453.
Glob. Chang. Biol. 10: 473–487. Boyle, J.R. and Ek, A.R. (1973). Whole tree har-
Bonga, J.M. and von Aderkas, P. (1993). Rejuvenation vesting: nutrient budget evaluation. J. For. 71:
of tissues from mature conifers and its implica- 760–762.
tions for propagation in vitro. In: Clonal Forestry Boyle, T., Liengsiri, C., and Piewluang, C. (1990). Ge-
I, Genetics and Biotechnology (ed. M.R. Ahuja and netic structure of black spruce on two contrasting
W.J. Libby), 182–199. Berlin: Springer-Verlag. sites. Heredity 65: 393–399.
References 659
Bradford, K.J. and Hsiao, T.C. (1982). Physiological Brayshaw, T.C. (1965). Native Poplars of Southern
responses to moderate water stress. In: Alberta and Their Hybrids. Canada: Dept. of
Physiological Plant Ecology II: Water Relations Forestry Publ No. 1109. 40 pp.
and Carbon Assimilation. Encyclopedia of Plant Brayton, R. and Mooney, G.A. (1966). Population
Physiology, vol. 12B (ed. O.L. Lange, P.S. Nobel, variability of Cercocarpus in the White Mountains
C.B. Osmond and H. Ziegler), 263–324. New York: of California as related to habitat. Evolution 20:
Springer-Verlag. 383–391.
Bradford, J.B., Birdsey, R.A., Joyce, L.A., and Ryan, Breckle, S.-W. (1974). Notes on alpine and nival flora
M.G. (2008a). Tree age, disturbance history, of the Hindu Kush, East Afganistan. Bot. Not. 127:
and carbon stocks and fluxes in subalpine 278–284.
Rocky Mountain forests. Glob. Chang. Biol. 14: Breckle, S.-W. (2002). Walter’s Vegetation of the Earth:
2882–2897. The Ecological Systems of the Geo-biosphere, 4e.
Bradford, M.A., Davies, C.A., Frey, S.D. et al. (2008b). Berlin: Springer-Verlag 527 pp.
Thermal adaptation of soil microbial respiration to Brehm, A.M., Mortelliti, A., Maynard, G.A., and
elevated temperature. Ecol. Lett. 11: 1316–1327. Zydlewski, J. (2019). Land-use change and the
Bradford, M.A., Warren, R.J. II, Baldrian, P. et al. (2014). ecological consequences of personality in small
Climate fails to predict wood decomposition at mammals. Ecol. Lett. 22: 1387–1395.
regional scales. Nat. Clim. Chang. 4: 625–630. Bremer, L.L. and Farley, K.A. (2010). Does plantation
Bradley, A.F., Fischer, W.C., and Noste, N.V. (1992). Fire forestry restore biodiversity or create green deserts?
ecology of the forest habitat types of eastern Idaho A synthesis of the effects of land-use transitions
and western Wyoming. USDA For. Serv. Gen. Tech. on plant species richness. Biodivers. Conserv. 19:
Report INT-290. Intermountain Res. Sta., Ogden, 3893–3915.
UT. 92 pp. Brewer, S. (2008). Declines in plant species richness and
Bradley, B.A., Houghton, R.A., Mustard, J.F., and Ham- endemic plant species in longleaf pine savannas
burg, S.P. (2006). Invasive grass reduces above- invaded by Imperata cylindrica. Biol. Invasions 10:
ground carbon stocks in shrublands of the western 1257–1264.
US. Glob. Chang. Biol. 12: 1815–1822. Brewer, R. and Merritt, P.G. (1978). Wind throw and
Bradshaw, K.E. (1965). Soil use and management in the tree replacement in a climax beech-maple forest.
National Forest of California. In: Forest–Soil Rela- Oikos 30: 149–152.
tionships in North America (ed. C.T. Young-berg), Brienen, R.J., Caldwell, L., Duchesne, L. et al. (2020).
413–424. Corvallis: Oregon State Univ. Press. Forest carbon sink neutralized by pervasive
Bradshaw, A.D. (1971). Plant evolution in extreme envi- growth-lifespan trade-offs. Nat. C ommun. 11: 1–10.
ronments. In: Ecological Genetics and Evolution (ed. Briffa, K.R., Schweingruber, F.H., Jones, P.D. et al.
R. Creed), 20–50. Oxford: Blackwell Press. (1998). Reduced sensitivity of recent tree-growth
Bradshaw, G.A. and Marquet, P.A. (ed.) (2002). How to temperature at high northern latitudes. Nature
Landscapes Change: Human Disturbance and 391: 678–682.
Fragmentation in the Americas. Ecological Studies Brinson, M.M. (1990). Riverine forests. In: Ecosystems of
162. New York: Springer 383 pp. the World, Forested Wetlands, vol. 15 (ed. D. Good-
Brady, N.C. (1974). The Nature and Properties of Soils, all, A. Lugo, M. Brinson and S. Brown), 87–141.
8e. New York: MacMillan 639 pp. New York: Elsevier.
Brady, N.C. (1990). The Nature and Properties of Soils, Brix, H. (1971). Effects of nitrogen fertilization on pho-
10e. New York: Macmillan 621 pp. tosynthesis and respiration in Douglas-fir. For. Sci.
Brandt, J., Antrop, M., de Blust, G. et al. (2009). Why a 17: 407–414.
European chapter of IALE? IALE Bull. 27: 1–5. Brix, H. and Ebell, L.F. (1969). Effects of nitrogen fer-
Brasier, C.M. (2001). Rapid evolution of introduced tilization on growth, leaf area, and photosynthesis
plant pathogens via interspecific hybridization: rate of Douglas-fir. For. Sci. 15: 189–196.
hybridization is leading to rapid evolution of Dutch Broadfoot, W.M. and Williston, H.L. (1973). Flooding
elm disease and other fungal plant pathogens. Bio- effects on southern forests. J. For. 71: 584–587.
science 51: 123–133. Bronson, D.R., Gower, S.T., Tanner, M. et al. (2008).
Braun, E.L. (1950). Deciduous Forests of Eastern North Response of soil surface CO2 flux in a boreal forest
America. New York: McGraw-Hill 596 pp. to ecosystem warming. Glob. Chang. Biol. 14:
Braun-Blanquet, J. (1921). Prinzipien einer Systematik 856–867.
der Pflanzengellschaften auf floristischer Grund- Bronstein, J.L. (1994). Our current understanding of
lage. Jahrb. St. Gllen Naturw. Ges. 57: 305–351. mutualism. Quarterly Review of Biology 69 (1): 31–51.
Braun-Blanquet, J. (1964). Pflanzensoziologie, Bronstein, J.L. (2009). The evolution of facilitation and
Grundzüge der Vegetationskunde, 3e. New York: mutualism. J. Ecol. 97: 1160–1170.
Springer-Verlag 865 pp. Bronstein, J.L. (2015). The study of mutualism. In:
Bray, J.R. and Gorham, E. (1964). Litter production in Mutualism (ed. J.L. Bronstein). Oxford, UK: Oxford
forests of the world. Adv. Ecol. Res. 2: 101–157. University Press 320 pp.
660 References
Brooks, M.L., D’Antonio, C.M., Richardson, D.M. et al. Bugmann, H. and Bigler, C. (2011). Will the CO2 fer-
(2004). Effects of invasive alien plants on fire tilization effect in forests be offset by reduced tree
regimes. Bioscience 54: 677–688. longevity? Oecologia 165: 533–544.
Brouillet, L. and Whetstone, R.D. (1993). Climate and Buma, B. (2015). Disturbance interactions: character-
physiography. In: Flora of North America, North of ization, prediction, and the potential for cascading
Mexico, vol. 1 (ed. R.R. Morin) conv. ed. New York: effects. Ecosphere 6: 70.
Oxford Univ. Press. Buma, B. and Wessman, C.A. (2012). Differential
Brown, C.D. and Johnstone, J.F. (2012). Once burned, species responses to compounded perturbations
twice shy: Repeat fires reduce seed availability and and implications for landscape heterogeneity and
alter substrate constraints on Picea mariana regen- resilience. For. Ecol. Manag. 266: 25–33.
eration. For. Ecol. Manag. 266: 34–41. Buma, B., Bisbing, S., Krapek, J., and Wright, G. (2017).
Brown, S. and Lugo, A.E. (1982). The storage and pro- A foundation of ecology rediscovered: 100 years of
duction of organic matter in tropical forests and succession on the William S. Cooper plots in Gla-
their role in the global carbon cycle. Biotropica 14: cier Bay, Alaska. Ecology 98: 1513–1523.
161–187. Buma, B., Weiss, S., Hayes, K., and Lucash, M. (2020).
Brown, C.L., McAlpine, R.G., and Kormanik, P.P. Wildland fire reburning trends across the US West
(1967). Apical dominance and form in woody suggest only short-term negative feedback and
plants: a reappraisal. Am. J. Bot. 54: 153–162. differing climatic effects. Environ. Res. Lett. 15:
Brundrett, M.C. (2002). Coevolution of roots and 034026.
mycorrhizas of land plants. New Phytol. 154: Bunn, A.G., Graumlich, L.J., and Urban, D.L. (2005).
275–304. Trends in twentieth-century tree growth at high
Brussard, P.F., Reed, J.M., and Tracy, C.R. (1998). Eco- elevations in the Sierra Nevada and White Moun-
system management: what is it really? Landsc. tains, USA. The Holocene 15: 481–488.
Urban Plan. 40: 9–20. Buol, S.W., Hole, F.D., and McCracken, R.J. (1980). Soil
Bryant, J.M. (1981). Phytochemical deterrence of Genesis and Classification. Ames: Iowa State Univ.
snowshoe hare browsing by adventitious shoots of Press 404 pp.
four Alaskan trees. Science 213: 889–890. Burger, D. (1972). Forest site classification in Canada.
Bryant, J.P., Chapin, F.S. III, Reichardt, P., and Clau- Mitt. Vereins forstl. Standortsk. Forstpflz. 21:
sen, T. (1985). Adaptation to resource availability 20–36.
as a determinant of chemical defense strategies in Burger, D. (1993). Revised site regions of Ontario: con-
woody plants. In: Chemically Mediated Interac- cepts, methodology and utility. Ontario For. Res.
tions Between Plants and Other Organisms. Recent Inst., Sault Ste. Marie, For. Res. Report No. 129.
Adv. Phytochemistry 19 (ed. G.A. Cooper-Driver, 24 pp.
T. Swain and E.E. Conn), 219–237. New York: Burke, I.C. (1989). Control of nitrogen mineraliza-
Plenum Press. tion in a sagebrush steppe landscape. Ecology 70:
Bryant, J.P., Provenze, F.D., Pastor, J. et al. (1991). Inter- 1115–1126.
actions between woody plants and browsing mam- Burke, M.K., Raynal, D., and Mitchell, M.J. (1991). Soil
mals mediated by secondary metabolites. Annu. nitrogen availability influences seasonal carbon
Rev. Ecol. Syst. 22: 431–446. allocation patterns in sugar maple (Acer saccha-
Bryson, R.A. (1966). Air masses, streamlines, and the rum). Can. J. For. Res. 22: 447–456.
boreal forest. Geogr. Bull. 8: 228–269. Burkhart, H.E., Avery, T.E., and Bullock, B.P. (2018).
Bryson, R.A. and Hare, F.K. (1974). The climates of Forest Measurements, 6e. Long Grove. IL: Waveland
North America. In: Climates of North America Press 434 pp.
(World Survey of Climatology, Vol. 11). (ed. R.A. Burley, J. (1966). Provenance variation in growth
Bryson and F.K. Hare), 1–47. New York: Elsevier. of seedling apices of Sitka spruce. For. Sci. 12:
Brzostek, E.R., Dragoni, D., Schmid, H.P. et al. (2014). 170–175.
Chronic water stress reduces tree growth and the Burns, G. P. (1923). Studies in tolerance of New
carbon sink of deciduous hardwood forests. Glob. England forest trees IV. Minimum light require-
Chang. Biol. 20: 2531–2539. ments referred to a definite standard. Univ. Vermont
Budde, K.B., Heuertz, M., Hernández-Serrano, A. et al. Agr. Exp. Sta. Bull 235.
(2014). In situ genetic association for serotiny, a Burns, G.R. (1942). Photosynthesis and absorption in
fire-related trait, in Mediterranean maritime pine blue radiation. Am. J. Bot. 29: 381–387.
(Pinus pinaster). New Phytol. 201: 230–241. Burns, R.M. and Honkala, B.H. (1990). Silvics of North
Buechling, A., Martin, P.H., Canham, C.D. et al. (2016). America, Vol. 2, Hardwoods. Washington, DC:
Climate drivers of seed production in Picea engel- U.S.D.A. For. Serv. Agr. Handbook 65 886 pp.
mannii and response to warming temperatures Burton, P.J., Balisky, A.C., Coward, L.P. et al. (1992).
in the southern Rocky Mountains. J. Ecol. 104: The value of managing for biodiversity. For. Chron.
1051–1062. 68: 225–237.
References 661
Busby, P.E., Motzkin, G., and Foster, D.R. (2008). flushing temperature and chilling. Bot. Gaz. 140:
Multiple and interacting disturbances lead to Fagus 223–231.
grandifolia dominance in coastal New England. J. Campbell, R.K., Pawuk, W.A., and Harris, A.S. (1989).
Torrey Bot. Soc. 135: 346–359. Microgeographic genetic variation of Sitka spruce
Büsgen, M. and Münch, E. (1929). The Structure and in southeastern Alaska. Can. J. For. Res. 19:
Life of Forest Trees, 3e (trans. T. Thomson). London: 1004–1013.
Chapman & Hall 436 pp. Campbell, C.R., Poelstra, J.W., and Yoder, A.D.
Busing, R.T. (1994). Canopy cover and tree regenera- (2018). What is speciation genomics? The roles
tion in old-growth cove forests of the Appalachian of ecology, gene flow, and genomic architecture
Mountains. Vegetatio 115: 19–27. in the formation of species. Biol. J. Linn. Soc.
Buttò, V., Deslauriers, A., Rossi, S. et al. (2020). The role 124: 561–583.
of plant hormones in tree-ring formation. Trees 34: Canada Inst. Forestry. (1992). The Forestry Chronicle
315–335. 68(1):21–120.
Byram, G.M. and Jemison, G.M. (1943). Solar radiation Canadell, J.G. and Raupach, M.R. (2008). Managing
and forest fuel moisture. J. Agric. Res. 67: 149–176. forests for climate change mitigation. Science 320:
Cain, S.A. (1935). Studies on virgin hardwood forests: 1456–1457.
III. Warren’s Woods, a beech-maple climax forest in Canadian Forest Ecosystem Classification System
Berrien County, Michigan. Ecology 16: 500–513. (CFEC) (2010). Natural Resources Canada.
Cain, S.A. (1939). The climax and its complexities. Am. Canadian Forest Service. Great Lakes Forestry
Midi. Nat. 21: 146–181. Centre, Sault Ste Marie, Ontario. Frontline Express
Cain, S.A. and de Oliveira Castro, G.M. (1959). Manual 38. 2pp.
of Vegetation Analysis. New York: Harper & Row Canadian Forest Service (2021). Mountain Pine
325 pp. Beetle factsheet. https://www.nrcan.gc.ca/forests/
Cain, M.L., Milligan, B.G., and Strand, A.E. (2000). fire-insects-disturbances/top-insects/13397#shr-pg0.
Long-distance seed dispersal in plant populations. Accessed August 25, 2021.
Am. J. Bot. 87: 1217–1227. Canham, C.D. (1985). Suppression and release during
Cajander, A. K. (1926). The theory of forest types. Ada canopy recruitment in Acer saccharum. Bull. Torrey
For. Fenn. 29. 108 pp. Bot. Club. 112: 134–145.
Calder, W.J., and St. Clair, S.B. (2012). Facilitation drives Canham, C.D. (1988). Growth and canopy architecture
mortality patterns along succession gradients of of shade-tolerant trees: response to canopy gaps.
aspen-conifer forests. Ecosphere 3: 1–11. Ecology 69: 786–795.
Callaham, R.Z. (1962). Geographic variability in growth Canham, C.D. and Loucks, O.L. (1984). Catastrophic
of forest trees. In: Tree Growth (ed. T.T. Kozlowski), windthrow in the presettlement forests of
311–325. New York: Ronald Press. Wisconsin. Ecology 65: 803–809.
Cameron, R.P. and Williams, D. (2011). Completing an Canham, C.D., Denslow, J.S., Platt, W.J. et al. (1990).
ecosystem classification system for Nova Scotia. Light regimes beneath closed canopies and tree-fall
Nat. Areas J. 31: 92–96. gaps in temperate and tropical forests. Can. J. For.
Campbell, R.S. (1955). Vegetational changes and Res. 20: 620–631.
management in the cutover longleaf pine-slash Canham, C.D., Finzi, A.C., Pacala, S.W., and Burbank,
pine area of the Gulf Coast. Ecology 36: 29–34. D.H. (1994). Causes and consequences of resource
Campbell, R.K. (1979). Genecology of Douglas-fir in heterogeneity in forests: interspecific variation in
a watershed in the Oregon Cascades. Ecology 60: light transmission by canopy trees. Can. J. For. Res.
1036–1050. 24: 337–349.
Campbell, R.K. (1986). Mapped genetic variation of Canham, C.D., Kobe, R.K., Latty, E.F., and Chazdon,
Douglas-fir to guide seed transfer in southwest R.L. (1999). Interspecific and intraspecific variation
Oregon. Silvae Genet. 35: 85–96. in tree seedling survival: effects of allocation to
Campbell, R.K. (1991). Soils, seed-zone maps, and phys- roots versus carbohydrate reserves. Oecologia 121:
iography: guidelines for seed transfer in Douglas-fir 1–11.
in southwestern Oregon. For. Sci. 37: 973–986. Cannell, M.G.R. and Smith, R.I. (1986). Climatic
Campbell, T.A. and Long, D.B. (2009). Feral swine warming, spring budburst, and forest damage on
damage and damage management in forested eco- trees. J. Appl. Ecol. 23: 177–191.
systems. For. Ecol. Manag. 257: 2319–2326. Cannon, H.L. (1960). The development of botanical
Campbell, R.K. and Sorensen, F.C. (1978). Effect of test methods of prospecting for uranium on the
environment on expression of clines and on delim- Colorado Plateau. U.S. Geol. Surv. Bull. 1085-A:
itation of seed zones in Douglas-fir. Theor. Appl. 1–50.
Genet. 51: 233–246. Cantor, L.F. and Whitham, T.G. (1989). Importance of
Campbell, R.K. and Sugano, A.I. (1979). Genecology of belowground herbivory: pocket gophers may limit
bud-burst phenology in Douglas-fir: response to aspen to rock outcrop refugia. Ecology 70: 962–970.
662 References
Carbó, M., Iturra, C., Correia, B. et al. (2019). Epige- Carneros, E., Yakovlev, I., Viejo, M. et al. (2017). The
netics in forest trees: keep calm and carry on. In: epigenetic memory of temperature during embryo-
Epigenetics in Plants of Agronomic Importance: genesis modifies the expression of bud burst-
Fundamentals and Applications (ed. R. Alvarez- related genes in Norway spruce epitypes. Planta
Venegas, C. De la Peña and J.A. Casas-Mollano), 246: 553–566.
381–403. New York: Springer. Carpenter, S., Caraco, N.F., Correll, D.L. et al. (1998).
Cardinale, B.J., Srivastava, D.S., Duffy, J.E. et al. (2006). Nonpoint pollution of surface waters with
Effects of biodiversity on the functioning of trophic phosphorus and nitrogen. Issues Ecol. 3: 1–12.
groups and ecosystems. Nature 443: 989–992. Carpenter, S., Walker, B., Andries, J.M., and Ablel, N.
Cardinale, B.J., Matulich, K.L., Hooper, D.U. et al. (2001). From metaphor to measurement: Resilience
(2011). The functional role of producer diversity in of what to what? Ecosystems 4: 765–781.
ecosystems. Am. J. Bot. 98: 572–592. Carpenter, S.R., Mooney, H.A., Agard, J. et al. (2009).
Cardinale, B.J., Duffy, J.E., Gonzalez, A. et al. (2012). Science for managing ecosystem services: beyond
Biodiversity loss and its impact on humanity. the millennium ecosystem assessment. Proc. Natl.
Nature 486: 59–67. Acad. Sci. 106: 1305–1312.
Carere, C. and Maestripieri, D. (2013). Animal Person- Carpenter, S.R., Cole, J.J., Pace, M.L. et al. (2011). Early
alities: Behavior, Physiology and Evolution. Chicago, warnings of regime shifts: a whole-ecosystem
IL: University of Chicago Press 507 pp. experiment. Science 332: 1079–1082.
Carey, A.B. and Johnson, M.L. (1995). Small mammals Carter, G.A. and Smith, W.K. (1985). Influence of shoot
in managed, naturally young, and old-growth for- structure on light interception and photosynthesis
ests. Ecol. Appl. 5: 336–352. in conifers. Plant Physiol. 79: 1038–1043.
Carlson, A.R., Sibold, J.S., Assal, T.J., and Negrón, J.F. Carter, R.E., Mackenzie, M.D., and Gjerstad, D.H.
(2017). Evidence of compounded disturbance (1999). Ecological land classification in the
effects on vegetation recovery following high- Southern Loam Hills of south Alabama. For. Ecol.
severity wildfire and spruce beetle outbreak. PLoS Manag. 114: 395–404.
One 12: e0181778. Casal, J.J. (2012). Shade avoidance. Arabidopsis Book
Carmean, W.H. (1967). Soil refinements for predicting 10: e0157.
black oak site quality in southeastern Ohio. Proc. Castello, J.D., Leopold, D.J., and Samllidge, P.J. (1995).
Soil Sci. Soc. Am. 31: 805–810. Pathogens, patterns and processes in forest ecosys-
Carmean, W.H. (1970a). Site quality for eastern hard- tems. BioScience 45: 16–24.
woods. In: The Silviculture of Oaks and Associated Certini, G. (2005). Effects of fire on properties of forest
Species. USDA For. Serv. Res. Paper NE-144. North- soils: a review. Oecologia 143: 1–10.
eastern For. Exp. Sta., Upper Darby, PA. Ceulemans, R.J. and Saugier, B. (1991). Photosynthesis.
Carmean, W.H. (1970b). Tree height-growth patterns in In: Physiology of Trees (ed. A.S. Raghavendra),
relation to soil and site. In: Tree Growth and Forest 21–50. New York: Wiley.
Soils (ed. C.T. Youngberg and C.B. Davey), 499–512. Chabot, B.F. and Hicks, D.J. (1982). The ecology of leaf
Corvallis: Oregon State Univ. Press. life spans. Annu. Rev. Ecol. Syst. 11: 233–257.
Carmean, W.H. (1975). Forest site quality evaluation in Chambers, J.Q., Fisher, J.I., Zeng, H. et al. (2007).
the United States. Adv. Agron. 27: 209–269. Hurricane Katrina’s carbon footprint on Gulf Coast
Carmean, W.H. (1977). Site classification for northern forests. Science 318: 1107.
forest species. In Proc. Symp Intensive Culture of Chambless, L.F. and Nixon, E.S. (1975). Woody
Northern Forest Types, USDA For. Serv. Gen. Tech. vegetation—soil relations in a bottomland forest of
Report NE-29. Northeastern For. Exp. Sta., Upper west Texas. Tex. J. Sci. 26: 407–416.
Darby, PA. Chapin, F.S. III, Van Cleve, K., and Tryon, P.R. (1986).
Carmean, W.H. (1996). Site-quality evaluation, site- Relationship of ion absorption to growth rate in
quality maintenance, and site-specific management taiga trees. Oecologia 69: 238–242.
for forest land in northwest Ontario. Ontario Chapin, F.S., III, Bloom, A.J., and Field, C.B. (1987).
Ministry Nat. Res., Northwest Sci. and Technology Plant response to multiple environmental factors.
Unit, NWST Tech. Report TR-105, Thunder Bay, BioScience 37: 49–57.
ON. 121 pp. Chapin, F.S. III, Walker, L.R., Fastie, C.L., and
Carmean, W.H., Hazenberg, G., and Niznowski, G.P. Sharman, L.C. (1994). Mechanisms of primary
(2001). Polymorphic site index curves for jack pine succession following deglaciation at Glacier Bay,
in Northern Ontario. For. Chron. 77: 141–150. Alaska. Ecol. Monogr. 64: 149–175.
Carmean, W.H., Hazenberg, G., and Deschamps, K.C. Chapin, F.S. III, Torn, M.S., and Tateno, M. (1996).
(2006). Polymorphic site index curves for black Principles of ecosystem sustainability. Am. Nat.
spruce and trembling aspen in northwest Ontario. 148: 1016–1037.
For. Chron. 82: 231–242.
References 663
Chapin, F.S. III, Zavaleta, E.S., Eviner, V.T. et al. (2000). Management for Parks and Wilderness (ed. J.K.
Consequences of changing biodiversity. Nature Agee and D.R. Johnson), 62–86. Seattle: Univ.
405: 234–242. Washington Press.
Chapin, F.S. III, Matson, P.A., and Vitousek, P.M. Christensen, N.L. and Peet, R.K. (1981). Secondary
(2012). Principles of Terrestrial Ecosystem Ecology, forest succession on the North Carolina piedmont.
2e. New York: Springer 544 pp. In: Forest Succession: Concepts and Applications
Chapman, H.H. (1932). Is the longleaf pine a climax? (ed. D.C. West, H.H. Shugart and D.B. Botkin),
Ecology 13: 328–335. 230–245. New York: Springer-Verlag.
Chapman, S.K., Whitham, T.G., and Powell, M. (2006). Christensen, K.M. and Whitham, T.G. (1991). Indirect
Herbivory differentially alters plant litter dynamics of herbivore mediation of avian seed dispersal in
evergreen and deciduous trees. Oikos 114: 566–574. pinyon pine. Ecology 72: 536–542.
Chapman, E.L., Chambers, J.Q., Ribbeck, K.F. et al. Christensen, K.M. and Whitham, T.G. (1993). Impact
(2008). Hurricane Katrina impacts on forest trees of insect herbivores on competition between birds
of Louisiana’s Pearl River basin. For. Ecol. Manag. and mammals for pinyon pine seeds. Ecology
256: 883–889. 74: 2270–2278.
Charlesworth, B. and Charlesworth, D. (2010). Elements Christensen, N.L., Agee, J.K., Brussard, P.F. et al.
of Evolutionary Genetics. New York: W.H. Freeman (1989). Interpreting the Yellowstone fires of 1988.
and Company 768 pp. Bioscience 39: 678–685.
Charney, N.D., Babst, F., Poulter, B. et al. (2016). Christensen, N.L., Bartuska, A.M., Brown, J.H. et al.
Observed forest sensitivity to climate implies large (1996). The report of the Ecological Society of
changes in 21st century North American forest America committee on the scientific basis for eco-
growth. Ecol. Lett. 19: 1119–1128.peterson. system management. Ecol. Appl. 6: 665–691.
Chauchard, S., Pille, G., and Carcaillet, C. (2006). Large Christy, H.R. (1952). Vertical temperature gradients
herbivores control the invasive potential of non- in a beech forest in central Ohio. Ohio J. Sci. 52:
native Austrian black pine in a mixed deciduous 199–209.
Mediterranean forest. Can. J. For. Res. 36: 1047–1053. Chung, H.H. and Barnes, R.L. (1977). Photosynthate
Chazdon, R. (1988). Sunflecks and their importance to allocation in Pinus taeda. I. Substrate requirements
forest understory plants. Adv. Ecol. Res. 18: 1–63. for synthesis of shoot biomass. Can. J. For. Res.
Chen, J., Franklin, J.F., and Spies, T.A. (1992). Vegeta- 7: 106–111.
tion responses to edge environments in old-growth Chung, M.Y., Son, S., Herrando-Moraira, S. et al. (2020).
Douglas-fir forests. Ecol. Appl. 2: 387–396. Incorporating differences between genetic diversity
Chen, J., Franklin, J.F., and Spies, T.A. (1995). of trees and herbaceous plants in conservation
Growing-season microclimatic gradients from strategies. Conserv. Biol. 34: 1142–1151.
clearcut edges into old-growth Douglas-fir forests. Churchill, D.J., Larson, A.J., Dahlgreen, M.C. et al.
Ecol. Appl. 5: 74–86. (2013). Restoring forest resilience: from reference
Chick, J.H., Cosgriff, R.J., and Gittinger, L.S. (2003). spatial patterns to silvicultural prescriptions and
Fish as potential dispersal agents for floodplain monitoring. For. Ecol. Manag. 291: 442–457.
plants: first evidence in North America. Can. J. Cicatelli, A., Todeschini, V., Lingua, G. et al.
Fish. Aquat. Sci. 60: 1437–1439. (2013). Epigenetic control of heavy metal stress
Chivian, E. and Bernstein, A. (ed.) (2008). Sustaining response in mycorrhizal versus non-mycorrhizal
Life: How Human Health Depends on Biodiversity. poplar plants. Environ. Sci. Pollut. Res. 21:
Oxford, UK: Oxford University Press 568 pp. 1723–1737.
Chomicki, G., Weber, M., Antonelli, A. et al. (2019). The Cieslar, A. (1895). Über die Erblichkeit des Zuwachs-
impact of mutualisms on species richness. Trends vermogens bei den Waldbäumen. Centralbl. gesam.
Ecol. Evol. 34: 698–711. Forstw. 21:7–29.
Christensen, N.L. (1977). Changes in structure, pattern Cieslar, A. (1899). Neues aus dem Gebiete der fostlichen
and diversity associated with climax forest matura- Zuchtwahl. Centbl. gesam. Forstw. 25:99–117.
tion in Piedmont, North Carolina. Am. Midl. Nat. Civitello, D.J., Cohen, J., Fatima, H. et al. (2015).
97: 176–188. Biodiversity inhibits parasites: broad evidence
Christensen, N.L. (1988a). Vegetation of the south- for the dilution effect. Proc. Natl. Acad. Sci. 112:
eastern Coastal Plain. In: North American Ter- 8667–8671.
restrial Vegetation (ed. M.G. Barbour and W.D. Clark, J.S. (1989). Ecological disturbance as a renewal
Billings), 317–363. New York: Cambridge Univ. process: theory and application to fire history.
Press. Oikos 56: 17–30.
Christensen, N.L. (1988b). Succession and natural Clark, J.S. (1990). Fire and climate change during the
disturbance: paradigms, problems, and pres- last 750 years in northwestern Minnesota. Ecol.
ervation of natural ecosystems. In: Ecosystem Monogr. 60: 135–169.
664 References
Clark, K.L., Skowronski, N., and Hom, J. (2010). Inva- Colautti, R.I., Ricciardi, A., Grigorovich, I.A., and
sive insects impact forest carbon dynamics. Glob. MacIsaac, H.J. (2004). Is invasion success explained
Chang. Biol. 16: 88–101. by the enemy release hypothesis? Ecol. Lett. 7:
Clason, T.R. and Sharrow, S.H. (2000). Silvopastoral 721–733.
practices. In: North American Agroforestry: An Cole, L.C. (1958). The ecosphere. Sci. Am. 198: 83–92.
Integrated Science and Practice (ed. H.E. Garrett, Cole, W.E. and Amman, G.D. (1980). Mountain Pine
W.J. Rietveld and R.F. Fisher), 119–147. Madison, Beetle Dynamics in Lodgepole Pine Forests, Part
Wisconsin, USA: American Society of Agronomy. 1: Course of An Infestation. USDA For. Serv. Gen.
Classen, A.T., Norby, R.J., Sides, K.E., and Weltzin, J.F. Tech. Rep. INT-89. Intermountain Res. Sta., Ogden,
(2010). Climate change alters seedling emergence UT. 56 pp.
and establishment in an old-field ecosystem. PLoS Cole, D.W. and Johnson, D.W. (1980). Mineral cycling
One 5: e13476. in tropical forests. In C. T. Youngberg (ed.), Forest
Clawson, M. (1978). The concept of multiple use for- Soils and Land Use. Proc. 5th North American
estry. Environ. Law 8: 281–308. Forest Soils Conf. Colorado State Univ., Fort
Cleland, D.T., Avers, P.E., McNab, W.H. et al. (1997). Collins.
National hierarchical framework of ecological Cole, D.W. and Rapp, M. (1981). Elemental cycling in
units. In: Ecosystem Management Applications for forest ecosystems. In: Dynamic Properties of Forest
Sustainable Forest and Wildlife Resources (ed. M.S. Ecosystems, Int. Biol. Programme 23 (ed. D.E.
Boyce and A. Haney), 181–200. New Haven, CT: Reichle), 341–409. Cambridge: Cambridge Univ.
Yale University Press. Press.
Cleland, E.E., Chuine, I., Menzel, A. et al. (2007). Shift- Coley, P.D. (1980). Effects of leaf age and plant life his-
ing plant phenology in response to global change. tory patterns on herbivory. Nature 284: 545–546.
Trends Ecol. Evol. 22: 357–365. Coley, P.D. (1987). Interspecific variation in plant
Clements, F.E. (1905). Research Methods in Ecology. anti-herbivore properties: the role of habitat quality
Univ. Publ. Co., Lincoln, NB. 334 pp. (reprinted and rate of disturbance. New Phytol. 106: 251–263.
1977, Arno Press, New York). Coley, P.D., Bryant, J.P., and Chapin, F.S. III (1985).
Clements, F.E. (1916). Plant succession: an analysis of Resource availability and plant anti-herbivore
the development of vegetation. Carneg. Inst. Wash. defense. Science 230: 895–899.
Publ. 242. 512 pp. Collinge, S.K. (2009). Ecology of Fragmented Land-
Clements, F.E. (1936). Nature and structure of the scapes. Baltimore, MD: Johns Hopkins University
climax. J. Ecol. 24: 252–284. Press 360 pp.
Clements, F.E. (1949). Dynamics of Vegetation: Selec- Colyvan, M., Linquist, S., Grey, W. et al. (2009).
tions from the Writings of Frederic E. Clements, Ph. Philosophical issues in ecology: recent trends and
D. The H. W. Wilson Co., New York. 296 pp. future directions. Ecol. Soc. 14: 22.
Cline, M. (1997). Concepts and terminology of apical Conkle, M.T. (1973). Growth data for 29 years from the
dominance. Am. J. Bot. 84: 1064. California elevational transect study of ponderosa
Cline, A.C. and Spurr, S.H. (1942). The virgin upland pine. For. Sci. 19: 31–39.
forest of central New England. Harvard For. Bull. Conkle, M.T. and Critchfield, W.B. (1988). Genetic
21.51 pp. variation and hybridization of ponderosa pine. In
Clinton, B.D. and Baker, C.R. (2000). Catastrophic D. Baumgartner and J.E. Lotan (eds.), Symp. Proc.
windthrow in the southern Appalachians: charac- Ponderosa Pine, The Species and Its Management.
teristics of pits and mounds and initial vegetation Washington State Univ., Dept. Nat. Res. Sciences.
responses. For. Ecol. Manag. 126: 51–60. Pullman.
Clout, M.N. and Williams, P.A. (ed.) (2009). Invasive Connell, J.H. (1978). Diversity in tropical rain forests
Species Management: A Handbook of Principles and and coral reefs. Science 199: 1302–1310.
Techniques. Oxford, UK: Oxford University Press Connell, J.H. and Slatyer, R.O. (1977). Mechanisms of
320 pp. succession in natural communities and their roles
Coates, P. (2007). American Perceptions of Immigrant in community stability and organization. Am. Nat.
and Invasive Species. Berkeley, CA: University of 111: 1119–1144.
California Press 266 pp. Conner, W.H., Mihalia, I., and Wolfe, J. (2002). Tree
Coetzee, B.W.T., Gaston, K.J., and Chown, S.L. (2014). community structure and changes from 1987 to
Local scale comparisons of biodiversity as a test for 1999 in three Louisiana and three South Carolina
global protected area ecological performance: A forested wetlands. Wetlands 22: 58–70.
meta-analysis. PLoS One 9: e105824. Cook, R.E. (1979). Asexual reproduction: a further
Coile, T.S. (1937). Distribution of forest tree roots in consideration. Am. Nat. 113: 769–772.
North Carolina Piedmont soils. J. For. 35: 247–257. Cook, E.R. and Kairiukstis, L.A. (ed.) (1990). Methods
Coile, T.S. (1952). Soil and the growth of forests. Adv. of Dendrochronology: Applications in the Environ-
Agron. 4: 330–398. mental Sciences. New York: Springer 406 pp.
References 665
Coomes, D.A. and Grubb, P.J. (2000). Impacts of root Costanza, R., De Groot, R., Sutton, P. et al. (2014).
competition in forests and woodlands: a theoret- Changes in the global value of ecosystem services.
ical framework and review of experiments. Ecol. Glob. Environ. Chang. 26: 152–158.
Monogr. 70: 171–207. Cottam, G. and Mclntosh, R.P. (1966). Vegetational
Coop, J.D., Parks, S.A., McClernan, S.R., and Hols- continuum. Science 152: 546–547.
inger, L.M. (2016). Influences of prior wildfires Cottingham, K.L., Brown, B.L., and Lennon, T.J. (2001).
on vegetation response to subsequent fire in a Biodiversity may regulate the temporal variability
reburned Southwestern landscape. Ecol. Appl. 26: of ecological systems. Ecol. Lett. 4: 72–85.
346–354. Coutts, M.P. and Grace, J. (ed.) (1995). Wind and Trees.
Coop, J.D., Parks, S.A., Stevens-Rumann, C.S. et al. Cambridge: Cambridge Univ. Press 485 pp.
(2020). Wildfire-driven forest conversion in west- Coutts, M.P. and Nicoll, B.C. (1991). Development of
ern North American landscapes. Bioscience 70: the surface roots of trees. In: L’Arbre, Biologie et
659–673. Development (ed. C. Edelin), 61–70. Montpellier:
Cooper, W.S. (1913). The climax forest of Isle Royale, Naturalia Monspeliensia.
Lake Superior, and it development. Bot. Gaz. 55 Covell, R.R. and McClurkin, D.C. (1967). Site index
(1–14): 189–235. of loblolly pine on Ruston soils in the southern
Cooper, W.S. (1923). The recent ecological history of Coastal Plain. J. For. 65: 263–264.
Glacier Bay, Alaska. Ecology 4: 93–128, 223–246, Covington, W.W. and Moore, M.M. (1994). Postsettle-
355–365. ment changes in natural fire regimes and forest
Cooper, W.S. (1931). A third expedition to Glacier Bay, structure: ecological restoration of old-growth pon-
Alaska. Ecology 12: 61–95. derosa pine forests. J. Sustain. For. 2: 153–181.
Cooper, W.S. (1939). A fourth expedition to Glacier Bay, Covington, W.W., Everett, R.L., Steele, R. et al. (1994).
Alaska. Ecology 20: 130–155. Historical and anticipated changes in forest
Cooper, C.F. (1960). Changes in vegetation, structure, ecosystems of the inland west of the United States.
and growth of southwestern pine forests since J. Sustain. For. 2: 13–63.
white settlement. Ecol. Monogr. 30: 129–164. Cowles, H.C. (1899). The ecological relations of the veg-
Cooper, C.F. (1961). The ecology of fire. Sci. Am. 204 etation on the sand dunes of Lake Michigan. Bot.
(4): 150–160. Gaz. 27: 95–117, 167–202, 281–308, 361–391.
Cooper, S.V., Neiman, K.E., and Roberts, D.W. (1991). Cowles, H.C. (1901). The physiographic ecology
Forest habitat types of northern Idaho: a second of Chicago and vicinity; a study of the origin,
approximation. USDA For. Serv. Gen. Tech. Report development, and classification of plant societies.
INT-236. Int. Res. Sta., Ogden, UT. 143 pp. Bot. Gaz. 31: 73–108, 145–182.
Corenblit, D., Tabacchi, E., Steiger, J., and Gurnell, A.M. Cowles, H.C. (1911). The causes of vegetative cycles.
(2007). Reciprocal interactions and adjustments Bot. Gaz. 51: 161–183.
between fluvial landforms and vegetation dynamics Cragg, J.B. (1953). Book review of natural communities
in river corridors: a review of complementary by LR. Dice. Bull. Inst. Biol. 1: 3.
approaches. Earth-Sci. Rev. 84: 56–86. Craine, J.M. and Reich, P.B. (2005). Leaf-level light
Corenblit, D., Steiger, J., Gurnell, A.M. et al. (2009). compensation points in shade-tolerant woody
Control of sediment dynamics by vegetation as a seedlings. New Phytol. 166: 710–713.
key function driving biogeomorphic succession Crane, P.R. and Lidgard, S. (1989). Angiosperm
within fluvial corridors. Earth Surf. Process. Landf. diversification and paleolatitudinal gradients in
34: 1790–1810. Cretaceous floristic diversity. Science 246: 675–678.
Corenblit, D., Steiger, J., Charrier, G. et al. (2016). Crankshaw, W.B., Qadir, S.A., and Lindsey, A.A. (1965).
Populus nigra L. establishment and fluvial land- Edaphic controls of tree species in presettlement
form construction: biogeomorphic dynamics Indiana. Ecology 46: 688–698.
within a channelized river. Earth Surf. Process. Crawford, D.J. (1974). A morphological and chemical
Landf. 41: 1276–1292. study of Populus acuminata Rydberg. Brittonia 26:
Cornwell, W.K., Cornelissen, J.H., Amatangelo, K. et al. 74–89.
(2008). Plant species traits are the predominant Crins, W. J., Gray, P. A., Uhlig, P. W. C., and Wester,
control on litter decomposition rates within biomes M. C. (2009). The Ecosystems of Ontario, Part
worldwide. Ecol. Lett. 11: 1065–1071. I: Ecozones and Ecoregions. Ontario Ministry
Correa, S.B., Winemiller, K.O., Lopez-Fernandez, H., of Natural Resources, Peterborough Ontario,
and Galetti, M. (2007). Evolutionary perspectives Inventory, Monitoring and Assessment, SIB TER
on seed consumption and dispersal by fishes. Bio- IMA TR-01, 71pp.
science 57: 748–756. Crisafulli, C.M. and Dale, V.H. (ed.) (2018). Ecological
Costanza, R., d’Arge, R., De Groot, R. et al. (1997). The Responses at Mount St. Helens: Revisited 35 years
value of the world’s ecosystem services and natural after the 1980 Eruption. New York, NY: Springer
capital. Nature 387: 253–260. 351 pp.
666 References
Critchfield, W. B. (1957). Geographic variation in Pinus Curtis, R.O., DeMars, D.J., and Herman, F.R. (1974a).
contorta. Maria Moors Cabot Found. Publ. No. 3. Which dependent variable in site index—height-
Harvard Univ., Cambridge, MA. 118 pp. age regressions. For. Sci. 20: 74–87.
Critchfield, W.B. (1985). The late Quaternary history of Curtis, R.O., Herman, F.R., and DeMars, D.J. (1974b).
lodgepole and jack pines. Can. J. For. Res. 15: 749–772. Height growth and site index for Douglas-fir in
Crone, E.E. and Rapp, J.M. (2014). Resource depletion, high-elevation forests of the Oregon-Washington
pollen coupling, and the ecology of mast seeding. Cascades. For. Sci. 20: 307–315.
Ann. NY Acad. Sci. 1322: 21–34. Cutter, A. (2019). A Primer of Molecular Population
Crookston, N.L. and Dixon, G.E. (2005). The Forest Genetics. Oxford, UK: Oxford University Press
Vegetation Simulator: A review of its structure, 272 pp.
content, and applications. Comput. Electron. Agric. Cwynar, L.C. and MacDonald, G.M. (1987).
49: 60–80. Geographical variation of lodgepole pine in rela-
Crouch, G.L. (1976). Wild animal damage to forests in tion to population history. Am. Nat. 129: 463–469.
the United States and Canada. In: Proc. XVI IUFRO Cypert, E. (1973). Plant succession on burned areas in
World Congress, Div. II, 468–478. Oslo, Norway. Okefenokee Swamp following the fires of 1954 and
Crowley, K.F., Lovett, G.M., Arthur, M.A., and 1955. In Proc. Annual Tall Timbers Fire Ecology Conf.
Weathers, K.C. (2016). Long-term effects of 12:199–217. Tall Timbers Res. Sta., Tallahassee, FL.
pest-induced tree species change on carbon and Daehler, C.C. (2003). Performance comparisons of
nitrogen cycling in northeastern US forests: A mod- co-occurring native and alien invasive plants:
eling analysis. For. Ecol. Manag. 372: 269–290. implications for conservation and restoration.
Crutzen, P.J. and Goldhammer, J.G. (ed.) (1993). Fire in Annu. Rev. Ecol. Evol. Syst. 34: 183–211.
the Environment: The Ecological, Atmospheric, and Daily, G.C. (ed.) (1997). Nature’s Services: Societal
Climatic Importance of Vegetation Fires. New York: Dependence on Natural Ecosystems. Washington,
Wiley 416 pp. DC: Island Press 412 pp.
Csilléry, K., Lalagüe, H., Vendramin, G.G. et al. (2014). Daily, G.C., Soderquist, S., Aniyar, S. et al. (2000). The
Detecting short spatial scale local adap tation and value of nature and the nature of value. Science
epistatic selection in climate-related candidate 289: 395–396.
genes in European beech (Fagus sylvatica) popula- Dale, V.H. and Crisafulli, C.M. (2018). Ecological
tions. Mol. Ecol. 23: 4696–4708. responses to the 1980 eruption of Mount St.
Cuevas, E. and Medina, E. (1988). Nutrient dynamics Helens: key lessons and remaining questions. In:
within Amazonian forest ecosystems. I. Nutrient Ecological Responses at Mount St. Helens: Revisited
flux in fine litter fall and efficiency of nutrient utili- 35 Years after the 1980 Eruption (ed. C. Crisafulli
zation. Oecologia 68: 466–472. and V. Dale), 1–18. New York, NY: Springer.
Culmsee, H., Schmidt, M., Schmiedel, I. et al. (2014). Dale, V.H. and Denton, E.M. (2018). Plant succession
Predicting the distribution of forest habitat types on the Mount St. Helens debris-avalanche deposit
using indicator species to facilitate systematic and the role of non-native species. In: Ecological
conservation planning. Ecol. Indic. 37: 131–144. Responses at Mount St. Helens: Revisited 35 Years
Cuny, H.E., Rathgeber, C.B., Frank, D. et al. (2014). after the 1980 Eruption (ed. C. Crisafulli and V.
Kinetics of tracheid development explain conifer Dale), 149–164. New York, NY: Springer.
tree-ring structure. New Phytol. 203: 1231–1241. Dale, V.H., Joyce, L.A., McNulty, S. et al. (2001). Cli-
Currie, D.J. and Paquin, V. (1987). Large-scale bio- mate change and forest disturbances. Bioscience 51:
geographical patterns of species richness of trees. 723–734.
Nature 329: 326–327. Dale, V.H., Crisafulli, C.M., and Swanson, F.J. (2005a).
Curtis, J.T. (1956). The modification of mid-latitude 25 years of ecological change at Mount St. Helens.
grasslands and forests by man. In: Man’s Role in Science 308: 961–962.
Changing the Face of the Earth (ed. W.L. Thomas Dale, V.H., Campbell, D.R., Adams, W.M. et al.
Jr.), 721–736. Chicago: Univ. Chicago Press. (2005b). Plant succession on the Mount St. Helens
Curtis, J.T. (1959). The Vegetation of Wisconsin. Madi- debris-avalanche deposit. In: Ecological Responses
son: Univ. Wisconsin Press 657 pp. to the 1980 Eruption of Mount St. Helens (ed. V.H.
Curtis, R.O. (1972). Yield tables past and present. J. For. Dale, F.J. Swanson and C.M. Crisafulli), 59–73.
70: 28–32. Springer. New York.
Curtis, J.D. and Lersten, N.R. (1974). Morphology, Dale, V.H., Swanson, F.J., and Crisafulli, C.M. (2005c).
seasonal variation and function of resin glands on Disturbance, survival, and succession: under-
buds and leaves of Populus deltoides (Salicaceae). standing ecological responses to the 1980 eruption
Am. J. Bot. 61: 835–845. of Mount St. Helens. In: Ecological Responses to
Curtis, J.T. and Mclntosh, R.P. (1951). An upland forest the 1980 Eruption of Mount St. Helens (ed. V.H.
continuum in the prairie-forest border region of Dale, F.J. Swanson and C.M. Crisafulli). New York:
Wisconsin. Ecology 32: 476–496. Springer.
References 667
D’Amato, A.W. and Palik, B.J. (2021). Building on the Daubenmire, R. F. (1989). The roots of a concept. In
last “new” thing: exploring the compatibility of D. E. Ferguson, P. Morgan, F. D. Johnson (comp.),
ecological and adaptation silviculture. Can. J. For. Proceedings—Land Classifications Based on Vegeta-
Res. 51: 172–180. tion: Applications for Resource Management. USDA
Damman, A. W. H. (1964). Some forest types of central For. Serv. Gen. Tech. Report INT-257, Int. Res. Sta.,
Newfoundland and their relation to environmental Ogden, UT.
factors. For. Sci. Monogr. 8. 62 pp. Daubenmire, R.F., and Daubenmire, J.B. (1968). Forest
Damman, A. W. H. (1975). Permanent changes in the vegetation of eastern Washington and northern
chronosequence of a boreal forest habitat induced Idaho. Washington Agric. Exp. Sta., Tech. Bull. 60.
by natural disturbances. In W. Schmidt (ed.), Int. 104 pp.
Proc. Symp. Sukzessionsforschung. Int. Verein. Veg- Davidson, E.A. and Janssens, I.A. (2006). Temperature
etationsk. J. Cramer, Vaduz. sensitivity of soil carbon decomposition and feed-
Daniel, T.W., Helms, J.A., and Baker, F.S. (1979). Princi- backs to climate change. Nature 440: 165–173.
ples of Silviculture. New York: McGraw-Hill 500 pp. Davidson, E.A. and Swank, W.T. (1987). Factors limiting
Dansereau, P. (1968). The continuum concept of vegeta- denitrification in soils from mature and dis-
tion: responses. Bot. Rev. 34: 253–332. turbed southeastern hardwood forests. For. Sci. 33:
D’Antonio, C.M. and Vitousek, P.M. (1992). Biological 135–144.
invasions by exotic grasses, the grass/fire cycle, and Davidson, E.A., Belk, E., and Boone, R.D. (1998). Soil
global change. Annu. Rev. Ecol. Syst. 23: 63–87. water content and temperature as independent or
Darabant, A., Rai, P.B., Tenzin, K. et al. (2007). Cattle confounded factors controlling soil respiration in
grazing facilitates tree regeneration in a conifer a temperate mixed hardwood forest. Glob. Chang.
forest with palatable bamboo understory. For. Ecol. Biol. 4: 217–227.
Manag. 252: 73–83. Davidson, A.M., Jennions, M., and Nicotra, A.B. (2011).
Darley-Hill, S. and Johnson, W.C. (1981). Dispersal of Do invasive species show higher phenotypic plas-
acorns by blue jays (Cyanocita cristata). Oecologia ticity than native species and, if so, is it adaptive? A
50: 231–232. meta-analysis. Ecol. Lett. 14: 419–431.
D’Arrigo, R.D., Kaufmann, R.K., Davi, N. et al. (2004). Davies, W.J. and Kozlowski, T.T. (1974). Stomatal
Thresholds for warming-induced growth decline at responses of five woody angiosperms to light inten-
elevational tree line in the Yukon territory. Canada. sity and humidity. Can. J. Bot. 52: 1525–1534.
Glob. Biogeochem. Cycles 18: GB3021. Davis, M.B. (1981). Quaternary history and the sta-
D’Arrigo, R., Wilson, R., Liepert, B., and Cherubini, P. bility of forest communities. In: Forest Succession:
(2008). On the “Divergence Problem” in northern Concepts and Application (ed. E.C. West, H.H.
forests: a review of the tree-ring evidence and pos- Shugart and D.B. Botkin), 132–153. New York:
sible causes. Glob. Planet. Chang. 60: 289–305. Springer-Verlag.
Dasmann, R.F. (1984). Environmental Conservation, 5e. Davis, M.B. (1983a). Holocene vegetational history of
New York: Wiley 486 pp. the eastern United States. In: The Late Quaternary
Daubenmire, R.F. (1936). The “Big Woods” of Minne- Environments of the United States. Vol. 2: The
sota: its structure, and relation to climate, fire, and Holocene (ed. H.E. Wright and S. Porter), 166–181.
soils. Ecol. Monogr. 6: 223–268. Minneapolis: Univ. Minnesota Press.
Daubenmire, R.F. (1952). Forest vegetation of northern Davis, M.B. (1983b). Quaternary history of deciduous
Idaho and adjacent Washington, and its bearing on forests of eastern North America and Europe. Ann.
concepts of vegetation classification. Ecol. Monogr. Missouri Bot. Card. 70: 550–563.
22: 301–330. Davis, M.B. (1986). Climatic instability, time lags, and
Daubenmire, R.F. (1954). Alpine timberlines in the community disequilibrium. In: Community Ecology
Americas and their interpretation. Butl. Univ. Bot. (ed. J. Diamond and T.J. Case), 269–284. New York:
Stud. 11: 119–136. Harper and Row.
Daubenmire, R.F. (1966). Vegetation: identification of Davis, M.B. (1987). Invasions of forest communities
typal communities. Science 151: 291–298. during the Holocene: beech and hemlock in the
Great Lakes region. In: Colonization, Succession,
Daubenmire, R.F. (1968). Plant Communities.
and Stability (ed. A.J. Gray, M.L. Crawley and P.J.
New York: Harper & Row 300 pp.
Edwards), 373–393. London: Blackwell Sci. Publ.
Daubenmire, R. F. (1970). Steppe vegetation of Wash-
Davis, P.H. and Heywood, V.H. (1963). Principles of
ington. Tech. Bull 62. Wash. State Agr. Exp. Sta.,
Angiosperm Taxonomy. New York: D. Van Nostrand
Pullman, WA. 131 pp.
558 pp.
Daubenmire, R.F. (1974). Plants and Environment: A Text-
Davis, R.B. and Jacobson, G.L. Jr. (1985). Late glacial
book of Plant Autecology. New York: Wiley 422 pp.
and early Holocene landscapes in northern New
Daubenmire, R.F. (1976). The use of vegetation in
England and adjacent areas of Canada. Quat. Res.
assessing the productivity of forest lands. Bot. Rev.
23: 341–368.
42: 115–143.
668 References
Davis, M.B., Woods, K.D., Webb, S.L., and Futyma, Del Vecchio, T.A., Gehring, C.A., Coob, N.S., and
R.P. (1986). Dispersal versus climate: expansion Whitham, T.G. (1993). Negative effects of scale
of Fagus and Tsuga into the Upper Great Lakes insect herbivory on the ectomoycorrhizae of
region. Vegetatio 67: 93–103. juvenile pinyon pine. Ecology 74: 2297–2302.
Davis, M.A., Chew, M.K., Hobbs, R.J. et al. (2011). Don’t Delcourt, H.R. (1980). Late quaternary vegetation
judge species on their origins. Nature 474: 153–154. history of the eastern Highland Rim and adjacent
Davis, K.T., Dobrowski, S.Z., Higuera, P.E. et al. (2019). Cumberland Plateau of Tennessee. Ecol. Monogr.
Wildfires and climate change push low-elevation 49: 225–280.
forests across a critical climate threshold for tree Delcourt, H.R. (2003). Forests in Peril: Tracking Decid-
regeneration. Proc. Natl. Acad. Sci. 116: 6193–6198. uous Trees from Ice-Age Refuges into the Greenhouse
Day, F.P. Jr. and Monk, C.D. (1974). Vegetation patterns World. Lincoln: University of Nebraska Press
on a southern Appalachian watershed. Ecology 55: 244 pp.
1064–1074. Delcourt, H.R. and Delcourt, P.A. (1974). Primeval
Dayton, P.K. (1972). Toward an understanding of magnolia—holly—beech climax in Louisiana.
community resilience and the potential effects of Ecology 55: 638–644.
enrichments to the benthos at McMurdo Sound, Delcourt, P.A. and Delcourt, H.R. (1977). The Tunica
Antarctica. In: Proceedings of the Colloquium on Hills, Louisiana—Mississippi: late-glacial locality
Conservation Problems in Antarctica (ed. B.C. for spruce and deciduous forest species. Quat. Res.
Parker), 81–96. Lawrence, KS: Allen Press. 7: 218–237.
De Bodt, S., Maere, S., and Van de Peer, Y. (2005). Delcourt, P.A. and Delcourt, H.R. (1979). Late Pleis-
Genome duplication and the origin of angio- tocene and Holocene distributional history of the
sperms. Trends Ecol. Evol. 20: 591–597. deciduous forest in the southeastern United States.
De Mazancourt, C. and Loreau, M. (2000). Effect of her- Veröff. Geobot. Inst. ETH, Stift. Rübel 68: 79–107.
bivory and plant species replacement on primary Delcourt, P.A., and Delcourt, H.R. (1981). Vegetation
production. Am. Nat. 155: 735–754. maps for eastern North America: 40,000 yr B. P. to
De Mazancourt, C., Isbell, F., Larocque, A. et al. (2013). the present. In R. Romans (ed.), Proc. 1980 Geo-
Predicting ecosystem stability from community botany Conference, Plenum, New York.
composition and biodiversity. Ecol. Lett. 16: Delcourt, P.A. and Delcourt, H.R. (1987). Long-Term
617–625. Forest Dynamics of the Temperate Zone. New York:
De Villemereuil, P., Gaggiotti, O.E., Mouterde, M., and Springer-Verlag 439 pp.
Till-Bottraud, I. (2016). Common garden experi- Delcourt, H.R. and Delcourt, P.A. (1988). Quaternary
ments in the genomic era: new perspectives and landscape ecology: relevant scales in space and
opportunities. Heredity 116: 249–254. time. Landsc. Ecol. 2: 23–44.
DeBano, L.F. (1969). The relationship between heat Delcourt, H.R. and Delcourt, P.A. (1991). Quaternary
treatment and water repellency in soils. In: Ecology, a Paleoecological Perspective. New York:
Water-Repellent Soils (ed. L.F. DeBano and J. Chapman and Hall 242 pp.
Letey), 265–279. Riverside: Univ. Calif. Delisle, Z.J. and Parshall, T. (2018). The effects of
DeBano, L.F., Osborn, J.F., Krammes, J.S., and Letey, J. oriental bittersweet on native trees in a New
Jr. 1967. Soil wettability and wetting agents . . . our England floodplain. Northeast. Nat. 25: 188–196.
current knowledge of the problem. USDA For. Serv. Denevan, W.M. (1992). The pristine myth: the landscape
Res. Paper PSW-43. Pacific Southwest For. and Rge. of the Americas in 1492. Ann. Assoc. Am. Geogr. 82:
Exp. Sta., Berkeley, CA 13 pp. 369–385.
DeBano, L.F., Savage, S.M., and Hamilton, D.A. (1976). Denton, S.R. and Barnes, B.V. (1988). An ecological
The transfer of heat and hydrophobic substances climatic classification of Michigan: a quantitative
during burning. Soil Sci. Soc. Am. J. 40: 779–782. approach. For. Sci. 34: 119–138.
DeBano, L.F., Neary, D.G., and Ffolliott, P.F. (1998). Fire D’Eon, R.G., Glenn, S.M., Parfitt, I., and Fortin, M.-J.
Effects on Ecosystems. New York: Wiley 352 pp. (2002). Landscape connectivity as a function
Decker, J.P. (1959). A system for analysis of forest of scale and organism vagility in a real forested
succession. For. Sci. 5: 154–157. landscape. Conserv. Ecol. 6: 10.
Deevey, E.S. Jr. (1949). Biogeography of the Pleistocene: Despain, D.G. (1990). Yellowstone Vegetation: Conse-
part I: Europe and North America. Geol. Soc. Am. quences of Environment and History in a Natural
Bull. 60: 1315–1416. Setting. Boulder, Colorado, USA: Rinehart 239 pp.
Del Moral, R. (1983). Initial recovery of subalpine veg- Desprez-Loustau, M.L., Marcais, B., Nageleisen, L.M.
etation on Mount St. Helens. Am. Midl. Nat. 109: et al. (2006). Interactive effects of drought and
72–80. pathogens in forest trees. Ann. For. Sci. 63: 597–612.
Del Moral, R. and Chang, C.C. (2015). Multiple assess- Dewan, S., Mijnsbrugge, K.V., De Frenne, P. et al.
ments of succession rates on Mount St. Helens. (2018). Maternal temperature during seed matu-
Plant Ecol. 216: 165–176. ration affects seed germination and timing of bud
References 669
set in seedlings of European black poplar. For. Ecol. Donato, D.C., Harvey, B.J., and Turner, M.G. (2016).
Manag. 410: 126–135. Regeneration of montane forests 24 years after
DeWoody, J., Rowe, C.A., Hipkins, V.D., and Mock, K.E. the 1988 Yellowstone fires: a fire-catalyzed shift in
(2008). “Pando” lives: molecular genetic evidence lower treelines? Ecosphere 7: e01410.
of a giant aspen clone in Central Utah. West. N. Am. Donnegan, J.A. and Rebertus, A.J. (1999). Rates and
Nat. 68: Article 8. mechanisms of subalpine forest succession along
Diamond, J.S., Epstein, J.M., Cohen, M.J. et al. (2021). an environmental gradient. Ecology 80: 1370–1384.
A little relief: ecological functions and autogenesis Dooley, S.R. and Treseder, K.K. (2012). The effect of
of wetland microtopography. Wiley Interdiscip. Rev. fire on microbial biomass: a meta-analysis of field
Water 8: e1493. studies. Biogeochemistry 109: 49–61.
Diaz, J.A., Carbonell, R., Virgos, E. et al. (2000). Effects Doolittle, W.T. (1957). Site index of scarlet and black
of forest fragmentation on the distribution of the oak in relation to southern Appalachian soil and
lizard Psammodromus algirus. Anim. Conserv. 3: topography. For. Sci. 3: 114–124.
235–240. Downs, A.A. (1938). Glaze damage in the
Dieterich, H. (1970). Die Bedeutung der Vegetation- birch-beech-maple-hemlock type of Pennsylvania
skunde für die forstliche Standortskunde. Der and New York. J. For. 36: 63–70.
Biologieunterricht. 6: 48–60. Downs, R.J. (1962). Photocontrol of growth and dor-
Dieterich, J.H. and Swetnam, T.W. (1984). Dendrochro- mancy in woody plants. In: Tree Growth (ed. T.T.
nology of a fire-scarred ponderosa pine. For. Sci. 30: Kozlowski), 131–149. New York: Ronald Press.
238–247. Downs, A.A. and McQuilken, W.E. (1944). Seed pro-
Dieterich, H., Müller, S., and Schlenker, G. (1970). duction of southern Appalachian oaks. J. For. 53:
Urwald von morgen. Stuttgart, Germany: Verlag 439–441.
Eugen Ulmer 174 pp. Drew, T.J. and Flewelling, J.W. (1977). Some recent
DiFazio, S.P., Slavov, G.T., Burczyk, J. et al. (2004). Gene Japanese theories of yield-density relationships and
flow from tree plantations and implications for their application to Monterey pine plantations. For.
transgenic risk assessment. In: Forest Biotechnology Sci. 23: 517–534.
for the 21st Century (ed. C. Walter and M. Carson). Drew, T.J. and Flewelling, J.W. (1979). Stand density
Kerala, India: Research Signpost 446 pp. management: an alternative approach and its
Divíšek, J., Chytrý, M., Beckage, B. et al. (2018). Sim- application to Douglas-fir plantations. For. Sci. 25:
ilarity of introduced plant species to native ones 518–532.
facilitates naturalization, but differences enhance Drobyshev, I., Goebel, P.C., Hix, D.M. et al. (2008).
invasion success. Nat. Commun. 9: 1–10. Pre-and post-European settlement fire history of
Dobbs, R.C. (1976). White spruce seed dispersal in red pine dominated forest ecosystems of Seney
central British Columbia. For. Chron. 52: 225–228. National Wildlife Refuge, Upper Michigan. Can. J.
Dobzhansky, T. (1951). Genetics and the Origin of For. Res. 38: 2497–2514.
Species, 3e. New York: Columbia Univ. Press Drury, W.H. and Nisbet, I.C.T. (1973). Succession. J.
364 pp. Arnold Arbor. 54: 331–368.
Dobzhansky, T. (1968). Adaptedness and fitness. Du Pisani, J.A. (2006). Sustainable development—
In: Population Biology and Evolution (ed. R.C. historical roots of the concept. Environ. Sci. 3:
Lewontin), 109–121. Syracuse, New York: Syracuse 83–96.
Univ. Press. Duffield, J.W. and Snyder, E.B. (1958). Benefits from
Dodd, J.C., Burton, C.C., Burns, R.G., and Jefferies, hybridizing American forest trees. J. For. 56:
P. (1987). Phosphatase activity associated with 809–815.
the roots and rhizosphere of plants infected with Duffy, D.C. and Meier, A.J. (1992). Do Appalachian her-
vesicular-arbuscular mycorrhizal fungi. New Phytol. baceous understories ever recover from clearcut-
107: 163–172. ting? Conserv. Biol. 6: 196–201.
Dodds, K.J., Aoki, C.F., Arango-Velez, A. et al. (2018). DuMerle, P. (1988). Phenological resistance of oaks to
Expansion of southern pine beetle into northeastern the green oak leafroller, Tortrix viridana (Lepidop-
forests: management and impact of a primary bark tera: Tortricidae). In: Mechanisms of Woody Plant
beetle in a new region. J. For. 116: 178–191. Defenses Against Insects: Search for Pattern (ed.
Dolan, B.J. and Parker, G.R. (2005). Ecosystem W.J. Mattson, J. Levieux and C. Bernard-Dagan),
classification in a flat, highly fragmented region of 215–226. New York: Springer-Verlag.
Indiana, USA. For. Ecol. Manag. 219: 109–131. Dumont, E., Bahrman, N., Goulas, E. et al. (2011). A
Domec, J.C., Rivera, L.N., King, J.S. et al. (2013). proteomic approach to decipher chilling response
Hemlock woolly adelgid (Adelges tsugae) infesta- from cold acclimation in pea (Pisum sativum L.).
tion affects water and carbon relations of eastern Plant Sci. 180: 86–98.
hemlock (Tsuga canadensis) and Carolina hemlock Dumroese, R.K., William, M.I., Stanturf, J.A., and
(Tsuga caroliniana). New Phytol. 199: 452–463. Clair, J.B.S. (2015). Considerations for restoring
670 References
temperate forests of tomorrow: Forest restoration, Ecological Land Classification. Ecol. Land Classif.
assisted migration, and bioengineering. New For. Series, No. 23, Can. Wildlife Serv., Env. Canada,
46: 947–964. Ottawa, Ontario. 119 pp. and map.
Dunn, P.H. and DeBano, L.F. (1977). Fire’s effect on Edwards, C.A. (1974). Macroarthropods. In: Biology
biological and chemical properties of chaparral of Plant Litter Decomposition (ed. C.H. Dickinson
soils. In H.A. Mooney, C.E. Conrad. (tech. coords.). and G.J.F. Pugh), 533–554. New York: Academic
Proceedings of a Symposium on Environmental Press.
Conservation: Fire and Fuel Management in Edwards, P.J. and Abivardi, C. (1998). The value of bio-
Mediterranean Ecosystems. August 4–5, 1988, Palo diversity: where ecology and economy blend. Biol.
Alto, CA. Washington, D.C. USDA For. Serv. WO-3. Conserv. 83: 239–246.
Dusenge, M.E., Duarte, A.G., and Way, D.A. (2019). Edwards, C.A. and Bohlen, P.J. (1996). Biology and
Plant carbon metabolism and climate change: Ecology of Earthworms. London: Chapman and
elevated CO2 and temperature impacts on photo- Hall 426 pp.
synthesis, photorespiration and respiration. New Egler, F.E. (1954). Vegetation science concepts I. initial
Phytol. 221: 32–49. floristic composition, a factor in old-field vegeta-
Duvigneaud, P. (1946). La variabilité des associations tion development. Vegetatio 4: 412–417.
végétales. Bull. Soc. R. Bot. Belg. 78: 107–134. Egler, F. E. (1968). The contumacious continuum. In P.
Dwyer, L.M. and Merriam, G. (1981). Influence of topo- Dansereau (ed.), The continuum concept of vegeta-
graphic heterogeneity on deciduous litter decompo- tion: responses. Bot. Rev. 34:253–332.
sition. Oikos 37: 228–237. Egler, F.E. (1977). The Nature of Vegetation, its
Dyksterhuis, E.J. (1949). Condition and management of Management and Mismanagement. Bridgewater,
range land based on quantitative ecology. J. Range CO: Conn. Cons. Assoc. 527 pp.
Manag. 2: 104–115. Eis, S. (1976). Association of western white pine cone
Dyksterhuis, E.J. (1958). Ecological principles in range crops with weather variables. Can. J. For. Res. 6:
evaluation. Bot. Rev. 24: 253–272. 6–12.
Dyrness, C.T. (1976). Effect of wildfire on soil wetta- Ekberg, I., Eriksson, G., and Gormling, I. (1979). Photo-
bility in the high Cascades of Oregon. USDA For. periodic reactions in conifer species. Holarct. Ecol.
Serv. Res. Paper PNW-202. Pacific Northwest For. 2: 255–263.
and Rge. Exp. Sta., Portland, OR. 18 pp. Ekins, P., Simon, S., Deutsch, L. et al. (2003). A frame-
Easley, A.T., Passineau, J.F., and Driver, B.L. (comps.). work for the practical application of the concepts
(1990). The use of wilderness for personal growth, of critical natural capital and strong sustainability.
therapy, and education. USDA. For. Serv., Gen. Ecol. Econ. 44: 165–185.
Tech. Rep. RM-193. Rocky Mountain Res. Sta., Ft. El-Kassaby, Y.A., Dunsworth, B.G., and Krakowski, J.
Collins, CO. 197 pp. (2003). Genetic evaluation of alternative silvicul-
Easterling, D.R., Kunkel, K.E., Arnold, J.R. et al. (2017). tural systems in coastal montane forests: western
Precipitation change in the United States. In: hemlock and amabilis fir. Theor. Appl. Genet. 107:
Climate Science Special Report: Fourth National Cli- 598–610.
mate Assessment, Volume I (ed. D.J. Wuebbles, D.W. Elkington, J. (2004). Enter the triple bottom line. In:
Fahey, K.A. Hibbard, et al.), 207–230. Washington, The Triple Bottom Line: Does it all Add Up? (ed. A.
DC, USA: U.S. Global Change Research Program. Henriques and J. Richardson). London: Routledge
Eckenwalder, J.E. (1977). North American cotton- Publishng 209 pp.
woods (Populus, Salicaceae) of sections Abaso and Elkins, J.W. and Rosen, R. (1989). Summary Report
Aigeiros. J. Arnold Arbor. 58: 193–208. 1988: Geophysical Monitoring for Climate Change.
Eckenwalder, J.E. (1996). Systematics and evolution of ERL, Boulder, CO: NOAA 142 pp.
Populus. In: Biology of Populus and its Implications Ellenberg, H. (1968). Wege der Geobotanik zum Ver-
for Management and Conservation (ed. R.F. Stettler, standnis der Pflanzendecke. Naturwissenschaften
H.D. Bradshaw Jr., P.E. Heilman and T.M. Hinck- 55: 462–470.
ley). Ottawa: NRC Res. Press 539 pp. Ellenberg, H. (1974). Zeigerwerte der Gefässpfanzen
Eckert, A.J., Bower, A.D., Gonzalez-Martinez, S.C. et al. Mitteleuropas. Scripta Geobot. 9: 1–97.
(2010). Back to nature: ecological genomics of Ellenberg, H. (1988). Vegetation Ecology of Central
loblolly pine (Pinus taeda, Pinaceae). Mol. Ecol. 19: Europe, 4e. Cambridge: Cambridge Univ. Press
3789–3805. 731 pp.
Eckert, A.J., Maloney, P.E., Vogler, D.R. et al. (2015). Ellis, M., von Dohlen, C. D., Anderson, J. E., and
Local adaptation at fine spatial scales: an example Romme, W. H. (1994). Some important factors
from sugar pine (Pinus lambertiana, Pinaceae). affecting density of lodgepole pine seedlings fol-
Tree Genet. Genomes 11: 42–58. lowing the 1988 Yellowstone fires. In E. G. Despain
Ecoregions Working Group (1989). Ecoclimatic regions (ed.), Conf. Proc. Plants and Their Environments.
of Canada, first approximation. Ecoregions Tech. Report NPS/NRYELL/NRTR-93/XX. USDI,
Working Group of the Canada Committee on Nat. Park Service.
References 671
Ellison, A.M., Bank, M.S., Clinton, B.D. et al. (2005). mid-Appalachian forested watersheds: role of
Loss of foundation species: consequences for the insect defoliation. Water Resour. Res. 34: 2005–2016.
structure and dynamics of forested ecosystems. Evans, K.L., Greenwood, J.J., and Gaston, K.J. (2005).
Front. Ecol. Environ. 3: 479–486. Dissecting the species–energy relationship. Proc. R.
Ellison, A.M., Barker Plotkin, A.A., Khalid, S., and Soc. B Biol. Sci. 272: 2155–2163.
(2016). Foundation species loss and biodiversity Ewel, K.C. (1990). Swamps. In: Ecosystems of Florida
of the herbaceous layer in New England forests. (ed. R.L. Myers and J.J. Ewel). Orlando: Univ.
Forests 7: 9. Central Florida Press.
Ellstrand, N.C. (1992). Gene flow among seed plant Eyre, F.H. (1980). Forest Cover Types of the United States
populations. In: Population Genetics of Forest Trees and Canada. Bethesda, MD: Soc. Am. For 148 pp.
(ed. W.T. Adams, S.H. Strauss, D.L. Copes and A.R. Faegri, K. and Iverson, J. (1989). Textbook of Pollen
Griffin), 241–256. Boston, MA: Kluwer. Analysis, 4e. New York: Wiley 338 pp.
Ellsworth, D.S. and Reich, P.B. (1993). Canopy struc- Fahrig, L. (2003). Effects of habitat fragmentation
ture and vertical patterns of photosynthesis and on biodiversity. Annu. Rev. Ecol. Evol. Syst. 34:
related leaf traits in a deciduous forest. Oecologia 487–515.
96: 69–178. Fairbridge, R.W. (ed.) (1972). The Encyclopedia of Geo-
Elston, J.J. and Hewitt, D.G. (2010). Intake of mast by chemistry and Environmental Sciences. New York:
wildlife in Texas and the potential for competition Van Nostrand Reinhold 1321 pp.
with wild boars. Southwest. Nat. 55: 57–66. Fall, M.W., Avery, M.L., Campbell, T.A. et al. (2011).
Endara, M.-J. and Coley, P.D. (2011). The resource avail- Rodents and other vertebrate invaders in the
ability hypothesis revisited: a meta-analysis. Funct. United States. In: Biological Invasions: Economic
Ecol. 25: 89–398. and Environmental Costs of Alien Plant, Animal,
Engeman, R.M., Stevens, A., Allen, J. et al. (2007). Feral and Microbe Species, 2e (ed. D. Pimentel), 381–410.
swine management for conservation of an imper- Boca Raton, FL: CRC Press.
iled wetland habitat: Florida’s vanishing seepage FAO (1980). Poplars and Willows in Wood Production
slopes. Biol. Conserv. 134: 440–446. and Land Use. Rome: Food and Agr. Organization
Engler, A. (1905). Einfluss der Provenienz des of the United Nations 328 pp.
Samens auf die Eigenschaften der forstlichen Farhat, P., Hidalgo, O., Robert, T. et al. (2019). Poly-
Holzgewächse. Mitt, schweiz. Centralanst. forstl. ploidy in the conifer genus Juniperus: an unexpect-
Versuchsw. 8:81–236. edly high rate. Front. Plant Sci. 10: 676.
Engler, A. (1908). Tatsachen, Hypothsen und Irrtümer Farley, J. (2012). Ecosystem services: the economics
auf dem Gebiete der Samenprovenienz-Frage. debate. Ecosyst. Serv. 1: 40–49.
Forstwiss. Centralbl. 30: 295–314. Fassnacht, K.S. and Gower, S.T. (1997). Interrelation-
Epanchin-Niell, R.S. and Liebhold, A.M. (2015). Ben- ships between the edaphic and stand characteris-
efits of invasion prevention: effect of time lags, tics, leaf area index and above-ground net primary
spread rates, and damage persistence. Ecol. Econ. productivity of upland forest ecosystems in north
116: 146–153. Central Wisconsin. Can. J. For. Res. 27: 1058–1067.
Erickson, A., Nitschke, C., Coops, N. et al. (2015). Past- Fastie, C. (1995). Causes and ecosystem consequences
century decline in forest regeneration potential of multiple pathways of primary succession at Gla-
across a latitudinal and elevational gradient in cier Bay, Alaska. Ecology 76: 1899–1916.
Canada. Ecol. Model. 313: 94–102. Fearer, T.M., Norman, G.W., Pack, J.C. Sr. et al. (2008).
Esau, K. (1977). Anatomy of Seed Plants. New York: Influence of physiographic and climatic factors on
Wiley 550 pp. spatial patterns of acorn production in Maryland
Eschner, A.R. and Patric, J.H. (1982). Debris avalanches and Virginia, USA. J. Biogeogr. 35: 2012–2025.
in eastern upland forests. J. For. 80: 343–347. Fearnside, P.M. (2021). The intrinsic value of Amazon
Eschrich, W., Burchardt, R., and Essiamah, S. (1989). biodiversity. Biodivers. Conserv. 30: 1199–1202.
The induction of sun and shade leaves of the Fedrowitz, K., Koricheva, J., Baker, S.C. et al. (2014).
European beech (Fagus sylvatica L.): anatomical Can retention forestry help conserve biodiversity?
studies. Trees 3: 1–10. A meta-analysis. J. Appl. Ecol. 51: 1669–1679.
Eshelman, K.R., Wagner, R.E., and Secrist, F.M. (1989). Feeny, P. (1976). Plant apparency and chemical defense.
Vegetation classification—problems, principles, In: Biochemical Interaction between Plants and
and proposals. In D. E. Ferguson, P. Morgan, Insects (ed. J.W. Wallace and R.L. Mansell), 1–40.
and F. D. Johnson (comps.), Symp. Proc. Land Boston, MA: Springer.
Classifications Based on Vegetation: Applications for Fei, S., Desprez, J.M., Potter, K.M. et al. (2017). Diver-
Resource Management. USDA For. Serv. Gen. Tech. gence of species responses to climate change. Sci.
Report INT-257. Intermountain Res. Sta., Ogden, Adv. 3: e1603055.
UT. 315 pp.
Fei, S., Jo, I., Guo, Q. et al. (2018). Impacts of climate on
Eshleman, K.N., Morgan, R.P., Webb, J.R. et al. (1998). the biodiversity-productivity relationship in natural
Temporal patterns of nitrogen leakage from forests. Nat. Commun. 9: 1–7.
672 References
Fenneman, N.M. (1931). Physiography of Western Management. USDA For. Serv. Gen. Tech. Report
United States. New York: McGraw-Hill 534 pp. INT-257. Intermountain Res. Sta., Ogden, UT.
Fenneman, N.M. (1938). Physiography of Eastern United Fisher, F.M., Parker, L.W., Anderson, J.P., and Whit-
States. New York: McGraw-Hill 691 pp. ford, W.G. (1987). Nitrogen mineralizaiton in a
Ferguson, D.E., Morgan, P., and Johnson, F.D. (comps.) desert soil: interacting effects of soil moisture and
(1989). Proceedings—Land Classifications Based on nitrogen fertilizer. Soil Sci. Soc. Am. J. 1033–1041.
Vegetation: Applications for Resource Management. Fisher, J.I., Mustard, J.F., and Vadeboncoeur, M.A.
USDA For. Serv. Gen. Tech. Report INT-257. Inter- (2006). Green leaf phenology at Landsat resolution:
mountain Res. Sta., Ogden, UT. 315 pp. scaling from the field to the satellite. Remote Sens.
Field, C. and Mooney, H.A. (1986). The photosy Environ. 100: 265–279.
nthesis-nitrogen relationship in wild plants. In: Fisichelli, N., Wright, A., Rice, K. et al. (2014a). First-
On the Economy of Plant Form and Function (ed. year seedlings and climate change: species-specific
T.J. Givnish), 25–55. Cambridge: Cambridge Univ. responses of 15 North American tree species. Oikos
Press. 123: 1331–1340.
Findlay, B.F. (1976). Recent developments in Fisichelli, N.A., Frelich, L.E., and Reich, P.B. (2014b).
eco-climatic classifications. In: Ecological Temperate tree expansion into adjacent boreal
(Biophysical) Land Classification in Canada (ed. forest patches facilitated by warmer temperatures.
J. Thie and G. Ironside), 121–127. Environment Ecography 37: 152–161.
Canada, Ottawa: Lands Directorate. Fitter, A.H. (1987). An architectural approach to the
Finegan, B. (1984). Forest succession. Nature 312: comparative ecology of plant root systems. New
109–114. Phytol. 106 (Supp): 61–77.
Finlay, R.D., Ek, H., Odham, G., and Söderström, B. Fitter, A.H. and Hay, R.K.M. (1987). Environmental
(1989). Uptake, translocation and assimilation of Physiology of Plants, 2e. New York: Academic Press
nitrogen from 15N-labeled ammonium and nitrate 423 pp.
sources by intact ectomycorrhizal systems of Fitter, A.H. and Stickland, T.R. (1991). Architectural
Fagus sylvatica and Paxillus involutus. New Phytol. analysis of plant root systems 2. Influence of
113: 47–55. nutrient supply on architecture in contrasting plant
Finney, M.A. and Martin, R.E. (1989). Fire history in a species. New Phytol. 118: 383–389.
Sequoia sempervirens forest at Salt Point State Park, Flanagan, P.W. and Van Cleve, K. (1983). Nutrient
California. Can. J. For. Res. 19: 1451–1457. cycling in relation to decomposition and organic
Finnigan, J.J. and Brunet, Y. (1995). Turbulent airflow matter quality in taiga ecosystems. Can. J. For. Res.
in forests on flat and hilly terrain. In: Wind and 13: 795–817.
Trees (ed. M.P. Coutts and J. Grace), 3–40. Cam- Flannigan, M.D. (1993). Fire regime and the abundance
bridge: Cambridge Univ. Press. of red pine. Int. J. Wildland Fire 3: 241–247.
Finzi, A.C., Raymer, P.C., Giasson, M.A., and Orwig, Fleischner, T.L. (1994). Ecological costs of livestock
D.A. (2014). Net primary production and soil respi- grazing in western North America. Conserv. Biol. 8:
ration in New England hemlock forests affected by 629–644.
the hemlock woolly adelgid. Ecosphere 5 (8): 1–16. Flint, H.R. (1930). Fire as a factor in the management of
Fiorucci, A.S. and Fankhauser, C. (2017). Plant strat- north Idaho national forests. Northwest Sci. 4: 12–15.
egies for enhancing access to sunlight. Curr. Biol. Flint, H.L. (1972). Cold hardiness of twigs of Quercus
27: R931–R940. rubra L. as a function of geographic origin. Ecology
Firestone, M.K. and Davidson, E.A. (1989). Microbial 53: 1163–1170.
basis of NO and N2O production and consump- Floate, K.D. and Whitham, T.G. (1993). The “hybrid
tion in soil. In: Exchange of Trace Gases between bridge” hypothesis: host shifting via plant hybrid
Terrestrial Ecosystems and the Atmosphere (ed. M.O. swarms. Am. Nat. 141: 651–662.
Andreae and D.S. Schimel). New York: Wiley. Floate, K.D., Kearsley, M.J.C., and Whitham, T.G.
Fischer, W.C. and Bradley, A. F. (1987). Fire ecology of (1993). Elevated herbivory in plant hybrid zones:
western Montana forest habitat types. USDA For. Chrysomela confluens, Populus and phenological
Serv. Gen. Tech. Report INT-223. Int. Res. Sta., sinks. Ecology 74: 2056–2065.
Odgen, UT. 95 pp. Flower, C.E., Knight, K.S., and Gonzalez-Meler, M.A.
Fischer, J., Brosi, B., Daily, G. et al. (2008). Should agri- (2013). Impacts of the emerald ash borer (Agrilus
cultural policies encourage land sparing or wildlife- planipennis Fairmaire) induced ash (Fraxinus spp.)
friendly farming? Ecol. Environ. 6: 380–385. mortality on forest carbon cycling and successional
Fisher, W.C. (1989). The fire effects information system: dynamics in the eastern United States. Biol. Inva-
a comprehensive vegetation knowledge base. sions 15: 931–944.
In D.E. Ferguson, P. Morgan, and F.D. Johnson Foggin, G.T. III and DeBano, L.F. (1971). Some geo-
(comps.), Proceedings—Land Classifications graphic implications of water-repellent soils. Prof.
Based on Vegetation: Applications for Resource Geogr. 23: 347–350.
References 673
Ford-Robertson, F.C. (ed.) (1983). Terminology of Fralish, J.S., Crooks, F.B., Chambers, J.L., and Harty,
Forest Science Technology Practice and Products, F.M. (1991). Comparison of presettlement, second-
2nd Printing. Washington, D.C: Soc. Am. Foresters growth and old-growth forest on six site types in
370 pp. the Illinois Shawnee Hills. Am. Midl. Nat. 125:
Forman, R.T.T. (1995). Land Mosaics, the Ecology of 294–309.
Landscapes and Regions. Cambridge: Cambridge Fralish, J.S., Mclntosh, R.P., and Loucks, O.L. (eds.).
Univ. Press 632 pp. (1993). John T. Curtis, Fifty Years of Wisconsin Plant
Forman, R.T.T. and Godron, M. (1986). Landscape Ecology. Wise. Acad. Sci., Arts & Letters, Madison.
Ecology. New York: Wiley New York. 619 pp. 339 pp.
Forrester, D.I., Ammer, C., Annighöfer, P.J. et al. (2018). Franklin, J.F. (1968). Cone production by upper-slope
Effects of crown architecture and stand structure conifers. USDA For. Serv. Res. Paper PNW-60.
on light absorption in mixed and monospecific Pacific Northwest For. and Rge. Exp. Sta., Portland,
Fagus sylvatica and Pinus sylvestris forests along a OR. 21 pp.
productivity and climate gradient through Europe. Franklin, E.C. (1970). Survey of Mutant Forms and
J. Ecol. 106: 746–760. Inbreeding Depression in Species of the Family Pina-
Foster, D.R. (1988a). Disturbance history, community ceae. Vol. 61, USDA For. Serv, Southeastern For.
organization and vegetation dynamics of the old- Exp. Sta., Asheville, NC. 21 pp.
growth Pisgah forest, southwestern New Hamp- Franklin, J.F. (1990). Biological legacies: a critical
shire, U.S.A. J. Ecol. 76: 105–134. management concept from Mount St. Helens.
Foster, D.R. (1988b). Species and stand response to In Trans. Fifty-fifth North American Wildlife and
catastrophic wind in central New England, U.S.A. Natural Resources Conference. Wildlife Manage.
J. Ecol. 76: 135–151. Inst., Washington, D.C.
Foster, D.R. and Boose, E.R. (1992). Patterns of forest Franklin, J.F. (1993). Preserving biodiversity: species,
damage resulting from catastrophic wind in central ecosystems, or landscapes? Ecol. Appl. 3: 202–205.
New England, USA. J. Ecol. 80: 70–98. Franklin, J.F. (1995). Sustainability of managed
Foster, D.R. and Boose, E.R. (1995). Hurricane distur- temperate forest ecosystems. In: Defining and
bance regimes in temperate and tropical forest Measuring Sustainability: The Biogeophysical
ecosystems. In: Wind and Trees (ed. M.P. Coutts Foundations (ed. M. Munasinghe and W. Shearer),
and J. Grace), 305–339. Cambridge: Cambridge 355–385. Washington, D.C: World Bank.
Univ. Press. Franklin, J.F. (1997). Ecosystem management: an
Foster, D.R., Knight, D.H., and Franklin, J.F. (1998). overview. In M. Boyce (ed.), Proc. Symp. Ecosystem
Landscape patterns and legacies resulting from large, Management: Applications for Sustainable Forest and
infrequent forest disturbances. Ecosystems 1: 497–510. Wildlife Resources. Yale Univ. Press, New Haven, CT.
Fotis, A.T., Morin, T.H., Fahey, R.T. et al. (2018). Forest Franklin, J.F. (2003). Challenges to temperate forest
structure in space and time: biotic and abiotic stewardship – focusing on the future. In: Towards
determinants of canopy complexity and their Forest Sustainability (ed. D.B. Lindenmayer and J.F.
effects on net pri mary productivity. Agric. For. Franklin), 1–13. Washington, DC: Island Press.
Meterol. 250: 181–191. Franklin, J.F. and Dyrness, C.T. (1973). Natural vegeta-
Fowler, D.P. (1965a). Effects of inbreeding in red pine, tion of Oregon and Washington. USDA For. Serv.
Pinus resinosa Ait. II. Pollination studies. Silvae Gen. Tech. Rep. PNW-8, Pacific Northwest For. Rge.
Genet. 14: 12–23. Exp. Sta., Portland OR. 417 pp.
Fowler, D.P. (1965b). Effects of inbreeding in red pine, Franklin, J.F. and Dyrness, C.T. (1980). Natural Vegeta-
Pinus resinosa Ait. IV. Comparison with other tion of Oregon and Washington. Corvallis: Oregon
Northeastern Pinus species. Silvae Genet. 14: 76–81. Univ. Press 452 pp.
Fowler, D. (1980). Removal of sulphur and nitrogen Franklin, J.F. and Johnson, K.N. (2012). A restora-
compounds from the atmosphere in rain and by tion framework for federal forests in the Pacific
dry deposition. In: Ecological Effects of Acid Pre- northwest. J. For. 110: 429–439.
cipitation (ed. D. Drablos and A. Tollan), 22–32. Franklin, J.F. and MacMahon, J.A. (2000). Messages
Oslo, Norway: SNSF (Acid Precipitation-Effects on from a mountain. Science 288: 1183–1184.
Forests and Fish) Project. Franklin, J.F., Cromack, K. Jr., Denison, W. et al. (1981).
Fowler, D.P. and Mullin, R.E. (1977). Upland-lowland Ecological characteristics of old-growth Douglas-fir
ecotypes not well developed in black spruce in forests. USDA For. Serv. Gen. Tech. Report PNW-
northern Ontario. Can. J. For. Res. 7: 35–10. 118. Pacific Northwest For. and Rge. Exp. Sta.,
Fox, J.T. (1977). Alternation and coexistence of tree Corvallis OR. 48 pp.
species. Am. Nat. 111: 69–89. Franklin, J.F., MacMahon, J.A., Swanson, F.J., and
Foy, C.D., Chaney, O., and White, M.C. (1978). The Sedell, J.R. (1985). Ecosystem responses to the
physiology of metal toxicity of plants. Ann. Rev. eruption of Mount St. Helens. Natl. Geogr. Res. 1:
Plant Physiol. 29: 511–566. 196–215.
674 References
Franklin, J.F., Shugart, H.H., and Harmon, M.E. (1987). Frissell, S.S. Jr. (1973). The importance of fire as a
Tree death as an ecological process. Bioscience 37: natural ecological factor in Itasca State Park, Min-
550–556. nesota. Quat. Res. 3: 397–407.
Franklin, J.F., Frenzen, P.M., and Swanson, F.J. (1988). Fritts, H. (1976). Tree Rings and Climate. New York:
Re-creation of ecosystems at Mount St. Helens: Academic Press 567 pp.
contrasts in artificial and natural approaches. In: Froelich, N.J. and Schmid, H.P. (2006). Flow divergence
Rehabilitating Damaged Ecosystems, vol. 2 (ed. and density flows above and below a deciduous
J. Cairns Jr.), 287–334. Boca Raton, FL: CRC Press. forest: part II. Below-canopy thermotopographic
Franklin, J.F., Berg, D.R., Thornburgh, D.A., and flows. Agric. For. Meteorol. 138: 29–43.
Tappeiner, J.C. (1997). Alternative silvicultural Frye, R.J. and Quinn, J.A. (1979). Forest development in
approaches to timber harvesting: variable retention relation to topography and soils on a floodplain of
harvest systems. In: Creating a Forestry for the the Raritan River, New Jersey. Bull. Torrey Bot. Club
21st Century (ed. K.A. Kohm and J.F. Franklin), 106: 334–345.
111–139. Washington, D.C: Island Press. Fuller, W.H., Shannon, S., and Burgess, P.S. (1955).
Franklin, J.F., Lindenmayer, D., MacMahon, J.A. et al. Effect of burning on certain forest soils of northern
(2000). Threads of continuity. Conserv. Biol. Pract. Arizona. For. Sci. 1: 44–50.
1: 8–16. Funk, J.L. and Vitousek, P.M. (2007). Resource-use
Franklin, J.F., Johnson, K.N., and Johnson, D.L. (2018). efficiency and plant invasion in low-resource sys-
Ecological Forest Management. Long Grove, IL: tems. Nature 446: 1079–1081.
Waveland Press 646 pp. Furnier, G.R., Knowles, P., Clyde, M.A., and Dancik,
Fraser, E.C., Lieffers, V.J., and Landhäusser, S.M. B.P. (1987). Effects of avian seed dispersal on the
(2006). Carbohydrate transfer through root grafts to genetic structure of whitebark pine populations.
support shaded trees. Tree Physiol. 26: 1019–1023. Evolution 4: 607–612.
Frederickson, M.E. (2017). Mutualisms are not on the Furniss, M. (1972). Observations of resistance and
verge of breakdown. Trends Ecol. Evol. 32: 727–734. susceptibility to Douglas-fir beetles. In Program
Fredriksen, R.L. (1972). Nutrient budget of a Douglas- Abstracts, Second North American For. Biol. Work-
fir forest on an experimental watershed in western shop, Oregon State Univ., Corvallis. p. 24.
Oregon. In J.F. Franklin. L.J. Dempster, and Gabriel, D., Sait, S.M., Hodgson, J.A. et al. (2010).
R.H. Waring (eds.), Symp. Proc. Research on Conif- Scale matters: the impact of organic farming on
erous Forest Ecosystems. USDA For. Serv., Pacific biodiversity at different spatial scales. Ecol. Lett. 13:
Northwest For. and Rge. Exp. Sta., Portland, OR. 858–869.
Frelich, L.E. (1992). The relationship of natural distur- Gadgil, M. and Berkes, F. (1991). Traditional resource
bances to white pine stand development. In R. A. management systems. Resour. Manage. Optimiza-
Stine, M.J. Baughman (eds.). White Pine Sym- tion 8: 127–141.
posium Proceedings: History, Ecology, Policy and Galbraith-Kent, S.L. and Handel, S.N. (2008). Invasive
Management. Department of Forest Resources, Acer platanoides inhibits native sapling growth in
College of Natural Resources, University of Min- forest understorey communities. J. Ecol. 96: 293–302.
nesota, and Minnesota Extension Service. St. Paul, Gallagher, R.V., Randall, R.P., and Leishman, M.R.
Minnesota. (2015). Trait differences between naturalized and
Frelich, L.E. (2002). Forest Dynamics and Disturbance invasive plant species independent of residence
Regimes. Cambridge, New York. 266 pp. time and phylogeny. Conserv. Biol. 29: 360–369.
Frelich, L.E. and Lorimer, C.G. (1985). Current and Galston, A.W. (1994). Life Processes of Plants. New York:
predicted long-term effects of deer browsing in W. H. Freeman 245 pp.
hemlock forests in Michigan, USA. Biol. Conserv. Gandhi, K.J. and Herms, D.A. (2010). Direct and
34: 99–120. indirect effects of alien insect herbivores on ecolog-
Frelich, L.E. and Lorimer, C.G. (1991). Natural distur- ical processes and interactions in forests of eastern
bance regimes in hemlock-hardwood forests of North America. Biol. Invasions 12: 389–405.
the upper Great Lakes region. Ecol. Monogr. 61: Garland, J.A. (1977). The dry deposition of sulphur
145–164. dioxide to land and water surfaces. Proc. R. Soc.
Frelich, L.E., Hale, C.M., Scheu, S. et al. (2006). Earth- Lond. 12: 245–268.
worm invasion into previously earthworm-free Garner, W.W. (1923). Further studies in photoperiodism
temperate and boreal forests. Biol. Invasions 8: in relation to hydrogen-ion concentration of the
1235–1245. cell-sap and the carbohydrate content of the plant.
Freud, S. (1917). Introductory Lectures on Psychoanal- J. Agric. Res. 23: 871–920.
ysis. New York: Liveright 656 pp. Garner, W.W. and Allard, H.A. (1920). Effect of the
Frey, S.J., Hadley, A.S., and Betts, M.G. (2016). Microcli- relative length of day and night and other factors
mate predicts within-season distribution dynamics of the environment on growth and reproduction in
of montane forest birds. Divers. Distrib. 22: 944–959. plants. J. Agric. Res. 18: 553–606.
References 675
Gashwiler, J.S. (1967). Conifer seed survival in a west- larch throughout 50 years in a common garden
ern Oregon clearcut. Ecology 48: 431–438. experiment. Tree Physiol. 37: 33–46.
Gašparíková, O., Čiamporová, M., Mistrík, I., and Gerhardt, F. and Foster, D.R. (2002). P hysiographical
Baluška, F. (2001). Recent Advances of Plant and historical effects on forest vegetation in
Root Structure and Function. New York: Kluwer central New England, USA. J. Biogeogr. 29:
Academic Publishers 192 pp. 1421–1437.
Gaston, K.J. and Spicer, J.I. (2014). Biodiversity: An Gersper, P.L. and Holowaychuk, N. (1971). Some effects
Introduction. Wiley 208 pp. of stem flow from forest canopy trees on chemical
Gaston, K.J., Warren, P.H., Devine-Wright, P. et al. properties of soils. Ecology 52: 691–702.
(2007). Psychological benefits of greenspace Ghazoul, J. (2007). Recognizing the complexities of
increase with biodiversity. Biol. Lett. 3: 390–394. ecosystem management and the ecosystem service
Gates, D.M. (1968). Energy exchange between concept. Gaia 16: 215–221.
organism and environment. In: Biometeorology Ghilarov, A.M. (2000). Ecosystem functioning and
(ed. W.P. Lowry). Corvallis: Oregon State Univ. intrinsic value of biodiversity. Oikos 90: 408–412.
Press. Gholz, H.L. (1982). Environmental limits on above-
Gates, D.M. (1980). Biophysical Ecology. New York: ground net primary production, leaf area,
Springer-Verlag 611 pp. and biomass in vegetation zones of the Pacific
Gaul, D., Hertel, D., and Leuschner, C. (2008). Effects Northwest. Ecology 63: 469–481.
of experimental soil frost on the fine-root system of Gholz, H.L. and Fisher, R.L. (1982). Organic matter
mature Norway spruce. J. Plant Nutr. Soil Sci. 171: production and distribution in slash pine (Pinus
690–698. elliottii) plantations. Ecology 63: 1827–1839.
Gaylord, M.L., Kolb, T.E., Pockman, W.T. et al. (2013). Gholz, H.L., Wedin, D.A., Smitherman, S.M. et al.
Drought predisposes piñon–juniper woodlands (2000). Long-term dynamics of pine and hard-
to insect attacks and mortality. New Phytol. 198: wood litter in contrasting environments: toward a
567–578. global model of decomposition. Glob. Chang. Biol.
Gehring, C.A. and Whitham, T.G. (1991). Herbivore- 6: 751–765.
driven mycorrhizal mutualism in insect-susceptible Gill, C.J. (1970). The flooding tolerance of woody
pinyon pine. Nature 353: 556–557. species—a review. For. Abstr. 31: 671–688.
Geiger, R., Aron, R.H., and Todhunter, P. (2009). The Gill, A. M. (1977). Plant traits adapted to fires in
Climate near the Ground, 7e. New York: Rowman & Mediterranean land ecosystems. In H. Mooney and
Littlefield Publishers 623 pp. C. E. Conrad (eds.), Proc. Symp. The Environmental
Genet, H., Bréda, N., and Dufrêne, E. (2010). Age- Consequences of Fire and Fuel Management in
related variation in carbon allocation at tree and Mediterranean Ecosystems. USDA, For. Serv. Gen.
stand scales in beech (Fagus sylvatica L.) and sessile Tech. Report WO-3, Washington, D.C.
oak (Quercus petraea (Matt.) Liebl.) using a chrono- Gill, D.S. and Marks, P.L. (1991). Tree and shrub seed-
sequence approach. Tree Physiol. 30: 177–192. ling colonization of old fields in Central New York.
Genre, A., Lanfranco, L., Perotto, S., and Bonfante, P. Ecol. Monogr. 61: 183–205.
(2020). Unique and common traits in mycorrhizal Gill, A.M., Groves, R.H., and Noble, I.R. (1981). Fire
symbioses. Nat. Rev. Microbiol. 18: 649–660. and the Australian Biota. Canberra: Australian
Gentry, A.H. (1986). Endemism in tropical versus Acad. Sci 582 pp.
temperate plant communities. In: Conservation Gill, D.S., Amthor, J.S., and Bormann, F.H. (1998). Leaf
Biology: The Science of Scarcity and Diversity (ed. phenology, photosynthesis, and the persistence of
M. Soulé). Sunderland, MA: Sinauer Associates. saplings and shrubs in a mature northern hard-
Gentry, A.H. and Dodson, C.H. (1987). Diversity and wood forest. Tree Physiol. 18: 281–289.
biogeography of neotropical vascular epiphytes. Gill, A.L., Gallinat, A.S., Sanders-DeMott, R. et al.
Ann. Missouri Bot. Gard. 74: 205–233. (2015). Changes in autumn senescence in northern
George, M.F. and Burke, M.J. (1976). The occurrence of hemisphere deciduous trees: a meta-analysis of
deep supercooling in cold hardy plants. Curr. Adv. autumn phenology studies. Ann. Bot. 116: 875–888.
Plant Sci. 22: 349–360. Gillespie, J.H. (2004). Population Genetics: A Concise
George, M.F. and Burke, M.J. (1977). Supercooling Guide, 2e. Baltimore, MD: Johns Hopkins Univer-
in overwintering azalea flower buds: Additional sity Press 232 pp.
freezing parameters. Plant Physiol. 59: 319–325. Gillett, N.P., Weaver, A.J., Zwiers, F.W., and Flannigan,
George, M.F., Burke, M.J., Pellet, H.M., and Johnson, M.D. (2004). Detecting the effect of climate change
A.G. (1974). Low temperature exotherms and on Canadian forest fires. Geophys. Res. Lett. 31:
woody plant distribution. Hortic. Sci. 6: 519–522. L18211.
George, J.P., Grabner, M., Karanitsch-Ackerl, S. et al. Gillman, L.N. and Wright, S.D. (2006). The influence
(2017). Genetic variation, phenotypic stability, of productivity on the species richness of plants: a
and repeatability of drought response in European critical assessment. Ecology 87: 1234–1243.
676 References
Girard, F., Payette, S., and Gagnon, R. (2009). Origin of Golley, F.B. (ed.) (1983). Ecosystems of the World,
the lichen–spruce woodland in the closed-crown Tropical Rain Forest Ecosystems, vol. 14A.
forest zone of eastern Canada. Glob. Ecol. Biogeogr. New York: Elsevier 382 pp.
18: 291–303. Golley, F.B. (1993). A History of the Ecosystem Concept in
Givnish, T.J. (1981). Serotiny, geography, and fire in the Ecology. New Haven, CT: Yale Univ. Press 254 pp.
pine barrens of New Jersey. Evolution 35: 101–123. Good, N.F. (1968). A study of natural replacement of
Givnish, T.J. (1988). Adaptation to sun and shade: a chestnut in six stands in the highlands of New
whole-plant perspective. Aust. J. Plant Physiol. 15: Jersey. Bull. Torrey Bot. Club 95: 240–253.
63–92. Goodall, D., Lugo, A., Brinson, M., and Brown, S.
Givnish, T.J. (2002). Adaptive significance of evergreen (1990). Ecosystems of the World, Forested Wetlands,
vs. deciduous leaves: solving the triple paradox. vol. 15. New York: Elsevier 527 pp.
Silva Fenn. 36: 703–743. Gordon, D., Rosati, A., Damiano, C., and Dejong, T.M.
Givnish, T.J., Montgomery, R.A., and Goldstein, G. (2006). Seasonal effects of light exposure, temper-
(2004). Adaptive radiation of photosynthetic ature, trunk growth and plant carbohydrate status
physiology in the Hawaiian lobeliads: light on the initiation and growth of epicormic shoots
regimes, static light responses, and whole-plant in Prunus persica. J. Hortic. Sci. Biotechnol. 81:
compensation points. Am. J. Bot. 91: 228–246. 421–428.
Gleason, H.A. (1917). The structure and development Gosz, J.R., Holmes, R.T., Likens, G.E., and Bormann,
of the plant association. Bull. Torrey Bot. Club 44: F.H. (1978). The flow of energy in a forest eco-
463–481. system. Sci. Am. 328: 92–102.
Gleason, H.A. (1926). The individualistic concept of the Gough, C.M., Atkins, J.W., Fahey, R.T., and Hardiman,
plant association. Bull. Torrey Bot. Club 53: 7–26. B.S. (2019). High rates of primary production in
Gleason, H.A. (1936). Is the synusia an association? structurally complex forests. Ecology 100: 1–6.
Ecology 17: 444–451. Gough, C.M., Atkins, J.W., Fahey, R.T. et al. (2020).
Gleason, H.A. (1939). The individualistic concept of the Community and structural constraints on the com-
plant association. Am. Midl. Nat. 21: 92–110. plexity of eastern North American forests. Glob.
Glenn-Lewin, D.C. and van der Maarel, E. (1992). Ecol. Biogeogr. 29: 2107–2118.
Patterns and processes of vegetation dynamics. In: Gowdy, J.M. (1997). The value of biodiversity: markets,
Plant Succession: Theory and Prediction (ed. D.C. society, and ecosystems. Land Econ. 73: 25–41.
Glenn-Lewin, R.K. Peet and T.T. Veblen), 13–59. Graber, R.E. (1970). Natural seed fall in white pine
London: Chapman and Hall. (Pinus strobus L.) stands of varying density. USDA
Glenn-Lewin, D.C., Peet, R.K., and Veblen, T.T. (ed.) For. Serv. Res. Note NE-119. Northeastern For. Exp.
(1992). Plant Succession: Theory and Prediction. Sta., Upper Darby, PA. 6 pp.
London: Chapman and Hall 352 pp. Grace, J. (1989). Tree lines. Phil. Trans. R. Soc. Lond. B
Glinski, R. L. (1977). Regeneration and distribution 324: 233–245.
of sycamores and cottonwood trees along Sonoita Grace, J.B., Anderson, T.M., Seabloom, E.W. et al.
Creek, Santa Cruz County, Arizona. In R. R. (2016). Integrative modelling reveals mechanisms
Johnson, D. A. Jones (tech. coords). Importance, linking productivity and plant species richness.
Preservation, and Management of Riparian Hab- Nature 529: 390–393.
itat: A Symposium. U.S. For. Serv. Gen. Tech. Rep. Gradmann, R. (1898). Das Pflanzenleben der
RM-43. Rocky Mountain For. Range Exp. Sta., Fort Schwäbischen Alb. Verlag Schwäbischen Albver-
Collins, CO. eins, Tübingen. Vo1 1:1–376; Vo1 2:1–376.
Glynn, C., Herms, D.A., Orians, C.M. et al. (2007). Test- Gradowski, T., Lieffers, V.J., Landhäusser, S.M. et al.
ing the growth differentiation balance hypothesis: (2010). Regeneration of Populus nine years after
dynamic responses of willows to nutrient avail- variable retention harvest in boreal mixedwood
ability. New Phytol. 176: 623–634. forests. For. Ecol. Manag. 259: 383–389.
Godwin, H. (1956). The History of the British Flora. Graham, A. (1999). Late Cretaceous and Cenozoic
Cambridge: Cambridge Univ. Press 384 pp. history of North American vegetation. New York:
Goebel, P.C., Palik, B.J., Kirkman, L.K. et al. (2001). Oxford University Press 350 pp.
Forest ecosystems of a Lower Gulf Coastal Plain Graham, B.F. Jr. and Bormann, F.H. (1966). Natural
landscape: multifactor classification and analysis. root grafts. Bot. Rev. 32: 255–292.
J. Torr. Bot. Soc. 128: 47–75. Granhall, U. (1981). Biological nitrogen fixation in rela-
Goldblatt, P. (1980). Polyploidy in angiosperms: tion to environmental factors and functioning of
monocotyledons. In: Polyploidy (ed. W.H. Lewis), natural ecosystems. In: Terrestrial Nitrogen Cycles.
219–239. New York: Plenum Press. Ecol. Bull. 33 (ed. F.E. Clark and T. Rosswell),
Golley, F.B. (1977). Ecological Succession. Stroudsburg, 131–144. Stockholm: Swedish Natural Sci. Res.
PA: Dowden, Hutchinson, and Ross 373 pp. Council.
References 677
Grant, V. (1963). The Origin of Adaptations. New York: Grier, C.C. (1975). Wildfire effects on nutrient distribu-
Columbia Univ. Press 606 pp. tion and leaching in a coniferous ecosystem. Can. J.
Grant, V. (1971). Plant Speciation. New York: Columbia For. Res. 5: 599–607.
Univ. Press 435 pp. Grier, C. C, Lee, K. M., Nadkarni, N. M. et al. (1989).
Grant, V. (1977). Organismic Evolution. San Francisco: Productivity of forests of the United States and its
W. H. Freeman 418 pp. relation to soil and site factors and management: a
Grant, M.C. (1993). The trembling giant. Discover 14: review. USDA For. Serv. Gen. Tech. Report PNW-
83–89. GTR-222. Pacific Northwest Res. Sta., Corvallis,
Grant, M.C., Mitton, J.B., and Linhart, Y.B. (1992). Even OR. 51 pp.
larger organisms. Nature 360: 216. Griffin, G.J. (1992). American chestnut survival in
Gratkowski, H.J. (1956). Windthrow around staggered understory mesic sites following the chestnut
settings in old-growth Douglas-fir. For. Sci. 2: blight pandemic. Can. J. Bot. 70: 1950–1956.
60–74. Grigal, D. F. (1984). Shortcomings of soil surveys for
Graumlich, L.J., Brubaker, L.B., and Grier, C.C. (1989). forest management. In J. G. Bockheim (ed.), Symp.
Long-term trends in forest net primary produc- Proc. Forest Land Classification: Experience, Prob-
tivity: Cascade Mountains, Washington. Ecology 70: lems, Perspectives. NCR-102 North Central For. Soils
405–410. Com., Soc. Am. For., USDA For. Serv., and USDA
Gray, S.B. and Brady, S.M. (2016). Plant developmental Soil Cons. Serv. Madison, WI.
responses to climate change. Dev. Biol. 419: 64–77. Grime, J.P. (1966). Shade avoidance and shade tolerance
Gray, L.K., Gylander, T., Mbogga, M.S. et al. (2011). in flowering plants. In: Light as an Ecological Factor
Assisted migration to address climate change: rec- (ed. R. Bainbridge, G.C. Evans and O. Rackham),
ommendations for aspen reforestation in western 187–207. England: Blackwell, Oxford.
Canada. Ecol. Appl. 21: 1591–1603. Grime, J.P. (1977). Evidence for the existence of three
Green, D.G. (1981). Time series and postglacial forest primary strategies in plants and its relevance to
ecology. Quat. Res. 15: 265–277. ecological and evolutionary theory. Am. Nat. 111:
1169–1194.
Green, D.C. and Grigal, D.F. (1979). Jack pine biomass
accretion on shallow and deep soils in Minnesota. Grime, J.P. (1979). Plant Strategies and Vegetation
Soil Sci. Soc. Amer. J. 43: 1233–1237. Processes. New York: Wiley 222 pp.
Green, D.C. and Grigal, D.F. (1980). Nutrient accumula- Grime, J.P. (1988). The C-S-R model of primary plant
tions in jack pine stands on deep and shallow soils strategies—origins, implications and tests. In: Plant
over bedrock. For. Sci. 26: 325–333. Evolutionary Biology (ed. L.D. Gottlieb and S.K.
Jain), 371–393. New York: Chapman and Hall.
Green, S.J. and Grosholz, E.D. (2021). Functional erad-
ication as a framework for invasive species control. Grimm, R.C. (1984). Fire and other factors controlling
Front. Ecol. Environ. 19: 98–107. the Big Woods vegetation of Minnesota in the mid-
nineteenth century. Ecol. Monogr. 54: 291–311.
Green, R.E., Connell, S.J., Scharlemann, J.P., and
Balmford, A. (2005). Farming and the fate of wild Grimm, V. and Wissel, C. (1997). Babel, or the ecolog-
nature. Science 307: 550–555. ical stability discussions: an inventory and analysis
of terminology and a guide for avoiding confusion.
Greenwood, M.S. and Hutchison, K.W. (1993).
Oecologia 109: 323–334.
Maturation as a developmental process. In:
Clonal Forestry I, Genetics and Biotechnology Groffman, P.M. and Tiedje, J.M. (1989). Denitrifica-
(ed. M.R. Ahuja and W.J. Libby), 14–33. Berlin: tion in north temperate forests soils: spatial and
Springer-Verlag. temporal patterns at the landscape and seasonal
scales. Soil Biol. Biochem. 21: 613–620.
Greer, B.T., Still, C., Howe, G.T. et al. (2016). Popula-
tions of aspen (Populus tremuloides Michx.) with Groffman, P.M., Tiedje, J.M., Mokma, D.L., and
Simkins, S. (1992). Regional scale analysis of deni-
different evolutionary histories differ in their cli-
trification in north temperate forest soils. Landsc.
mate occupancy. Ecol. Evol. 6: 3032–3039.
Ecol. 7: 45–53.
Greer, B.T., Still, C., Cullinan, G.L. et al. (2018). Poly-
Groner, E. and Novoplansky, A. (2003). Reconsidering
ploidy influences plant–environment interactions
diversity–productivity relationships: directness
in quaking aspen (Populus tremuloides Michx.).
of productivity estimates matters. Ecol. Lett. 6:
Tree Physiol. 38: 630–640.
695–699.
Gregory, S.V., Swanson, F.J., McKee, W.A., and Cum-
Groom, M.J., Meffe, G.K., Carroll, C.R., and Andelman,
mins, K.W. (1991). An ecosystem perspective of
S.J. (2006). Principles of Conservation Biology, 3e.
riparian zones. Bioscience 41: 540–551.
Sunderland, MA: Sinauer Associates 816 pp.
Gregow, H., Laaksonen, A., and Alper, M. (2017).
Gross, H.L. (1972). Crown deterioration and reduced
Increasing large scale windstorm damage in
growth associated with excessive seed production
western, central and northern European forests,
by birch. Can. J. Bot. 50: 2431–2437.
1951–2010. Sci. Rep. 7: 46397.
678 References
Gross, H.L. and Harnden, A. A. (1968). Dieback and Gustafson, F.G. (1943). Influence of light upon tree
abnormal growth of yellow birch induced by heavy growth. J. For. 41: 212–213.
fruiting. Canada Dept. For. and Rural Develop., Gustafson, E.J. and Crow, T.R. (1994). Modeling the
For. Res Lab. Info. Report O-X-79, Sault Ste. Marie, effects of forest harvesting on landscape structure
Ontario. 12 pp. and the spatial distribution of cowbird brood para-
Grotkopp, E. and Rejmánek, M. (2007). High seedling sitism. Landsc. Ecol. 9: 237–248.
relative growth rate and specific leaf area are traits of Haack, R.A. and Byler, J.W. (1993). Insects and path-
invasive species: phylogenetically independent con- ogens, regulators of forest ecosystems. J. For. 91:
trasts of woody angiosperms. Am. J. Bot. 94: 526–532. 32–37.
Grubb, P.J. (1977). The maintenance of species-richness Haase, E.F. (1970). Environmental fluctuations on
in plant communities: the importance of the regen- south-facing slopes in the Santa Catalina Moun-
eration niche. Biol. Rev. 52: 107–145. tains of Arizona. Ecology 51: 959–974.
Grubb, P.J. (1985). Plant populations and vegetation in Habeck, J.R. (1958). White cedar ecotypes in Wisconsin.
relation to habitat, disturbance and competition: Ecology 39: 457–463.
problems of generalization. In: The Population Habeck, J.R. and Mutch, R.W. (1973). Fire-dependent
Structure of Vegetation (Handbook of Vegetation forests in the northern Rocky Mountains. Quat.
Science, Part III) (ed. J. White), 23–36. Dordrecht: Res. 3: 408–424.
W. Junk. Hack, J.T. and Goodlett, J.C. (1960). Geomorphology
Grubb, P.J. (1988). The uncoupling of disturbance and and forest ecology of a mountain region in the
recruitment, two kinds of seed bank, and persis- central Appalachians. U. S. Geol. Surv. Prof. Paper
tence of plant populations at the regional and local 347. 66 pp + map.
scales. Ann. Zool. Fenn. 25: 23–36. Hacke, U.G., Sperry, J.S., Pockman, W.T. et al. (2001).
Grumbine, R.E. (1994). What is ecosystem Trends in wood density and structure are linked
management? Conserv. Biol. 8: 27–38. to prevention of xylem implosion by negative
Grytnes, J.A. and Vetaas, O.R. (2002). Species richness pressure. Oecologia 126: 457–461.
and altitude: a comparison between null models Hacket-Pain, A.J., Lageard, J.G.A., and Thomas, P.A.
and interpolated plant species richness along the (2017). Drought and reproductive effort interact
Himalayan altitudinal gradient, Nepal. Am. Nat. to control growth of a temperate broadleaved tree
159: 294–304. species (Fagus sylvatica). Tree Physiol. 37: 744–754.
Gu, L., Hanson, P.J., Post, W.M. et al. (2008). The 2007 Hagemeier, M. and Leuschner, C. (2019). Leaf and
eastern US spring freeze: increased cold damage in crown optical properties of five early-, mid-and
a warming world? Bioscience 58: 253–262. late-successional temperate tree species and their
Guarino, R., Willner, W., Pignatti, S. et al. (2018). relation to sapling light demand. Forests 10: 925.
Spatio-temporal variations in the application of the Hagen, J.B. (1992). An Entangled Bank. New Bruns-
Braun-Blanquet approach in Europe. Phytocoenolo- wick, NJ: Rutgers Univ. Press 245 pp.
gia 48: 239–250. Hagglund, B. (1981). Evaluation of forest site produc-
Gunderson, C.A., Edwards, N.T., Walker, A.V. et al. tivity. For. Abstr. 42: 515–527.
(2012). Forest phenology and a warmer climate– Hagmann, R.K., Hessburg, P.F., Prichard, S.J. et al.
growing season extension in relation to climatic (2021). Evidence for widespread changes in the
provenance. Glob. Chang. Biol. 18: 2008–2025. structure, composition, and fire regimes of western
Guo, J.P. and Wang, J.R. (2006). Comparison of height North American forests. Ecol. Appl. 31: e02431.
growth and growth intercept models of jack pine Halbwachs, H., Brandl, R., and Bässler, C. (2015). Spore
plantations and natural stands in northern Ontario. wall traits of ectomycorrhizal and saprotrophic
Can. J. For. Res. 36: 2179–2188. agarics may mirror their distinct lifestyles. Fungal
Gurevitch, J. and Padilla, D. (2004). Are invasive species Ecol. 17: 197–204.
a major cause of extinctions? Trends Ecol. Evol. 19: Hale, C.M., Frelich, L.E., Reich, P.B., and Pastor, J.
470–474. (2005a). Effects of European earthworm invasion
Gurnell, A.M. and Grabowski, R.C. (2016). Vegetation– on soil characteristics in northern hardwood forests
hydrogeomorphology interactions in a low-energy, of Minnesota, USA. Ecosystems 8: 911–927.
human-impacted river. River Res. Appl. 32: 202–215. Hale, C.M., Frelich, L.E., and Reich, P.B. (2005b).
Gurnell, A.M., Bertoldi, W., and Corenblit, D. (2012). Exotic European earthworm invasion dynamics
Changing river channels: the roles of hydrological in northern hardwood forests of Minnesota, USA.
processes, plants and pioneer landforms in humid Ecol. Appl. 15: 848–860.
temperate, mixed load, gravel bed rivers. Earth-Sci. Hale, C.M., Frelich, L.E., and Reich, P.B. (2006).
Rev. 111: 129–141. Changes in hardwood forest understory plant
Gurnell, A.M., Corenblit, D., García de Jalón, D. communities in response to European earthworm
et al. (2016). A conceptual model of vegetation– invasions. Ecology 87: 1637–1649.
hydrogeomorphology interactions within river Hall, F.C. (1989). Plant community classification:
corridors. River Res. Appl. 32: 142–163. from concept to application. In D.E. Ferguson,
References 679
E.E., P. Morgan, and F.D. Johnson (comps.), Hamrick, J.L. and Godt, M.J.W. (1996). Effects of life
Proceedings—Land Classifications Based on Vegeta- history traits on genetic diversity in plant species.
tion: Applications for Resource Management. USDA Philos. Trans. R. Soc. B: Biol. Sci. 351: 1291–1298.
For. Serv. Gen. Tech. Report INT-257. Intermoun- Hamrick, J.L., Godt, M.J.W., and Sherman-Broyles, S.L.
tain Res. Sta., Ogden, UT. 315 pp. (1992). Factors influencing levels of genetic diver-
Hall, A.R., Miller, A.D., Leggett, H.C. et al. (2012). sity in woody plant species. In: Population Genetics
Diversity-disturbance relationships: frequency and of Forest Trees (ed. W.T. Adams, S.H. Strauss, D.L.
intensity interact. Biol. Lett. 8: 768–771. Copes and A.R. Griffin), 95–124. Boston, MA:
Hallé, F., Oldeman, R.A.A., and Tomlinson, P.B. (1978). Kluwer.
Tropical Trees and Forests. Berlin: Springer-Verlag Han, Q. and Kabeya, D. (2017). Recent developments
441 pp. in understanding mast seeding in relation to
Haller, J.R. (1965). The role of 2–needle fascicles in the dynamics of carbon and nitrogen resources in
adaptation and evolution of ponderosa pine. Brit- temperate trees. Ecol. Res. 32: 771–778.
tonia 17: 354–382. Handler, S., Pike, C., and St. Clair, B. (2018). Assisted
Halliday, W.E.D. (1937). A forest classification for migration. USDA Forest Service Climate Change
Canada. For. Serv. Bull. 89; Can. Dept. Mines and Resource Center. http://www.fs.usda.gov/ccrc/
Resources. 50 pp. topics/assisted-migration. Accessed 01-13-2022.
Halofsky, J.E., Andrews-Key, S.A., Edwards, J.E. et al. Hanes, T. (1988). California chaparral. In M.G. Bar-
(2018). Adapting forest management to climate bour and J. Major (eds.), Terrestrial Vegetation of
change: the state of science and applications in California. California Native Plant Society, Special
Canada and the United States. For. Ecol. Manag. Publ. No. 9. Berkeley, CA.
421: 84–97. Hanks, J.P., Fitzhugh, E., and Hanks, S. R. (1983).
Halpern, C.B. and Harmon, M.E. (1983). Early plant A habitat type classification system for ponderosa
succession on the Muddy River mudflow, Mount pine forests of northern Arizona. USDA For. Serv.
St. Helens. Am. Midl. Nat. 110: 97–106. Gen. Tech. Report RM-97. Rocky Mountain For.
Halpern, C.B. and Spies, T.A. (1995). Plant species and Rge. Exp. Sta., Fort Collins, CO. 22 pp.
diversity in natural and managed forest of the Hanley, N. and Perrings, C. (2019). The economic
Pacific Northwest. Ecol. Appl. 5: 913–934. value of biodiversity. Ann. Rev. Resour. Econ. 11:
Hamanishi, E.T. and Campbell, M.M. (2011). Genome- 355–375.
wide responses to drought in forest trees. Forestry Hanley, M.E., Lamont, B.B., Fairbanks, M.M., and
84: 273–283. Rafferty, C.M. (2007). Plant structural traits and
Hamilton, M.B. (2009). Population Genetics. Hoboken, their role in anti-herbivore defence. Perspect. Plant
NJ: Wiley Blackwell 424 pp. Ecol. Evol. Syst. 8: 157–178.
Hammerson, G.A. (1994). Beaver (Caston canadensis): Hannah, L. (2021). Climate Change Biology, 3e.
ecosystem alterations, management, and moni- Cambridge, MA: Academic Press 528 pp.
toring. Nat. Areas J. 14: 44–57. Hanover, J.W. (1975). Physiology of tree resistance to
Hammitt, W. E. and Barnes, B.V. (1989). Composition insects. Annu. Rev. Entomol. 20: 75–95.
and structure of an old-growth oak-hickory forest Hansen, H.P. (1947). Postglacial forest succession,
in southern Michigan over 20 years. In G. Rink climate, and chronology in the Pacific Northwest.
and C. A. Budelsky (eds.), Proc. Seventh Central Trans. Am. Philos. Soc. 37 (1): 130 pp.
Hardwoods Conference. USDA For. Serv. Gen. Tech. Hansen, H.P. (1955). Postglacial forests in south central
Report NC-132. North Central For. Exp. Sta., St. and central British Columbia. Am. J. Sci. 253:
Paul, MN. 640–658.
Hampf, F.E. (ed.). (1965). Site index curves for some Hansen, E.M. (2014). Forest development and carbon
forest species in the eastern United States. dynamics after mountain pine beetle outbreaks.
USDA For. Serv., Eastern Region. Upper Darby, For. Sci. 60: 476–488.
PA. 43 pp. Hansen, A.J., Spies, T.A., Swanson, F.J. et al. (1991).
Hamrick, J.L. (1989). Isozymes and the analysis of ge- Conserving biodiversity in managed forests. Biosci-
netic structure in plant populations. In: Isozymes ence 41: 382–392.
in Plant Biology (ed. D.E. Soltis and P.S. Soltis), Hansen, J.W., Challinor, A., Ines, A. et al. (2006). Trans-
87–105. Portland, OR: Dioscorides Press. lating climate forecasts into agricultural terms:
Hamrick, J.L. (2004). Response of forest trees to global advances and challenges. Clim. Res. 33: 27–41.
environmental changes. For. Ecol. Manag. 197: Hanson, A.D. and Hitz, W.D. (1982). Metabolic
323–335. responses of mesophytes to plant water deficits.
Hamrick, J.L. and Godt, M.J.W. (1989). Allozyme diver- Annu. Rev. Plant Physiol. 33: 163–203.
sity in plant species. In: Plant Populations Genetics, Hardig, T.M., Brunsfeld, S.J., Fritz, R.S. et al. (2000).
Breeding, and Genetic Resources (ed. A.H.D. Brown, Morphological and molecular evidence for hybrid-
M.T. Clegg, A.L. Kahler and B.S. Weir), 43–63. ization and introgression in a willow (Salix) hybrid
Sunderland, MA: Sinauer. zone. Mol. Ecol. 9: 9–24.
680 References
Hardin, J.W. (1979). Patterns of variation in foliar Harper, J.L. and White, J. (1974). The demography of
trichomes of eastern North American Quercus. Am. plants. Annu. Rev. Ecol. Syst. 5: 419–463.
J. Bot. 66: 576–585. Harrington, T.C. (1986). Growth decline of wind-
Hardy, J.P., Groffman, P.M., Fitzhugh, R.D. et al. (2001). exposed red spruce and balsam fir in the White
Snow depth manipulation and its influence on soil Mountains. Can. J. For. Res. 16: 232–238.
frost and water dynamics in a northern hardwood Harrington, W. (1991). Wildlife: severe decline and
forest. Biogeochemistry 56: 151–174. partial recovery. In: America’s Renewable Resources:
Hare, F.K. and Ritchie, J.C. (1972). The boreal biocli- Historical Trends and Current Challenges (ed. K.D.
mates. Geogr. Rev. 62: 333–365. Frederick and R.A. Sedjo), 205–248. Washington,
Hari, P. and Häkkinen, R. (1991). The utilization of old D.C: Resources for the Future.
phenological time series of budburst to compare Harrington, C.A. and Gould, P.J. (2015). Tradeoffs bet-
models describing annual cycles of plants. Tree ween chilling and forcing in satisfying dormancy
Physiol. 8: 281–287. requirements for Pacific Northwest tree species.
Harley, J.L. and Smith, S.E. (1983). Mycorrhizal Symbi- Front. Plant Sci. 6: 120.
osis. New York: Academic Press 483 pp. Harrington, C.A., Gould, P.J., and Clair, J.B.S. (2010).
Harley, P., Guenther, A., and Zimmerman, P. (1996). Modeling the effects of winter environment on
Effects of light, temperature and canopy position dormancy release of Douglas-fir. For. Ecol. Manag.
on net photosynthesis and isoprene emission from 259: 798–808.
sweetgum (Liquidambar styraciflua) leaves. Tree Harris, T.M. (1958). Forest fire in the mesozoic. J. Ecol.
Physiol. 16: 25–32. 46: 447–453.
Harlow, W.M., Harrar, E.S., Hardin, J.W., and White, Harris, L.D. (1984). The Fragmented Forest. Chicago:
F.M. (1996). Textbook of Dendrology, 8e. New York: Univ. Chicago Press 211 pp.
McGraw-Hill 501 pp. Harris, J.R. and Fanelli, J. (1999). Root and shoot
Harmon, M.E. (1984). Survival of trees after low- growth of pot-in-pot red and sugar maple.
intensity surface fires in Great Smoky Mountains J. Environ. Hortic. 17: 80–83.
National Park. Ecology 65: 796–802. Harris, L.D. and Silva-Lopez, G. (1992). Forest
Harmon, M.E. (1987). The influence of litter and fragmentation and the conservation of biological
humus accumulations and canopy openness on diversity. In: Conservation Biology (ed. P.L. Fiedler
Picea sitchensis (Bong.) Carr. and Tsuga hetero- and S.K. Jain), 197–237. New York: Chapman and
phylla (Raf.) Sarg. seedlings growing on logs. Can. Hall.
J. For. Res. 17: 1475–1479. Harris, J.R., Bassuk, N.L., Zobel, R.W., and Whitlow,
Harmon, M.E. and Franklin, J.F. (1989). Tree seedlings T.H. (1995). Root and shoot growth periodicity of
on logs in Picea-Tsuga forests of Oregon and Wash- green ash, scarlet oak, Turkish hazelnut, and tree
ington. Ecology 70: 48–59. lilac. J. Am. Soc. Hortic. Sci. 120: 211–216.
Harmon, M.E. and Hua, C. (1991). Coarse woody debris Harris, N.L., Gibbs, D.A., Baccini, A. et al. (2021).
dynamics in two old-growth ecosystems. Bioscience Global maps of twenty-first century forest carbon
41: 604–610. fluxes. Nat. Clim. Chang. 11: 234–240.
Harmon, M.E., Franklin, J.F., Swanson, F.J. et al. (1986). Harshberger, J.W. (1899). Thermotropic movement
Ecology of coarse woody debris in temperate eco- of leaves of Rhododendron maximum. Proc. Natl.
systems. Adv. Ecol. Res. 15: 133–302. Acad. Sci. U. S. A. 214–222.
Harmon, M.E., Ferrell, W.K., and Franklin, J.F. Hart, J.W. (1988). Light and Plant Growth. Boston, MA:
(1990). Effects on carbon storage of conversion of Unwin Hyman 204 pp.
old-growth forests to young forests. Science 247: Hart, S.C., Stark, J. M., Davidson, E. A., and Firestone,
699–702. M. K. (1994). Nitrogen mineralization, immobiliza-
Harmon, M.E., Bond-Lamberty, B., Tang, J., and Vargas, tion, and nitrification. In R. W. Weaver, S. Angle, P.
R. (2011). Heterotrophic respiration in disturbed Bottomley, D. Bezdicek, S. Smith, A. Tabatabai, and
forests: a review with examples from North A. Wollum (eds.), Methods of Soil Analysis: Part 2.
America. J. Geophys. Res. 116: G00K04. Microbiological and Biochemical Properties. Soil Sci.
Harper, J.L. (1961). Approaches to the study of plant Soc. Amer. Book Series, No. 5., Soil Sci. Soc. Amer.,
competition. In F.L. Milthorpe (ed.), Mechanisms Segoe, WS.
in Biological Competition. Symp. Soc. Exp. Biology Hart, S.J., Henkelman, J., McLoughlin, P.D. et al. (2019).
15:1–39. Examining forest resilience to changing fire fre-
Harper, R. M. (1962). Historical notes on the relation quency in a fire-prone region of boreal forest. Glob.
of fires to forests. In Proc. Annual Tall Timbers Chang. Biol. 25: 869–884.
Fire Ecology Conf. 1:11–29. Tall Timbers Res. Sta. Hartesveldt, R.J., and Harvey, H.T. (1967). The fire
Talahassee, FL. ecology of sequoia regeneration. In Proc. Annual
Harper, J.L. (1977). Population Biology of Plants. Lon- Tall Timbers Fire Ecology Conf. 7:65–77. Tall Tim-
don: Academic Press 892 pp. bers Res. Sta., Tallahassee, FL.
References 681
Hartl, D.L. (2020). A Primer of Population Genetics and productivity through population asynchrony and
Genomics, 4e. Oxford, UK: Oxford University Press overyielding. Ecology 91: 2213–2220.
320 pp. Hedrick, P.W. (2011). Genetics of Populations, 4e. Sud-
Hartl, D.L. and Clark, A.G. (2006). Principles of bury, MA: Jones and Bartlett 675 pp.
Population Genetics, 4e. Sunderland, MA: Sinauer Heikkenen, H.J., Scheckler, S.E., Egan, P.J.J. Jr., and
Associates 672 pp. Williams, C.B. Jr. (1986). Incomplete abscission of
Harvey, C.A. (2000). Colonization of agricultural wind- needle clusters and resin release from artificially
breaks by forest trees: effects of connectivity and water-stressed loblolly pine (Pinus taeda): a com-
remnant trees. Ecol. Appl. 10: 1762–1773. ponent for plant-animal interactions. Am. J. Bot. 73:
Harvey, B.J., Donato, D.C., and Turner, M.G. (2016a). 1384–1392.
Burn me twice, shame on who? Interactions bet- Heinselman, M.L. (1973). Fire in the virgin forests of
ween successive forest fires across a temperate the Boundary Waters Canoe Area, Minnesota.
mountain region. Ecology 97: 2272–2282. Quat. Res. 3: 329–382.
Harvey, B.J., Donato, D.C., and Turner, M.G. (2016b). Heinselman, M.L. (1981). Fire and succession in the
High and dry: Post-fire tree seedling establish- conifer forests of northern North America. In:
ment in subalpine forests decreases with post-fire Forest Succession: Concepts and Application (ed.
drought and large stand-replacing burn patches. D.C. West, H.H. Shugart and D.B. Botkin), 374–405.
Glob. Ecol. Biogeogr. 25: 655–669. New York: Springer-Verlag.
Hase, H. and Foelster, H. (1983). Impact of planta- Heinze, J., Gensch, S., Weber, E., and Joshi, J. (2017).
tion forestry with teak (Tectonia grandis) on the Soil temperature modifies effects of soil biota on
nutrient status of young alluvial soils in west plant growth. J. Plant Biol. 10: 808–821.
Venezuela. For. Ecol. Manag. 6: 33–57. Helgath, S. F. (1975). Trial deterioration in the Selway-
Hattas, D., Scogings, P.F., and Julkunen-Tiitto, R. Bitterroot Wilderness. USDA For. Serv. Res. Note
(2017). Does the growth differentiation balance INT-193. Intermountain For. and Rge. Exp. Sta.,
hypothesis explain allocation to secondary metabo- Ogden, UT. 15 pp.
lites in Combretum apiculatum, an African savanna Helgerson, O.T. (1989). Heat damage in tree seedlings
woody species? J. Chem. Ecol. 43: 153–163. and its prevention. New For. 3: 333–358.
Haugo, R.D., Kellogg, B.S., Cansler, C.A. et al. (2019). Hellmers, H. (1962). Temperature effect upon optimum
The missing fire: quantifying human exclusion of tree growth. In: Tree Growth (ed. T.T. Kozlowski),
wildfire in Pacific Northwest forests, USA. Eco- 275–287. New York: Ronald Press.
sphere 10: e02702. Hellmers, H. (1966a). Temperature action and interac-
Hawk, G.M. and Zobel, D.B. (1974). Forest succession tion of temperature regimes in the growth of red fir
on alluvial landforms of the McKenzie River valley, seedlings. For. Sci. 12: 90–96.
Oregon. Northwest Sci. 48: 245–265. Hellmers, H. (1966b). Growth response of redwood
Hawkins, W.A. and Graham, R.C. (2017). Soil min- seedlings to thermoperiodism. For. Sci. 12:
eralogy of a vernal pool catena in southern 276–283.
California. Soil Sci. Soc. Am. J. 81: 214–223. Hellmers, H. and Sundahl, W.P. (1959). Response of
Hayhoe, K., Wuebbles, D.J., Easterling, D.R. et al. Sequoia sempervirens (D. Don) Endl. and Pseu-
(2018). Our changing climate. In: Impacts, Risks, dotsuga menziesii (Mirb. Franco) seedlings to tem-
and Adaptation in the United States: Fourth perature. Nature 184: 1247–1248.
National Climate Assessment, Volume II (ed. D.R. Hellmers, H., Genthe, M.K., and Ronco, F. (1970).
Reidmiller, C.W. Avery, D.R. Easterling, et al.), Temperature affects growth and development of
72–144. Washington, DC, USA: U.S. Global Change Engelmann spruce. For. Sci. 16: 447–452.
Research Program. Hember, R.A., Kurz, M.A., and Coops, N.C. (2017).
He, H.S., Gustafson, E.J., and Lischke, H. (2017). Relationships between individual-tree mortality
Modeling forest landscapes in a changing cli- and water-balance variables indicate positive trends
mate: theory and application. Landsc. Ecol. 32: in water stress-induced tree mortality across North
1299–1305. America. Glob. Chang. Biol. 23: 1691–1710.
Heberling, J.M. and Fridley, J.D. (2013). Resource-use Henderson, G.S., Swank, W.T., Waide, J.B., and Grier,
strategies of native and invasive plants in eastern C.C. (1978). Nutrient budgets of Appalachian and
North American forests. New Phytol. 200: 523–533. cascade region watersheds: a comparison. For. Sci.
Heberling, J.M., Cassidy, S.T., Fridley, J.D., and Kalisz, 24: 385–397.
S. (2019). Carbon gain phenologies of spring- Hendrick, R.L. and Pregitzer, K.S. (1992). The demog-
flowering perennials in a deciduous forest indicate raphy of fine roots in a northern hardwood forest.
a novel niche for a widespread invader. New Phytol. Ecology 73: 1094–1104.
221: 778–788. Hendricks, J.J., Nadelhoffer, K.J., and Aber, J.D. (1993).
Hector, A., Hautier, Y., Saner, P. et al. (2010). General Assessing the role of fine roots in carbon and
stabilizing effects of plant diversity on grassland nitrogen cycling. Trends Ecol. Evol. 8: 174–178.
682 References
Henkel, T.K., Chambers, J.Q., and Baker, D.A. (2016). Heyerdahl, E.K., Brubaker, L.B., and Agee, J.K. (2001).
Delayed tree mortality and Chinese tallow (Tri- Spatial controls of historical fire regimes: a multi-
adica sebifera) population explosion in a Louisiana scale example from the Interior West, USA. Ecology
bottomland hardwood forest following Hurricane 82: 660–678.
Katrina. For. Ecol. Manag. 378: 222–232. Hibbs, D.E. (1981). Leader growth and the architecture
Henry, H.A.L. and Aarssen, L.W. (2001). Inter-and of three North American hemlocks. Can. J. Bot. 59:
intraspecific relationships between shade tolerance 476–480.
and shade avoidance in temperate trees. Oikos 93: Hicke, J.A., Allen, C.D., Desai, A.R. et al. (2012). Effects
477–487. of biotic disturbances on forest carbon cycling in
Henry, J.A., Portier, J.M., and Coyne, J. (1994). The the United States and Canada. Glob. Chang. Biol.
Climate and Weather of Florida. Sarasota, FL: 18: 7–34.
Pineapple Press 279 pp. Hicke, J.A., Meddens, A.J., and Kolden, C.A. (2016).
Hermann, R.K. and Lavender, D.P. (1968). Early growth Recent tree mortality in the western United States
of Douglas-fir from various altitudes and aspects in from bark beetles and forest fires. For. Sci. 62:
southern Oregon. Silvae Genet. 17: 143–151. 141–153.
Hermes, K. (1955). Die Lage der oberen Waldgrenze Higuera, P.E., Shuman, B.N., and Wolf, K.D. (2021).
in den Gebirgen der Erde und ihr Abstand zur Rocky Mountain subalpine forests now burning
Schneegrenze (KoÈlner geographische Arbeiten 5). more than any time in recent millennia. Proc. Natl.
Geographisches Institut, UniversitaÈt KoÈln. Acad. Sci. 118: e2103135118.
Herms, D.A. and Mattson, W.J. (1992). The dilemma Hilker, M. and Schmülling, T. (2019). Stress priming,
of plants: to grow or defend. Q. Rev. Biol. 67: memory, and signalling in plants. Plant Cell
283–335. Environ. 42: 753–761.
Herms, D.A. and McCullough, D.G. (2014). Emerald Hills, G.A. (1952). The classification and evaluation of
ash borer invasion of North America: history, site for forestry. Ontario Dept. Lands and For., Res.
biology, ecology, impacts, and management. Annu. Rept. 24. 41 pp.
Rev. Entomol. 59: 13–30. Hills, G.A. (1960). Regional site research. For. Chron.
Herrel, A., Joly, D., and Danchin, E. (2020). Epigenetics 36: 401–423.
in ecology and evolution. Funct. Ecol. 34: 381–384. Hills, G.A. (1977). An integrated iterative holistic
Heslop-Harrison, J. (1964). Forty years of genecology. approach to ecosystem classification. In J. Thie
Adv. Ecol. Res. 2: 159–247. and G. Ironside (eds.), Ecological (Biophysical)
Heslop-Harrison, J. (1967). New Concepts in Flowering- Land Classification in Canada. Ecological land
Plant Taxonomy. Cambridge, MA: Harvard Univ. classification series, No. 1, Lands Directorate, Env.
Press 134 pp. Canada, Ottawa.
Hessburg, P.F., Miller, C.I., Parks, S.A. et al. (2019). Cli- Hills, G. A. and Pierpoint, G. (1960). Forest site evalua-
mate, environment, and disturbance history govern tion in Ontario. Ontario Dept. Lands and For., Res.
resilience of western North American forests. Rept. 42. 64 pp.
Front. Ecol. Evol. 7: 239. Hilmers, T., Friess, N., Bässler, C. et al. (2018). Biodiver-
Hett, J.M. (1971). A dynamic analysis of age in sugar sity along temperate forest succession. J. Appl. Ecol.
maple seedlings. Ecology 52: 1071–1074. 55: 2756–2766.
Heusser, C. J. (1960). Late-Pleistocene environments Himelick, E.B. and Neely, D. (1962). Root-grafting of
of North Pacific North America. Amer. Geog. Soc. city-planted American elms. Plant Dis. Rep. 46:
Spec. Publ. 35. 308 pp. 86–87.
Heusser, C.J. (1965). A Pleistocene phytogeographical Hinckley, T.M., Teskey, R.O., Duhme, F., and Richter, H.
sketch of the Pacific Northwest and Alaska. In: The (1981). Temperate hardwood forests. In: Water Def-
Quaternary of the United States (ed. H.E. Wright Jr. icits and Plant Growth, vol. 6 (ed. T.T. Kozlowski),
and D.G. Frey), 469–483. Princeton, NJ: Princeton 153–208. New York: Academic Press.
Univ. Press. Hinds, T.E., Hawksworth, F.G., and Davidson, R.W.
Heusser, C.J. (1983). Vegetational history of the north- (1965). Beetle-killed Engelmann spruce: its deterio-
western United States including Alaska. In: The Late ration in Colorado. J. For. 63: 536–542.
Pleistocene, Vol. 1 of H. E. Wright, Jr., (ed.), Late- Hinesley, L.E. (1981). Initial growth of Fraser fir seed-
Quaternary Environments of the United States (ed. lings at different day/night temperatures. For. Sci.
S.C. Porter). Minneapolis: Univ. Minnesota Press. 27: 545–550.
Hewitt, N. and Kellman, M. (2002). Tree seed dispersal Hix, D.M. (1988). Multifactor classification and anal-
among forest fragments: II. Dispersal abilities and ysis of upland hardwood forest ecosystems of the
biogeographical controls. J. Biogeogr. 29: 351–363. Kickapoo River watershed, southwestern Wiscon-
Hewlett, J.D. and Hibbert, A.R. (1963). Moisture and sin. Can. J. For. Res. 18: 1405–1415.
energy conditions within a sloping soil mass during Hix, D.M. and Barnes, B.V. (1984). Effects of clear-
drainage. J. Geophys. Res. 68: 1081–1087. cutting on the vegetation and soil of an eastern
References 683
hemlock dominated ecosystem, western Upper elevational gradient, bedrock effects and migration
Michigan. Can. J. For. Res. 14: 914–923. rates. Plant Ecol. 195: 179–196.
Hobbie, E.A., Tingey, D.T., Rygiewicz, P.T. et al. (2002). Hook, D.D. (1984). Waterlogging tolerance of lowland
Contributions of current year photosynthate to fine tree species of the South. South. J. Appl. For. 8:
roots estimated using a 13 C-depleted CO2 source. 136–149.
Plant Soil 247: 233–242. Hooper, D.U., Chapin, F.S. III, Ewel, J.J. et al. (2005).
Hobbs, R.J. (2007). Setting effective and realistic restora- Effects of biodiversity on ecosystem functioning:
tion goals: key directions for research. Restoration a consensus of current knowledge. Ecol. Monogr.
Ecology 15 (2): 354–357. 75: 3–35.
Hobbs, R.J. and Humphries, S.E. (1995). An Hooper, D.U., Adair, E.C., Cardinale, B.J. et al. (2012).
integrated approach to the ecology and A global synthesis reveals biodiversity loss as a
management of plant invasions. Conserv. Biol. 9: major driver of ecosystem change. Nature 486:
761–770. 105–108.
Hobbs, R.J., Arico, S., Aronson, J. et al. (2006). Novel Horn, M.H.S.B., Correa, P., Parolin, B.J.A. et al. (2011).
ecosystems: theoretical and management aspects of Seed dispersal by fishes in tropical and temperate
the new ecological world order. Glob. Ecol. Biogeogr. fresh waters: the growing evidence. Acta Oecol. 37:
15: 1–7. 561–577.
Hocker, H.W. Jr. (1956). Certain aspects of climate as Horton, K.W. (1956). The ecology of lodgepole pine in
related to the distribution of loblolly pine. Ecology Alberta and its role in forest succession. For. Br.
37: 824–834. Can. Tech. Note 45. 29 pp.
Hoegh-Guldberg, O., Hughes, L., McIntyre, S. et al. Hosie, R.C. (1969). Native Trees of Canada, 7e. Ottawa,
(2008). Assisted colonization and rapid climate Canada: Can. For. Serv. 380 pp.
change. Science 321: 345–346. Host, G.E., Pregitzer, K.S., Ramm, C.W. et al. (1987).
Holderidge, L.R. (1947). Determination of world plant Landform-mediated differences in successional
formations from simple climatic data. Science 105: pathways among upland forest ecosystems in
367–368. northwestern Lower Michigan. For. Sci. 33:
Holderidge, L.R. (1967). Life Zone Ecology. Tropical Sci. 445–457.
Center, San Jose, Costa Rica. 206 pp. Host, G.E., Pregitzer, K.S., Ramm, C.W. et al. (1988).
Hollender, C.A. and Dardick, C. (2015). Molecular basis Variation in over-story biomass among glacial land-
of angiosperm tree architecture. New Phytol. 206: forms and ecological land units in northwestern
541–556. Lower Michigan. Can. J. For. Res. 18: 659–668.
Holliday, J.A., Zhou, L., Bawa, R. et al. (2016). Evidence Hough, A.F. and Forbes, R.D. (1943). The ecology
for extensive parallelism but divergent genomic and silvics of forests in the high plateaus of
architecture of adaptation along altitudinal and Pennsylvania. Ecol. Monogr. 13: 299–320.
latitudinal gradients in Populus trichocarpa. New Houlton, B.Z., Morford, S.L., and Dahlgren, R.A. (2008).
Phytol. 209: 1240–1251. Convergent evidence for nitrogen sources in the
Holling, C.S. and Meffe, G.K. (1996). Command and Earth’s surface environment. Science 360: 58–62.
control and the pathology of natural resource Howe, H.F. (1993). Specialized and generalized dis-
management. Conserv. Biol. 10: 328–337. persal systems: where does “the paradigm” stand?
Holly, D.C., Ervin, G.N., Jackson, C.R. et al. (2009). Vegetatio 107/108: 3–13.
Effect of an invasive grass on ambient rates of Hsu, C.Y., Liu, Y., Luthe, D.S., and Yuceer, C. (2006).
decomposition and microbial community structure: Poplar FT2 shortens the juvenile phase and pro-
a search for causality. Biol. Invasions 11: 1855–1868. motes seasonal flowering. Plant Cell 18: 1846–1861.
Holmsgaard, E. and Bang, C. (1989). Loss of volume Hu, Z., Michaletz, S.T., Johnson, D.J. et al. (2018). Traits
increment due to cone production in Norway drive global wood decomposition rates more than
spruce. Forstl. Forsøgsvawsen i Danmark 42: climate. Glob. Chang. Biol. 24: 5259–5269.
217–231. Hubbes, M. (1999). The American elm and Dutch elm
Holsinger, L., Parks, S.A., and Miller, C. (2016). disease. For. Chron. 75: 265–273.
Weather, fuels, and topography impede wildland Huber, O. (1978). Light compensation point of vascular
fire spread in western US landscapes. For. Ecol. plants of a tropical cloud forest and an ecological
Manag. 380: 59–69. interpretation. Photosynthetica 12: 382–390.
Holthuijzen, A.M.A., Sharik, T.L., and Fraser, J.D. Huberty, A.F. and Denno, R.F. (2004). Plant water stress
(1987). Dispersal of eastern red cedar (Juniperus and its consequences for herbivorous insects: a new
virginiana) into pastures: an overview. Can. J. Bot. synthesis. Ecology 85: 1383–1398.
65: 1092–1095. Huenneke, L.F. and Sharitz, R.R. (1986). Microsite
Holzinger, B., Hulber, K., Camenisch, M., and Grab- abundance and distribution of woody seedlings in
herr, G. (2008). Changes in plant species richness a South Carolina cypress-tupelo swamp. Am. Midi.
over the last century in the eastern Swiss Alps: Nat. 115: 328–335.
684 References
Hufkens, K., Friedl, M.A., Keenan, T.F. et al. (2012). Ibáñez, I., Katz, D.S., and Lee, B.R. (2017). The contrast-
Ecological impacts of a widespread frost event fol- ing effects of short-term climate change on the early
lowing early spring leaf-out. Glob. Chang. Biol. 18: recruitment of tree species. Oecologia 184: 701–713.
2365–2377. Illick, J. S. (1914). Pennsylvania Trees. Pa. Dept. For.
Huggett, R.J. (1995). Geoecology, an Evolutionary Bull. 11, 231 pp.
Approach. New York: Routledge 320 pp. Illick, J.S. (1921). Replacement of the chestnut. J. For.
Hulme, P.E. (2009). Trade, transport and trouble: 19: 105–114.
managing invasive species pathways in an era of Intergovernmental Panel on Climate Change (2013).
globalization. J. Appl. Ecol. 46: 10–18. IPCC, 2013: Climate Change 2013: The Physical
Hunter, M.L. Jr. and Gibbs, J.P. (2021). Fundamen- Science Basis. In: Fifth Assessment Report of the
tals of Conservation Biology, 4e. New York: Wiley Intergovernmental Panel on Climate Change (ed.
Blackwell 672 pp. T.F. Stocker, D. Qin, G.-K. Plattner, et al.). Cam-
Hunter, A.F. and Lechowicz, M.J. (1992). Predicting the bridge, UK: Cambridge University Press.
timing of budburst in temperate trees. J. Appl. Ecol. Intergovernmental Panel on Climate Change (2014).
29: 597–604. Climate change 2014: Impacts, adaptation, and vul-
Hunter, M.L. and Schmiegelow, F.K. (2011). Wildlife, nerability. In: Fifth Assessment Report of the Inter-
Forests, and Forestry Principles of Managing Forests governmental Panel on Climate Change (ed. V.R.
for Biological Diversity, 2e. Upper Saddle River, New Barros, C.B. Field, D.J. Dokken, et al.). Cambridge,
Jersey, USA: Prentice Hall 259 pp. UK: Cambridge University Press.
Huntly, N. (1991). Herbivores and the dynamics of Intergovernmental Panel on Climate Change (2021).
communities and ecosystems. Annu. Rev. Ecol. Syst. Climate change 2021: The physical science basis.
22: 477–503. In: Sixth Assessment Report of the Intergovern-
Hupp, C.R. and Osterkamp, W.R. (1985). Bottom- mental Panel on Climate Change (ed. V.P. Masson-
land vegetation distribution along Passage Creek, Delmotte, P. Zhai, A. Pirani, et al.). Cambridge,
Virginia, in relation to fluvial landforms. Ecology UK: Cambridge University Press.
66: 670–681. Isaac, L. A. (1938). Factors affecting establishment of
Huston, M. (1979). A general hypothesis of species Douglas-fir seedlings. U.S. Dept. Agr. Circ. 486. 45 pp.
diversity. American Naturalist 113: 81–101. Ishii, H.T., Ford, E.D., and Kennedy, M.C. (2007).
Huston, M.A. (1993). Biological diversity, soils, and eco- Physiological and ecological implications of
nomics. Science 262: 1676–1680. adaptive reiteration as a mechanism for crown
Huston, M.A. (1994). Biological Diversity. Cambridge: maintenance and longevity. Tree Physiol. 27:
Cambridge Univ. Press 681 pp. 455–462.
Huston, M.A. (2014). Disturbance, productivity, and Iversen, C.M., Murphy, M.T., Allen, M.F. et al. (2012).
species diversity: empiricism vs. logic in ecological Advancing the use of minirhizotrons in wetlands.
theory. Ecology 95: 2382–2396. Plant Soil 352: 23–39.
Huston, M. and Smith, T. (1987). Plant succession: life Iverson, L.R., Dale, M.F., Scott, C.T., and Prasad, A.
history and competition. Am. Nat. 130: 168–198. (1997). A GIS-derived integrated moisture index to
Huston, M., DeAngelis, D., and Post, W. (1988). New predict forest composition and productivity of Ohio
computer models unify ecological theory. Biosci- forests (USA). Landsc. Ecol. 12: 331–348.
ence 38: 682–691. Iverson, L.R., Schwartz, M.W., and Prasad, A.M. (2004).
Hutchings, M.J., John, E.A., and Stewart, A.J.A. (ed.) How fast and far might tree species migrate in the
(2000). The Ecological Consequences of Environ- eastern United States due to climate change? Glob.
mental Heterogeneity. Oxford, UK: Blackwell Ecol. Biogeogr. 13: 209–219.
Science 434 pp. Iverson, L.R., Prasad, A.M., Matthews, S.N., and Peters,
Hutchins, H.E. and Lanner, R.M. (1982). The central M. (2008). Estimating potential habitat for 134
role of Clark’s nutcracker in the dispersal and estab- eastern US tree species under six climate scenarios.
lishment of whitebark pine. Oecologia 55: 192–201. For. Ecol. Manag. 254: 390–406.
Hutchinson, B.A. and Matt, D.R. (1976). Beam enrich- Iverson, L.R., Peters, M.P., Prasad, A.M., and Mat-
ment of diffuse radiation in a deciduous forest. thews, S.N. (2019a). Analysis of climate change
Agric. Meteorol. 18: 255–265. impacts on tree species of the eastern US: Results
of DISTRIB-II modeling. Forests 10: 302.
Huxley, J.S. (1938). Clines, an auxiliary taxonomic prin-
ciple. Nature 142: 219–220. Iverson, L.R., Prasad, A.M., Peters, M.P., and Mat-
thews, S.N. (2019b). Facilitating adaptive forest
Huxley, J.S. (1939). Clines, an auxiliary method in tax-
management under climate change: A spatially
onomy. Bijd. Dierkunde 27: 491–520.
specific synthesis of 125 species for habitat changes
Huxley, C.R. and Cutler, D.F. (1991). Ant–Plant Interac-
and assisted migration over the eastern United
tions. Oxford: Oxford Univ. Press 601 pp.
States. Forests 10: 989.
References 685
Ives, A.R. and Carpenter, S.R. (2007). Stability and Jaramillo-Correa, J.P., Beaulieu, J., Khasa, D.P., and
diversity of ecosystems. Science 317: 58–62. Bousquet, J. (2009). Inferring the past from the pre-
Jackson, L.W.R. (1952). Radial growth of forest trees in sent phylogeographic structure of North American
the Georgia Piedmont. Ecology 33: 336–341. forest trees: seeing the forest for the genes. Can. J.
Jackson, W.R. (1967). Effect of shade tolerance on leaf For. Res. 39: 286–307.
structure of deciduous tree species. Ecology 48: Jarvis, P.G. and Laverenz, J.W. (1983). Productivity of
498–499. temperate, deciduous, and evergreen forests. In:
Jackson, S.T. (2004). Late Quaternary biogeography: Encyclopedia of Plant Physiology, New Series, Vol.
linking biotic responses to environmental vari- 12D (ed. O.L. Lange, P.S. Nobel, C.B. Osmond and
ability across timescales. In: Frontiers of Biogeog- H. Ziegler), 233–280. New York: Springer-Verlag.
raphy (ed. M. Lomolino and L. Heaney), 47–65. Jenkins, C.N., Van Houtan, K.S., Pimm, S.L., and
Sunderland, MA: Sinauer Associates. Sexton, J.O. (2015). US protected lands mismatch
Jackson, R.B., Mooney, H.A., and Schulze, E.D. (1997). biodiversity priorities. Proc. Natl. Acad. Sci. 112:
A global budget for fine root biomass, surface area, 5081–5086.
and nutrient contents. Proc. Natl. Acad. Sci. 94: Jenny, H. (1941). Factors of Soil Formation. New York:
7362–7366. McGraw-Hill 281 pp.
Jackson, S.T., Webb, R.S., Anderson, K.H. et al. (2000). Jenny, H. (1980). The Soil Resource: Origin and Behavior.
Vegetation and environment in eastern North New York: Springer-Verlag 377 pp.
America during the last glacial maximum. Quart. Jenny, H., Arkley, R.J., and Schultz, A.M. (1969). The
Sci. Rev. 19: 489–508. pygmy forest-podsol ecosystem and its dune asso-
Jacobs, D.F., Cole, D.W., and McBride, J.R. (1985). Fire ciates of the Mendocino coast. Madroño 20: 60–74.
history and perpetuation of natural coast redwood Jiang, L. and Pu, Z.C. (2009). Different effects of species
ecosystems. J. For. 83: 494–497. diversity on temporal stability in single-trophic and
Jacobson, R.B., Miller, A.J., and Smith, J.A. (1989). The multitrophic communities. Am. Nat. 174: 651–659.
role of catastrophic geomorphic events in central Jiao, Y., Wickett, N.J., Ayyampalayam, S. et al. (2011).
Appalachian landscape evolution. Geomorphology Ancestral polyploidy in seed plants and angio-
2: 257–284. sperms. Nature 473: 97–100.
Jacoby, G.C., Williams, P.L., and Buckley, B.M. (1992). Jiggins, C.D. (2019). Can genomics shed light on the
Tree ring correlation between prehistoric landslides origin of species? PLoS Biol. 17: e3000394.
and abrupt tectonic events in Seattle, Washington. Jo, I., Fridley, J.D., and Frank, D.A. (2015). Linking
Science 258: 1621–1623. above-and belowground resource use strategies for
Jakubos, B. and Romme, W.H. (1993). Invasion of subal- native and invasive species of temperate deciduous
pine meadows by lodgepole pine in Yellowstone forests. Biol. Invasions 17: 1545–1554.
National Park, Wyoming, USA. Arct. Alp. Res. 25: Jo, I., Fridley, J.D., and Frank, D.A. (2017). Invasive
382–390. plants accelerate nitrogen cycling: evidence from
James, S. (1984). Lignotubers and burls: their struc- experimental woody monocultures. J. Ecol. 105:
ture, function, and ecological significance in 1105–1110.
Mediterranean ecosystems. Bot. Rev. 50: 225–245. Jochner, S. and Menzel, A. (2015). Urban phenological
Janos, D.P. (1987). VA mycorrhizas in humid tropical studies–Past, present, future. Environ. Pollut. 203:
systems. In: Ecophysiology of VA Mycorrhizal Plants 250–261.
(ed. G.R. Safer). Boca Raton, FL: CRC Press 224 pp. Johnsen, O., Daehlen, O.G., Østreng, G., and Skrøppa,
Janowiak, M.K., Swanston, C.W., Nagel, L.M. et al. T. (2005). Daylength and temperature during
(2014). A practical approach for translating cli- seed production interactively affect adaptive
mate change adaptation principles into forest performance of Picea abies progenies. New Phytol.
management actions. J. For. 112: 424–433. 168: 589–596.
Janzen, D.H. (1969). Seed-eaters versus seed size, Johnson, P.W. (1972). Factors affecting buttressing in
number, toxicity and dispersal. Evolution 23: 1–27. Triplochiton scleroxylon K. Schum. Ghana J. Agric.
Janzen, D.H. (1983). Dispersal of seeds by vertebrate Sci. 5: 13–21.
guts. In: Coevolution (ed. D.J. Futuyma and M. Slat- Johnson, E.A. (1992). Fire and Vegetation Dynamics:
kin), 232–262. Sunderland, MA: Sinauer Assoc. Studies from the North American Boreal Forest.
Janzen, D.H. (1985). The natural history of mutualisms. Cambridge: Cambridge Univ. Press 129 pp.
In: The Biology of Mutualism (ed. D.H. Boucher), Johnson, W.C. and Adkisson, C.S. (1985). Airlifting the
40–99. London: Croom Helm. oaks. Nat. Hist. 10: 41–46.
Janzen, D.H. and Martin, P.S. (1982). Neotropical ana- Johnson, E.A. and Miyanishi, K. (ed.) (2001). Forest
cronisms: the fruits the gomphotheres ate. Science Fires: Behavior and Ecological Effects. New York:
215: 19–27. Academic Press 594 pp.
686 References
Johnson, W.C. and Webb, T. III (1989). The role of blue Jouzel, J., Masson-Delmotte, V., Cattani, O. et al. (2007).
jays (Cyanocitta cristata L.) in the postglacial dis- Orbital and millennial Antarctic climate variability
persal of fagaceous trees in eastern North America. over the past 800,000 years. Science 317: 793–796.
J. Biogeogr. 16: 561–571. Jucker, T., Bouriaud, O., and Coomes, D.A. (2015).
Johnson, W.C., Burgess, R.L., and Keammerer, W.R. Crown plasticity enables trees to optimize canopy
(1976). Forest overstory vegetation and environ- packing in mixed-species forests. Funct. Ecol. 29:
ment on the Missouri River floodplain in North 1078–1086.
Dakota. Ecol. Monogr. 46: 59–84. Jull, L.G., Frank, A., Blazich, L., and Hinesley, E.
Johnson, M.G., Phillips, D.L., Tingey, D.T., and Storm, (1999). Seedling growth of Atlantic white-cedar as
M.J. (2000). Effects of elevated CO2, N-fertilization, influenced by photoperiod and day/night tempera-
and season on survival of ponderosa pine fine ture. J. Environ. Hortic. 17: 107–113.
roots. Can. J. For. Res. 30: 220–228. Jung, J.-H., Domijan, M., Klose, C. et al. (2016). Phyto-
Johnson, E.A., Miyanishi, K., and K. (2008). Testing the chromes function as thermosensors in Arabidopsis.
assumptions of chronosequences in succession. Science 354: 886–889.
Ecol. Lett. 11: 419–431. Jurasinski, G., Retzer, V., and Beierkuhnlein, C. (2009).
Johnson, P.S., Shifley, S.R., Rogers, R. et al. (2019). The Inventory, differentiation, and proportional diver-
Ecology and Silviculture of Oaks, 3e. Boston, MA: sity: a consistent terminology for quantifying
CAB International 648 pp. species diversity. Oecologia 159: 15–26.
Johnstone, J.F., Hollingsworth, T.N., Chapin, F.S. III, Jurik, T.W., Webber, J.A., and Gates, D.M. (1988).
and Mack, M.C. (2010). Changes in fire regime Effects of temperature and light on photosynthesis
break the legacy lock on successional trajectories of dominant species of a northern hardwood forest.
in Alaskan boreal forest. Glob. Chang. Biol. 16: Bot. Gaz. 149: 203–208.
1281–1295. Justus, J., Colyvan, M., Regan, H., and Maguire, L.
Johnstone, J.F., Allen, C.D., Franklin, J.F. et al. (2016). (2009). Buying into conservation: intrinsic versus
Changing disturbance regimes, ecological memory, instrumental value. Trends Ecol. Evol. 24: 187–191.
and forest resilience. Front. Ecol. Environ. 14: Kalberer, S.R., Wisniewski, M., and Arora, R. (2006).
369–378. Deacclimation and reacclimation of cold-hardy
Jolley, D.B., Ditchkoff, S.S., Sparklin, B.D. et al. (2010). plants: current understanding and emerging con-
Estimate of herpetofauna depredation by a cepts. Plant Sci. 171: 3–16.
population of wild pigs. J. Mammal. 91: 519–524. Kalela, E.K. (1957). Über Veränderungen in den Wur-
Jones, J.R. (1969). Review and comparison of site evalu- zelverhältnissen der Kiefernbestände im Laufe der
ation methods. USDA For. Serv. Res. Paper RM-51. Vegetationsperiode. Acta For. Fenn. 65: 1–41.
Rocky Mountain For. and Rge. Exp. Sta., Fort Col- Kamminga-Van Wijk, C. and Prins, H. (1993). The
lins, CO. 27 pp. kinetics of NH4+ and NO3− uptake by Douglas-fir
Jones, S.M. (1991). Landscape ecosystem classification from single N-solutions and from solutions con-
for South Carolina. In D.L. Mengel and D. T. Tew taining both NH4+ and NO3−. Plant Soil 151: 91–96.
(eds.), Symp. Proc., Ecological Land Classification: Kaplan, S. (1992). Environmental preference in a
Applications to Identify the Productive Potential knowledge-seeking, knowledge-using organism.
of Southern Forests. USDA For. Serv. Gen. Tech. In: The Adapted Mind: Evolutionary Psychology and
Report SE-68. Southeastern For. Exp. Sta., Ashe- the Generation of Culture (ed. J.H. Barkow, L. Cos-
ville, NC. mides and J. Tooby), 581–598. New York: Oxford
Jones, S.M. and Lloyd, F.T. (1993). Landscape ecosystem Univ. Press.
classification: the first step toward ecosystem Kashian, D.M. (2016). Sprouting and seed produc-
management in the southeastern United States. In: tion may promote persistence of green ash in the
Defining Sustainable Forestry (ed. G.H. Aplet, N. presence of the emerald ash borer. Ecosphere 7:
Johnson, J.T. Olson and V.A. Sample). Washington e01332.
D.C.: Island Press 231 pp. Kashian, D.M. and Barnes, B.V. (2021). Tree growth and
Jones, R.K., Pierpoint, G., Wickware, G.M. et al. (1983). survival over 61 years at the Second International
Field guide to forest ecosystem classification for Larch Provenance Test in southeastern Michigan,
the Clay Belt. Site Region 3E. Ont. Min. Nat. Res. USA. Silvae Genet. 70: 9–21.
Toronto, ON. 123 pp. Kashian, D.M. and Witter, J.A. (2011). Assessing the
Jordan, C., Caskey, W., Escalante, G. et al. (1982). The potential for ash canopy tree replacement via
nitrogen cycle in a “Terra Firme” rainforest on current regeneration following emerald ash borer-
oxisol in Amazon territory of Venezuela. Plant Soil caused mortality on southeastern Michigan land-
67: 325–332. scapes. For. Eco. Manag. 261: 480–488.
Jorgensen, J.R. and Hodges, C.S. Jr. (1970). Microbial Kashian, D.M., Barnes, B.V. (2000). Landscape
characteristics of a forest soil after twenty years of influence on the spatial and temporal distribution
prescribed burning. Mycologia 62: 721–726. of the Kirtland’s warbler at the Bald Hill burn,
References 687
northern Lower Michigan, USA. Can. J. For. Res. Kay, C.E. (1994). Aboriginal overkill, the role of Native
30: 1895–1904. Americans in structuring western ecosystems.
Kashian, D.M., Barnes, B.V., and Walker, W.S. (2003). Hum. Nat. 5: 359–398.
Landscape ecosystems of northern Lower Michigan Kay, C.E. (1997). Is aspen doomed? J. For. 95: 4–11.
and the occurrence and management of the Kirt- Kaye, M.W., Binkley, D., and Stohlgren, T.J. (2005).
land’s warbler. For. Sci. 49: 140–159. Effects of conifers and elk browsing on quaking
Kashian, D.M., Tinker, D.B., Turner, M.G., and Scar- aspen forests in the central Rocky Mountains, USA.
pace, F.L. (2004). Spatial heterogeneity of lodgepole Ecol. Appl. 15: 1284–1295.
pine sapling densities following the 1988 fires in Kayll, A.J. (1974). Use of fire in land management.
Yellowstone National Park, Wyoming, U.S.A. Can. In: Fire and Ecosystems (ed. T.T. Kozlowski and
J. For. Res. 34: 2263–2276. C.E. Ahlgren), 483–511. New York: Academic
Kashian, D.M., Turner, M.G., and Romme, W.H. Press.
(2005a). Variability in leaf area and stemwood Keane, R.E., Morgan, P., and Running, S.W. (1996).
increment along a 300-year lodgepole pine chrono- FIRE-BGC—a mechanistic ecological process
sequence. Ecosystems 8: 48–61. model for simulating fire succession on coniferous
Kashian, D.M., Turner, M.G., Romme, W.H., and forest landscapes of the northern Rocky Moun-
Lorimer, C.G. (2005b). Variability and conver- tains. USDA For. Serv. Res. Paper INT-RP-484. Int.
gence in stand structural development on a Res. Sta., Odgen, UT. 122 pp.
fire-dominated subalpine landscape. Ecology 86: Keane, R.E., Ryan, K.C., Veblen, T.T. et al. (2002). Cas-
643–654. cading Effects of Fire Exclusion in Rocky Mountain
Kashian, D.M., Romme, W.H., Tinker, D.B. et al. (2006). Ecosystems: A Literature Eeview. USDA For. Serv.
Carbon storage on landscapes with stand-replacing Gen. Tech. Rep. RMRS-GTR-91. Rocky Mountain
fires. Bioscience 56: 598–606. Res. Sta. Fort Collins, CO. 24 p.
Kashian, D.M., Romme, W.H., and Regan, C.M. (2007). Kearsley, M.J.C. and Whitham, T.G. (1989). Develop-
Reconciling divergent interpretations of quaking mental changes in resistance to herbivory: implica-
aspen decline on the northern Colorado Front tions for individuals and populations. Ecology 70:
Range. Ecol. Appl. 17: 1296–1311. 422–434.
Kashian, D.M., Corace, R.G. III, Shartell, L.M. et al. Keeley, J.E. and Keeley, S.C. (1988). Chaparral. In:
(2012). Variability and persistence of post-fire North American Terrestrial Vegetation (ed. M.G.
biological legacies in jack pine-dominated eco- Barbour and W.D. Billings), 203–254. Cambridge:
systems of northern Lower Michigan. For. Ecol. Cambridge Univ. Press.
Manag. 263: 148–158. Keeley, J.E. and Pausas, J.G. (2019). Distinguishing
Kashian, D.M., Romme, W.H., Tinker, D.B. et al. (2013). disturbance from perturbations in fire-prone eco-
Post-fire changes in forest carbon storage over systems. Int. J. Wildland Fire 28: 282–287.
a 300-year chronosequence of Pinus contorta- Keeling, E.G., Sala, A., and DeLuca, T.H. (2006). Effects
dominated forests. Ecol. Monogr. 83: 49–66. of fire exclusion on forest structure and compo-
Kashian, D.M., Sosin, J.R., Huber, P.W. et al. (2017). sition in unlogged ponderosa pine/Douglas-fir
A neutral modeling approach for designing forests. For. Ecol. Manag. 237: 418–428.
spatially heterogeneous jack pine plantations in Keenan, T.F., Hollinger, D.Y., Bohrer, G. et al. (2013).
northern Lower Michigan, USA. Landsc. Ecol. 32: Increase in forest water-use efficiency as atmo-
1117–1131. spheric carbon dioxide concentrations rise. Nature
Kauffman, J.B. and Krueger, W.C. (1984). Livestock 499: 324–327.
impacts on riparian ecosystems and streamside Keenan, T.F., Gray, J., Friedl, M.A. et al. (2014). Net
management implications: A review. J. Range carbon uptake has increased through warming-
Manag. 37: 430–437. induced changes in temperate forest phenology.
Kauffman, J.B. and Uhl, C. (1990). Interactions of Nat. Clim. Chang. 4: 598–604.
anthropogenic activities, fire, and rain forests in the Keesing, F. and Ostfeld, R.S. (2015). Is biodiversity good
Amazon Basin. In: Fire in the Tropical Biota. (Eco- for your health? Science 349: 235–236.
logical Studies 84) (ed. J.G. Goldammer), 117–134. Keesing, F. and Ostfeld, R.S. (2021). Impacts of biodi-
New York: Springer-Verlag. versity and biodiversity loss on zoonotic diseases.
Kauffman, J.B., Krueger, W.C., and Vavra, M. (1983). Proc. Natl. Acad. Sci. 118: e2023540118.
Effects of late season cattle grazing on riparian Keesing, F., Belden, L.K., Daszak, P. et al. (2010).
plant communities. Rangel. Ecol. Manag./J. Range Impacts of biodiversity on the emergence and
Manag. Archives 36: 685–691. transmission of infectious diseases. Nature 468
Kaufmann, J., Bork, E.W., Alexander, M.J., and Blenis, (7324): 647–652.
P.V. (2013). Habitat selection by cattle in foothill Keever, C. (1953). Present composition of some stands
landscapes following variable harvest of aspen of the former oak, chestnut forest in the southern
forest. For. Ecol. Manag. 306: 15–22. Blue Ridge Mountains. Ecology 34: 44–54.
688 References
Kellogg, L.F. (1939). Site index curves for plantation Kilgore, B.M. (1976c). From fire control to fire
black walnut in the Central States region. Central management: an ecological basis for policies. In
States For. Exp. Sta., Res. Note 35: 3 pp. Trans. 41st North American Wildlife and Natural
Kelly, D. and Sullivan, J. J. (1997). Quantifying the Resources Conference. Wildlife Management Insti-
benefits of mast seeding on predator satiation and tute, Washington, D.C.
wind pollination in Chionochloa pallens (Poaceae). Kilgore, B.M. and Sando, R.W. (1975). Crown-fire
78:143–150. potential in a sequoia forest after prescribed
Kelly, D., Hart, D.E., and Allen, R.B. (2001). Evaluating burning. For. Sci. 21: 83–87.
the wind pollination benefits of mast seeding. Killingbeck, K.T. (1996). Nutrients in senesced leaves:
Ecology 82: 117–126. keys to the search for potential resorption and
Kelsall, J.P., Telfer, E.S., and Wright, T.D. (1977). The resorption proficiency. Ecology 77: 1716–1727.
effects of fire on the ecology of the Boreal Forest, Kilvitis, H.J., Alvarez, M., Foust, C.M. et al. (2014).
with particular reference to the Canadian north: Ecological epigenetics. In: Ecological Genomics.
a review and selected bibliography. Can. Wildlife Advances in Experimental Medicine and Biology,
Serv., Ottawa. Occ. Paper 32. 58 pp. vol. 781 (ed. C. Landry and N. Aubin-Horth),
Kelsey, E.P., Cann, M.D., Lupo, K.M., and Haddad, L.J. 191–210. New York: Springer.
(2019). Synoptic to microscale processes affecting King, D.I. and Schlossberg, S. (2014). Synthesis of the
the evolution of a cold-air pool in a northern New conservation value of the early-successional stage
England forested mountain valley. J. Appl. Meteorol. in forests of eastern North America. For. Ecol.
Climatol. 58: 1309–1324. Manag. 324: 186–195.
Kemp, K.B., Higuera, P.E., Morgan, P., and Abatzoglou, Kinzig, A.P. (2009). Ecosystem services. In: The Prince-
J.T. (2019). Climate will increasingly determine ton Guide to Ecology (ed. S.A. Levin, S.R. Carpenter,
post-fire tree regeneration success in low-elevation H.C.J. Godfray, et al.), 573–578. Princeton, NJ:
forests, Northern Rockies, USA. Ecosphere 10: Princeton University Press.
e02568. Kitajima, K. (1994). Relative importance of photosyn-
Kemperman, J.A. and Barnes, B.V. (1976). Clone size in thetic traits and allocation patterns as correlates
American aspens. Can. J. Bot. 54: 2603–2607. of seedling shade tolerance of 13 tropical trees.
Kenoyer, L.A. (1933). Forest distribution in Oecologia 98: 419–428.
southwestern Michigan as interpreted from the Klappa, C.F. (1980). Rhizoliths in terrestrial carbonates:
original land survey (1826–32). Pap. Mich. Acad. classification. recognition, genesis and significance.
Sci. Arts Lett. 19: 107–111. Sedimentology 27: 613–629.
Keyes, M.R. and Grier, C.C. (1981). Above-and below- Klein Goldewijk, K. (2001). Estmating global land use
ground net production in 40–year-old Douglas-fir over the past 300 years: the HYDE database. Glob.
stands on low and high productivity sites. Can. J. Biogeochem. Cycles 15: 417–433.
For. Res. 11: 599–605. Kleinschmit, J. and Bastien, J.C. (1992). IUFRO’s role in
Khoshoo, T.N. (1959). Polyploidy in gymnosperms. Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco)
Evolution 13: 24–39. tree improvement. Silvae Genet. 41: 161–173.
Kikuzawa, K. (2003). Phenological and morphological Klijn, F. and Udo de Haes, H.A. (1994). A hierarchical
adaptations to the light environment in two woody approach to ecosystems and its implications
and two herbaceous plant species. Funct. Ecol. for ecological land classification. Landsc. Ecol.
17: 29–38. 9: 89–104.
Kilburn, P.D. (1960). Effects of logging and fire on xero- Kling, M.M. and Ackerly, D.D. (2021). Global wind
phytic forests in northern Michigan. Bull. Torrey patterns shape genetic differentiation, asymmetric
Bot. Club 6: 402–405. gene flow, and genetic diversity in trees. Proc. Natl.
Kilgore, B.M. (1972). Fire’s role in a sequoia forest. Acad. Sci. 118: e2017317118.
Naturalist 23: 26–37. Klinge, H. and Rodrigues, W.A. (1968a). Litter produc-
Kilgore, B.M. (1973). The ecological role of fire in tion in an area of Amazonian terra firme forest.
Sierran conifer forests: its application to national Part I. Litter-fall, organic carbon and total nitrogen
park management. Quat. Res. 3: 496–513. contents of litter. Amazoniana 1: 287–302.
Kilgore, B.M. (1975). Restoring fire to National Park Klinge, H. and Rodrigues, W.A. (1968b). Litter produc-
wilderness. Am. For. 81: 16–19. tion in an area of Amazonian terra firme forest.
Kilgore, B.M. (1976a). Fire management in the National Part II. Mineral nutrient content of the litter. Ama-
Parks: an overview. In Proc. Annual Tall Timbers zoniana 1: 303–310.
Fire Ecology Conference and Fire and Land Klinge, H., Rodrigues, W.A., Brunig, E., and Fittkau,
Management Symposium, 14:45–57. Tall Timbers E.J. (1975). Biomass and structure in a central
Res. Sta., Tallahassee, FL. Amazon rain forest. In: Tropical Ecological Systems:
Kilgore, B.M. (1976b). America’s renewable resource Trends in Terrestrial and Aquatic Research (ed.
potential—1975: the turning point. Proc. Soc. Am. F.B. Golley and E. Medina), 115–122. New York:
For. 1975: 178–188. Springer-Verlag.
References 689
Klinka, K., Krajina, V.J., Ceska, A., and Scagel, A.M. Komarek, E.V. (1971). Principles of fires ecology
(1989). Indicator Plants of Coastal British Columbia. and fire management in relation to the Alaskan
Vancouver: Univ, British Columbia Press. 288 pp. environment. In C.W. Slaughter, R.A. Barney,
Knapp, A.K. and Carter, G.A. (1998). Variability in leaf and G.M. Hansen (eds.), Symp. Proc. Fire in the
optical properties among 26 species from a broad Northern Environment. USDA For. Serv. Pacific
range of habitats. Am. J. Bot. 85: 940–946. Northwest For. and Range Exp. Sta., Portland, OR.
Knight, H. (1966). Loss of nitrogen from the forest floor Komarek, E.V. (1973). Ancient fires. In Proc. Annual
by burning. For. Chron. 42: 149–152. Tall Timbers Fire Ecology Conf. 12:219–240. Tall
Knight, D.H., Jones, G.P., Reiners, W.A., and Romme, Timbers Res. Sta., Tallahassee, FL.
W.H. (2014). Mountains and Plains, The Ecology Köppen, W. (1931). Grundriss der Klimakunde. Berlin:
of Wyoming Landscapes, 2e. New Haven, CT: Yale Walter de Gruyter 388 pp.
Univ. Press 404 pp. Körner, C. (1998). A re-assessment of high elevation
Knoepp, J.D., Turner, D.P., and Tingey, D.T. (1993). treeline positions and their explanation. Oecologia
Effects of ammonium and nitrate on nutrient uptake 115: 445–459.
and activity of nitrogen assimilation enzymes in Körner, C. (2017). A matter of tree longevity. Science
western hemlock. For. Ecol. Manag. 59: 179–191. 355: 130–131.
Knops, J.M., Koenig, W.D., and Carmen, W.J. (2007). Körner, C. and Paulsen, J. (2004). A world-wide study
Negative correlation does not imply a tradeoff of high-altitude treeline temperatures. J. Biogeogr.
between growth and reproduction in California 31: 713–732.
oaks. Proc. Natl. Acad. Sci. 104: 16982–16985. Korstian, C.F. and Coile, T.S. (1938). Plant competi-
Knowles, R. (1981). Denitrification. In: Soil Biochem- tion in forest stands. Duke Univ. School For. Bull. 3.
istry Vol. 5 (ed. E.A. Paul and J.N. Ladd), 315–329. 125 pp.
New York: Dekker. Korstian, C.F. and Stickel, P.W. (1927). The natural
Knowles, P. and Grant, M.C. (1985). Genetic variation replacement of blight-killed chestnut in the hard-
of lodgepole pine over time and microgeographical wood forests of the Northeast. J. Agric. Res. 34:
space. Can. J. Bot. 63: 722–727. 631–648.
Kobe, R.K. (1997). Carbohydrate allocation to storage as Koski, V. (1970). A study of pollen dispersal as a mech-
a basis of interspecific variation in sapling survivor- anism of gene flow in conifers. Comm. Inst. For.
ship and growth. Oikos 80: 226–233. Fenn. 70: 1–78.
Kobe, R.K., Pacala, S.W., Silander, J.A., and Canham, Koski, V. and Tallqvist, J.R. (1978). Results of long-time
C.D. (1995). Juvenile tree survivorship as a compo- measurements of the quantity of flowering and seed
nent of shade tolerance. Ecol. Appl. 5: 517–532. crop of forest trees. Folia Forestalia No. 364:1–60
Koch, J., Pearson, D.E., Huebner, C.D. et al. (2021). Kosola, K.R., Dickmann, D.I., Paul, E.A., and Parry,
Restoration of landscapes and habitats affected by D. (2001). Repeated insect defoliation effects on
established invasive species. In: Invasive Species growth, nitrogen acquisition, carbohydrates, and
in Forests and Rangelands of the United States (ed. root demography of poplars. Oecologia 129: 65–74.
T.M. Poland, T. Patel-Weynand, D.M. Finch, et al.), Kovač, M., Kutnar, L., and Hladnik, D. (2016). Assessing
185–202. New York: Springer. biodiversity and conservation status of the Natura
Koenig, W.D. and Knops, J.M.H. (1998). Scale of mast- 2000 forest habitat types: Tools for designated
seeding and tree-ring growth. Nature 396: 225–226. forestlands stewardship. For. Ecol. Manag. 359:
Kohm, K.A. and Franklin, J.F. (1997). Introduction. In: 256–267.
Creating a Forestry for the 21st Century (ed. K.A. Kovač, M., Gasparini, P., Notarangelo, M. et al. (2020).
Kohm and J.F. Franklin), 1–5. Washington, D.C.: Towards a set of national forest inventory indica-
Island Press. tors to be used for assessing the conservation status
Kolb, T.E. and Teulon, D.A.J. (1991). Relationship bet- of the habitats directive forest habitat types. J. Nat.
ween sugar maple budburst phenology and pear Conserv. 53: 125747.
thrips damage. Can. J. For. Res. 21: 1043–1048. Kozlowski, T.T. (1949). Light and water in relation to
Kolb, T.E., Fettig, C.J., Ayres, M.P. et al. (2016). growth and competition of Piedmont forest tree
Observed and anticipated impacts of drought on species. Ecol. Monogr. 19: 207–231.
forest insects and diseases in the United States. For. Kozlowski, T.T. (1971a). Growth and Development of
Ecol. Manag. 380: 321–334. Trees, I. Seed Germination, Ontogeny, and Shoot
Kolka, R.K., D’Amato, A.W., Wagenbrenner, J.W. et al. Growth. New York: Academic Press 443 pp.
(2018). Review of ecosystem level impacts of Kozlowski, T.T. (1971b). Growth and Development of
emerald ash borer on black ash wetlands: what Trees, II. Cambial Growth, Root Growth, and Repro-
does the future hold? Forests 9: 179. ductive Growth. New York: Academic Press 520 pp.
Kollas, C., Körner, C., and Randin, C.F. (2014). Spring Kozlowski, T.T. and Ahlgren, C.E. (ed.) (1974). Fire and
frost and growing season length co-control the cold Ecosystems. New York: Academic Press 542 pp.
range limits of broad-leaved trees. J. Biogeogr. 41: Kozlowski, T.T. and Keller, T. (1966). Food relations of
773–783. woody plants. Bot. Rev. 32: 293–382.
690 References
Kozlowski, T.T. and Pallardy, S.G. (1997a). Physiology of concentration, and inorganic phosphorus. Plant
Woody Plants, 2e. San Diego, CA: Academic Press Soil 105: 3–10.
411 pp. Krzic, M., Newman, R.F., Broersma, K., and Bomke,
Kozlowski, T.T. and Pallardy, S.G. (1997b). Growth A.A. (1999). Soil compaction of forest planta-
Control in Woody Plants. San Diego, CA: Academic tions of interior British Columbia. Rangel. Ecol.
Press 641 pp. Manag./J. Range Manag. Archives 52 (6): 671–677.
Kozlowski, T.T. and Scholtes, W.H. (1948). Growth of Kubiszewski, I., Costanza, R., Anderson, S., and Sutton,
roots and root hairs of pine and hardwood seed- P. (2020). The future value of ecosystem services:
lings in the Piedmont. J. For. 46: 750–754. Global scenarios and national implications. Ecosyst.
Kozlowski, T.T., Kramer, P.J., and Pallardy, S.G. Serv. 26: 289–301.
(1991). The Physiological Ecology of Woody Plants. Küchler, A. W. (1964). The potential natural vegetation
New York: Academic Press 657 pp. of the conterminous United States. Am. Geogr. Soc.
Kraft, N.J., Adler, P.B., Godoy, O. et al. (2015). Spec. Publ. No. 36. 154 pp.
Community assembly, coexistence and the environ- Kulakowski, D. and Jarvis, D. (2013). Low-severity fires
mental filtering metaphor. Funct. Ecol. 29: 592–599. increase susceptibility of lodgepole pine to moun-
Kramer, P.J. (1943). Amount and duration of growth of tain pine beetle outbreaks in Colorado. For. Ecol.
various species of tree seedlings. Plant Physiol. 18: Manag. 289: 544–550.
239–251. Kulakowski, D. and Veblen, T.T. (2007). Effect of
Kramer, P.J. (1957). Some effects of various combina- prior disturbances on the extent and severity of
tions of day and night temperatures and photope- wildfire in Colorado subalpine forests. Ecology 88:
riod on the growth of loblolly pine seedlings. For. 759–769.
Sci. 3: 45–55. Kulakowski, D., Jarvis, D., Veblen, T.T., and Smith, J.
Kramer, P.J. (1969). Plant and Soil Water Relationships: A (2012). Stand-replacing fires reduce susceptibility
Modern Synthesis. New York: McGraw-Hill 482 pp. to mountain pine beetle outbreaks in Colorado. J.
Kramer, P.J. (1983). Water Relations of Plants. Biogeogr. 39: 2052–2060.
New York: Academic Press 489 pp. Kulakowski, D., Matthews, C., Jarvis, D., and Veblen,
Kramer, P.J. and Boyer, J.S. (1995). Water Relations of T.T. (2013). Compounded disturbances in sub-
Plants and Soils. New York: Academic Press 495. alpine forests in western Colorado favour future
Kramer, P.J. and Clark, W.S. (1947). A comparison of dominance by quaking aspen Populus tremuloides.
photosynthesis in individual pine needles and J. Veg. Sci. 24: 168–176.
entire seedlings at various light intensities. Plant Kulakowski, D., Kaye, M.W., and Kashian, D.M. (2013).
Physiol. 22: 51–57. Long-term aspen cover change in the western US.
Kramer, P.J. and Decker, J.P. (1944). Relation between For. Ecol. Manag. 299: 52–59.
light intensity and rate of photosynthesis of loblolly Kulakowski, D., Veblen, T.T., and Bebi, P. (2016). Fire
pine and certain hardwoods. Plant Physiol. 19: severity controlled susceptibility to a 1940s spruce
350–358. beetle outbreak in Colorado, USA. PLoS One 11:
Krawchuk, M.A., Haire, S.L., Coop, J. et al. (2016). e0158138.
Topographic and fire weather controls of fire Kumschick, S., Gaertner, M., Vila, M. et al. (2015).
refugia in forested ecosystems of northwestern Ecological impacts of alien species: Quantification,
North America. Ecosphere 7: e01632. scope, caveats, and recommendations. Bioscience
Krawchuk, M.A., Meigs, G.W., Cartwright, J.M. et al. 65: 55–63.
(2020). Disturbance refugia within mosaics of Kunkel, K.E., Palecki, M., Ensor, L. et al. (2009). Trends
forest fire, drought, and insect outbreaks. Front. in twentieth-century U.S. snowfall using a quality-
Ecol. Environ. 18: 235–244. controlled dataset. J. Atmos. Ocean. Technol. 26:
Krebs, C.J. (1989). Ecological Methodology. New York: 33–44.
Harper and Row 654 pp. Kupfer, J.A., Myers, A.T., McLane, S.E., and Melton,
Krebs, C.J. (1998). Ecological Methodology, 2e. San Fran- G.N. (2008). Patterns of forest damage in a
cisco: Benjamin Cummings 624 pp. southern Mississippi landscape caused by Hurri-
cane Katrina. Ecosystems 11: 45–60.
Kreyling, J., Buhk, C., Backhaus, S. et al. (2014). Local
adaptations to frost in marginal and central popula- Kurz, W.A., Dymond, C.C., Stinson, G. et al. (2008a).
tions of the dominant forest tree Fagus sylvatica L. Mountain pine beetle and forest carbon feedback to
as affected by temperature and extreme drought climate change. Nature 452: 987–990.
in common garden experiments. Ecol. Evol. 4: Kurz, W.A., Stinson, G., Rampley, G.J. et al. (2008b).
594–605. Risk of natural disturbances makes future con-
Kriebel. H.B. (1957). Patterns of genetic variation in tribution of Canada’s forests to the global carbon
sugar maple. Ohio Agr. Exp. Sta. Res. Bull. 791. 56 pp. cycle highly uncertain. Proc. Natl. Acad. Sci. 105:
1551–1555.
Kroehler, C.J. and Linkins, A.E. (1988). The root
surface phosphatases of Eriophorum vagi- Kusbach, J.V.A. and Mikeska, M. (2003). Czech forest
natum: effects of temperature, pH, substrate ecosystem classification. J. For. Sci. 49: 85–93.
References 691
Kvaalen, H. and Johnsen, Ø. (2008). Timing of bud set Lanner, R.M. (1982). Adaptations of whitebark pine for
in Picea abies is regulated by a memory of temper- seed dispersal by Clark’s nutcracker. Can. J. For.
ature during zygotic and somatic embryogenesis. Res. 12: 391–402.
New Phytol. 177: 49–59. Lanner, R.M. (1990). Biology, taxonomy, evolution,
Ladefoged, K. (1952). The periodicity of wood and geography of stone pines of the world. In
formation. Copenh. Biol. Skr. 7 (3): 1–98. W. Schmidt and K. McDonald (comps.), Proc.
Lahti, T. (1995). Understorey vegetation as an indicator Symp. Whitebark Pine Ecosystems: Ecology and
of forest site potential in southern Finland. Finnish Management of a High-Mountain Resource, USDA
Soc: For. Sci., Finnish For. Res. Inst 68 pp. For. Serv. Gen. Tech. Report INT-270, Intermoun-
LaMarche, V.C. and Hirschboeck, K.K. (1984). Frost tain Exp. Sta., Odgen, UT.
rings in trees as records of major volcanic erup- Lapin, M. (1990). The Landscape Ecosystem Groups
tions. Nature 307: 121–126. of the University of Michigan Biological Station:
Lamarque, L.J., Delzon, S., and Lortie, C.J. (2011). Tree Classification, Mapping, and Analysis of Ecological
invasions: a comparative test of the dominant Diversity. Master’s Thesis, University of Michigan,
hypotheses and functional traits. Biol. Invasions 13: School of Natural Resources. 140 pp.
1969–1989. Lapin, M. and Barnes, B.V. (1995). Using the landscape
Lambers, H. and Oliveira, R.S. (2019). Plant ecosystem approach to assess species and eco-
Physiological Ecology, 3e. New York: Springer- system diversity. Conserv. Biol. 9: 1148–1158.
Verlag 763 pp. Lapointe, M. and Brisson, J. (2011). Tar spot disease
Lambert, J.M. and Dale, M.B. (1964). The use of on Norway maple in North America: quantifying
statistics in phytosociology. Adv. Ecol. Res. 2: 55–99. the impacts of a Reunion between an invasive
Lämke, J. and Bäurle, I. (2017). Epigenetic and tree species and its adventive natural enemy in an
chromatin-based mechanisms in environmental urban forest. Ecoscience 18: 63–69.
stress adaptation and stress memory in plants. Larcher, W. (1969). The effect of environmental and
Genome Biol. 18: 124. physiological variables on the carbon dioxide gas
Lamothe, K.A., Somers, K.M., and Jackson, D.A. (2019). exchange of trees. Photosynthetica 3: 167–198.
Linking the ball-and-cup analogy and ordination Larcher, W. (1995). Physiological Plant Ecology, 3e.
trajectories to describe ecosystem stability, resis- New York: Springer-Verlag 506 pp.
tance, and resilience. Ecosphere 10: e02629. Larcher, W. (2003). Physiological Plant Ecology, 4e.
Landres, P.B., Morgan, P., and Swanson, F.J. (1999). New York: Springer-Verlag 534 pp.
Overview of the use of natural variability concepts Larcher, W. and Bauer, H. (1981). Ecological sig-
in managing ecological systems. Ecol. Appl. 9: nificance of resistance to low temperature. In:
1179–1188. Physiological Plant Ecology I, Responses to the
Landsberg, J.J. (1995). Forest canopies. In: Encyclopedia Physical Environment, New Series, vol. 12A (ed.
of Environmental Biology (ed. W.A. Nierenberg). O.L. Lange, P.S. Nobel, C.B. Osmond and H.
New York: Academic Press. Ziegler). New York: Springer-Verlag.
Lang, G.A., Early, J.D., Martin, G.C., and Darnell, Larsen, B.B., Miller, E.C., Rhodes, M.K., and Wiens,
R.L. (1987). Endo-, Para-, and eco dormancy: J.J. (2017). Inordinate fondness multiplied and
physiological terminology and classification for redistributed: the number of species on earth and
dormancy research. Hortic. Sci. 22: 371–378. the new pie of life. Q. Rev. Biol. 92: 229–265.
Langford, A.N. and Buell, M.F. (1969). Integration, Larson, P.R. (1963b). Stem form development of forest
identity and stability in the plant association. Adv. trees. For. Sci. Monogr. 5. 42 pp.
Ecol. Res. 6: 83–135. Larson, M.M. (1967). Effect of temperature on initial
Langlet, O. (1959). A cline or not a cline—a question of development of ponderosa pine seedlings from
scots pine. Silvae Genet. 8: 13–22. three sources. For. Sci. 13: 286–294.
Langlet, O. (1971). Two hundred years’ genecology. Larson, M.M. (1970). Root regeneration and early
Taxon 20: 653–721. growth of red oak seedlings: influence of soil
Langston, N. (1995). Forest Dreams, Forest Nightmares, temperature. For. Sci. 16: 442–446.
the Paradox of Old Growth in the Inland West. Larsson, S. (1989). Stressful times for the plant stress—
Seattle: Univ. Washington Press 368 pp. insect performance hypothesis. Oikos 56: 277–283.
Lanner, R.M. (1966). Needed: a new approach to the Lassoie, J.P. (1982). Physiological processes in Douglas-fir.
study of pollen dispersion. Silvae Genet. 15: 50–52. In: Analysis of Coniferous Forest Ecosystems in the
Lanner, R.M. (1980). Avian seed dispersal as a factor in Western United States. US/IBP Synthesis Series
the ecology and evolution of limber and whitebark (ed. R.L. Edmonds), 126–185. Stroudsburg. PA:
pines. In B.P. Dancik and K.O. Higginbotham Dowden, Hutchinson and Ross.
(eds.), Proc. Sixth North American For. Biol. Work- Latham, R.E. and Ricklefs, R.E. (1993a). Global patterns
shop. Univ. Alberta, Edmonton, Alberta, Canada. of tree species richness in moist forests: energy-
Lanner, R.M. (1981). The Piñon Pine: A Natural and diversity theory does not account for variation in
Cultural History. Reno: Univ, Nevada Press 208 pp. species richness. Oikos 67: 325–333.
692 References
Latham, R.E. and Ricklefs, R.E. (1993b). Continental Lenoir, J., Gégout, J.C., Marquet, P.A. et al. (2008). A
comparisons of temperate-zone tree species diver- significant upward shift in plant species optimum
sity. In: Species Diversity in Ecological Communities: elevation during the 20th century. Science 320:
Historical and Geographical Perspectives (ed. R. 1768–1771.
Ricklefs and D. Schluter), 294–313. Chicago: Univ. Leopold, A. (1949). A Sand County Almanac. Reprinted
Chicago Press. 1966. Oxford Univ. Press, New York. 189 pp.
Lavender, D.P., and Overton, W.S. (1972). Thermope- Lesser, M.R. and Jackson, S.T. (2013). Contributions
riods and soil temperatures as they affect growth of long-distance dispersal to population growth in
and dormancy of Douglas-fir seedlings of different colonising Pinus ponderosa populations. Ecol. Lett.
geographic origin. Oregon State Univ., For. Res. 16: 380–389.
Lab. Res. Paper 13. 26 pp. Letourneau, D.K., Jedlicka, J.A., Bothwell, S.G., and
Lavoie, N., Venzina, L.-P., and Margolis, H. (1992). Moreno, C.R. (2009). Effects of natural enemy
Absorption and assimilation of nitrate and biodiversity on the suppression of arthropod her-
ammonium ions by jack pine seedlings. Tree bivores in terrestrial ecosystems. Annu. Rev. Ecol.
Physiol. 11: 171–183. Evol. Syst. 40: 573–592.
Lawrence, W.H. and Rediske, J.H. (1962). Fate of sown Levin, S.A. (2012). The challenge of sustainability:
Douglas-fir seed. For. Sci. 8: 210–218. lessons from an evolutionary perspective. In: Sus-
Lawton, J.H. (1993). Range, population abundance and tainability Science: The Emerging Paradigm and the
conservation. Trends Ecol. Evol. 8: 409–413. Urban Environment (ed. M.P. Weinstein and R.E.
Layser, E.F. (1974). Vegetative classification: its applica- Turner). New York: Springer.
tion to forestry in the northern Rocky Mountains. Levine, J.M., Adler, P.B., and Yelenik, S.G. (2004). A
J. For. 73: 354–357. meta-analysis of biotic resistance to exotic plant
Leak, W.B. (1975). Age distribution in virgin red spruce invasions. Ecol. Lett. 7: 975–989.
and northern hardwoods. Ecology 56: 1451–1454. Levins, R. (1969). Some demographic and genetic
Leak, W. B. (1978). Relationship of species and site consequences of environmental heterogeneity
index to habitat in the White Mountains of New for biological control. Bull. Entomol. Soc. Am. 15:
Hampshire. USDA For. Serv. Res. Paper NE-397. 237–240.
Northeastern For. Exp. Sta., Broomall, PA. 9 pp. Levins, R. (1970). Extinction. In: Some Mathematical
Leak, W. B. (1982). Habitat mapping and interpretation in Problems in Biology (ed. M. Gerstenhaber). RI:
New England. USDA For. Serv. Res. Paper NE-496. American Mathematical Society, Providence.
Northeastern For. Exp. Sta. Upper Darby, PA. 10 pp. Levitt, J. (1980a). Responses of Plants to Environmental
Ledig, F.T. (1992). Human impacts on genetic diversity Stresses. Vol. 1. Chilling, Freezing, and High Temper-
in forest ecosystems. Oikos 63: 87–108. ature Stress. New York: Academic Press 497 pp.
Ledig, F.T. and Fryer, J.H. (1972). A pocket of variability Levitt, J. (1980b). Responses of Plants to Environmental
in Pinus rigida. Evolution 26: 259–266. Stresses. Vol. 2. Water, Radiation, Salt, and Other
Ledig, F.T. and Perry, T.O. (1969). Net assimilation rate Stresses. New York: Academic Press 607 pp.
and growth in loblolly pine seedlings. For. Sci. 15: Levsen, N.D. and Mort, M.E. (2009). Inter-simple
431–438. sequence repeat (ISSR) and morphological var-
Lee, B.R. and Ibáñez, I. (2021a). Spring phenological iation in the western North American range of
escape is critical for the survival of temperate tree Chrysosplenium tetrandrum (Saxifragaceae). Botany
seedlings. Funct. Ecol. 35: 1848–1861. 87 (8): 780–790.
Lee, B.R. and Ibáñez, I. (2021b). Improved phenological Lev-Yadun, S. and Sprugel, D. (2011). Why should trees
escape can help temperate tree seedlings maintain have natural root grafts? Tree Physiol. 31: 575–578.
demographic performance under climate change Lexer, C., Fay, M.F., Joseph, J.A. et al. (2005). Barrier
conditions. Glob. Chang. Biol. 27: 3883–3897. to gene flow between two ecologically divergent
Lefcheck, J.S., Byrnes, J.E., Isbell, F. et al. (2015). Biodi- Populus species, P. alba (white poplar) and P.
versity enhances ecosystem multifunctionality across tremula (European aspen): the role of ecology and
trophic levels and habitats. Nat. Commun. 6: 1–7. life history in gene introgression. Mol. Ecol. 14:
Legris, M., Klose, C., Burgie, E.S. et al. (2016). Phyto- 1045–1057.
chrome B integrates light and temperature signals Li, W.-L., Berlyn, G.P., and Ashton, P.M.S. (1996).
in Arabidopsis. Science 354: 897–900. Polyploids and their structural and physiological
Lemieux, G.J. (1965). Soil-vegetation relationships in characteristics relative to water deficit in Betula
the northern hardwoods of Quebec. In: Forest–Soil papyrifera (Betulaceae). Am. J. Bot. 83: 15–20.
Relationships in North America (ed. C.T. Young- Li, J., Li, G., Wang, H., and Wang, D.X. (2011). Phyto-
berg). Corvallis: Oregon State Univ. Press. chrome signaling mechanisms. Arabidopsis Book
Lennartz, M.R. (1988). The red-cockaded woodpecker: 9: e0148.
old-growth species in a second-growth landscape. Liang, J., Watson, J.V., Zhou, M., and Lei, X. (2016a).
Nat. Areas J. 8: 160–165. Effects of productivity on biodiversity in forest
References 693
ecosystems across the United States and China. Lindenmayer, D.B. and Franklin, J.F. (ed.) (2003).
Conserv. Biol. 30: 308–317. Towards Forest Sustainability. Washington, DC:
Liang, J., Crowther, T.W., Picard, N. et al. (2016b). Island Press 212 pp.
Positive biodiversity-productivity relationship pre- Lindenmayer, D., Franklin, J., Lõhmus, A. et al. (2012).
dominant in global forests. Science 354: 6309. A major shift to the retention approach for forestry
Liao, C., Peng, R., Luo, Y. et al. (2008). Altered eco- can help resolve some global forest sustainability
system carbon and nitrogen cycles by plant inva- issues. Conserv. Lett. 5: 421–431.
sion: a meta-analysis. New Phytol. 177: 706–714. Linder, S. (1971). Photosynthetic action spectra of scots
Lichti, N.I., Steele, M.A., and Swihart, R.K. (2017). Seed pine needles of different ages from seedlings grown
fate and decision-making processes in scatter- under different nursery conditions. Physiol. Plant.
hoarding rodents. Biol. Rev. 92: 474–504. 25: 58–63.
Lieberman, M., Lieberman, D., and Peralta, R. (1989). Linder, S. and Axelsson, B. (1982). Changes in carbon
Forests are not just Swiss cheese: canopy stereoge- uptake and allocation patterns as a result of irri-
ometry of non-gaps in tropical forests. Ecology 70: gation and fertilization in a young Pinus sylvestris
550–552. stand. In R. H. Waring (ed.), Carbon Uptake and
Liebhold, A.M., MacDonald, W.L., Bergdahl, D., and Allocation: Key to Management of Subalpine Forest
Mastro, V.C. (1995). Invasion by exotic forest pests: a Ecosystems. IUFRO Workshop, Forest Resources
threat to forest ecosystems. For. Sci. Monogr. 30: 49 pp. Laboratory, Oregon State Univ., Corvallis.
Liebhold, A.M., McCullough, D.G., Blackburn, L.M. Linder, S., McDonald, J., and Lohammar, T. (1981).
et al. (2013). A highly aggregated geographical dis- Effects of nitrogen status and irradiance during cul-
tribution of forest pest invasions in the USA. Divers. tivation on photosynthesis and respiration of birch
Distrib. 19: 1208–1216. seedlings. Energy Forest Project Tech. Rep. 12.
Liebhold, A.M., Brockerhoff, E.G., Kalisz, S. et al. Swedish Univ. Agricultural Science, Uppsala,
(2017). Biological invasions in forest ecosystems. Sweden. 19 pp.
Biol. Invasions 19: 3437–3458. Lindsey, A. A., and Escobar, L. K. (1976). Eastern
Lieth, H. (1973). Primary production in terrestrial eco- Deciduous Forest, II, Beech-Maple Region. U.S. Dep.
systems. Hum. Ecol. 1: 303–332. Interior, Natl. Park Serv., Nat. Hist. Theme Stud.
Ligon, J.D. (1978). Reproductive interdependence of No. 3. NPS Publ. No. 148. 238 pp.
Pinon jays and pinon pines. Ecol. Monogr. 48: Lindsey, A.A. and Sawyer, J.O. Jr. (1970). Vegetation-
111–126. climate relationships in the eastern United States.
Ligon, J.D. and Stacey, P.B. (1995). Land use, lag times Proc. Indiana Acad. Sci. 80: 210–214.
and the detection of demographic change: the Lindsey, A.A., Petty, R.O., Sterling, D.K., and Van
case of the acorn woodpecker. Conserv. Biol. 10: Asdall, W. (1961). Vegetation and environment
840–846. along the Wabash and Tippecanoe Rivers. Ecol.
Likens, G.E., Bormann, F.H., Pierce, R.S. et al. (1977). Monogr. 31: 105–156.
Biogeochemistry of a Forested Ecosystem. New York: Linhart, Y.B. (1989). Interactions between genetic
Springer-Verlag 146 pp. and ecological patchiness in forest trees and their
Likens, G.E., Bormann, F.H., Pierce, R.S., and Reiners, dependent species. In: Evolutionary Ecology of
W.A. (1978). Recovery of a deforested ecosystem. Plants (ed. J.H. Bock and Y.B. Linhart), 393–430.
Science 199: 492–496. Boulder, CO: Westview Press.
Lind, B.M., Friedline, C.J., Wegrzyn, J.L. et al. (2017). Linhart, Y.B., Mitton, J.B., Sturgeon, K.B., and Davis,
Water availability drives signatures of local M.L. (1981). Genetic variation in space and time
adaptation in whitebark pine (Pinus albicaulis in a population of ponderosa pine. Heredity 48:
Engelm.) across fine spatial scales of the Lake 407–426.
Tahoe Basin, USA. Mol. Ecol. 26: 3168–3185. Lippincott, C.L. (2000). Effects of Imperata cylindrica
Lindberg, S.E. and Garten, C.T. (1988). Sources of (L.) Beauv. (Cogongrass) invasion on fire regime in
sulphur in forest canopy throughfall. Nature 336: Florida sandhill (USA). Nat. Areas J. 20: 140–149.
148–151. Litschert, S.E., Brown, T.C., and Theobald, D.M.
Lindenmayer, D.B. and Cunningham, S.A. (2013). Six (2012). Historic and future extent of wildfires in
principles for managing forests as ecologically sus- the southern Rockies Ecoregion, USA. For. Ecol.
tainable ecosystems. Landsc. Ecol. 28: 1099–1110. Manag. 269: 124–133.
Lindenmayer, D.B. and Fischer, J. (2013). Habitat Little, S. (1974). Effects of fire on temperate forests:
Fragmentation and Landscape Change: An Ecolog- northeastern United States. In: Fire and Ecosystems
ical and Conservation Synthesis. Washington, DC: (ed. T.T. Kozlowski and C.E. Ahlgren), 225–250.
Island Press 352 pp. New York: Academic Press.
Lindenmayer, D.B. and Franklin, J.F. (2002). Conserving Little, E.L., Jr. (1979). Checklist of United States Trees.
Forest Biodiversity: A Comprehensive Multiscaled USDA For. Serv. Handbook No. 541, Washington
Approach. Washington, DC: Island Press 352 pp. D.C. 374 pp.
694 References
Little, E.L. Jr., Brinkman, A., and McComb, A.L. (1957). saccharum) seedlings grown in various light inten-
Two natural Iowa hybrid poplars. For. Sci. 3: sities. Physiol. Plant. 22: 104–116.
253–262. Logan, J.A. and Powell, J.A. (2001). Ghost forests, global
Little, S. and Moore, E.B. (1949). The ecological role warming and the mountain pine beetle (Coleop-
of prescribed burns in the pine-oak forests of tera: Scolytidae). Am. Entomol. 47: 160–173.
southern New Jersey. Ecology 30: 223–233. Logan, J.A., Macfarlane, W.W., and Willcox, L. (2010).
Litton, C.M., Ryan, M.G., Knight, D.H., and Stahl, P.D. Whitebark pine vulnerability to climate-driven
(2003). Soil-surface carbon dioxide efflux and mountain pine beetle disturbance in the greater
microbial biomass in relation to tree density 13 Yellowstone ecosystem. Ecol. Appl. 20: 895–902.
years after a stand replacing fire in a lodgepole pine Long, J.N. (1980). Productivity of western coniferous
ecosystem. Glob. Chang. Biol. 9: 680–696. forests. In: Analysis of Coniferous Forest Ecosys-
Liu, H., Bauer, L.S., Gao, R. et al. (2003). Exploratory tems in the Western United States. US/IBP Synthesis
survey for the emerald ash borer, Agrilus planipen- Series (ed. R.L. Edmonds), 89–125. Stroudsburg,
nis (Coleoptera: Buprestidae), and its natural PA: Dowden, Hutchinson and Ross.
enemies in China. Gt. Lakes Entomol. 36: 11. Lonsdale, W.M. (1990). The self-thinning rule: dead or
Liu, H., Platt, S.G., and Borg, C.K. (2004). Seed dispersal alive? Ecology 71: 1373–1388.
by the Florida box turtle (Terrapene carolina bauri) Loope, L., Duever, M., Herndon, A. et al. (1994). Hur-
in pine Rockland forests of the lower Florida keys, ricane impact on uplands and freshwater swamp
United States. Oecologia 138: 539–546. forest. BioScience 44: 238–246.
Liu, J.G., Han, X., Yang, T. et al. (2019). Genome-wide Loreau, M. and De Mazancourt, C. (2013). Biodiversity
transcriptional adaptation to salt stress in Populus. and ecosystem stability: a synthesis of underlying
BMC Plant Biol. 19: 367. mechanisms. Ecol. Lett. 16: 106–115.
Lloyd, A.H. and Fastie, C.L. (2002). Spatial and Loreau, M., Naeem, S., Inchausti, P. et al. (2001). Biodiver-
temporal variability in the growth and climate sity and ecosystem functioning: current knowledge
response of treeline trees in Alaska. Clim. Chang. and future challenges. Science 294: 804–808.
52: 481–509. Lorimer, C.G. (1977). The presettlement forest and
Lloyd, W.J. and Lemmon, P.E. (1970). Rectifying azi- natural disturbance cycle of northeastern Maine.
muth (of aspect) in studies of soil-site index rela- Ecology 58: 139–148.
tionships. In: Tree Growth and Forest Soils (ed. C.T. Lorimer, C.G. (1980). Age structure and disturbance
Youngberg and C.B. Davey), 435–448. Corvallis: history of a southern Appalachian virgin forest.
Oregon State Univ. Press. Ecology 61: 1169–1184.
Loach, K. (1967). Shade tolerance on tree seedling. I. Lorimer, C.G. (1984). Development of the red maple
Leaf photosynthesis and respiration in plants raised understory in northeastern oak forests. For. Sci. 30:
under artificial shade. New Phytol. 66: 607–621. 3–22.
Loehle, C. (1986). Phototropism of whole trees: effects Lorimer, C.G. (1993). Causes of the oak regeneration
of habitat and growth form. Am. Midi. Nat. 116: problem. In D. Loftis, D. and C.E. McGee (eds.),
190–196. Symp. Proc., Oak Regeneration: Serious Problems,
Loehle, C. and Jones, R.H. (1990). Adaptive significance Practical Recommendations. USDA For. Serv. Gen.
of root grafting in trees. Funct. Ecol. 4: 268–271. Tech. Report SE-84. Southeastern For. Exp. Sta.,
Loehle, C. and Wein, G. (1994). Landscape hab- Asheville, NC.
itat diversity: a multiscale information theory Lorimer, C.G., Frelich, L.E., and Nordheim, E.V. (1988).
approach. Ecol. Model. 73: 311–329. Estimating gap origin probabilities for canopy
Logan, K.T. (1965). Growth of tree seedlings as affected trees. Ecology 69: 778–785.
by light intensity. I. White birch, yellow birch, Lotspeich, F.B. (1980). Watersheds as the basic eco-
sugar maple, and silver maple. Dept. For. Canada, system: this conceptual framework provides a basis
Publ. 1121. 16 pp. for a natural classification system. Water Resour.
Logan, K.T. (1966a). Growth of tree seedlings as Bull. 16: 581–586.
affected by light intensity. II. Red pine, white pine, Loucks, O.L. (1962). A Forest Classification for the Mari-
jack pine and eastern larch. Dept. For. Canada, time Provinces. For. Res. Br., Can. Dept. For. 167 pp.
Publ. 1160. 19 pp. Loucks, O.L. (1970). Evolution of diversity, efficiency
Logan, K.T. (1966b). Growth of tree seedlings as and community stability. Am. Zool. 10: 17–25.
affected by light intensity. III. Basswood and white Louma, D.L., Stockdale, C.A., Molina, R., and Eberhart,
elm. Dept. For. Canada, Publ. 1176. 15 pp. J.L. (2006). The spatial influence of Pseudotsuga
Logan, K.T. (1970). Adaptations of the photosynthetic menziesii retention trees on ectomycorrhizal diver-
apparatus of sun-and shade-grown yellow birch (Bet- sity. Can. J. For. Res. 36: 2561–2573.
ula alleghaniensis Britt.). Can. J. Bot. 48: 1681–1688. Lovett, G.M. (1994). Atmospheric deposition of nutri-
Logan, K.T. and Krotkov, G. (1969). Adaptations of the ents and pollutants in North America: an ecological
photosynthetic mechanism of sugar maple (Acer perspective. Ecol. Appl. 4: 629–650.
References 695
Lovett, G.M. and Kinsmann, D. (1990). Atmospheric Lyr, H. and Hoffmann, G. (1967). Growth rates and
pollutant deposition to high-elevation ecosystems. growth periodicity of tree roots. Int. Rev. For. Res. 2:
Atmos. Environ. 24A: 2767–2786. 181–236.
Lovett, G.M., Jones, C.G., Turner, M.G., and Weathers, Lytle, D.A. and Poff, N.L. (2004). Adaptation to natural
K.C. (ed.) (2005). Ecosystem Function in Heteroge- flow regimes. Trends Ecol. Evol. 19: 94–100.
neous Landscapes. New York: Springer 489 pp. MacArthur, R.H. and MacArthur, J. (1961). On bird
Lovett, G.M., Canham, C.D., Arthur, M.A. et al. (2006). species diversity. Ecology 42: 594–598.
Forest ecosystem responses to exotic pests and MacArthur, R.H. and Wilson, E.O. (1967). The Theory
pathogens in eastern North America. Bioscience 56: of Island Biogeography. Princeton, NJ: Princeton
395–405. University Press 224 pp.
Lovett, G.M., Arthur, M.A., Weathers, K.C., and Griffin, MacDonald, G.M. and Cwynar, L.C. (1985). A fossil
J.M. (2010). Long-term changes in forest carbon pollen based reconstruction of the late quaternary
and nitrogen cycling caused by an introduced pest/ history of lodgepole pine (Pinus contorta ssp.
pathogen complex. Ecosystems 13: 1188–1200. latifolia) in the western interior of Canada. Can.
Lowry, W.P. (1966). Apparent meteorological require- J. For. Res. 15: 1039–1044.
ments for abundant cone crop in Douglas-fir. For. Mace, G.M., Norris, K., and Fitter, A.H. (2012). Bio-
Sci. 12: 185–192. diversity and ecosystem services: a multilayered
Luckman, B.H. (1986). Reconstruction of Little ice age relationship. Trends Ecol. Evol. 27: 19–26.
events in the Canadian Rocky Mountains. Géog. Mack, R.N., Simberloff, D., Lonsdale, W.M. et al.
Phys. Quatern. 40: 17–28. (2000). Biotic invasions: causes, epidemiology,
Lüdi, W. (1945). Besiedlung und Vegetationsentwick- global consequences, and control. Ecol. Appl. 10:
lung auf den jungen Seitenmoränen des Gros- 689–710.
sen Aletschgletschers mit einem Vergleich der Mackey, R.L. and Currie, D.J. (2001). The diversity-
Besiedlung im Vorfeld des Rhonegletschers und disturbance relationship: is it generally strong and
des Oberen Grindelwaldgletschers. Bericht über das peaked? Ecology 82: 3479–3492.
Geobotanische Forschungsinstitut Rübel. Zürich, Mackey, H.E. Jr. and Sivec, N. (1973). The present
1944:35–112. composition of a former oak-chestnut forest in the
Lüthi, D., Le Floch, M., Bereiter, B. et al. (2008). High- Allegheny Mountains of western Pennsylvania.
resolution carbon dioxide concentration record Ecology 54: 915–919.
650,000–800,000 years before present. Nature 453: MacKinnon, A., Meidinger, D., and Klinka, K. (1992).
379–382. Use of biogeoclimatic ecosystem classification
Lutz, H.J. (1930). The vegetation of Heart’s Content, a system in British Columbia. For. Chron. 68:
virgin forest in northwestern Pennsylvania. Ecology 100–120.
11: 1–29. Maguire, W.P. (1955). Radiation, surface temperatures,
Lutz, H.J. (1934). Ecological relations in the pitch pine and seedling survival. For. Sci. 1: 277–285.
plains of southern New Jersey. Yale School of For. Magurran, A.E. (1988). Ecological Diversity and Its
Bull. 38. 80 pp. Measurement. Princeton, NJ: Princeton Univ. Press
Lutz, H.J. (1940). Disturbance of forest soil resulting 179 pp.
from the uprooting of trees. Yale Univ. School For. Magurran, A.E. (2013). Measuring Biological Diversity.
Bull. 45. 37 pp. New York: Wiley 264 pp.
Lutz, H.J. (1956). Ecological effects of forest fires in the Magurran, A.E. and McGill, B.J. (ed.) (2011). Biological
interior of Alaska. U.S. Dept. Agr. Tech. Bull. 1133. Diversity: Frontiers in Measurement and Assessment.
121 pp. Oxford, UK: Oxford University Press 368 pp.
Lyford, W.H. (1980). Development of the root system Maherali, H., Walden, A.E., and Husband, B.C.
of northern red oak (Quercus rubra L.). Harvard (2009). Genome duplication and the evolution of
Forest Paper No. 21. 30 pp. physiological responses to water stress. New Phytol.
Lyford, W.H. and MacLean, D.W. (1966). Mound and pit 184: 721–731.
microrelief in relation to soil disturbance and tree Mahrt, L., Lee, X., Black, A. et al. (2000). Nocturnal
distribution in New Brunswick, Canada. Harvard mixing in a forest subcanopy. Agric. For. Meteorol.
Forest Paper No. 15. 18 pp. 101: 67–78.
Lyford, W.R. and Wilson, B.F. (1966). Controlled growth Mailly, D. and Gaudreault, M. (2005). Growth intercept
of forest tree roots: technique and application. Har- models for black spruce, jack pine and balsam fir in
vard Forest Paper No. 16. Harvard Univ. Petersham, Quebec. For. Chron. 81: 104–113.
Mass. 12 pp. Major, J. (1951). A functional, factorial approach to
Lynch, H.J., Renkin, R.A., Crabtree, R.L., and Moorcroft, plant ecology. Ecology 32: 392–412.
P.R. (2006). The influence of previous mountain Major, E.J. (1990). Water stress in Sitka spruce and its
pine beetle (Dendroctonus ponderosae) activity on effect on the green spruce aphid Elatobium. Popul.
the 1988 Yellowstone fires. Ecosystems 9: 1318–1327. Dyn. Insects 1: 85–93.
696 References
Malanson, G.P. (1993). Riparian Landscapes. Cam- Attribute data. Agriculture and Agri-Food Canada,
bridge: Cambridge Univ. Press 296 pp. Research Branch, Centre for Land and Biological
Mallet, J. (2006). What does Drosophila genetics tell us Resources Research and Environment Canada,
about speciation? Trends Ecol. Evol. 21: 386–393. State of the Environment Directorate, Ecozone
Mallet, J., Besansky, N., and Hahn, M.W. (2016). How Analysis Branch. Ottawa/Hull. http://sis.agr.
reticulated are species? Bioessays 38: 140–149. gc.ca/cansis/nsdb/ecostrat/1999report/index.html
Marchand, D.E. (1971). Rates and modes of denudation, (accessed July 13, 2021).
White Mountains, eastern California. Am. J. Sci. Martin, P.H. (1999). Norway maple (Acer platanoides)
270: 109–135. invasion of a natural forest stand: understory
Marcot, B.C. (1997). Biodiversity of old forests of the consequence and regeneration pattern. Biol. Inva-
west: lessons from our elders. In: Creating a For- sions 1: 215–222.
estry for the 21st Century (ed. K.A. Kohm and J.F. Martin, K.L. and Goebel, P.C. (2013). The foundation
Franklin), 87–105. Washington, DC: Island Press. species influence of eastern hemlock (Tsuga
Margalef, R. (1972). Homage to Evelyn Hutchinson, or canadensis) on biodiversity and ecosystem function
why is there an upper limit to diversity. Trans. Con- on the unglaciated Allegheny Plateau. For. Ecol.
necticut Acad. Arts Sci. 44: 211–235. Manag. 289: 143–152.
Marion, G.M. and Black, C.H. (1988). Potentially available Martin, R.E., Cooper, R.W., Crow, A.B. et al. (1977).
nitrogen and phosphorus along a chaparral fire cycle Report of task force on prescribed burning. J. For.
chronosequence. Soil Sci. Soc. Am. J. 52: 1155–1162. 75: 297–301.
Marks, P.L. (1974). The role of pin cherry (Prunus Martin, J.L., Gower, S.T., Plaut, J., and Holmes, B.
pensylvanica L.) in the maintenance of stability in (2005). Carbon pools in a boreal mixedwood
northern hardwood ecosystems. Ecol. Monogr. 44: logging chronosequence. Glob. Chang. Biol. 11:
73–88. 1883–1894.
Marks, P.L. (1975). On the relation between extension Martín, D., Vázquez-Piqué, J., Carevic, F.S. et al. (2015).
growth and successional status of deciduous trees Trade-off between stem growth and acorn produc-
of the northeastern United States. Bull. Torrey Bot. tion in holm oak. Trees 29: 825–834.
Club 102: 172–177. Marx, D.H. and Beattie, D.J. (1977). Mycorrhizae—
Marks, P.L. and Harcombe, P.A. (1981). Community promising aid to timber growers. For. Farmer 36:
diversity of coastal plain forests in southern East 6–9.
Texas. Ecology 56: 1004–1008. Maser, C. and Maser, Z. (1988). Interactions among
Marlon, J.R., Bartlein, P.J., Daniau, A.-L. et al. (2013). squirrels, mycorrhizal fungi, and coniferous forests
Global biomass burning: a synthesis and review in Oregon. Great Basin Nat. 48: 358–369.
of Holocene paleofire records and their controls. Maser, C., Trappe, J.M., and Nussbaum, R.A. (1978).
Quat. Sci. Rev. 65: 5–25. Fungal—small mammal interrelationships with
Marquis, D.A. (1974). The impact of deer browsing on emphasis on Oregon conifer forests. Ecology 59:
Allegheny hardwood regeneration. USDA For. Serv. 799–809.
Res. Paper NE-308. Northeastern For. Exp. Sta., Maser, C.R., Tarrant, F., Trappe, J.M., and Franklin, J.F.
Upper Darby, PA. 8 pp. (eds.). (1988). From the forest to the sea: A story
Marquis, D.A. 1975. The Allegheny hardwood forests of of fallen trees. USDA For. Serv. Gen. Tech. Report
Pennsylvania. USDA For. Serv. Gen. Tech. Report PNW-229. Washington, D.C. 153 pp.
NE-15. Northeastern For. Exp. Sta., Upper Darby, Masterson, J. (1994). Stomatal size in fossil plants: evi-
PA. 32 pp. dence for polyploidy in majority of angiosperms.
Marquis, D.A. (1981). Effect of deer browsing on timber Science 264: 421–424.
in Allegheny hardwood forests of northwestern Matthews, J.D. (1963). Factors affecting the production
Pennsylvania. USDA For. Serv. Res. Paper NE-475. of seed by forest trees. For. Abstr. 24 (1): i–xiii.
Northeastern For. Exp. Sta., Upper Darby, PA. 10 pp. Matthews, J.A. (1992). The Ecology of Recently Deglaciated
Marr, J.W. (1977). The development and movement of Terrain. Cambridge: Cambridge Univ. Press 386 pp.
tree islands near the upper limit of tree growth Mattson, W.J. and Addy, N.D. (1975). Phytophagous
in the southern Rocky Mountains. Ecology 58: insects as regulators of forest primary production.
1159–1164. Science 190: 515–522.
Marshall, P.E. and Kozlowski, T.T. (1976). Importance Mattson, W.J. and Haack, R.A. (1987). The role of
of photosynthetic cotyledons for early growth of drought in outbreaks of plant-eating insects. Biosci-
woody angiosperms. Physiol. Plant. 37: 336–340. ence 37: 110–118.
Marshall, J.D. and Toffel, M.W. (2005). Framing the Mattson, W.J. and Shriner, D.S. (eds.). (2001). Northern
elusive concept of sustainability: a sustainability Minnesota Independence Day Storm: A Research
hierarchy. Environ. Sci. Technol. 39: 673–682. Needs Assessment. USDA For. Serv. Gen. Tech.
Marshall, I.B., Schut, P.H., and Ballard, M. (1999). Report NC-GTR-216. North Central Res. Sta., St.
A national ecological framework for Canada: Paul, MN. 65 pp.
References 697
Mattson, W.J., Lawrence, R.K., Haack, R.A. et al. (1988). ports of entry and border crossings over a 17-year
Defensive strategies of woody plants against differ- period. Biol. Invasions 8: 611–630.
ent insect-feeding guilds in relation to plant eco- McCune, B. and Allen, T.F.H. (1985). Forest dynamics
logical strategies and intimacy of association with in the bitterroot canyons, Montana. Can. J. Bot. 63:
insects. In: Mechanism of Woody Plant Defenses 377–383.
against Insects (ed. W.J. Mattson, J. Levieux and C. McCune, B. and Cottam, G. (1985). The successional
Bernard-Dagan), 3–38. New York: Springer-Verlag. status of a southern Wisconsin oak woods. Ecology
Mattson, W. J., Herms, D. A., Witter, J. A., and Allen, D. C. 66: 1270–1278.
(1991). Woody plant grazing systems: North American McDermott, R. E. (1954). Seedling tolerance as a factor
outbreak folivores and their host plants. In Y. N. in bottomland timber succession. Mo. Agr. Exp. Sta.
Baranchikov, W. J. Mattson, F. P. Hain, and T. L. Payne Res. Bull. 557. 11 pp.
(eds.), Forest Insect Guilds: Patterns of Interaction with McDowell, N.G., Allen, C.D., Anderson-Teixeira, K.
Host Trees. USDA For. Serv. Gen. Tech. Report NE-153. et al. (2020). Pervasive shifts in forest dynamics in a
Northeastern For. Exp. Sta., Radnor, PA. changing world. Science 368: eaaz9463.
Mattson, W., Vanhanen, H., Veteli, T. et al. (2007). Few McElhinny, C., Gibbons, P., and Brack, C. (2006). An
immigrant phytophagous insects on woody plants objective and quantitative methodology for con-
in Europe: legacy of the European crucible? Biol. structing an index of stand structural complexity.
Invasions 9: 957–974. For. Ecol. Manag. 235: 54–71.
Mátyás, C. and Yeatman, C.W. (1992). Effect of McEvoy, P.M., McAdam, J.H., Mosquera-Losada, M.,
geographical transfer on growth and survival of and Rigueiro-Rodriguez, A. (2006). Tree regener-
jack pine (Pinus banksiana Lamb.) populations. ation and sapling damage of pedunculate oak Quer-
Silvae Genet. 41: 370–376. cus robur in a grazed forest in Galicia, NW Spain: A
May, R.M. (1974). Stability and Complexity in Model comparison of continuous and rotational grazing
Ecosystems. Princeton, NJ: Princeton University systems. Agrofor. Syst. 66: 85–92.
Press 304 pp. McFarlane, R.W. (1992). A Stillness in the Pines: The
May, R.M. (2011). Why worry about how many species Ecology of the Red-Cocaded Woodpecker. New York:
and their loss? PLoS Biol. 9: e1001130. W. W. Norton 270 pp.
May, R.M. and McLean, A.R. (2007). Theoretical McIntire, E.J. and Fajardo, A. (2014). Facilitation as a
Ecology: Principles and Applications, 3e. Oxford, ubiquitous driver of biodiversity. New Phytol. 201:
UK: Oxford University Press. 403–416.
Mayer, P.M., Reynolds, S.K. Jr., McCutchen, M.D., and McIntosh, R.P. (1967). The continuum concept of vege-
Canfield, T.J. (2007). Meta-analysis of nitrogen removal tation. Bot. Rev. 33: 130–187.
in riparian buffers. J. Environ. Qual. 36: 1172–1180. McIntosh, R.P. (1968). Reply. In P. Dansereau (ed.), The
Mayerfeld, D., Rickenbach, M., and Rissman, A. continuum concept of vegetation: responses. Bot.
(2016). Overcoming history: attitudes of resource Rev. 34:253–332.
professionals and farmers toward silvopasture in McIntosh, R.P. (1993). The continuum continued:
southwest Wisconsin. Agrofor. Syst. 90: 723–736. John T. Curtis’ influence on ecology. In: Fifty Years
Mayfield, A.E. III, Seybold, S.J., Haag, W.R. et al. (2021). of Wisconsin Plant Ecology (ed. J.S. Fralish, R.P.
Impacts of invasive species in terrestrial and McIntosh and O.L. Loucks), 122–195. Madison:
aquatic systems in the United States. In: Invasive Univ. Wisconsin. Press.
Species in Forests and Rangelands of the United McIntyre, L. (1985). Humboldt’s way. Natl. Geogr. 168
States (ed. T.M. Poland, T. Patel-Weynand, D.M. (3): 318–351.
Finch, et al.), 5–40. New York: Springer.
McKenney, D.W., Pedlar, J.H., Lawrence, K. et al.
McAndrews, J.H. (1966). Postglacial history of prairie, (2007a). Potential impacts of climate change on the
savanna and forest in northwestern Minnesota. distribution of North American trees. Bioscience 57:
Torreya Bot. Club Mem. 22 (2): 72 pp. 939–948.
McBride, J. (1973). Natural replacement of disease- McKenney, D.W., Pedlar, J.H., Lawrence, K. et al.
killed elms. Am. Midi. Nat. 90: 300–306. (2007b). Beyond traditional hardiness zones: using
McCann, K.S. (2000). The diversity-stability debate. climate envelopes to map plant range limits. Biosci-
Nature 405: 228–233. ence 57: 929–937.
McClanahan, T.R. (1986). The effect of a seed source on McKenney, D.W., Pedlar, J.H., Rood, R.B., and Price, D.
primary succession in a forest ecosystem. Vegetatio (2011). Revisiting projected shifts in the climate
65: 175–178. envelopes of North American trees using updated
McCree, K.J. (1972). The action spectrum, absorptance general circulation models. Glob. Chang. Biol. 17:
and quantum yield of photosynthesis in crop 2720–2730.
plants. Agric. Meteorol. 9: 191–216. McKenzie, D., Miller, C., and Falk, D.A. (ed.) (2011).
McCullough, D.G., Work, T.T., Cavey, J.F. et al. (2006). The Landscape Ecology of Fire. New York: Springer
Interceptions of nonindigenous plant pests at US 332 pp.
698 References
McKevlin, M.R., Hook, D.D., and McKee, W.H. Jr. INT-111. Intermountain For. and Rge. Exp. Sta.,
(1995). Growth and nutrient use efficiency of water Ogden, UT. 20 pp.
tupelo seedlings in flooded and well-drained soil. Meffe, G.K. and Carroll, C.R. (1997). Principles of
Tree Physiol. 15: 753–758. Conservation Biology, 2e, 673. Sunderland, MA:
McKnight, J.S., Hook, D.D., Langdon, O.G., and Sinauer Associates.
Johnson, R.L. (1981). Flood tolerance and related Meier, A.R., Saunders, M.R., and Michler, C.H. (2012).
characteristics of trees of the bottomland forests Epicormic buds in trees: a review of bud establish-
of the southern United States. In: Wetlands of ment, development, and dormancy release. Tree
Bottomland Hardwood Forests, vol. 11 (ed. J.R. Physiol. 32: 565–584.
Clark and J. Benforado), 29–69. Amsterdam: Meigs, G.W., Dunn, C.J., Parks, S.A., and Krawchuk,
Elsevier. M.A. (2020). Influence of topography and fuels on
McLachlan, J.S. and Clark, J.S. (2004). Reconstructing fire refugia probability under varying fire weather
historical ranges with fossil data at continental conditions in forests of the Pacific northwest, USA.
scales. For. Ecol. Manag. 197: 139–147. Can. J. For. Res. 50: 636–647.
McLachlan, J.S., Clark, J.S., and Manos, P.S. (2005). Meiners, T.M., Smith, D.W., Sharik, T.L., and Beck,
Molecular indicators of tree migration capacity D.E. (1984). Soil and plant water stress in an
under rapid climate change. Ecology 86: Appalachian oak forest in relation to topography
2088–2098. and stand age. Plant Soil 80: 171–179.
McLane, S.C. and Aitken, S.N. (2012). Whitebark pine Melgoza, G., Nowak, R.S., and Tausch, R.J. (1990). Soil
(Pinus albicaulis) assisted migration potential: test- water exploitation after fire: competition between
ing establishment north of the species range. Ecol. Bromus tectorum (cheatgrass) and two native
Appl. 22: 142–153. species. Oecologia 83: 7–13.
McLaughlin, B.C., Ackerly, D.D., Klos, P.Z. et al. (2017). Melillo, J.M. (1981). Nitrogen cycling in deciduous
Hydrologic refugia, plants, and climate change. forests. In: Terrestrial Nitrogen Cycles. Ecol. Bull.
Glob. Chang. Biol. 23: 2941–2961. 33 (ed. F.E. Clark and T. Rosswell), 427–442. Stock-
McNab, W.H. (1987). Yellow-poplar site quality related holm: Swedish Natural Sci. Res. Council.
to slope type in mountainous terrain. Northern J. Melillo, J.M., Aber, J.D., and Muratore, J.F. (1982).
Appl. For. 4: 189–192. Nitrogen and lignin control of hardwood leaf litter
McNab, W.H. (1989). Terrain shape index: quantifying decomposition dynamics. Ecology 63: 621–626.
effect of minor landforms on tree height. For. Sci. Melillo, J.M., Aber, J.D., Steudler, P.A., and Schimel, J.P.
35: 91–104. (1983). Denitrification potentials in a successional
McNab, W.H. (1991). Land classification in the Blue sequence of northern hardwood forest stands. In:
Ridge Province: state-of-the-science report. In D.L. Environmental Biogeochemistry. Ecol. Bull. 35. (ed.
Mengel and D.T. Tew (eds.), Symp. Proc. Ecological R. Hallberg), 217–228. Stockholm: Swedish Natural
Land Classification: Applications to Identify the Sci. Res. Council.
Productive Potential of Southern Forests. USDA For. Melillo, J.M., McGuire, A.D., Kicklighter, D.W. et al.
Serv. Gen. Tech. Report SE-68. Southeastern For. (1993). Climate change and terrestrial net primary
Exp. Sta., Asheville, NC. productivity. Nature 363: 234–363.
McNab, W.H. (1993). A topographic index to quantify Melles, S.J., Fortin, M.-J., Lindsay, K., and Badzinski,
the effect of mesoscale landform on site produc- D. (2011). Expanding northward: influence of cli-
tivity. Can. J. For. Res. 23: 1100–1107. mate change, forest connectivity, and population
McNab, W.H., Browning, S.A., Simon, S.A., and Fouts, processes on a threatened species’ range shift. Glob.
P.E. (1999). An unconventional approach to eco- Chang. Biol. 17: 17–31.
system unit classification in western North Caro- Mendelsohn, R. and Balick, M.J. (1995). The value of
lina, USA. For. Ecol. Manag. 114: 405–420. undiscovered pharmaceuticals in tropical forests.
Meades, W.J. and Roberts, B.A. (1992). A review of Econ. Bot. 49: 223–228.
forest site classification activities in Newfoundland Méndez-Toribio, M., Meave, J.A., Zermeño-Hernández,
and Labrador. For. Chron. 68: 25–33. I., and Ibarra-Manríquez, G. (2016). Effects of slope
Meehl, G.A., Washington, W.M., Ammann, C.M. et al. aspect and topographic position on environmental
(2004). Combinations of natural and anthropogenic variables, disturbance regime and tree community
forcings in twentieth-century climate. J. Clim. 17: attributes in a seasonal tropical dry forest. J. Veg.
3721–3727. Sci. 27: 1094–1103.
Meeker, D.O. Jr. and Merkel, D.L. (1984). Climax Messier, C. and Bellefleur, P. (1988). Light quantity and
theories and a recommendation for vegetation quality on the forest floor of pioneer and climax
classification—a viewpoint. J. Range Manag. 37: stages in a birch-beech-sugar maple stand. Can. J.
427–430. For. Res. 18: 615–622.
Meeuwig, R. O. (1971). Infiltration and water repel- Metz, L.J., Lotti, T., and Klawitter, R.A. (1961). Some
lency in granitic soils. USDA For. Serv. Res. Paper effects of prescribed burning on coastal plain forest
References 699
soil. USDA For. Serv. Sta. Paper 133. Southeastern States (ed. T.M. Poland, T. Patel-Weynand, D.M.
For. Exp. Sta., Asheville, NC. 10 pp. Finch, et al.), 41–55. New York: Springer.
Meyers, R. K., Zahner, R., and Jones, S. M. (1986). Minteer, B.A. and Collins, J.P. (2010). Move it or lose it?
Forest habitat regions of South Carolina. Clem- The ecological ethics of relocating species under
son Univ., Dept. Forestry Res. Series No. 42. 31 climate change. Ecol. Appl. 20: 1801–1804.
pp. + map. Mirkin, B.M., Naumova, L.G., Martynenko, V.B., and
Michalak, J.L., Lawler, J.J., Roberts, D.R., and Carroll, Shirokikh, P.S. (2015). Contribution of the Braun-
C. (2018). Distribution and protection of climatic Blanquet syntaxonomy to research on successions
refugia in North America. Conserv. Biol. 32: of plant communities. Russ. J. Ecol. 46: 303–308.
1414–1425. Mirov, N.T. (1967). The Genus “Pinus.”. New York:
Miles, J. (1979). Vegetation Dynamics. London: Chap- Ronald Press 602 pp.
man and Hall 80 pp. Mittelbach, G.G., Steiner, C.F., Scheiner, S.M. et al.
Miles, J. (1985). The pedogenic effects of different (2001). What is the observed relationship between
species and vegetation types and the implications species richness and productivity? Ecology 82:
of succession. J. Soil Sci. 36: 571–584. 2381–2396.
Miles, J. (1987). Vegetation succession: past and present Mitton, J.B. and Ferrenberg, S.M. (2012). Mountain pine
perceptions. In: Colonization, Succession and Sta- beetle develops an unprecedented summer genera-
bility (ed. A.J. Gray, M.J. Crawley and P.J. Edwards), tion in response to climate warming. Am. Nat. 179:
1–30. Oxford: Blackwell. E163–E171.
Millar, C.I., Swanston, C.W., and Peterson, D.L. (2014). Mitton, J.B. and Grant, M.C. (1980). Observations on
Adapting to climate change. In: Climate Change the ecology and evolution of quaking aspen, Popu-
and United States Forests (ed. D.L. Peterson, J.M. lus tremuloides, in the Colorado front range. Am. J.
Vose and T. Patel-Weynand), 183–222. New York: Bot. 67: 202–209.
Springer. Miyazaki, Y. (2013). Dynamics of internal carbon
Millennium Ecosystem Assessment (2005). Ecosystems resources during masting behavior in trees. Ecol.
and Human Well Being: Current State and Trends. Res. 28: 143–150.
Washington, DC: Island Press 815 pp. Mladenoff, D.J. and Steams, F. (1993). Eastern hemlock
Miller, J.R. and Bestelmeyer, B.T. (2016). What’s wrong regeneration and deer browsing in the northern
with novel ecosystems, really? Restor. Ecol. 24: Great Lakes region: a re-examination and model
577–582. simulation. Conserv. Biol. 7: 889–900.
Miller, G.T. Jr. (1992). Living in the Environment: An Mock, K.E., Rowe, C.A., Hooten, M.B. et al. (2008).
Introduction to Environmental Science, 7e. Belmont, Clonal dynamics in western north American aspen
CA: Wadsworth 705 pp. (Populus tremuloides). Mol. Ecol. 17: 4827–4844.
Miller, H.A. and Lamb, S.H. (1985). Oaks of North Mock, K.E., Callahan, C.M., Islam-Faridi, M.N. et al.
America. Happy Camp, CA: Naturegraph Publ (2012). Widespread triploidy in western north
327 pp. American aspen (Populus tremuloides). PLoS One
Miller, C. and Urban, D.L. (2000). Connectivity of forest 7: e48406.
fuels and surface fire regimes. Landsc. Ecol. 15: Mody, K., Eichenberger, D., and Dorn, S. (2009). Stress
145–154. magnitude matters: different intensities of pulsed
Miller, B., Reading, R., Trittholt, J.S. et al. (1999). Using water stress produce non-monotonic resistance
focal species in the design of nature reserve net- responses of host plants to insect herbivores. Eco-
works. Wild Earth 8: 81–92. logical Entomology 34 (1): 133–143.
Miller, T.E., Burns, J.H., Munguia, P. et al. (2005). A Moehring, D.H., Grano, C.X., and Bassett, J.R. (1966).
critical review of twenty years’ use of the resource- Properties of forested loess soils after repeated
ratio theory. Am. Nat. 165: 439–448. prescribed burns. USDA For. Serv. Res. Note SO-40.
Miller, T.E., Burns, J.H., Munguia, P. et al. (2007). Eval- Southern For. Exp. Sta., New Orleans, LA. 4 pp.
uating support for the resource-ratio hypothesis: a Mohn, C.A. and Pauley, S. (1969). Early performance of
reply to Wilson et al. Am. Nat. 169: 707–708. cottonwood seed sources in Minnesota. Minn. For.
Miller, A., Reilly, D., Bauman, S., and Shea, K. (2012). Inter- Notes 207: 4 pp.
actions between frequency and size of disturbance Monsen, S.B. and Shaw, N.L. (2000). Big sagebrush
affect competitive outcomes. Ecol. Res. 27: 783–791. (Artemisia tridentata) communities-ecology, impor-
Mills, L.S., Soulé, M.E., and Doak, D.F. (1993). tance and restoration. In L. Wagner, D. Neuman
The keystone-species concept in ecology and (eds.). Striving for Restoration, Fostering Technology
conservation. Bioscience 43: 219–224. and Policy for Reestablishing Ecological Function.
Miniat, C.F., Fraterrigo, J.M., Brantley, S.T. et al. (2021). Proceedings of the Billings Land Reclamation
Impacts of invasive species on forest and grassland Symposium, March 20–24, 2000, Billings, MT.
ecosystem processes in the United States. In: Inva- Reclamation Research Unit Publication No. 00–01.
sive Species in Forests and Rangelands of the United Bozeman, MT: Montana State University.
700 References
Monserud, R.A. (1984). Problems with site index: an Morgan, P. and Bunting, S. C. (1990). Fire effects in
opinionated review. In J.G. Bockheim (ed.), Symp whitebark pine forests. In W.C. Schmidt and K. J.
Proc. Forest Land Classification: Experience, Prob- McDonald (comps.), Symp. Proc. Whitebark Pine
lems, Perspectives, NCR-102 North Central For. Soil Ecosystems: Ecology and Management of a High-
Com., Soc. Am. For., USDA For. Serv., USDA Cons. Mountain Resource. USDA For. Serv. Gen Tech.
Serv., Madison, WI. Rep. INT-270, Intermountain Res. Sta., Odgen, UT.
Monserud, R.A. and Rehfeldt, G.E. (1990). Genetic and Morgenstern, E.K. (1996). Geographic Variation in
environmental components of variation of site Forest Trees. Vancouver, B.C.: Univ. British Colum-
index in inland Douglas-fir. For. Sci. 36: 1–9. bia Press 209 pp.
Montgomery, D.R., Grant, G.E., and Sullivan, K. (1995). Morgenstern, E.K. and Mullin, T.J. (1990). Growth and
Watershed analysis as a framework for implement- survival of black spruce in the range wide prove-
ing ecosystem management. Water Resour. Bull. 31: nance study. Can. J. For. Res. 20: 130–143.
369–386. Mori, A.S., Lertzman, K.P., and Gustafsson, L. (2017).
Montgomery, R.A., Rice, K.E., Stefanski, A. et al. (2020). Biodiversity and ecosystem services in forest
Phenological responses of temperate and boreal ecosystems: a research agenda for applied forest
trees to warming depend on ambient spring tem- ecology. J. Appl. Ecol. 54: 12–27.
peratures, leaf habit, and geographic range. Proc. Morin, R.S. and Liebhold, A.M. (2016). Invasive forest
Natl. Acad. Sci. 117: 10397–10405. defoliator contributes to the impending downward
Moola, F.M. and Vasseur, L. (2008). The importance trend of oak dominance in eastern North America.
of clonal growth to the recovery of Gaultheria Forestry 89: 284–289.
procumbens L. (Ericaceae) after forest disturbance. Morrison, P.H. and Swanson, F.J. (1990). Fire history
Plant Ecol. 201: 319–337. and pattern in a Cascade Range landscape. USDA
Mooney, H.A. (1972). The carbon balance of plants. For. Serv. Gen. Tech. Report PNW-GTR-254, Pacific
Annu. Rev. Ecol. Syst. 3: 315–346. Northwest Res. Sta., Corvallis, OR. 77 pp.
Moore, J.R. and Watt, M.S. (2015). Modelling the Mosley, J.C., Bunting, S.C., and Manoukian, M.E.
influence of predicted future climate change on the (1999). Cheatgrass. In: Biology and Management
risk of wind damage within New Zealand’s planted of Noxious Rangeland Weeds (ed. R.L. Sheley and
forests. Glob. Chang. Biol. 21: 3021–3035. J.K. Petroff), 175–188. Corvallis, OR: Oregon State
Moos, C., Khelidj, N., Guisan, A. et al. (2021). A University Press.
quantitative assessment of rockfall influence on Mosquin, T. (1966). Reproductive specialization as a
forest structure in the Swiss Alps. Eur. J. For. Res. factor in the evolution of the Canadian flora. In:
140: 91–104. The Evolution of Canada’s Flora (ed. R.L. Taylor
Mopper, S. and Whitham, T.G. (1992). The plant stress and R.A. Ludwig), 43–65. Toronto: Univ. Toronto
paradox: effects on pinyon sawfly sex ratios and Press.
fecundity. Ecology 73: 515–525. Moss, C.E. (1913). Vegetation of the Peak District. Cam-
Mopper, S., Maschinski, J., Cobb, N., and Whitham, T.G. bridge: Cambridge Univ. Press 235 pp.
(1991a). A new look at habitat structure: conse- Mott, K.A. and Parkhurst, D.F. (1991). Stomatal
quences of herbivore-modified plant architecture. responses to humidity in air and helox. Plant Cell
In: Habitat Complexity: The Physical Arrangement Environ. 14: 509–515.
of Objects in Space (ed. S. Bell, E. McCoy and H. Mount, A.B. (1964). The interdependence of the euca-
Mushinsky), 260–280. New York: Chapman and lyptus and forest fires in southern Australia. Aust.
Hall. For. 28: 166–172.
Mora, C., Tittensor, D.P., Adl, S. et al. (2011). How many Mueggler, W.F. (1985). Vegetation associations. In N.V.
species are there on earth and in the ocean? PLoS DeByle and R. P. Winokur (eds.), Aspen: Ecology
Biol. 9: e1001127. and Management in the Western United States.
Moran, E., Lauder, J., Musser, C. et al. (2017). The USDA For. Serv. Gen. Tech. Report RM-119. Rocky
genetics of drought tolerance in conifers. New Mountain For. and Rge. Exp. Sta., Ft. Collins, CO.
Phytol. 216: 1034–1048. Mueggler, W.R. and Campbell, R.B. (1986). Aspen
Morecroft, M.D., Taylor, M.E., and Oliver, H.R. (1998). community types of Utah. USDA For. Serv. Res.
Air and soil microclimates of deciduous woodland Paper INT-362. Intermountain Res. Sta., Ogden,
compared to an open site. Agric. For. Meteorol. 90: UT. 69 pp.
141–156. Mueller-Dombois, D. (1987). Natural dieback in forests.
Morelli, T.L., Daly, C., Dobrowski, S.Z. et al. (2016). Bioscience 37: 575–583.
Managing climate change refugia for climate Mueller-Dombois, D. and Ellenberg, H. (1974). Aims
adaptation. PLoS One 11: e0159909. and Methods of Vegetation Ecology. New York:
Morelli, T.L., Barrows, C.W., Ramirez, A.R. et al. (2020). Wiley 547 pp.
Climate-change refugia: biodiversity in the slow Muffler, L., Beierkuhnlein, C., Aas, G. et al. (2016).
lane. Front. Ecol. Environ. 18: 228–234. Distribution ranges and spring phenology explain
References 701
late frost sensitivity in 170 woody plants from the chilling advances the time to budburst in North
northern hemisphere. Glob. Ecol. Biogeogr. 25: American tree species. Tree Physiol. 37: 1727–1738.
1061–1071. Nanson, G.C. and Croke, J.C. (1992). A genetic
Mühlhäusser, G. and Müller, S. (1995). Das südwest- classification of floodplains. Geomorphology 4:
deutsche Verfahren der Forstlichen Standorts- 459–486.
kartierung. In Arbeitsgruppe Biozönosenkunde. National Drought Mitigation Center (2021). Maps and
1995. Ansätze für eine Regionale Biotop und data. https://droughtmonitor.unl.edu. Accessed
Biozönosenkunde von Baden-Württemberg. Mitt. 01-21-2021.
Forst. Versuchs u. Forschungsanst. Baden- National Oceanic and Atmospheric Administration
Württemberg. Freiburg, Germany. 166 pp. (2021a). National Centers for Environmental
Mühlhäusser, G., Hubner, W., and Sturmmer, G. (1983). Information. https://www.ncei.noaa.gov. Accessed
Die Forstliche Standortskarte 1:10,000 nach dem 09-19-2021.
Baden-Württembergischen Verfahren. Mitt. Verein National Oceanic and Atmospheric Administration
forstl. Standortsk. u. Forstpflz. 30:3–13. (2021b). U.S. Climate Extremes Index. http://www.
Muir, P.S. and Lotan, J.E. (1985). Disturbance history ncdc.noaa.gov/extremes/cei. Accessed 09-19-2021.
and serotiny in Pinus contorta in western Montana. Naveh, Z. and Lieberman, A.S. (1994). Landscape
Ecology 66: 1658–1668. Ecology: Theory and Application, 2e. New York:
Munoz, S.E., Gajewski, K., and Peros, M.C. (2010). Springer 360 pp.
Synchronous environmental and cultural change Neary, D.G., Klopatek, C.C., DeBano, L.F., and Ffolliott,
in the prehistory of the northeastern United States. P.F. (1999). Fire effects on belowground sustain-
Proc. Natl. Acad. Sci. 107: 22008–22013. ability: a review and synthesis. For. Ecol. Manag.
Musselman, R.C., Lester, D.T., and Adams, M.S. (1975). 122: 51–71.
Localized ecotypes of Thuja occidentalis L. in Wis- Neary, D.G., Ryan, K.C., and DeBano, L.F. (2005). Wild-
consin. Ecology 56: 647–655. land fire in ecosystems: effects of fire on soils and
Mutch, R.W. (1970). Wildland fires and ecosystems—a water. USDA For. Serv. Gen. Tech. Rep. RMRS-
hypothesis. Ecology 51: 1046–1051. GTR-42-vol. 4. Rocky Mountain Research Station,
Mutch, R.W. (1976). Fire management and land use Ogden, UT. 250 pp.
planning today: tradition and change in the Forest Neilson, R.P. and Wullstein, L.H. (1980). Catkin
Service. Western Wildlands 3: 13–19. freezing and acorn production in gambel oak in
Myers, R.L. (1990). Scrub and high pine. In: Ecosystems Utah, 1978. Am. J. Bot. 67: 426–428.
of Florida (ed. R.L. Myers and J.J. Ewel). Orlando: Neilson, R.P., Pitelka, L.F., Solomon, A.M. et al. (2005).
Univ. Central Florida Press. Forecasting regional to global plant migration in
Myers, R.L. and Ewel, J.J. (ed.) (1990). Ecosystems of response to climate change. Bioscience 55: 749–759.
Florida. Orlando: Univ. Central Florida Press 765 pp. Neiman, K.E. (1988). Soil characteristics as an aid to
Myers, J.A. and Kitajima, K. (2007). Carbohydrate identifying forest habitat types in northern Idaho.
storage enhances seedling shade and stress toler- USDA For. Serv. Res. Pap. INT-390. Intermountain
ance in a neotropical forest. J. Ecol. 95: 383–395. Res. Sta., Ogden, UT. 16 pp.
Myneni, R., Keeling, C., Tucker, C. et al. (1997). Nesbitt, L., Hotte, N., Barron, S. et al. (2017). The social
Increased plant growth in the northern high lati- and economic value of cultural ecosystem services
tudes from 1981 to 1991. Nature 386: 698–702. provided by urban forests in North America: A
Nadelhoffer, K.J., Aber, J.D., and Melillo, J.M. (1985). review and suggestions for future research. Urban
Fine roots, net primary production, and soil For. Urban Green. 25: 103–111.
nitrogen availability: a new hypothesis. Ecology 66: Neumayer, E. (2013). Weak Versus Strong Sustainability:
1377–1390. Exploring the Limits of Two Opposing Paradigms,
Naeem, S. (2009). Biodiversity, ecosystem function, 4e. Cheltenham, UK: Edward Elgar Publishing 296
and ecosystem services. In: The Princeton Guide pp.
to Ecology (ed. S.A. Levin, S.R. Carpenter, H.C.J. Newman, J.A., Anand, M., Henry, H.A.L. et al. (2011).
Godfray, et al.), 584–590. Princeton, NJ: Princeton Climate Change Biology. Boston, MA: CAB
University Press. International 288 pp.
Naeem, S., Duffy, J.E., and Zavaleta, E. (2012). The Nichols, G.E. (1923). A working basis for the ecolog-
functions of biological diversity in an age of ical classification of plant communities. Ecology 4:
extinction. Science 336: 1401–1406. 154–179.
Naess, A. (1986). Intrinsic value: will the defenders of Nicolai, V. (1986). The bark of trees: thermal properties,
nature please rise? In: Conservation Biology: The microclimate and fauna. Oecologia 69: 148–160.
Science of Scarcity and Diversity (ed. M.E. Soulé), Nicolini, E., Chanson, B., and Bonne, F. (2001). Stem
504–515. Sunderland, MA: Sinauer Associates. growth and epicormic branch formation in under-
Nanninga, C., Buyarski, C.R., Pretorius, A.M., and storey beech trees (Fagus sylvatica L.). Ann. Bot. 87:
Montgomery, R.A. (2017). Increased exposure to 737–750.
702 References
Nicoll, B.C. and Ray, D. (1996). Adaptive growth of tree Norby, R.J., Warren, J.M., Iversen, C.M. et al. (2010).
root systems in response to wind action and site CO2 enhancement of forest productivity con-
conditions. Tree Physiol. 16: 891–898. strained by limited nitrogen availability. Proc. Natl.
Nicolson, M. (1990). Henry Allan Gleason and the indi- Acad. Sci. 107: 19368–19373.
vidualistic hypothesis: the structure of a botanist’s Norman, S.A., Schaetzl, R.J., and Small, T.W. (1995).
career. Bot. Rev. 56: 91–161. Effects of slope angle on mass movement by tree
Nielsen, U.N., Wall, D.H., and Six, J. (2015). Soil bio- uprooting. Geomorphology 14: 19–27.
diversity and the environment. Annu. Rev. Env. North, M.P., Stephens, S.L., Collins, B.M. et al. (2015).
Resour. 40: 63–90. Reform forest fire management. Science 349:
Nielsen, U.N., Wall, D.H., and Six, J. (2015). Soil biodi- 1280–1281.
versity and the environment. Annu. Rev. Environ. Norton, B.G. (1982). Environmental ethics and non-
Resour. 40: 63–90. human rights. Environ. Ethics 4: 17–36.
Nienstaedt, H. (1974). Genetic variation in some pheno- Norton, D.A. and Kelly, D. (1988). Mast seeding over
logical characteristics of forest trees. In: Phenology 33 years by Dacrydium cupressinum lamb, (rimu)
and Seasonality Modeling (ed. H. Lieth), 589–400. (Podocarpaceae) in New Zealand: the importance
New York: Springer-Verlag. of economies of scale. Funct. Ecol. 2: 399–408.
Niering, W.A. and Egler, F.E. (1955). A shrub Noss, R. F., LaRoe, E. T., III, and Scott, J. M. (1995).
community of Viburnum lentago, stable for twenty- Endangered ecosystems of the United States: a
five years. Ecology 36: 356–360. preliminary assessment of loss and degradation.
Niering, W.A. and Goodwin, R.H. (1974). Creation USDI, Nat. Biol. Serv., Biol. Report 28. Washington
of relatively stable shrublands with herbicides: D.C. 58 pp.
arresting succession on rights-of-way and pasture- Nowacki, G.J. and Abrams, M.D. (1991). Community
land. Ecology 55: 784–795. and edaphic analysis of mixed oak forests in the
Niinemets, Ü. and Valladares, F. (2006). Tolerance to ridge and valley province of central Pennsylvania.
shade, drought, and waterlogging of temperate In L.H. McCormick and K.W. Gottschalk (eds.),
northern hemisphere trees and shrubs. Ecol. Proc. Eighth Central Hardwood Conference. USDA,
Monogr. 76: 521–547. For. Serv. Gen. Tech. Report NE-148. Northeastern
Nikles, D.G. (1970). Breeding for growth and yield. For. Exp. Sta., Broomall, PA.
Unasylva 24 (2–3): 9–22. Nowacki, G.J. and Abrams, M.D. (2008). The demise of
Nilsen, E.T. (1985). Seasonal and diurnal leaf move- fire and “mesophication” of forests in the eastern
ments of Rhododendron maximum L. in contrasting United States. Bioscience 58: 123–138.
irradiance environments. Oecologia 65: 296–302. Nowacki, G.J., Abrams, M.D., and Lorimer, C.G. (1990).
Nilsen, E.T. (1987). Influence of water relations and Composition, structure, and historical development
temperature on leaf movements of Rhododendron of northern red oak stands along an edaphic
species. Plant Physiology 83: 607–612. gradient in north-Central Wisconsin. For. Sci. 36:
Nilsen, E.T. (1992). Thermonastic leaf movements: a 276–292.
synthesis of research with Rhododendron. Bot. J. Noy-Meir, I. and van der Maarel, E. (1987). Relations
Linn. Soc. 110: 205–233. between community theory and community anal-
Nilsson, S.G. and Wästljung, U. (1987). Seed predation ysis in vegetation science: some historical perspec-
and cross-pollination in mast-seeding beech (Fagus tives. Vegetatio 69: 5–15.
sylvatica) patches. Ecology 68: 260–265. Nybom, H. (2004). Comparison of different nuclear
Nobel, P.S. (2009). Physicochemical and Biophysical DNA markers for estimating intraspecific genetic
Plant Physiology, 4e. New York: Academic Press diversity in plants. Mol. Ecol. 13: 1143–1155.
604 pp. Nye, P.H. (1977). The rate-limiting step in plant nutrient
Noble, I.R. and Slatyer, R.O. (1977). Post-fire succession absorption from soil. Soil Sci. 123: 292–297.
of plants in Mediterranean ecosystems. In H.A. Nye, P.H. and Tinker, P.B. (1977). Solute Movement in
Mooney and C. E. Conrad (eds.), Symp. Proc. the Soil–Root System. Oxford: Blackwell 342 pp.
Environmental Consequences of Fire and Fire Nyland, R.D., Kenefic, L.S., Bohn, K.K., and Stout, S.L.
Management in Mediterranean Ecosystems. USDA (2016). Silviculture: Concepts and Applications, 3e.
For. Serv. Gen. Tech. Report WO-3. Washington, Long Grove, IL: Waveland Press 680 pp.
D.C. Oberdorfer, E. (1990). Pflanzensoziologische Exkursions-
Noble, I.R. and Slatyer, R.O. (1980). The use of vital flora. Stuttgart: Verlag Eugen Ulmer 1050 pp.
attributes to predict successional changes in plant Obeso, J.R. (2002). The costs of reproduction in plants.
communities subject to recurrent disturbances. New Phytol. 155: 321–348.
Vegetatio 43: 5–21. O’Dea, M.E., Zasada, J.C., and Tappeiner, J.C. III (1995).
Norby, R.J. and Zak, D.R. (2011). Ecological lessons Vine maple clone growth and reproduction in
from free-air CO2 enrichment (FACE) experiments. managed and unmanaged coastal Oregon Douglas-
Annu. Rev. Ecol. Evol. Syst. 42: 181–203. fir forests. Ecol. Appl. 5: 63–73.
References 703
Oehri, J., Schmid, B., Schaepman-Strub, G., and O’Neill, R.V. and DeAngelis, D.L. (1981). Comparative
Niklaus, P.A. (2017). Biodiversity promotes primary productivity and biomass relations of forest eco-
productivity and growing season lengthening at systems. In D.E. Reichle (ed.), Dynamic Properties
the landscape scale. Proc. Natl. Acad. Sci. 114: of Forest Ecosystems, Int. Biol. Programme 23.
10160–10165. Cambridge Univ. Press, Cambridge.
Ohmart, R.D. and Anderson, B.W. (1982). North Ontl, T.A., Janowiak, M.K., Swanston, C.W. et al.
American desert riparian ecosystems. In: Reference (2020). Forest management for carbon sequestra-
Handbook on the Deserts of North America (ed. tion and climate adaptation. J. For. 118: 86–101.
G.L. Bender), 433–479. Westport, CT: Greenwood Oosting, H.J. (1956). The Study of Plant Communities,
Press. 2e. San Francisco CA: Freeman 440 pp.
Oke, T.R. (1987). Boundary Layer Climates, 2e. Oxford- Opdam, P. and Wascher, D. (2004). Climate change
shire, England, UK: Routledge Publishing 464 pp. meets habitat fragmentation: linking landscape
Oldeman, R.A.A. (1990). Forests: Elements of Silvology. and biogeographical scale levels in research and
New York: Springer-Verlag 624 pp. conservation. Biol. Conserv. 117: 285–297.
Oldemeyer, J.L., Franzmann, A.W., Brundage, A.L. Ordonez, A. (2014). Functional and phylogenetic simi-
et al. (1977). Browse quality and the Kenai moose larity of alien plants to cooccurring natives. Ecology
population. J. Wildl. Manag. 41: 533–542. 95: 1191–1202.
Oliver, C.D. (1981). Forest development in North Orodho, A.B., Trlica, M.J., and Bonham, C.D. (1990). Long-
America following major disturbances. For. Ecol. term heavy-grazing effects on soil and vegetation in
Manag. 3: 153–168. the four corners region. Southwest. Nat. 35: 9–14.
Oliver, C.D. and Larson, B.C. (1996). Forest Stand Orwig, D.A., Cobb, R.C., D’Amato, A.W. et al. (2008).
Dynamics: Updated Edition. New York: Wiley 519 Multi-year ecosystem response to hemlock woolly
pp. adelgid infestation in southern New England for-
Öllerer, K., Varga, A., Kirby, K. et al. (2019). Beyond the ests. Can. J. For. Res. 38: 834–843.
obvious impact of domestic livestock grazing on Osawa, A. and Allen, R.B. (1993). Allometric theory
temperate forest vegetation–A global review. Biol. explains self-thinning relationships of mountain
Conserv. 237: 209–219. beech and red pine. Ecology 74: 1020–1032.
Ollinger, S.V., Aber, J.D., Lovett, G.M. et al. (1993). A Osterkamp, W.R. and Hupp, E.R. (1984). Geomor-
spatial model of atmospheric deposition for the phic and vegetative characteristics along three
northeastern U.S. Ecol. Appl. 3: 459–472. northern Virginia streams. Bull. Geol. Soc. Am. 95:
Olson, J.S. (1958). Rates of succession and soil changes 1093–1101.
on southern Lake Michigan sand dunes. Bot. Gaz. Ostfeld, R.S. and Keesing, F. (2000). Biodiversity and
119: 125–170. disease risk: the case of Lyme disease. Conserv. Biol.
Olson, J.S. (1963). Energy storage and the balance of 14: 722–728.
producers and decomposers in ecological systems. Oswalt, S., Oswalt, C., Crall, A. et al. (2021). Inventory
Ecology 44: 322–331. and monitoring of invasive species. In: Invasive
Olson, S.R., Reiners, W.A., Cronan, C.S., and Lang, G.E. Species in Forests and Rangelands of the United
(1981). The chemistry and flux of through-fall and States (ed. T.M. Poland, T. Patel-Weynand, D.M.
stemflow in subalpine balsam fir forests. Holarct. Finch, et al.), 231–242. New York: Springer.
Ecol. 4: 291–300. Otto, S.P. and Whitton, J. (2000). Polyploid incidence
Omernik, J.M. (1987). Ecoregions of the conterminous and evolution. Annu. Rev. Genet. 34: 401–437.
United States. Ann. Assoc. Am. Geogr. 77: 118–125. Ouarmim, S., Asselin, H., Hely, C. et al. (2014). Long-
Omernik, J.M. (1995). Ecoregions: A spatial frame- term dynamics of fire refuges in boreal mixedwood
work for environmental management. In: forests. J. Quat. Sci. 29: 123–129.
Biological Assessment and Criteria: Tools for Water Ovington, J.D. (ed.) (1983). Ecosystems of the World,
Resource Planning and Decision Making (ed. W.S. Temperate Broad-Leaved Evergreen Forests, Vol. 10.
Davis and T.P. Simon). Boca Raton, FL: Lewis New York: Elsevier 242 pp.
Publishers 432 pp. Owens, J.N. (1991). Flowering and seed set. In: Phys-
Omernik, J.M. (2004). Perspectives on the nature and iology of Trees (ed. A.S. Raghavendra), 247–271.
definition of ecological regions. Environ. Manag. 34 New York: Wiley.
(Supplement 1): S27–S38. Owens, J.N., Molder, M., and Langer, H. (1977). Bud
Omernik, J.M. and Bailey, R.G. (1997). Distinguishing development in Picea glauca. I. Annual growth
between watersheds and ecoregions. J. Am. Water cycle of vegetative buds and shoot elongation as
Res. Assoc. 33: 935–949. they relate to date and temperature sums. Can. J.
Omernik, J.M. and Griffith, G.E. (2014). Ecoregions Bot. 55: 2728–2745.
of the conterminous United States: evolution of a Pabst, R.J. and Spies, T.A. (2001). Ten years of vegeta-
hierarchical spatial framework. Environ. Manag. tion succession on a debris-flow deposit in Oregon.
54: 1249–1266. J. Am. Water Res. Assoc. 37: 1693–1708.
704 References
Paillet, F.L. (2002). Chestnut: history and ecology of a Parker, I.M., Simberloff, D., Lonsdale, W.M. et al.
transformed species. J. Biogeogr. 29: 1517–1530. (1999). Impact: toward a framework for under-
Paine, R.T. (1969). A note on trophic complexity and standing the ecological effects of invaders. Biol.
community stability. Am. Nat. 103: 91–93. Invasions 1: 3–19.
Paine, R.T., Tegner, M.J., and Johnson, E.A. (1998). Parks, S.A., Miller, C., Nelson, C.R., and Holden, Z.A.
Compounded perturbations yield ecological sur- (2014). Previous fires moderate burn severity of
prises. Ecosystems 1: 535–545. subsequent wildland fires in two large western US
Palik, B.J. and Pregitzer, K.S. (1994). White pine seed- wilderness areas. Ecosystems 17: 29–42.
tree legacies in an aspen landscape: influences on Parsons, W.F., Knight, D.H., and Miller, S.L. (1994).
post-disturbance white pine population structure. Root gap dynamics in lodgepole pine forest:
For. Ecol. Manag. 67: 191–201. nitrogen transformations in gaps of different size.
Palik, B.J., Goebel, P.C., Kirkman, L.K., and West, L. Ecol. Appl. 4: 354–362.
(2000). Using landscape hierarchies to guide res- Pascual, J., Cañal, M.J., Correia, B. et al. (2014). Can
toration of disturbed ecosystems. Ecol. Appl. 10: epigenetics help forest plants to adapt to climate
89–202. change? In: Epigenetics in Plants of Agronomic
Palik, B.J., D’Amato, A.W., Franklin, J.F., and Johnson, Importance: Fundamentals and Applications (ed. R.
K.N. (2021). Ecological Silviculture. Long Grove, IL: Alvarez-Venegas, C. De la Peña and J.A. Casas-
Waveland Press 343 pp. Mollano), 125–146. New York: Springer.
Palik, B.J., D’Amato, A.W., and Slesak, R.A. (2021). Pastor, J. and Post, W.M. (1986). Influence of climate,
Wide-spread vulnerability of black ash (Fraxinus soil moisture, and succession on forest carbon and
nigra marsh.) wetlands in Minnesota USA to loss nitrogen cycles. Biogeochemistry 2: 3–27.
of tree dominance from invasive emerald ash borer. Pastor, J., Aber, J.D., McClaugherty, C.A., and Melillo,
Forestry 94: 455–463. J.M. (1984). Aboveground production and N and
Pallardy, S.G. (2008). Physiology of Woody Plants, 3e. P cycling along a nitrogen mineralization gradient
Burlington, MA: Academic Press 454 pp. on Blackhawk Island, Wisconsin. Ecology 65:
Pallardy, S.G., Nigh, T.A., and Garrett, H.E. (1988). 256–268.
Changes in forest composition in central Missouri: Patric, J. H. and Black, P.E. (1968). Potential evapotrans-
1968–1982. Am. Midi Nat. 120: 380–389. piration and climate in Alaska by Thornthwaite’s
Pan, Y., Birdsey, R.A., Fang, J. et al. (2011). A large classification. USDA For. Serv. Res. Paper PNW-71.
and persistent carbon sink in the world’s forests. Pacific Northwest For. and Rge. Exp. Sta., Portland,
Science 33: 988–993. OR. 28 pp.
Paquette, A. and Messier, C. (2010). The role of plan- Patric, J.H. and Helvey, J.D. (1986). Some effects of graz-
tations in managing the world’s forests in the ing on soil and water in the eastern forest. USDA
Anthropocene. Front. Ecol. Environ. 8: 27–34. For. Serv. Gen. Tech. Report NE-115, Northeastern
For. Exp. Sta., Broomall, PA. 24 pp.
Paquette, A. and Messier, C. (2011). The effect of bio-
diversity on tree productivity: from temperate to Patterson, W.A. III and Backman, A.E. (1988). Fire and
boreal forests. Glob. Ecol. Biogeogr. 20: 170–180. disease history of forests. In: Vegetation History (ed.
B. Huntley and T. Webb III), 603–632. Dordrecht:
Paradis, A., Elkinton, J., Hayhoe, K., and Buonaccorsi,
Kluwer.
J. (2008). Role of winter temperature and climate
change on the survival and future range expansion Paul, E.A. and Clark, F.E. (1996). Soil Microbiology and
of the hemlock woolly adelgid (Adelges tsugae) in Biochemistry, 2e. New York: Academic Press 340
eastern North America. Mitig. Adapt. Strateg. Glob. pp.
Chang. 13: 541–554. Paul, C., Hanley, N., Meyer, S.T. et al. (2020). On the
Park, Y.S. and Fowler, D.P. (1988). Geographic variation functional relationship between biodiversity and
of black spruce tested in the maritimes. Can. J. For. economic value. Sci. Adv. 6: eaax7712.
Res. 18: 106–114. Pauley, S.S. (1958). Photoperiodism in relation to tree
Park, A. and Talbot, C. (2018). Information underload: improvement. In: The Physiology of Forest Trees
ecological complexity, incomplete knowledge, (ed. K.V. Thimann), 557–571. New York: Ronald
and data deficits create challenges for the assisted Press.
migration of forest trees. Bioscience 68: 251–263. Pauley, S.S. and Perry, T.O. (1954). Ecotypic variation
Park, A., Puettmann, K., Wilson, E. et al. (2014). Can of the photoperiodic response in Populus. J. Arnold
boreal and temperate forest management be Arbor. 35: 167–188.
adapted to the uncertain ties of 21st century cli- Pausas, J.G. and Ribeiro, E. (2013). The global fire-
mate change? Crit. Rev. Plant Sci. 33: 251–285. productivity relationship. Glob. Ecol. Biogeogr. 22:
Parker, A.J. and Parker, K.C. (1994). Structural vari- 728–736.
ability of mature lodgepole pine stands on gently Pavlik, B.M., Muick, P.C., Johnson, S.G., and Popper,
sloping terrain in Taylor Park basin, Colorado. Can. M. (1991). Oaks of California. Los Olivos, CA:
J. For. Res. 24: 2020–2029. Cachuna Press 184 pp.
References 705
Pearce, F. (1997). Lightning sparks pollution rethink. along a soil gradient: an isotopic perspective. Eco-
New Scientist 153(2066): 15. systems 24: 1976–1990.
Pearcy, R.W., Chazdon, R.L., Gross, L.J. et al. (1994). Penning de Vries, F.W.T. (1975). The cost of mainte-
Photosynthetic utilization of sunflecks: a tempo- nance processes in plant cells. Ann. Bot. 39: 77–92.
rally patchy resource on a time scale of seconds Peñuelas, J., Prieto, P., Beier, C. et al. (2007). Response of
to minutes. In: Exploitation of Environmental plant species richness and primary productivity in
Heterogeneity by Plants. (ed. M.M.Caldwell and shrublands along a north–south gradient in Europe
R.W. Pearcy), 175–208, San Diego: Academic Press. to seven years of experimental warming and drought:
Pearsall, D.R. (1995). Landscape Ecosystems of the Uni- reductions in primary productivity in the heat and
versity of Michigan Biological Station: Ecosystem drought year of 2003. Glob. Chang. Biol. 13: 2563–2581.
Diversity and Ground-Cover Diversity. Ph.D. Thesis. Perlin, J. (1989). A Forest Journey, the Role of Wood in
University of Michigan, Ann Arbor. 396 pp. the Development of Civilization. Cambridge, MA:
Pearse, I.S., Koenig, W.D., and Kelly, D. (2016). Mech- Harvard Univ. Press 445 pp.
anisms of mast seeding: resources, weather, cues, Perrings, C. (2007). Future challenges. Proc. Natl. Acad.
and selection. New Phytol. 212: 546–562. Sci. 104: 15179–15180.
Pearson, G.A. (1936). Some observations on the Perry, P.O., Sellers, H.E., and Blanchard, C.O. (1969).
reaction of pine seedlings to shade. Ecology 17: Estimation of photosynthetically active radiation
270–276. under a forest canopy with chlorophyll extracts and
Pearson, G.A. (1940). Shade effects in ponderosa pine. J. from basal area measurements. Ecology 50: 39–44.
For. 38: 778–780. Perry, T.O. (1962). Racial variation in the day and
Pec, G.J., Karst, J., Sywenky, A.N. et al. (2015). Rapid night temperature requirements or red maple and
increases in forest understory diversity and produc- loblolly pine. For. Sci. 8: 336–344.
tivity following a mountain pine beetle (Dendrocto- Perry, T.O. (1971). Dormancy of trees in winter. Science
nus ponderosae) outbreak in pine forests. PLoS One 171: 29–36.
10: e0124691. Perry, T.O. and Wang, C.W. (1960). Genetic variation
Pedlar, J.H., McKenney, D.W., Beaulieu, J. et al. (2011). in the winter chilling requirement for date of dor-
The implementation of assisted migration in mancy break for Acer rubrum. Ecology 41: 790–794.
Canadian forests. For. Chron. 87: 766–777. Perry, D.A., Hessburg, P.F., Skinner, C.N. et al. (2011).
Pedlar, J.H., McKenney, D.W., Aubin, I. et al. (2012). The ecology of mixed severity fire regimes in Wash-
Placing forestry in the assisted migration debate. ington, Oregon, and northern California. For. Ecol.
Bioscience 62: 835–842. Manag. 262: 703–717.
Peet, R.K. (1974). The measurement of species diversity. Pesendorfer, M.B., Sillett, T.S., Koenig, W.D., and Mor-
Annu. Rev. Ecol. Syst. 5: 285–307. rison, S.A. (2016). Scatter-hoarding corvids as seed
Peet, R.K. (1978). Forest vegetation of the Colorado dispersers for oaks and pines: a review of a widely
front range: patterns of species diversity. Vegetatio distributed mutualism and its utility to habitat res-
37: 65–78. toration. Condor: Ornithol. Appl. 118: 215–237.
Peet, R.K. (1981). Forest vegetation of the Colorado Peters, M.P., Prasad, A.M., Matthews, S.N., and Iverson, L.R.
front range: composition and dynamics. Vegetatio (2020). Climate change tree atlas, Version 4. U.S.
45: 3–75. Forest Service, Northern Research Station and
Peet, R.K. (1992). Community structure and eco- Northern Institute of Applied Climate Science,
system function. In: Plant Succession: Theory Delaware, OH. https://www.nrs.fs.fed.us/atlas.
and Prediction (ed. D.C. Glenn-Lewin, R.K. Peet Accessed 01-17-2022.
and T.T. Veblen), 103–151. London: Chapman Peterson, C.J. (2000). Catastrophic wind damage to
and Hall. North American forests and the potential impact of
Peet, R.K. and Christensen, N.L. (1980a). Succession: a climate change. Sci. Total Environ. 262: 287–311.
population process. Vegetatio 43: 131–140. Peterson, C.J. and Pickett, S.T.A. (1995). Forest reor-
Peet, R.K., and Christensen, N.L. (1980b). Hardwood ganization: a case study in an old-growth forest
forest vegetation of the North Carolina piedmont. catastrophic blowdown. Ecology 76: 763–774.
Veröff. Geobot. Inst. Eddg. Tech. Hochsch., Stift. Peterson, C.J., Carson, W.P., McCarthy, B.C., and
Rübel, Zürich 69:14–39. Pickett, S.T.A. (1990). Microsite variation and soil
Peet, R.K. and Christensen, N.L. (1987). Competition dynamics within newly created treefall pits and
and tree death. Bioscience 37: 586–595. mounds. Oikos 58: 39–46.
Pellitier, P.T. and Zak, D.R. (2018). Ectomycorrhizal Peterson, B.J., Wollheim, W.M., Mulholland, P.J. et al.
fungi and the liberation of nitrogen from soil (2001). Control of nitrogen export from watersheds
organic matter. New Phytol. 217: 68–73. by headwater streams. Science 292: 86–90.
Pellitier, P.T., Zak, D.R., Argiroff, W.A., and Upchurch, Petit, R. and Hampe, A. (2006). Some evolutionary
R.A. (2021). Coupled shifts in ectomycorrhizal consequences of being a tree. Annu. Rev. Ecol. Evol.
communities and plant uptake of organic nitrogen Syst. 37: 187–214.
706 References
Petit, J.R., Jouzel, J., Raynaud, D. et al. (1999). Climate Piotto, D. (2008). A meta-analysis comparing tree
and atmospheric history of the past 420,000 years growth in monocultures and mixed plantations.
from the Vostok ice core, Antarctica. Nature 399: For. Ecol. Manag. 255: 781–786.
429–436. Piovesan, G. and Adams, J.M. (2001). Masting behav-
Petit, R.J., Bialozyt, R., Garnier-Géré, P., and P, and iour in beech: linking reproduction and climatic
A. Hampe. (2004). Ecology and genetics of tree variation. Can. J. Bot. 79: 1039–1047.
invasions: from recent introductions to quaternary Platnik, N.I. (1991). Patterns of biodiversity: tropical vs.
migrations. For. Ecol. Manag. 197: 117–137. temperate. J. Nat. Hist. 25: 1083–1088.
Petrie, M.D., Bradford, J.B., Hubbard, R.M. et al. (2017). Platnik, N.I. (1992). Patterns of biodiversity. In: Sys-
Climate change may restrict dryland forest regener- tematics, Ecology, and the Biodiversity Crisis (ed. N.
ation in the 21st century. Ecology 98: 1548–1559. Eldregde), 15–24. New York: Columbia University
Petty, R.O., and Jackson, M.T. (1966). Plant commu- Press.
nities. In: Natural Features of Indiana. (ed. Platt, W.J. (1975). The colonization and formation of
A.A. Lindsey), 264–296, Indianapolis: Ind. Acad. Sci. equilibrium plant species associations on badger
Pfister, R.D. and Arno, S.F. (1980). Classifying forest disturbances in a tall-grass prairie. Ecol. Monogr.
habitats based on potential climax vegetation. For. 45: 285–305.
Sci. 26: 52–70. Platt, W.J. and Gottschalk, R.M. (2001). Effects of exotic
Pfister, R., Kovalchik, B. L., Arno, S. F., and Presby, R. C. grasses on potential fine fuel loads in the ground-
(1977). Forest habitat types of Montana. USDA For. cover of south Florida slash pine savannas. Int. J.
Serv. Gen. Tech. Report INT-34. Intermountain For. Wildland Fire 10: 155–159.
and Rge. Exp. Sta., Ogden, UT. 174 pp. Plotkin, A.B., Schoonmaker, P., Leon, B., and Foster, D.
Phalan, B., Balmford, A., Green, R.E., and Schar- (2017). Microtopography and ecology of pit-mound
lemann, J.P.W. (2011). Minimising the harm to structures in second-growth versus old-growth for-
biodiversity of producing more food globally. Food ests. For. Ecol. Manag. 404: 14–23.
Policy 36: S62–S71. Podani, J. (2006). Braun-Blanquet’s legacy and data
Phillips, J. (1931). The biotic community. J. Ecol. 19: analysis in vegetation science. J. Veg. Sci. 17:
1–24. 113–117.
Phillips, J. (1934-35). Succession, development, the Poland, T.M., Patel-Weynand, T., Finch, D.M. et al. (ed.)
climax, and the complex organism; an analysis of (2021). Invasive Species in Forests and Rangelands
concepts. J. Ecol. 22/23: 554–571. /210–246. of the United States: A Comprehensive Science Syn-
Phillips, W.S. (1963). Depth of roots in soil. Ecology 44: thesis for the United States Forest Sector. New York:
424. Springer 500 pp.
Piao, S., Liu, Q., Chen, A. et al. (2019). Plant phenology Poore, M.E.D. (1955). The use of phytosociological
and global climate change: current progresses and methods in ecological investigations. J. Ecol. 43:
challenges. Glob. Chang. Biol. 25: 1922–1940. 226–269.
Pickett, S.T.A. (1988). Space-for-time substitution as Porte, A., Huard, F., and Dreyfus, P. (2004). Microcli-
an alternative to long-term studies. In: Long-Term mate beneath pine plantation, semi-mature pine
Studies in Ecology: Approaches and Alternatives (ed. plantation and mixed broadleaved-pine forest.
G.E. Likens), 110–135. New York: Springer-Verlag. Agric. For. Meteorol. 126: 175–182.
Pickett, S.T.A. and White, P.S. (ed.) (1985). The Ecology Post, E. (2013). Ecology of Climate Change. Princeton,
of Natural Disturbance and Patch Dynamics. NJ: Princeton University Press 408 pp.
New York: Academic Press 472 pp. Pott, R. (2011). Phytosociology: A modern geobotanical
Pickett, S.T.A., Collins, S.L., and Armesto, J.J. (1987). method. Plant Biosyst. 145: 9–18.
Models, mechanisms and pathways of succession. Potts, A.S. and Hunter, M.D. (2021). Unraveling the
Bot. Rev. 53: 335–371. roles of genotype and environment in the expres-
Pigott, C.D. and Huntley, J.P. (1981). Factors controlling sion of plant defense phenotypes. Ecol. Evol. 11:
the distribution of Tilia cordata at the northern 8542–8561.
limits of its geographical range. III. Nature and Poulson, T.L. and Platt, W.J. (1996). Replacement pat-
causes of seed sterility. New Phytol. 87: 817–839. terns of beech and sugar maple in Warren Woods,
Pimentel, D., Zuniga, R., and Morrison, D. (2005). Michigan. Ecology 77: 1234–1253.
Update on the environmental and economic costs Poulson, T.L. and White, W.B. (1969). The cave environ-
associated with alien-invasive species in the United ment. Science 165: 971–981.
States. Ecol. Econ. 52: 273–288. Povak, N.A., Hessburg, P.F., and Salter, R.B. (2018). Evi-
Pimm, S.L., Davis, G.E., Loope, L. et al. (1994). Hurri- dence for scale-dependent topographic controls on
cane Andrew. Bioscience 44: 224–229. wildfire spread. Ecosphere 9: e02443.
Pinter, N., Fiedel, S., and Keeley, J.E. (2011). Fire and Power, M.E., Tilman, D., Estes, J.A. et al. (1996). Chal-
vegetation shifts in the Americas at the vanguard of lenges in the quest for keystones. Bioscience 46:
Paleoindian migration. Quat. Sci. Rev. 30: 269–272. 609–620.
References 707
Prach, K. and Walker, L.R. (2020). Comparative Plant Youngberg and C.B. Davey). Corvallis: Oregon
Succession among Terrestrial Biomes of the World. State Univ. Press.
Cambridge, UK: Cambridge University Press Puettmann, K.J., Coates, K.D., and Messier, C. (2008).
412 pp. A Critique of Silviculture: Managing for Complexity.
Prasad, A., Pedlar, J., Peters, M. et al. (2020). Combining Washington, DC: Island Press 206 pp.
US and Canadian forest inventories to assess hab- Pyne, S.J. (1982). Fire in America. Princeton, NJ: Princ-
itat suitability and migration potential of 25 tree eton Univ. Press 654 pp.
species under climate change. Divers. Distrib. 26: Pyne, S.J. (2019). Fire: A Brief History, 2e. Seattle: Uni-
1142–1159. versity of Washington Press 240 pp.
Pratt, R.B., Jacobsen, A.L., Ewers, F.W., and Davis, S.D. Pyšek, P., Jarošík, V., Hulme, P.E. et al. (2012). A global
(2007). Relationships among xylem transport, bio- assessment of invasive plant impacts on resident
mechanics and storage in stems and roots of nine species, communities and ecosystems: the inter-
Rhamnaceae species of the California chaparral. action of impact measures, invading species’ traits
New Phytol. 174: 787–798. and environment. Glob. Chang. Biol. 18: 1725–1737.
Pregitzer, K.S. and Barnes, B.V. (1984). Classification Quamme, H.A. (1985). Avoidance of freezing injury in
and comparison of the upland hardwood and woody plants by deep supercooling. Acta Hortic.
conifer ecosystems of the Cyrus H. McCormick 168: 11–30.
experimental Forest. Upper peninsula, Michigan. Quick, B.E. (1923). A comparative study of the distribu-
Can. J. For. Res. 14: 362–375. tion of the climax association in southern Michi-
Pregitzer, K.S. and Friend, A.L. (1996). The structure gan. Pap. Mich. Acad. Sci. Arts Lett. 3: 211–244.
and function of Populus root systems. In: Biology Quijas, S., Schmid, B., and Balvanera, P. (2010). Plant
of Populus and its Implications for Management diversity enhances provision of ecosystem services:
and Conservation. Nat. Res (ed. R.F. Stettler, H.D. A new synthesis. Basic Appl. Ecol. 11: 582–593.
Bradshaw Jr., P.E. Heilman and T.M. Hinckley), Quine, C., Coutts, M., Gardiner, B., and Pyatt, G. (1995).
331–354. Ottawa: Council Canada Press. Forests and wind: management to minimize damage.
Pregitzer, K.S., Barnes, B.V., and Lemme, G.E. (1983). For. Comm. Bull. 114. HMSO, London. 27 pp.
Relationship of topography to soils and vegetation Radeloff, V.C., Mladenoff, D.J., Guries, R.P., and Boyce,
in an upper Michigan ecosystem. Soil Sci. Soc. Am. M.S. (2004). Spatial patterns of cone serotiny in
J. 47: 117–123. Pinus banksiana in relation to fire disturbance. For.
Pregitzer, K.S., Hendrick, R.L., and Fogel, R. (1992). The Ecol. Manag. 189: 133–141.
demography of fine roots in responses to patches of Radeloff, V.C., Hammer, R.B., and Stewart, S.I. (2005).
water and nitrogen. New Phytol. 125: 575–580. Rural and suburban sprawl in the US Midwest
Pregitzer, K.S., Zak, D.R., Curtis, P.S. et al. (1995). from 1940 to 2000 and its relation to forest
Atmospheric CO2, soil nitrogen, and turnover of fragmentation. Conserv. Biol. 19: 793–805.
fine roots. New Phytol. 129: 579–585. Raffa, K.F., Aukema, B.H., Bentz, B.J. et al. (2008).
Pregitzer, K.S., DeForest, J.L., Burton, A.J. et al. (2002). Cross-scale drivers of natural disturbances prone to
Fine root architecture of nine North American anthropogenic amplification: the dynamics of bark
trees. Ecol. Monogr. 72: 293–309. beetle eruptions. Bioscience 58: 501–517.
Pregitzer, C.C., Bailey, J.K., Schweitzer, J.A., and J. A. Raghavendra, A.S. (ed.) (1991). Physiology of Trees.
(2013). Genetic by environment interactions affect New York: Wiley 509 pp.
plant-soil linkages. Ecol. Evol. 3: 2322–2333. Raich, J.W. and Nadelhoffer, K.J. (1989). Belowground
Pretzsch, H. (2014). Canopy space filling and tree carbon allocation in forest ecosystems: global
crown morphology in mixed-species stands trends. Ecology 70: 1346–1354.
compared with monocultures. For. Ecol. Manag. Ralph, C.J. (1985). Habitat association patterns of forest
327: 251–264. and steppe birds of northern Patagonia, Argentina.
Prevéy, J.S., Harrington, C.A., and Clair, J.B.S. (2018). Condor 87: 471–483.
The timing of flowering in Douglas-fir is deter- Ralston, C.W. (1964). Evaluation of forest site produc-
mined by cool-season temperatures and genetic tivity. Int. Rev. For. Res. 1: 171–201.
variation. For. Ecol. Manag. 409: 729–739.
Rammer, W., Braziunas, K.H., Hansen, W.D. et al.
Prieto, I., Almagro, M., Bastida, F., and Querejeta, J.I. (2021). Widespread regeneration failure in forests
(2019). Altered leaf litter quality exacerbates the of greater Yellowstone under scenarios of future
negative impact of climate change on decomposi- climate and fire. Glob. Chang. Biol. 27: 4339–4351.
tion. J. Ecol. 107: 2364–2382.
Ratnam, W., Rajora, O.P., Finkeldey, R. et al. (2014).
Primack, R.B. (2010). Essentials of Conservation Biology, Genetic effects of forest management practices:
5e. Sunderland, MA: Sinauer Associates 601 pp. global synthesis and perspectives. For. Ecol. Manag.
Pritchett, W.L. and Smith, W.H. (1970). Fertilizing 333: 52–65.
slash pine on sandy soils of the lower coastal Raup, D.M. (1986). Biological extinction in earth his-
plain. In: Tree Growth and Forest Soils (ed. C.T. tory. Science 231: 1528–1533.
708 References
Read, R.A. (1952). Tree species occurrences as influ- Reinartz, J.A. and Popp, J.W. (1987). Structure of clones
enced by geology and soil on an Ozark north slope. of northern prickly-ash (Xanthoxylum america-
Ecology 33: 239–246. num). Am. J. Bot. 74: 415–428.
Read, R.A. (1980). Genetic variation in seedling progeny Reineke, L.H. (1933). Perfecting a stand-density index
of ponderosa pine provenances. For. Sci. Monogr. for even-aged forests. J. Agric. Res. 46: 627–638.
23: 59 pp. Reiners, W.A. and Lang, G.E. (1979). Vegetational
Regan, T. (1981). The nature and possibility of an envi- patterns and processes in the balsam fir zone,
ronmental ethic. Environ. Ethics 3: 19–34. White Mountains, New Hampshire. Ecology 60:
Rehfeldt, G.E. (1986a). Adaptive variation in Pinus pon- 403–417.
derosa from intermountain regions, I. Snake and Rejmánek, M. and Richardson, D.M. (1996). What
Salmon River basins. For. Sci. 32: 79–92. attributes make some plant species more invasive?
Rehfeldt, G.E. (1986b). Adaptive variation in Pinus Ecology 77: 1655–1661.
ponderosa from intermountain regions. II. Middle Remington, C.L. (1968). Suture-zones of hybrid inter-
Columbia River system. USDA For. Serv. Res. Paper action between recently joined biotas. Evol. Biol. 2:
INT-373. Intermountain For. Res., Sta., Odgen, UT. 321–428.
9 pp. Renkin, R., Despain, D., and Clark, D. (1994). Aspen
Rehfeldt, G.E. (1988). Ecological genetics of Pinus con- seedlings following the 1988 Yellowstone fires. In
torta from the Rocky Mountains (USA): a synthesis. E. G. Despain (ed.), Conf. Proc., Plants and Their
Silvae Genet. 37: 131–135. Environments. Tech. Report NPS/NRYELL/NRTR-
Rehfeldt, G.E. (1989). Ecological adaptation in Douglas- 93/XX. USDI, Nat. Park Service.
fir (Pseudotsuga menziesii var. glauca): a synthesis. Rennie, P.J. (1962). Methods of assessing forest site
For. Ecol. Manag. 28: 203–215. capacity. Trans. 7th Inter. Soc. Soil Sci., Comm. IV
Rehfeldt, G.E. (1990). Genetic differentiation among and V, pp. 3–18.
populations of Pinus ponderosa from the upper Rey, O., Eizaguirre, C., Angers, B. et al. (2020). Linking
Colorado River basin. Bot. Gaz. 151: 125–137. epigenetics and biological conservation: towards a
Rehfeldt, G.E., Stage, A.R., and Bingham, R.T. (1971). conservation epigenetics perspective. Funct. Ecol.
Strobili development in western white pine: period- 34: 414–427.
icity, prediction, and association with weather. For. Reyer, C.P., Brouwers, N., Rammig, A. et al. (2015).
Sci. 17: 454–461. Forest resilience and tipping points at different
Reich, P.B., Uhl, C., Walters, M.B., and Ellsworth, D.S. spatio-temporal scales: approaches and challenges.
(1991). Leaf life-span as a determinant of leaf J. Ecol. 103: 5–15.
structure and function among 23 species in Amazo- Reyers, B., Polasky, S., Tallis, H. et al. (2012). Finding
nian forest communities. Oecologia 86: 16–24. common ground for biodiversity and ecosystem
Reich, P.B., Walters, M.B., and Ellsworth, D.S. (1992). services. Bioscience 62: 503–507.
Leaf life-span in relation to leaf, plant and stand Rhoades, D.F. (1976). The anti-herbivore defenses
characteristics among diverse ecosystems. Ecol. of Larrea. In: The Biology and Chemistry of the
Monogr. 62: 365–392. Creosote Bush, A Desert Shrub (ed. T.J. Mabry, J.H.
Reich, P.B., Walters, M.B., Kloeppel, B.D., and Hunziker and D.R. DiFeo), 3–54. Stroudsburg, PA:
Ellsworth, D.S. (1995). Different photosynthesis- Dowden, Hutchinson, and Ross.
nitrogen relations in deciduous hardwood and Rhoades, D.F. and Cates, R.G. (1976). Toward a gen-
evergreen coniferous tree species. Oecologia 104: eral theory of plant antiherbivore chemistry. In:
24–30. Biochemical Interaction between Plants and Insects
Reich, P.B., Grigal, D.F., Aber, J.D., and Gower, S.T. (ed. J.W. Wallace and R.L. Mansell). Boston, MA:
(1997). Nitrogen mineralization and productivity Springer.
in 50 hardwood and conifer stands on diverse soils. Ricart, R.D., Pearsall, D.R., and Curtis, P.S. (2020).
Ecology 78: 335–347. Multidecadal shifts in forest plant diversity and
Reich, P.B., Ellsworth, D.S., Walters, M.B. et al. (1999). community composition across glacial landforms
Generality of leaf trait relationships: A test across in northern lower Michigan, USA. Can. J. For. Res.
six biomes. Ecology 80: 1955–1969. 50: 126–135.
Reich, P.B., Wright, I.J., Cavender-Bares, J. et al. (2003). Ricciardi, A. and Simberloff, D. (2009). Assisted coloni-
The evolution of plant functional variation: zation is not a viable conservation strategy. Trends
traits, spectra, and strategies. Int. J. Plant Sci. 164: Ecol. Evol. 24: 248–253.
S143–S164. Rice, S.K., Westerman, B., and Federici, R. (2004).
Reich, P.B., Hobbie, S.E., Lee, T. et al. (2006). Nitrogen Impacts of the exotic, nitrogen-fixing black locust
limitation constrains sustainability of ecosystem (Robinia pseudoacacia) on nitrogen-cycling in a
response to CO2. Nature 440: 922–925. pine–oak ecosystem. Plant Ecol. 174: 97–107.
Reichle, D.E. (ed.) (1981). Dynamic Properties of Forest Rice, A., Šmarda, P., Novosolov, M. et al. (2019). The
Ecosystems. International Biosphere Programme global biogeography of polyploid plants. Nat. Ecol.
23. Cambridge: Cambridge Univ. Press 683 pp. Evol. 3: 265–273.
References 709
Richards, E.J. (2006). Inherited epigenetic variation – Robledo-Arnuncio, J.J. and Gil, L. (2005). Patterns of
revisiting soft inheritance. Nat. Rev. Genet. 7: pollen dispersal in a small population of Pinus
395–401. sylvestris L. revealed by total-exclusion paternity
Richards, N.A. and Stone, E.L. (1964). The application analysis. Heredity 94: 13–22.
of soil survey to planting site selection: an example Rodman, K.C., Veblen, T.T., Battaglia, M.A. et al.
from the Allegheny uplands of New York. J. For. 62: (2020). A changing climate is snuffing out post-fire
475–480. recovery in montane forests. Glob. Ecol. Biogeogr.
Richards, C.L., Bossdorf, O., and Pigliucci, M. (2010). 29: 2039–2051.
What role does heritable epigenetic variation play Roe, A.L. (1967). Seed dispersal in a bumper spruce
in phenotypic evolution? Bioscience 60: 232–237. seed year. USDA For. Serv. Res. Paper INT-39.
Richardson, S.D. (1956). Studies of root growth of Acer Intermountain For. and Rge. Exp. Sta., Ogden,
saccharinum L. IV: the effect of differential shoot UT. 10 pp.
and root temperature on root growth. Proc. K. Ned. Roesch, L.F., Fulthorpe, R.R., Riva, A. et al. (2007).
Akad. Wet. 59: 428–438. Pyrosequencing enumerates and contrasts soil
Richardson, A.D., Hufkens, K., Milliman, T. et al. microbial diversity. ISME J. 1: 283–290.
(2018). Ecosystem warming extends vegetation Rogers, W.S. and Booth, G.A. (1960). The roots of fruit
activity but heightens vulnerability to cold temper- trees. Sci. Hortic. 14: 27–34.
atures. Nature 560: 368–371. Rohde, K. (1992). Latitudinal gradients in species diver-
Ricketts, T.H. (2001). The matrix matters: effective isola- sity: the search for the primary cause. Oikos 65:
tion in fragmented landscapes. Am. Nat. 158: 87–99. 514–527.
Ricklefs, R. and Schluter, D. (1993). Species Diver- Rohde, A. and Bhalerao, R.P. (2007). Plant dormancy in
sity in Ecological Communities: Historical and the perennial context. Trends Plant Sci. 12: 217–223.
Geographical Perspectives. Chicago: Univ. Chicago Röhrig, E. and Ulrich, B. (1991). Ecosystems of the
Press 414 pp. World, Temperate Deciduous Forests. Vol. 7.
Ripple, W.J. and Beschta, R.L. (2003). Wolf reintroduc- New York: Elsevier 635 pp.
tion, predation risk, and cottonwood recovery in Roland, J. (1993). Large-scale forest fragmentation
Yellowstone National Park. For. Ecol. Manag. 184: increases the duration of tent caterpillar outbreak.
299–313. Oecologia 93: 25–30.
Ripple, W.J. and Beschta, R.L. (2012). Trophic cascades Román-Palacios, C. and Wiens, J.J. (2020). Recent
in Yellowstone: the first 15 years after wolf reintro- responses to climate change reveal the drivers of
duction. Biol. Conserv. 145: 205–213. species extinction and survival. Proc. Natl. Acad.
Ripple, W.J., Beschta, R.L., and Painter, L.E. (2015). Sci. 117: 4211–4217.
Trophic cascades from wolves to alders in Yellow- Romme, W.H. (1982). Fire and landscape diversity in
stone. For. Ecol. Manag. 354: 254–260. subalpine forests of Yellowstone National Park.
Risser, P.G., Karr, J.R., and Forman, R.T.T. (1984). Ecol. Monogr. 52: 199–221.
Landscape ecology: directions and approaches. Romme, W.H. and Despain, D.G. (1989). Historical per-
Illinois Natural History Survey, Number 2. Cham- spective on the Yellowstone fires of 1988. Bioscience
paign, IL. 39: 695–699.
Ritchie, M.E., Tilman, D., and Knops, J.M.H. (1998). Romme, W.H. and Knight, D.H. (1982). Landscape
Herbivore effects and plant and nitrogen dynamics diversity, the concept applied to Yellowstone Park.
in oak savanna. Ecology 79: 165–177. Bioscience 32: 664–670.
Robbins, C.S., Dawson, D.K., and Dowell, B.A. (1989). Romme, W.H., Knight, D.H., and Yavitt, J.B. (1986).
Habitat area requirements of breeding forest birds in Mountain pine beetle outbreaks in the Rocky
the middle Atlantic states. Wildl. Monogr. 103: 1–34. Mountains: regulators of primary productivity?
Robertson, G.P. (1982). Factors regulating nitrification Am. Nat. 127: 484–494.
in primary and secondary succession. Ecology 63: Romme, W.H., Turner, M.G., Tuskan, G.A., and Reed,
1561–1573. R.A. (2005). Establishment, persistence, and
Robertson, G.P. and Tiedje, J.M. (1984). Denitrification growth of aspen (Populus tremuloides) seed-
and nitrous oxide production in successional and lings in Yellowstone National Park. Ecology 86:
old-growth Michigan forests. Soil Sci. Soc. Am. J. 404–418.
48: 383–389. Ronov, A.B. and Yaroshevsky, A.A. (1972). Earth’s
Robertson, G.P. and Tiedje, J.M. (1988). Denitrification crust geochemistry. In: The Encyclopedia of Geo-
in a humid tropical rainforest. Nature 336: 756–759. chemistry and Environmental Sciences (ed. R.W.
Robertson, G.P. and Vitousek, P.M. (1981). Nitrification Fairbridge), 243–254. New York: Van Nostrand
potentials in primary and secondary succession. Reinhold.
Ecology 62: 376–386. Roschanski, A.M., Csilléry, K., Liepelt, S. et al. (2016).
Robinson, S.K., Thompson, R.R. III, Donovan, T.M. et al. Evidence of divergent selection at landscape
(1995). Regional forest fragmentation and the nesting and local scales in Abies alba mill. in the French
success of migratory birds. Science 267: 1987–1990. Mediterranean Alps. Mol. Ecol. 25: 776–794.
710 References
Rosenvald, R. and Lõhmus, A. (2008). For what, when, Rowe, J.S. (1956). Uses of undergrowth species in for-
and where is green-tree retention better than estry. Ecology 37: 461–473.
clear-cutting? A review of the biodiversity aspects. Rowe, J.S. (1961a). The level-of-integration concept and
For. Ecol. Manag. 255: 1–15. ecology. Ecology 42: 420–427.
Rosenzweig, M.L. (1992). Species diversity gradients: Rowe, J.S. (1961b). Critique of some vegetational con-
we know more and less than we thought. J. Mam- cepts as applied to forests of northwestern Alberta.
mal. 73: 715–730. Can. J. Bot. 39: 1007–1017.
Rosenzweig, M.L. (1995). Species Diversity in Space and Rowe, J.S. (1962). Soil, site and land classification. For.
Time. Cambridge, UK: Cambridge University Press Chron. 38: 420–432.
460 pp. Rowe, J.S. (1964). Environmental preconditioning, with
Rosenzwieg, M. (1968). Net primary productivity of ter- special reference to forestry. Ecology 45: 399–403.
restrial communities: prediction from climatolog- Rowe, J.S. (1966). Phytogeographic zonation: an eco-
ical data. Am. Nat. 102: 67–74. logical appreciation. In: The Evolution of Canada’s
Roskoski, J.P. (1980). Nitrogen fixation hardwood for- Flora (ed. R.L. Taylor and R.A. Ludwig), 12–27.
ests of the northeastern United States. Plant Soil Univ. Toronto Press.
54: 33–44. Rowe, J.S. (1969). Plant community as a landscape
Ross, M.S., Sharik, T.L., and Smith, D.Wm. (1982). Age feature. In K.N.H. Greenidge (ed.), Symp. Proc.
structural relationships of tree populations in an Terrestrial Plant Ecology, Nova Scotia Museum,
Appalachian oak forest. Bull. Torrey Bot. Club 109: Halifax.
287–298. Rowe, J.S. (1972). Forest Regions of Canada. Can. For.
Ross, S.D., Pharis, R.P., and Binder, W.D. (1983). Growth Serv., Dept. Env. Publ. No. 1300, Ottawa. 172 pp +
regulators and conifers: their physiology and poten- map.
tial uses in forestry. In: Plant Growth Regulating Rowe, J.S. (1983). Concepts of fire effects on plant indi-
Chemicals, Vol. 2 (ed. L.G. Nickell). Boca Raton, FL: viduals and species. In: The Role of Fire in Northern
CRC Press. Circumpolar Ecosystems (ed. R.W. Wein and D.A.
Ross, M.S., O’Brien, J.J., and Flynn, L.J. (1992). Eco- MacLean), 135–154. New York: Wiley.
logical site classification of Florida keys terrestrial Rowe, J.S. (1984a). Understanding forest landscapes:
habitats. Biotropica 24: 486–502. what you conceive is what you get. The Leslie L.
Roth, I. (1990). Leaf structure of a Venezuelan cloud Schaffer Lectureship in Forest Science. Vancouver,
forest in relation to microclimate. Encyclopedia of B. C., Canada.
Plant Anatomy, Vol. 14, Part 1. Gebrüder Borntrae- Rowe, J.S. (1984b). Forestland classification: limitations
ger, Berlin. 244 pp. of the use of vegetation. In J.G. Bockheim (ed.),
Rother, M.T., Veblen, T.T., and Furman, L.G. (2015). Symp. Proc. Forest Land Classification: Experience,
A field experiment informs expected patterns Problems, Perspectives. NCR-102 North Central For.
of conifer regeneration after disturbance under Soils Comm., Soc. Am. For., USDA For. Serv., and
changing climate conditions. Can. J. For. Res. 45: USDA Soil Cons. Serv., Madison, WI.
1607–1616. Rowe, J.S. (1989). The importance of conserving
Rothstein, D.E., Zak, D.R., and Pregitzer, K.S. (1996). systems. In: Endangered Spaces: The Future for
Nitrate deposition in northern hardwood forests Canada’s Wilderness (ed. M. Hummel), 228–235.
and the nitrogen metabolism of Acer saccharum Toronto: Key Porter Books.
marsh. Oecologia 108: 338–344. Rowe, J. S. (1990). Home Place. NeWest Publishers,
Rothstein, D.E., Yermakov, Z., and Buell, A.L. (2004). Edmonton, Alberta, Canada. Canadian Parks and
Loss and recovery of ecosystem carbon pools fol- Wilderness Society, Henderson Book Series No. 12.
lowing stand-replacing wildfire in Michigan jack 253 pp.
pine forests. Can. J. For. Res. 34: 1908–1918. Rowe, J.S. (1992a). Prologue. For. Chron. 68: 22–24.
Rousi, M., Tahvanainen, J., and Uotila, I. (1989). Rowe, J.S. (1992b). The ecosystem approach to forest-
Inter-and intraspecific variation in the resistance land management. For. Chron. 68: 222–224.
of winter-dormant birch (Betula spp.) against Rowe, J.S. (1992c). The integration of ecological studies.
browsing by the mountain hare. Holarct. Ecol. 12: Funct. Ecol. 6: 115–119.
187–192. Rowe, J.S. (1994). A new paradigm for forestry. For.
Rousi, M., Tahvanainen, J., and Uotila, I. (1991). A Chron. 70: 565–568.
mechanism of resistance to hare browsing in Rowe, J.S. (1997). The necessity of protecting ecoscapes.
winter-dormant European white birch (Betula Global Biodivers. 7: 9–12.
pendula). Am. Nat. 137: 64–82.
Rowe, J.S. (1998). “Earth” as the metaphor for “life”.
Rout, T.M., Moore, J.L., and McCarthy, M.A. (2014). Bioscience 48: 428–429.
Prevent, search or destroy? A partially observable
Rowe, J.S. (2000). An earth-based ethic for humanity.
model for invasive species management. J. Appl.
Natur und Kultur: Transdisciplinare Zeitschrift fur
Ecol. 51: 804–813.
okologische. Nachhaltigkeit 1: 106–120.
References 711
Rowe, J.S. and Barnes, B.V. (1994). Geo-ecosystems and Range basin: synoptic to local-scale controls. J. Geo-
bio-ecosystems. Ecol. Soc. Am. Bull. 75: 40–41. phys. Res. Atmos. 125: e2020JD032686.
Rowe, J.S. and Scotter, G.W. (1973). Fire in the boreal Rushton, B.S. (1993). Natural hybridization within the
forest. Quat. Res. 3: 444–464. genus Quercus L. Ann. Sci. For. 50: 73s–90s.
Rowe, J.S. and Sheard, J.W. (1981). Ecological land Rustad, L.E.J.L., Campbell, J., Marion, G. et al. (2001).
classification: a survey approach. Environ. Manag. A meta-analysis of the response of soil respiration,
5: 451–464. net nitrogen mineralization, and aboveground
Royall, P.D., Delcourt, P.A., and Delcourt, H.R. (1991). plant growth to experimental ecosystem warming.
Late quaternary paleoecology and paleoenviron- Oecologia 126: 543–562.
ments of the Central Mississippi alluvial valley. Ryan, M.G. (1988). The importance of maintenance res-
Bull. Geol. Soc. Am. 103: 157–170. piration by the living cells in sapwood of subalpine
Rubec, C.D.A. (1992). Thirty Years of Ecological Land conifers. Ph.D. Dissertation. Oregon State Univ.,
Surveys in Canada, from 1960 to 1990 Symp. Proc. Corvallis. 104 pp.
Landscape Approaches to Wildlife and Ecosystem Ryan, M.G. (1991). Effects of climate change on plant
Management (ed. G.B. Ingram and M.R. Moss). respiration. Ecol. Appl. 1: 157–167.
Morin Heights, Canada: Polysci. Publ. Inc. Ryan, M.G., Linder, S., Vose, J.M., and Hubbard, R.M.
Rudis, V.A. (1995). Regional forest fragmentation effects (1994). Dark respiration in pines. Ecol. Bull. 43: 50–63.
on bottomland hardwood community types and Ryan, M.G., Harmon, M.E., Birdsey, R.A. et al. (2010).
resource values. Landsc. Ecol. 10: 291–307. A synthesis of the science on forests and carbon for
Rudolph, T.D. (1964). Lammas growth and prolepsis US forests. Issues Ecol. 13: 1–16.
in jack pine in the Lake States. For. Sci. Monogr. 6: Ryan, K.C., Knapp, E.E., and Varner, J.M. (2013). Pre-
70 pp. scribed fire in North American forests and wood-
Rundel, P.W. (1972). Habitat restriction in giant lands: history, current practice, and challenges.
sequoia: the environmental control of grove bound- Front. Ecol. Environ. 11: 15–24.
aries. Am. Midi. Nat. 87: 81–99. Rychert, R., Skujins, J., Sorensen, D., and Prochella, D.
Rundel, P.W. (1981). Fire as an ecological factor. In: (1978). Nitrogen fixation by lichens and free-living
Physiological Plant Ecology I. Responses to the microorganisms in deserts. In: Nitrogen in Desert
Physical Environment, New Series Vol. 12A (ed. Ecosystems (ed. N.E. West and J. Skujins), 20–30.
O.L. Lange, P.S. Nobel, C.B. Osmond and H. Stroudsburg, PA: Dowden, Hutchinson, and Ross.
Ziegler). New York: Springer-Verlag. Sackett, S.S. and Haase, S.M. (1992). Measuring soil
Rundel, P.W. (1983). Impact of fire on nutrient cycles in and tree temperatures during prescribed fires with
Mediterranean-type ecosystems with reference to thermocouple probes. USDA For. Serv. Gen. Tech.
chaparral. In: Mediterranean-Type Ecosystems: The Rep. PSW-131, Pacific Southwest Res. Sta., Albany,
Role of Nutrients (ed. F.J. Kruger, D.T. Mitchell and CA. 15 pp.
J.U.M. Jarvis). New York: Springer-Verlag 570 pp. Saeki, I., Dick, C.W., Barnes, B.V., and Murakami, N.
Runge, M. and Rode, M.W. (1991). Effects of soil (2011). Comparative phylogeography of red maple
acidity on plant associations. In: Soil Acidity (ed. (Acer rubrum L.) and silver maple (Acer sacchari-
B. Ulrich and M.E. Sumner), 183–202. Berlin: num L.): impacts of habitat specialization, hybrid-
Springer-Verlag. ization and glacial history. J. Biogeogr. 38: 992–1005.
Runkle, J.R. (1981). Gap regeneration in some old- Sáenz-Romero, C., O’Neill, G., Aitken, S.N., and Lindig-
growth forests of the eastern United States. Ecology Cisneros, R. (2021). Assisted migration field tests
62: 1041–1051. in Canada and Mexico: lessons, limitations, and
Runkle, J.R. (1982). Patterns of disturbance in some challenges. Forests 12: 9.
old-growth mesic forests of eastern North America. Sagnard, F., Oddou-Muratorio, S., Pichot, C. et al.
Ecology 63: 1533–1546. (2011). Effects of seed dispersal, adult tree and
Runkle, J.R. (1985). Disturbance regimes in temperate seedling density on the spatial genetic structure of
forests. In: The Ecology of Natural Disturbance and regeneration at fine temporal and spatial scales.
Patch Dynamics (ed. S.T.A. Pickett and P.S. White), Tree Genet. Genomes 7: 37–48.
17–33. New York: Academic Press. Sakai, A. and Larcher, W. (1987). Frost Survival of
Runkle, J.R. (1990). Gap dynamics in an Ohio Acer- Plants. New York: Springer-Verlag 321 pp.
Fagus forest and speculations on the geography of Sakai, A. and Weiser, C.J. (1973). Freezing resistance
disturbance. Can. J. For. Res. 20: 632–641. of trees in North America with reference to tree
Running, S.W., Nemani, R.R., Peterson, D.I. et al. regions. Ecology 54: 118–126.
(1989). Mapping regional forest evapotranspiration Salisbury, F.B. and Ross, C.W. (1992). Plant Physiology,
and photosynthesis by coupling satellite data with 4e. Belmont, CA: Wadsworth 682 pp.
ecosystem simulation. Ecology 70: 1090–1101. Salles, J.M. (2011). Valuing biodiversity and ecosystem
Rupp, D.E., Shafer, S.L., Daly, C. et al. (2020). Tempera- services: why put economic values on nature? C. R.
ture gradients and inversions in a forested Cascade Biol. 334: 469–482.
712 References
Samman, S. and Logan, J.A. (2000). Assessment and Schaetzl, R.J., Burns, S.F., Small, T.W., and Johnson, D.L.
Response to Bark Beetle Outbreaks in the Rocky (1990). Tree uprooting: review of types and patterns
Mountain Area. USDA For. Serv. Gen. Tech. Rep. of soil disturbance. Phys. Geogr. 11: 277–291.
RMRS-GTR-62. Rocky Mountain Res. Sta., Ogden, Schaffalitzky De Muckadell, M. (1959). Investigations
Utah. 46 pp. on aging of apical meristems in woody plants and
Šamonil, P., Schaetzl, R.J., Valtera, M. et al. (2013). its importance in silviculture. Det Forstl. Forggsv. i
Crossdating of disturbances by tree uprooting: can Danmark 25:309–455.
treethrow microtopography persist for 6000 years? Schaffalitzky De Muckadell, M. (1962). Environmental
For. Ecol. Manag. 307: 123–135. factors in development stages of trees. In: Tree
Šamonil, P., Daněk, P., Schaetzl, R.J. et al. (2015). Soil Growth (ed. T.T. Kozlowski), 289–298. New York:
mixing and genesis as affected by tree uprooting in Ronald Press.
three temperate forests. Eur. J. Soil Sci. 66: 589–603. Scheller, R.M. (2018). The challenges of forest modeling
Sampson, H.C. (1930). Succession in the swamp given climate change. Landsc. Ecol. 33: 1481–1488.
forest formation in northern Ohio. Ohio J. Sci. 30: Scheller, R.M., Domingo, J.B., Sturtevant, B.R. et al.
340–356. (2007). Design, development, and application of
Sanchez, P.A. (1976). Properties and Management of LANDIS-II, a spatial landscape simulation model
Tropical Soils. New York: Wiley 619 pp. with flexible temporal and spatial resolution. Ecol.
Sánchez-Pinillos, M., Leduc, A., Ameztegui, A. et al. Model. 201: 409–419.
(2019). Resistance, resilience or change: post- Schier, G.A. (1975). Deterioration of aspen clones in
disturbance dynamics of boreal forests after insect the Middle Rocky Mountains. USDA For. Serv. Res.
outbreaks. Ecosystems 22: 1886–1901. Paper INT-170. Intermountain For. and Rge. Exp.
Sankey, T.T., Montagne, C., Graumlich, L. et al. (2006). Sta., Ogden, Utah. 14 pp.
Twentieth century forest–grassland ecotone shift Schimel, J.P., Firestone, M.K., and Killham, K.S. (1984).
in Montana under differing livestock grazing Identification of heterotrophic nitrification in a
pressure. For. Ecol. Manag. 234: 282–292. Sierran forest soil. Appl. Environ. Microbiol. 48:
Sarvas, R. (1962). Investigations on the flowering and 802–806.
seed crop of Pinus silvestris. Comm. Inst. For. Fenn. Schimel, D.S., Coleman, D.C., and Horton, K.A. (1985).
53: 1–198. Organic matter dynamics in paired range-land
Sarvas, R. (1969). Genetical adaptation of forest trees to and cropland toposequences in North Dakota.
the heat factor of the climate. Second world con- Geoderma 36: 201–214.
sultation on forest tree breeding, Washington, DC. Schimper, A.F.W. (1898). Pflanzen-Geographie auf Physi-
FO-FTB-69–2/15, 11 pp. ologischer Grundlage. Jena: Gustav Fischer 876 pp.
Sato, K. and Iwasa, Y. (1993). Modelling of wave regenera- Schlenker, G. (1960). Zum Problem der Einordnung
tion in subalpine Abies forests: population dynamics klimatischer Unterschiede in das System der Wald-
with spatial structure. Ecology 74: 1538–1550. standorte Baden-Württembergs. Mitt. Vereins forstl.
Saunders, D.A., Hobbs, R.H., and Margules, C.R. Standortsk. Forstpflz. 9:3–15.
(1991). Biological consequences of ecosystem Schlenker, G. (1964). Entwicklung des in Südwest-
fragmentation: a review. Conserv. Biol. 5: 18–32. deutschland angewandten Verfahrens der
Savage, M. and Swetnam, T.W. (1990). Early 19th- forstlichen Standortskunde. In Standort, Wald und
century fire decline following sheep pasturing Waldwirschaft in Oberschwabèn, “Oberschwäbische
in a Navajo ponderosa pine forest. Ecology 71: Fichtenreviere.” Stuttgart.
2374–2378. Schlesinger, W.H. (1978). On the relative dominance
Savolainen, O., Pyhäjärvi, T., and Knürr, T. (2007). Gene of shrubs in Okefenokee Swamp. Am. Nat. 112:
flow and local adaptation in trees. Annu. Rev. Ecol. 949–954.
Evol. Syst. 38: 595–619. Schlesinger, W.H. and Marks, P.L. (1977). Mineral
Sax, D.F. (2002). Native and naturalized plant diversity cycling and the niche of Spanish moss, Tillandsia
are positively correlated in scrub communities of usneoides L. Am. J. Bot. 64: 1254–1262.
California and Chile. Divers. Distrib. 8: 193–210. Schlichting, C.D. (1986). The evolution of pheno-
Schaetzl, R.J. and Follmer, L.R. (1990). Longevity of typic plasticity in plants. Annu. Rev. Ecol. Syst. 17:
treethrow microtopography: implications for mass 667–693.
wasting. Geomorphology 3: 113–123. Schmidt, H. (1970). Versuche über die Pollenverteilung
Schaetzl, R.J., Johnson, D.L., Burns, S.F., and Small, in einem Kiefernbestand. Diss. Forstl. Fak. Univ.
T.W. (1989a). Tree uprooting: review of termi- Göttingen, Germany.
nology, process, and environmental implications. Schmidt, W.C. and Dufour, W.P. (1975). Building a
Can. J. For. Res. 19: 1–11. natural area system for Montana. Western Wild-
Schaetzl, R.J., Burns, S.F., Johnson, D.L., and Small, lands 2: 20–29.
T.W. (1989b). Tree uprooting: review of impact on Schmidt, W.C. and Shearer, R.C. (1971). Ponderosa pine
forest ecology. Vegetatio 79: 165–176. seed—for animals or trees? USDA For. Serv. Res.
References 713
Paper INT-112. Intermountain For. and Rge. Exp. Schwertmann, U. and Taylor, R.M. (1989). Iron oxides.
Sta., Ogden, Utah. 14 pp. In: Minerals in Soil Environments, 2e (ed. J.B. Dixon
Schmitz, O.J., Kalies, E.L., and Booth, M.G. (2006). and S.B. Weed). Madison, WI: Soil Sci. Soc. Amer.
Alternative dynamic regimes and trophic control of Sconiers, W.B. and Eubanks, M.D. (2017). Not all
plant succession. Ecosystems 9: 659–672. droughts are created equal? The effects of stress
Schoenike, R.R. (1976). Geographical variations in jack severity on insect herbivore abundance. Arthropod
pine (Pinus banksiana). Univ. Minnesota. Agr. Exp. Plant Interact. 11: 45–60.
Sta. Tech. Bull. 304. 47 pp. Scott, D.A., Proctor, J., and Thompson, J. (1992). Eco-
Schoennagel, T., Turner, M.G., and Romme, W.H. logical studies on a lowland evergreen rain forest
(2003). The influence of fire interval and serotiny on Maraca Island, Brazil. II. Litter and nutrient
on postfire lodgepole pine density in Yellowstone cycling. J. Ecol. 80: 705–717.
National Park. Ecology 84: 2967–2978. Searle, E.B. and Chen, H.Y. (2018). Temporal declines
Schoennagel, T., Veblen, T.T., and Romme, W.H. (2004). in tree longevity associated with faster lifetime
The interaction of fire, fuels, and climate across growth rates in boreal forests. Environ. Res. Lett. 13:
Rocky Mountain forests. Bioscience 54: 661–676. e125003.
Schowalter, T.D. (1989). Canopy arthropod community Sebald, O. (1964). Ökologische Artengruppen für den
structure and herbivory in old-growth and regen- Wuchsbezirk “Oberer Neckar.” Mitt. Vereins forstl.
erating forests in western Oregon. Can. J. For. Res. Standortsk. Forstpflz. 14:60–63.
19: 318–322. Segraves, K.A. and Anneberg, T.J. (2016). Species
Schrodt, F., Bailey, J.J., Kissling, W.D. et al. (2019). To interactions and plant polyploidy. Am. J. Bot. 103:
advance sustainable stewardship, we must docu- 1326–1335.
ment not only biodiversity but geodiversity. Proc. Seidl, R., Schelhaas, M.J., Rammer, W., and Verkerk, P.J.
Natl. Acad. Sci. 116: 16155–161658. (2014). Increasing forest disturbances in Europe
Schüle, W. (1990). Landscape and climate in pre- and their impact on carbon storage. Nat. Clim.
history: interactions of wildlife, man, and fire. Chang. 4: 806–810.
In: Fire in the Tropical Biota (Ecological Studies Seidl, R., Thom, D., Kautz, M. et al. (2017). Forest dis-
84) (ed. J.G. Goldammer), 273–318. New York: turbances under climate change. Nat. Clim. Chang.
Springer-Verlag. 7: 395–402.
Schuler, T.M. and Smith, F.W. (1988). Effect of species Sercu, B.K., Baeten, L., van Coillie, F. et al. (2017). How
mix on size/density and leaf-area relations in tree species identity and diversity affect light trans-
southwest pinyon/juniper woodlands. For. Ecol. mittance to the understory in mature temperate
Manag. 25: 211–220. forests. Ecol. Evol. 7: 10861–10870.
Schullery, P. (1989). The fires and fire policy. Bioscience Seymour, R.S. and Hunter, M.L. (1999). Principles of
39: 686–694. ecological forestry. In: Maintaining Biodiversity in
Schulze, E.D., Fuchs, M.I., and Fuchs, M. (1977). Forest Ecosystems (ed. M.L. Hunter), 22–61. Cam-
Spacial distribution of photosynthetic capacity bridge: Cambridge University Press.
and performance in a montane spruce forest Shaffer, M. (2015). Changing filters. Conserv. Biol. 29:
of northern Germany. I. Biomass distribution 611–612.
and daily CO2 uptake in different crown layers. Sharik, T.L. and Barnes, B.V. (1976). Phenology of shoot
Oecologia 29: 43–61. growth among diverse populations of yellow birch
Schupp, E.W., Jordano, P., and Gómez, J.M. (2010). (Betula alleghaniensis) and sweet birch (B. lenta).
Seed dispersal effectiveness revisited: a conceptual Can. J. Bot. 54: 2122–2129.
review. New Phytol. 188: 333–353. Sharik, T.L., Feret, P.P., and Dyer, R.W. (1990).
Schupp, E.W., Jordano, P., and Gómez, J.M. (2017). Recovery of the endangered Virginia round-
A general framework for effectiveness concepts in leaf birch (Betula uber): a decade of effort. In
mutualisms. Ecol. Lett. 20: 577–590. R.S. Mitchell, C.J. Sheviak, and D.J. Leopold
Schuster, R.L., Logan, R.L., and Pringle, P.T. (1992). (eds.), Ecosystem Management: Rare Species and
Prehistoric rock avalanches in the Olympic Moun- Significant Habitats. Proc. 15th Ann. Natural Areas
tains, Washington. Science 258: 1620–1621. Conference. New York State Museum Bull. 471.
Schuur, E.A., McGuire, A.D., Schädel, C. et al. (2015). Sharov, A.A., Leonard, D., Liebhold, A.M. et al. (2002).
Climate change and the permafrost carbon “Slow the spread”: a national program to contain
feedback. Nature 520: 171–179. the gypsy moth. J. For. 100: 30–36.
Schwartz, M.D., Ahas, R., and Aasa, A. (2006). Onset Sharrow, S.H. (2007). Soil compaction by grazing live-
of spring starting earlier across the northern hemi- stock in silvopastures as evidenced by changes in
sphere. Glob. Chang. Biol. 12: 343–351. soil physical properties. Agrofor. Syst. 71: 215–223.
Schweingruber, F.H. (1989). Tree Rings: Basics and Shaw, K.L. and Mullen, S.P. (2011). Genes versus phe-
Applications of Dendrochronology. New York: notypes in the study of speciation. Genetica 139:
Springer 290 pp. 649–661.
714 References
Shea, K.L. (1990). Genetic variation between and within time, aeration, and moisture content. Can. J. For.
populations of Engelmann spruce and sub-alpine Res. 12: 646–652.
fir. Genome 33: 1–8. Simard, A.J., Haines, D.A., Blank, R.W., and Frost, J.S.
Sheil, D. and Bongers, F. (2020). Interpreting forest (1983). The Mack Lake fire. USDA For. Serv. Gen.
diversity-productivity relationships: volume values, Tech. Report NC-83. North Central For. Exp. Sta. St.
disturbance histories and alternative inferences. Paul, MN. 36 pp.
For. Ecosyst. 7: 6. Simard, M., Romme, W.H., Griffin, J.M., and Turner,
Shimwell, D.W. (1971). The Description and M.G. (2011). Do mountain pine beetle outbreaks
Classification of Vegetation. Seattle: Univ. Washing- change the probability of active crown fire in lodge-
ton Press 322 pp. pole pine forests? Ecol. Monogr. 81: 3–24.
Shirley, H.L. (1945b). Reproduction of upland conifers Simberloff, D. (2013). Invasive Species: What Everyone
in the Lake states as affected by root competition Needs to Know. Oxford University Press 352 pp.
and light. Am. Midi. Nat. 33: 537–612. Simberloff, D. (2015). Non-native invasive species and
Shiva, V. (1990). Biodiversity, biotechnology, and profit: novel ecosystems. F1000Prime Reports 7:47.
the need for a peoples’ plan to protect biological Simberloff, D., Murcia, C., and Aronson, C. (2015).
diversity. Ecologist 20: 44–17. “Novel ecosystems” are a Trojan horse for
Short, H.L. (1976). Composition and squirrel use of conservation. http://ensia.com/voices/novel-
acorns of black and white oak groups. J. Wildl. ecosystems-are-a-trojanhorse-for-conservation
Manag. 40: 479–483. (accessed 20 Nov 2021).
Shotola, S.J., Weaver, G.T., Robertson, P.A., and Ashby, Simpson, E.H. (1949). Measurement of diversity. Nature
W.C. (1992). Sugar maple invasion of an old-growth 163: 688.
oak-hickory forest in southwestern Illinois. Am. Simpson, G.G. (1952). How many species? Evolution 6:
Midl. Nat. 127: 125–138. 342.
Show, S.B., and Kotok, E.I. (1929). Cover type and Simpson, T.A. (1990). Landscape Ecosystems and Cover
fire control in the National Forests of northern Types of the Reserve Area and Adjoining Land of the
California. USDA, Dept. Bull. 1495. Washington, Huron Mountain Club, Marquette Co., Michigan.
D.C. 35 pp. Ph.D. Thesis, University of Michigan, Ann Arbor.
Shrader-Frechette, K.S. and McCoy, E.D. (1993). Method 384 pp.
in Ecology: Strategies for Conservation. Cambridge: Simpson, T.A., Stuart, P.E., and Barnes, B.V. (1990).
Cambridge Univ. Press 328 pp. Landscape ecosystems and cover types of the
Shuman, B., Webb III, T., Bartlein, P. et al. (2002). The Reserve Area and adjoining lands of the Huron
anatomy of a climatic oscillation: vegetation change Mountain Club, Marquette Co., MI. Huron
in eastern North America during the Younger Mountain Wildlife Foundation Occasional Paper
Dryas chronozone. Quaternary Science Reviews 21: No. 4. 128 pp.
1777–1791. Sims, R.A. and Uhlig, P. (1992). The current status of
Siemann, E., Carrillo, J.A., Gabler, C.A. et al. (2009). forest site classification in Ontario. For. Chron. 68:
Experimental test of the impacts of feral hogs on 64–77.
forest dynamics and processes in the southeastern Sims, R.A., Towill, W.D., Baldwin, K.A., and Wickware, G.M.
US. For. Ecol. Manag. 258: 546–553. (1989). Field Guide to the Forest Ecosystem
Silen, R.R. (1978). Genetics of Douglas-fir. USDA For. Classification for Northwestern Ontario. Ont. Min.
Serv. Res. Paper W0–35, Washington D.C. 34 pp. Nat. Res. Toronto, ON. 191 pp.
Silkworth, D.R. and Grigal, D.F. (1982). Determining Singer, J.A., Turnbull, R., Foster, M. et al. (2019).
and evaluating nutrient losses from whole-tree Sudden aspen decline: a review of pattern and pro-
harvesting of aspen. Soil Sci. Soc. Am. J. 46: cess in a changing climate. Forests 10: 671.
626–631. Sirén, G. (1955). The development of spruce forest
Sillett, S.C., McCune, B., Peck, J.E. et al. (2000). Dis- on raw humus sites in northern Finland and its
persal limitations of epiphytic lichens result in ecology. Acta For. Fenn. 62 (4): 1–408.
species dependent on old-growth forests. Ecol. Sittaro, F., Paquette, A., Messier, C., and Nock, C.A.
Appl. 10: 789–799. (2017). Tree range expansion in eastern North
Silvertown, J.W. (1980). The evolutionary ecology America fails to keep pace with climate warming
of mast seeding in trees. Biol. J. Linn. Soc. 14: at northern range limits. Glob. Chang. Biol. 23:
235–250. 3292–3301.
Silvertown, J.W. and Lovett-Doust, J. (1993). Introduc- Skroch, M. and López-Hoffman, L. (2009). Saving
tion to Plant Population Biology. Oxford: Blackwell nature under the big tent of ecosystem services: a
210pp. response to Adams and Redford. Conserv. Biol. 24:
Silvester, W.B., Sollins, P., Verhoeven, T., and Cline, S.P. 325–327.
(1982). Nitrogen fixation and acetylene reduction Slaughter, C.W., Barnes, R.J., and Hansen, G.M. (eds.).
in decaying conifer boles: effects of incubation (1971). Fire in the Northern Environment—A
References 715
Symposium. USDA For. Serv. Pacific Northwest For. range-expanding mangroves. Mar. Ecol. Prog. Ser.
and Rge. Exp. Sta., Portland, OR. 275 pp. 644: 65–74.
Smalley, G.W. (1984). Classification and evaluation of Snaydon, R.W. and Davies, M.S. (1972). Rapid
forest sites in the Cumberland Mountains. USDA population differentiation in a mosaic environ-
For. Serv. Gen. Tech. Report SO-50. Southern For. ment. II. Morphological variation in Anthoxanthum
Exp. Sta., New Orleans, LA. 84 pp. odoratum. Evolution 26: 390–405.
Smalley, G.W. (1991). No more plots; go with what you Snaydon, R.W. and Davies, M.S. (1976). Rapid
know: developing a forest land classification system population differentiation in a mosaic environ-
for the Interior Uplands. In D.L. Mengel and D.T. ment. IV. Populations of Anthoxanthum odoratum
Tew (eds.), Symp. Proc. Ecological Land Classification: at sharp boundaries. Heredity 37: 9–25.
Applications to Identify and Productive Potential of Sobhani, V.M., Barrett, M., and Peterson, C.J. (2014).
Southern Forests. USDA For. Serv. Gen. Tech. Report Robust prediction of treefall pit and mound sizes
SE-68. Southeastern For. Exp. Sta., Asheville, NC. from tree size across 10 forest blowdowns in east-
Smilanich, A.M., Fincher, R.M., and Dyer, L.A. (2016). ern North America. Ecosystems 17: 837–850.
Does plant apparency matter? Thirty years of data Soil Survey Staff (1975). Soil Taxonomy. USDA Agr.
provide limited support but reveal clear patterns of Handbook 436. Washington, D.C. 754 pp.
the effects of plant chemistry on herbivores. New Solbrig, O.T. (1970). Principles and Methods of Plant
Phytol. 210: 1044–1057. Biosystematics. Toronto: Collier-Macmillan 226 pp.
Smith, D.M. (1951). The influence of seedbed condi- Solbrig, O.T. (1991). The origin and function of biodi-
tions on the regeneration of eastern white pine. versity. Environment 33 (16–20): 34–38.
Conn. Agr. Exp. Sta. Bull. 545: 61 pp. Sollars, E.S.A. and Buggs, R.J.A. (2018). Genome-wide
Smith, R.H. (1966). Resin quality as a factor in the epigenetic variation among ash trees differing in
resistance of pines to bark beetles. In: Breeding susceptibility to a fungal disease. BMC Genom.
Pest-Resistant Trees (ed. H.D. Gerhold, R.E. McDer- 19: 502.
mott, E.J. Schreiner and J.A. Winieski), 189–196. Sollins, P., Robertson, G.P., and Uehara, G. (1988).
New York: Pergamon Press. Nutrient mobility in variable-and permanent-
Smith, R.H. (1977). Monoterpenes of ponderosa pine charge soils. Biogeochemistry 6: 181–199.
xylem resin in western United States. USDA For. Solórzano, L.A. and Páez-Acosta, G.I. (2009). Forests.
Serv., Tech Bull. 1532. 48 pp. In: The Princeton Guide to Ecology (ed. S.A. Levin,
Smith, H. (1982). Light quality, photoperception S.R. Carpenter, H.C.J. Godfray, et al.), 606–613.
and plant strategy. Annu. Rev. Plant Physiol. 33: Princeton, NJ: Princeton University Press.
481–518. Soltis, P.S. (2005). Ancient and recent polyploidy in
Smith, C.C. (1990). The advantage of mast years for angiosperms. New Phytol. 166: 5–8.
wind pollination. Am. Nat. 136: 154–166. Soltis, D.E., Morris, A.B., McLachlan, J.S. et al. (2006).
Smith, B.R. and Blumstein, D.T. (2008). Fitness con- Comparative phylogeography of unglaciated
sequences of personality: a meta-analysis. Behav. eastern North America. Mol. Ecol. 15: 4261–4293.
Ecol. 19: 448–455. Soltis, D.E., Albert, V.A., Leebens-Mack, J. et al. (2009).
Smith, W.B., and Brand, G.J. (1983). Allometric biomass Polyploidy and angiosperm diversification. Am. J.
equations for 98 species of herbs, shrubs and small Bot. 96: 336–348.
trees. USDA For. Serv. Note NC-299. North Central Soltis, D.E., Visger, C.J., and Soltis, P.S. (2014). The
For. Exp. Sta., St. Paul, MN. 8 pp. polyploidy revolution then. . . and now: Stebbins
Smith, D.D. and Follmer, D. (1972). Food preferences of revisited. Am. J. Bot. 101: 1057–1078.
squirrels. Ecology 53: 82–91. Soltis, P.S., Marchant, D.B., Van de Peer, Y., and Soltis,
Smith, T. and Huston, M. (1989). A theory of the spatial D.E. (2015). Polyploidy and genome evolution in
and temporal dynamics of plant communities. plants. Curr. Opin. Genet. Dev. 35: 119–125.
Plant Ecol. 83: 49–69. Soolanayakanahally, R.Y., Guy, R.D., Silim, S.N., and
Smith, T.J. III, Robblee, M.B., Wanless, H.R., and Doyle, Song, M. (2013). Timing of photoperiodic com-
T.W. (1994). Mangroves, hurricanes, and lightning petency causes phenological mismatch in balsam
strikes. Bioscience 44: 256–262. poplar (Populus balsamifera L.). Plant Cell Environ.
Smith, D.M., Larson, B.C., Kelty, M.J., and Ashton, 36: 116–127.
P.M.S. (1997). The Practice of Silviculture, 9e. Sork, V.L., Bramble, J., and Owen, S. (1993). Ecology of
New York: Wiley 537 pp. mast-fruiting in three species of North American
Smith, N.R., Kishchuk, B.E., and Mohn, W.W. (2008). deciduous oaks. Ecology 74: 528–541.
Effects of wildfire and harvest disturbances on Sork, V.L., Aitken, S.N., Dyer, R.J. et al. (2013).
forest soil bacterial communities. Appl. Environ. Putting the landscape into the genomics of trees:
Microbiol. 74: 216–224. approaches for understanding local adaptation
Smith, R.S., Blaze, J.A., and Byers, J.F. (2020). Negative and population responses to changing climate.
indirect effects of hurricanes on recruitment of Tree Genet. Genomes 9: 901–911.
716 References
Sow, M.D., Allona, I., Ambroise, C. et al. (2018). Epige- Sprugel, D.G. and Bormann, F.H. (1981). Natural distur-
netics in forest trees: state of the art and poten- bance and the steady state in high-altitude balsam
tial implications for breeding and management fir forests. Science 211: 390–393.
in a context of climate change. Adv. Bot. Res. 88: Spurr, S.H. (1945). A new definition of silviculture.
387–453. J. For. 43: 44.
Sow, M.D., Le Gac, A.-L., Fichot, R. et al. (2021). RNAi Spurr, S.H. (1952a). Origin of the concept of forest
suppression of DNA methylation affects the succession. Ecology 33: 426–427.
drought stress response and genome integrity in Spurr, S.H. (1952b). Forest Inventory. New York: Ronald
transgenic poplar. New Phytol. 232: 80–97. Press 476 pp.
Speer, J.H. (2010). Fundamentals of Tree-Ring Research. Spurr, S.H. (1954). The forests of Itasca in the
Tucson, AZ: University of Arizona Press 360 pp. nineteenth century as related to fire. Ecology 35:
Sperry, J.S., Hacke, U.G., and Pittermann, J. (2006). Size 21–25.
and function in conifer tracheids and angiosperm Spurr, S.H. (1956a). Natural restocking of forests fol-
vessels. Am. J. Bot. 93: 1490–1500. lowing the 1938 hurricane in central New England.
Sperry, J.S., Venturas, M.D., Todd, H.N. et al. (2019). Ecology 37: 433–451.
The impact of rising CO2 and acclimation on the Spurr, S.H. (1956b). Forest associations in the Harvard
response of US forests to global warming. Proc. Forest. Ecol. Monogr. 26: 245–262.
Natl. Acad. Sci. 116: 25734–25744. Spurr, S.H. (1957). Local climate in the Harvard Forest.
Spies, T.A. (1983). Classification and Analysis of Forest Ecology 38: 37–46.
Ecosystems of the Sylvania Recreation Area, Upper Spurr, S.H. (1963). Growth of Douglas-fir in New Zea-
Michigan. Ph.D. Thesis, University of Michigan, land. New Zealand For. Serv. For. Res. Inst. Tech.
Ann Arbor. 321 pp. Paper 43. 54 pp.
Spies, T.A. (1997). Forest stand structure, composition, Spurr, S.H. (1964). Forest Ecology. New York: Ronald
and function. In: Creating a Forestry for the 21st Press 352 pp.
Century (ed. K.A. Kohm and J.F. Franklin), 11–30. Spurr, S.H. and Barnes, B.V. (1973). Forest Ecology, 2e.
Washington, D.C: Island Press. New York: Ronald Press 571 pp.
Spies, T.A. and Barnes, B.V. (1981). A morphological Spurr, S.H. and Barnes, B.V. (1980). Forest Ecology, 3e.
analysis of Populus alba, P. grandidentata and their New York: Wiley 687 pp.
natural hybrids in southeastern Michigan. Silvae
Squillace, A.E. (1966). Geographic variation in slash
Genet. 30: 102–106.
pine. For. Sci. Monogr. 10. 56 pp.
Spies, T.A. and Barnes, B.V. (1985a). A multi-factor
St. Clair, B.J., Mandel, N.L., and Vance-Borland, K.W.
ecological classification of the northern hardwood
(2005). Genecology of Douglas-fir in western
and conifer ecosystems of the Sylvania Recreation
Oregon and Washington. Ann. Bot. 96: 1199–1214.
Area, Upper Peninsula, Michigan. Can. J. For. Res.
Staaf, H. and Berg, B. (1982). Accumulation and release
15: 949–960.
of plant nutrients in decomposing Scots pine
Spies, T.A. and Barnes, B.V. (1985b). Ecological species
needle litter. Long-term decomposition in a Scots
groups of upland northern hardwood-hemlock
pine forest II. Can. J. Bot. 60: 1561–1568.
forest ecosystems of the Sylvania Recreation Area,
Stacey, P.B. and Koenig, W.D. (1984). Cooperative
Upper Peninsula, Michigan. Can. J. For. Res. 15:
breeding in the acorn woodpecker. Sci. Am. 252:
961–972.
114–121.
Spies, T.A. and Franklin, J.F. (1989). Gap characteristics
Stage, A.R. (1989). Utility of vegetation-based land
and vegetation response in tall coniferous forests.
classes for predicting forest regeneration and
Ecology 70: 543–545.
growth. In D.E. Ferguson, P. Morgan, F.D. Johnson
Spies, T.A. and Franklin, J.F. (1991). The structure of
(comps.), Proceedings—Land Classifications
natural young, mature, and old-growth forests in
Based on Vegetation: Applications for Resource
Washington and Oregon. In L.F. Ruggiero,
Management. USDA For. Serv. Gen. Tech. Report
K.B. Aubry, A.B. Carey, and M. H. Huff (tech.
INT-257. Intermountain Res. Sta., Ogden, UT.
coords.), Wildlife and Vegetation of Unmanaged
Staley, D.M., Kean, J.W., and Rengers, F.K. (2020). The
Douglas-fir forests. USDA For. Serv. PNW-GTR-385.
recurrence interval of post-fire debris-flow gener-
Pacific Northwest Exp. Sta., Portland, OR.
ating rainfall in the southwestern United States.
Spies, T.A., Franklin, J.F., and Klopsch, M. (1990).
Geomorphology 370: 107392.
Canopy gaps in Douglas-fir forests of the Cascade
Stambaugh, M.C., Guyette, R.P., Grabner, K., and
Mountains. Can. J. For. Res. 20: 649–658.
Kolaks, J. (2006). Understanding Ozark forest litter
Sposito, G. (1989). The Chemistry of Soils. New York:
variability through a synthesis of accumulation
Oxford Univ. Press 277 pp.
rates and fire events. In B.W. Butler, P.I. Andrews
Sprugel, D.G. (1976). Dynamic structure of wave- (comps.). Fuels Management-How To Measure
regenerated Abies balsamea forests in the north- Success. Conference Proceedings. 28–30 March,
eastern United States. J. Ecol. 64: 889–911. 2006. Portland, OR. USDA For. Serv. Gen. Tech.
References 717
Rep. RMRS-P-41. Rocky Mountain Res. Sta., Fort Stevens-Rumann, C.S. and Morgan, P. (2019). Tree
Collins, CO. regeneration following wildfires in the western US:
Stamp, N. (2003). Out of the quagmire of plant defense a review. Fire Ecol. 15: 1–17.
hypotheses. Quart. Rev. Bot. 78: 23–55. Stevens-Rumann, C.S., Kemp, K.B., Higuera, P.E. et al.
Stamp, N. (2004). Can the growth-differentiation (2018). Evidence for declining forest resilience
balance hypothesis be tested rigorously? Oikos 107: to wildfires under climate change. Ecol. Lett. 21:
439–448. 243–252.
Stanturf, J.A., Wade, D.D., Waldrop, T.A. et al. 2002. Stewart, G.H., Rose, A.B., and Veblen, T.T. (1991).
Fire in southern forest landscapes. In D.M. Wear, J. Forest development in canopy gaps in old-growth
Greis (eds.). Southern Forest Resource Assessment. beech (Nothofagus) forests, New Zealand. J. Veg. Sci.
USDA For. Serv. Gen. Tech. Rep. SRS-53. Southern 2: 679–690.
Res. Sta., Asheville, NC. Stewart, J.R., Lister, A.M., Barnes, I., and Dalén, L.
Stark, N. (1968). Seed ecology of Sequoiadendron gigan- (2010). Refugia revisited: individualistic responses
teum. Madroño 19: 267–277. of species in space and time. Proc. R. Soc. B Biol.
Stark, N.M. (1977). Fire and nutrient cycling in a Sci. 277: 661–671.
Douglas-fir/larch forest. Ecology 58: 16–30. Stoeckeler, J.H. (1948). The growth of quaking aspen
Stebbins, G.L. (1950). Variation and Evolution in Plants. as affected by soil properties and fire. J. For. 46:
New York: Columbia Univ. Press 643 pp. 727–737.
Stebbins, G.L. (1969). The significance of hybridiza- Stoeckeler, J.H. (1960). Soil factors affecting the
tion for plant taxonomy and evolution. Taxon 18: growth of quaking aspen forests in the Lake
26–35. states. Univ. Minnesota Agr. Exp. Sta. Tech. Bull.
Stebbins, G.L. (1970). Variation and evolution in 233. 48 pp.
plants: progress during the past twenty years. Stohlgren, T.J. (2007). Measuring Plant Diversity: Lessons
In: Essays in Evolution and Genetics (ed. M.K. from the Field. New York: Oxford University Press
Hecht and W.C. Steere), 173–208. New York: 408 pp.
Appleton-Century-Crofts. Stone, E.L. (1974). The communal root system of
Stebbins, G.L. (1971). Processes of Organic Evolution, 2e. red pine: growth of girdled trees. For. Sci. 20:
Englewood Cliffs, NJ: Prentice-Hall 193 pp. 294–305.
Stebbins, G.L. (1985). Polyploidy, hybridization, and the Stone, E.L. (1975). Windthrow influences on spatial
invasion of new habitats. Ann. Missouri Bot. Gard. heterogeneity in a forest soil. Mitt. Eid. Anst. forstl.
72: 824–832. Versuchsw. 51:77–87.
Steele, B.M. and Cooper, S.V. (1986). Predicting site Stone, D.M. (1977). Leaf dispersal in a pole-sized maple
index and height for selected tree species of stand. Can. J. For. Res. 7: 189–192.
northern Idaho. USDA For. Serv. Res. Paper INT- Stone, E.L. and Kalisz, P.J. (1991). On the maximum
126. Int. For. and Rge. Exp. Sta., Ogden, UT. 16 p. extent of tree roots. For. Ecol. Manag. 46: 59–102.
Steele, M.A., Knowles, T., Bridle, K., and Simms, E.L. Stone, E.L. and Stone, M.H. (1954). Root collar sprouts
(1993). Tannins and partial consumption of acorns: in pine. J. For. 52: 487–491.
implications for dispersal of oaks by seed predators. Storch, F., Dormann, C.F., and Bauhus, J. (2018). Quan-
Am. Midl. Nat. 130: 229–238. tifying forest structural diversity based on large-
Steinbrenner, E.C. (1951). Effect of grazing on floristic scale inventory data: a new approach to support
composition and soil properties of farm woodlands biodiversity monitoring. For. Ecosyst. 5: 34.
in southern Wisconsin. J. For. 49: 906–910. Stottlemyer, A.D., Shelburne, V.B., Waldrop, T.A. et al.
Stephenson, S.L. (1974). Ecological composition of (2009). Fuel characterization in the southern
some former oak-chestnut communities in western Appalachian Mountains: an application of
Virginia. Castanea 39: 278–286. landscape ecosystem classification. Int. J. Wildland
Stephenson, N. L. and Brigham, C. (2021). Preliminary Fire 18: 423–429.
estimates of sequoia mortality in the 2020 Castle Stowe, K., Marquis, R., Hochwender, C., and Simms,
Fire. Sequoia and Kings Canyon National Parks: E. (2000). The evolutionary ecology of tolerance
National Park Service. https://www.nps.gov/ to consumer damage. Annu. Rev. Ecol. Syst. 31:
articles/000/preliminary-estimates-of-sequoia- 565–595.
mortality-in-the-2020-castle-fire.htm. Accessed Stralberg, D., Carroll, C., Pedlar, J.H. et al. (2018).
01-18-2022. Macrorefugia for North American trees and song-
Sterba, H. and Monserud, R.A. (1993). The maximum birds: climatic limiting factors and multi-scale
density concept applied to uneven-aged mixed- topographic influences. Glob. Ecol. Biogeogr. 27:
species stands. For. Sci. 39: 432–452. 690–703.
Sterck, F. (2005). Woody tree architecture. In: Plant Strauss, S.Y., Webb, C.O., and Salamin, N. (2006). Exotic
Architecture and its Manipulation (ed. C. Turnbull), taxa less related to native species are more invasive.
209–237. Oxford, UK: Blackwell. Proc. Natl. Acad. Sci. 103: 5841–5845.
718 References
Strimbeck, R.G., Schaberg, P.G., Fossdal, C.G. et al. 14 (ed. R.L. Edmonds), 267–291. Stroudsburg, PA:
(2015). Extreme low temperature tolerance in Hutchinson Ross.
woody plants. Front. Plant Sci. 6: 1–15. Swanson, F.J., Kratz, T.K., Caine, N., and Woodma-
Stromberg, J.C., Chew, M.K., Nagler, P.I., and Glenn, nsee, R.G. (1988). Landform effects on ecological
E.P. (2009). Changing perceptions of change: the processes and features. Bioscience 38: 92–98.
role of scientists in Tamarix and river management. Swanson, F.J., Franklin, F.J., and Sedell, J.R. (1990).
Restor. Ecol. 17: 177–186. Landscape patterns, disturbance, and management
Strong, W.L., Oswald, E.T., and Downing, D.J. (eds.). in the Pacific Northwest, USA. In: Changing Land-
(1990). The Canadian vegetation classification scapes: An Ecological Perspective (ed. I.S. Zonneveld
system. First Approx. Ecological Land Classif. and R.T.T. Forman), 191–213. New York:
Series, No. 25, Env. Canada, Ottawa. 22 pp. Springer-Verlag.
Stuckey, R.L. (1981). Origin and development of the Swanson, M.E., Franklin, J.F., Beschta, R.L. et al.
concept of the Prairie Peninsula. In R.L. Stuckey (2011). The forgotten stage of forest succession:
and K.J. Reese (eds.), The Prairie Peninsula—In the early-successional ecosystems on forest sites. Front.
“Shadow” of Transeau. Ohio Biol. Surv. Biol. Notes Ecol. Environ. 9: 117–125.
15. Swanston, C.W., Janowiak, M.K., Brandt, L.A. et al.
Sturrock, R.N., Frankel, S.J., Brown, A.V. et al. (2011). (2016). Forest Adaptation Resources: Climate
Climate change and forest diseases. Plant Pathol. Change Tools and Approaches for Land Man-
60: 133–149. agers. USDA For. Serv. Gen. Tech. Rep. NRS-
Sturtevant, B.R. and Seagle, S.W. (2004). Comparing GTR-87-2. Northern Res. Sta., Newtown Square,
estimates of forest site quality in old second-growth PA. 161 pp.
oak forests. For. Ecol. Manag. 191: 311–328. Swenson, J.J. and Waring, R.H. (2006). Modelled pho-
Sturtevant, B.R., Bissonette, J.A., Long, J.N., and tosynthesis predicts woody plant richness at three
Roberts, D.W. (1997). Coarse woody debris as a geographic scales across the northwestern United
function of age, stand structure, and disturbance in States. Glob. Ecol. Biogeogr. 15: 470–485.
boreal Newfoundland. Ecol. Appl. 7: 702–712. Swetnam, T.W. (1993). Fire history and climate change
Suarez-Gonzalez, A., Lexer, C., and Cronk, Q.C. (2018). in giant sequoia groves. Science 262: 885–889.
Adaptive introgression: a plant perspective. Biol. Swetnam, T.W., Baisan, C.H., Caprio, A.C. et al. (2009).
Lett. 14: 20170688. Multi-millennial fire history of the Giant Forest,
Sukachev, V.N. and Dylis, N.V. (1964). Fundamentals of Sequoia National Park, California, USA. Fire Ecol.
Forest Biogeocoenology. (trans. J. M. MacLennan). 5: 120–150.
Edinburgh: Oliver and Boyd, Ltd. 672 pp. Swift, M.J., Heal, O.W., and Anderson, J.M. (1979).
Svensson, J.R., Lindegarth, M., Jonsso, P.R., and Pavia, Decomposition in Terrestrial Ecosystems. Oxford,
H. (2012). Disturbance-diversity models: what do UK: Blackwell Scientific 372 pp.
they really predict and how are they tested? Proc. R. Swihart, R.K. and Bryant, J.P. (2001). Importance of bio-
Soc. B Biol. Sci. 279: 2163–2170. geography and ontogeny of woody plants in winter
Swain, A.M. (1973). A history of fire and vegetation in herbivory by mammals. J. Mammal. 82: 1–21.
northeast Minnesota as recorded in lake sediments. Taiz, L. and Zeiger, E. (1991). Plant Physiology. Red-
Quat. Res. 3: 383–396. wood City, CA: Benjamin/Cummings 559 pp.
Swain, A.M. (1978). Environmental changes during the Tallis, J.H. (1991). Plant Community History. New York:
past 2000 years in north-central Wisconsin: anal- Chapman and Hall 398 pp.
ysis of pollen, charcoal, and seeds from varved lake Tansley, A.G. (1929). Succession: the concept and its
sediments. Quat. Res. 10: 55–68. values. In B.M. Duggar (ed.), Proc. International
Swank, W.T. and Crossley, D.A. (ed.) (1988). Forest Congress of Plant Sciences, Ithaca, NY, 1926. George
Hydrology and Ecology at Coweeta. New York: Banta Publ. Co., Menasha, WS.
Springer-Verlag 469 pp. Tansley, A.G. (1935). The use and abuse of vegetational
Swanson, F.J. (1981). Fire and geomorphic processes. In concepts and terms. Ecology 16: 284–307.
Proc. Fire Regimes and Ecosystems. USDA For Serv. Tansley, A.G. (1939). The British Islands and their Vege-
Gen. Tech;. Report WO-26. Washington, D.C. tation. Cambridge: Cambridge Univ. Press 930 pp.
Swanson, F.J., Fredriksen, R.L., and McCorison, F.M. Tappeiner, J., Zasada, J., Ryan, P., and Newton, M.
(1982a). Material transfer in a western Oregon for- (1991). Salmonberry clonal and populations struc-
ested watershed. In: Analysis of Coniferous Forest ture: the basis for a persistent cover. Ecology 72:
Ecosystems in the Western United States. US/IBP 609–618.
Synthesis Series 14 (ed. R.L. Edmonds), 233–266. Tardif, J., Flannigan, M., and Bergeron, Y. (2001). An
Stroudsburg, PA: Hutchinson Ross. analysis of the daily radial activity of 7 boreal tree
Swanson, F.J., Gregory, S.V., Sedell, J.R., and Campbell, species, northwestern Quebec. Environ. Monit.
A.G. (1982b). Land-water interactions: the riparian Assess. 67: 141–160.
zone. In: Analysis of Coniferous Forest Ecosystems in Tarrant, R.F. (1956b). Effect of slash burning on some
the Western United States. US/IBP Synthesis Series physical soil properties. For. Sci. 2: 18–22.
References 719
Tate, K.W., Dudley, D.M., McDougald, N.K., and Thom, D., Rammer, W., and Seidl, R. (2017). The impact
George, M.R. (2004). Effect of canopy and grazing of future forest dynamics on climate: interactive
on soil bulk density. J. Range Manag. 57: 411–417. effects of changing vegetation and disturbance
Taub, D.R. and Goldberg, D. (1996). Root system regimes. Ecol. Monogr. 87: 665–684.
topology of plants from habitats of differing soil Thomas, S.C. and Winner, W.E. (2000). Leaf area index
resource availability. Funct. Ecol. 10: 258–264. of an old-growth Douglas-fir forest estimated from
Taulavuori, K.M.J., Laine, K., and Taulavuori, E.B. direct structural measurements in the canopy. Can.
(2002). Artificial deacclimation response of Vac- J. For. Res. 30: 1922–1930.
cinium myrtillus in mid-winter. Ann. Bot. Fenn. 39: Thompson, S.C.G. and Barton, M.A. (1994). Ecocentric
143–147. and anthropocentric attitudes toward the environ-
Taylor, A.R. (1974). Forest fire. In: Yearbook of Science ment. J. Environ. Psychol. 14: 149–158.
and Technology (ed. D. N. Lapedes and). London: Thompson, M.P., MacGregor, D.G., Dunn, C.J. et al.
McGraw-Hill. (2018). Rethinking the wildland fire management
Taylor, T.M.C. (1959). The taxonomic relationship bet- system. J. For. 116: 382–390.
ween Picea glauca (Moench) Voss and P. engelman- Thomsen, M.S., Altieri, A.H., Angelini, C. et al. (2018).
nii parry. Madroño 15: 111–115. Secondary foundation species enhance biodiversity.
Taylor, S.W. and Carroll, A.L. (2004). Disturbance, forest Nat. Ecol. Evol. 2: 634–639.
age, and mountain pine beetle outbreak dynamics Thomson, A.M., Dick, C.W., and Dayanandan, S.
in BC: a historical perspective. In T.L. Shore, (2015). A similar phylogeographical structure
J.E. Brooks, J.E. Stone (eds.). Mountain Pine Beetle among sympatric North American birches (Betula)
Symposium: Challenges and Solutions. Information is better explained by introgression than by shared
Report BC-X-399. Natural Resources Canada, biogeographical history. J. Biogeogr. 42: 339–350.
Canadian Forest Service, Pacific Forestry Center, Thoreau, H.D. (1993). Faith in a Seed. Washington D.C:
Victoria, BC. Island Press 283 pp.
Taylor, A.H. and Skinner, C.N. (1998). Fire history and Tiedemann, A.R., Clary, W.P., and Barbour, R.J.
landscape dynamics in a late-successional reserve, (1987). Underground systems of Gambel oak
Klamath Mountains, California, USA. For. Ecol. (Quercus gambelii) in Central Utah. Am. J. Bot. 74:
Manag. 111: 285–301. 1065–1071.
Taylor, A.H. and Skinner, C.N. (2003). Spatial patterns Tiedje, J.M. (1988). Ecology of denitrification and
and controls on historical fire regimes and forest dissimilatory nitrate reduction to ammonium. In:
structure in the Klamath Mountains. Ecological Biology of Anaerobic Microorganisms (ed. A.J.B.
Applications 13 (3): 704–719. Zehnder), 179–244. New York: Wiley.
Taylor, S., Carroll, A., Alfaro, R., and Safranyik, L. Tiedje, J.M., Simkins, S., and Groffman, P.M. (1989).
(2006). Forest, climate and mountain pine beetle Perspectives on measurement of denitrification in
outbreak dynamics in western Canada. In: The field including recommended protocols for acety-
Mountain Pine Beetle: A Synthesis of Biology, lene based methods. Plant Soil 115: 261–284.
Management, and Impacts on Lodgepole Pine Tilghman, N.G. (1989). Impacts of white-tailed deer on
(ed. L. Safranyik and B. Wilson). Pacific Forestry forest regeneration in northwestern Pennsylvania.
Centre, Victoria, BC: Natural Resources Canada, J. Wildl. Manag. 53: 524–532.
Canadian Forest Service. Tilman, D. (1982). Resource Competition and
Teipner, C.L., Garton, E.O., and Nelson, L. Jr. (1983). Community Structure. Princeton, NJ: Princeton
Pocket gophers in forest ecosystems. USDA For. University Press 296 pp.
Serv. Gen. Tech. Report INT-154. Intermountain Tilman, D. (1985). The resource-ratio hypothesis of
For. and Rge. Exp. Sta., Ogden, UT. 53 pp. plant succession. Am. Nat. 125: 827–852.
Tepley, A.J., Thompson, J.R., Epstein, H.E., and Tilman, D. (1988). Plant Strategies and the Dynamics
Anderson-Teixeira, K.J. (2017). Vulnerability and Structure of Plant Communities. Princeton, NJ:
to forest loss through altered postfire recovery Princeton Univ. Press 360 pp.
dynamics in a warming climate in the Klamath
Tilman, D. (1996). Biodiversity: population versus eco-
Mountains. Glob. Chang. Biol. 23: 4117–4132.
system stability. Ecology 77: 350–363.
Terborgh, J. (1989). Where Have all the Birds Gone?
Tilman, D., Reich, P.B., and Knops, J.M.H. (2006).
Princeton, NJ: Princeton Univ. Press 207 pp.
Biodiversity and ecosystem stability in a decade-
The Nature Conservancy (1996). Conservation by design: long grassland experiment. Nature 441: 629–632.
A framework for mission success. Arlington, VA
Tilman, D., Reich, P.B., and Isbell, F. (2012). Biodiver-
(Updated 2011). Spatial data available at http://
sity impacts ecosystem productivity as much as
maps.tnc.org. Accessed 05-09-2021.
resources, disturbance, or herbivory. Proc. Natl.
Thom, D. and Seidl, R. (2016). Natural disturbance Acad. Sci. 109: 10394–10397.
impacts on ecosystem services and biodiversity
Tinker, D.B., Romme, W.H., Hargrove, W.W. et al.
in temperate and boreal forests. Biol. Rev. 91:
(1994). Landscape-scale heterogeneity in lodgepole
760–781.
pine serotiny. Can. J. For. Res. 24: 897–903.
720 References
Tinker, D.B., Romme, W.H., and Despain, D.G. (2003). Troll, C. (1963b). Über Landschafts-Sukession. Arbeiten
Historic range of variability in landscape structure zur Rheinschen Landeskunde Bonn. 19:5–12.
in subalpine forests of the Greater Yellowstone Troll, C. (1968). Landschaftsökologie. In R. Tüxen (ed.),
Area, USA. Landsc. Ecol. 18: 427–439. Pflanzensoziologie und Landschaftsökologie. Beri-
Titus, J.H. (1990). Microtopography and woody plant chte das Internalen Symposiums der Internationalen
regeneration in a hardwood floodplain swamp in Vereinigung für Vegetationskunde, Stolzenau/Weser
Florida. Bull. Torrey Bot. Club 117: 429–127. 1963. W. Junk, The Hague.
Tomlinson, P.B. (1983). Tree architecture. Am. Scientist Troll, C. (1971). Landscape ecology (geo-ecology) and
71: 141–149. bioceonology—a terminology study. Geoforum 8:
Toumey, J.W. and Kienholz, R. (1931). Trenched plots 43–46.
under forest canopies. Yale Univ. School For. Bull. Trouet, V. (2020). Tree Story: The History of the World
30. 31 pp. Written in Rings. Baltimore, MD: Johns Hopkins
Tovar-Sanchez, E. and Oyama, K. (2006). Effect of University Press 256 pp.
hybridization of the Quercus crassifolia× Quercus Tubbs, C.H. (1965). Influence of temperature and early
crassipes complex on the community structure of spring conditions on sugar maple and yellow birch
endophagous insects. Oecologia 147: 702–713. germination in upper Michigan. USDA For. Serv.
Townsend, A.M., Bentz, S.E., and Johnson, G.R. (1995). Res. Note LS-72. North Central For. Exp. Sta.,
Variation in response of selected American elm St. Paul, MN. 2 pp.
clones to Ophlostoma ulmi. J. Environ. Hortic. 13: Tucker, G.F. and Emmington, W.H. (1977). Morphological
126–128. changes in leaves of residual western hemlock after
Transeau, E.N. (1935). The prairie peninsula. Ecology clear and shelterwood cutting. For. Sci. 23: 195–203.
16: 423–137. Tucker, M.M., Corace, R.G., Cleland, D.T., and Kashian,
Trappe, J.M. and Fogel, R.D. (1977). Ecosystematic D.M. (2016). Long-term effects of managing for an
function of mycorrhizae. In J.K. Marshall (ed.), endangered songbird on the heterogeneity of a
The Belowground Ecosystem: A Synthesis of Plant- fire-prone landscape. Landsc. Ecol. 31: 2445–2458.
Associated Processes. Range Sci. Dep. Sci. Ser. No. Tukey, H.B. (1970). The leaching of substances from
26. Colorado State Univ., Ft. Collins. plants. Ann. Rev. Plant Physiology 21: 305–324.
Trense, D. and Tietze, D.T. (2018). Studying speciation: Turetsky, M.R., Abbott, B.W., Jones, M.C. et al. (2019).
genomic essentials and approaches. In: Bird Species Permafrost collapse is accelerating carbon release.
(ed. D. Tietze), 39–61. New York: Springer. Nature 569: 32–34.
Treseder, K.K., Mack, M.C., and Cross, A. (2004). Rela- Turnbull, C.G. (ed.) (2005). Plant Architecture and its
tionships among fires, fungi, and soil dynamics in Manipulation. Oxford, UK: Blackwell 336 pp.
Alaskan boreal forests. Ecol. Appl. 14: 1826–1838. Turner, M.G. (2010). Disturbance and landscape
Treshow, M. (1970). Environment and Plant Response. dynamics in a changing world. Ecology 91:
New York: McGraw-Hill 422 pp. 2833–2849.
Trewartha, G.T. (1968). An Introduction to Climate, 4e. Turner, M.G. and Chapin, F.S. III (2005). Causes and
New York: McGraw-Hill 408 pp. consequences of spatial heterogeneity in ecosystem
Trimble, G.R. Jr. and Weitzman, S. (1956). Site index function. In: Ecosytem Function in Heterogeneous
studies of upland oaks in the northern Appala- Landscapes (ed. G.M. Lovett, C.G. Jones, M.G.
chians. For. Sci. 2: 162–173. Turner and K.C. Weathers). New York: Springer.
Tritton, L.M. and Hornbeck, J.W. (1982). Biomass Turner, M.G. and Gardner, R.H. (ed.) (1991).
equations for major tree species of the Northeast. Quantitative Methods in Landscape Ecology: The
USDA For. Serv. Gen. Tech. Report NE-69. North- Analysis and Interpretation of Landscape Heteroge-
eastern For. Exp. Sta., Upper Darby, PA. 52 pp. neity. New York: Springer-Verlag 536 pp.
Troeger, R. (1960). Kiefernprovenienzversuche. I. Teil. Turner, M.G. and Gardner, R.H. (2015). Landscape
Der grosse Kiefernprovenienzversuch im südwürt- Ecology in Theory and Practice, 2e. New York:
tembergischen Forstbezirk Schussenried. AFJZ 131 Springer-Verlag 482 pp.
(3–4): 49–59. Turner, M.G. and Romme, W.H. (1994). Landscape
Trofymow, J.A., Addison, J., Blackwell, B.A. et al. dynamics in crown fire ecosystems. Landsc. Ecol.
(2003). Attributes and indicators of old growth 9: 59–77.
and successional Douglas-fir forests on Vancouver Turner, M.G., Gardner, R.H., Dale, V.H., and O’Neill,
Island. Environ. Rev. 11: S187–S204. R.V. (1989). Predicting the spread of disturbance
Troll, C. (1939). Luftbildplan and ökologische Boden- across heterogeneous landscapes. Oikos 55:
forschung, 241–298. Berlin: Zeitschrift der Gesell- 121–129.
schaft für Erdkunde. Turner, M.G., Hargrove, W.W., Gardner, R.H., and
Troll, C. (1963a). Landscape ecology and land Romme, W.H. (1994). Effects of fire on landscape
development with special reference to the tropics. heterogeneity in Yellowstone National Park, Wyo-
J. Trop. Geogr. 17: 1–11. ming. J. Veg. Sci. 5: 731–742.
References 721
Turner, J.R.G., Lennon, J.J., and Greenwood, J.J.D. http://www.wcc.nrcs.usda.gov/snow/index.html.

