Geomorphology For Engineers

GFE-Prelims.
qxd 6/27/05 2:35 PM Page i
Geomorphology for Engineers

GFE-Prelims.qxd 6/27/05 2:35 PM Page ii
GFE-Prelims.qxd 6/27/05 2:35 PM Page iii
Geomorphology for Engineers
Edited by
P. G. Fookes,
E. M. Lee and
G. Milligan
Whittles Publishing
CRC PRESS
GFE-Prelims.qxd 6/27/05 2:35 PM Page iv
Typeset by
Mizpah Publishing Services Private Limited, Chennai, India
Published by
Whittles Publishing,
Dunbeath Mains Cottages,
Dunbeath,
Caithness KW6 6EY,
Scotland, UK
www.whittlespublishing.com
Distributed in North America by

CRC Press LLC,
Taylor and Francis Group,
6000 Broken Sound Parkway NW,
Suite 300,
Boca Raton, FL 33487, USA
© 2005 P. G. Fookes, E. M. Lee and G. Milligan
ISBN 1-870325-03-6
USA ISBN 0-8493-9641-7
Every effort has been made to correctly acknowledge illustrative material. The publishers
ask that any omissions be reported to them for incorporation in future editions.
Printed and bound in Poland, EU

GFE-Prelims.qxd 6/27/05 2:35 PM Page v
Contents
Foreword xi 2.14 Weathering and geohazards 53

Preface xiii 2.15 Summary 54
Dedication xv References 55
Biographies xvii
3. Sedimentology
1. Introduction to Engineering Colin J. R. Braithwaite
Geomorphology
Peter Fookes and Mark Lee 3.1 Introduction 57
3.2 Weathering 57
1.1 Preface 1 3.3 Sediment transport and deposition 58
1.2 Geomorphology, landforms and 3.4 Sedimentary rocks 74
the nature of modern engineering 3.5 Sedimentary facies and
geomorphology 4 environments 82
1.3 Techniques in engineering 3.6 Summary 83
geomorphology 16 References 83
1.4 Summary 26 Further reading 83
References 26
Further Reading 26 4. Tectonics
Tony Waltham
PART I CONTROLS 4.1 Geological structures 85

4.2 Tectonic landforms 97
2. Climate and Weathering 4.3 Lithology and landforms 102
Mark Lee and Peter Fookes 4.4 Summary 106
References 106
2.1 Climate 31 Further reading 106
2.2 Tropical climates 31
2.3 Temperate climates 33 5. Stratigraphy
2.4 Polar climates 35 Colin J. R. Braithwaite
2.5 Climate classification 37
2.6 Climate variability 41 5.1 Introduction and basic principles 107
2.7 Extreme events 41 5.2 Correlation 109
2.8 Climate and flooding 42 5.3 Bedding and time 109
2.9 Global climate change 44 5.4 Three kinds of stratigraphy 111
2.10 The hydrological cycle 44 5.5 Non-radiometric chronology 113
2.11 Weathering 47 5.6 Seismic stratigraphy 114
2.12 Weathering and landscape 47 5.7 The geological time scale 114
2.13 Weathering products: 5.8 Correlating and interpreting
engineering soils 52 sequences 116
v
GFE-Prelims.qxd 6/27/05 2:35 PM Page vi
vi CONTENTS
5.9 Geological maps 121 8.7 Landslides: susceptibility,

5.10 Summary 122 hazard and risk 198
References 122 8.8 Planning and undertaking a
landslide investigation 205
8.9 Landslide stabilisation 206
6. The Quaternary
8.10 Conclusion 214
Andrew Goudie
References 215
6.1 Introduction 124 Acknowledgements 217
6.2 The late Cainozoic 124
6.3 Abrupt change 126
9. Active Tectonic Environments
6.4 The last glacial cycle 129
and Seismic Hazards
6.5 Interglacials 130
James V. Hengesh, and
6.6 Holocene 130
William R. Lettis
6.7 The Quaternary in low latitudes 131
6.8 Sea-level changes 131 9.1 Introduction 218
6.9 Summary 134 9.2 Earthquake occurrence and
References 135 magnitude scaling 220
Further reading 136 9.3 Types of fault displacement 223
9.4 Seismic source characterisation 232
7. Engineering Behaviour of Soils 9.5 Probabilistic seismic hazard
and Rocks assessment and treatment of
George Milligan, Peter Fookes and uncertainty 251
Mark Lee 9.6 Liquefaction and related
ground deformation 253
7.1 Introduction 137 9.7 Implications for engineering 259
7.2 Elements of behaviour 138 References 259
7.3 Slope instability 149
7.4 Geotechnical problems associated
with different soil types 153 10. Rivers
7.5 Typical values of soil properties 163 Mark Lee
7.6 Closing remarks 167
10.1 Introduction 263
References 168
10.2 River channel changes 263
Further reading 169
10.3 Channel form: regime theory 266
10.4 Planform types 270
10.5 Bank erosion and instability 273
PART II GEOMORPHOLOGICAL
10.6 Bed scour 274
PROCESSES
10.7 Floodplain accretion 277
8. Landslides 10.8 River terraces: phases of
James S Griffiths aggradation and incision 278
10.9 Floods 281
8.1 Introduction 173 10.10 Sediment transport 283
8.2 Description 176 10.11 The significance for conservation 284
8.3 Landslide cause 178 10.12 Soft engineering and
8.4 Landslide recognition 180 river restoration 284
8.5 Landslide activity and behaviour 180 10.13 Summary: the catchment
8.6 Landslide distribution and approach to river engineering 284
occurrence 186 References 285
GFE-Prelims.qxd 6/27/05 2:35 PM Page vii
CONTENTS vii
11. Soil Erosion 13.3 Glacial erosion and entrainment

Mark Lee and John Charman of sediment 349
13.4 Depositional processes and
11.1 Introduction 287 classification of tills 351
11.2 Types of erosion features 288 13.5 Contemporary glacial
11.3 Detachment: erosivity of rainfall 293 environments: land system
11.4 Transport: running water 293 models 353
11.5 Transport: erosivity of wind 296 13.6 Engineering in glacial
11.6 Resistance to erosion: soil environments 361
erodibility and plant cover 297 13.7 Conclusions: recommendations
11.7 Erosion hazard assessment 301 to engineers working in glacial
11.8 The Universal Soil Loss Equation 307 or glaciated environments 372
11.9 Erosion control methods and References 372
materials 307
11.10 Summary 316
References 316
14. Periglacial Forms and
Processes
H. Jesse Walker
12. Subsidence
Tony Waltham 14.1 Introduction 376
14.2 Permafrost distribution 376
12.1 Subsidence environments 318 14.3 Periglacial processes 378
12.2 Subsidence on clay 318 14.4 Ground ice and permafrost 382
12.3 Subsidence on peat 324 14.5 Periglacial landforms 386
12.4 Hydrocompaction of 14.6 Periglacial soils 389
collapsing soils 326 14.7 River ice, lake ice, and sea ice 391
12.5 Subsidence on permafrost 327 14.8 Engineering and fossil
12.6 Subsidence on limestone and periglacial forms 392
gypsum 329 14.9 Engineering and present-day
12.7 Salt subsidence 329 permafrost environments 394
12.8 Volcanic subsidence 331 14.10 Oil industry 395
12.9 Implications for engineering on 14.11 Summary and future
naturally cavernous ground 332 challenges 396
12.10 Subsidence due to mining 335 References 397
12.11 Tectonic subsidence 339 Further reading 399
12.12 Summary 341
References 341
Further reading 342
15. Temperate Environments
Kenneth J. Gregory
15.1 Deposits, sediments and the

PART III ENVIRONMENTS AND landscape 400
LANDSCAPES 15.2 Water movement 412
13. Glacial Environments 15.3 Human activity 415
Lewis A. Owen and 15.4 Summary 417
Edward Derbyshire References 417
Further reading 418
13.1 Introduction 345
13.2 Physics of glaciers 348
GFE-Prelims.qxd 6/27/05 2:35 PM Page viii
viii CONTENTS
16. Hot Drylands 18.8 Slope instability and landsliding 496

Mark Lee and Peter Fookes 18.9 Extreme events and major
geomorphological problems in
16.1 Introduction 419 hot wetlands 496
16.2 Desert landscapes 419 18.10 Summary 498
16.3 Zone I: the uplands 421 References 498
16.4 Zone II: footslopes and fans 423 Further reading 500
16.5 Zone III: plains 428
16.6 Zone IV: base level plains 434
16.7 Surface erosion and instability 437
19. Mountain Environments
John Charman and Mark Lee
16.8 Excavation 438
16.9 Behaviour of desert soils 438 19.1 Introduction 501
16.10 Aggregates 439 19.2 Rocks, climate and weathering 502
16.11 Water and sediment problems 439 19.3 Mountain terrain model 505
16.12 Dune mobility 443 19.4 Access and routing 506
16.13 Aggressive salty ground 446 19.5 Large landslides 508
16.14 Summary 450 19.6 Debris flows and torrents 513
References 451 19.7 Snow melt and flash floods 514
19.8 Landslide dams 517
17. Savanna 19.9 Glacial lake outburst floods 519
Michael F. Thomas 19.10 Snow avalanches 524
19.11 Ice avalanches 527
17.1 Introduction to savanna 19.12 Construction materials 528
environments 455 19.13 Summary 529
17.2 The nature of the regolith in References 532
savanna areas 456
17.3 Landform assemblages and
surface materials 460
20. Estuaries and Deltas
Warren E. Grabau, H. Jesse Walker
17.4 Surface processes and
and Mark Lee
geohazards in the savanna 465
17.5 Summary 471 20.1 Estuaries and deltas: introduction 535
References 471 20.2 Estuaries 535
Further reading 473 20.3 Deltas 552
Acknowledgements 473 20.4 Depositional environments 556
20.5 Summary 562
18. Hot Wetlands References 563
Ian Douglas Further reading 565
18.1 The tropical rainforest

environment 473
21. Coastal Environments
Julian Orford
18.2 Weathering and soil
characteristics 479 21.1 Introduction 566
18.3 Soil characteristics on 21.2 Perspectives affecting our under-
different rock types 485 standing of the coastal zone 566
18.4 Rivers and alluvial landforms 487 21.3 Process and response controls
18.6 Volcanic terrain 490 in the coastal zone 570
18.7 Karst terrain 492 21.4 Sediments in the coastal zone 576
GFE-Prelims.qxd 6/27/05 2:35 PM Page ix
CONTENTS ix
21.5 Coastal organisation 577 24. Karst Terrains

21.6 Effects of morphology on Tony Waltham
coastal zone structure 585
21.7 Controls on coastal morphology 587 24.1 Soluble rocks terrains 662
21.8 Key clastic coastal depositional 24.2 Surface landforms of karst 663
morphology 587 24.3 Underground features of karst 670
21.9 Conclusion 599 24.4 Types of karst terrain 679
21.10 Summary 599 24.5 Summary 683
References 599 References 686
Further reading 602 Further reading 687
22. Continental Shelves 25. Loess

Colin Jago Edward Derbyshire and Xingmin Meng
22.1 Introduction 603 25.1 Introduction 688

22.2 Morphology of continental 25.2 Loess Terrain 690
margins 603 25.3 Engineering properties of loess 697
22.3 Major controls of shelf 25.4 Key engineering
sedimentation 605 geomorphological issues 711
22.4 Shelf hydrodynamic regime 607 25.5 Conclusions 723
22.5 Properties of shelf sediments 611 References 724
22.6 Organism–sediment interaction 614 Acknowledgements 728
22.7 Patterns of shelf sedimentation 614
22.8 The shoreface and coastal zone
26. Chalk Landscapes
bypassing 615
Fred Bell and Martin Culshaw
22.9 Muddy shelves 617
22.10 Tide-dominated sandy shelves 622 26.1 Introduction 729
22.11 Wave and storm-dominated 26.2 Geomorphology of the Chalklands 730
sandy shelves 625 26.3 Chalk engineering 742
22.12 Preservation of relict features 626 26.4 Summary and conclusions 753
22.13 Preservation of nearshore References 754
marine patterns 627 Acknowledgements 756
22.14 Summary 629
References 629
Further reading 632 27. Urban Geomorphology
Ian Douglas
23. Volcanic Landscapes 27.1 Introduction 757

Tony Waltham 27.2 Problems of siting and planning
urban developments 758
23.1 Plate tectonics and volcanoes 633 27.3 Impact of urban activities on
23.2 Basaltic volcanic landscapes 636 geomorphic processes 768
23.3 Explosive volcanoes 648 27.4 Landform creation by
23.4 Mud volcanoes 658 urbanisation 772
23.5 Summary 659 27.5 Remediation and Prevention 775
References 659 27.6 Summary 776
Further reading 661 References 777
Further reading 779
GFE-Prelims.qxd 6/27/05 2:35 PM Page x
x CONTENTS
APPENDICES Appendix A4 Geomorphological

Mapping 816
Appendix A1 Visual Identification 782 Appendix A5 Soil Erosion Assessment 825
Appendix A2 Identification, Description Appendix A6 Units, Scales and
and Classification of Soils 789 Conversions 834
Appendix A3 Identification, Description
and Classification of Rocks 804 INDEX 843
GFE-Prelims.qxd 6/27/05 2:35 PM Page xi
Foreword
Denys Brunsden OBE, DSc (Hons), FKC

Emeritus Professor,
University of London, King’s College
Ever since the Stone Age, when man hit his from which they are made; predictions about their
thumb with a rock, human beings have of neces- future form, behaviour and status. Engineering geo-
sity been interested in the earth sciences. Our ear- morphology is the application of these geomorpho-
liest activities involved an understanding of the logical techniques and analysis to an engineering or
environments in which we lived, sought our food, environmental management problem.
sheltered and survived. Inevitably, we have The practice of the subject has provided real
observed, contemplated and even worshipped nat- opportunities for fundamental research and many of
ural phenomena. As technology has advanced we the most important discoveries have emerged from
have manipulated, altered, exploited and some- the pragmatic arena. Applied studies have strength-
times irrevocably destroyed the systems upon ened theoretical developments and improved the
which we depend. Humans are rational beings, quality of inventory, mapping, calibration, evalua-
however, who are capable of learning from their tion of hazard calculations and stability studies. The
mistakes and one of our greatest intellectual need for technical competence has involved geo-
achievements has been our growing knowledge of morphologists in the use of advanced techniques
the natural world and our increasing willingness such as scanning electron microscopy, interpreta-
to apply that knowledge to the ‘proper’ manage- tion of satellite imagery for structural geology and
ment of the earth – the sustainable human use of hydrocarbon exploration, side scan sonar for land-
the land, air, living things and water. Central to slide identification on the sea bed, hydraulic geom-
this work is the cooperation between earth scien- etry and discharge of river systems for irrigation,
tists, planners and engineers, a key element being sediment yield monitoring on high mountain rivers
the part that applied geomorphology can make to and other process modelling.
the well being of the earth and its inhabitants. The recognition that human beings were
Geomorphology is the study of the forms of the powerful agents of geomorphological change has
earth’s surface: their origin; the processes involved a long history and the public concern over the
in their development; the properties of the materials economic costs for remedial works, declining
XI
GFE-Prelims.qxd 6/27/05 2:35 PM Page xii
xii FOREWORD
yields and obvious damage led to the establish- standards of the profession, the quality of the con-
ment of conservation services, erosion control sultants and the employment of geomorphologists
boards, official resource surveys and government within the engineering, planning, environmental
research establishments all over the world such as management, insurance and related professions.
the US Soil Conservation Service, the New Attempts have been made to establish a code of
Zealand Soils and River Control Council (now ethics, standards of education and training, repre-
Landcare Ltd) and the Land Research Division of sentation in national and local government,
the CSIRO, Australia. The work was technical and legislative, employment and career structures
professional especially where process monitoring, and to provide a good information dissemination
air photograph interpretation, remote sensing, service.
laboratory analysis, geomorphological mapping A major influence on the development of a
and land surveying were required. This work was professional engineering geomorphology in the
so successful with resource development, engi- UK was recognition by several of Britain’s oldest
neering and Second World War reconstruction and most respected engineering consultancies
programmes that during the 1960s and 70s there who established the first real sponsorships of geo-
was a rapid expansion of activity sponsored by morphology in the UK by retaining the services
government agencies such as ODA or UNESCO. of geomorphologists and by awarding, in 1971,
These fundamental changes to the science of the first engineering geomorphology studentships
geomorphology were accompanied by a rapid at British universities. During the 1970s major
growth in international aid that adopted a knowl- projects were carried out by geomorphologists
edge-based approach to the solution of envi- from many British universities and by students
ronmental problems. Geomorphology was used who gained professional positions. They worked
in the reconnaissance, site investigation and for numerous national governments and most of
construction, operational and decommissioning the major engineering companies in Britain. This
stages of engineering developments. The need of British experience was matched elsewhere, in the
newly independent countries to map their natural USA, Canada, Australia, India and China and
resources and plan their development program- geomorphologists found their way into agencies
mes, the availability of aid money, the occurrence such as the Department of the Environment, FAO,
of natural disasters, the oil-boom development of UNESCO, World Bank, the Environment Agency
desert coasts and the spread of development into and the British Insurance Association.
hostile terrains such as Alaskan permafrost and The growth of the subject in the UK was
Middle East sabkhas – all increased public aware- largely due to the pioneering vision of Professor P.
ness that geomorphology had an important role to G. Fookes who is the real champion of engineer-
play in human development. ing geomorphology. Due to his untiring promotion
The work that evolved was intensely interdis- of the subject it is now standard good practice to
ciplinary. Geomorphology suddenly became an specify and carry out geomorphological studies
essential component of geotechnical education. within a planning, engineering, resource or hazard
Engineers, geologists and geomorphologists assessment programme and to cooperate with
found themselves involved in the same project planners, developers, architects, engineers, politi-
and all went through a rapid, sometimes chasten- cians, lawyers and administrators. He demon-
ing, learning curve. Geomorphologists brought to strated that an essential task is to cooperate in
the geoteam the ability to think in spatial terms, to interdisciplinary teams. To achieve this, a good,
detect spatial correlations and patterns and to clear and well-explained guidebook is an essential.
understand and determine the significance of the Most engineering companies employ geomorphol-
time dimension in the operation of processes and ogists on their fulltime staff and all require a text
the evolution of the landscape. that directs them to the core of the subject. This
Perhaps the most important aspect has been the book fulfils that need and is a fitting testament to
growth of professionalism both in terms of the Peter Fookes and his many collaborators.
GFE-Prelims.qxd 6/27/05 2:35 PM Page xiii
Preface
This book has been written by geomorphologists, due to the limited print run before the demise of
together with some engineers and geologists, and the university press and, in any case, major
is mainly for engineers. The aim was to present a advances have been made in geomorphology in
basic but authoritative handbook of geomorphol- the last two decades. It was therefore thought
ogy for engineers concerned with the implications timely for the new book. It was also thought bet-
of earth surface processes (e.g. landslides, erosion ter to call it ‘Geomorphology for Engineers’ to
and floods) and ground conditions. It should also reflect its contents.
appeal to others with an interest in understanding The book has been several years in gestation,
the landscape. It does not attempt to be compre- in part due to its multi-author nature, some
hensive and is not particularly systematic. It is not authors replying to requests immediately, others
therefore, a text book of geomorphology. The taking somewhat longer. Even with the advent of
main purpose is to act as an aide memoire on a electronic office wizardry, producing this book
worldwide scale of the different landscapes in has taken as long as, if not longer than, the previ-
which engineers might find themselves working. ous book before the advent of such technology.
It is assembled in three parts: firstly the controls Professor Peter Vaughan has now retired (or
that have helped shape the landscape; secondly, nearly so) to East Anglia. His role as the engineer
the processes, geomorphological and geological, on the original team has been ably filled, for this
which, sometimes with the influence of man, have book, by Dr George Milligan, formerly an Oxford
helped shape landforms and can lead to the academic, now a pragmatic consulting geotechni-
occurence of natural hazards that are of increas- cal engineer and a director of Geotechnical
ing concern as people rapidly spread over the Consulting Group (GCG), London.
planet; finally the various environments and land- The editorial team has been strenghthened by
scapes that are the product of the controls and the addition of a geomorphologist, Dr Mark Lee,
processes described earlier. one of the leaders of the current generation in the
A hope for this book is that it will help to alert newly developing field of engineering geomor-
engineers of the issues they may need to take into phology which now has a handful of full time
account especially when working in unfamiliar practitioners in U.K. Mark has worked for a major
environments — forewarned is forearmed. civil engineering consultancy, has spent time in
This book is the successor rather than a second academia at Newcastle University, and is now an
edition of the orginal Fookes and Vaughan 1985 independent consultant. Both Mark and I owe
‘Engineering Geomorphology’ published by the much to the inspiration of Professors Denys
now ceased Surrey University Press. Copies of Brunsden OBE, Sir Ron Cooke and David Jones
the latter are rather difficult to get hold of now, and to Dr John Doornkamp. They formed the
XIII
GFE-Prelims.qxd 6/27/05 2:35 PM Page xiv
xiv PREFACE
excellent pioneering Geomorphological Services colleagues in academia and consulting ‘too numer-
Limited (GSL) in the 1970s, undoubtedly the first ous to mention’.
British firm (perhaps first in the world?) to spe- The editors and publisher thank the many
cialise in geomorphology for engineers, and from authors for their efforts and stamina without
a professional viewpoint, a very successful com- which this book could not have been written.
pany that inspired much of the current practice of Special thanks go to Mr Mike Sweeney of
applied geomorphology and was a timely training British Petroleum, their senior geotechnical engi-
ground for many of the current generation of neer who has long been a supporter of geomor-
practitioners. Sadly GSL was taken over and then phology for engineers and who kindly arranged
closed down by accountants. However, its spirit for the Project Manager Local Environmental
lives on in the way that engineering geomorphol- Impact, Green Operations, BP, Mr Jim Clarke to
ogy in Britain is becoming an established adjunct make a small financial sponsorship to help keep
to civil engineering and environmental practice. such a large handbook to a reasonable price.
In addition to the gurus of Geomorphological The original book was dedicated to five given
Services Limited, Mark Lee would like to thank in names. Most of my friends knew that these were
particular Professor John Pethick (now pretending my children, but there were many questions ask-
to be retired) for his friendship and inspiration for ing who they were. This book is dedicated to
what is now one of Mark’s principal areas of inter- the wives and children or grandchildren of the
est and consulting — coastal engineering. editors.
George Milligan would like to thank Professors
Hugh Sutherland and Peter Wroth, Drs. David Professor P. G. Fookes, FR Eng., Hon. FRGS
Hight and Laurie Richards, along with many other Winchester
The printing of this book has been supported by BP in recognition of the importance of
geomorphology for engineering of major energy industry projects, and the key role that
the authors have played in developing and applying the thinking described herein, with
successful results.
GFE-Prelims.qxd 6/27/05 2:35 PM Page xv
Dedication
Edna Fookes and grandchildren Matthew, Christopher, James, Caitriona, Alexander and Niamh
Claire Lee and children Rebecca and James Lee
Barbara Milligan and grandchildren Fraser, Struan, Isabel, Sandy and Carrie
XV
GFE-Prelims.qxd 6/27/05 2:35 PM Page xvi
GFE-Prelims.qxd 6/27/05 2:35 PM Page xvii
Biographies
Fred Bell is a fellow of the Royal Society of underlined the importance of sea-level history,
South Africa, a fellow of the Institution of Civil and thus of climatic change, to the development
Engineers and the Institution of Mining and of carbonate systems of the more distant past.
Metallurgy, and a fellow of the Geological Over the years additional observations of lime-
Society. He is a Visiting Research Associate at the stones throughout the Phanerozoic from Britain,
British Geological Survey. Previously, he was Belgium, France, Germany, Norway, Montenegro
Professor and Head of the Department of Geology and Spain have been added and consultancy has
and Applied Geology, University of Natal, provided experience of hydrocarbon reservoirs in
Durban, South Africa. During this time he Canada, the UK and Libya, and engineering
also was a Distinguished Visiting Professor, applications in the UK.
Department of Geological Engineering,
University of Missouri-Rolla, USA. Professor John Charman has over 30 years experience as a
Bell’s consultancy research subjects have practising engineering geologist. Before embark-
included ground stability, subsidence, ground ing on a career as an independent consultant he
treatment, engineering behaviour of soils and worked for a large civil engineering contractor
rocks, cement, lime and PFA stabilisation of clay before becoming technical director of a specialist
soils, acid mine drainage, mining impacts, land- geotechnical consultancy. He now applies his
fills, derelict and contaminated ground, rock knowledge of geomorphological and engineering
durability in relation to tunnelling and dam foun- geology mapping, site investigation, landslides
dations, slope stability, aggregates, bricks, build- and soil erosion, soil and rock slope stability,
ing stone, and geohazards. Professor Bell is engineering behaviour of temperate, tropical and
author/editor of twenty books and over 200 papers arid soils, and the location and utilisation of con-
on geotechnical subjects. He has served on the struction materials to civil engineering projects.
editorial boards of five international journals and He works extensively overseas including techni-
has been a series editor for three publishers. cal assistance and training assignments for the
funding agencies.
Colin J. R. Braithwaite has a wide research
experience centred on the origins and charac- Martin Culshaw is in charge of the British
teristics of carbonate sediments and rocks in Geological Survey’s Urban Geoscience and
marine, lacustrine and subaerial environments Geological Hazards Programme and is the
in the Seychelles, Florida, the Bahamas, the Red Survey’s Chief Engineering Geologist. He has
Sea (Sudan) and Tobago, detailing the facies been involved in engineering geological research,
geometry and architecture of reefs and larger environmental and engineering geological map-
scale carbonate systems. The models developed ping, geohazard assessment and site investigation
have been tested on relatively recent (Pleistocene, for over thirty years. During this time he has spent
⬍ 2 million years old) limestones on Aldabra, in a number of years overseas in Asia, Africa, Europe
Kenya and Mauritius and subsurface in Australia and Central America. He is a former Chair of
(the Great Barrier Reef), in studies that have the Engineering Group, of the East Midlands
XVII
GFE-Prelims.qxd 6/27/05 2:35 PM Page xviii
xviii BIOGRAPHIES
Regional Group, and of the Geological Society. (IUGS). He has authored or edited more than 240
He is Chair of the Engineering Geological Maps scientific articles, including 6 books and several
Commission No. 1 of the International conference volumes.
Association for Engineering Geology and the
Environment and is a former editor of Engineering Ian Douglas, Professor of Physical Geography at
Geological Special Publications for the Geological Manchester University, studied erosion rates in
Society. He has served on various steering groups tropical rain forests in North Queensland for his
for the Geological Society, the Institution of Civil doctorate, and has subsequently worked on
Engineers and the Construction Industry Research applied problems in south-east Asia.
and Information Association. He is Visiting
Professor in Engineering Geology in the School of Peter Fookes is an independent consultant in
the Built Environment at the Nottingham Trent engineering geology. He started his private prac-
University. He has published over ninety papers tice in 1971, after two decades working in indus-
and articles and more than a hundred technical try and lecturing at Imperial College, and has
reports, edited eight conference proceedings, as worked in more than 90 countries.
well as organising numerous meetings and confer- He is a visiting or honorary professor at sev-
ences. He has served on the editorial boards of two eral British universities and has published over
international journals. 170 papers and books, and continues to lecture
and run university field trips. He has chaired
Edward Derbyshire is Research Professor many international technical working parties and
in Quaternary Science in the Department of committees. His work is field investigation and
Geography at Royal Holloway, University of heavy construction orientated and therefore more
London, England, Emeritus Professor of the pragmatic than experimental, and has been recog-
University of Leicester, Honorary Professor of nised by numerous awards from concrete, engi-
the Academy of Science of Gansu Province, neering, geographical, geological and geotechnical
China, and an Honorary Member of both the institutions worldwide. In 1991, he was made a
International Union for Quaternary Research Fellow of the Royal Academy of Engineering and
(INQUA) and the Quaternary Research in 2001 a distinguished research assoicate of
Association (UK). He is a holder of the Antarctic Oxford University.
Service Medal of the United States, and has He claims that his work is his hobby and is
conducted field research in all six continents, in currently resisting strong attempts by his wife to
pursuit of his interest in the geomorphology and get him to retire, even in part.
sedimentology of glacial and wind-blown
deposits. He was Chairman of the British Geo- Andrew Goudie is Professor of Geography in the
morphological Research Group in 1982–83, University of Oxford and has worked extensively
President of Section E (Geography) of the British as a geomorphologist not only in the UK but also
Association for the Advancement of Science in the world’s deserts and savanna lands.
in 1990, Secretary-General of the International
Union for Quaternary Research (INQUA) Warren E. Grabau, geologist/geographer, retired
1991–95, and Chairman of the Scientific Board from the US Army Corps of Engineers in 1984
of the International Geological Correlation after a career devoted primarily to multidiscipli-
Programme (IUGS-UNESCO) 1996–2001. He nary research involving the enhancement of
has served as Editor of Quaternary Perspectives the performances of military materiel, chiefly
and is on the editorial board of Glacial Geology weapons systems and vehicles, in harsh environ-
and Geomorphology and Geografia Fisica e ments. His interests extend to the dynamics and
Dynamica Quaternaria. He is currently Chairman geomorphology of large alluvial rivers, with
of the Committee for Research Directions of the special attention to the floodplain and delta of the
International Union of Geological Sciences lower Mississippi. He is also a student of military
GFE-Prelims.qxd 6/27/05 2:35 PM Page xix
BIOGRAPHIES xix
history and has published material on the impact three edited books, ‘Landslides’, ‘Land Surface
of geography on military operations. Evaluation for Engineering Practice’ and ‘Mapping
in Engineering Geology’.
Kenneth J. Gregory was Lecturer (1962–72) and
Reader in Physical Geography (1972–76) at the James V. Hengesh has been a Senior Geologist
University of Exeter, Professor of Geography since 1986 specializing in seismic hazard investi-
(1976–92), Dean of the Faculty of Science gations. He has performed investigations for criti-
(1984–87), and Deputy Vice Chancellor (1988– cal facilities and lifelines such as nuclear power
92) at the University of Southampton, and Warden plants, liquefied natural gas (LNG) terminals, off-
of Goldsmiths College University of London shore platforms, pipelines, dams and military
(1992–98). From 1998–2001 he held a Leverhulme installations. During these projects he has con-
Emeritus Fellowship. His research and publications ducted geological and seismotectonic assessments
are in the fields of river channel changes, palaeo- to characterize seismic hazards for input to design
hydrology, river channel management, and the of new structures or engineering evaluation of
nature of physical geography. He is President of the existing facilities.
Global Continental Palaeohydrology Commission He has been funded by the U.S. National
of INQUA (1999–2005). He has received the Earthquake Hazard Reduction Program (NEHRP)
Back Award (1980) and the Founder’s Medal to investigate faults in northern California and
(1993) of the Royal Geographical Society, the Puerto Rico. He has participated in a number of
Linton award (1999) of the British Geomor- USGS Working Groups to evaluate seismic haz-
phological Research Group, and the Geographical ards in the San Francisco Bay area and has been
Medal (2000) of the Royal Scottish Geographical an invited speaker on seismic hazard issues at
Society. conferences in Europe, the Pacific Islands, and
New Zealand.
James S. Griffiths has worked as a natural haz-
ard assessment specialist in the UK, Eire, Spain, Colin F. Jago graduated in Petroleum Geology
Switzerland, Chile, California, Liberia, Ethiopia, from Imperial College, University of London. His
Syria, Saudi Arabia, Georgia, Pakistan, Hong doctoral research, also at Imperial, London, was
Kong, Papua New Guinea and Australia. After in beach and estuarine sedimentation. He has
completing his PhD in flood hazard assessment maintained research interests in the sedimentary
for highway design in hot deserts he joined dynamics of macrotidal estuaries but his primary
Rendel Palmer & Tritton, Consulting Engineers research activity is concerned with suspended
in 1979. For the next 14 years he worked in sediment dynamics of continental shelves. This
industry on a range of projects throughout the research is focused on bio-physical interactions
world specialising in flood estimation for bridge that govern the properties and flux of suspended
and culvert design, highway earthworks design, sediments and the geotechnical and hydro-
slope stabilisation and engineering geomorpho- dynamic properties of seabed sediments. Dr Jago
logical mapping. For the period 1987–90 he was is Senior Lecturer in Geological Oceanography in
technical director of Geomorphological Services the School of Ocean Sciences, University of
Ltd. In 1993 he joined the University of Wales Bangor.
Plymouth as a lecturer in engineering geology,
was made Head of Geological Sciences in 1995, Mark Lee is an engineering geomorphologist
and in 2002 became Enterprise Co-ordinator in working in private practice. He was formerly
the Faculty of Science. He has continued to work Principal Research Associate at the University of
as a consultant and in 1998–99 was part of Newcastle, UK and prior to that had worked in
the team carrying out a global assessment of industry for 18 years. He has had extensive expe-
landslide risk in Hong Kong. He is the author or rience of landslide hazard and risk assessment in
co-author of over 60 publications including both natural and man-made terrain throughout the
GFE-Prelims.qxd 6/27/05 2:35 PM Page xx
xx BIOGRAPHIES
world. He is actively involved in research into the George Milligan has an MA in Engineering
role of geomorphology in Quantitative Risk Science from Oxford University, an MEng in Geo-
Assessment. In recent years he has specialised in technical Engineering from Glasgow University,
the development and application of probabilistic and a PhD from Cambridge University, where his
methods of modelling coastal cliff recession and research was on flexible retaining walls. He
landslide reactivation scenarios. gained experience as a structural and geotechnical
engineer with Ove Arup and Partners, Scott
William R. Lettis is responsible for conducting Wilson Kirkpatrick and Partners, and Golder
regional and site investigations to assess seismic Associates, including working in Iraq for two
hazards in both active plate margin and stable years. He was then a University Lecturer at
intraplate tectonic environments. These investi- Oxford University for 17 years, conducting
gations typically involve the deterministic and research in soil reinforcement and soil nailing,
probabilistic assessment of ground motions, applications of geotextiles, and small diameter
surface fault rupture, liquefaction, dynamic tunnelling, particularly pipe jacking. Since 1997
slope stability, and tsunami. He has conducted he has worked for the Geotechnical Consulting
field and office studies for nuclear facilities, Group, on a wide range of projects in the UK and
transportation systems and bridges, hydroelec- abroad, and has been a Director since 1998.
tric facilities, pipelines and tunnels, and LNG
and oil platform facilities throughout the United Julian Orford graduated in Geography and
States and abroad. Many of these projects were Sociology from Keele University, and undertook
successfully completed in a regulatory environ- research at both Salford and Reading Universities
ment. He is currently a member of the into aspects of beach gravel, beach sedimentation
California Earthquake Prediction Evaluation and numerical methods of analysing facies vari-
Council, a scientific advisory council to the ability in coastal sedimentary environments. He
Governor and Office of Emergency Services in has pursued coastal research in the UK, Ireland,
California. Canada, and the USA. His recent research areas
are in coastal development and beach behaviour
Xingmin Meng is a Geographic Database over decade–century scales, with a specific interest
Manager and Geographic Information System in the behaviour of transgressive sand and gravel
(GIS) Analyst in the Department of Geography, barriers and regressive dune systems. He is also
Royal Holloway, University of London. His PhD concerned with developing effective means of
thesis was on loess and loess instability in North coastal zone management that combine geomor-
China. He was involved in a major (eight year) phology and engineering in a sustainable fashion.
international collaborative research programme He has been lecturing in Geography at Queen’s
on geohazards in the loess terrain of North China, University, Belfast, since 1977, where he is now
financed by the Commission of the European Professor of Physical Geography and head of the
Communities and the Government of Gansu School of Geography.
Province, China. He has also held the positions of
Postgraduate Research Assistant and Postdoctoral Lewis A. Owen received his doctorate at the
Research Assistant in several NERC-funded University of Leicester in 1988 for his research on
research programmes in both the University of the Quaternary geology and geomorphology of the
Leicester and the University of London (Royal Karakoram Mountains in the northern areas of
Holloway), as well as being Postdoctoral Research Pakistan. Since receiving his doctorate he has con-
Assistant under a grant from the Leverhulme tinued to work in high mountains and drylands to
Trust. His current research interests include: examine aspects of landscape evolution and pale-
Geographic Information System (GIS) and oenvironmental change. This has included research
related applications; landslides and slope instabil- in the Himalaya, Tibet, Sierra Nevada, Alaska and
ity; aeolian dust and loess deposits. the deserts of Central Asia and Western USA. He
GFE-Prelims.qxd 6/27/05 2:35 PM Page xxi
BIOGRAPHIES xxi
has been on the faculty at the Hong Kong Baptist Arctic Institute, and the International Association of
University, Royal Holloway University of London, Geomorphologists. His major research topics have
the University of California and the University of been devoted to arctic geomorphology and coastal
Cincinnati. He is presently an Associate Professor modification.
in the Department of Geology at the University of
Cincinnati. Tony Waltham was a Senior Lecturer in
Engineering Geology at Nottingham Trent
Michael F. Thomas is Professor (Emeritus) in University in the UK. After completing degrees at
Environmental Science, University of Stirling, Imperial College London, he turned his geological
Scotland, UK. His research includes the role of interests away from mining and towards engineer-
saprolites in landscape development, and the ing. His main research has been into the process of
impact of Quaternary climate changes on sedi- ground subsidence and catastrophic failures, espe-
mentation in humid tropics. He is the author of cially those related to cave collapse and sinkhole
Geomorphology in the Tropics (Wiley, 1994), and development in limestone karst. This stemmed
was a member of Geological Society working from his many years as an active caver when he
party on Tropical Residual Soils. explored and mapped caves while trying to under-
stand their geomorphology. He has travelled
H. Jesse Walker, Boyd Professor Emeritus, widely through almost every country in Europe,
has taught in and served as Chairman of the Asia and North America, and finds active volca-
Department of Geography and Anthropology at noes nearly as exciting as underground rivers. He
Louisiana State University, Baton Rouge, USA. He has written numerous books, including the widely
has been a visiting professor at the Universities of used student textbook Foundations of Engineering
Hawaii, California, and Tsukuba (Japan), and a Geology, and now also runs Geophotos, his own
Liaison Scientist with the US Office of Naval picture library that specialises in geology and
Research, London. He is a Fellow of the AAAS, the geomorphology.
GFE-Prelims.qxd 6/27/05 2:35 PM Page xxii
“Marcley Hill in the year 1575 after shaking and roaring for the space of three days, to the great
horror, fright, and astonishment of the neighbouring inhabitants, began to move about 6 a clock
on a Sunday evening and continued moving or walking till 2 a clock on Monday morning; it then
stood still and moved no more. It carried along with it the trees that grew upon it, and the sheep
folds and flocks of sheep grazing on it. In the place from whence it remov’d, it left a gap of 400
wide, and 320 foot long. The whole spot whereon the Hill stood contained about 20 acres.”
Contemporary description of the 1575 ‘Wonder Landslide’ in Herefordshire, written by Fuller

and quoted in Coole (1882).
GFE-01.qxd 6/24/05 2:43 PM Page 1
1. Introduction to Engineering Geomorphology