(1996). Does climate cause the global biodiversity Accessed 09-21-2021.
gradient? In: Aspects of the Genesis and Mainte- U. S. Global Change Research Program (2017). Climate
nance of Biological Diversity (ed. M.E. Hochberg, science special report: Fourth National Climate
J. Clobert and R. Barbault), 199–220. Oxford, UK: Assessment, Volume I. https://science2017.global-
Oxford University Press. change.gov. Accessed 9/19/2021.
Turner, M.G., Romme, W.H., Gardner, R.H., and Har- U. S. National Vegetation Classification (2019). United
grove, W.W. (1997). Effects of fire size and pattern States National Vegetation Classification Database,
on early succession in Yellowstone National Park. V2.03. Federal Geographic Data Committee, Veg-
Ecol. Monogr. 67: 411–433. etation Subcommittee, Washington DC. [http://
Turner, M.G., Romme, W.H., and Gardner, R.H. (1999). usnvc.org] (accessed 05-09-2021).
Prefire heterogeneity, fire severity and plant rees- Uehara, G. and Gillman, G. (1981). The Mineralogy,
tablishment in subalpine forests of Yellowstone Chemistry and Physics of Tropical Soils with Var-
National Park, Wyoming. Int. J. Wildland Fire 9: iable Charge Clays. Boulder, CO: Westview Press
21–36. 170 pp.
Turner, M.G., Romme, W.H., and Tinker, D.B. (2003a). Uemura, M., Tominaga, Y., Nakagawara, C. et al. (2006).
Surprises and lessons from the 1988 Yellowstone Responses of the plasma membrane to low temper-
fires. Front. Ecol. Environ. 1: 351–358. atures. Physiol. Plant. 126: 81–89.
Turner, M.G., Romme, W.H., and Reed, R.A. (2003b). United Nations Environment Programme-World
Postfire aspen seedling recruitment across the Conservation Monitoring Centre, Internationla
Yellowstone (USA) landscape. Landsc. Ecol. 18: Union for Conservation of Nature, National Geo-
127–140. graphic Society. 2020. Protected Planet Live Report
Turner, M.G., Tinker, D.B., Romme, W.H. et al. (2004). 2020. UNEP-WCMC, IUCN and NGS: Cambridge
Landscape patterns of sapling density, leaf area, UK: Gland, Switzerland and Washington, DC, USA.
and aboveground net primary production in post- https://livereport.protectedplanet.net. Accessed
fire lodgepole pine forests, Yellowstone National 07-28-2021.
Park (USA). Ecosystems 7: 751–775. Urban, M.C. (2015). Accelerating extinction risk from
Turner, M.G., Donato, D.C., and Romme, W.H. (2013). climate change. Science 348: 571–573.
Consequences of spatial heterogeneity for eco- Urgenson, L.S., Halpern, C.B., and Anderson, P.D.
system services in changing forest landscapes: (2013). Twelve-year responses of planted and nat-
priorities for future research. Landsc. Ecol. 28: urally regenerating conifers to variable-retention
1081–1097. harvest in the Pacific Northwest, USA. Can. J. For.
Turner, M.G., Whitby, T.G., Tinker, D.B., and Romme, Res. 43: 46–55.
W.H. (2016). Twenty-four years after the Yel- Usbeck, T., Wohlgemuth, T., Dobbertin, M. et al. (2010).
lowstone Fires: are postfire lodgepole pine stands Increasing storm damage to forests in Switzerland
converging in structure and function? Ecology 97: from 1858 to 2007. Agric. For. Meteorol. 150: 47–55.
1260–1273. Ustin, S.L., Woodward, R.A., and Barbour, M.G. (1984).
Turner, M.G., Braziunas, K.H., Hansen, W.D., and Har- Relationships between sunfleck dynamics and red
vey, B.J. (2019). Short-interval severe fire erodes the fir seedling distribution. Ecology 65: 1420–1428.
resilience of subalpine lodgepole pine forests. Proc. Vale, T.R. (1982). Plants and People: Vegetation Change
Natl. Acad. Sci. 116: 11319–11328. in North America. Washington, DC: Associatio of
Turreson, G. (1922a). The species and the variety as eco- American Geographers 88 pp.
logical units. Hereditas 3: 100–113. Valladares, F. and Niinemets, Ü. (2008). Shade toler-
Turreson, G. (1922b). The genotypical response of the ance, a key plant feature of complex nature and
plant species to the habitat. Hereditas 3: 211–350. consequences. Annu. Rev. Ecol. Evol. Syst. 39:
Turreson, G. (1923). The scope and import of genecol- 237–257.
ogy. Hereditas 4: 171–176. Van Auken, O.W. (2009). Causes and consequences of
U. S. Department of Agriculture (1974). Seeds of woody plant encroachment into western North
Woody Plants in the United States. USDA For. American grasslands. Journal of environmental
Serv. Agr. Handbook No. 450. Washington, D.C. management 90 (10): 2931–2942.
883 pp. Van Buskirk, E.K., Reddy, A.K., Nagatani, A., and Chen,
U. S. Department of Agriculture (1990). Silvics of North M. (2014). Photobody localization of phytochrome
America. USDA For. Serv. Agr. Handbook 654. B is tightly correlated with prolonged and light-
Washington, D.C. Vol. 1, Conifers, 675 pp; Vol. 2, dependent inhibition of hypocotyl elongation in
Hardwoods, 877 pp. the dark. Plant Physiol. 165: 595–607.
U. S. Department of Agriculture Natural Resources Van Cleve, K., Oliver, L., Schlentner, R. et al. (1983).
Conservation Service (2020). Snow telemetry Productivity and nutrient cycling in taiga forests.
(SNOTEL) and snow course data and products. Can. J. For. Res. 13: 747–766.
722 References
Van Cleve, K., Chapin, F.S. III, Flanagan, P.W. et al. efficient seed harvest and dispersal. Ecol. Monogr.
(ed.) (1986). Forest Ecosystems in the Alaskan Taiga. 47: 89–111.
New York: Springer-Verlag 230 pp. Vanderwel, M.C., Coomes, D.A., and Purves, D.W.
Van de Peer, Y., Mizrachi, E., and Marchal, K. (2017). (2013). Quantifying variation in forest disturbance,
The evolutionary significance of polyploidy. Nat. and its effects on aboveground biomass dynamics,
Rev. Genet. 18: 411–424. across the eastern United States. Glob. Chang. Biol.
Van den Berg, L.J.L., Bullock, J.M., Clarke, R.T. et al. 19: 1504–1517.
(2001). Territory selection by the Dartford warbler Vankat, J.L. (1979). The Natural Vegetation of North
(Sylvia undata) in Dorset, England: the role of veg- America. New York: Wiley 486 pp.
etation type, habitat fragmentation and population Varela, E., Górriz-Mifsud, E., Ruiz-Mirazo, J., and
size. Biol. Conserv. 101: 217–228. López-i-Gelats, F. (2018). Payment for targeted
Van der Heijden, M.G., Martin, F.M., Selosse, M.A., and grazing: integrating local shepherds into wildfire
Sanders, I.R. (2015). Mycorrhizal ecology and evo- prevention. Forests 9: 464.
lution: the past, the present, and the future. New Varnell, L.M. (1998). The relationship between inun-
Phytol. 205: 1406–1423. dation history and baldcypress stem form in a
Van der Maarel, E. (1975). The Braun-Blanquet Virginia floodplain swamp. Wetlands 18: 176–183.
approach in perspective. Vegetatio 30: 213–219. Vašutová, M., Mleczko, P., López-García, A. et al.
Van der Pijl, L. (1957). The dispersal of plants by bats. (2019). Taxi drivers: the role of animals in trans-
Acta Bot. Neerl. 6: 291–315. porting mycorrhizal fungi. Mycorrhiza 29: 413–434.
Van der Pijl, L. (1972). Principles of Dispersal in Higher Veatch, J.O. (1953). Soils and Land of Michigan. East
Plants. New York: Springer-Verlag 162 pp. Lansing: Michigan State College Press 241 pp.
Van Dyke, F. and Lamb, R.L. (2020). Conservation Veatch, J.O. (1959). Presettlement forest in Michigan.
Biology: Foundations, Concepts, Applications, 3e. Michigan State Univ., Dept. Res Development. East
New York: Springer 644 pp. Lansing. Map.
Van Kleunen, M., Weber, E., and Fischer, M. (2010). Veblen, T.T. (1987). Trees of the trembling earth. Nat.
A meta-analysis of trait differences between inva- Hist. 96: 43–46.
sive and non-invasive plant species. Ecol. Lett. 13: Veblen, T.T. (1992). Regeneration dynamics. In:
235–245. Plant Succession, Theory and Prediction (ed. D.C.
Van Mantgem, P.J., Stephenson, N.L., Byrne, J.C. et al. Glenn-Lewin, R.K. Peet and T.T. Veblen), 152–187.
(2009). Widespread increase of tree mortality London: Chapman & Hall.
rates in the western United States. Science 323: Veblen, T.T. (2000). Disturbance patterns in southern
521–524. Rocky Mountain forests. Forest fragmentation in
Van Ommen Kloeke, A.E.E., Douma, J.C., Ordoñez, J.C. the southern Rocky Mountains. University Press of
et al. (2012). Global quantification of contrasting Colorado, Boulder, Colorado, USA, pp.31-54.
leaf life span strategies for deciduous and evergreen Veblen, T.T. and Lorenz, D.C. (1986). Anthropogenic
species in response to environmental conditions. disturbance and recovery patterns in montane for-
Global Ecol. Biogeogr. 21: 224–235. ests, Colorado Front Range. Phys. Geogr. 7: 1–24.
Van Veen, J.A., Ladd, J.N., and Frissel, M.J. (1984). Veblen, T.T. and Lorenz, D.C. (1988). Recent vegeta-
Modelling C & N turnover through the microbial tion changes along the forest/steppe ecotone of
biomass in soil. Plant Soil 76: 257–274. northern Patagonia. Annals Assoc. Amer. Geogra-
Van Wagner, C.E. (1970). Fire and red pine. In Proc. phers 78: 93–111.
Annual Tall Timbers Fire Ecology Conf. 10:211–219. Veblen, T.T. and Lorenz, D.C. (1991). The Colorado
Tall Timbers Res. Sta., Tallahassee, FL. Front Range, A Century of Ecological Change. Salt
Van Wagner, C.E. (1983). Fire behaviour in northern Lake City: Univ. Utah Press 210 pp.
conifer forests and shrublands. In: The Role of Fire Veblen, T.T., Hadley, K.S., and Reid, M.S. (1991a).
in Northern Circumpolar Ecosystems (ed. R.W. Wein Disturbance and stand development of a Colorado
and D.A. MacLean), 65–80. New York: Wiley. subalpine forest. J. Biogeogr. 18: 707–716.
Van Zandt, P.A. (2007). Plant defense, growth, and hab- Veblen, T.T., Hadley, K.S., Reid, M.S., and Rebertus, A.J.
itat: a comparative assessment of constitutive and (1991b). The response of subalpine forests to spruce
induced resistance. Ecology 88: 1984–1993. beetle outbreak in Colorado. Ecology 72: 213–231.
Vander Wall, S.B. (1990). Food Hoarding in Animals. Veblen, T.T., Hadley, K.S., Nel, E.M. et al. (1994). Dis-
Chicago, IL, USA: University of Chicago Press 453 turbance regime and disturbance interactions in
pp. a Rocky Mountain subalpine forest. J. Ecol. 82:
Vander Wall, S.B. (2010). How plants manipulate the 125–135.
scatter-hoarding behaviour of seed-dispersing ani- Veblen, T.T., Kitzberger, T., and Donnegan, J. (2000).
mals. Philos. Trans. R. Soc. B 365: 989–997. Climatic and human influences on fire regimes
Vander Wall, S.B. and Balda, R.P. (1977). Co-adaptations in ponderosa pine forests in the Colorado Front
of the Clark’s nutcracker and the piñon pine for Range. Ecol. Appl. 10: 1178–1195.
References 723
Vegis, A. (1964). Dormancy in higher plants. Annu. Rev. Viro, P.J. (1974). Effects of forest fire on soil. In: Fire
Plant Physiol. 15: 185–224. and Ecosystems (ed. T.T. Kozlowski and C.E. Ahl-
Venette, R.C., Gordon, D.R., Juzwik, J. et al. (2021). gren), 7–45. New York: Academic Press.
Early intervention strategies for invasive species Vitasse, Y., Lenz, A., and Körner, C. (2014). The interac-
management: connections between assessment, tion between freezing tolerance and phenology in
prevention efforts, eradication, and other rapid temperate deciduous trees. Front. Plant Sci. 5: 541.
responses. In: Invasive Species in Forests and Vitousek, P.M. (1982). Nutrient cycling and nutrient use
Rangelands of the United States (ed. T.M. Poland, efficiency. Am. Nat. 119: 553–572.
T. Patel-Weynand, D.M. Finch, et al.), 111–131. Vitousek, P.M. (1984). Litterfall, nutrient cycling, and
New York: Springer. nutrient limitation in tropical forests. Ecology 65:
Venterea, R.T., Groffman, P.M., Verchot, L.V. et al. 285–298.
(2003). Nitrogen oxide gas emissions from Vitousek, P.M. and Matson, P.A. (1984). Mechanisms
temperate forest soils receiving long-term nitrogen of nitrogen retention in forest ecosystems: a field
inputs. Glob. Chang. Biol. 9: 346–357. experiment. Science 225: 51–52.
Verrall, A.F. and Graham, T.W. (1935). The transmis- Vitousek, P.M. and Melillo, J.M. (1979). Nitrate loss
sion of Ceratostomella ulmi through root grafts. from disturbed forests: patterns and mechanisms.
Phytopathology 25: 1039–1040. For. Sci. 25: 605–619.
Vézina, P.E. and Boulter, D.W.K. (1966). The spectral Vitousek, P.M. and Reiners, W.A. (1975). Ecosystem
composition of near ultraviolet and visible radi- succession and nutrient retention: a hypothesis.
ation beneath forest canopies. Can. J. Bot. 44: Bioscience 25: 376–381.
1267–1284. Vitousek, P.M. and Sanford, R.L. Jr. (1986). Nutrient
Viereck, L.A. (1970). Forest succession and soil cycling in moist tropical forests. Annu. Rev. Ecol.
development adjacent to the Chena River in Syst. 17: 137–167.
interior Alaska. Arct. Alp. Res. 2: 1–26. Vitousek, P.M., Gosz, J.R., Grier, C.C. et al. (1982). A
Viereck, L.A. (1973). Wildfire in the taiga of Alaska. comparative analysis of potential nitrification and
Quat. Res. 3: 465–495. nitrate mobility in forest ecosystems. Ecol. Monogr.
Viereck, L.A. (1982). Effects of fire and firelines on 52: 155–177.
active layer thickness and soil temperatures in Vitousek, P.M., Walker, L.R., Whiteaker, L.D., and
interior Alaska. In H. M. French (ed.), Proc. Fourth Mueller-Dombois, D. (1987). Biological invasion
Can. Permafrost Conference, Calgary. Natl. Res. by Myrica faya alters ecosystem development in
Council Can. Ottawa. Hawaii. Science 238: 802–804.
Viereck, L.A. (1983). The effects of fire in black spruce Vitousek, P.M., Mooney, H.A., Lubchenco, J., and
ecosystems of Alaska and northern Canada. In: Melillo, J.M. (1997). Human domination of Earth’s
The Role of Fire in Northern Circumpolar Ecosys- ecosystems. Science 277: 494–499.
tems (ed. R.W. Wein and D.A. MacLean), 210–220. Vitra, A., Lenz, A., and Vitasse, Y. (2017). Frost hard-
New York: Wiley. ening and dehardening potential in temperate trees
Viereck, L.A. and Foote, J.M. (1970). The status of from winter to budburst. New Phytol. 216: 113–123.
Populus balsamifera and P. trichocarpa in Alaska. Vitt, P., Havens, K., Kramer, A. et al. (2010). Assisted
Canad. Field-Nat. 84: 169–174. migration of plants: changes in latitudes, changes
Viereck, L. A. Dyrness, C. T., and Van Cleve, K. (1984). in attitudes. Biol. Conserv. 143: 18–27.
Potential use of the Alaska vegetation system as Vizcaíno-Palomar, N., Revuelta-Eugercios, B., Zavala,
an indicator of forest site productivity in interior M.A. et al. (2014). The role of population origin
Alaska. In M. Murray (ed.), Proc. Workshop, and microenvironment in seedling emergence and
Forest Classification at High Latitudes as an Aid to early survival in Mediterranean maritime pine
Regeneration. USDA For. Serv. Gen. Tech. Report (Pinus pinaster Aiton). PLoS One 9: e109132.
PNW-177. Pacific Northwest For. and Rge. Exp.
Vogelmann, J.E. (1995). Assessment of forest
Sta., Portland, OR.
fragmentation in southern New England using
Vierling, L.A. and Wessman, C.A. (2000). Photosyn- remote sensing and geographic information sys-
thetically active radiation heterogeneity within tems technology. Conserv. Biol. 9: 439–449.
a monodominant Congolese rain forest canopy.
Vogl, R.J. (1964). Vegetational history of the Crex
Agric. For. Meteorol. 103: 265–278.
Meadows, a prairie savanna in northwestern Wis-
Viers, S.D. (1980). The influence of fire in coast red- consin. Am. Midi. Nat. 72: 157–175.
wood forests. In M.A. Stokes and J.H. Dieterich
Vogl, R. J. (1970). Fire and the northern Wisconsin pine
(eds.), Proc. Fire History Workshop. USDA For Serv.
barrens. In Proc. Annual Tall Timbers Fire Ecology
Gen Tech. Report. RM-81, Rocky Mountain For.
Conf. 6:47–96. Tall Timbers Res. Sta., Tallahassee, FL.
and Rge. Exp. Sta., Fort Collins, CO.
Vogl, R.J. (1973). Ecology of knobcone pine in the
Vince-Prue, D. (1975). Photoperiodism in Plants.
Santa Ana Mountains, California. Ecol. Monogr. 43:
New York: McGraw-Hill 444 pp.
125–143.
724 References
Vogl, R.J. and Ryder, C. (1969). Effects of slash burning Wagner, W.H. Jr., Taylor, S.R., Grieve, G. et al. (1988).
on conifer reproduction in Montana’s Mission Simple-leaved ashes (Fraxinus: Oleaceae) in Michi-
Range. Northwest Sci. 43: 135–147. gan. Mich. Bot. 27: 119–134.
Vogl, R.J., Armstrong, W.P., White, K.L., and Cole, K.L. Wahlenberg, W.G. (1949). Forest succession in the
(1988). The closed-cone pines and cypresses. In southern Piedmont region. J. For. 47: 713–715.
M.G. Barbour and J. Major (eds), Terrestrial Vege- Waide, J.B., Caskey, W.H., Todd, R.L., and Boring, L.R.
tation of California. CA. Native Plant Soc, Special (1988). Changes in soil nitrogen pools and trans-
Publ. No. 9. Berkeley. formations following forest clearcutting. In: Forest
Vogt, K.A., Grier, C.C., Meier, C.E., and Keyes, M.R. Hydrology and Ecology at Coweeta. (Ecological
(1983). Organic matter and nutrient dynamics in Studies 66) (ed. W.T. Swank and D.A. Crossley Jr.),
forest floors in young and mature Abies amabilis 221–232. New York: Springer-Verlag.
stands in western Washington. Ecol. Monogr. 53: Waide, R.B., Willig, M.R., Steiner, C.F. et al. (1999). The
139–157. relationship between productivity and species rich-
Vogt, K.A., Grier, C.C., and Vogt, D.J. (1986). Produc- ness. Annu. Rev. Ecol. Syst. 30: 257–300.
tion, turnover, and nutrient dynamics of above- Wakeley, P.C. (1954). Planting the southern pines.
and belowground detritus in world forests. Adv. USDA For. Serv. Agr. Monogr. No. 18. Washington,
Ecol. Res. 15: 303–377. D. C. 233 pp.
Voigt, G.K. (1968). Variation in nutrient uptake by trees. Wakeley, P.D. and Marrero, J. (1958). Five-year intercept
In Forest Fertilization. Tennessee Valley Authority, as site index in southern pine plantations. J. For. 56:
Muscle Shoals, Ala. 332–336.
Von Arx, G., Dobbertin, M., and Rebetez, M. (2012). Waldrop, M.P. and Harden, J.W. (2008). Interactive
Spatio-temporal effects of forest canopy on under- effects of wildfire and permafrost on microbial
story microclimate in a long-term experiment in communities and soil processes in an Alas-
Switzerland. Agric. For. Meteorol. 166-7: 144–155. kan black spruce forest. Glob. Chang. Biol. 14:
Von Arx, G., Pannatier, E.G., Thimonier, A., and 2591–2602.
Rebetez, M. (2013). Microclimate in forests with Walker, L.R. and Chapin, F.S. III (1986). Physiological
varying leaf area index and soil moisture: potential controls over seedling growth in primary
implications for seeding establishment in a chang- succession on an Alaskan floodplain. Ecology 67:
ing climate. J. Ecol. 101: 1201–1213. 1508–1523.
Vose, J. M., Clinton, B. D., and Swank, W. T. (1994). Walker, L.R. and Chapin, F.S. III (1987). Interactions
Fire, drought, and forest management influences among processes controlling successional change.
on pine/hardwood ecosystems in the southern Oikos 50: 131–135.
Appalachians. In Proc. 12th Int. Conf. on Fire and Walker, L.R. and Wardle, D.A. (2014). Plant succession
Forest Meteorology, Jekyll Island, GA. Soc. Am. For., as an integrator of contrasting ecological time
Bethesda, MD. scales. Trends Ecol. Evol. 29: 504–510.
Vose, J.M., Clark, J.S., Luce, C.H., and Patel-Weynand, T. Walker, L.R., Zasada, J.C., and Chapin, F.S. III (1986).
(ed.) (2016). Effects of Drought on Forests and Range- The role of life history processes in primary
lands in the United States: A Comprehensive Science succession on an Alaskan floodplain. Ecology 67:
Synthesis. USDA For. Serv. Gen Tech. Rep. WO 93b. 1243–1253.
Washington, DC: Washington Office 289 pp. Walker, W.S., Barnes, B.V., and Kashian, D.M. (2003).
Voss, E.G. (1972). Michigan Flora. I. Gymnosperms and Landscape ecosystems of the Mack Lake burn,
Monocots. Cranbrook lnst. Sci. and Univ. Mich. northern Lower Michigan, and the occurrence of
Herbarium. Bloomfield Hills, MI. 488 pp. the Kirtland’s warbler. For. Sci. 49: 119–139.
Voss, E.G. and Crow, G.E. (1976). Across Michigan by Walker, L.R., Wardle, D.A., Bardgett, R.D., and Clark-
covered wagon: a botanical expedition in 1888. son, B.D. (2010). The use of chronosequences
Mich. Bot. 15: 3–70. in studies of ecological succession and soil
Wagner, F.H. (1969). Ecosystem concepts in fish and development. J. Ecol. 98: 725–736.
game management. In: The Ecosystem Concept Wall, D.H., Bradford, M.A., St. John, M.G. et al. (2008).
in Natural Resource Management (ed. G.M. Van Global decomposition experiment shows soil
Dyne), 259–307. New York: Academic Press. animal impacts on decomposition are climate-
Wagner, W.H. Jr. (1983). Reticulistics: the recogni- dependent. Glob. Chang. Biol. 14: 2661–2677.
tion of hybrids and their role in cladistics and Wallace, L.L. and Dunn, E.L. (1980). Comparative
classification. In: Advances in Cladistics, Vol. 2 (ed. photosynthesis of three gap phase successional tree
N.I. Platnick and V.A. Funk), 63–79. New York: species. Oecologia 45: 331–340.
Columbia Univ. Press. Walters, M.B. and Reich, P.B. (1996). Are shade toler-
Wagner, W.H. Jr. and Schoen, D.J. (1976). Shingle oak ance, survival, and growth linked? Low light and
(Quercus imbricaria) and its hybrids in Michigan. nitrogen effects on hardwood seedlings. Ecology 77:
Mich. Bot. 15: 141–155. 841–853.
References 725
Walters, M.B. and Reich, P.B. (1997). Growth of Acer Wareing, P.F. and Robinson, L.W. (1963). Juvenility
saccharum seedlings in deeply shaded problems in woody plants. Rep. For. Res. 125–127.
understories of northern Wisconsin: effects of Waring, R.H. (1989). Ecosystems: fluxes of matter and
nitrogen and water availability. Can. J. For. Res. energy. In: Ecological Concepts (ed. J.M. Cherrett),
27: 237–247. 17–41. Oxford: Blackwell.
Walters, M.B. and Reich, P.B. (2000a). Seed size, Waring, R.H. (1991). Responses of evergreen trees to
nitrogen supply, and growth rate affect tree seed- multiple stresses. In: Response of Plants to Multiple
ling survival in deep shade. Ecology 81: 1887–1901. Stresses (ed. H.A. Mooney, W.E. Winner and E.J.
Walters, M.B. and Reich, P.B. (2000b). Trade-offs in Pell), 371–390. New York: Academic Press.
low-light CO2 exchange: a component of variation Waring, R.H. and Pitman, G.B. (1985). Modifying
in shade tolerance among cold temperate tree seed- lodgepole pine stands to change susceptibility to
lings. Funct. Ecol. 14: 155–165. mountain pine beetle attack. Ecology 66: 889–897.
Walters, M.B., Kruger, E.L., and Reich, P.B. (1993). Waring, R.H. and Schlesinger, W.H. (1985). Forest
Growth, biomass distribution and CO2 exchange of Ecosystems: Concepts and Management. New York:
northern hardwood seedlings in high and low light: Academic Press 340 pp.
relationships with successional status and shade Waring, R.H., McDonald, A.J.S., Larsson, S. et al.
tolerance. Oecologia 94: 7–16. (1985). Differences in chemical composition of
Wan, S., Hui, D., and Luo, Y. (2001). Fire effects on plants grown at constant relative growth rates with
nitrogen pools and dynamics in terrestrial ecosys- stable mineral nutrition. Oecologia 66: 157–160.
tems: a meta-analysis. Ecol. Appl. 11: 1349–1365. Warming, E. (1909). Oecology of Plants, an Introduction
Wang, F. and Xu, Y.J. (2009). Hurricane Katrina- to the Study of Plant Communities. Oxford: Claren-
induced forest damage in relation to ecological don Press 422 pp.
factors at landscape scale. Environ. Monit. Assess. Watson, D.M. and Rawsthorne, J. (2013). Mistletoe
156: 491–507. specialist frugivores: latterday ‘Johnny Appleseeds’
Wang, W.J., He, H.S., Thompson, F.R. et al. (2017). or self-serving market gardeners? Oecologia 172:
Changes in forest biomass and tree species distri- 925–932.
bution under climate change in the northeastern Watt, A.S. (1925). On the ecology of British beech
United States. Landsc. Ecol. 32: 1399–1413. woods with special reference to their regeneration.
Wang, S., Zhang, Y., Ju, W. et al. (2020). Recent global Part II: the development and structure of beech
decline of CO2 fertilization effects on vegetation communities on the Sussex Downs. J. Ecol. 13:
photosynthesis. Science 370: 1295–1300. 27–73.
Wani, M.C., Taylor, H.L., Wall, M.E. et al. (1971). Plant Watt, A.S. (1947). Pattern and process in the plant
antitumor agents. VI. The isolation and struc- community. J. Ecol. 35: 1–22.
ture of taxol, a novel antileukemic and antitumor Watts, W.A. (1970). The full-glacial vegetation of north-
agent from Taxus brevifolia. J. Am. Chem. Soc. 93: western Georgia. Ecology 51: 17–33.
2325–2327. Watts, W.A. (1979). Late quaternary vegetation of
Ward, J.V., Tockner, K., Arscott, D.B., and Claret, C. central Appalachia and the New Jersey Coastal
(2002). Riverine landscape diversity. Freshw. Biol. Plain. Ecol. Monogr. 49: 427–469.
47: 517–539. Watts, W.A. and Stuiver, M. (1980). Late Wisconsin cli-
Wardle, P. (1968). Engelmann spruce (Picea engelman- mate of northern Florida and the origin of species-
nii Engel.) at its upper limits in the Front Range, rich deciduous forest. Science 210: 325–327.
Colorado. Ecology 49: 483–495. Waughman, G.J., French, R.J., and Jones, K. (1981).
Wardle, P. (1985). New Zealand timberlines. 3. A syn- Nitrogen fixation in some terrestrial environments.
thesis. New Zealand J. Bot. 23: 263–271. In: Nitrogen Fixation, Vol. I: Ecology (ed. W.J.
Wardle, D.A. (1992). A comparative assessment of the Broughton), 135–192. Oxford: Clarendon Press.
factors which influence microbial biomass carbon Way, D.A. and Oren, R. (2010). Differential responses to
and nitrogen levels in soil. Biol. Rev. 67: 321–358. changes in growth temperature between trees from
Wardle, D.A. and Peltzer, D.A. (2017). Impacts of different functional groups and biomes: a review
invasive biota in forest ecosystems in an above- and synthesis of data. Tree Physiol. 30: 669–688.
ground–belowground context. Biol. Invasions 19: Way, D.A. and Pearcy, R.W. (2012). Sunflecks in trees
3301–3316. and forests: from photosynthetic physiology to
Wareing, P.F. (1959). Problems of juvenility and global change biology. Tree Physiol. 32: 1066–1081.
flowering in trees. J. Linn. Soc. Lond. Bot. 56: Weaver, H. (1951). Fire as an ecological factor in
282–289. southwestern ponderosa pine forests. J. For. 49: 93–98.
Wareing, P.F. (1987). Phase change and vegetative Weaver, H. (1974). Effects of fire on temperate forest:
propagation. In: Improving Vegetatively Propagated western United States. In: Fire and Ecosystems
Crops (ed. A.J. Abbott and R.K. Atkin), 262–270. (ed. T.T. Kozlowski and C.E. Ahlgren), 279–319.
London: Academic Press. New York: Academic Press.
726 References
Weaver, J.E. and Clements, F.E. (1929). Plant Ecology. Wellner, C.A. (1989). Classification of habitat types in the
New York: McGraw-Hill 520 pp. western United States. In D.E. Ferguson,
Webb, S.L. (1986). Potential role of passenger pigeons P. Morgan, and F.D. Johnson (comps.),
and other vertebrates in the rapid Holocene migra- Proceedings—Land Classifications Based on Vegeta-
tions of nut trees. Quat. Res. 26: 367–375. tion: Applications for Resource Management. USDA
Webb, T. III, Cushing, E.J., and Wright, H.E. Jr. (1983). For. Serv. Gen. Tech. Report INT-257. Intermountain
Holocene changes in the vegetation of the Mid- Res. Sta., Ogden, UT.
west. In: Late-Quaternary Environments of the Wells, P.V. (1965). Scarp woodlands, transported grass-
United States, Vol. 2, The Holocene (ed. H.E. Wright land soils, and concept of grassland climate in the
Jr.), 142–165. Minneapolis: Univ. Minnesota Press. Great Plains region. Science 148: 246–249.
Webb, L.J., Tracey, J.G., and Williams, W.T. (1972). Wells, P.V. (1970). Postglacial vegetational history of the
Regeneration and pattern in the subtropical rain Great Plains. Science 167: 1574–1582.
forest. J. Ecol. 6: 675–695. Wells, G.L. (1992). The Aeolian Landscapes of North
Webb, T.I.I.I., Bartlein, P.J., Harrison, S.P., and Anderson, America for the Late Pleistocene. PhD Thesis, Uni-
K.H. (1993). Vegetation, lake levels, and climate in versity of Oxford, U.K. 256 pp.
eastern North America for the past 18,000 years. In: Wells, O.O. and Wakeley, P.C. (1966). Geographic varia-
Global Climates since the Last Glacial Maximum (ed. tion in survival, growth and fusiform rust infection
H.E. Wright Jr., J.E. Kutzbach, T. Webb III, et al.), of planted loblolly pine. For. Sci. Monogr. 11. 40 pp.
415–467. Minneapolis: Univ. Minnesota Press. Wells, O.O., Switzer, G.L., and Schmidtling, R.C. (1991).
Webster, C.R. and Lorimer, C.G. (2005). Minimum Geographic variation in Mississippi loblolly pine
opening sizes for canopy recruitment of midtoler- and sweetgum. Silvae Genet. 40: 105–119.
ant tree species: a retrospective approach. Ecol. Wendel, G.W. (1987). Abundance and distribution of
Appl. 15: 1245–1262. vegetation under four hardwood stands in north-
Webster, C.R., Nelson, K., and Wangen, S.R. (2005). Stand central West Virginia. USDA For. Serv. Res. Paper
dynamics of an insular population of an invasive NE-607. Northeastern For. Exp. Sta., Broomall,
tree, Acer platanoides. For. Ecol. Manag. 208: 85–99. PA. 6 pp.
Webster, J.R., Morkeski, K., Wojculewski, C.A. et al. Wertz, W.A. and Arnold, J.F. (1972). Land systems
(2012). Effects of hemlock mortality on streams in inventory. USDA For. Serv., Intermountain Region,
the southern Appalachian Mountains. Am. Midl. Ogden, UT. 12 pp.
Nat. 168: 112–131. Wertz, W.A. and Arnold, J.F. (1975). Land stratification
Wehner, M.F., Arnold, J.R., Knutson, T. et al. (2017). for land-use planning. In: Forest Soils and Forest
Droughts, floods, and wildfires. In: Climate Science Land Management (ed. B. Bernier and C.H. Win-
Special Report: Fourth National Climate Assessment, get), 617–629. Quebec: Laval Univ. Press.
Volume I (ed. D.J. Wuebbles, D.W. Fahey, K.A. Hib- West, D.C., Shugart, H.H., and Botkin, D.B. (1981).
bard, et al.), 231–256. Washington, DC: U.S. Global Introduction. In: Forest Succession: Concepts and
Change Research Program. Application (ed. D.C. West, H.H. Shugart and D.B.
Weidman, R.H. (1939). Evidences of racial influence Botkin). New York: Springer-Verlag.
in a 25–year test of ponderosa pine. J. Agric. Res. Westerling, A.L., Hidalgo, H.G., Cayan, D.R., and
59: 855–887. Swetnam, T.W. (2006). Warming and earlier spring
increase western US forest wildfire activity. Science
Weigelt, A., Mommer, L., Andraczek, K. et al. (2021). An
313: 940–943.
integrated framework for plant form and function: a
Westerling, A.L., Turner, M.G., Smithwick, E.A.H.
belowground perspective. New Phytol. 232: 42–59.
et al. (2011). Continued warming could trans-
Weil, R.R. and Brady, N.C. (2017). The Nature and Prop-
form Greater Yellowstone fire regimes by mid-21st
erties of Soils, 15e. Essex, England, UK: Pearson
century. Proc. Natl. Acad. Sci. 108: 13165–13170.
1104 pp.
Westhoff, V. and van der Maarel, E. (1973). The Braun-
Wein, R.W. and MacLean, D.A. (1983). The Role of Fire
Blanquet approach. In: Handbook of Vegetation
in Northern Circumpolar Ecosystems. New York:
Science, Part V, Ordination and Classification of
Wiley 322 pp.
Communities (ed. R. Tüxen), 287–399. The Hague:
Weiser, C.J. (1970). Cold resistance and injury in woody Junk.
plants. Science 169: 1269–1278.
Westman, W.E. (1990). Managing for biodiversity. Bio-
Welbourn, M.L., Stone, E.L., and Lassoie, J.P. (1981). science 40: 26–33.
Distribution of net litter inputs with respect to slope
Westoby, J. (1989). Introduction to World Forestry: People
position and wind direction. For. Sci. 27: 651–659.
and their Trees. Hoboken, NJ: Blackwell 240 pp.
Wellner, C.A. (1970). Fire history in the northern Rocky
Westphal, M.I., Browne, M., MacKinnon, K., and Noble,
Mountains. In Symp. Proc. The Role of Fire in the
I. (2008). The link between international trade and
Intermountain West. Intermountain Fire Research
the global distribution of invasive alien species.
Council, Univ. Montana, School of Forestry,
Biol. Invasions 10: 391–398.
Missoula.
References 727
Wharton, C.H., Kitchens, W.M., Pendleton, E.C., and and mosaics of genetic variability: the adaptive sig-
Sipe, T.W. (1982). The ecology of bottomland nificance of somatic mutations in plants. Oecologia
hardwood swamps of the Southeast: a community 49: 287–292.
profile. U. S. Fish and Wildlife Service, FWS/OBS- Whitham, T.G., Morrow, P.A., and Potts, B.M. (1991).
81/37. 133 pp. Conservation of hybrid plants. Science 254:
Wheeler, N.C. and Critchfield, W.B. (1985). The distri- 779–780.
bution and botanical characteristics of lodgepole Whitham, T.G., Morrow, P.A., and Potts, B.M. (1994).
pine: biogeographical and management impli- Plant hybrid zones as centers of biodiversity: the
cations. In D.M. Baumgartner, R.G. Krebill, J.T. herbivore community of two endemic Tasmanian
Arnott, and G.F. Weetman (eds.), Symp. Proc. eucalypts. Oecologia 97: 481–190.
Lodgepole Pine, The Species and Its Management. Whitlock, C. and Bartlein, P.J. (1997). Vegetation and
Washington State Univ., Pullman. climate change in northwest America during the
Whelan, R.J. (1995). The Ecology of Fire. Cambridge: past 125 kyr. Nature 388: 57–61.
Cambridge Univ. Press 346 pp. Whitney, G.G. (1994). From Coastal Wilderness to
Whitcomb, R.F., Robbins, C.S., Lynch, J.F. et al. (1981). Fruited Plain: A History of Environmental Change
Effects of forest fragmentation on avifauna of in Temperate North America from 1500 to the Pre-
the eastern deciduous forest. In: Forest Island sent. Cambridge: Cambridge Univ. Press 451 pp.
Dynamics in Man-Dominated Landscapes (ed. R.L. Whitney, H.E. and Johnson, W.C. (1984). Ice storms and
Burgess and D.M. Sharpe), 125–205. New York: forest succession in southwestern Virginia. Bull.
Springer-Verlag. Torrey Bot. Club 111: 429–437.
White, T.C.R. (1978). The importance of a relative Whitney, K., Randell, R.A., and Rieseberg, L. (2006).
shortage of food in animal ecology. Oecologia 33: Adaptive introgression of herbivore resistance
71–86. traits in the weedy sunflower Helianthus annuus.
White, P.S. (1979). Pattern, process and natural distur- Am. Nat. 167: 794–807.
bance in vegetation. Bot. Rev. 45: 229–299. Whittaker, R.H. (1953). A consideration of climax
White, A.S. (1983). The effects of thirteen years of theory: the climax as a population and pattern.
annual prescribed burning on a Quercus ellipsoidalis Ecol. Monogr. 23: 41–78.
community in Minnesota. Ecology 64: 1081–1085. Whittaker, R.H. (1956). Vegetation of the Great Smoky
White, J. (1985). The thinning rule and its application to Mountains. Ecol. Monogr. 26: 1–80.
mixtures of plant populations. In: Studies on Plant Whittaker, R.H. (1960). Vegetation of the Siskiyou
Demography (ed. J. White), 291–309. New York: Mountains, Oregon and California. Ecol. Monogr.
Academic Press. 30: 279–338.
White, A.S. (1986). Prescribed burning for oak savanna Whittaker, R.H. (1962). Classification of natural com-
restoration in central Minnesota. USDA For. Serv. munities. Bot. Rev. 28: 1–239.
Res. Paper NC-266. North Central For. Res. Sta., St. Whittaker, R.H. (1966). Forest dimensions and produc-
Paul, MN. 12 pp. tion in the Great Smoky Mountains. Ecology 47:
White, C.S. (1986). Volatile and water-soluble inhibi- 103–121.
tors of nitrogen mineralization and nitrification Whittaker, R.H. (1967). Gradient analysis of vegetation.
in a ponderosa pine ecosystem. Biol. Fertil. Soils 2: Biol. Rev. 42: 207–264.
97–104.
Whittaker, R.H. (1972). Evolution and measurement of
White, C.S. (1988). Nitrification inhibition by mono- species diversity. Taxon 21: 213–251.
terpeniods: theoretical mode of action based on
Whittaker, R.H. (1975). Communities and Ecosystems,
molecular structures. Ecology 69: 1631–1633.
2e. New York: MacMillan 385 pp.
White, P.S. and Jentsch, A. (2001). The search for gen-
Whittaker, R.H. (1977). Evolution of species diversity in
erality in studies of disturbance and ecosystem
land communities. In: Evolutionary Biology, vol. 10
dynamics. Prog. Bot. 62: 399–450. (ed. M.K. Hecht, W.C. Steere and B. Wallace), 1–67.
White, T.L., Adams, W.T., and Neale, D.B. (2007). Forest New York: Plenum.
Genetics. Cambridge, MA: CABI Publishing 704 pp. Whittet, R., Cavers, S., Cottrell, J., and Ennos, R. (2016).
Whitehead, D.R. (1981). Late-Pleistocene vegetational Seed sourcing for woodland creation in an era of
changes in northeastern North Carolina. Ecol. uncertainty: an analysis of the options for Great
Monogr. 51: 451–471. Britain. Forestry 90: 163–173.
Whitham, T.G. (1989). Plant hybrid zones as sinks for Wicken, E. B. (1986). Terrestrial ecozones of
pests. Science 244: 1490–1493. Canada. Environment Canada. Ecological Land
Whitham, T.G. and Mopper, S. (1985). Chronic her- Classification Series No. 19. Lands Directorate,
bivory: impacts on architecture and sex expression Ottawa. 26 pp.
of pinyon pine. Science 228: 1089–1091. Wicklow, D. (1975). Fire as an environmental cue initi-
Whitham, T.G. and Slobodchikoff, C.N. (1981). Evolu- ating ascomycete development in a tallgrass prairie.
tion by individuals, plant-herbivore interactions, Mycologia 67: 852–862.
728 References
Wickware, G., and Rubec, C.D.A. (1989). Ecoregions of Williams, J.W. and Jackson, S.T. (2007). Novel climates,
Ontario. Ecological Land Classification Series No. no-analog communities, and ecological surprises.
26, Env. Canada, Ottawa. 37 pp. Front. Ecol. Environ. 5: 475–482.
Wiens, J.A. (2005). Toward a unified landscape ecology. Williams, J.W., Shuman, B.N., Webb, T. III et al. (2004).
In: Issues and Perspectives in Landscape Ecology (ed. Late-Quaternary vegetation dynamics in North
J.A. Wiens and M.R. Moss), 365–373. Cambridge, America: scaling from taxa to biomes. Ecol. Monogr.
UK: Cambridge University Press. 74: 309–334.
Wiens, J.A. and Moss, M.R. (2005). Issues and Per- Williams, C.A., Collatz, G.J., Masek, J., and Goward,
spectives in Landscape Ecology. Cambridge, UK: S.N. (2012). Carbon consequences of forest dis-
Cambridge University Press 412 pp. turbance and recovery across the conterminous
Wiersma, J.H. (1962). Enkele quantitative aspecten United States. Glob. Biogeochem. Cycles 26: GB1005.
van het exotenvraagstuk. Ned. Bosb. Tijdschr. 34: Williams, C.A., Gu, H., MacLean, R. et al. (2016). Dis-
175–184. turbance and the carbon balance of US forests: a
Wiersma, J.H. (1963). A new method of dealing with quantitative review of impacts from harvests, fires,
results of provenance tests. Silvae Genet. 12: 200–205. insects, and droughts. Glob. Planet. Chang. 143:
Wiersum, K.F. (1995). 200 years of sustainability in 66–80.
forestry: lessons from history. Environ. Manag. 19: Williamson, M. (1996). Biological Invasions. New York:
321–329. Springer 256 pp.
Wignall, T.A., Browning, G., and Mackenzie, K.A.D. Williamson, G.B. and Black, E.M. (1981). High tem-
(1987). The physiology of epicormic bud emer- perature of forest fires under pines as a selective
gence in pedunculate oak (Quercus robur L.). advantage over oaks. Nature 293: 643–644.
Responses to partial notch girdling in thinned and Willmer, P. (2011). Pollination and Floral Ecology. Princ-
unthinned stands. Forestry 60: 45–56. eton, NJ: Princeton University Press 832 pp.
Wilcove, D.S. (1989). Protecting biodiversity in Wilmers, C.C., Post, E., Peterson, R.O., and Vucetich,
multiple-use lands: lessons from the US Forest Ser- J.A. (2006). Predator disease out-break modulates
vice. Trends Ecol. Evol. 4: 385–388. top-down, bottom-up and climatic effects on herbi-
Wilcove, D.S., McLellan, C.H., and Dobson, A.P. (1986). vore population dynamics. Ecol. Lett. 9: 383–389.
Habitat fragmentation in the temperate zone. In: Wilson, B.F. (1984). The Growing Tree. Amherst: Univ.
Conservation Biology, the Science of Scarcity and Massachusetts Press 138 pp.
Diversity (ed. M.E. Soulé), 237–256. Sunderland, Wilson, E.O. (1992). The Diversity of Life. Harvard Uni-
MA: Sinauer Associates. versity, Cambridge, MA: Belknap Press 432 pp.
Wilcove, D.S., Rothstein, D., Dubow, J. et al. (1998). Wilson, B.F. (2000). Apical control of branch growth
Quantifying threats to imperiled species in the and angle in woody plants. Am. J. Bot. 87: 601–607.
United States. Bioscience 48: 607–615. Wilson, E.O. (2010). The Diversity of Life, 2e. Cam-
Wilde, S.A. (1958). Forest Soils: Their Properties and Rela- bridge, MA: Belknap Press, Harvard University
tion to Silviculture. New York: Ronald Press 537 pp. 440 pp.
Wilkin, K., Ackerly, D., and Stephens, S. (2016). Climate Wilson, M.V. and Shmida, A. (1984). Measuring beta
change refugia, fire ecology and management. diversity with presence-absence data. J. Ecol. 72:
Forests 7: 77. 1055–1064.
Will, R.E., Wilson, S.M., Zou, C.B., and Hennessey, Wilson, C.M., Schaeffer, R.N., Hickin, M.L. et al. (2018).
T.C. (2013). Increased vapor pressure deficit due to Chronic impacts of invasive herbivores on a foun-
higher temperature leads to greater transpiration dational forest species: a whole-tree perspective.
and faster mortality during drought for tree seed- Ecology 99: 1783–1791.
lings common to the forest–grassland ecotone. New Wingfield, M.J., Garnas, J.R., Hajek, A. et al. (2016).
Phytol. 200: 366–374. Novel and co-evolved associations between insects
Williams, A.P. and Abatzoglou, J.T. (2016). Recent and microorganisms as drivers of forest pestilence.
advances and remaining uncertainties in resolving Biol. Invasions 18: 1045–1056.
past and future climate effects on global fire Wistendahl, W.A. (1958). The flood plain of the Raritan
activity. Curr. Clim. Change Rep. 2: 1–14. River, New Jersey. Ecol. Monogr. 28: 129–153.
Williams, M.A. and Baker, W.L. (2012). Spatially exten- With, K.A. (2019). Essenitals of Landscape Ecology.
sive reconstructions show variable-severity fire Oxford, UK: Oxford University Press 656 pp.
and heterogeneous structure in historical western With, K.A., Cadaret, S.J., and Davis, C. (1999).
United States dry forests. Glob. Ecol. Biogeogr. 21: Movement responses to patch structure in experi-
1042–1052. mental fractal landscapes. Ecology 80: 1340–1353.
Williams, M.I. and Dumroese, R.K. (2013). Preparing Witter, J.A. and Waisanen, L.A. (1978). The effect of
for climate change: forestry and assisted migration. differential flushing times among trembling aspen
J. For. 111: 287–229. clones on tortricid caterpillar populations. Environ.
Williams, G.J. III and McMillan, C. (1971). Phenology Entomol. 7: 139–143.
of six United States provenances of Liquidamhar Wofsy, S.P., Goulden, M.L., Munger, J.W. et al. (1993).
stryraciflua under controlled conditions. Am. J. Bot. Net exchange of CO2 in a mid-latitude forest.
58: 24–31. Science 260: 1314–1317.
References 729
Wohl, E. (2013). Fluvial geomorphology. In: Treatise on Worrall, J.J., Egeland, L., Eager, T. et al. (2008). Rapid
Geomorphology, vol. 9 (ed. J. Schroder). San Diego, mortality of Populus tremuloides in southwestern
CA: Academic Press 860 pp. Colorado, USA. For. Ecol. Manag. 255: 686–696.
Wolda, H. (1981). Similarity indices, sample size and Worrall, J.J., Rehfeldt, G.E., Hamann, A. et al. (2013).
diversity. Oecologia 50: 296–302. Recent declines of Populus tremuloides in North
Wolf, M. and Weissing, F.J. (2012). Animal person- America linked to climate. For. Ecol. Manag. 299:
alities: consequences for ecology and evolution. 35–51.
Trends Ecol. Evol. 27: 452–461. Worster, D. (1994). Nature’s Economy, 2e. Cambridge:
Wolfe, J., Bryant, G., and Koster, K.L. (2002). What is Cambridge Univ. Press 505 pp.
‘unfreezable water’, how unfreezable is it and how Wright, J.W (1970). Genetics of eastern white pine
much is there? CryoLetters 23: 157–166. (Pinus strobus L.). USDA For. Serv. Res. Paper
Wood, D.M. and del Moral, R. (1987). Mechanisms of WO-9. Washington, D.C. 16 pp.
early primary succession in subalpine habitats on Wright, J.W. (1976). Introduction to Forest Genetics.
Mount St. Helens. Ecology 68: 780–790. New York: Academic Press 463 pp.
Wood, T., Bormann, F.H., and Voigt, G.T. (1984). Phos- Wright, H.E. Jr. (1971). Late quaternary vegetational
phorus cycling in a northern hardwood forest: history of North America. In: The Late Cenozoic
biological and chemical control. Science 223: 391–393. Glacial Ages (ed. K.K. Turekian), 425–464. New
Wood, T.E., Takebayashi, N., Barker, M.S. et al. (2009). Haven, CT: Yale Univ. Press.
The frequency of polyploid speciation in vascular Wright, H.E. Jr. and Heinselman, M.L. (ed.) (1973). The
plants. Proc. Natl. Acad. Sci. 106: 13875–13879. ecological role of fire in natural conifer forests of
Woodall, C.W., Oswalt, C.M., Westfall, J.A. et al. western and northern North America. Quat. Res. 3:
(2009). An indicator of tree migration in forests of 317–513.
the eastern United States. For. Ecol. Manag. 257: Wright, J.W., Pauley, S.S., Polk, R.B. et al. (1966).
1434–1444. Performance of Scotch pine varieties in the North
Woodall, C.W., Zhu, K., Westfall, J.A. et al. (2013). Central Region. Silvae Genet. 15: 101–110.
Assessing the stability of tree ranges and influence Wright, J.W., Wilson, L.F., and Randall, W. (1967). Dif-
of disturbance in eastern US forests. For. Ecol. ferences among Scotch pine varieties in suscepti-
Manag. 291: 172–180. bility to European pine sawfly. For. Sci. 13: 175–181.
Woodall, C.W., Westfall, J.A., D’Amato, A.W. et al. Wright, R.A., Wein, R.W., and Dancik, B.P. (1992).
(2018). Decadal changes in tree range stability Population differentiation in seedling root size
across forests of the eastern US. For. Ecol. Manag. between adjacent stands of jack pine. For. Sci. 38:
429: 503–510. 777–785.
Woodcock, D. and Shier, A. (2002). Wood specific Wright, D.H., Curie, D.J., and Maurer, B.A. (1993).
gravity and its radial variations: the many ways to Energy supply and patterns of species richness
make a tree. Trees 16: 437–443. on local and regional scales. In: Species Diver-
Woods, F.W. (1953). Disease as a factor in the evolution sity in Ecological Communities: Historical and
of forest composition. J. For. 51: 871–873. Geographical Perspectives (ed. R.E. Ricklefs and
Woods, F.W. (1957). Factors limiting root penetration D. Schluter), 66–74. Chicago, IL: University of
in deep sands of the southeastern coastal plain. Chicago Press.
Ecology 38: 357–359. Wright, I.J., Recih, P.B., Westoby, M. et al. (2004).
Woods, K.D. (1979). Reciprocal replacement and the The worldwide leaf economics spectrum. Nature
maintenance of codominance in a beech-maple 428: 821.
forest. Oikos 33: 31–39. Wu, J. (2013). Landscape sustainability science: eco-
Woods, K.D. (1984). Patterns of tree replacement: system services and human well-being in changing
canopy effects on understory pattern in hemlock- landscapes. Landsc. Ecol. 28: 999–1023.
northern hardwood forests. Vegetatio 56: 87–107. Wu, J. and Hobbs, R.J. (ed.) (2007). Key Topics in
Woods, F.W. and Shanks, R.E. (1959). Natural Landscape Ecology. Cambridge, UK: Cambridge
replacement of chestnut by other species in he University Press 314 pp.
Great Smoky Mountains National Park. Ecology 40: Wuerthner, G. (1988). Yellowstone and the Fires of
349–361. Change. Salt Lake City, UT: Haggis House Publ.
Woodward, F.I. (1987). Climate and Plant Distribution. Yaffee, S.L. (1994). The Wisdom of the Spotted Owl.
Cambridge: Cambridge Univ. Press 174 pp. Washington, D.C.: Island Press 430 pp.
Woodwell, G. M., and D. B. Botkin. (1970). Metabolism Yahner, R.H. (1988). Changes in wildlife communities
of terrestrial ecosystems by gas exchange tech- near edges. Conserv. Biol. 2: 333–339.
niques: The Brookhaven approach. In D. Reichle Yahner, R.H. (1995). Eastern Deciduous Forest, Ecology
(ed.), Analysis of Temperate Forest Ecosystems. (Eco- and Wildlife Conservation. Minneapolis: Univ. Min-
logical Studies, Vol. 1) Springer-Verlag, New York. nesota Press 220 pp.
Worrall, J. (1983). Temperature–bud-burst relationships Yakovlev, I.A., Fossdal, C.G., and Johnsen, O. (2010).
in amabilis and subalpine fir provenance tests MicroRNAs, the epigenetic memory and climatic
replicated at different elevations. Silvae Genet. 32: adaptation in Norway spruce. New Phytol. 187:
203–209. 1154–1169.
730 References
Yakovlev, I., Fossdal, C., Skrøppa, T. et al. (2012). An Zhang, D., Hui, D., Luo, Y., and Zhou, G. (2008). Rates
adaptive epigenetic memory in conifers with of litter decomposition in terrestrial ecosystems:
important implications for seed production. Seed global patterns and controlling factors. J. Plant
Sci. Res. 22: 63–76. Ecol. 1: 85–93.
Yeboah, D. and Chen, H.Y.H. (2016). Diversity– Zhang, Y., Chen, H.Y.H., and Reich, P.B. (2012). Forest
disturbance relationship in forest landscapes. productivity increases with evenness, species rich-
Landsc. Ecol. 31: 981–987. ness and trait variation: a global meta-analysis.
Yocom, H.A. (1968). Shortleaf pine seed dispersal. J. Ecol. 100: 742–749.
J. For. 66: 422. Zheng, Z., Li, Y., Li, M. et al. (2021). Whole-genome
Youngberg, C.T. and Wollum, A.G. (1976). Nitrogen diversification analysis of the hornbeam species
accretion in developing Ceonothus velutinus stands. reveals speciation and adaptation among closely
Soil Sci. Soc. Am. Proc. 40: 109–112. related species. Front. Plant Sci. 12: 1–10.
Youngblood, A.P. and Mauk, R.L. (1985). Coniferous Zhu, K., Woodall, C.W., and Clark, J.S. (2012). Failure to
forest habitat types of central and southern Utah. migrate: lack of tree range expansion in response to
USDA For. Serv. Gen. Tech. Report INT-187. Inter- climate change. Glob. Chang. Biol. 18: 1042–1052.
mountain For. Exp. Sta., Ogden, UT. 89 pp. Zhu, K., Woodall, C.W., Ghosh, S. et al. (2014). Dual
Zachariassen, K.E. and Kristiansen, E. (2000). Ice impacts of climate change: forest migration and
nucleation and antinuclea tion in nature. Cryobi- turnover through life history. Glob. Chang. Biol. 20:
ology 41: 257–279. 251–264.
Zahner, R. (1958). Site-quality relationships of pine for- Zimmermann, M.H. (1971). Transport in the xylem.
ests in southern Arkansas and northern Louisiana. In: Trees: Structure and Function (ed. M.H. Zim-
For. Sci. 4: 162–176. mermann and C.L. Brown), 169–220. New York:
Zahner, R. (1968). Water deficits and growth of trees. Springer-Verlag.
In: Water Deficits and Plant Growth. II (ed. T.T. Zimmermann, M.H. and Brown, C.L. (1971). Trees:
Kozlowski), 191–254. New York: Academic Press. Structure and Function. New York: Springer-Verlag
Zahner, R. and Crawford, N.A. (1965). The clonal con- 336 pp.
cept in aspen site relations. In: Tree Growth and Zobel, D.B. (1969). Factors affecting the distribution
Forest Soils (ed. C.T. Youngberg and C.B. Davey), of Pinus pungens, an Appalachian endemic. Ecol.
230–243. Corvallis: Oregon State Univ. Press. Monogr. 39: 303–333.
Zahner, R. and Stage, A.R. (1966). A procedure for calcu- Zobel, D.B. and Antos, J.A. (1992). Survival of plants
lating daily moisture stress and its utility in regres- buried for eight growing seasons by volcanic
sions of tree growth on weather. Ecology 47: 64–74. tephra. Ecology 73: 698–701.
Zak, D.R. and Pregitzer, K.S. (1990). Spatial and Zobel, B. and Talbert, J. (1984). Applied Forest Tree
temporal variability of nitrogen cycling in northern Improvement. New York: Wiley 505 pp.
Lower Michigan. For. Sci. 36: 367–380. Zogg, G.P. and Barnes, B.V. (1995). Ecological
Zak, D.R., Pregitzer, K.S., and Host, G.E. (1986). classification and analysis of wetland ecosystems,
Landscape variation in nitrogen mineralization and northern lower Michigan. Can. J. For. Res. 25:
nitrification. Can. J. For. Res. 16: 1258–1263. 1865–1875.
Zak, D.R., Host, G.E., and Pregitzer, K.S. (1989). Zohner, C.M., Benito, B.M., Svenning, J.-C., et al.
Regional variability in nitrogen mineralization, (2016). Day length unlikely to constrain climate-
nitrification, and overstory biomass in northern driven shifts in leaf-out times of northern woody
Lower Michigan. Can. J. For. Res. 19: 1521–1526. plants. Nat. Clim. Chang. 6: 1120–1123.
Zak, D.R., Groffman, P.M., Pregitzer, K.S. et al. (1990). Zohner, C.M., Mo, L., and Renner, S.S. (2018). Global
The vernal dam: plant-microbe competition for warming reduces leaf-out and flowering synchrony
nitrogen in northern hardwood forests. Ecology 71: among individuals. elife 7: e40214.
651–656. Zou, X., Theiss, C., and Barnes, B.V. (1992). Pattern
Zak, D.R., Tilman, D., Parmenter, R.R. et al. (1994). of Kirtland’s warbler occurrence in relation to
Plant production and soil microorgansims in late- the landscape structure of its summer habitat
successional ecosystems: a continental-scale study. in northern Lower Michigan. Landsc. Ecol. 6:
Ecology 75: 2333–2347. 221–231.
Zasada, J. (1985). Production, dispersal, and germina- Zverev, V., Zvereva, E.L., and Kozlov, M.V. (2017). Onto-
tion, and first year seedling survival of white spruce genetic changes in insect herbivory in birch (Betula
and birch in the Rosie Creek burn. In G.P. Juday pubesecens): the importance of plant apparency.
and C.T. Dyrness (eds.), Early Results of the Rosie Funct. Ecol. 31: 2224–2232.
Creek Research Project—1984. Agr. and For. Exp. Zweifel, R., Bohm, J.P., and Hasler, R. (2002). Midday
Sta., Univ. Alaska, Fairbanks, AK. Misc. Publ. 85–2. stomatal closure in Norway spruce-reactions
Zasada, J.C., Sharik, T.L., and Nygren, M. (1992). The in the upper and lower crown. Tree Physiol. 22:
reproductive process in boreal forest trees. In: 1125–1136.
A Systems Analysis of the Global Boreal Forest Zwolak, R. and Sih, A. (2020). Animal personalities and
(ed. H.H. Shugart, R. Leemans and G.B. Bonan), seed dispersal: a conceptual review. Funct. Ecol. 34:
211–233. New York: Cambridge Univ. Press. 1294–1310.
Scientific Names of Trees
and Shrubs
Acacia Acacia spp.