Peter Fookes and Mark Lee
1.1 Preamble Engineering and geomorphology

The application of some geomorphological acumen
The earth’s surface is dynamic and landforms by non geomorphologists (i.e. those not specifically
change through time in response to weathering trained or experienced in the subject) dates back
and surface processes (e.g. erosion, mass move- thousands of years. Ancient examples would be the
ment and deposition). Geomorphology is the siting of London (at the last fording place before the
study of landforms and landform change. The lit- sea); the siting of the Nile/Red Sea waterway, fin-
eral meaning of the word is the study of the form ished in the second century BC; the siting of Chinese
of the earth. Geomorphology, therefore, studies canals dating back to about 1000 BC or the Qanat
the earth’s surface forms and processes. It is by no systems of Persia dating back to about 1500 BC.
means a new subject, having its modern origins More recent examples include aspects of the siting
in the 19th century but its specific application of canals in the canal era in England (c. AD 1760 to
to civil engineering (as engineering geomor- 1830), Brunel’s or Stephenson’s railway alignments
phology) is a much more recent development. (c. 1820 to 1860); aspects of the reclamation of
Engineering geomorphology complements Fenlands (c. AD 600 to present) or polders in
engineering geology in providing a spatial con- Holland (c. AD 600 to present). It is clear that virtu-
text for explaining the nature and distribution ally every form of civil engineering involving the
of particular ground-related problems (e.g. the ground benefits from some understanding of sur-
presence of duricrusts, aggressive soils) and face forms and processes (i.e. geomorphology),
resources (e.g. sand and gravel deposits). whether applied or not.
Importantly, engineering geomorphology is also Although geomorphology and engineering
concerned with evaluating the implications of geology are complementary subjects, they have
landform changes for society and the environ- different histories and, hence, have experienced
ment. The focus is primarily on the risks from different paths towards acceptance by the engi-
surface processes (i.e. the impact of so-called neering profession. Engineering geology, which
‘geohazards’) and the effects of development on has some of its roots in the attempts to quantify the
the environment, notably the operation of surface behaviour of soils, commenced in the early 18th
processes and the resulting changes to landforms century almost entirely in France (Skempton,
or the level of risks. 1979; 1985). The engineering works associated
These views form the substance and thrust of with the industrial revolution provided much
the book. We also make the distinction between empirical experience of the behaviour of geotech-
applied geomorphology — the use of specia- nical materials. A paradox then developed. While
list academic expertise to address particular the exposures of the ground provided by engineer-
problems — and engineering geomorphology — ing works played a major part in the development
the integration of the geomorphological app- of geology as a separate subject, the engineering
roach into engineering project planning, design scientists of the day were unable to provide coher-
and construction. ent analyses to formulate the behaviour of soils,
1
GFE-01.qxd 6/24/05 2:43 PM Page 2
2 GEOMORPHOLOGY FOR ENGINEERS
particularly clays. While such notable scientists as et al., 1982); Geomorphological Hazards in
Navier (1833), Poncelet (1840), Collin (1846), Los Angeles (Cooke, 1984) and Applied Fluvial
Darcy (1856), Rankine (1862), Boussinesq (1883) Geomorphology for River Engineering and
and Reynolds (1887) — all reviewed by Skempton Management (Thorne et al., 1997) are among some
(1985) — made contributions, it was not until the of the relatively recent distinguished exceptions.
1920s when Terzaghi appreciated the role of pore- Geomorphology is without doubt of primary
water pressure in controlling the behaviour of soil, importance to civil engineers involved in ground
that such analyses became possible. Thus soil engineering. However, the authors know of
mechanics did not become a full part of engineer- no case where ‘geomorphology for engineers’ is
ing science until the 1930s, when courses in the being taught to engineers. Such a course should
subject were first run at Harvard. Its worldwide incorporate subjects such as basic geomorphol-
dissemination did not occur until the 1950s. The ogy, with the principal topic areas treated with
importance of geology to engineers was appreci- emphasis on surface form, process and near-
ated and taught to engineers, however, long before surface materials; basic climatology and climate
geology had gone its separate way as a natural sci- change; Quaternary studies; techniques involving
ence primarily concerned with the history of the photo-interpretation and remote sensing, plus the
earth. It had relatively little interest in the superfi- appropriate geomorphological mapping training in
cial deposits, landforms and the processes that the field, together with the relevant elements of
shape them — of primary concern to engineers — basic geology such as physical geology, historical
and could not offer any quantitative basis for geology, sedimentology, petrology and structural
predicting engineering behaviour. It sat uneasily geology.
with other quantitative subjects in an engineering This is a book largely by geomorphologists for
education. Much has changed in the last five engineers. It may occasionally show a parallel
decades but current attitudes and practices still need for a book on engineering for geomorpholo-
reflect earlier difficulties. gists, but that is a different question. It is not our
Geomorphology has, however, developed intention to try and turn engineers into expert
somewhat as a separate subject related to, but not geomorphologists, but to help engineers and
a part of, geology. This is especially so in Britain, others to understand the subject and to appreciate
where it has its roots in physical geography — the part that geomorphology and geomorpholo-
a long way removed from the worldly business gists can play in engineering. The importance of
of construction. Geomorphology has also devel- first understanding a site in broad context, before
oped separately from geotechnical engineering. engaging in detailed investigations and predic-
Inevitably, there are problems of communication, tions, cannot be overstressed.
in both directions, some of which may be apparent Problems have to be identified before they can
in this book. be solved. This book should thus be of most
It has been suggested by Cooke and Doornkamp value at the project planning and site investi-
(1974) that applied geomorphology studies date gation stages. One cautionary comment may
back to about the turn of the last century. In general be prudent for the inexperienced reader. Many
such work was done by academic geographers with diagrams illustrating landscapes in this book
little background in the world of civil engineering, attempt to include most if not all the features
although some learned quickly. Unfortunately, even which may be associated with a particular land-
today there is a general scarcity of papers and text- scape. Nature is not always so obliging, or some-
books specifically related to applied geomorphol- times not so complicated, and this should be
ogy; the books Environmental Geomorphology remembered when applying these models to a
(Coates, ed., 1971); Geomorphology in Environ- particular site.
mental Management (Cooke and Doornkamp, The authors can see great benefit in utilising
1974; 1990); Applied Geomorphology (Hails, ed., and incorporating geomorphology and geomor-
1977); Urban Geomorphology in Drylands (Cooke phologists within the framework of geotechnical
GFE-01.qxd 6/24/05 2:43 PM Page 3
INTRODUCTION TO ENGINEERING GEOMORPHOLOGY 3
engineering. This is already happening. Such a Often, landscapes have been formed predomi-
move will benefit from more specific training nantly by past climatic conditions, rather than by
and education of both engineers and geomor- present ones. Britain is a good example. The
phologists. It is hoped that this geomorpho- form of the landscape has for the most part been
logy for engineers will be found useful in this inherited from the Tertiary, but also bears the
process. imprint of relatively recent glacial and periglacial
conditions that are no longer present. In terms of
The book engineering at least, there is a need to separate
Many, if not most, difficulties in geotechnical the past effects that have helped form the land-
engineering arise either from an unawareness of scape from current effects which may only be
ground conditions or a failure to appreciate the modifying. Current tectonic activity also affects
influence of known ground conditions on a active geomorphological processes, for example
particular engineering problem. A basic under- if young mountains are slowly moving upwards
standing of geomorphology is a considerable then landslides, erosion and so on are typically
aid to the solution of the first problem. The more active. In old mountains which have ceased
second problem requires a specific engineering to go up and are being worn down, their slopes
input. are, in the long term, becoming progressively
The aim of this book is, therefore, to present a more stable. Similarly, ground level changes in
basic yet authoritative handbook of geomorphol- relation to sea level control river erosion and
ogy for geotechnical engineers and others. It does deposition, the risk of coastal inundation and so
not attempt to be comprehensive, is not particu- on. These effects are discussed mainly in the first
larly systematic, and as such is not a textbook of part of the book and their results highlighted in
geomorphology. Its main purpose is to act as an the second and third parts.
aide-memoire, on a worldwide scale, of different The second part of the book (Chapters 8–12)
landscapes in which engineers may find them- is concerned with processes — geomorphologi-
selves working. It is hoped that each chapter will cal and geological, with and without the influ-
help them understand a little better what may be ence of man, which help shape landforms and
unfamiliar environments. lead to the occurrence of natural hazards such as
The essence of the book is in the diagrams landslides and floods. The usually slower contin-
around which each chapter is built. The first ual wear of wind, ice and water and the often
part (Chapters 2–7) is concerned with controls more spectacular but irregular effects of neotec-
that have helped shape the landscape — climate tonics, subsidence and landslips are brought
and weathering, past and present, acting on the together in this part.
land surface; the sediments and sedimentary The third part (Chapters 13–27) of the book is
processes which help build the landscape; stratig- concerned with various environments and land-
raphy, which is concerned with the history of forms, mainly natural but also some influenced
soils and rocks; and tectonics, geological move- by man. It starts in the colder, higher-latitude
ments, often from deep in the Earth’s crust, parts of the world, the glacial and periglacial, and
which may form the bones as well as some of the moves generally towards the middle and lower
wrinkles of the landscaped skin of the Earth. This latitudes, through temperate environments, hot
part includes the very important chapter on the dryland (‘desert’), savanna and hot wetland
Quaternary, the history of the last 2 or 3 million (‘tropical rain forest’). It goes on to outline spe-
years of the Earth’s surface, a study that should cific geomorphological features, mountains,
figure large in the geological and geomorpholog- estuaries, coasts and shallow seas around the
ical training of any civil engineer. It ends with a coasts. Three particular terrains which present
chapter on engineering soils in which soil and special problems to engineering are considered:
rock mechanics and their relation to geomor- volcanic, soluble rock (mainly limestones) and
phology are explored. loess (windblown silt). This part is completed
GFE-01.qxd 6/24/05 2:43 PM Page 4
with a chapter about the influence of man on the 1.2 Geomorphology, landforms and
urban landscape. the nature of modern engineering
However, the three parts should not be seen as geomorphology
rigorous — each chapter contains bits that would
fit into other chapters. One of the key strengths of geomorphology for
The engineering content of the book has neces- engineering is the ability to provide a framework
sarily to be brief, yet as useful and relevant as pos- for understanding the current and historical
sible. Following from the above, it was thought behaviour of physical systems — slopes, terrain
that a realistic way of dealing with engineering units, catchments or coastal sediment transport
situations was to produce a checklist of the type of cells. Morphological adjustments (i.e. erosion and
engineering difficulties that can arise, with as deposition) can occur in response to variations in
much emphasis on the relatively unusual problems energy inputs into these systems, including varia-
that can catch you out, as well as the more well tions in river flows (discharge and sediment load)
known. It is assumed the reader will either know, and wave/tidal energy arriving at the coast, over
or will seek elsewhere, the full answer to the prob- a range of timescales. A number of distinctive
lem. The book tries to show and make the reader modes of change may be recognised:
aware of problems that might arise. It is clear that Episodic, progressive change, where no recov-
a separate checklist for each chapter would invoke ery or retrogressive change occurs. For example,
excess duplication and length. A tabulated master cliff erosion involves sequences of episodic
checklist is therefore given in Chapter 7, on landslide events that result in loss of cliff top
Engineering Soils, while each chapter contains land. Cliffs do not recover by advancing for-
further discussion of the relevant issues. ward. Figure 1.1 presents an example of the
Each chapter thus becomes a summary of the recession process from the glacial till cliffs of
particular types of landforms arising from the Holderness, Yorkshire, UK.
particular environments, which concentrate on Periodic or cyclical change, where the land-
geometry of form and on types of materials. Each form responds to an event (e.g. a storm or large
chapter also summarises the particular types of flood) by altering the morphology, but then grad-
geomorphological agency, with the influence of ually recovers (either partially or completely) its
these agencies, when current, on engineering original pre-event form. For example, variations
design and construction. in wave energy can cause major changes to beach
It should be noted that the specific determina- form, with sediment being eroded from the upper
tion of material types, geometric boundaries, shore and transported offshore under high wave
engineering properties and methods of design and energy conditions and brought onshore in lower
analysis, are part of the geotechnical engineering energy conditions, leading to beach accumulation
processes which are outside the intent of this (Figure 1.2). Thus, there may be an envelope of
book, though difficulties arising in site investiga- beach positions varying between steeper gradi-
tion are not. However, much of the geotechnical ents associated with periods of beach deposition
process and virtually all the design and analysis and flatter profiles associated with periods of
process depends on a presumption of the prob- beach erosion.
lems with which the proposed structure will have The availability of sediment for exchange or
to cope. The engineering aim of this book is, transfer between different ‘stores’ is critical to
therefore, to aid the engineer in helping to ensure cyclical change. Considering a beach-sand dune
that any relevant problems are not overlooked, system, sand may be deposited on the upper
hence the checklist method adopted. The experi- beach during low wave energy events, providing a
enced reader will doubtless be able to add further source of sediment for aeolian transport and fore-
points to each list. Each chapter ends with specific dune growth. During high energy events, the
references given in the text and a bibliography of beach will flatten and the fore-dunes will erode.
related but fuller texts. This provides an additional source of sand that is
GFE-01.qxd 6/24/05 2:43 PM Page 5
Figure 1.1 Representative annual cliff recession measurements on the Holderness coast, UK (redrawn from
Pethick, 1996).
On the coast, the Holocene sea level rise and

changes in sediment availability have initiated
important long-term trends in landform develop-
ment. During the early Holocene, for example,
rapidly rising relative sea level resulted in onshore
Figure 1.2 A simple beach profile change model (from movement of large volumes of sand and gravel
Hansom, 1998). onto the Scottish coast and is associated with
periods of beach and dune formation (Hansom,
1998). However, after around 6500 BP, relative sea
transported seaward and deposited on the lower level began to fall on parts of the Scottish coast
beach. because of the increasing significance of non-
Secular change, where there is a longer-term uniform isostatic uplift. The resultant sediment
trend in the evolution of the systems. For example, deficit had two main effects on coastal evolution.
Warburton et al. (1993) describe how the Ashley First, erosion of the beach and back-beach dune
River braid-plain, New Zealand, is migrating systems, as sediments were removed from these
northwards over a timescale of decades due to areas into deeper water in order to maintain
bank erosion on the northern margins and bar a near-shore gradient capable of dissipating
stabilisation on the southern margin (Figure 1.3). wave energy. The widespread trend of beach ero-
They speculate that this trend may be in response sion and eroding dune systems throughout the
to large-scale sedimentation cycles. Western and Northern Isles is believed to be the
GFE-01.qxd 6/24/05 2:43 PM Page 6
Figure 1.3 Ashley River channel changes: (A) composite map of all channel positions 1950–1992 (B) age of
vegetated bar surfaces showing a migration northwards (from Warburton et al., 1993).
consequence of the long-term sediment deficit. 2. High to relatively high probability events, such
Second, reorganisation of the coastline into pro- as wind blown sand, soil erosion, shallow hill-
gressively smaller sediment transport cells as side failures, flooding and river bank erosion,
the emergence of headlands on the retreating coastal erosion and deposition.
shoreline created barriers to long-shore transport. 3. Low probability events that would have a major
These cells have tended to evolve subsequently by impact on the project or development, such as
internal reorganisation, erosion and deposition. flash floods, major first-time landslides, reacti-
The limited duration of engineering time vation of pre-existing landslides, channel bed
imposes a constraint on the types of landform scour, fault rupturing and neotectonics, the col-
change that are relevant to engineering geomor- lapse of solution cavities (especially in areas of
phology, viz.: gypsum karst) and tsunamis.
1. Abrupt and dramatic landscape changes that Specific questions that frequently have to be
are likely to be significant over a 1–100⫹ year addressed include:
timescale. Examples include the establishment
of gully systems, migration of sand dunes, 1. What could happen? — The nature and scale
river planform changes, coastal cliff recession of the changes or events that might occur over
and the growth and breakdown of shingle the next 50–100 years or so (longer for major
barriers. projects).
GFE-01.qxd 6/24/05 2:43 PM Page 7
2. Why might such events happen? — The circum- regionalised associations between landforms,
stances associated with particular changes or ground conditions, geohazards and resources —
events (e.g. the occurrence of extreme intensity a characteristic geomorphological environment.
rainfall events, wet years, earthquakes, storm Some of the major geomorphological environ-
surges, human interference). ments are discussed later in the book e.g. glacial,
3. What is the chance of it happening? — The periglacial, temperate, hot drylands, savannah,
probability of such events occurring in hot wetlands, mountains and highlands. How-
any given year (the annual probability) or over ever, over geological timescales the boundaries
a particular time period (the cumulative of these environments are dynamic, drifting
probability). across the continents in response to major
4. What damage could be caused? — The conse- climatic or tectonic changes. As these environ-
quences of such events in terms of losses (e.g. ments have shifted, so the intensity and charac-
loss of life, property, services, infrastructure ter of geomorphological processes will have
damage). changed.
Present day surface processes are acting on a
For these reasons, engineering geomorphology landscape fashioned by ancient processes. In
is directed towards understanding the way land- some areas current processes may have behaved
forms or physical systems respond to relatively in the same or a similar manner for extremely
short–medium term impulses (e.g. climatic vari- long periods of time, as in the ancient cratonic
ability, changes in sediment supply, land use regions of Australia and South Africa. In other
change, the effects of man, neotectonics in seis- areas, current processes may be actively re-shaping
mically active regions) rather than longer-term the inherited landscape, destroying it or simply
landscape denudation and evolution. However, if having no effect at all.
the focus is on what in geological terms could be It follows that many landscapes are compos-
called the ‘here and now’, there is a need to be ite forms, containing a range of features of
aware of the significance of longer term trends varying antiquity inherited from different envi-
(e.g. the Holocene decline in sediment availability ronments as major climatic or tectonic changes
experienced on many temperate coastlines) and left their imprint. The landscape rarely reflects
the presence of potential geohazards or resources any one climate or period of geomorphological
inherited from the distant past (e.g. ancient land- change, rather it is usually an assemblage of
slides, karst features). landforms and systems that have superimposed
As engineering geomorphology needs to address histories — geomorphologists often describe the
short- and medium-term changes, notions of grad- landscape as a palimpsest (i.e. like a surface
ual landscape denudation (e.g. progressing through which has been written on many times after pre-
the youth, maturity and old age of the so-called vious inscriptions have only been partially
Davisian cycle) over millions of years has little erased, e.g. Chorley et al., 1984; Twidale, 1985).
relevance. Instead the foundations of engineering Present day conditions do not necessarily
geomorphology are awareness of and the distribu- give a good guide to the conditions that created
tion of inherited landforms, geomorphological particular landforms. The implications for engi-
events, interconnected physical systems and their neering geomorphology include landscape ele-
complex behaviour. ments of different ages and stabilities; a legacy
of pre-existing geohazards; the presence of
Inherited landforms inherited near surface materials and inherited
Landscapes are the product of extremely long resources.
periods of weathering and surface processes. It Landscape elements under present day climate
is widely recognised that these processes oper- and environmental controls include:
ate with different intensities in different climatic Actively dynamic landforms: subject to almost
and tectonic regimes, tending to produce highly continuous adjustment to changes in energy inputs
GFE-01.qxd 6/24/05 2:43 PM Page 8
(e.g. rainfall, river discharges, wave attack) or pediment slopes in deserts and ancient surfaces in
sediment availability. River channels and the cratonic areas.
coastline are the best examples of this type of Although pre-existing geohazards may have
landform. been created under very different conditions,
Episodically active landforms: subject to these features may continue to present significant
occasional adjustments to extreme events, such risks to engineering projects. Examples include
as hillsides prone to shallow debris slides in trop- pre-existing landslides prone to reactivation, and
ical regions; sand dune fields and steep mountain karst terrain where there remains a potential for
catchments prone to debris flows and flash collapse.
floods. Inherited near-surface materials that have
Stable, relict landforms: although they are sta- accumulated under former environmental condi-
ble under present conditions, they would not have tions include spreads of wind-blown loess derived
evolved under them. For example, changes in from glacial outwash plains that bordered the
wind velocity and increased rainfall have sta- mid-latitude Pleistocene ice sheets, glacial tills
bilised many dune systems that would have been and solifluction deposits, and the deeply and
active earlier in the Pleistocene (e.g. the stabilised intensively weathered regoliths or relict duricrusts
dune fields of the Nebraska Sand Hills, covering that can be found beneath many low relief sur-
57 000 km2). Periods of extreme aridity and high faces, especially in stable cratonic zones.
wind speeds led to the development of the Inherited resources, such as fluvio-glacial or
100–200⫹m high primary dunes of the Saharan relict floodplain sands and gravels, are essentially
sand seas. In the current dry phase, sand dune non-renewable under contemporary conditions
mobility is restricted by relatively low wind and processes.
speeds and, hence, the primary dunes are effec-
tively stable. However, smaller-scale dune forms Geomorphological events
(secondary dunes) that are superimposed onto the A geomorphological event can be described as the
larger immobile primary dunes are mobile under occurrence of a process that results in landform
current conditions. Elsewhere, many stable river change, such as a landslide, flash flood or erosion
channels in temperate regions are relict forms, event. By their nature, those event that are signif-
adjusted to suit climatic and sediment load condi- icant for engineers tend to be infrequent episodic
tions during and after deglaciation. events. In the intervening years the landscape
Unstable, relict landforms: landforms that might appear completely benign, often instilling a
developed under past conditions, but are being false sense of security before a major damaging
progressively destabilised by present conditions. event occurs. A good analogy for event history of
Many beaches, for example, were created from a landscape is:
material moved onshore as sea levels rose early in
the Holocene and, thus, are essentially relict fea- ‘like the life of a soldier, long periods of boredom
tures often with no obvious contemporary source interrupted by brief moments of terror’ (Ager, 1976).
of sediment. Slapton Beach, Devon, comprising
chert and flint but surrounded by cliffs of The size of these events is not constant, as
Devonian-aged sedimentary rocks is a good illustrated by desert streams where discharge
example. The implication for coastal management can vary by 2–3 orders of magnitude. However,
is clear: unless the erosion of cliffs along adjacent high-intensity events tend to be rare, whereas
sections of a coastline is able to counter the natu- low- to medium-intensity events are more fre-
ral wastage due to attrition, beach mining, loss quent. Typical magnitude/frequency distribu-
offshore or along the coast, then these beaches tions tend to be positively skewed in form
will diminish in size. (Figure 1.4a). However, some geomorphological
Fossil landforms: essentially unchanging events such as landslides do not conform to this
elements of the landscape, such as low angled simple frequency distribution as the events are
GFE-01.qxd 6/24/05 2:43 PM Page 9
often constrained within narrow size-bands by The probability of an event occurring over a
internal slope factors (e.g. discontinuity spacing, particular time period (n years) is:
disposition of strata; Figure 1.4b).
Pn ⫽ 1 ⫺ (1 ⫺ Po)n
Historical event frequency is often used to
estimate the likelihood (probability) of such an where Po is the annual probability of the event.
event occurring in the future. For example, in The flood that is expected to be equalled or
flood studies the likelihood of a flood of a partic- exceeded, on average, every 100 years, has a return
ular magnitude is generally expressed by the period of 100 years and an annual probability of
return period or recurrence interval: 0.01. As floods are assumed to be random, rather
than regular in occurrence, the 100-year flood event
N ⫹1 could occur any year (or not), but the chance of its
Return period ⫽ occurrence during a 100-year period is greater
m
than during a single year.
Most estimates of event probability are
where N is the number of records in the time
extremely sensitive to the quality of the historical
series and m is the order or rank number within
data set. To achieve 95% reliability on an estimate
the series of the event in question.
of the 50-year flood requires 110 years of records.
Such lengthy data sets are rare, placing consider-
able limitations on our ability to predict extreme
events.
Where there is no historical record of geomor-
phological events, the combined body of evidence
(site-specific and landform assemblages) does
allow certain judgements to be made about the
potential for geohazard activity in an area. As the
field evidence alone rarely supports the use of
precise quantitative results for particular sites,
generalised statements applicable to a variety of
sites are often used (e.g. high, medium, low
potential). In other instances, an absence of peak
flow measurements can be partly compensated for
by geomorphological or sedimentological infor-
mation that permits the reconstruction of the
maximum heights reached by floodwaters.
The causes of geomorphological events can be
very complex, but often relate to a combination of
the following factors:
Preparatory factors which make the system sus-
ceptible to geomorphological events — for exam-
ple, antecedent conditions determine the amount of
a catchment that is saturated prior to a rainstorm or
snowmelt event and, hence, the amount of runoff
that will be generated.
Triggering factors which actually initiate the
event (‘the straw that breaks the camel’s back’) —
Figure 1.4 Typical event frequency distributions: (a) for example, gully erosion events are often trig-
positively skewed (e.g. wave height) and (b) in narrow, gered by intense rainstorms, although land use
constrained bands (e.g. landslides). change is often an important preparatory factor.
GFE-01.qxd 6/24/05 2:43 PM Page 10
Controlling factors that determine the nature, regarded as the product of broader controls that
scale and intensity of the event when it occurs — influence the behaviour of hillsides, river catch-
for example, the catchment’s network characteris- ments (drainage basins) or coastal cells. For
tics will influence the speed at which water is example, to understand the flood character of a
transmitted through the channel system. Dendritic river, something must be known of the climatic,
networks tend to produce a marked concentration geological, topographic and land use controls on
of flow in the lower catchment as floodwaters are the supply of water and sediments from the sur-
delivered down the main tributaries at a similar rounding hills. On the coast, the development of
speed. Trellis networks tend to produce a more beaches or shingle banks needs to be seen as the
muted response. product of sediment transport within dynamic
Events may often be significantly delayed coastal systems.
after the initial preparation. For example, In addition to providing a framework for
although the majority of major rock slope fail- understanding the flood behaviour of rivers and
ures in Scotland appear to be associated with streams, the catchment concept can help explain
glacially steepened slopes, they occurred during the way water and sediments are transported
the period 10 000–5000 BP (i.e. they are failures from supply areas towards the coastal zone.
delayed many thousands of years after glaciation). Sediment supply occurs in areas of hill-slope
These slides are believed to be have been triggered erosion and where river channel migration cuts
by high magnitude seismic activity associated with through areas of stored sediments resulting from
differential isostatic uplift following glacial retreat past phases of erosion and deposition under dif-
(e.g. Ballantyne, 1991). The concept of a lagged ferent climatic conditions, or from an extreme
response to impulses of change is illustrated in flood in the recent past (e.g. spreads of glacial
Figure 1.5, which also highlights that there will be deposits or floodplain alluvium). The supply of
a period of adjustment (the relaxation time), before sediment is generally intermittent, with rare
the system achieves a stable form. floods among the most effective events in deliv-
When past records are used to predict the ering sediment from these stores into the river
probability of future events, there is an underlying channel network. Once in the channel, the sedi-
assumption that there will be no change in the ment size is important in determining how far it
factors causing the events (e.g. rainfall, wave is carried before being temporarily stored in fea-
climate). This assumption of stasis in the histori- tures such as point bars or as spreads on the
cal record ignores the possibility of climate riverbed. In short rivers the suspended load
change or changes to other environmental con- may reach the estuary in a single flood, but
trols. However, over ‘engineering time’ there can coarser sediments may become incorporated in
be significant changes in the probability of trig- the floodplain.
gering events. For example, significant changes in Similar considerations apply on the coast,
climate and the pattern of geomorphological which can be viewed as a series of interlinked
events are known to be associated with the occur- physical systems, comprising both offshore and
rence of the North Atlantic Oscillation (NAO) and onshore elements. Sediment (e.g. clay, silt, sand,
El Niño Southern Oscillation (ENSO) events and gravel) is moved around the coast by waves and
volcanic eruptions (Chapters 2 and 23). currents in a series of linked systems. Simple
systems comprise an arrangement of sediment
Interconnected physical systems source areas (e.g. eroding cliffs, the sea bed),
The causes and effects of geomorphological areas where sediment is moved by coastal
events can be complex and should be viewed as processes, and sediment sinks (e.g. beaches, estu-
expressions of the operation of large physical aries or offshore sinks). Along a particular stretch
systems and not as isolated processes. Indeed, of coast there may be a series of such systems,
while the effects are readily apparent and well often operating at different scales (Figure 1.6). In
appreciated at a local level, they are not often contrast to catchments, coastal systems have no
GFE-01.qxd 6/24/05 2:43 PM Page 11
Figure 1.5 Illustration of the lag time following an impulse of change (triggering event), with a subsequent relax-
ation period over which the landforms adjust by morphological change (redrawn from Brunsden, 1980).
obvious boundaries. Suspended sediments, for The interconnected nature of physical systems
example, may be carried thousands of miles present a number of issues that will need to be
around the coast. Although headlands which addressed by most engineering projects:
appear to mark the limits of coarse sediment Changes at one location can result in a variety
transport can be identified, they are often not per- of responses elsewhere — for example, an increase
manent boundaries as material may be moved in the rate of cliff recession may be of benefit to
around these sediment divides in severe storm a nearby low-lying coast. The sand and gravel
conditions. released by erosion may build up and sustain
GFE-01.qxd 6/24/05 2:43 PM Page 12
Open boundary Cliff