Ailanthus; tree-of-heaven Ailanthus altissima (Mill.) Swingle
Alaska-cedar Chamaecyparis nootkatensis (D. Don) Spach
Alder, green or Sitka Alnus crispa (Ait.) Pursh.
Alder, red or Oregon Alnus rubra (Bong.)
Alder, Sitka Alnus sinuata (Reg.) Rydb.
Alder, speckled Alnus rugosa (Du Roi) Spreng.
Apple, common Malus pumila Miller
Arborvitae Thuja occidentalis L.
Ash, black Fraxinus nigra Marsh.
Ash, blue Fraxinus quadrangulata Michx.
Ash, European Fraxinus excelsior L.
Ash, red or green Fraxinus pennsylvanica Marsh.
Ash, white Fraxinus americana L.
Ash, prickly- Zanthoxylum americanum Mill.
Aspen, bigtooth Populus grandidentata Michx.
Aspen, European Populus tremula L.
Aspen, trembling or quaking Populus tremuloides Michx.
Baldcypress Taxodium distichum (L.) Rich.
Basswood Tilia americana L.
Beech, American Fagus grandifolia Ehrh.
Beech, Antarctic Nothofagus antarctica (Forst.) Oerst.
Beech, blue- Carpinus caroliniana Walt.
Beech, European Fagus sylvatica L.
Bigtree; giant sequoia Sequoiadendron giganteum (Lindl.) Buchholz
Birch, black or sweet Betula lenta L.
Birch, bog, swamp, or low Betula pumila L.
Birch, European or silver Betula pendula Roth.
Birch, gray Betula populifolia Marsh.
731
732 S c i e n t i f i c Names of Trees a n d S h ru bs
Birch, paper or white Betula papyrifera Marsh.