erosion
Biome
Landslide
System 1 Beach
River
sediment
Offshore Onshore
Redistribution
Aeolian
sediment
NEARSHORE
Dunes
Headland
complete or
Deep
partial barrier ?
water
Submarine
canyon
Sink
System 2
Cliff erosion
Beach erosion
& deposition
pply
r su
Rive
The sea
Storm beach
Beach storage storage
Bar moves
onshore with
sea level rise
Offshore
relict storage
Variable Sediment Onshore

offshore exchange stores
stores Current
state Cliff Net loss
Offshore Also
Either gain longshore
bars Beach
or loss Storm Either
(renewable) current coastal beach gain or loss
sediment store
Relict River Net loss
offshore Volume
Net loss
deposits Estuary Either
(finite) marsh gain or loss
Strong sea level
rise effects
Figure 1.6 A simple coastal cell model (from Brunsden and Goudie, 1997).
beaches or shingle ridges which can form import- indirectly lead to beach decline and an increase in
ant components of flood defence solutions, either flood or erosion risk elsewhere.
alone or where they front embankments or sea Man can significantly alter the magnitude and
walls. Preventing cliff recession will cause a sig- frequency of geomorphological events — upland
nificant reduction in sediment supply and, hence, forestry or land clearance for infrastructure can
GFE-01.qxd 6/24/05 2:43 PM Page 13
cause a dramatic increase in the rate of soil ero- prisms of the Blyth and Dunwich estuaries, the
sion and the frequency of landslides, and may also southward growth of a shingle spit across their
lead to changes in flood behaviour further down a mouths and the collapse of the tidal deltas. By
catchment. 1328 the port had been rendered virtually useless
The cumulative effects of development can and over 400 buildings, including churches were
lead to significant changes to physical systems. In lost in one night in 1347.
Western Europe, extensive land reclamation has A widely used concept in geomorphology is that
taken place in many estuaries, since Roman times, physical systems tend towards equilibrium condi-
often involving the enclosure of saltmarshes to tions where the inputs of mass and energy to a spe-
control tidal flooding and improve grazing condi- cific system are equal to the outputs from the same
tions. Over time urban and industrial uses have system. The gross form will remain unchanged
replaced agriculture in many estuaries. The throughout these transfers. Unfortunately the con-
impacts on estuary processes have included the cept has tended to be misunderstood, generating a
modification of the tidal prism with resultant rather simpler view of the behaviour of systems
changes in the pattern of saltmarsh/mudflat (e.g. ‘beaches or rivers are in equilibrium with the
erosion and accretion. Tidal delta structures have forcing conditions’). This is probably because there
collapsed in response to the changing tidal prisms, are a variety of types of equilibrium that are asso-
resulting in accelerated erosion on the open coast- ciated with quite different patterns of behaviour
line. Indeed, it is suspected that collapse of an ebb (Figure 1.7; Chorley and Kennedy, 1971):
tide delta structure may have triggered the onset of
accelerated erosion at the former port of Dunwich, 1. Static equilibrium: no change over time
Suffolk in the UK (Lee, 2001a). Medieval land 2. Stable equilibrium: the tendency for the form
reclamation probably led to changes in the tidal to return to its original value through internal
Figure 1.7 Types of equilibrium (after Chorley and Kennedy, 1971).

GFE-01.qxd 6/24/05 2:43 PM Page 14
feedback operations within the system follow- Complex behaviour

ing a disturbance Engineers are familiar with the organised sim-
3. Unstable equilibrium: the tendency for the plicity of machines or engineering works that can
variable to respond to system disturbance by be analysed using deterministic mathematical
adjustment to a new value functions. Often, they are comfortable with those
4. Metastable equilibrium: a combination of systems that behave in a random manner (unor-
stable and unstable equilibrium except that the ganised complexity) and can be modelled using
variable settles on a new value only after having stochastic methods (e.g. floods, wave climate).
crossed some threshold value — otherwise it Most geomorphological systems lie between these
returns to the original value two states (i.e. organised complexity), showing
5. Steady-state equilibrium: the variable has elements of both deterministic and random
short-term fluctuations with a longer-term behaviour but not conforming to either.
constant mean value Geomorphological events occur when system
6. Dynamic equilibrium: the variable has short- thresholds are exceeded. These thresholds can be
term fluctuations with a changing longer-term external or internal.
mean value (i.e. an increasing or decreasing An external threshold is where an event does
trend). not occur until an external variable exceeds a crit-
ical value. For example, the minimum rainfall
intensity required to trigger debris flows and shal-
It is worth stressing that steady state is a spe- low landslides in parts of New Zealand has been
cific form of dynamic equilibrium when the mean found to be (Caine, 1980):
rate is unchanging. Perhaps the most important
ir ⫽ 14.82 D⫺0.39
aspect of dynamic equilibrium is the rate of
change of the mean rate, rather than the short- where ir ⫽ rainfall intensity (mm/hour) and
term oscillations around it. D ⫽ rainfall duration (hours).
In most systems equilibrium is achieved by Internal thresholds are where gradual changes
feedback mechanisms, involving negative feed- to the system may result in an event, without a
back mechanisms that tend to counteract and change of external influences. In semi-arid areas,
stabilise system changes by means of internal the accumulation of sediment in a valley floor
adjustments (e.g. increased erosion or storage of gradually increases the valley slope angle (long
sediment). If movement along an active fault profile) until failure (erosion) occurs. Figure 1.8
crossing an alluvial river channel causes an illustrates how in the oil shale region of western
increase in channel gradient, the resulting Colorado it has been possible to identify thresh-
increase in flow velocity will tend to promote old slope angles for drainage basins greater than
bed lowering and, hence, a reduction in channel 4 square miles; above the threshold line, valley
gradient. Positive feedback mechanisms tend to slopes tend to be unstable and affected by gully-
propagate or intensify the effects of a change, ing. Such incipient instability is common in allu-
for a limited time period. Removal of the bed vial river channels, where channel sinuosity and
armouring of gravel bed rivers can lead to accel- meander amplitude increase until a cut-off or
erated erosion of the underlying finer sediments. channel avulsion occurs. This is due to channel
An intense rainstorm can strip the upper, perme- lengthening and gradient reduction and is not nec-
able soil horizons and expose less permeable essarily related to external changes.
subsoil. This can lead to the generation of Often geomorphological events are the result of
greater runoff depths and velocities and acceler- the variable interaction of a range of factors. Figure
ated erosion, exposing another soil horizon with 1.9 highlights the complex relationship between
even lower infiltration capacity. In this way the preparatory and triggering factors that are active in
whole soil profile can be stripped off in a single promoting coastal cliff recession. There are rapid
rainstorm. temporal changes in the margin of stability of
GFE-01.qxd 6/24/05 2:44 PM Page 15
coastal cliffs due to the superimposition of trigger-

ing factors on the trends imposed by relatively
steady erosion at the base of the cliff. As the mar-
gin of stability is progressively reduced by the
operation of preparatory factors, so the minimum
size of triggering event required to initiate reces-
sion becomes smaller. Thus, triggering events of a
particular magnitude may be redundant (i.e. do not
initiate cliff recession) until preparatory factors
lower the margin of stability to a critical value.
As Figure 1.9 indicates, this can mean variable
time periods (epochs) between coastal landslides,
depending on the sequences of storm or rainfall
events. In addition, the same size triggering events
may not necessarily lead to landslides.
Another example of this complex behaviour is
Figure 1.8 Threshold line separating stable and given in Figure 1.10. As mentioned above, sediment
unstable valley floors for drainage basins of different build up causes the gradual steepening of a valley
size, Piceance Creek basin, Colorado (redrawn and floor, reducing its ability to withstand the effects of
adapted from Patton and Schumm, 1975). potentially erosive flood events. Eventually a flood
A. Interaction between declining stability and triggering events

Minimum size of triggering
event required to initiate
0.00
failure
Medium
Triggering event
Redundant event
1
High
Time
Low
B. Event sequences: redundant events and variable epochs

High
Minimum size of triggering
event required to initiate
0.00
Margin of stability
Medium
failure
1
Low
Epoch 1 Epoch 2 Epoch 3
Figure 1.9 A schematic illustration of the variable interaction between potential triggering events and coastal land-
slide (from Lee et al., 2001).
GFE-01.qxd 6/24/05 2:44 PM Page 16
Figure 1.10 The relationship between random flood events of the same magnitude and valley floor erosion. When
a flood event coincides with a period of low stability of valley floor alluvium (solid line) it causes incision and
removal of sediment (upper dashed line) (redrawn from Schumm, 1977).
of a particular magnitude will exceed the threshold of linkages between landforms within a system
of valley stability and erosion will occur. In Figure (i.e. sediment budget). A stable landscape, for
1.10, all the random flood events are of the same example, is one in which the controlling resist-
size (and return period), however only events 3, 6, 8 ances are sufficient to prevent a storm event from
and 10 will trigger valley floor erosion, the remain- having any effect.
der are redundant.
Many geomorphological events are not inde-
pendent but are influenced by the size and loca- 1.3 Techniques in engineering
tion of previous events. In other words landslides geomorphology
or gully formation, for example, are processes
with a ‘memory’, insofar as the current and future Regardless whether an engineering geomorpholo-
behaviour is influenced by the effects of past gist is concerned with the risks from geohazards
events on the system. or the impact of a project on the environment,
The response of different landscapes to the it will be necessary to develop a ‘ground model’
effects of triggering events or human activity or an understanding of the ‘total geological and
can vary considerably. For example, a storm geomorphological history’ of the site or region
passing across two catchments may result in (Fookes, 1997; Fookes, et al., 2000), at an appro-
very different patterns of erosion, deposition priate level of detail. In doing so, an engineering
and flooding. This complexity of response is a geomorphologist will often seek to establish the
measure of the sensitivity of a landscape, which geohazard potential in the area, the event history
can range from fast responding systems, which (i.e. frequency and magnitude) and the likely
are very sensitive to disturbing events (this knock-on effects of any modification or distur-
type of system can be morphologically complex bance to landforms (e.g. slope re-profiling) or
because the landforms are subject to rapid physical systems (e.g. disruption or alteration of
change, such as active sand dunes) to slowly sediment transport pathways). The techniques that
responding insensitive systems, such as ancient, can be used to achieve these goals are many and
cratonic surfaces. varied, but may include:
The main controls of sensitivity include mate-
rial strength, morphological resistance (e.g. rela- 1. terrain evaluation
tive relief, altitude, slope angle) and the importance 2. engineering geomorphological mapping
GFE-01.qxd 6/24/05 2:44 PM Page 17
3. measurement and monitoring of current rates also been used successfully on major infrastructure
of change projects in developed regions (e.g. the Channel
4. field measurements Tunnel Rail Link; Waller and Phipps, 1996).
5. analysis of climate records At the broadest scale, landscape types (terrain
6. historical research models or land systems) can be defined (e.g.
7. analysis of historical map sources mountains, desert plateaux, coastal plains); this
8. identification of event triggering conditions level of sub-division may be suitable for pre-
9. the use of tracers feasibility overviews of very large areas. Within
10. the dating of events. a landscape type it will be possible to identify
a variety of landform assemblages (terrain units),
Terrain evaluation such as river floodplains, escarpment faces,
Terrain evaluation is a method of summarising the extensive areas of unstable hill-slopes and ridge
principal physical characteristics of an area and crests; this level of detail may be sufficient for
their implications for engineering projects (e.g. corridor assessment. Within a terrain unit there
Phipps, 2001). The principles of terrain evaluation will be numerous individual landforms (terrain
involve defining areas of terrain that have similar sub-units) that will each present slightly different
physical characteristics (i.e. a typical range of levels of challenge to a project; an escarpment
topographic, geohazard and constructability fac- face, for example, may contain a variety of sub-
tors). Thus, these areas of terrain can be considered units, including bare rock faces and discrete land-
to present consistent levels of challenge that will slide systems separated by stable ridges and spurs
need to be addressed by a project. (e.g. Figures 1.11a and 1.11b).
Terrain evaluation has its origins in the need to In providing a spatial framework of terrain units
organise and communicate specific earth science the terrain specialist seeks to identify clear associ-
information or intelligence in a way that is of ations between surface forms (i.e. the terrain units),
direct relevance to the end user (e.g. civil and mil- near-surface materials and processes; simplify the
itary engineers, land use planners, agriculturalists complexity of ground conditions and surface
and foresters). An interesting summary of the processes within a particular area, highlighting
development of the approach can be found in those of significance to the planning of projects,
Mitchell (1973). Perhaps the most well-known and provide a tool for predicting terrain character-
examples of the use of terrain evaluation have istics within a particular area or region — this is
been the Commonwealth Scientific and Industrial based on the assumption that the terrain units are
Research Organisation (CSIRO) applications in sufficiently homogeneous and mutually distinctive
Australia between the 1940s and 1960s to support to allow valid prediction.
agricultural development (the Land Research and Some basic properties of terrain units and sub-
Regional Survey division) and engineering (the units can be derived from available geological and
Division of Soil Mechanics), and the Military soil maps (other useful maps include hydrogeology,
Engineering Experimental Establishment (MEXE) hydrology, geomorphology, land use and vegeta-
and Soils Laboratory at Oxford’s system for mili- tion, but are rarely available). However, it will be
tary terrain evaluation, developed in the 1960s. The necessary to make judgements about the nature and
military need for terrain information includes both significance of many issues, including geohazards
strategic (e.g. the gross distribution of terrains and and landscape sensitivity. These judgements should
transport corridors) and tactical (e.g. visibility, be based on the specialists’ experience of the broad
cross-country mobility, reaction of the terrain to association between materials (i.e. geology), land-
deformation and the availability of water and con- forms (i.e. the terrain units and sub-units) and
struction resources). processes (i.e. geohazards) that are known to occur
Terrain evaluation has tended to be associated in different climatic regions (e.g. tundra, temperate,
with rapid assessments of remote, inaccessible savannah, hot drylands, hot wetlands) and tectonic
regions (e.g. Fookes et al., 2001). However, it has zones (e.g. plate margins, cratonic areas).
GFE-01.qxd 6/24/05 2:44 PM Page 18
Figure 1.11a Terrain model 1: Tademait plateau (from Fookes et al., 2001).
Engineering geomorphological mapping 2. The characterisation of these units in terms

The production of some form of map underpins of the significant surface processes (these
many engineering geomorphological studies. may be active or relict) and the near-surface
Training and experience are the keys to producing materials.
good quality geomorphological maps. As experi- 3. The interpretation of the significance of these
enced field geologists produce the best geological forms, processes and materials to the problem
maps, so field geomorphologists produce the best facing the geomorphologist and the clients.
geomorphological maps. In addition, the ability to This may involve producing some form of
talk the same technical language as the clients derivative map (e.g. a landslide hazard map,
will help overcome the all too inevitable misun- aggregate resource map) or an extended map
derstandings over what precisely the landscape legend suitable for use by the clients or their
means to the end users. technical advisors.
The map might be the product of an intensive
The Geological Society Working Party Report
fieldwork programme, a sketch map of part of an
on maps and plans (Anon., 1972) identified exam-
area based on a walk-over survey, or interpreta-
ples of geomorphological mapping that could be of
tion of remote imagery (e.g. aerial photographs or
use to engineers. However, the value of the tech-
satellite images). Most map making follows three
nique was best highlighted by its application to
stages:
road projects in unstable terrain in Nepal and South
1. The recognition of landforms or landform ele- Wales during the early 1970s (e.g. Brunsden et al.,
ments (units) that provide a practical frame- 1975a; 1975b; Doornkamp et al., 1979; Cooke
work for addressing the problem in hand. et al., 1982; Jones et al., 1986); soil erosion (e.g.
GFE-01.qxd 6/24/05 2:44 PM Page 19
Figure 1.11b Terrain model 2: Grand Erg Occidental, Algeria (from Fookes et al., 2001).
Morgan, 1986), and river management (e.g. potential for landscape change, either through
Doornkamp, 1982; Richards et al., 1987). Despite major ‘natural’ events such as neotectonics, gully
these applications, BS5930 — Code of Practice for formation, landslides or river channel migration or
Site Investigations, contains little reference to geo- as a result of human intervention (e.g. unloading
morphological maps and the technique remains a pre-existing landslides or solifluction sheets). Field
marginal skill practised by a few experienced engi- geomorphologists should seek to understand the
neering geomorphologists. Lee (2001b) presents inheritance from previous geological periods and
an introduction to the preparation of engineering the sensitivity of the landscape to change from
geomorphology maps. Stylised symbols are often evidence of surface morphology (e.g. back-tilted
used (Demek and Embleton, 1978; Gardiner and ground is often a good indicator of deep-seated
Dackcombe, 1983; Cooke and Doornkamp, 1990). landslides), landform assemblages (e.g. the pres-
The old adage, ‘site and situation’, is clearly ence of steep, gullied terrain can be interpreted, in
important in ensuring that potential geohazards are some environments, as an indication of the potential
identified at an early stage in the investigation for channelled debris flows) and superficial
process (e.g. Fookes, 1997). Hillside sites need to deposits (e.g. clast-supported or matrix-supported
be seen in the context of the whole slope, river val- deposits can indicate flash flood or debris flow
leys as part of the whole catchment, coastal sites as deposits, respectively).
part of a sediment transport cell. Geomorphological Engineering geomorphological maps, like
maps should also provide an indication of the most other forms of map, incorporate judgements
GFE-01.qxd 6/24/05 2:44 PM Page 20
based on the available evidence (i.e. they are based Analysis of climate records
on ‘reading the ground’, Dearmen and Fookes Time series climate records can provide input data
(1974)). It is important that geomorphologists to various empirical models, such as the use of
convey an indication of the reliability of the rainfall intensity to predict soil erosion loss (e.g.
information and the uncertainties associated with EI30, the energy associated with the maximum
their judgement to the users. 30 minute rainfall intensity) or debris flow activity
(e.g. critical rainfall thresholds). Potential sand drift
Measurement and monitoring of calculation can be calculated using Fryberger’s
current rates of change (1979) method, from the average wind velocity at
The direct measurement of river or coastal change 10 m and the duration of wind.
is the most obvious method of obtaining informa- A common approach to analysing data series is
tion on current erosion or accretion rates. For to rank the records (e.g. daily rainfall in order of
example, systematic surveys undertaken on a regu- decreasing magnitude). This can yield the recur-
lar basis are a common approach to establishing rence interval (return period), as described earlier.
cliff recession rates. On the Holderness Coast, The data can be plotted on probability paper: data
Yorkshire, for example, the local authority initiated conforming to the log–normal former distribution,
a programme of cliff recession measurement in plot as a straight line on log–probability graph
1951 which has been continued on an annual basis paper; data having a Gumbel (type 1) distribution,
ever since. A series of seventy-one marker posts, plot as a straight line on extreme probability paper.
termed ‘erosion posts’ by the local authority were Often analysis of the data series can reveal
installed at 500 m intervals along 40 km of the trends (e.g. increased winter rainfall) or cyclic
coastline, each post located at a distance of fluctuations (e.g. wet-year sequences) that may be
between 50 m and 100 m normal to the coast. These important factors in controlling the frequency and
posts are replaced further inland from time to time magnitude of geomorphological events. For
if they become too close to the cliff top. Annual example, on the West Dorset coast, there is a
measurements from each post to the cliff top — direct relationship between wet-year sequences
defined as the lip of the most recent failure scar — and landslide activity. Analysis of rainfall records
commenced in 1953. The resulting database pro- from Pinhay (3 km west of Lyme Regis) for the
vides an invaluable source of medium- to long- period 1868–1998 has revealed that the annual
term measurements of cliff recession (Figure 1.1). rainfall has increased over time by around 75 mm
(i.e. there has been a 10 per cent increase in rain-
Field measurements fall over this period; the mean value is 914 mm).
On-site measurement or estimates of surface fea- A similar trend has occurred for the annual effec-
tures and materials can provide the input data for a tive rainfall (precipitation minus evapotranspira-
range of modelling or predictive tools that are used tion). This provides an indication of the amount of
by engineering geomorphologists. For example, in rainfall actually available to the groundwater
the absence of hydrological records, it is possible to tables within the slopes, i.e. a moisture balance
use careful observations of channel cross-section, index (Figure 1.12a; the mean annual value is
bed gradient and bed sediment sizes to provide an 319 mm). In order to identify wet-year sequences
approximation of the flooding and scour hazard and their frequency, a method of moving averages
(Chapter 16). Measurements of the wave climate, was used to smooth the effective rainfall (i.e. the
beach profiles and sediment sizes can be used to moisture balance index) series by replacing each
predict longshore sediment transport rates using observation with a weighted mean. Figure 1.12b
the CERC (Coastal Engineering Research Center) presents the 9-year moving mean of the data
formula (CERC, 1984). Field assessment of soil series (expressed as values relative to the mean
erodibility, slope length and steepness and crop effective rainfall) and indicates several periods of
management are key input data in the Universal increased rainfall with a possible frequency of
Soil Loss Equation for predicting soil erosion rates. 20–25 years.
GFE-01.qxd 6/24/05 2:44 PM Page 21
800
Regression line
700
600
500
Effective rainfall (mm)
400
300
200
100
0
1868 1875 1882 1889 1896 1903 1910 1917 1924 1931 1938 1945 1952 1959 1966 1973 1980 1987 1994
⫺100
⫺200
Calendar date (years)
Figure 1.12a Lyme Regis: annual effective rainfall: moisture balance index.
600
9-year moving mean
500
400
300
Effective rainfall (mm)
200
100
0
1868 1875 1882 1889 1896 1903 1910 1917 1924 1931 1938 1945 1952 1959 1966 1973 1980 1987 1994
⫺100
⫺200
⫺300
⫺400
⫺500
Figure 1.12b Lyme Regis: moisture balance index relative to the mean annual effective rainfall.
GFE-01.qxd 6/24/05 2:44 PM Page 22
7 Recorded landslide events

Number of consecutive above average years
Marine Parade—Cobb Road

6
0
1868 1875 1882 1889 1896 1903 1910 1917 1924 1931 1938 1945 1952 1959 1966 1973 1980 1987 1994
Figure 1.13 Lyme Regis: wet-year sequences and landslide events (after Bray, 1986).
Wet-year sequences were identified by calculat- allowed the identification of distinct phases of
ing the cumulative number of years with effective landslide activity, often corresponding with peri-
rainfall (moisture balance) greater than the mean ods of prolonged heavy winter rainfall, a range of
value (Figure 1.13). This value is set to zero every ground movement problems in different parts of
time it falls below the mean, on the assumption that the town and vulnerable areas affected by fre-
the groundwater levels only become critical when quent ground movement.
they rise above the average level. When researching particular events that have
been recorded in local newspapers or documents,
it is necessary to make a judgement on the relia-
Historical research bility of the data source. Three questions need to
In many countries, there is a wide range of sources be borne in mind:
that can provide useful information on the past
occurrence of events, including: aerial photo- 1. What is the nature of the event being recorded,
graphs, topographic maps, satellite imagery, pub- and with what detail, and is it pertinent to the
lic records, local newspapers, consultants, reports, stated objectives?
scientific papers, journals and diaries (e.g. 2. Who is making the report, in particular what are
Brunsden et al., 1995). his or her qualifications to know of the event, i.e.
In Ventnor, Isle of Wight, the pattern of con- is it a personal observation based on his or her
temporary ground movement was established own experience; an editing of reports from other
from a systematic search of local newspapers people, who themselves may have edited the
from 1855 to present day (e.g. Lee and Moore, information; a plausible rumour; a complete
1991). Over 200 individual incidents of ground invention, or falsification?
movement and coastal erosion were identified, 3. In the light of knowledge of this type of event,
allowing a detailed model of landslip potential to is the report credible, in whole, in part, or not
be developed. Close inspection of these records at all?
GFE-01.qxd 6/24/05 2:44 PM Page 23
Analysis of historical map sources 20 mm, irrespective of antecedent conditions,

Historical topographical maps, charts and aerial implying that combinations of antecedent soil
photographs provide a record of the former posi- moisture and rainfall below 20 mm have never
tions of various features. In many cases, historical been sufficient to attain the critical water content.
maps and charts may provide the only evidence of In contrast, every storm greater than 140 mm has
evolution over the last 100 years or more. When triggered landslides in the past. Figure 1.14b
compared with recent surveys or photographs, presents the return periods for different rainfall
these sources can give an estimate of the cumula- magnitudes — this provides a link between the
tive land loss and the average annual erosion rate thresholds and the probability of landsliding, with
between the survey dates. However, great care is the return period of the 140 mm rainfall event (the
needed in their use because of the potential accu- maximum probability threshold) around 20 years.
racy and reliability problems (e.g. Carr, 1962; Figure 1.14c highlights the probability of land-
1980; Hooke and Kain, 1982; Hooke and slides occurring within particular time periods.
Redmond, 1989). For example, the maximum threshold rainfall of
As part of the Brahmaputra River Training 140 mm has a probability of occurrence on any
Study, Thorne et al. (1993) describe how day of 0.02%, within 30 days of 0.2% and within
Bangladesh survey maps and aerial photographs a year of 5%.
from the 1950s, together with 1989 ‘SPOT’ satellite
imagery were used to assess bank-line migration. The use of tracers
The section of the river that was being studied was A wide variety of methods can be used to model
subdivided into a series of reaches, and point meas- sediment transport and water flow pathways
urements of right and left bank erosion were made through physical systems. In broad terms, the
at 0.5 km intervals. Mean erosion rates were estab- approach can help assess sediment provenance
lished by dividing the cumulative losses by the time and transport and water movement.
interval. The survey revealed that bank erosion was Fine sediment transport and sediment prove-
more severe on the right bank, where the average nance can be assessed using, for example, mineral
annual rate was 90 m (in some reaches the rate was magnetic signatures, stable isotopes and radionu-
double this figure). clides. For example, Walling and He (1999) used
fallout radionuclides to estimate rates of floodplain
Identification of event triggering accretion for a number of British rivers, using 137Cs
conditions to date the depositional sequences.
The identification of event triggering conditions Coarse sediment transport along rivers and the
often involves establishing initiating thresholds coast can be assessed using artificial sediments
between various parameters (e.g. rainfall, seismic (e.g. acoustic, electronic, magnetic or aluminium
activity) and geomorphological events. The most pebbles), painted or fluorescent coated grains
readily defined threshold is one that identifies the (luminophors), radioactive isotopes or exotic
minimum conditions (or envelope) for activity; materials. Tracer experiments on Chesil Beach,
above this, the conditions are necessary, but not for example, were used to demonstrate a net east-
always sufficient to trigger events and, below this, ward drift along the shingle beach (Carr, 1971).
there is insufficient impetus. Tsoar (1978) used coloured sand to determine
For example, Glade (1998) and Crozier and sand flow directions in desert dunes.
Glade (1999) describe how a rainfall threshold for Water movement through soil, groundwater
regions of New Zealand was defined and used to systems and caves, and in river channels can be
calculate probabilities of landslide occurrence. traced using, for example, fluorescent dyes, opti-
Figure 1.14a presents the minimum and maxi- cal brighteners, lycopodium spores and lithium
mum probability rainfall thresholds for the acetate. A useful summary of the geomorpholog-
Wellington region. The minimum threshold ical application of tracing methods can be found
required to trigger landslides appears to be in Foster (2000).
GFE-01.qxd 6/24/05 2:44 PM Page 24
Minimum probability threshold Maximum probability threshold

1,000,000
100,000 0
Number of counts
10,000 0.3
1.6
1,000 4.3
100 13.0
20.7
40.0 100
10
100
1
0.1
0.1 - 20.0
20.1 - 40.0
40.1 - 60.0
60.1 - 80.0
80.1 - 100.0
100.1 - 120.0
120.1 - 140.0
140.1 - 160.0
160.1 - 180.0
180.1- 200.0
200.1 - 220.0
220.0 - 240.0
240.1 -260.0
260.0 - 280.0
>300.1
280.1 - 300.0
Precipitation classes (mm)
Recorded landslide-triggering daily precipitation Actual daily precipitation
Figure 1.14a Probabilities (%) of landslide occurrence associated with rainfall of a given magnitude in
Wellington, New Zealand, 1862–1995 (note: a value of 50 means that 50% of all measured daily rainfalls in a given
category produced landslides in the past) (after Crozier and Glade, 1999).
400
y = 27.695 Ln(x) + 64.571

R2 = 0.9832
300
Precipitation (mm)
200
100
y = 16.825 Ln(x) + 75.102
R2 = 0.9723
0
0.001 0.01 0.01 1 10 100 1000
Returm period
Figure 1.14b Return periods of daily precipitation and fitted logarithmic trend lines in the Wellington region,
New Zealand (after Crozier and Glade, 1999).
Dating of events of the time elapsed since biological or inorganic

A number of methods are available to obtain dates fixation. Specialist applications include the use
for episodic events and thereby to build up a pic- of an accelerator mass spectrometer to date the
ture of the event history of the area, including: organic material included in varnish layers char-
Radioactive dating, in which measurement of acteristic of many desert rock surfaces or stone
the ratio of stable (12C) to radioactive (14C) iso- pavements. The age range for radiocarbon dat-
topes in an organic sample (e.g. charcoal, wood, ing is from a few centuries to around 40 000
peat, roots, macrofossils) allows the determination years).
GFE-01.qxd 6/24/05 2:44 PM Page 25
100
Maximum threshold of
landslide-triggering rainfail
Probability of occurence (%)
10
1
P365
P180
P90
0.1
P30
P10
0.01
Minimum threshold of
landslide-triggering rainfail
P1
0.001
0 50 100 150 200 250
Precipitation (mm)
Figure 1.14c Probability of occurrence of daily precipitation equalling or exceeding given values in Wellington,
New Zealand. Note: the different lines refer to the probability of occurrence of specific rainfall magnitude on each
single day (P1), within a period of 10 days (P10), within a month (P30), etc. The empirically established minimum
and maximum thresholds of landslide triggering rainfall (20 mm and 140 mm) are shown by the thin vertical lines
(after Crozier and Glade, 1999).
Lichenometric dating, involving the measure- optically stimulated luminescence. The maximum
ment of the maximum lichen size on exposed age range for these methods depends on the min-
surfaces (e.g. boulders) and comparing with eral: for quartz it is around 100 000–150 000
lichen-growth curves. This is assumed to indicate years; for feldspars it is around 800 000 years.
the time elapsed since a particular event (e.g. Loesses have been successfully dated up to
rockfall, debris flow, flash floods). 800 000 years old. Other applications include
Dendrochronology, whereby the age of death dating alluvial fans and aeolian deposits.
of a tree (e.g. buried beneath a landslide) may be Uranium series dating is applicable for the time
determined by cross-reference to a master tree- range ⬍1000 to over 500 000 years. It involves
ring chronology for that region. The oldest rings testing surface sediments to establish the radioac-
of trees that have colonised a now stabilised slide tive disequilibrium in the decay chains of Uranium
or debris fan will give a minimum age for the and related isotopes such as Thorium (e.g. the ratio
event. between 230Th and 234U). The method has been
Luminescence dating, based on estimating the used to date calcretes and halite deposits, or crusts
time since a sediment was last exposed to day- beneath landslides.
light, which ‘zeroes’ the previously accumulated Cosmogenic nuclide dating, whereby the con-
radiation damage to minerals (e.g. quartz or centration of cosmogenic nuclides (e.g. 3He,
10
feldspar) in the sample. The age of a sample is Be, 14C, 21Ne, 26Al and 36Cl) produced by in situ
derived from: nuclear reactions (the interaction between cosmic
rays and terrestrial atoms) in the upper few
Palaeodose metres of the ground surface are detected by
Age ⫽
Dose rate accelerator mass spectrometry. The concentra-
tions can be used to determine the length of time
where the palaeodose is the accumulated radia- that material has spent at or near the ground
tion damage and the dose rate is the rate at which surface. These methods have been used to date
the sample absorbs energy from the immediate rockfalls, basalt flows and alluvial fan surfaces.
proximity. Palaeodoses are calculated by thermo- The datable range is from 1000 years to several
luminescence (energy is supplied by an oven) or million years.
GFE-01.qxd 6/24/05 2:44 PM Page 26
SOME ENGINEERING GEOLOGY / GEOMORPHOLOGY landslide studies are provided by Stokes (1997)
RELATIONSHIPS and Lang et al. (1999) respectively.
DEEPER SHALLOWER
STRONGER WEAKER 1.4 Summary
ENGINEERING ENGINEERING
GEOLOGY GEOMORPHOLOGY
Geomorphology studies the earth’s landforms and
ER SURFA landform change. Engineering geomorphology
WAT CE WA
UND URFA
CE TER
complements engineering geology in providing a
GRO S U B -S S SUR
ESSE PRO FACE
ROCK PROC CES SOIL spatial context for explaining the nature and dis-
SES
SOIL
MECHANICS ROCK IALS MECHANICS
MATE
R MATER
IALS
tribution of particular ground-related problems
DATIO
N ENGINE
ERING
and resources.
FOUN
Importantly, engineering geomorphology is also
concerned with evaluating the implications of
CIVIL landform changes for society and the environment.
ENGINEERING
This book is largely by geomorphologists and
ENGINE S
its aim is to produce a basic yet authoritative hand-
ERING MATERIAL
book for geotechnical engineers and others. Its
essence is in its diagrams around which each
STRUCTURAL HYDRAULIC
ENGINEERING ENGINEERING chapter is built. The first part is concerned with
controls which have shaped the landscape, the sec-
ARCHITECTURAL ond with processes and the third is concerned with
BUILDING
the various environments and landforms.
The application of geomorphological knowl-
Figure 1.15 Some engineering, geology and geomor-
phology relationships.
edge to civil engineering will come about by the
education of both engineers and geomorpho-
logists. A relationship between engineering,
Cation-ratio rock varnish dating method, based geomorphology and geology is summarised in
on the observation that the ratio of certain cations Figure 1.15.
(K, Ca, Ti) in rock varnish decreases with age as
a result of preferential leaching. Cation-ratios can
be calculated by photon-induced X-ray emission References
spectrometry, electron microprobe or inductively
coupled plasma spectroscopic (ICP) methods. Ager, D. L. (1976) The Nature of the Stratigraphic
Comparing ratios from a number of samples can Record. Methuen, London.
generate relative ages. Calibrated ages can be Anon. (1972) The preparation of maps and plans in
terms of engineering geology. Quarterly Journal of
determined by comparison with curves generated Engineering Geology 5, 293–381.
for samples of known age. Ballantyne, C. K. (1991) Holocene geomorphic activity
Amino acid dating methods provide relative in the Scottish Highlands. Scottish Geographical
ages by measuring the extent to which certain Magazine 107, 84–98.
amino acids within protein (e.g. bones, mollusc Bray, M. J. (1986) A Geomorphological Investigation of
the South West Dorset Coast, Volume 1: Patterns of
shells, eggshells) residues has transformed from Sediment Transport. Report to Dorset Country
one of two chemically identical forms to the other Council, 798 pp.
(e.g. the transformation from L-form to D-form Brunsden, D. (1980) Applicable models of long-term
amino acids) until equilibrium is reached. The landform evolution. Zeitschrift für Geomorphologie
methods can be used for timescales from a few Supplementband 36, 16–26.
Brunsden, D. (2002) Geomorphological roulette for
years to hundreds of thousands of years. engineers and planners: some insights into an old
Useful discussions of the suitability of the game. Quarterly Journal of Engineering Geology
different methods in desert environments and and Hydrology 35, 101–142.
GFE-01.qxd 6/24/05 2:44 PM Page 27
Brunsden, D. and Goudie, A.S. (1997) Classic Dearman, W. R. and Fookes P. G. (1974) Engineering
Landforms of the West Dorset Coast. Geographical geological mapping for civil engineering practice in
Asociation, Sheffield, 56 p. the United Kingdom. Quarterly Journal of
Brunsden, D., Doornkamp, J. C., Fookes, P. G., Jones, Engineering Geology 7, 223–256.
D. K. C. and Kelly, J. M. N. (1975a) Large scale Demek, J. and Embleton, C. (eds) (1978) Guide
geomorphological mapping and highway engineer- to Medium-scale Geomorphological Mapping.
ing design. Quarterly Journal of Engineering International Geographical Union, Brno.
Geology 8, 227–253. Doornkamp, J. C. (1982) The physical basis for plan-
Brunsden, D., Doornkamp, J. C., Hinch, L. W. and ning in the Third World. Third World Planning
Jones, D. K. C. (1975b) Geomorphological map- Review 4, 11–31.
ping and highway design. Sixth Regional Doornkamp, J. C. Brunsden, D., Jones, D. K. C.,
Conference for Africa on Soil Mechanics and Cooke, R. U. and Bush, P. R. (1979) Rapid geomor-
Foundation Engineering, 3–9. phological assessments for engineering. Quarterly
Brunsden, D., Ibsen, M-L., Lee, E. M. and Journal of Engineering Geology 12, 189–204.
Moore, R. (1995) The validity of temporal archive Fookes, P. G. (1997) The First Glossop Lecture:
records for geomorphological purposes. Quaestiones Geology for engineers: the geological model, predic-
Geographicae Special Issue 4, 79–92. tion and performance. Quarterly Journal of
Caine, N. (1980) The rainfall intensity-duration control Engineering Geology 30, 290–424.
of shallow landslides and debris flows. Geografiska Fookes, P. G., Baynes, F. J. and Hutchinson, J. N. (2000)
Annaler 62, 23–27. Keynote Lecuture: Total geological history: a model
Carr, A. P. (1962) Cartographic error and historical approach to the anticipation, observation and under-
accuracy. Geography 47, 135–144. standing of site conditions. GeoEng 2000, an
Carr, A. P. (1971) Experiments on longshore International Conference on Geotechnical and
transport and sorting of pebbles: Chesil Beach, Geological Engineering, Melbourne 1, 370–460.
England. Journal of Sedimentary Petrology 41(4), Fookes, P. G., Lee E. M. and Sweeney, M. (2001)
1084–1104. Terrain evaluation for pipeline route selection and
Carr, A. P. (1980) The significance of cartographic characterisation, Algeria. In Griffiths, J. S. (ed.)
sources in determining coastal change. In Land Surface Evaluation for Engineering Practice.
Cullingord, R. A., Davidson, D. A. and Lewin, J. Geol. Soc. Special Publication 18, 115–121.
(eds) Timescales in Geomorphology. Wiley, Foster, I. D. L. (2000) Tracers in Geomorphology.
Chichester, 67–78. Wiley, Chichester.
CERC (1984) Shore Protection Manual (4th edn). Fryberger, S. G. (1979) Dune forms and wind regimes. In
Coastal Engineering Research Center, US Corps of McKee, E. D. (ed.) A Study of Global Sand Seas. US
Engineers. Geological Survey Professional Paper 1052, 137–169.
Chorley, R. J. and Kennedy, B. A. (1971) Physical Gardiner, V. and Dackombe, R. (1983) Geomorpho-
Geography: A Systems Approach. Prentice-Hall, logical Field Manual. Allen & Unwin, London.
London. Glade, T. (1998) Establishing the frequency and magni-
Chorley, R. J., Schumm, S. A. and Sugden, D. E. (1984) tude of landslide-triggering rainstorm events in
Geomorphology. Methuen, London. New Zealand. Environmental Geology 35, 16–174.
Coates, D. R. (ed.) (1971) Environmental Geomor- Hails, J. R. (ed.) (1977) Applied Geomorphology.
phology. Publications in Geomorphology, State Elsevier, Amsterdam.
University of New York, Binghampton, NY. Hansom, J. (1998) The coastal geomorphology of
Cooke, R. U. (1984) Geomorphological Hazards in Los Scotland: understanding sediment budgets for effec-
Angeles. London Research Series in Geography 7. tive coastal management. In Scotland’s Living
Allen & Unwin, London. Coastline, 34–44. Stationery Office.
Cooke, R. U. and Doornkamp, J. C. (1974) Geomor- Hooke, J. M. and Kain, R. J. P. (1982) Historical
phology in Environmental Management. Oxford Change in the Physical Environment: a Guide to
University Press. Sources and Techniques. Butterworth, Seven Oaks.
Cooke, R. U. and Doornkamp, J. C. (1990) Geomor- Hooke, J. M. and Redmond, C. E. (1989) Use of carto-
phology in Environmental Management (2nd edn). graphic sources for analysing river channel change
Oxford University Press. with examples from Britain. In Petts, G. E., Moller, H.
Cooke, R. U., Brunsden, D., Doornkamp, J. C. and and Roux, A. L. (eds) Historical Change of Large
Jones, D. K. C. (1982) Urban Geomorphology in Alluvial Rivers: Western Europe. Wiley, Chichester,
Drylands. Oxford University Press. 79–94.
Crozier, M. J. and Glade, T. (1999) Frequency and Jones, D. K. C., Cooke, R. U. and Warren, A. (1986)
magnitude of landsliding: fundamental research Geomorphological investigation, for engineering
issues. Zeitschrift für Geomorphologie NF Suppl. purposes, of blowing sand and dust hazard. Quarterly
Bd. 115, 141–155. Journal of Engineering Geology 19, 251–270.
GFE-01.qxd 6/24/05 2:44 PM Page 28
Lang, A., Moya, J., Corominas, J., Schrott, L. and Bristow, C. S. (eds) Braided Rivers. Geological
Dikau, R. (1999) Classic and new dating methods Society Special Publication 75, 257–276.
for assessing the temporal occurrence of mass Thorne, C. R., Hey, R. D. and Newson, M. D. (eds)
movements. Geomorphology 30, 33–52. (1997) Applied Fluvial Geomorphology for River
Lee, E. M. (2001a) Estuaries and coasts: morphological Engineering and Management. Wiley, Chichester.
adjustment and process domains. In Higgett, D. and Tsoar, H. (1978) The dynamics of longitudinal dunes.
Lee, E. M. (eds) Geomorphological Processes and US Army European Research Office Grant No.
Landscape Change: Britain in the Last 1000 Years. DA-ERO 76-G-072. Department of Geography, Ben
Blackwell, Oxford, 147–189. Gurion University of the Negev.
Lee, E. M. (2001b) Geomorphological mapping. In Twidale, C. R. (1985) Ancient landscapes: their nature
Griffiths, J. S. (ed.) Land Surface Evaluation for and significance for the question of inheritance. In
Engineering Practice. Geological Society Special Hayden, R. S. (ed.) Global Mega-Geomorphology.
Publication 18, 53–56. NASA Conference Publication, 2312, 29–40.
Lee, E. M. and Moor, R. (1991) Coastal Landslip Waller, A. M. and Phipps P. J. (1996) Terrain systems
Potential: Ventnor, Isle of Wight. GSL Publications, mapping and geomorphological studies for the
Newport Pagnell. Channel Tunnel Rail Link. In Craig C. (ed.)
Lee, E. M., Hall, J. W. and Meadowcroft, I. C. (2001) Advances in Site Investigation Practice. Thomas
Coastal cliff recession: the use of probabilistic pre- Telford, London, 25–38.
diction methods. Geomorphology 40, 253–269. Walling, D. E. and He, Q. (1999) Changing rates of over-
Mitchell, C. W. (1973) Terrain Evaluation. Longman, bank sedimentation on the floodplains of British
London. rivers during the past 100 years. In Brown, A. G.
Morgan, R. P. C. (1986) Soil Erosion and Conservation. and Quine, T. A. (eds) Fluvial Processes and
Longman, London. Environmental Change. Wiley, Chichester, 207–222.
Patton, P. C. and Schumm, S. A. (1975) Gully erosion, Warburton, J., Davies, T. R. H. and Mandl, M. G.
Northwestern Colorado: a threshold phenomenon. (1993) A meso-scale field investigation of channel
Geology 1, 88–90. change and floodplain characteristics in an upland
Pethick, J. (1996) Coastal slope development: temporal braided gravel-bed river, New Zealand. In Best, J. L.
and spatial periodicity in the Holderness Cliff and Bristow, C. S. (eds) Braided Rivers. Geological
Recession. In Anderson, M. G. and Brooks, S. M. Society Special Publication 75, 241–255.
(eds) Advances in Hillslope Processes 2, 897–917.
Phipps, P. J. (2001) Terrain systems mapping. In
Griffiths, J. S. (ed.) Land Surface Evaluation for
Engineering Practice. Geol. Soc. Special Publication Further reading
18, 59–61.
Richards, K. S., Brunsden, D., Jones, D. K. C. and Coates, D. R. (1976) Geomorphology and Engineering.
McCaig, M. (1987) Applied fluvial geomorphology: Allen & Unwin, London.
river engineering project appraisal in its geomorpho- Coates, D. R. (ed.) (1976) Urban Geomorphology.
logical context. In Richards, K. S. (ed.) River Geological Society of America Special Paper 174.
Channels: Environment and Process. Blackwell, Costa, J. E and Fleisher, P. J. (eds) (1984) Developments
Oxford, 348–382. and Applications of Geomorphology. Springer-
Schumm, S. A. (1977) The Fluvial System. Wiley, Verlag, Berlin.
Chichester. Cooke, R. U. and Doornkamp, J. C. (1990) Geomor-
Skempton, A. W. (1979) Landmarks in early soil phology in Environmental Management (2nd edn.)
mechanics. Proceedings of 7th European Conference Oxford University Press.
on Soil Mechanics and Foundation Engineering, Cooke, R. U., Brunsden, D., Doornkamp, J. C. and
Brighton, Vol. 5, 1–26. Jones, D. K. C. (1982) Urban Geomorphology in
Skempton, A. W. (1985) A history of soil properties, Drylands. Oxford University Press.
1717–1927. Proceedings of 11th International Craig, R. G. and Croft, J. L. (1982) Applied
Conference on Soil Mechanics and Foundation Geomorphology. Allen & Unwin, London.
Engineering, San Francisco. Gardiner, V. and Dackombe, R. (1983) Geomorpho-
Stokes, S. (1997) Dating of desert sequences. In logical Field Manual. Allen & Unwin, London.
Thomas, D. S. G. (ed.) Arid Zone Geomorphology. Gregory, K. J. and Walling, D. E. (1979) Man and
Wiley, Chichester, 607–638. Environmental Processes. W. Dawson, Folkestone,
Thorne, C. R., Russell, A. P. G. and Alam, M. K. (1993) England.
Planform pattern and channel evolution of the Verstappen, H. Th. (1983) Applied Geomorphology.
Brahmaputra River, Bangladesh. In Best, J. L. and Elsevier, Amsterdam.
GFE-02.qxd 6/24/05 2:45 PM Page 29
Part I
Controls
GFE-02.qxd 6/24/05 2:45 PM Page 30
GFE-02.qxd 6/24/05 2:45 PM Page 31
2. Climate and Weathering