Birch, river or red Betula nigra L.
Birch, Virginia round-leaf Betula uber (Ashe) Fern.
Birch, yellow Betula alleghaniensis Britton
Bitterbrush Purshia tridentata (Pursh) DC.
Bittersweet, Asian or Oriental Celastrus orbiculatus Thunb.
Blackgum Nyssa sylvatica Marsh.
Blueberry, lowbush Vaccinium angustifolium Aiton
Boxelder Acer negundo L.
Buckeye, Ohio Aesculus glabra Willd.
Buckeye, painted Aesculus sylvatica Bartr.
Buckeye, red Aesculus pavia L.
Buckeye, yellow Aesculus octandra Marsh.
Buckthorn, alder Rhamnus alnifolia L’Héritier
Buckthorn, common Rhamnus cathartica L.
Buckthorn, glossy Frangula alnus Mill
Butternut Juglans cinerea L.
Catalpa, northern Catalpa speciosa Warder
Catalpa, southern Catalpa bignonioides Walter
Ceanothus, redstem Ceanothus sanguineus Pursh.
Cedar, Alaska- Chamaecyparis nootkatensis (D. Don) Spach
Cedar, Atlantic white- Chamaecyparis thyoides (L.) B. S. P.
Cedar, eastern red- Juniperus virginiana L.
Cedar, incense Libocedrus decurrens Torr.
Cedar, northern white- Thuja occidentalis L.
Cedar, Port Orford- Chamaecyparis lawsoniana (A. Murr.) Parl.
Cedar, western red- Thuja plicata Donn
Cherry, black Prunus serotina Ehrh.
Cherry, choke Prunus virginiana L.
Cherry, pin or fire Prunus pensylvanica L. f.
Chestnut, American Castanea dentata (Marsh.) Borkh.
Chinquapin, golden Chrysolepis chrysophylla Douglas ex Hook.
Chinquapin, Ozark Castanea ozarkensis Ashe
Cottonwood, black Populus trichocarpa Torr. & Gray
Cottonwood, eastern Populus deltoides Bartr.
Cottonwood, European; black poplar Populus nigra L.
Cottonwood, Fremont Populus fremontii Wats.
Cottonwood, narrowleaf Populus angustifolia James
Cottonwood, plains Populus deltoides var. occidentalis Rydb.
Creeping strawberry-bush Euonymus obovata Nutt.
Creosote bush Larrea spp.
Cypress, Arizona Cupressus arizonica Greens
Cypress, Mexican Cupressus lusitanica Miller