Mark Lee and Peter Fookes
2.1 Climate example, can only occur where ground tempera-

tures fall below freezing, while permafrost will
Climate is the average, long-term summary of the only develop where the mean annual temperature
weather patterns in a region, with the short-term is at least below 0 °C. Wind action is most effect-
extremes and transient weather conditions ive in dry regions with less than 200 mm mean
smoothed out to give average values, for example, annual rainfall. However, in many regions it may
of annual or monthly temperature, precipitation be the extremes, such as high intensity rainfall
and wind regime. It is a major influence on the events, that are responsible for triggering processes
rate, scale and significance of current geomor- or the development of landforms and the behaviour
phological processes and weathering. The imprint of man-made structures.
of past climates can be seen in the variety of The variations in surface temperature and avail-
inherited features that are characteristic of many able kinetic energy, along with geological structure
landscapes, from duricrusts in hot drylands to the (tectonics) and lithology (stratigraphy), and local
relict, periglacial landslides of many temperate differences in the potential energy arising from
regions and solution caverns in limestone terrain. the height of material above base level, provide the
Climate is also an important consideration in impetus for global and regional differences in the
many aspects of engineering, from the prepara- intensity of geomorphological processes. Three
tion and performance of concretes, to the timing major climate types can be recognised, reflecting
of earth-moving operations and weathering of global variations in precipitation (Figure 2.3) and
building stone. temperature (Figure 2.4).
The importance of climate stems from the fact
that solar radiation is a key source of kinetic
energy for geomorphological processes, along 2.2 Tropical climates
with tidal energy generated by the gravitational
attraction of the Sun and the Moon and rota- Equatorial Trough (Inter-Tropical Convergence
tional energy derived from the momentum of the Zone): this represents an area of low pressure near
Earth’s rotation (Figure 2.1; note that endogenic the equator towards which persistent winds (trade
processes such as volcanism, seismic activity and winds) blow — from the NE in the northern hemi-
tectonics are driven by geothermal energy). Solar sphere and from the SE in the southern hemi-
energy drives the hydrological cycle which sphere. This zone is associated with a relatively
involves the continuous exchange of water from uniform climate (i.e. little monthly variation) with
oceans to the continents via the atmosphere and high rainfall, temperature and humidity, and sup-
back again as runoff (Figure 2.2). The input of porting characteristic tropical rainforest vegetation
solar radiation varies between the equator and the (Figure 2.5a).
poles, with the resulting thermal gradients giving Sub-Tropical High Pressure Zone: this is char-
rise to the circulation of air in the atmosphere acterised by subsiding warm, dry air from air cir-
(winds) and water in the ocean (ocean currents). culation cells centred on the equator, known as
There are clear links between climate and Hadley cells. These regions are generally cloud
geomorphological processes. Frost action, for free and arid, giving rise to major deserts (e.g. the
31
GFE-02.qxd 6/24/05 2:45 PM Page 32
Solar radiation intercepted by the earth

17.8ⴛ 1016 W
Exogenic processes
Solar energy to the atmosphere Hydrological cycle Fluvial and glacial

12.3ⴛ 1016 W 4 ⴛ 1016 W processes
Kinetic energy (Wind)

Aeolian processes
7ⴛ 1014 W
Coastal waves and
currents 5ⴛ 1012 W
Ocean tides Coastal processes
2.2ⴛ 1012 W
Transfer of heat to
Rotational energy of the Volcanism
surface
solar system 2.9 ⴛ 1012 W
Seismicity
Geothermal heat flow
Solid earth tides Global 36 ⴛ 1012 W
0.4 ⴛ 1012 W Continents 11 ⴛ 1212 W
Oceans 25 ⴛ 1012 W Tectonics
Earth cooling and

Endogenic processes
radioactive decay
Figure 2.1 Estimated energy flows relevant to different geomorphological processes (from Summerfield, 1991).
Horizontal atmospheric transfer of water

vapour
79 Precipitation on 12
Evaporation from oceans
continents
Evapotranspiration from
Precipitation in oceans continents
21
88
Runoff to oceans
Figure 2.2 The hydrological cycle, with figures representing percentages of the mean annual global precipitation
of c. 1000 mm (from Summerfield, 1991).
GFE-02.qxd 6/24/05 2:45 PM Page 33
CLIMATE AND WEATHERING 33
Figure 2.3 Global mean annual precipitation (from Briggs and Smithson, 1992).
Sahara, Kalahari and Great Australian Deserts). experience arid or semi-arid conditions. In
The climate typically involves extreme tempera- Somalia, mean annual precipitation is below
tures — the desert can be very cold at night, espe- 250 mm, although the area is only just north of the
cially in winter — and very rare, intense equator (Figure 2.5d).
cloudbursts (Figure 2.5b).
Monsoons: the equatorial climates of West
Africa, India and South East Asia are charac- 2.3 Temperate climates
terised by marked wet and dry seasons. The wet
season, the monsoon, is generated by the sea- Mediterranean: these regions are dominated by
sonal migration of the equatorial low-pressure variable westerly winds with depressions and
trough as the adjacent continental masses heat anticyclones. In winter, depressions are more fre-
up in summer months. This allows moist south quent because of the migration of the sub-tropical
and south-west winds to bring intense rain high-pressure zones towards the equator. This
(Figure 2.5c). leads to a climate of mild wet winters and hot dry
Hurricanes, Typhoons and Tropical Cyclones: summers (Figure 2.5e). Rainfall can be highly
these intense storms develop over the warmer variable, with very dry years interspersed with
parts of the oceans during summer months. Once sudden flooding or prolonged heavy rainfall often
developed they move westwards, within the trade resulting in widespread hillside erosion and land-
winds, gradually increasing in intensity. The slide activity.
storms decay over land, but not before causing Maritime: on the western side of continents in
widespread erosion and flooding on exposed higher latitudes, the influence of depressions is felt
coastlines, or landslides in mountainous areas throughout the year, producing less seasonality to
such as Hong Kong and the Philippines. the rainfall pattern (Figure 2.5f). However, extreme
Equatorial Arid Zone: a few tropical regions storms can occur. In Britain, the heaviest daily falls
lie beyond the influence of tropical storms and recorded include 297.4 mm on 18 July 1955 at
GFE-02.qxd 6/24/05 2:45 PM Page 34
Figure 2.4 Mean sea-level temperatures in January (above) and July (below) in degrees centigrade (from Briggs
and Smithson, 1992).
GFE-02.qxd 6/24/05 2:45 PM Page 35
Manaos (Brazil), 44m New Delhi (India), 218m

40 50
Highest
35
30 Mean max
˚C
40
25 Mean min
20 Lowest
30
250
Highest
200
˚C
20 Mean max
150 Mean annual
mm
precipitation–1811mm
100
10
50
Mean min
0
J FMAMJ J A SOND
0
Figure 2.5a Climate graphs for the equatorial trough Lowest
zone: Manaos, Brazil (from Briggs and Smithson, 200
1992).
Mean annual
mm
100 precipitation - 640mm

Arica (Chile), 29m
35
30
Highest 0
J F M A M J J A S O N D
25 Mean max
Figure 2.5c Climate graphs for the monsoon climate:
20
New Delhi, India (from Briggs and Smithson, 1992).
˚C
Mean min
15
Lowest
10 coast, this climate is associated with hot sum-
5
mers, cold winters and precipitation spread
evenly through the year. Away from the coast,
0 temperatures become more extreme. Winnipeg,
Mean annual Canada experiences a temperature range from
mm
10
precipitation–⬍0.3mm
42 °C in the summer to ⫺48 °C in the winter.
0
J F M A M J J A S O N D Rainfall gradually decreases away from the coast,
leading to semi-arid or arid deserts (Figure 2.5g).
Figure 2.5b Climate graphs for the sub-tropical high Northern Extremes: relatively warm ocean cur-
pressure zone: Arica, Chile (from Briggs and Smithson,
1992). rents and high latitude depression tracks allow the
fringes of the maritime climate zone to extend
towards the poles. In these areas the climate is cool,
Martinstown, Dorset; 242.8 mm on 28 June 1917 at moist and cloudy with frequent gales (Figure 2.5h).
Bruton, Somerset and 238.8 mm on 18 August The coastal waters generally remain ice-free.
1924 at Cannington, Somerset. The cooling influ-
ence of the sea reduces summer temperature but
generally minimises the duration and severity of 2.4 Polar climates
winter freezing.
Continental: on the eastern side of high lati- Tundra: this region is associated with permanently
tude continents the prevailing airflow is offshore, frozen ground (permafrost) and a characteristic
so the maritime influence is reduced. Near the vegetation of mosses, lichens, sedges, grasses and
GFE-02.qxd 6/24/05 2:45 PM Page 36
Galcaio (Somalia), 240m

Perth (Western Australia), 60m
50
40 Highest
40 30
Highest
Mean max
Mean max
˚C
20
˚C
30
Mean min
10
Lowest
20 Mean min
0
Lowest
10 Mean annual
mm
100 precipitation–881mm
Mean annual 0
mm
50 precipitation–150mm J F MA M J J A S O N D
Figure 2.5e Climate graphs for the mediterranean cli-

mate: Perth, Australia (from Briggs and Smithson, 1992).
0
J F MA M J J A SO N D
Vancouver (Canada), 14m
Figure 2.5d Climate graphs for the equatorial dry
zone: Galcaio, Somalia (from Briggs and Smithson, 30
1992).
20
a few birch trees. Winter temperatures are very Highest
low. Snowfalls may be low because of continental 10
˚C
effects. Summer temperatures can rise as high Mean max

as 30 °C, although precipitation remains low Mean min
(Figure 2.5i). 0
Arctic: this region is dominated by sea-ice

and experiences low winter temperatures and pre- ⫺10
cipitation. In summer, 24-hour daylight allows
Lowest
increased receipt of solar radiation. However,
much of the energy is spent on melting surface ice 200
and evaporation, so that temperatures rarely rise
above 5 °C. The combination of moisture from Mean annual
mm
precipitation–1458mm
melting ice and relatively low temperatures leads 100
to dense, saturated air and abundant cloud cover.

However, precipitation remains low because con- 0
ditions for rising air are infrequent. JFMAMJ JASOND
Antarctic: the effects of latitude, altitude (over Figure 2.5f Climate graphs for the maritime cli-
50% of the continent is above 2000 m) and isolation mate, Vancouver, Canada (from Briggs and Smithson,
from other continents results in much of the region 1992).
GFE-02.qxd 6/24/05 2:45 PM Page 37
Kazalinsk (USSR), 63m Reykjavik (Iceland), 28m
40 20
Highest
30 10
Mean max
20
˚C
0 Mean min
Highest
10
⫺10
˚C
Lowest
0
⫺20
Mean max
100
⫺10 Mean min Mean annual
mm
0
⫺20
Figure 2.5h Climate graphs for the maritime northern
extreme, Reykjavik, Iceland (from Briggs and Smithson,
⫺30 Lowest 1992).
50
Mean annual
mm
Yakutz (USSR), 163m
0 40
30
Figure 2.5g Climate graphs for the continental climate,
Kazalinsk (CIS, former Soviet Union) (from Briggs and 20
Smithson, 1992). 10
0
being ice covered. Depressions rarely penetrate Highest
into the region, although low pressure prevails ⫺10
˚C
because of the dense cold air. Precipitation is very ⫺20

low. Cold temperatures prevail throughout the year,
⫺30
seldom rising above 0 °C (Figure 2.5j). Persistent,
strong winds blow off the ice caps, reaching aver- ⫺40 Mean max
age speeds of over 6 m/s. Near the coast the winds Mean min
⫺50
strengthen as they are funnelled by steep glacier
valleys. At Cape Denison, for example, the annual ⫺60 Lowest
mean wind speed is around 20 m/s. 40
Mean annual
mm
20 precipitation - 188mm
2.5 Climate classification 0
The classifications developed by Köppen (1931; Figure 2.5i Climate graphs for the tundra climate,
1936) and Thornthwaite (1933) provide a useful Yakutz (Siberia) (from Briggs and Smithson, 1992).
GFE-02.qxd 6/24/05 2:45 PM Page 38
South Pole, 2800m by the total monthly evaporation summed for each
⫺10 of the twelve months of the year. Six temperature
provinces can be defined:
⫺20 Highest
Mean max Temperature province T/E index

⫺30 Mean min
A⬘ Tropical ⬎ 128
Lowest
⫺40 B⬘ Mesothermal 64–127
C⬘ Microthermal 32–63
⫺50 D⬘ Taiga 16–31
E⬘ Tundra 1–15
⫺60 F⬘ Frost 0
⫺70
These two indexes were combined to produce a
⫺80
world map showing thirty-two climatic types
(Figure 2.6). For example, the climate of western
⫺90 Britain is described as BC⬘r, whereas the Amazon
Basin is BA⬘r.
Figure 2.5j Climate graphs for the Antarctic climate: The important influence of climate on geo-
South Pole (from Briggs and Smithson, 1992). morphological processes has been expressed in
terms of morphoclimatic zones within which
guide to support the development of a preliminary there are distinctive landform assemblages and
ground model for a region. The latter is based on a dominant processes (e.g. Figure 2.7; Table 2.1).
combination of effective precipitation (i.e. precip- Two broad groupings of morphoclimatic regions
itation minus evaporation) and thermal efficiency. can be identified:
The former is related to the P/E index (i.e. the total
monthly rainfall is divided by the total monthly 1. First-order morphoclimatic zones: glacial,
evaporation summed for each of the twelve arid and humid tropical. These have non-
months of the year). Five humidity provinces were seasonal processes, generally low average ero-
proposed: sion rates, highly infrequent and episodic
erosive activity (e.g. desert rainstorms).
Humidity Characteristic P/E index
2. Second-order morphoclimatic zones: humid–
province vegetation
arid tropical, semi-arid tropical, dry continen-
A Wet Rainforest ⬎ 128 tal, wet mid-latitude and periglacial. These
B Humid Forest 64–127 have distinctly seasonal processes. In regions
C Sub-humid Grassland 32–63 with warmer climates, the processes differ
D Semi-arid Steppe 16–31 most significantly in terms of the length of the
E Arid Desert ⬍ 16 wet season. In cooler regions, variations in
summer temperature and precipitation are
These five provinces can be further sub-divided important.
on the basis of rainfall seasonality:
r rainfall abundant in all seasons While these morphoclimatic zones provide a

s rainfall deficient in summer starting point for developing global or site-scale
w rainfall deficient in winter models of anticipated ground conditions, it can be
d rainfall deficient in all seasons. misleading to seek a link between present-day
climate and a particular landscape. All current
The thermal efficiency is expressed as the T/E processes act on a stage fashioned by ancient
index, defined as the monthly temperature divided processes (i.e. many landscapes comprise a
GFE-02.qxd 6/24/05 2:45 PM Page 39
Figure 2.6 World climate based on Thornthwaite’s classification (from Doornkamp, 1986).
Figure 2.7 World morphoclimatic zones, excluding highlands (from Kiewiet de Jonge, 1984).
GFE-02.qxd 6/24/05 2:45 PM Page 40
Table 2.1 Major morphoclimatic zones (from Fookes et al., 2000).
Morphoclimatic Mean annual Mean annual Relative importance of