Cypress, pond Taxodium distichum var. imbricarium (Nutt.)
Croon (syn = T. ascendens Brong.)
Dawn redwood Metasequoia glyptostroboides H. H. Hu & Cheng
Dogwood, alternate-leaf Cornus alternifolia L.
Dogwood, flowering Cornus florida L.
Dogwood, gray Cornus foemina Miller
Dogwood, red-osier Cornus stolonifera Michx.
Dogwood, roundleaf Cornus rugosa Lamarck
Dogwood, silky Cornus amomum Miller
Douglas-fir Pseudotsuga menziesii (Mirb.) Franco
Elm, American Ulmus americana L.
Elm, rock Ulmus thomasii Sarg.
Elm, Siberian Ulmus pumila L.
Elm, slippery Ulmus rubra Muhl.
Elm, winged Ulmus alata Michx.
Eucalyptus Eucalyptus spp.
Fetterbush Leucothoe spp.
Fir, alpine or subalpine Abies lasiocarpa (Hook.) Nutt.
Fir, balsam Abies balsamea (L.) Mill.
Fir, European silver Abies alba Mill.
Fir, Fraser Abies fraseri (Pursh.) Poir.
Fir, grand or lowland white Abies grandis (Dougl.) Lindl.
Fir, noble Abies procera Rehd.
Fir, Pacific silver Abies amabilis (Dougl.) Forbes
Fir, red Abies magnifica A. Murr.
Fir, white Abies concolor (Gord. & Glend.) Lindl.
Giant sequoia; bigtree Sequoiadendron giganteum (Lindl.) Buchholz
Ginkgo; maidenhair tree Ginkgo biloba L.
Gum, black; black tupelo Nyssa sylvatica Marshall
Hackberry Celtis occidentalis L.
Hackberry, dwarf Celtis tenuifolia Nuttall
Hackberry, iguana Celtis iguanaea (Jacq.) Sarg.
Haw, black Viburnum prunifolium L.
Haw, dotted Crataegus punctata Jacquin
Hawthorn Crataegus spp.
Hazel, beaked Corylus cornuta Marsh.
Hemlock, eastern Tsuga canadensis (L.) Carr.
Hemlock, mountain Tsuga mertensiana (Bong.) Carr.
Hemlock, western Tsuga heterophylla (Raf.) Sarg.
Hickory, bitternut Carya cordiformis (Wangenh.) K. Koch
Hickory, mockernut Carya tomentosa Nutt.
Hickory, pignut Carya glabra (Mill.) Sweet