zone temperature (⬚C) precipitation (mm) geomorphological processes
Azonal mountain Highly variable Highly variable Rates of all processes vary significantly with
zone altitude; mechanical and glacial action become
significant at high elevations.
Glacial ⬍0 0–1000 Mechanical weathering rates (especially frost
action) high; chemical weathering rates low; mass
movement rates low except locally; fluvial action
confined to seasonal melt; glacial action at a
maximum; wind action significant.
Periglacial ⫺1 to ⫹2 100–1000 Mechanical weathering very active with frost
action at a maximum; chemical weathering rates
low to moderate; mass movement very active;
fluvial processes seasonally active; wind action
rates locally high. Effects of repeated formation
and decay of permafrost.
Wet mid-latitude 0–20 400–1800 Chemical weathering rates moderate, increasing
to high at lower latitudes; mechanical weathering
activity moderate with frost action important at
higher latitudes; mass movement activity moderate
to high; moderate rates of fluvial processes;
wind action confined to coasts.
Dry continental 0–10 100–400 Chemical weathering rates low to moderate;
mechanical weathering, especially frost action,
seasonally active; mass movement moderate and
episodic; fluvial processes active in wet season;
wind action locally moderate.
Hot dry (arid 10–30 0–300 Mechanical weathering rates high (especially
tropical) salt weathering); chemical weathering minimal;
mass movement minimal; rates of fluvial activity
generally very low but sporadically high;
wind action at maximum.
Hot semi-dry 10–30 300–600 Chemical weathering rates moderate to low;
(semi-arid tropical) mechanical weathering locally active especially
on drier and cooler margins; mass movement
locally active but sporadic; fluvial action rates
high but episodic; wind action moderate to high.
Hot wet-dry 20–30 600–1500 Chemical weathering active during wet season;
(humid-arid rates of mechanical weathering low to moderate;
tropical) mass movement fairly active; fluvial action high
during wet season with overland and channel flow;
wind action generally minimal but locally
moderate in dry season.
Hot wet (humid 20–30 ⬎ 1500 High potential rates of chemical weathering;
tropical) mechanical weathering limited; active, highly
episodic mass movement; moderate to low rates
of stream corrasion but locally high rates of
dissolved and suspended load transport.
GFE-02.qxd 6/24/05 2:45 PM Page 41
mosaic of features of different origins and ages, to sunspot cycles, North Atlantic Oscillation
some of which may be of considerable antiquity). (NAO) and El Niño Southern Oscillation (ENSO)
Current processes may be gradually changing this events, volcanic eruptions, and stochastic variations
inherited relief, rapidly destroying it or having no (Barber et al., 1994; Roberts, 1998; Waple, 1999).
effect on it at all. The presence of relict, deep- There are also sustained periods of climate change
seated rotational landslides on scarp faces in the lasting several hundreds of years or more. Over the
Sahara (Figure 16.6, Chapter 16) highlights the last millennium the UK has experienced two
importance of looking beyond the present-day important natural perturbations on the 100-year
climate–process associations in developing a scale, the so-called ‘Little Optimum’ (or Medieval
comprehensive ground model. Warm Period) warming of c. 700–1200 AD (Lamb,
1997) and the ‘Little Ice Age’ (widest definition
1420–1850 AD with a period of amelioration from
2.6 Climate variability 1500–1550 AD). There is quite strong coincidental
evidence that these are related to the ‘Grand
Climate has not remained constant throughout the Maximum’ and the two great minima of sunspot
long history of the Earth (Figure 6.1), primarily activity, the Spörer and the Maunder Minima
due to variations in the orbit around the Sun and (Adams et al., 1999). The ‘Little Ice Age’ was asso-
solar radiation. Other possible mechanisms caus- ciated with an increase in debris flow and avalanche
ing rapid climatic changes include the cold ice- activity (Grove, 1988).
berg melting events (7000–13 000 year intervals: Shorter episodes such as droughts or
so-called ‘Heinrich’ events) and the warm cycles sequences of wet years can give rise to noticeable
(1000–3000 year intervals, 1500-year duration: changes in the rate of geomorphological pro-
so-called ‘Dansgaard-Oeschger’ cycles). How- cesses. In the Caspian Sea a post-1975 rise in sea-
ever, there has been a growing appreciation of the level of over 2 m (reversing the trend of falling
link between plate tectonics and climate changes sea-levels; Figure 2.10) that is believed to be
over tens of millions of years (Figure 2.8). This associated with the development of a more humid
arises from variations in the rate of sea-floor climate in the region has been accompanied by a
spreading and the release of CO2 that, along 40% increase in the extent of coastal erosion
with other factors, leads to adjustments in global (Bird, 2000).
temperatures.
In addition to the climatic changes that occur
over geological time (e.g. the repeated glacial and 2.7 Extreme events
interglacial episodes of the Quaternary), there are
important climatic variations that influence geo- It is common for significant geomorphological
morphological processes over shorter periods changes and natural hazards to be associated with
(Figure 2.9; Brunsden, 1996). Over a 10 000-year extreme climatic events, be they hurricanes,
period average temperatures can change by intense rainstorms or periods of heavy rainfall or
c. 5–10 °C with recognisable sequences of cool– strong winds. These climatic events can be
wet and warm–dry periods. For example, there regarded as a random part of a natural series of
have been several major fluctuations over the events of varying magnitude and frequency,
Holocene including the cooling of c. 8000–4800 BP, whose distribution can be established from a
the cool event at c. 3500 BP and the move from sequence of records and whose probability of
the warm-dry sub-Boreal to the cooler-wetter occurrence in a given period can be calculated
sub-Atlantic of c. 2500 BP. These events caused using standard statistical methods. For example,
significant changes in rates of geomorphological the likelihood of a rainfall of a particular magni-
activity. tude is generally expressed by the return period or
Over the 1000-year period, major climatic recurrence interval. Thus, the rainfall that is
trends may be due to solar activity variation related expected to be equalled or exceeded on average
GFE-02.qxd 6/24/05 2:45 PM Page 42
Continental drift
Change in Global distribution of Thermally isolated

paleolatitude Global albedo
continents poles
MORPHOLOCLIMATIC
REGIMES
Orogenesis
Change in atmospheric - oceanic Ice sheet
circulation and climatic patterns development
Full ice age climate

Continentality
Global sea level

change
Global temperature Short-term glacio-eustacy
Long-term tectono-
Volcanicity Atmospheric CO2 eustacy
Volume of global
Sea-floor spreading mid-oceanic ridge
rate system
Figure 2.8 A schematic representation of some of the ways in which tectonic processes generate morphoclimatic
zone change (from Summerfield, 1991).
every 100 years has a return period of 100 years. rainfall with return period T during a time period
This event could occur any year, but the probabil- of N years is given by the following equation:
ity of its occurrence during 100 years is much
greater than during a one-year period. P ⫽ 1 ⫺ (1 ⫺ 1/T)N
The likelihood of a particular rainfall event
(e.g. an annual rainfall total) occurring in any sin- This provides the ‘encounter probability’ which is
gle year (i.e. the return period) may be calculated useful for communicating risk as it is less likely to
from a series of events (i.e. annual rainfall totals be misunderstood than return period.
over many years) as follows:
Number or years
2.8 Climate and flooding
in the sequence +1
Return period ⫽ Climate is a major control of flooding events,
Ranking of the event along with other factors such as landslides and
in the sequence dam failures, and flood intensifying factors
(Figure 2.11). Figure 2.12 presents a global
The relationship between probability (expressed classification of flood zones. The latitude, posi-
as a percentage), return period and the length of tion and strength of the circumpolar vortex (or jet
period under consideration is shown in Table 2.2. stream) that controls depression tracks govern the
This indicates that a 1000-year event has a 9% occurrence of floods in the high mid latitudes.
chance of occurring during the 100-year lifetime Changes in atmospheric circulation can lead to
of an engineering structure. The likelihood of changes in flood frequency. More frequent floods
encountering an event equal to or exceeding a in the Upper Mississippi valley during the late
GFE-02.qxd 6/24/05 2:45 PM Page 43
Figure 2.9 Timescales and terminology of geomorphological events (redrawn from Brunsden, 1996).
Metres below Mediterranean sea level
26 SEA LEVEL CHANGES 26
27 27
28 28
29 29
1930 1935 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000
Year
Figure 2.10 Changes in the level of the Caspian Sea since 1930, as indicated by the Baku tide gage, Azerbaijan
(from Bird, 2000).
1800s and since 1950 were due to a weak westerly latter is associated with high sea-surface temper-
circulation in mid latitudes (Knox, 1984; 1988). ature anomalies that spread eastwards across the
In low latitudes floods and/or drought events are equatorial Pacific Ocean. ENSO events occur
associated with tropical storms, the monsoon and every 3–10 years and trigger extreme floods and
the El Niño Southern Oscillation (ENSO). The droughts in many regions (Figure 2.13).
GFE-02.qxd 6/24/05 2:45 PM Page 44
Table 2.2 Percentage probability of the N-year rainfall occurring in a particular period.
Number of N ⫽ average return period in years

years in
period 5 10 20 50 100 200 500 1000
1 20 10 5 2 1 0.5 0.2 0.1

5 67 41 23 10 4 2 1 0.5
10 89 65 40 18 10 5 2 1
30 99 95 79 45 26 14 6 3
60 – 98 95 70 31 26 11 6
100 – 99.9 99.4 87 65 39 18 9
300 – – – 99.8 95 78 45 26
600 – – – – 99.8 95 70 45
1000 – – – – – 99.3 87 64
Where no figure is inserted the percentage probability ⬎99.9.
In the shaded box: there is, on average, a 9% chance that a 1000-year event (annual proba-
bility of 0.001) would occur within a 100 year time period.
An important consequence of short-term cli- zones poleward by about 550 km. Precipitation has
matic variation is that the probability of flood increased by 0.5 to 1% per decade in the twenti-
events may not be constant over time. Many of the eth century over the Northern Hemisphere conti-
world’s major rivers experienced ‘humid’ and ‘dry’ nents and there has been a 2 to 4% increase in the
periods throughout the last century (Probst and frequency of heavy precipitation events.
Tardy, 1987). A degree of caution is needed, there- The impacts of climate change will vary from
fore, when using return period statistics as they region to region. The UK Climate Impacts
assume that the probability of flooding (or any other Programme (Hulme et al., 2002), for example, has
geohazard) of a given magnitude remains constant predicted that the climate changes as a result of
from year to year, over the historical record. human activity over the next century are expected
to increase risks from coastal flooding and erosion
in two ways:
2.9 Global climate change
1. by the 2050s the rise in sea level is predicted to
Global climatic changes are occurring as the result increase the frequency of extreme high water
of man-induced accumulation of so-called green- levels from once a century to, typically, once a
house gasses such as CO2 in the atmosphere. decade — this situation would be further exac-
Evidence from ice cores supplemented by direct erbated if storminess were to increase
measurements since the mid-1950s shows a steady 2. days with heavy rainfall will become typically
rise in greenhouse gas concentrations from the three or four times more common, increasing
late 1700s changing to a rapid rise post 1950. the risk of non-tidal floods in estuaries and
Atmospheric concentrations of carbon dioxide promoting accelerated cliff instability.
(CO2), the primary anthropogenic greenhouse gas,
have risen from about 270 ppm in pre-industrial
times to over 360 ppm. 2.10 The hydrological cycle
It is expected that the mean temperature will
rise by 1–3.5 °C by 2100, accompanied by a sea- The hydrological cycle (Figure 2.2) provides an
level rise of 15–95 cm (IPCC, 2001). Warming at important link between climate, geomorphologi-
the higher end of this range would shift climatic cal processes and weathering, particularly the
GFE-02.qxd 6/24/05 2:45 PM Page 45
Figure 2.11 Causes of floods and flood intensifying factors (after Ward, 1978).
GFE-02.qxd 6/24/05 2:45 PM Page 46
Figure 2.12 Flood climate regions of the world (from Hayden, 1988; Macklin and Lewin, 1997).
Figure 2.13 Regions affected month by month by the El Niño Southern Oscillation (from Folland, et al., 1990;
Macklin and Lewin, 1997).
balance between runoff, infiltration into the soil flooding (e.g. the speed of transmission of rain-
and subsequent evaporation, through-flow or per- fall into the drainage channels), gully forma-
colation towards the groundwater table (Figure tion (e.g. the concentration of runoff), soil piping
2.14). The balance between these processes can (e.g. concentration of through-flow), landslides
control the occurrence of geohazards such as (e.g. the rapid build up of pore water pressures in
GFE-02.qxd 6/24/05 2:45 PM Page 47
Figure 2.14 Idealised characteristics of near-surface hydrological environments (from Fookes, 1997).
soils prone to shallow slides or the slow rise of the expressed the global variation in weathering in
groundwater table in areas prone to deep-seated terms of a series of weathering zones related to
slides), high groundwater tables, the height of cap- temperature and precipitation (Figures 2.16 and
illary rise and associated aggressive soil problems. 2.17). Peltier (1950) used mean annual tempera-
Table 2.3 highlights some engineering-related ture and mean annual precipitation to define
issues associated with climate and groundwater. zones characterised by different intensity of
chemical and frost weathering (Figure 2.18).
While both approaches allow a broad indication
2.11 Weathering of the anticipated weathering regime in a region,
at a detailed level weathering will reflect site
Weathering involves the gradual breakdown (i.e. conditions, notably lithology and structure of the
disintegration) and alteration (i.e. decomposition) parent material (including the features inherited
of materials through a combination of chemical from past climates), availability of water (e.g.
and physical (mechanical) processes (Table 2.4; groundwater or through-flow), vegetation and
Figure 2.15), leading to the development of human activity.
weathering profiles and residual soils. The rela-
tive significance of any weathering process
depends on the weathering environment (e.g. cli- 2.12 Weathering and landscape
mate), the nature of the geological materials and
biological conditions. Weathering plays an important role in land-
Climate has an important control on both the scape development, notably etchplain formation,
nature and rate of weathering. Strakhov (1967) duricrusts and karst landforms.
GFE-02.qxd 6/24/05 2:45 PM Page 48
Table 2.3 Some engineering issues associated with climate and groundwater (modified from Fookes, 1997).
Condition Comments
Annual evapotranspiration Low water tables controlled by stream levels and presence of more permeable
exceeds infiltration strata, upward flow due to evaporation. Partly saturated soils, sometimes
to great depth.
Long-term desiccation. Possibility of collapse on wetting (or when evaporation is
prevented by sealing the ground surface) in porous residual soils, weathered
sands, loess, and heave on wetting of plastic clay.
Soil suctions facilitate construction (e.g. steep temporary slopes, shafts)
Erosion protection of slopes by vegetation often impractical.
Chemically active zone at ground level due to evaporation of ground water,
recementation and cap-rock.
Tendency for flashy sheet floods and ephemeral channel flows in
response to cloudbursts.
Tendency for shallow landslides (e.g. rock and debris slides) and debris flows in
response to cloudbursts.
Annual infiltration exceeds Deep percolation dominates. High groundwater tables usual in fine-grained soils,
evapotranspiration even in slopes. Saturated soils.
Flow into excavations and underground works.
Ground movements due to dewatering and internal erosion.
Erosion protection of slopes by vegetation usually practical.
Flood character varies with catchment type and antecedent conditions, but can
range from flash floods in steep catchments to prolonged lowland floods.
Landslide character varies with slope conditions (e.g. geology, slope angle,
antecedent conditions), but includes deep-seated failures on clay slopes or
slopes with mudrock beds.
Seasonal shorter Evapotranspiration and infiltration rates often vary seasonally (and from year
term variations to year) and also over shorter timescales. Surface conditions and water contents
(and related engineering properties of superficial fine-grained soils)
vary accordingly.
Seasonal heave and settlement of superficial clays and foundations. Variation of
superficial undrained strength of clays and suitability as embankment
foundations. Heave and settlement of superficial structural foundations depend
on conditions prior to construction. Undrained strength and suitability of
superficial fine-grained soils as fill also varies.
Conditions during site investigation may vary from those during construction.
Earthmoving of fine-grained soils very difficult when rainfall exceeds
evapotranspiration. Seasonal variations control earth moving season and
short-term variation controls working time. These variations depend on
temperature, wind and rainfall duration, rather than on rainfall amount,
problems usually worst in temperate climates and are often strongly
influenced by site elevation.
Supplies of water for conditioning fills may vary seasonally.
Seasonal variation in probability of flood, erosion and landslide events. Risk
greatest during intense cyclonic storms and periods of snowmelt or prolonged
rainfall, although antecedent conditions will determine whether events
actually occur.
GFE-02.qxd 6/24/05 2:45 PM Page 49
Table 2.4 Classification of weathering processes (from Cooke and Doornkamp, 1990).
Processes of disintegration Processes of decomposition
Crystallisation processes: Hydration and hydrolysis

Salt weathering (crystal growth,
hydration, thermal expansion)
Frost weathering
Temperature/pressure change processes: Oxidation and reduction
Insolation weathering
Sheeting, unloading
Weathering by wetting and drying: Solution, carbonation, sulphation
Moisture swelling
Alternate wetting and drying
Water-layer weathering
Organic processes: Chelation
Root wedging
Colloidal plucking
Lichen activity
Bio-chemical changes
Micro-organism decay
Bacteria
Lichens
Figure 2.15 Factors and processes important in weathering (adapted from Brunsden, 1979).
GFE-02.qxd 6/24/05 2:45 PM Page 50
Figure 2.16 Simple cross-section from pole to equator showing climate and weathering characteristics (adapted
from Strakhov, 1967).
Figure 2.17 World weathering zones (adapted from Strakhov, 1967).

GFE-02.qxd 6/24/05 2:45 PM Page 51
Figure 2.18 Characteristic climate and weathering patterns, A–C redrawn from Peltier (1950); D redrawn from
Fookes (1980).
Etchplain development Duricrusts

In many areas, especially the humid tropics, land- Duricrusts are indurated horizons at or near the
scape denudation involves a combination of ground surface, generally formed through the
ongoing chemical decomposition at the weather- accumulation of iron or aluminium (ferricrete or
ing front under the influence of groundwater alcrete), silica (silcrete), calcium carbonate (cal-
through-flow, and the removal of weathered mate- crete), magnesium carbonate (dolocrete) or gyp-
rial from the ground surface by sheet-wash or sum (gypcrete). Ferricretes and alcretes form by
wind erosion (Figure 2.19). Etch forms can cut the relative accumulation of iron and aluminium
across lithologies and can result in deep and oxides in the soil, as more mobile compounds are
intensely weathered residuals soils (Twidale, leached out of the weathering profile. For this rea-
1990). Acting over millions of years, etching can son, they tend to be associated with high rainfall
result in remarkably flat plains and plateau sur- climates (Chapter 18). The other duricrusts form
faces, such as those developed in limestones in through an absolute accumulation in the weather-
the Sahara and in granites and gneiss in Western ing profile. Accumulation may occur as a result of
Australia and Namibia. Climate change or tec- capillary rise (per ascensum model), downward
tonic uplift can stimulate incision through parts of percolation (per descensum model) or through-
the etchplain surface, resulting in a pattern of flow of solute-rich groundwater. These duricrusts
soils that appear to be unrelated to the current cli- are frequently associated with desert environ-
mate or landscape. ments (Chapter 16). Duricrusts are generally
GFE-02.qxd 6/24/05 2:45 PM Page 52
stronger than the underlying or surrounding mate-

rials and, hence, tend to armour the landscape
by forming a hard cap to flat-topped hills
and plateaux surfaces. The presence of these
indurated horizons may have significant implica-
tions for excavation operations and sources of
construction aggregates.
Karst landforms
Karst landforms are distinctive landscapes that
have developed where chemical weathering of
soluble rocks such as limestones, gypsum and
halite has created features such as thin soil covers,
ground depressions and underground cave sys-
tems. These landscapes present a unique range
of engineering problems that are described in
Chapter 24.
2.13 Weathering products:

engineering soils
In situ weathering produces a weathering profile

comprising various grades of weathered rock and
residual soil (Appendix Tables A3.3a, A3.3b and
A3.3c and Chapter 7 Table 7.4). These changes
are accompanied by significant changes in mate-
rial properties, notably a decrease in strength and
changes in permeability, porosity and water con-
tent. The nature of the weathering profiles tends
to vary with rock type and climate, as will be dis-
cussed in later Chapters. The thickness of the
weathering profile reflects the relative balance
Figure 2.19 The development of different types of etch-
between the rate of bedrock weathering and the
plains and etchsurfaces. The diagrams do not necessarily
represent an evolutionary sequence as repeated episodes rate of removal by soil erosion or landslides.
of accelerated erosion may only succeed in partially Depths of weathering may exceed 100 m, espe-
removing the weathering mantle. The types of etchplains cially in humid tropical environments.
and etchsurfaces illustrated are: (A) lateritized etchplains The type of clay mineral generated through
comprising a surface of low local relief underlain by a
weathering depends, in part, on the climate. For
thick weathering mantle, including indurated lateritic
horizons (ferricretes), which has been subject to only lim- example, the relative abundance of kaolinite,
ited stream incision; (B) dissected etchplains in which gibbsite and iron and aluminium oxides is related
accelerated stream downcutting promoted by climatic to the mean annual precipitation (Figure 2.20),
change or uplift leads to the development of well-defined the nature of the original parent material
valleys, fringed in places by duricrust breakaways, and
and the timescale over which the material has
the very localized exposure of bedrock and the formation
of tors; (C) partially stripped etchplains characterized by experienced weathering activity. In addition,
widespread stream dissection and the extensive stripping the intensity of soil leaching (a function of
of the weathering mantle (after Thomas, 1994). temperature, precipitation and soil drainage)
GFE-02.qxd 6/24/05 2:45 PM Page 53
Figure 2.20 Variations in the clay mineral and residual hydroxide composition of soils in relation to mean annual
precipitation: (A) surface soil samples on feldspar and quartz rich igneous rocks; (B) olivine, amphibole and pyroxene-
rich igneous rocks; (C) soils developed under an alternating wet and dry climate and (D) soils developed in a
continuously humid climate (from Summerfield, 1991).
by water through-flow is an important factor minerals. Calcium carbonate often accumulates

(Table 2.5). with soils in arid environments, with more soluble
Intense leaching environments are associated salts such as gypsum or halite able to build up in
with the removal of metal cations, iron and alu- areas of extreme aridity.
minium oxides and hydroxides from the soil in
solution. Kaolinites tend to be the dominant clay
mineral. 2.14 Weathering and geohazards
Moderate leaching environments are associ-
ated with the build up of cations released by Weathering and the resultant changes to material
weathering, leading to the formation of cation- properties often act as a preparatory factor in the
bearing clays such as illite and smectite. generation of geohazards, especially debris flows
Low intensity leaching environments are asso- and landslides. For example, in many mountain-
ciated with very high concentrations of dissolved ous areas, landslide activity follows a progressive
GFE-02.qxd 6/24/05 2:45 PM Page 54
Table 2.5 Weathering products and environmental controls (from Loughnan, 1969; Summerfield, 1991).
Environment pH Redox potential Behaviour of Mineralogy of

(Eh) major elements weathering products
Non-leaching. Alkaline Oxidising Some loss of Partly decomposed parent minerals.