Hickory, shagbark Carya ovata (Mill.) K. Koch
Hickory, shellbark Carya laciniosa (Michaux f.) G. Don
Hickory, water Carya aquatica (Michx. f.) Nutt.
Hobblebush Viburnum alnifolium Marsh.
Holly, American Ilex opaca Ait.
Honeysuckle, Amur Lonicera maackii (Rupr.) Maxim.
Hop-hornbeam; ironwood Ostrya virginiana (Mill.) K. Koch
Hornbeam Carpinus betulus L.
Horse-chestnut Aesculus hippocastanum L.
Ironwood; hop-hornbeam Ostrya virginiana (Mill.) K. Koch
Juniper, alligator Juniperus deppeana Steud.
Juniper, common Juniperus communis L.
Juniper, ground Juniperus communis L. var. depressa Pursh.
Juniper, one-seed Juniperus monosperma (Engelm.) Sarg.
Juniper, Rocky Mountain Juniperus scopulorum Sarg.
Juniper, Utah Juniperus osteosperma (Torr.) Little
Juniper, western Juniperus occidentalis Hook.
Kentucky coffeetree Gymnocladus dioicus (L.) K. Koch
Kudzu Pueraria lobata (Willd.) Ohwi
Larch, eastern; tamarack Larix laricina (Du Roi) K. Koch
Larch, European Larix decidua Mill.
Larch, subalpine Larix lyallii Parl.
Larch, western Larix occidentalis Nutt.
Laurel, mountain Kalmia latifolia L.
Leatherleaf Chamaedaphne calyculata (L.) Moench
Lilac Syringa vulgaris L.
Locust, black Robinia pseudoacacia L.
Locust, honey Gleditsia triacanthos L.
Madrone, Pacific Arbutus menziesii Pursh.
Magnolia, Fraser Magnolia fraseri Walt.
Magnolia, southern or evergreen Magnolia grandiflora L.
Magnolia, sweetbay Magnolia virginiana L.
Magnolia, umbrella Magnolia tripetala L.
Maidenhair tree; gingko Ginkgo biloba L.
Mangrove, black Avicennia nitida Jacq.
Mangrove, red Rhizophora mangle L.
Maple, ash-leaf Acer negundo L.
Maple, bigleaf Acer macrophyllum Pursh.
Maple, bigtooth Acer grandidentatum Nutt.
Maple, black Acer nigrum Michaux f.
Maple, mountain Acer spicatum Lam.
Maple, Norway Acer platanoides L.

Maple, red Acer rubrum L.
Maple, silver Acer saccharinum L.
Maple, striped Acer pensylvanicum L.
Maple, sugar Acer saccharum Marsh.
Maple, vine Acer circinatum Pursh.
Mesquite Prosopis juliflora (Sw.) DC.
Mexican cypress Cupressus lusitanica Miller
Mimosa Albizia julibrissin Durazz.
Mountain-ash, American Sorbus americana Marsh.
Mountain-ash, showy Sorbus decora (Sarg.) Schneid.
Mountain laurel Kalmia latifolia L.
Mulberry, red Morus rubra L.
Mulberry, white Morus alba L.
Myrtle, California wax Myrica californica Cham. & Schltdl.
Nannyberry Viburnum lentago L.
Oak, bear Quercus ilicifolia Wangenh.
Oak, black Quercus velutina Lam.
Oak, blackjack Quercus marilandica Muenchh.
Oak, bur Quercus macrocarpa Michx.
Oak, California black Quercus kelloggii Newb.
Oak, California live Quercus agrifolia Née
Oak, canyon live Quercus chrysolepis Liebm.
Oak, cherrybark Quercus pagoda Raf. (syn. Q. falcata var. pagodifolia 111.)
Oak, chestnut or rock chestnut Quercus prinus L.
Oak, chinquapin or yellow Quercus muehlenbergii Engelm.
Oak, diamond leaf Quercus laurifolia Michx.
Oak, dwarf chestnut Quercus prinoides Willd.
Oak, Emory Quercus emoryi Torr.
Oak, English, European, or common Quercus robur L.
Oak, Gambel Quercus gambelii Nutt.
Oak, holm Quercus ilex L.
Oak, interior live Quercus wislizeni A. DC.
Oak, live Quercus virginiana Mill.
Oak, northern pin Quercus ellipsoidalis E. J. Hill
Oak, Oregon Quercus garryana Dougl.
Oak, overcup Quercus lyrata Walt.
Oak, pin Quercus palustris Muenchh.
Oak, post Quercus stellata Wangenh.
Oak, northern red Quercus rubra L.
Oak, sand live Quercus geminata Small, [syn. Q. virginiana var. maritima
(Michx.) Sarg.]
Oak, scarlet Quercus coccinea Muenchh.

Oak, shingle Quercus imbricaria Michx.
Oak, southern red Quercus falcata Michx.
Oak, swamp chestnut Quercus michauxii Nutt.
Oak, swamp laurel Quercus laurifolia Michx.
Oak, swamp white Quercus bicolor Willd.
Oak, turkey Quercus laevis Walt.
Oak, water Quercus nigra L.
Oak, white Quercus alba L.
Oak, willow Quercus phellos L.
Olive, Russian Elaeagnus angustifolia L.
Osage-orange Madura pomifera (Raf.) C. K. Schneid.
Pacific silver fir Abies amabilis (Dougl.) Forbes
Palm, cabbage Sabal palmetto (Walt.) Lodd.
Palmetto, cabbage Sabal palmetto (Walt.) Lodd.
Palmetto, saw Serenoa repens (Bartr.) Small
Pawpaw Asimina triloba (L.) Dunal
Pine, bishop Pinus muricata D. Don
Pine, black or Austrian Pinus nigra Arnold
Pine, bristlecone Pinus aristata Engelm.
Pine, Caribbean Pinus caribaea Morelet
Pine, digger Pinus sabiniana Dougl.
Pine, eastern white Pinus strobus L.
Pine, erectcone or Calabrian Pinus brutia Ten.
Pine, Jack Pinus banksiana Lamb.
Pine, Jeffrey Pinus jeffreyi Grev. & Balf.
Pine, knobcone Pinus attenuata Lemmon
Pine, limber Pinus flexilis James
Pine, loblolly Pinus taeda L.
Pine, lodgepole Pinus contorta Dougl.
Coastal lodgepole pine Pinus contorta ssp. contorta
Mendocino White Plains lodgepole pine Pinus contorta ssp. bolanderi
Rocky Mountain-Intermountain Pinus contorta ssp. latifolia
lodgepole pine
Sierra–Cascade lodgepole pine Pinus contorta ssp. murrayana
Pine, longleaf Pinus palustris Mill.
Pine, Mexican pinyon (piñon) Pinus cembroides Zucc.
Pine, Mexican white Pinus ayacahuite Ehrenb.
Pine, Monterey or radiata Pinus radiata D. Don
Pine, mugo or mountain Pinus mugo Turra. [P. montana Mill.]
Pine, patula Pinus patula Schl. & Cham.
Pine, pinyon (piñon) Pinus edulis Engelm.

or P. monophylla Torr. & Frem.
or P. cembroides Zucc.
or P. quadrifolia Parl. ex Sudw.
Pine, pitch Pinus rigida Mill.
Pine, pond Pinus serotina Michx.
Pine, ponderosa Pinus ponderosa Laws.
Pine, red Pinus resinosa Ait.
Pine, sand Pinus clausa (Chapm.) Vasey
Pine, Scots or Scotch Pinus sylvestris L.
Pine, shortleaf Pinus echinata Mill.
Pine, slash Pinus elliottii Engelm. var. elliottii
Pine, sugar Pinus lambertiana Dougl.
Pine, stone or Swiss stone Pinus cembra L.
Pine, table mountain Pinus pungens Lamb.
Pine, Torrey Pinus torreyana Parry ex Carr.
Pine, Virginia Pinus virginiana Mill.
Pine, western white Pinus monticola Dougl.
Pine, whitebark Pinus albicaulis Engelm.
Planetree, oriental Platanus orientalis L.
Plum, Allegheny Prunus alleghaniensis Porter
Poplar, balsam Populus balsamifera L.
Poplar, black hybrid Populus x euramericana
Poplar, European white Populus alba L.
Poplar, Lombardy Populus nigra ‘italica’ Muenchh.
Prickly-ash Zanthoxylum americanum Mill.
Privet, swamp Forestiera acuminata (Michx.) Poir.
Redbud, eastern Cercis canadensis L.
Redwood Sequoia sempervirens (D. Don) Endl.
Redwood, dawn Metasequoia glyptostroboides H. H. Hu & Cheng
Rhododendron, rosebay Rhododendron maximum L.
Rose Rosa spp.
Sagebrush, big Artemisia tridentata Nutt.
Saguaro; giant cactus Cereus giganteus Engelm.
Salmonberry Rubus spectabilis Pursh.
Sassafras Sassafras albidum (Nutt.) Nees
Sequoia, giant; bigtree Sequoiadendron giganteum (Lindl.) Buchholz
Serviceberry, downy Amelanchier arborea (Michaux f.) Fernald
Serviceberry, smooth Amelanchier laevis Wieg.
Sheepberry Viburnum lentago L.
Silver bell; Carolina silverbell Halesia carolina L.
Snowbrush Ceanothus velutinus Doug. ex Hook
Sourwood Oxydendrum arboreum (L.) DC.

Spruce, black Picea mariana (Mill.) B. S. P.
Spruce, Colorado blue Picea pungens Engelm.
Spruce, Engelmann Picea engelmannii Parry
Spruce, Norway or European Picea abies (L.) Karst.
Spruce, red Picea rubens Sarg.
Spruce, Sitka Picea sitchensis (Bong.) Carr.
Spruce, white Picea glauca (Moench) Voss
Sugarberry Celtis laevigata Willd.
Sumac, smooth Rhus glabra L.
Sweetgum Liquidambar styraciflua L.
Sycamore, American Platanus occidentalis L.
Tallow, Chinese Triadica sebifera (L.) Small
Tamarack; eastern larch Larix laricina (Du Roi) K. Koch
Tamarix, five-stamen Tamarix pentandra Pall.
Torreya, Florida Torreya taxifolia Arn.
Tree-of-heaven; ailanthus Ailanthus altissima (Mill.) Swingle
Tuliptree; yellow-poplar Liriodendron tulipifera L.
Tupelo, water or swamp Nyssa aquatica L.
Virginia creeper Parthenocissus quinquefolia (L.) Planchon
Walnut, black Juglans nigra L.
Willow, black Salix nigra Marsh.
Willow, crack Salix fragilis L.
Willow, peachleaf Salix amygdaloides Andersson
Willow, sandbar Salix interior Rowlee
Willow, weeping Salix babylonica L.
Witch-hazel Hamamelis virginiana L.
Yellow-poplar; tuliptree Liriodendron tulipifera L.
Yew, Canada Taxus canadensis Marshall
Yew, Florida Taxus floridana Nutt.
Yew, Pacific Taxus brevifolia Nutt.
Index
Note: Page numbers followed by “f” and “t” figures and tables, respectively.
A allopatric speciation, 57
abiotic environment, 4 allozymes, 40
aboveground net primary productivity (ANPP), 473, alpha diversity, 332, 333–335, 334f
477f, 506, 635, 637 alpine tree lines, 302–303
for alpine, boreal, temperate, and tropical terrestrial altered fire regimes, 600–601
ecosystems, 474f alumina octrahedra structure, 208, 208f
relationship with aboveground litter American beech, 444–446, 445f
production, 507f American chestnut trees, 411, 603, 605f
soil properties, forest biomass, and, 475–476 American elm, 411, 605
terrestrial ecosystems, 473f amino acids, 515
water balance and, 476f andisols, 214
absorption spectra, 133 angiosperms, 71n1
abstract space, 299 animals in forest ecosystems, 269
acidity of soil, 210–212 animal predation effect on seed availability, 288t
actinobacteria, 498 change in white tailed deer populations, 286f
action spectra, 133 elk browsing, 287f
active seed bank, 71t influence of livestock on forest ecosystems,
actual evapotranspiration (AET), 474, 474f 288–290
adaptation anthesis, 74, 75
to climate change effects on forests, 581–587 antifreeze, 159
of fire in forest trees, 224–227 apical control, 95, 97
local, 55 apical dominance, 96, 97
adaptedness, 35 apomixis, 39
adaptive introgression, 62 apparency hypothesis, 270–271
adhesion, 203 arbuscular mycorrhizae (AM), 504–505, 505f
adult phase, 71–73 architectural models of trees, 97–102, 98f
advanced regeneration, 425 canopy architecture of individual species, 98–99
aerial roots, 104 patterns of intermittent growth, 99–102
aggradation, 426 short and long shoots, 98
aggrading, 423 sylleptic and proleptic shoots, 102
aggregates of soil, 201–202, 233 asexual reproduction of woody plants, 70, 71, 388. See
agricultural landscapes, 647 also sexual reproduction of woody plants
albedo, 124 asexual systems, 39
alfisols, 214 aspect of sloping terrain, 171, 172–173
algal crusts, 497 aspen–birch forests, 563
alien non‐native species, 590 assimilation, 500
alleles, 37 assisted colonization. See assisted species migration
allogenic coexistence, 444 assisted gene flow. See assisted population migration
allogenic succession, 416 assisted genetic migration. See assisted
allometric biomass equations, 470 population migration
739
bindex.indd 739 11-14-2022 20:52:31

740 Index
assisted long‐distance migration. See assisted bole, 226

species migration boundary layer of leaf, 151
assisted population migration, 581 branch habit and self‐pruning ability, 225–226
assisted range expansion, 581–582 brown‐rot fungi, 517
assisted species migration, 581–583, 582f browsers, 269, 288
associations of species, 296 bud burst, 46, 145, 161–163, 551
atmospheric CO2, 540 building. See aggrading
atmospheric deposition, 492–495, 494f, 496f buttresses, 110
atmospheric water potential, 204
autecology, 12 C
autogenic succession, 416 Cajander’s method, 263
available water content, 205 California acorn woodpecker (Melanerpes
avoiders, 227, 228 formicivorus), 279, 279f
Azotobacter, 497 California and Pacific Northwest mountainous
terrain, 176–180
B Canadian Forest Ecosystem Classification System
balsam poplar, 43 (CFEC), 263
beech bark disease, 599 canopy closure, 308
belowground net primary productivity, 482–485. canopy packing, 129
See also aboveground net primary canopy structural complexity (CSC), 130
productivity (ANPP) canopy structure, 125, 127f
beta diversity, 332, 335–336 CSC, 130
biodiversity, 327, 332 effects on light, 125
in ecosystem stability, 356–357 and LAI, 126, 127–129
preservation, 358, 644–645, 649 structural differences in, 126
value of, 329–331 carbohydrates, 458
biodiversity–productivity relationship, 354–356 carbon (C), 449, 487
bioecosystem approach, 4 biochemical constituents of needles and
biological control for invasive species, 610 shoots, 457t
biological diversity. See biodiversity budget of 14‐year‐old Scots pine tree, 459t
biological legacies, 385–386, 414, 415t dynamics, 635
biomass emissions for forests, 586f
distribution, 467t flows, 449
of forest ecosystems, 466–469 gain, 574
measurement, 469–473 gain hypothesis, 319
biomass accumulation during ecosystem light and C allocation, 461
development, 477, 480f loss, 575–577
biomass pools in second‐and old‐growth and nutrient cycling, 601–602
ecosystems, 482t pools, 575
changes in processes controlling, 478f sequestration, 573
natural disturbance, 481 sink, 637
net ecosystem production, 479–480 soil nitrogen availability and C allocation, 461–465
biomes, 20–25, 291 source, 637
distribution of, 23f carbon balance, 449
of Eurasia and North America, 24f controlling processes of terrestrial ecosystems, 468f
biota, 237 of ecosystems, 465–482
biotic community, 3 of trees, 450–465
biotic composition changes, 411 carbon dioxide (CO2), 13, 123, 207, 540
addition of species, 412 exchange system, 470, 471f
elimination of species, 411–412 fertilization, 549–550
biotic regions. See biomes carnivorous bird species, 622
birds, 279–281 catastrophic and local land movements, 407
bitternut hickory, 378–379, 379f cation exchange capacity of soil, 209–210
black locust, 498, 602 cellulose, 515
bindex.indd 740 11-14-2022 20:52:31

Index 741
central Appalachians forests and physiography, coevolutionary systems, 42–43

180–183 cogongrass, 600, 601f
cheatgrass, 600, 601f, 602 cohesion, 203
chemical control for invasive species, 610 cold acclimation process, 156
chemical defenses, 270 cold‐air drainage effects on forests, 148–150
chestnut blight, 598, 603–606 cold injury to plants, 153–154
chlorite, 208 Colorado Front Range, 647
chloroplast DNA markers (cpDNA markers), 62–63 common garden approach, 37
chronosequence, 416 communities, 291, 296
clay mineralogy, 207–209 competition and niche differentiation, 305
Clements concept of ecology, 297–298 composition and structure, 349
Clementsian succession, 419–420, 420f continuum concept, 298–299
climate, 12, 473–475, 539 grid of species interactions, 307t
of broadscale ecosystem, 25, 26 grounding, 292–296
climate‐disturbance–carbon interactions, 580 interactions among organisms, 307–315
envelopes, 558 as landscape ecosystem property, 299–300
regulation, 647 pattern of co‐occurrence of species on
climate change, 331, 449, 539 landscape, 300f
adapting to climate change effects on potential and optimum range of tree species, 306f
forests, 581–587 schools and terminology, 296–298
effects on forest carbon, 573–581 spatial variation in forest communities, 300–305
effects on forest disturbances, 563–573 type, 297
effects on forest tree, 548–556 understory tolerance, 315–324
effects on tree species distributions, 556–563 compensatory dynamics, 356
global temperature changes, 541f competition, 305, 308, 600
precipitation effects, 545–547 changing composition in northern hardwood
refugia, 585f forest, 314t
temperature effects, 542–544 differentiation of trees of pure even‐aged stand, 309f
variation in temperature and atmospheric CO2 forest community structure and
concentration, 540f composition, 308–313
climatic classification, 18 and overstory composition, 313–314
ecoclimatic zones of world, 18f tree diameter distributions, 310f
vertical zonation, 19f in understory, 314–315
zonal relationships between climate, soil orders, and competitive exclusion principle, 308
vegetation, 19t complementarity, 354
climatic maximum. See climax compounded disturbances, 409
climax, 297, 419 coniferous trees. See conifers
climatic, 431 conifers, 73, 75, 115, 135, 222, 548
monoclimax, 419 growth rings, 112
pattern of vegetation, 432 pollination and fertilization in, 74
physiographic, 431 vegetative reproduction by, 92, 93
site, 432 connectivity, 622–624
clines consolidated parent material, 195–197, 196f
of glaucousness, 57 construction respiration, 456–458
of trees, 41 continental relationships, 343–344
clone, 39 continental scales, diversity at, 343–346
formation, 70 continuum concept, 298–299
maintenance or expansion, 71t corridors, 618
trembling aspen clones, 91, 92f, 92 cowbirds, 621
vine maple clone, 92, 93f crown classes, 309–310
closed‐cone pines, 228–230 crown dieback, 80
cloud deposition, 492 crown fires, 220, 221
coarse woody debris, 577 cultural control for invasive species, 609
co‐dominant crown classes, 309 cultural services, 644, 645t
bindex.indd 741 11-14-2022 20:52:31

742 Index
D catastrophic and local land movements, 407

damage as ecosystem process, 384–386
avoiding fire damage, 224–226 effects on heterogeneity, 627–632
and death, 286–288 events, 15, 16, 16f
heat damage in seedlings, 151–152 fire, 387–400
principles of wind damage, 401–403 floodwater and ice storms, 404–405
recovering from fire damage, 226 in forest ecosystems, 387
severity, 624 heterogeneity effects on, 624–627
dark respiration, 133, 450–452 historical range of variability, 632–634
de‐acclimation, 159 insects and disease, 406–407
dead wood, 479–480 interactions, 409–411, 410f
deciduousness and temperature, 164 land clearing, 409
declining. See senescing on landscapes, 624
decomposers, 269 linked, 409
decomposition, 285–286, 575–577 lodgepole pine trees, 406f
decomposition dynamics, 517 logging, 407–408
abscised leaf, 519f refugia, 584
changes in C and N contents of leaf litter, 522f regeneration wave, 405f
rate constants, 518t, 520f regimes, 16, 16f, 383
decurrent form of trees, 96f, 97 source, 386
deep rooting, 225, 226 wave‐regenerated fir species, 404
definite habitat, 42 by wind, 174, 175, 400–404
degenerating. See senescing diversity, 327. See also biodiversity
deglaciated terrain, primary succession on, 434–437, biological and ecosystem, 327
435f, 436f causes of species, 343–351
degree days, 144–145 conserving ecosystem and biological, 363–364
de‐hardening, 159 ecosystem, 327, 342
dehydration, 159 focal species in conserving, 351–354
deliquescent form of trees. See decurrent form of trees forest management and, 357–363
dendrochronology, 115 and functioning of ecosystems, 354–357
denitrification, 526–528 ground‐cover species, 337–341
density‐dependent competition, 312 index of, 334
deposition velocity, 495 landscape ecosystems, 336–337
derecho, 401, 573 levels, 331–333
determinate growth of plants, 98 measuring, 331–342
detritivores, 269 value of species, 328–331
diameter at breast height (DBH), 444, 469–470 dominant crown classes, 309
differentiation, 335–336 dominant unions, 247
differentiation diversity, 332 dormant/dormancy, 155–156
diffuse‐porous species, 112 degree of, 161
diffusion, 499 and persistent seed bank, 71t
digestibility reducers, 270 transitions in seasonal growth–dormancy cycling in
digital elevation models (DEMs), 172, 183 trees, 162f
dioecious condition, 73 of woody plants, 82–83, 82f
direct exploitation of species, 331 Douglas‐fir, 590
discrete forest communities, 300. See also merging forests, 643
forest communities genetic differentiation of, 52–54
alpine tree lines, 302–303 germination and survival of seeds of, 86f
Coastal California, 301–302 in Pacific Northwest, 394
forest–grassland ecotone, 302 pollination and fertilization in, 74
marine terraces, 301f regeneration, 555f
disease, 406–407 reproductive cycle of, 75f
disturbance(s), 349–351, 381–384, 425 study of seedling survival of, 85
ball‐in‐cup diagrams, 384f downgrading. See senescing
biotic composition changes, 411–412 droughts, 547, 572
bindex.indd 742 11-14-2022 20:52:31

Index 743
drought tolerance, 117 ecotype of trees, 42

dry deposition, 492, 494–495 ectomycorrhizal fungi (ECM), 504–505, 505f
dry‐matter accumulation, 388 edaphic characteristics of soil, 12
Dutch elm disease, 598–599, 603–606 eddy correlation technique, 472
dwarf shoots of plants. See short shoots of plants eddy covariance method, 471
edge effects, 621
E edge species, 621
early‐successional species, 85, 86 effective depth, 254, 256
earlywood formation control, 114–115 elements situated in landscape, 613
Earth as metaphor for life, 650–651 elevation of landforms, 170, 344–346
earthworms, 607 elm–ash–cottonwood, 563
eastern North American species, 47–49 emerald ash borer, 596, 603, 604f
eastern redcedar, 281, 281f endangered species, 353–354
ecoclimatic zones, 27 endemics, 353–354
ecological/ecology, 3, 4 endemism, 353
processes, 634–640 endurers, 227, 228
regionalization, 26 enemy release hypothesis, 591
responses, 16 entisols, 214
separation, 58 environmental factors as measure of site, 251
species groups, 245–247 aspect, slope steepness, and site index for
system. See ecosystem black oak, 255f
ecological forestry, 359‐363, 643 climatic factors, 251–252
ecological land classification (ELC), 262–264 land types characteristic of dissected forest
ecological site units (ESUs), 292, 293f terrain, 253f
economy functions, 647 physiographic and soil factors, 252–256
ecosystem change, 431–432 physiographic land classification, 252
light and, 141 productivity curve with site‐index ratio, 255f
in southern Appalachians, 373 soil surveys, 256
ecosystems, 3, 4, 7, 228, 642, 651. See also forest environmental filters, 595
ecosystems epicormic branches, 139
carbon balance and retention and loss of epicormic sprouting, light and, 139
nutrients, 531–532 epigenetics, 66–67
complete, 13 epigeous seed germination, 83, 84f, 85
delineation, 28 epsilon diversity, 332
disturbances as ecosystem process, 384–386 erosion, 234–236, 234f
diversity, 327, 342 establishment stage. See stand initiation
functioning, 354–357 ethological separation, 58
geography, 25 Eurasian wild boars, 607
management, 643 evaders, 227, 228
and mapping in Baden‐Württemberg, 257–258 evapotranspiration (ET), 635, 636f, 475
and mapping in Michigan, 258–260 evenness of species, 331–334, 629
non‐exhaustive list, 645t evolutionary biology in sexual populations, 38
services, 329, 644–645, 650 natural selection, 38–39
in southeastern United States, 261–262 sexual and asexual systems, 39
in southwestern United States, 262 excurrent form of trees, 95–96, 96f
types, 433 exotic non‐native species, 590
ecosystems, carbon balance of, 465 exposure time, 151
biomass accumulation during ecosystem extensive tree mortality, 603
development, 477–482
biomass and productivity of, 466–469 F
climate and productivity, 473–475 facilitation, 307, 354, 428–431, 429f
measurement of biomass and fascicles, 98
productivity, 469–473 feral swine, 607
soil properties, forest biomass, and ANPP, 475–476 fertilization effect, 549
ecotone, 293, 301 fiber farms, 649
bindex.indd 743 11-14-2022 20:52:31

744 Index
field capacity, 204, 205 floodplains, 186

fine‐grained mica, 208 changes in physiography of, 190
fine roots, 105–107, 107f, 463 forests, 563
influence of soil N availability, 465f landform zonation pattern, 187, 188
life span, 463 Mississippi River floodplain, 187f
mortality, 464 in outwash plains and outwash over moraine
firebreaks, physiography and, 190–191 physiography, 189f
fire cycle. See fire rotation similar effects, principle of, 187
fire frequency, 220, 221 variation in types of landforms, 189
fire‐induced opening of cones, 71t Florida Keys, 292–293
fire in forest trees, 220, 230, 387, 565. See also flowering, 226
forest trees mast, 76
adaptations, 224–227 of woody plants, 72, 73, 75
avoiding fire damage, 224–226 flushing, 46
beneficial effects of, 236 focal species in conserving diversity, 351
in boreal forest and taiga, 397–398 endemics and rare and endangered species, 353–354
causes, 220 foundation species, 352
colonizing burned‐over areas, 226–227 keystone species, 352–353
direct effects of, 232–233 tree species richness, 353f, 355f
exclusion, 399–400 foliage retention near ground, 227
fire‐scar cross section of ponderosa pine, 223f forcing, 161
fire regime, 220–224, 225f forest, 3, 197
in forest ecosystems, 387–395 applicability to forest management, 12–13
history and behavior, 395–399 biomass, 475–476
indirect effects of, 230–232 canopy closure event, 425
key characteristics of forest species, 224 direct measurement of forest productivity, 238–239
mean annual number, 567f forest–grassland ecotone, 302
and oak dominance, 446–447 giant and pygmy, 301–302
organic matter removal and erosion, 234–236 harvesting and nutrient loss, 532
processes directing forest conversion due interior species, 621
to, 568f projected changes in forest type
promoting fire occurrence, 227 distributions, 561–563
recovering from fire damage, 226 short‐term changes in flow of water and nutrients, 534f
roles in landscape ecosystems, 219 temperature within, 147–148
strategies of species persistence, 227–230 types, 3, 263, 297
suppression, 399–400, 634, 647 forest carbon
types, 221 carbon gain, 574
fire in northern Lake States, 395–397 carbon loss, 575–577
fire in northern Rocky Mountains, 398–399 climate change effects, 573
fire intensity, 220, 221, 222 dynamics, 637–638
fire‐resistant live foliage, 226 feedback among disturbance, climate change, and
fire rotation, 220 carbon, 577–581
fire‐triggered dehiscence of fruits and cones, 227 hypothetical trajectory, 578f
fire–vegetation feedbacks, disruption of, 569–570 management, 585–587
fish, 278 forest community, 3
fitness–flexibility compromise, 66 discrete forest communities, 300–303
flagship species, 351 merging forest communities, 303–305
flammable foliage and bark, 227 numbers of trees per hectare, 311f
flatlands, 183 stand density, 312–313
floodplains, 186–190 structure and composition, 308
Great Plains, 184–185 vertical structure, 310–312
pine savannas of western Great Lakes Region, 185 forest disturbances, climate change effects on, 563.
till plains of Midwest, 185 See also disturbance(s)
southeastern and southern Coastal Plain, 186 fires, 565–570
flat terrain, 171 insects and pathogens, 570–572
flooding/floodwater, 404–405, 628 wind, 572–573
bindex.indd 744 11-14-2022 20:52:31

Index 745
forest ecology, 3, 641. See also landscape ecology genecology, 40–57

applicability to forest management, 12–13 genetic diversity of woody species, 39–40
approach to study of, 11–12 growth and mortality, 548–550
landscape ecosystems, 3, 4–11 hybridization, 60–63
prevalence of human values in, 641–642 maximum photosynthetic capacity of, 450
forest ecosystems, 12, 449. See also ecosystems; long‐ niche, 60
term forest ecosystem and vegetation change phenology, 550–553
biomass and productivity of, 466–469 polyploidy, 63–65
influence of livestock on, 288–290 regeneration, 553–556
instrumentation, 472f resources for growth and development, 195
mean residence time of organic matter, nitrogen sources of genetic variation, 37–38
and phosphorus, 514t formal ecological theory, 417
mycorrhiza, 504–506 fossil charcoal, 220
nitrogen content of leaf and fine‐root litter, 511t foundation species, 351–352
nutrient cycling within, 499 Frankia, 498
nutrients in forest floor, 511–513 free‐air CO2 enrichment (FACE), 549
nutrient transport to roots, 499–500 free‐living cyanobacteria, 497
nutrient uptake and assimilation by roots, frost hardiness and cold resistance, 156
500–503 freeze avoidance by plants, 157, 159
organic matter decomposition and nutrient hardening and de‐hardening of twigs, 158f
mineralization, 513–524 mechanisms of freezing resistance, 158f
plant litter and return of nutrients to forest floor plants cope with freezing temperatures, 157
and soil, 506–511 regression of killing temperature, 157f
plant root systems, 503f thermotropic movements in rhododendrons, 159–161
root architecture, 503–504 frost pockets, 148
understory tolerance examples in, 319 fuel moisture, 623
forest fragmentation, 614, 619 functional component of niche, 60
connectivity, 622–624
distribution of eastern white pine, 619f G
ecological effects, 619–622 gametic incompatibility, 58
habitat loss on landscape pattern, 620f gamma diversity, 332
patches in forest ecology, 618–619 gap dynamics, 439–446
forest management gap‐phase species, 87–88, 423
and diversity, 357 gap specialists, 444–446
ecological forestry, 359–363 gene
effects on diversity, 357–358 flow in plants, 38, 56–57, 59–60
in North America, 643 mutation, 37–38
preserving diversity in managed forests, 358–359 recombination, 37, 38
forestry‐assisted migration, 583 upregulation, 50
forest succession, 413, 603–606 genecology, 40–41
change in ecosystems, 431–432 clinal variation of critical night length, 47f
fire and oak dominance, 446–447 examples of genecological differentiation, 46–54
following eruption of Mount St. Helens, factors affecting differentiation, 56–57
437–439 factors eliciting genecological differentiation,
primary succession on deglaciated terrain, 42–46
434–437 genecological categories, 42
secondary succession and gap dynamics, local genecological differentiation, 55
439–446 patterns of genecological differentiation, 41–42
secondary succession following fire in ponderosa genet, 71
pine forests, 439 genetic diversity/variation of forest trees
forest trees. See fire in forest trees of ponderosa pine, 51
components of phenotypic variation, 35–37 sources of, 37–38
ecological considerations at species level, 57–60 of woody species, 39–40
epigenetics, 66–67 genome sequencing, 59
evolutionary sequence, 38–39 genomics, 59
fitness–flexibility compromise, 66 genotype, 11, 35
bindex.indd 745 11-14-2022 20:52:31

746 Index
genotype–environment interaction, 35, 36 H

geoecosystems, 4 habitat loss and fragmentation, 331
geographic ecosystems. See landscape ecosystems habitat type, 247
geographic separation, 57 heartwood, 111
geometric series, 335 heat capacity of soils, 146
geomorphic processes, 16 heat‐induced germination, 227
geomorphology, 168 heat sum. See temperature sum
germination heavy‐seeded species, 81
and establishment, 284–285 hemicellulose, 517
of woody plants, 82–83, 82f hemlock seedlings, 60
giant sequoia, 394–395, 396f hemlock woolly adelgid, 352, 572, 602, 603, 606, 610
gibberellic acid application, 73 herb diversity, 345–346
glacial microrefugia, migration from, 377–379 herbivores, 269
glacial refugia, 583 heritability, 36
glaciated landscapes, 171 heterogeneity effects on disturbances, 624–627
glaciation, 344 heterotrophic respiration, 575–577
Gleason concept of ecology, 297–298 high earthworm biomass, 607
global warming, 542. See also climate change Hills’ physiographic approach, 263, 264f
glucose, 457 historical range of variability (HRV), 632–634
gradient analysis approach, 298 histosols, 214
grain, 15 horizons of soil profile, 198, 198f
grass stage, 225, 393 A horizon, 199
gravitational potential of soil water, 204 C horizon, 199
grazers, 269 O horizon, 198, 199
Greater Yellowstone Ecosystem, 569, 578 R horizon, 199
Great Plains ecosystems, 184–185 host resistance, 610
Great Smoky Mountains, 303 human‐caused disturbances, 381
greenhouse effect, 540 hurricanes, 401, 572
green plant chloroplasts, 123 1938 Hurricane, 404, 624–625
gross photosynthesis, 450 broadscale disturbance by, 403–404
gross primary production (GPP), 467, 468t, 470, 531 generalized pathways, 625f
ground‐cover species diversity in northern Lower Harvard Tract of Pisgah National Forest, 626f
Michigan, 337 Hurricane Andrew, 403
alpha diversity indices for, 339t in New England, 624–627
alpha diversity measure for upland ecosystem post‐hurricane damage classes, 626f
groups, 341t hybrid bridge hypothesis, 276
beta diversity measures for ecosystem groups, 340t hybrid inviability, 59
ecosystem groups, 337–340, 338f hybridization, 38, 59, 60–63
of ecosystems, 335, 336f hybrid sterility, 59
ecosystem types, 340–341 hybrid weakness, 59
site conditions of six landscape ecosystem hydrarch succession, 420, 421t
groups, 338t hydrogen (H), 487
ground fires, 220, 221 hypertrophy, 110
grounding communities, 292 hypogeous seed germination, 83, 84f, 85
Florida Keys, 292–293
interior Alaska, 294 I
southern Illinois, 294–296 ice‐contact terrain, 171
growing season length, 44, 45, 144, 543, 550 ice storms, 404–405
growth cessation of trees, 43–46 igneous rocks, 195, 197t
growth–differentiation balance hypothesis, 271–272 illuminance, 124
growth‐intercept method, 243 illuviation, 199
growth resumption of trees, 46 importance value, 559
growth rings of trees, 111–112 inceptisols, 214
gusts, 401, 402f indeterminate growth of plants, 98
gypsy moth. See spongy moth indicator plants/species, 246, 351
bindex.indd 746 11-14-2022 20:52:31

Index 747
individualistic concept of ecology, 297 knee roots, 110

inhibition mechanism, 428–431, 429f kudzu, 592, 592f, 600
initial floristic composition (IFC), 422–423
injury of plants, 273–274 L
insects, 406–407, 570–572, 571f lammas growth, 102
herbivores with distributions in hybrid zone, 276f land clearing, 409
lower bole of lodgepole pine, 273f landforms, 168, 170, 186, 433. See also physiographic/
performance and simultaneous and sustained physiography
plant stress, 275f aspect of slope, 172–173
plant defense against, 272–276 elevation of, 170
insolation, 124, 124f levels of landscape ecosystems, 434f
integrated moisture index, 183, 184f level terrain, 170–171
integrated pest management (IPM), 610 parent material in relation to, 173
intermediate crown classes, 309 position in landscape, 174
intermediate disturbance hypothesis, 349 position on slope, 172
intermittent growth patterns of trees, 99–102 shape of, 170
International Union for Conservation of Nature slope characteristics, 171
(IUCN), 358 slope inclination, 173
internodes, 99 sloping terrain, 171
interspecific competition, 308 spatial position of, 175
intolerant. See understory intolerant zonation pattern, 187, 188
intraspecific competition, 308 landscape, 15, 614
introgression, 61, 62f agricultural, 647
introgressive hybridization. See introgression change, 616
invaders, 227–228 function, 615
invasive grasses, 600 occupied, 563
invasive pathogens, 598 relationships of soil, 215–216
invasive species, 331, 589–591 structure, 614–615
characteristic traits of invasive plant landscape ecology, 15, 613
species, 591–595 association of mineralization and nitrification
growing conditions, 595f rates, 639t
impacts of invasive animals on forests, 606–607 carbon balance, 638f
impacts of invasive insects and pathogens on concepts, 614–616
forests, 602–606 disturbances on landscapes, 624–634
impacts of invasive plants on forests, 599–602 forest fragmentation and connectivity, 614–627
non‐plant invasive species in forests, 596–599 housing density, 616f
performance of native vs. invasive plants, 594f interactions of landscape patterns and ecological
primer of invasive species management in processes, 634–640
forests, 607–612 post‐fire landscape structure characterization, 615f
trees invasive in North America and native to North landscape ecosystems, 3, 4–7
America, 591t approach, 640
inventory diversity, 332, 333–335 and community, 7
inversion in mountainous terrain, 148 distinguishing and mapping at multiple spatial
island biogeography theory, 618 scales, 25–32
isomorphic substitution, 208–209 diversity of, 336–337, 342, 342f
flow of energy and matter, 8f
J objects of study, 5f
Jaccard similarity coefficients, 335, 339 spatial scales of, 15–17, 18–20
jack pine, 501, 632, 634, 634f structure and function, 7–9
jack pine barrens, 185 temporal scales, 15–17
juvenile phase, 71–73, 277 three‐dimensional, volumetric nature of, 5f
types, 9–11, 10f
K of upper and lower slopes, 6f
kaolinite, 208 vegetation types and biomes, 20–25
keystone species, 351–353 landscape‐mosaic view, 619
bindex.indd 747 11-14-2022 20:52:31

748 Index
landscape pattern lignotubers, 226

effects on disturbance spread, 627 lipids, 457
interactions, 634–640 live biomass accumulation of plants, 467
large disturbances in forests, 628f livestock influence on forest ecosystems, 288–290
land‐sparing, 649 live vegetation removal effect, 231
landtype association (LTA), 263 loblolly pine, 393
lapse rate, 170 local genecological differentiation, 55
latent buds, 74 local landscape ecosystems, 30. See also regional
lateral bud primordia, 74 landscape ecosystems
lateral transfers, 635 regionalization, 30
latewood formation control, 114–115 in Upper Michigan, 30–32, 31f
latitude, 344–346 locally dispersed seeds, 71t
layering, 39 local scales, diversity at, 346–351
leaching, 525–526, 526t location‐specific ecosystems, 413
leaf absorption, 125, 126f lodgepole pine forests, 579, 580f, 631, 631f
leaf and root litter production, 506–508, 507f, 508t logging, 407–408
leaf area and productivity, 635 long shoots of plants, 98
leaf area index (LAI), 126, 127–129, 635, 636f long‐term forest ecosystem and vegetation change, 367.
leaf C gain See also forest ecosystems
light and, 452–454 change before Pleistocene age, 367–368
photosynthesis, dark respiration, and, 450–452 eastern North America, 370–373, 372f
soil nitrogen availability and, 456, 462f independent migration and similarity of
temperature and, 454–455, 454f communities, 379–380
water and, 455, 455f last glacial maximum, 370–375
leaf curling, 160 Mt. LeConte, Great Smoky Mountains National
leaf drooping, 160, 161 Park, 374f
Leeuwenberg’s model of trees, 98f patterns of tree genera and species
legacy creation event, 425 migrations, 375–379
legumes, 497 Pleistocene glaciations, 368–369
level terrain, 170–171 western North America, 373–375
lifeboating, 361 lumbering, 407
light, 124, 452–454
and C allocation, 461 M
canopy structure and leaf area, 125–130 MacArthur’s broken‐stick model, 335
distribution reaching ecosphere, 124 macroecosystem, 26t
and ecosystem change, 141 macrolevel landscape ecosystems, 17, 20
energy, 123 macronutrients, 206, 206t, 207
and epicormic sprouting, 139 biochemical function of, 207t
and growth of trees, 133–136 maintenance respiration, 456–458
insolation distribution, 124, 124f mammals, 282–284
photocontrol of plant response, 139–141 mammals, plant defense against, 276–277
photosynthetic light saturation curves, 134f managed relocation. See assisted species migration
photosynthetic response to changing light mapping, 25
intensity, 453f in Baden‐Württemberg, 257–258
plant interception of radiation, 125 ecosystem units at multiple scales, 26
quality beneath forest canopy, 130–133, 131f of landscape ecosystems, 237, 246
relative irradiance, 125f Massart’s model of trees, 98f
respiration, 133 mass flow, 499
saturation point, 134 mast flowering, 76
and seedling survival and growth, 136–137 mast(ing), 76, 280
and tree morphology and anatomy, 137–138 matric potential of soil water, 204
light compensation point, 133 matrix, 618, 619f
Light Detection and Ranging (LiDAR), 172 mature forest stage, 426
light to far‐red light ratio (R/FR ratio), 140 McClure’s model of trees, 98f
lignin, 517 mean annual precipitation, 473, 473f
bindex.indd 748 11-14-2022 20:52:31

Index 749
mean annual temperature, 473, 473f mountainous physiography, 176

merging forest communities, 303. See also discrete comparison of physiographic provinces, 178–179f
forest communities and forests of central Appalachians, 180–183
Eastern Deciduous Forest, 303–304 interfingering of vegetation zones, 180f
New England, 304 mountainous terrain of California and Pacific
topographic distribution of vegetation types, 304f Northwest, 176–180
transitional middle‐aged stands, 305t physiographic diagram of Sierra Nevada Mountains, 177f
mesarch succession, 420, 421t multiple‐factor methods of site and ecosystem
mesic forests, 88, 563 classification, 256
mesoecosystem, 26t applications of multi‐factor methods in United
mesolevel landscape ecosystems, 17 States and Canada, 258–264
mesophication, 590 ecosystem classification and mapping in Baden‐
metamorphic rocks, 195, 197t Württemberg, 257–258
metapopulation theory, 618 landscape ecosystem types, 259f, 260f, 261f, 262f
meteorites, 220 multiple‐scale approaches, 238
methane (CH4), 540 multiple spatial scales, landscape ecosystems at, 25–26.
micelles, 200 See also spatial scales of landscape ecosystems
microbial abundance, 232 hierarchical ecosystem classification, 26t
microbial activity, 575 local landscape ecosystems, 30–32, 31t
microclimate, 147, 149f regional landscape ecosystems, 26–30
microecosystem, 26t mutualisms in forest ecosystems, 307
microlandforms, 191–194 nonsymbiotic mutualisms, 307–308
microlevel landscape ecosystems, 17 symbiotic mutualisms, 307
microrefugia, 378 mycorrhiza, 307, 504–506
microtopography, 237
pit‐and‐mound, 191–193, 193f N
and regeneration in hardwood swamps, 193–194 national classification systems in United
migration of forest trees, 367 States, 262–263
from glacial microrefugia, 377–379 National Hierarchical Framework of Ecological Units
irregularities and disturbance, 377 (NHFEU), 262
routes of tree species in eastern North America, 376f natural disturbances, 381, 385–386, 481
sequence and patterns, 371–373 natural hybridization, 61
and similarity of communities, 379–380 naturalized species, 590, 593
Millennium Ecosystem Assessment (MEA), 644 natural plant distributions
mineral cycling, 285–286 and cold hardiness, 162–164
minerals in ash, 233 freezing resistance, 163f
mineral weathering, 488, 493f natural range of variability. See historical range of
in ecosystems, 489–490 variability (HRV)
nutrient inputs, 491 natural selection, 38–39, 57
rates, 489, 492t net ecosystem carbon balance (NECB). See net
mini‐rhizotron system, 463, 463f ecosystem productivity (NEP)
mitigation, 581 net ecosystem productivity (NEP), 469, 479–480,
mixed fire regime, 220 531, 532f, 637
mixed‐severity fires, 220 net photosynthesis, 450, 452, 456
modern refugia, 584 net primary productivity (NPP), 232, 467, 487, 574
modular unit of construction, 99 allocation, 484t
moisture availability, 554 estimates of total plant, construction, and
moisture gradients, 303 maintenance respiration, 468t
molecular sieve, 203 hypotheses, 483f
monoclimax, 419 net secondary production, 467, 469
monoecious condition, 73 New England, merging forest communities in, 304
monomorphic gene, 39 niche, 60
montmorillonite, 208, 209 differentiation, 58, 60, 305
Morista–Horn coefficient, 336 spatial component of, 60
mosaic ofpatch types, 614 temporal component of, 60
bindex.indd 749 11-14-2022 20:52:31

750 Index
nitrification, 523–524, 638 streamwater nitrate concentrations, 533f

nitrogen (N), 456, 487 transformation of N by soil microorganisms, 527f
availability and belowground net primary nutrition, 274–275, 388
productivity, 482–485
availability in forest ecosystems, 523 O
immobilization and mineralization, 521–523 oaks
leaf N concentration, 456f black oak, 255f
leakage, 606 northern red oak, 553f, 590
nitrogen‐fixing actinobacteria, 497 red oak, 295
removal, 615 at risk, 446–447
nodes, 99 old‐growth stage of forest trees, 426
nongenetic variation, 38 old forest stage. See old‐growth stage of forest trees
non‐indigenous non‐native species, 590 on‐plant seed storage, 227
nonlethal understory fires, 220 open system of plant growth, 99
non‐plant invasive species in forests, 596 organic matter, 234–236
cumulative detections of established forest pests decomposition, 8, 513–524
over time, 596f heating and burning of, 233
forest insects and diseases, 597t removal by fire, 231, 234–236
geographical variation, 598f of soil, 212–213
tree pathogens, 599 organic soils, 146
nonsymbiotic mutualisms, 307–308 orographic precipitation, 170
North American ash, 596, 598 orthotropic models of trees, 97
Norway maple, 592–593 osmosis, 203
Nostoc, 497 osmotic potential of soil water, 204
nothospecies, 61 overstory biomass, 466, 476, 638
novel ecosystems, 610–612 overstory composition, 313–314
nurse logs, 193 overtopped crown classes, 309
nutrient overyielding effect, 356
dynamics, 638–640 oxisols, 214
in forest floor, 511–513, 512t oxygen (O), 487
mineralization, 513–524
nutrient‐use efficiency of litter, 509, 510f P
retranslocation, 508–511 paleogeography, 343–344
storage in boreal, temperate, and tropical forests, Pando clone, 92
528–529, 529t parent material, 167, 195
supply of, 209–210 consolidated, 195–196, 196f
transport to roots, 499–500 in relation to landform, 173
uptake and assimilation by roots, 500–503, 501f unconsolidated, 195–197, 196f
nutrient additions to forest ecosystems, 487 patch(es), 613–614
accumulation of N in forest ecosystems, 498t in forest ecology, 618–619
atmospheric deposition, 492–495 patch‐based ecological theory, 618
biological fixation of nitrogen, 495–498 patch‐based theory, 618
flow of nutrients, 488f patch–corridor–matrix model, 618
mineral weathering, 488–492 size, 615
nutrient cycling, 487, 647 pathogens, 570–572
ecosystem C balance and retention and loss of pathway analysis, 609
nutrients, 531–532 peatlands, 575
within forest ecosystems, 499–524 peds. See aggregates of soil
forest harvesting and nutrient loss, 532–535 people–planet–profit model, 646
nutrient additions to forest ecosystems, 487–498 percentage similarity (PS), 335, 339
nutrient loss from forest ecosystems, 525–528 percent base saturation of soil, 210
and storage of nutrients in forest periodicity of woody plant seed crops, 76–78, 77f
ecosystems, 528–531 permafrost, 398, 575
nutrient loss from forest ecosystems, 525, 532f permanent wilting point, 204
denitrification, 526–528 persistent juveniles, 70, 71t
leaching, 525–526 perturbations, 382
bindex.indd 750 11-14-2022 20:52:31

Index 751
phases of tree species, 249 pinyon ips bark beetle, 572

phenological/phenology, 550 pioneer species. See early‐successional species
escape, 552 pit‐and‐mound microtopography, 191–193, 193f
of flowering, 56 plagiotropic models of trees, 97
separation, 58 plant
phenotype, 11, 35 formation, 296
green, 57 hybrid zones as reservoirs for insect
plasticity of, 36–37 diversity, 275–276
relationship to genotype and environment, 35–36, 36f interception of radiation, 125
phenotypic plasticity. See plasticity of phenotype physiognomy, 23
phloem, 111 succession, 417
photocontrol of plant response, 139–141 plant association, 297
photoperiodism, 43, 46, 139 at climax, 250
photosensory systems of plants, 123 and habitat types in western United States, 247–250
photosynthate, 78 plant defense, 269
photosynthesis, 123, 151, 450–452, 548–549, 635 hypothesized relationship of plant investment into
photosynthetically active radiation (PAR), defense, 271f
124, 129, 130 against insects, 272–276
photosynthetic capacity of trees, 450, 451–452t investment in, 270–272
phototropism, 97 against mammals, 276–277
phyllosilicate clays, 490 relationship of allocation, 272f
phyllosilicate minerals, 200, 208 plant growth, temperature and, 150
building blocks of, 208f cold injury to plants, 153–154
clays, 201, 208 dormancy, 155–156
physical and chemical properties of, 208t external influences and internal interactions, 155f
physical control for invasive species, 609 frost hardiness and cold resistance, 156–161
physical legacies, 386 growth resumption, 161–162
physiognomy, 3 heat damage in seedlings, 151–152
physiographic climax, 431 phases of temperature regime, 152
physiographic/physiography, 12, 20, 167–168, 346–349. photosynthesis role, 151
See also landforms seasonal pattern of cold resistance, 156f
of broadscale ecosystem, 25, 26 stems, roots and leaves, 150–151
characteristics, 168–169 thermoperiod, 152–153
cross section in Upper Peninsula of Michigan, 215f winter chilling and growth resumption, 161–162
eastern North American map of deciduous forest plant life history, 277–288
regions, 22f plant litter
and firebreaks, 190–191 biochemical constituents of, 514–516, 516f
flatlands, 183–190 and return of nutrients to forest floor and soil, 506–511
landforms, 168, 170–174, 186 plant species
landscape development since deglaciation, 169f addition of, 412
effect of local topography, 175f availability and arrival sequence of, 428
microlandforms and microtopography, 191–194 elimination, 411–412
mountainous, 176–183 groups of ground cover, 245–247
multiple roles of, 174–176, 176f migrations, 375–379
regions of North America, 21f rescue. See assisted species migration
setting, 169–170 plasticity
site types, 263 in leaf anatomy, 138
physiological amplitude of species, 305, 306f of phenotype, 36–37
physiological optimum range of species, 305, 306f Pleistocene glaciations, 368
phytochrome, 140–141 maximum extent of continental glaciers in North
phytometers, 244 America, 369f
phytosociology in Europe, 298 temperature data determined from Vostok ice
pignut hickory, 378–379, 379f core, 369f
pines point diversity, 332
in New England and Lake States, 389–390 point processes, 635
pine savannas of western Great Lakes Region, 185 pollen analyses, 375
bindex.indd 751 11-14-2022 20:52:32

752 Index
pollination, 644 R
animal, 277 races of trees, 42
of woody plants, 75–76 local races, 55
polyclimatic climax theory, 432 ponderosa pine, 51f, 52f
polyclimax theory, 431–432 rain shadow, 170
polygenic genes, 38 ramets, 71, 90, 92
polymerase chain reaction (PCR), 40 rapid foliage decomposition, 226
polymorphic gene, 39 rapid juvenile growth, 225
polymorphic patterns, 241 rare species, 353–354
polyploids, 63–64 Rauh’s model of trees, 98f
polyploidy, 63–65 recruitment process, 70
ponderosa pine forests, 3, 611 red maple, 590
average frequencies of two‐needle fascicles on, 53f red to far‐red ratio (R/FR ratio), 130
disturbance and regeneration, 55 reflectance, 125, 126f
eastern geographic races of, 51f refugia, 378, 583–585
genetic differentiation of, 49–52 regionalization, 25, 258
interval between fires, 222 ecological, 26
landscape ecosystems, 262 hierarchical levels of North America, 27
monoterpene components in xylem resin of, 54f of local landscape ecosystems, 30
secondary succession following fire in, 439, 440f regional landscape ecosystems, 26–28. See also local
western geographic races of, 52f landscape ecosystems
populations, 38n1 ecoregion map for North America, 27f
diversity in response to genetic differentiation, 57 of Michigan, 28–30, 29f
of modular units, 71 regression tree analysis (RTA), 559–560
post‐disturbance spatial heterogeneity, 648 regulating services, 644, 645t
post‐dormancy, 155 regulatory control for invasive species, 609
post‐zygotic barriers, 58t, 59 reiteration, 97
potential evapotranspiration (PET), 474, 545 relay floristics, 422
prairie landscapes, 167 reorganization. See stand initiation
pre‐dormancy, 155 reproductive control for invasive species, 610
predominant height, 240 reproductive cycles of woody plants, 74–75
Pre‐European colonization forest communities, 186f reptiles, 278
preforest stage, 425 resilience of forests, 383, 384f, 569, 573, 583
press disturbances, 382–383 resistance of forests, 383, 384f
pre‐zygotic barriers, 57–58, 58t resisters, 227, 228
primary productivity, 467, 574–575 resource‐availability hypothesis, 271
primary succession, 414, 420–423. See also secondary resource‐ratio hypothesis, 428
succession respiration, 123, 133, 548
on deglaciated terrain, 434–437 retranslocation, 508
stages, 421t return interval, 220
primer of invasive species management in forests, 607 rhizomes, 226
early intervention strategies, 609 rhododendrons, thermotropic movements in, 159–161
management approaches, 609–610, 611t richness of species, 331–332
novel ecosystems and invasive species, 610–612 ring‐porous angiosperms, 112, 113f
timeline for invasion, 608f riparian buffers, 615
proleptic shoots of plants, 102 riparian zones, 187
proportional diversity, 332 risk assessment, 609
provenance test, 41 riverine ecosystems, 342
provisioning services, 644, 645t root architecture, 503–504
pulse disturbances, 382 root collar or crown, 226
root development of seedlings, 136
Q root pressure, 115
qualitative defense chemicals, 270 roots of trees, 102–103
quaternary biogeography, 370 fine root relations, 105–107
quaternary paleoecology, 370 horizontal and vertical root development, 107–108
bindex.indd 752 11-14-2022 20:52:32

Index 753
kinds, forms, and occurrence, 103–105 semi‐natural landscapes, 647

modification of root systems, 104f senescing, 423
periodicity of primary root growth, 108 seral stage, 421
recovering from fire damage, 226 sere, 421
root development of seedlings, 103t series order habitat types, 249
root grafting, 108–110, 109f serotinous cones, 81
specialized roots and buttresses, 110 severity of fire, 220, 221, 222
temperature of, 150 sexual reproduction of woody plants, 70, 71, 388
rule of tens, 590 correlational studies, 78, 79f
distribution of seeds of Engelmann spruce, 80f
S effects on vegetative growth, 78–80
sand dunes, 302 establishment following, 83–90
sapwood, 111 flowering, 72, 73, 75
saturation water content, 205 increasing seed production, 73
scale, 15. See also spatial scales of landscape maturation and ability to flower, 71–72
ecosystems; temporal scales, landscape periodicity of seed crops, 76–78, 77f
ecosystems in pollination, 75–76
scatter‐hoarding, 280 post‐establishment development, 90
scots pine, 49–50, 463, 464f reproductive cycles, 74–75
scrub “islands”, 353, 354 seed bank, dormancy, and germination, 82–83, 82f
seasonality of fire, 220 seed dispersal, 80–82
secondary chemicals, 270, 282 sexual systems, 39
secondary foundation species, 352 shade‐tolerant species, 88, 129, 590
secondary growth of trees, 2, 110, 112–115 Shannon–Wiener index, 333–334, 629
secondary minerals, 200 shape of landforms, 170
secondary succession, 414, 423–427. See also primary short‐interval fires, 569
succession shortleaf pine, 393
fire and oak dominance, 446–447 short shoots of plants, 98
fire in ponderosa pine forests of Western short stature, 227
Montana, 439 silica tetrahedra structure, 208, 208f
and gap dynamics, 439–446 silviculture, 13
stages, 423t similar effects, principle of, 187
stand development stages, 424f simulation model, 579, 623
sedimentary rocks, 195, 197t single‐gene mutations, 38
seed bank of woody plants, 82–83, 82f site, 12
seedbeds, 388 climax, 432
seed dispersal, 278 districts, 263
by birds, 279–281 index, 239
effectiveness, 280 regions, 263
effects of personality on seed dispersal distance, 283f site‐index curves, 239f, 240–243
by fish and reptiles, 278 site quality and ecosystem, 237
in forest trees, 56–57 advantages and limitations, 243
by mammals, 282–284 comparisons between species, 243
of woody plants, 80–82 direct measurement of forest productivity, 238–239
seedlings, 554 environmental factors as measure of site, 251–256
of gap‐phase and shade‐tolerant species, 88 multiple‐factor methods of site and, 256–264
heat damage in, 151–152 tree height as measure of site, 239–243
height and top/root weight ratio, 137t vegetation as indicator of, 244–251
root development of, 136 slash pine, 393
of species, 85 slope
sugar maple, 88, 90 aspect, 172–173
seed production, 226 characteristics, 171
of woody plants, 76, 77–78 inclination, 171, 173
seed‐source test, 41 length, 171
self‐thinning. See density‐dependent competition position, 172, 172f
bindex.indd 753 11-14-2022 20:52:32

754 Index
sloping terrain, 171 average number of plant species in local floras, 330f
snowpack, 546 causes of, 343
soil‐moisture availability, 554 at continental and subcontinental scales, 343–346
soil N availability, 484f and disturbance, 350f
and belowground net primary productivity, 482–485 estimates of relative proportion of groups of
leaf litter and fine‐root production, 485f organisms, 328f
soil(s), 195, 346–349 increase of biological diversity over geological
acidity, 209, 210–212 time, 329f
cation exchange capacity of, 209–210, 210t moist temperate forest trees in Northern
chemical properties of, 206, 207 Hemisphere, 343t
clay mineralogy, 207–209, 210t patterns of plant species’ richness in Siskiyou
color, 202 Mountains, 347f
formation, 197 plant species with increasing elevation in European
improvement, 285–286 Alps, 346f
landscape relationships, 215–216 predicted number of species along forest
microbes, 231 succession, 351f
nitrogen availability, 456, 461–465, 462f species richness of North American trees, 345f
non‐wettability before, during, and after fire, 235f threats to, 331
organic matter, 212–213 value of biodiversity, 329–331
parent material, 195–197 spodosols, 214, 216
profile development, 197–199 spongy moth, 386, 602, 603, 606
properties, 200, 475–476 spring vessel formation, 115–116
structure, 201–202 sprouting, 226
supply of nutrients, 209–210 spruce‐fir forests, 563
surveys, 256 spur shoots of plants, 98
taxonomy, 213–215 stability of ecosystem, 383
temperature, 145–147 stable state of ecosystem, 383
texture, 200–201, 201f, 205, 206f, 210t, 476 stand‐replacing fires, 220, 221, 387, 638
water, 202–6 stand‐replacing wildfires in, 629–632
soil‐site studies, 252–256 stand density, 312–313
soil‐water stand habit, 226
potential, 204 stand initiation, 425
stress effects, 116 starch, 515
solar radiation, 124 steady‐state stage. See old‐growth stage of forest trees
interception of, 125, 126 stems of trees, 110–111
plant interception of, 125 analysis, 243
Sorensen similarity coefficients, 335 earlywood and latewood formation control,
southeastern and southern Coastal Plain, 186 114–115
southern Appalachians elongation, 37
ecosystem change in, 373 exclusion, 426
merging forest communities in, 303–304 periodicity and control of secondary growth, 112–114
southern Illinois, 294–296 recovering from fire damage, 226
southern pines, 393, 563 temperature of, 150
space‐for‐time substitution, 416 winter freezing and water transport, 115–116
spatial heterogeneity, 613, 648 xylem cells and growth rings, 111–112
spatial scale of physiography, 169 stratification, 161, 162
spatial scales of landscape ecosystems, 15–17. See also stress tolerance hypothesis, 322
multiple spatial scales, landscape ecosystems at stringers, 222
climatic classification, 18–19 strong sustainability, 646–648
physiography, 20–22 subcontinental scales, diversity at, 343–346
spatial variability, 639 subordinate unions, 247
specialized roots, 110 succession, 349–351, 367, 388–389, 414, 416
speciation, 59 autogenic and allogenic, 416
species diversity. See also plant species availability and arrival sequence of species, 428
bindex.indd 754 11-14-2022 20:52:32

Index 755
biological legacies, 414 natural plant distributions and cold

causes, 418t, 427–431 hardiness, 162–164
as ecosystem process, 432–434 observed changes in annual average
end point, 431–432 temperature, 543f
eruption of Mount St. Helens, 437–439, 438f percentage of land area of contiguous United
facilitation, tolerance, and inhibition, 428–431 States, 544f
facilitative and inhibitory effects, 430f and plant growth, 150–162
key characteristics and regeneration at soil surface, 145–147
strategies, 427–428 sum, 46, 46n2, 144
mechanisms, 414–416, 427–431 variation with local topography, 148–150
models, 414–416, 422f, 427–431 temporal scales, landscape ecosystems in, 15–17
pathway, 414–416, 415f, 433f terrain age, 434
primary and secondary, 414 thermal conductivity, 146
stages, 420–427 thermoperiod, 152–153
work, 418–431 thermotropic movements in rhododendrons,
sudden aspen decline, 393 159–161
sugar maple, 39, 444–446, 445f thinning, 388
canopy, 130 till plains of Midwest, 185
seedlings, 88, 90, 312, 319 tolerance mechanism, 428–431, 429f. See also
stand in Upper Peninsula, 321f understory tolerance of species
sunflecks, 130–133, 132f Tomlinson’s model of trees, 98f
supercooling, 159 top‐down approach, 432
supporting services, 644, 645t top height concept, 240
surface backfires, 221 topography, 216. See also microtopography
surface fires, 220, 221–222 temperature variation with local
surface headfires, 221 topography, 148–150
sustainability, 641–642, 646–648 varying leaf phenology related to, 552f
Earth as metaphor for life, 650–651 tornadoes, 401, 628
ecological crisis, 648–650 total soil‐water potential, 204
ecosystem services, 644–645 toxins, 270
of forest ecosystems, 641 tracheids, 112
in forests, 649 transmittance, 125, 126f
historical perspective of sustainability in forests, 643 tree‐fall gaps, 442, 443f
landscapes, 648f tree form, 95
prevalence of human values in forest architectural models, 97–102
ecology, 641–642 climate role, 96
sustainable ecosystem, 642 decurrent form, 96f, 97
sustained yield, 643 excurrent form, 95–96, 96f
sustaining natural resources, 643 woody plant, 95
sylleptic shoots of plants, 102 tree‐line position around globe, 144, 145f
symbiotic mutualisms, 307 tree(s)
sympatric speciation, 57 architecture, 95
genera patterns, 375–379
T light and growth of, 133–136
taxonomy of soil, 213–215 light and tree morphology and anatomy,
temperate forest ecosystem, nitrogen and calcium cycle 137–138
of, 529–531, 530f line, 302
temperature, 12, 143, 454–455, 454f, 540 pathogens, 599
climate effects on, 542 primary growth, 2
deciduousness and, 164 regeneration, 554
within forest, 147–148 roots, 102–110
geographical patterns of, 143–145 stems, 110–116
mean weekly maxima, minima, and mean uprooting, 191–193
temperatures, 147t water deficits and tree growth, 116–117
bindex.indd 755 11-14-2022 20:52:32

756 Index
trees, carbon balance of, 450 V

allocation to structure, storage, and defense, variable fire regime, 220
459–461, 460f variable‐retention harvest system, 360–363, 360f
construction and maintenance respiration, 456–458 vascular plant diversity, 350–351
light and C allocation, 461 vegetation, 299, 413, 417
light and leaf C gain, 452–454 applications and limitations of vegetation, 250–251
photosynthesis, dark respiration, and leaf C change or succession, 16
gain, 450–452 distribution of forest trees, 249f
soil nitrogen availability and C allocation, 461–465 ecological species groups and ecosystem types, 248t
soil nitrogen availability and leaf C gain, 456 height‐over‐age curves of Scots pine, 245f
temperature and leaf C gain, 454–455, 454f as indicator of site quality, 244
water and leaf C gain, 455, 455f operational site classification, 247–250
tree species distributions, climate change effects on, 556 plant associations and habitat types, 247–250
climate envelope richness, 561f reduction of competing, 388
future projections of distribution of sugar maple, 559f species groups of ground cover, 245–247
habitat for current forest types of eastern United types, 20–25
States, 564f units, 16
modeled changes in distance and direction of mean vegetative reproduction of woody plants, 71t
centers, 562f conceptual model of, 91f
observed range shifts, 556–558 by conifers, 92, 93
projected changes in, 558–560 sprouting, 90–91, 94
tree biomass and seedling densities, 557f survival of single genet, 93–94
tree‐species line, 302 trembling aspen clones, 91–92, 92f
trembling aspen, 65, 390–393 vine maple clone, 92, 93f
clones, 91, 92f, 92 vermiculite, 208
ecosystem, 127f vertical sinker roots, 103
juvenile sprouts of, 277 vessels, 112
triple bottom line, 646, 646f vital attributes concept, 428
Troll’s model of trees, 98f volatilization, 233
trophic systems, 269 volcanic eruptions, 628
true dormancy, 155
W
U water, 202, 455, 455f
ultisols, 214 physical properties of, 203
umbrella species, 351 purification, 647
unconsolidated parent material, 195–197, 196f relations and dynamics of, 12
understory intolerant species, 315, 316–317 transport, 115–116
understory reinitiation, 426 vapor, 540
understory tolerance of species, 315, 316–317 watershed protection, 649
competition in, 314–315 wave‐regenerated fir species, 404
environmental factors relating to, 322–323 weak sustainability, 646–647
examples in forest ecosystems, 319 weather(ing), 195, 208, 209, 539
nature, 319–324 western pines, 390–393
photosynthesis and respiration for tolerant and wet deposition, 492, 495
intolerant species, 323t white‐red‐jack pine forests, 563
physiological processes, 323–324 white‐rot fungi, 517
ratings of tree species, 317–319 whitebark pine, 583
of selected North American forest trees, 318t whole genome duplication, 64
unicellular bacteria, 497 widely wind‐dispersed seeds, 71t
United States National Vegetation Classification wilderness areas, 647, 650
(USNVC), 23, 25 wildfires, large, 628
upgrading. See aggrading wildland–urban interface (WUI), 647
bindex.indd 756 11-14-2022 20:52:32

Index 757
wildlife habitat, 389 woody plant, 69, 275

Douglas‐fir in Pacific Northwest, 394, 395f asexual reproduction, 70, 71
fire and ponderosa pine, 392f plant life cycle with trees, 70f
giant sequoia, 394–395 populations, 62
lodgepole pine regeneration, 390f regenerative strategies, 71t
natural longleaf pine seedlings, 394f sexual regeneration, 72f
pines in New England and Lake States, 389–390 sexual reproduction, 70–90
southern pines, 393 vegetative reproduction, 90–94
western pines and trembling aspen, 390–393
wind, 400, 572–573 X
broadscale disturbance by hurricanes, 403–404 xerarch succession, 420, 421t
principles of wind damage, 401–403 xylem cells, 110
widespread and local effects, 401 small‐diameter, 115
wind‐pollinated species, 75 of stem, 111–112
windthrow of standing trees, 103, 402f, 403
winter Y
chilling, 161–162 Yellowstone National Park, 628–629
freezing, 115–116 fire, carbon, and climate change in forests
temperatures, 572 of, 579–581
bindex.indd 757 11-14-2022 20:52:32

WILEY END USER LICENSE AGREEMENT
Go to www.wiley.com/go/eula to access Wiley’s ebook EULA.

Forest Ecology

Uploaded by

Copyright:

Available Formats

Forest Ecology

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Forest Ecology

Uploaded by

Copyright:

Available Formats

Forest Ecology

Forest Ecology FIFTH EDITION

Burton V. Barnes (deceased)

Stephen H. Spurr (deceased)

Limit of Liability/Disclaimer of Warranty

Library of Congress Cataloging-­in-­Publication Data

Cover Design: Wiley

Set in 9.5/12pt STIXTwoText by Straive, Pondicherry, India

PA R T 1 Forest Ecology and Landscape Ecosystems

1 Concepts of Forest Ecology 3

2 Landscape Ecosystems at Multiple Scales 15

PA R T 2 The Forest Tree

3 Forest Tree Variation 35

Periodicity of Seed Crops, 76

5 Tree Structure and Growth 95

PA R T 3 The Physical Environment

0005430834.INDD 9 01-05-2023 17:43:22

Position on slope, 172

11 Site Quality and Ecosystem Evaluation

0005430834.INDD 12 11-23-2022 12:32:40

Ecosystem Classification in the Southwestern United States, 262

12 Animals in Forest Ecosystems 269

13 Forest Communities 291

0005430834.INDD 13 11-23-2022 12:32:40

Phytosociology in Europe, 298

14 Diversity in Forests 327

0005430834.INDD 14 11-23-2022 12:32:40

Diversity of Landscape Ecosystems, 336

PA R T 5 Forest Ecosystem Dynamics

15 Long-­Term Forest Ecosystem and Vegetation Change 367

0005430834.INDD 15 11-23-2022 12:32:40

Eastern North America, 370

0005430834.INDD 16 11-23-2022 12:32:41

Biotic Composition Changes, 411

17 Forest Succession 413

18 Carbon Balance of Trees and Ecosystems 449

0005430834.INDD 17 11-23-2022 12:32:41

Temperature and Leaf C Gain, 454

19 Nutrient Cycling 487

0005430834.INDD 18 11-23-2022 12:32:41

PA R T 6 Forests of the Future

20 Climate Change and Forest Ecosystems 537

0005430834.INDD 19 11-23-2022 12:32:41

21 Invasive Species in Forest Ecosystems 587

22 Forest Landscape Ecology 611

0005430834.INDD 20 11-23-2022 12:32:41

23 Sustainability of Forest Ecosystems 639

Scientific Names of Trees and Shrubs 731

0005430834.INDD 21 11-23-2022 12:32:41

F orest ecologists work to understand the dynamics, structure, and function of

F I G U R E 1 . 1 The three-­dimensional, volumetric nature of a landscape ecosystem. Ecosystems comprise the

Upper Air Meteorology

Lower Air Topoclimatology

Mineral Strata Bedrock Geology

Water Table Hydrology

LANDSCAPE ECOSYSTEM AND COMMUNITY

ECOSYSTEM STRUCTURE AND FUNCTION

dead living material or energy flow other influences

Humans as a partner *relations omitted

EXAMPLES OF LANDSCAPE ECOSYSTEMS

Library of Congress Cataloging-in-Publication Data

1 Concepts of Forest Ecology 3

2 Landscape Ecosystems at Multiple Scales 15

3 Forest Tree Variation 35

5 Tree Structure and Growth 95

12 Animals in Forest Ecosystems 269

13 Forest Communities 291

14 Diversity in Forests 327

15 Long-Term Forest Ecosystem and Vegetation Change 367

17 Forest Succession 413

18 Carbon Balance of Trees and Ecosystems 449

19 Nutrient Cycling 487

20 Climate Change and Forest Ecosystems 537

21 Invasive Species in Forest Ecosystems 587

22 Forest Landscape Ecology 611

23 Sustainability of Forest Ecosystems 639

Scientific Names of Trees and Shrubs 731

F I G U R E 1 . 1 The three-dimensional, volumetric nature of a landscape ecosystem. Ecosystems comprise the

continuous. By contrast, a discontinuous pattern may result if the species’ distribution or

F I G U R E 3.1 0 Average frequencies of two-needle fascicles on (mature/young) native ponderosa pines

F I G U R E 3 .1 2 Hypothetical example of introgression—the interbreeding of species followed by backcross-

material—the process is often called adaptive introgression (Suarez-Gonzalez et al. 2018).