Mean annual Na⫹ and K⫹. Iron Illite, chlorite, smectite and
precipitation present in ferric mixed layered clay minerals.
⬍ 300 m. Hot state. Hematite, carbonates, secondary
silica and salts. Organic matter
absent or sparse.
Non-leaching Alkaline to Reducing Some loss of Partly decomposed parent minerals.
below water neutral Na⫹ and K⫹. Iron Illite, chlorite, smectite and mixed
table present in ferrous layered clay minerals. Siderite
state. (iron carbonate) and pyrite (iron
sulphide). Organic matter present.
Moderate leaching. Acid Oxidising to Loss of Na⫹, K⫹, Kaolinite with or without degraded
Mean annual reducing Ca2⫹, Mg2⫹ and (K-deficient) illite. Some hematite
precipitation some loss of SiO2. present. Organic matter generally
600–1300 mm. Concentration of present.
Temperate Al2O3, Fe2O3, TiO2.
Intense leaching. Acid Oxidising Loss of Na⫹, K⫹, Hematite, goethite, gibbsite and
Mean annual Ca2⫹, Mg2⫹ and SiO2. boehmite with some kaolinite.
precipitation Concentration of Organic matter absent or sparse.
⬎ 1300 mm. Al2O3, Fe2O3, TiO2.
Hot
Intense leaching. Very acid Reducing Loss of Na⫹, K⫹, Ca2⫹, Kaolinite, possibly with some
Mean annual Mg2⫹ and some iron gibbsite or degraded Illite.
precipitation and Al2O3. SiO2. and Organic matter abundant.
⬎ 1300 mm. TiO2 retained.
Cold
soil-stripping model (Crozier, 1986), with the pat- other circumstances, frost weathering may act as
tern and frequency of failures controlled by the a trigger for rockfalls.
availability of unstable material. Soil development
proceeds until a critical depth and a landslide-
triggering event (e.g. intense rainfall) occur in 2.15 Summary
combination, generating shallow hillside failures.
After the initial failure, the rate of weathering and Climate — both current and past — has a strong
soil formation will be the limiting factor on the influence on engineering through the operation of
frequency of failures at a particular site. geomorphological processes and geohazards that
In cuttings in stiff clay, the combination of the might impact on the site, and through the produc-
slow recovery in pore water pressures, progres- tion of a weathering profile and residual soil. All
sive failure and strain-softening will result in a of these have a significant influence on both engi-
gradual decline in effective shear strength and neering design and construction. Climate should
may result in the eventual failure of the slope. On not be regarded as constant over ‘engineering
London Clay railway cuttings in the UK, delayed time’, with obvious implications for assessments
failures have been reported over 100 years after of the probability of climatically triggered geo-
the initial excavation (e.g. Skempton, 1964). In hazards. Global climate change and sea-level rise
GFE-02.qxd 6/24/05 2:45 PM Page 55
may have important implications for geohazards Fookes, P. G., Baynes, F. J. and Hutchinson, J. N. (2000)
and associated risks. The various climatic, morpho- Keynote Lecture: Total geological history: a model
approach to the anticipation, observation and
climatic, weathering zone and soils maps pre-
understanding of site conditions. GeoEng 2000,
sented in the Chapter should provide a starting International Conference on Geotechnical and
point for developing a ground model for the area Geological Engineering, Melbourne 1, 370–460.
of interest. In the following Chapters more spe- Grove, J. M. (1988) The Little Ice Age. Methuen,
cific accounts are provided of the particular con- London.
Hayden, B. P. (1988) Flood climates. In Baker, V. R.,
ditions to be found, and engineering problems that
Kochel, R. C. and Patton, P. C. (eds) Flood
can occur, in selected environments. Geomorphology. Wiley, New York, 13–26.
Hulme, M., Jenkins, G. J., Lu, X., Turnpenny, J. R.,
Mitchell, T. D., Jones, R. G., Lowe, J., Murphy, J. M.,
References Hassell, D., Boorman, P., McDonald, R. and Hill, S.
(2002) Climate Change Scenarios for the United
Adams, J., Maslin, M. and Thomas, E. (1999) Sudden Kingdom: The UKCIPO2 Scientific Report. Tyndall
climatic transitions during the Quaternary. Progress Centre for Climate Change Research, University of
in Physical Geography 23, 1–36. East Anglia, UK.
British Standards Institution (1999) BS5930-1999 IPCC (Intergovernmental Panel on Climate Change)
Code of Practice for Site Investigations. (2001) The IPCC third assessment report – sum-
Bird, E. (2000) Coastal Geomorphology: An mary for policy markers. http://www/ipcc.ch/
Introduction. Wiley, London. index.html.
Barber, K. E., Chambers, F. M., Maddy, D., Stoneman, Kiewiet de Jonge, C. J. (1984) Budel’s geomorphology II.
R. and Brew, J. S. (1994) A sensitive high resolution Progress in Physical Geography 8, 365–397.
record of late Holocene climatic change from a Knox, J. C. (1984) Flurial responses to small scale cli-
raised bog in northern England. The Holocene 4, matic changes. In Costa, J. E. and Fliesher, P. J. (eds)
198–205. Developments and Application of Geomorphology.
Briggs, D. and Smithson, P. (1992) Fundamentals of Springer-Verlag, New York, 318–342.
Physical Geography. Routledge, London. Knox, J. C. (1988) Climatic influence on Upper
Brunsden, D. (1979) Weathering. In Embleton, C. and Mississippi Valley. Annals of the Association of
Thornes, J. (eds), Process in Geomorphology. American Geographers 77, 224–244.
Arnold, London, 73–129. Köppen, W. (1931) Grundriss der Klimakunde. De
Brunsden, D. (1996) Geomorphological events and Gruyter, Berlin.
landform change. Zeitschrift für Geomorphologie Köppen, W. (1936) Das Geographischem System der
40, 273–288. Klimate. Gebr. Bortraeger, Berlin.
Cooke, R. U. and Doornkamp, J. C. (1990) Geomor- Lamb, H. H. (1997) Climate, History and the Modern
phology in Environmental Management. Oxford World. Routledge, London.
University Press. Loughnan, F. C. (1969) Chemical Weathering of the
Crozier, M. J. (1986) Landslides: Causes, Consequences Silicate Minerals. Elsevier, Amsterdam.
and Environment. Routledge, London. Macklin, M. G. and Lewin, J. (1997) Channel, flood-
Doornkamp, J. C. (1986) Climate and weathering. In plain and drainage basin response to environmental
Fookes, P. G. and Vaughan, P. R. (eds) A Handbook change. In Thorne, C. R., Hey, R. D. and Newson,
of Engineering Geomorphology. Surrey University M. D. (eds) Applied Fluvial Geomorphology for
Press, Blackie and Son, London, 10–24. River Engineering and Management. Wiley,
Folland, C. K., Karl, T. R. and Vinnikov, V. (1990) Chichester, 15–45.
Observed climate variations and change. In Peltier, L. C. (1950) The geographic cycle in periglacial
Houghton, J. T., Jenkins, G. J. and Ephraums, J. (eds) regions as related to climatic geomorphology.
Climate Change: the IPCC Scientific Assessment. Annals of the Association of American Geographers
Cambridge University Press, 200–228. 40, 214–236.
Fookes, P. G. (1980) An introduction to the influence of Probst, J. L. and Tardy, Y. (1987) Long range stream-
natural aggregates on the performance and durabil- flow and world continental runoff fluctuations since
ity of concrete. Quarterly Journal of Engineering the beginning of this century. Journal of Hydrology
Geology 13, 207–229. 94, 289–311.
Fookes, P. G. (1997) First Glossop Lecture: Geology for Roberts, N. (1998) The Holocene: an Environmental
engineers: the geological model, prediction and per- History. Blackwell, Oxford.
formance. Quarterly Journal of Engineering Skempton, A. W. (1964) Long term stability of clay
Geology 30, 293–424. slopes. Geotechnique 14, 77–101.
GFE-02.qxd 6/24/05 2:45 PM Page 56
Strakhov, N. M. (1967) Principles of Lithogenesis, Vol. Twidale, C. R. (1990) The origin and implications of
1. Oliver and Boyd, Edinburgh. some erosional landforms. Journal of Geology 98,
Summerfield, M. A. (1991) Global Geomorphology. 343–364.
Longman, Harlow. Waple, A. M. (1999) The sun–climate relationship in
Thomas, M. F. (1974) Tropical Geomorphology. recent centuries: a review. Progress in Physical
Macmillan, London and Halstead Press, New York. Geography 51, 309–328.
Thornthwaite, C. W. (1933) The climates of the earth. Ward, R. (1978) Floods: A Geographical Perspective.
Geographical Review 23, 433–440. Macmillan, London.
GFE-03.qxd 6/24/05 2:48 PM Page 57
3. Sedimentology
Colin J. R. Braithwaite
3.1 Introduction understand how surface environments have come

about in the last few hundred thousand years, and
Sedimentology first grew as a distinct branch of to appreciate that they are dynamic systems that
Earth Science in response to the demands of the experience changes observable within our lifetimes
petroleum exploration industry for more detailed and affecting all human constructions. This per-
information on the nature and origin of sedimen- spective may be gained through Part II of this book.
tary rocks involved in the formation and accumu- However, there are two important caveats that must
lation of hydrocarbons. It is concerned with the be applied. Present environmental conditions do
ways in which sedimentary grains may be gener- not necessarily represent all of those that may be
ated, transported, deposited and buried to become possible and that may have been present in the past.
rocks. Observations may be at scales varying There is good evidence, for example, that climate
from microns (using the scanning electron micro- and atmospheric and ocean compositions have
scope) to global. The geometry of sedimentary been different in the past. In addition, there are
rock sequences can be used to define environ- now clear indications of major events, shaping
mental systems, their assembly in basins and their Earth environments that have never been directly
global distribution. Almost all of these data are observed. The evidence of major meteorite impacts
also, to some degree, of concern to the engineer falls into this category. This chapter reviews con-
involved in site investigation and in the evaluation cepts surrounding the processes of weathering,
of sediments or sedimentary rocks as construction transport and deposition that form sediments,
materials. together with the processes and products of their
A principal goal of the sedimentologist is to conversion to rocks.
interpret sedimentary sequences in order to recon- From an engineering perspective, the environ-
struct the environment in which they formed. This ments described are those in which structures will
might seem an academic exercise but it provides a be built, and to a large extent each develops a
framework for predicting the specific characteris- characteristic architecture. These models provide
tics and distribution of the sediments or rocks a basis on which the distributions of sedimentary
involved. These are important issues in site investi- materials can be predicted. The petrography of
gation, mineral resources and hydrology. The sedimentary rocks addresses the major influences
approach is based on the concept of uniformitari- on their strength and behavioural characteristics.
anism, which assumes that the rules governing the
physics and chemistry of environmental systems
have remained the same throughout geological 3.2 Weathering
time. From this premise, study of the processes and
products of present-day environments allows sedi- The group of processes referred to as ‘weathering’
mentologists to interpret sequences that formed is discussed in relation to climate in Chapter 2.
over thousands of millions of years. The preserved Igneous and metamorphic rocks, and by inference
record of these ancient environments (Chapter 5) a proportion of sediments, consist of mineral
may be used to reconstruct the geological history assemblages that formed at elevated temperatures
of the Earth. For this reason it is important to and pressures, for the most part deep within the
57
GFE-03.qxd 6/24/05 2:48 PM Page 58
Earth’s crust. They are commonly unstable when may dramatically reduce strength without any
exposed to surface environments and as a conse- change being visible to the naked eye. In areas
quence slowly disintegrate. Their breakdown may such as Brazil the early effects of weathering can
be by physical processes that include the unloading be detected at depths of well over 100 metres.
resulting from their exposure, or result from chem-
ical attack by percolating surface waters, decom-
posing and changing their mineralogical and 3.3 Sediment transport and
chemical compositions. Although weathering is deposition
normally described as ‘physical’ or ‘chemical’,
neither operates in isolation. One or the other Introduction
may dominate at different times and places, but Erosion is the progressive removal from the land
more often they aid each other and are commonly surface of material that is typically weathered.
augmented by biological processes. Rock surfaces If erosion did not take place, a thickening layer of
freshly exposed by physical weathering are typi- weathered debris would accumulate over rock
cally attacked by chemical processes. It can be surfaces. The relative protection afforded by this
argued that the main function of physical weather- cover would progressively slow the weathering
ing is to increase the surface area available for process until eventually little further change
chemical attack. Dividing a cube into eight smaller would take place. The rate at which weathering
cubes by only three cuts at right angles doubles the can occur therefore depends in part on the rate at
surface area. Generally, chemical weathering is which weathering products are removed by ero-
most effective in temperate and tropical regions sion. Conversely, however, the rate of erosion is
where temperatures are moderate or high and itself influenced by the rate at which particles are
water is plentiful. In such environments, a thick released from the rock surface by weathering. The
cover of weathered material may mantle land sur- same factors, principally climate and relief, influ-
faces. Table 3.1 summarises the weathering prod- ence the rates of both weathering and erosion. It is
ucts that are produced from various rock-forming notoriously difficult to measure rates of erosion,
minerals. It is important to realise that in all rocks, and regional patterns of denudation are com-
weathering may begin along grain boundaries and monly derived from estimates of sediment yield
Table 3.1 Summary of weathering processes and products.
Process Materials attacked Soluble components Insoluble residue
Physical or mechanical Outcrops of all Successively smaller

disintegration rock types rock fragments
Unaltered silicate grains
Chemical reactions, Olivine Mg2, SiO2 Fe3 oxides and hydroxides
hydrolysis and dissolution Pyroxene Mg2, SiO2 Fe3 oxides and hydroxides,
Amphiboles Mg2, Ca2, Na, SiO2 plus hydrous aluminium
silicates (clay minerals)
Biotite mica Mg2, SiO2 Fe3 oxides and hydroxides,
plus hydrous aluminium
silicates (clay minerals)
Muscovite mica K, SiO2 No change
Feldspars Ca2, Na, K, SiO2 Hydrous aluminium
silicates (clays)
Quartz Quartz (no change)
Calcite and Ca2, Mg2, HCO3 None
dolomite
GFE-03.qxd 6/24/05 2:48 PM Page 59
SEDIMENTOLOGY 59
Table 3.2 Comparison of rates of denudation for high-altitude (mountainous) and lowland regions under
contrasting climatic regimes (data from Kukal, 1971, Blatt et al., 1980 and others).
Climate regime High altitude Denudation rate Lowland Denudation rate

(mm/1000 years) (mm/1000 years)
Arid to semi-arid. Low latitude desert 100–200 Low latitude 1

Very low precipitation, mountains Sahara Desert
Sporadic floods, California, Fort 50 Eyre Peninsula, 0.5–1
no seasonal pattern Sage Mountains South Australia
Warm temperate: Mid latitude, 3000–8000 Mid-latitude 30
moderate precipitation Coastal Range Rhine Basin
throughout the year of the USA
Warm temperate, Tropical to low 720 Tropical to low 20
moderate to high latitudes latitudes,
precipitation throughout e.g. Himalayas Northern India
the year
Cool Temperate, moderate New Zealand Alps 50–130
to heavy precipitation European Alps 50–910
throughout the year
(Table 3.2). However, weathering is greatest obvious that other factors act to reduce friction.
where rainfall is moderately high all year, and Among the most important of these are porewater
lowest in arid to semi-arid regions, regardless of pressures and the expansion and contraction of
altitude, whereas rates of denudation are greater the soil due to temperature changes, including
in high altitude mountainous regions, regardless freeze and thaw. Triggering mechanisms that
of climate. Recent years have seen the develop- overcome friction most commonly relate to varia-
ment of direct methods of dating exposed surfaces tions in these factors but include events such as
by the use of apatite fission track measurements exceptional rainfall and earthquakes. To empha-
and cosmogenic nucleides, and these promise sise the importance of these events, between 1950
more detailed and more accurate measures in the and 1993 a succession of wet years in southern
future. Rates of both weathering and erosion are California resulted in damage from mass move-
modified by vegetation, the distribution of which ments that on average totalled $500M.
generally parallels that of rainfall. The organic Some of the varied forms of mass movement
acids generated by plant decay enhance weather- are illustrated in Figure 3.1 but see also Chapter 8.
ing, while the formation of a canopy of shrubs or The principal features that differentiate these
trees may stabilise slopes and slow erosion. In from one another are the volume of water
some areas there have been dramatic (and delete- included with the rock debris, the speed at which
rious) changes in rates of erosion in response to they move, and the degree of internal deformation
deforestation and changes in farming practice. within the moving mass. In one respect, mass
movement can be regarded as gradational between
river transport, where the movement of debris is
Mass movement the direct result of the movement of water, and the
In the absence of vegetation, weathered debris on movement of rock debris by gravity alone.
a slope is usually stable until the friction between Creep is a slow and almost continuous downs-
the debris and the slope is overcome. In theory, lope movement, typically occurring in a moderately
friction or pore suction can hold weathered debris thick soil layer. The displacement that occurs repre-
on slopes of up to about 45, but in practice move- sents the cumulative effect of individual small
ment may occur on slopes as low as 1 and it is movements resulting from wetting and drying or
GFE-03.qxd 6/24/05 2:48 PM Page 60
a)
b)
Block Glide Block Topple

d)
c)
Slump Debris Flow
Figure 3.1 Summary of the varied mechanisms of slope failure. In block glide (A) movement takes place along a
bedding plane, cleavage or foliation surface inclined at a lower angle than the slope. Elevated pore pressures on the
plane may aid movement. Block topple (B) occurs where steeply inclined rectilinear jointing frees blocks that top-
ple from the exposed face. In slumping (C), failure occurs along arcuate surfaces that daylight on the exposed face.
The detached material rotates to push forwards a raised toe of debris. In debris flows (D) high pore fluid pressures
mean that the material of the slope loses cohesion and moves as a semi-fluid mass of variable viscosity.
freezing and thawing. The mechanism is illustrated masses of rock on relatively planar surfaces such
in Figure 3.2. Expansion causes displacements per- as bedding or foliation planes. Mapping the
pendicular to the ground surface while subsequent distribution and orientation of such planes is
shrinkage, aided by gravity, is vertical, so that there important in assessing slope stability. Typically
is a net movement downslope. Although such there is little or no internal deformation within the
movements are slow, their presence may be moving mass. Slumps are more common in less
revealed by bulging soil surfaces, by bent trees, cohesive materials such as soils or clay-rich rocks,
where growth is compensating for realignment, or and involve displacements on curved, concave-up
by disturbed fencing. Any of these indicators shear surfaces (Figure 3.1c). These create ten-
should be a cause for caution. sional stress in the material at the upper end of the
The movement of landslides is generally much shear plane, and compression at the foot. The
more rapid and is of two kinds. Block glides are moving mass rotates, and the toe of the slide
the displacement of coherent and competent moves upwards until its mass counterbalances the
GFE-03.qxd 6/24/05 2:48 PM Page 61
SEDIMENTOLOGY 61
debris moves downslope buoyed up only by the

interaction of grain impacts. Such material has
Level of frozen ground surfa
ce little internal friction and emerging from a steep
slope may ‘run out’ for some considerable distance.
More commonly a debris flow is the movement of
90º
a viscous mass of water-saturated debris in which
pebbles or boulders are suspended in a slurry con-
Level of unfrozen ground sisting of sand, silt or clay. Rates of movement
surface
range from barely perceptible in relatively dry sys-
tems to almost instantaneous where large volumes
of water are incorporated and the mass becomes
Figure 3.2 The process of soil creep under the influ- fluid, travelling downslope at speeds in excess of
ence of freeze-and-thaw. Note the expansion of the 4 m/s. Debris avalanches at Mount Huscarán in
active layer at 90 to the slope surface as the contained Peru in 1962 are thought to have travelled at speeds
water freezes and the vertical return under gravity as
the ice melts. in excess of 320 kph, forming a wave 80 m high.
As densities may reach 1.5–2.0 they are capable
of moving blocks weighing many tonnes and of
downward force exerted by the body of the slide destroying large structures.
and movement ceases. Arcuate fractures and a Mass movements may result in significant
concave scar are left on the hillside where material hazards, either within the area of failure or as a
has been removed, and a convex hump forms result of subsequent deposition. The distinctive
where it comes to rest. There may be little internal characteristics of the deposits formed allow
deformation within the rotating mass but slumps the contrasting mechanisms of transport to be
are also transitional to other movements. identified. A mass of rock moved by a block glide
Both block glides and slumps move rapidly, and (Figure 3.1a) usually retains its internal structure
their ‘instantaneous’ failure, the reason why they and any original stratification. By contrast, strata
are considered hazards, is commonly due to rising contained within a slump are typically tilted by
water seepage pressures. Although small amounts the rotational movement (Figure 3.1c) and, where
of water in dry materials may actually increase strength begins to be lost, may become folded.
friction, pore pressures that approach overburden As the mass continues to move downslope inter-
pressure cause a dramatic reduction. Pore pres- nal structures may be completely destroyed.
sures may be focused in joints or bedding planes, The deposits of debris flows usually contain a
or along the junction between two rock or sedi- wide range of particle sizes, from clay-sized par-
ment (soil) types of varying permeability, but in ticles through to pebbles or boulders, and are
poorly consolidated materials are commonly more described as ‘poorly sorted’ (see ‘siliciclastic sed-
dispersed. Once initiated, slides may occur inter- iments and sedimentary rocks’). Those from rock
mittently along the same shear plane over many falls (Figure 3.1b) or avalanches are likely to con-
years, whenever pore pressures rise sufficiently. sist of angular fragments of varying sizes piled
As more and more water is added to the system, into a heap and lacking any form of layering.
by direct rainfall, by groundwater seepage or from However, although debris-flow deposits are typi-
leaking pipelines, pore pressures rise and the mate- cally poorly sorted, they may sometimes develop
rial loses strength, either by the dilation of coarser a crude stratification as a result of shear generated
materials or the loss of suction in clays. This may by their movement. The more water a flow con-
mean that the slump deforms internally, sometimes tains the more likely it is to develop stratification
generating complex folds of existing sedimentary as the conditions approach those of fluid flow.
layers, but with increasing volumes of water the In areas where environmental factors are
moving mass becomes a debris flow. Examples are closely monitored some element of prediction
known of rock flows in which unconsolidated may be possible. In some areas in California, for
GFE-03.qxd 6/24/05 2:48 PM Page 62
example, it is known that where rainfall exceeds Regime’ conditions, required for two discrete sets
about 15 mm/hour the threshold time for the onset of bedforms. Like the Reynolds number this is a
of debris flows varies from 8–14 hours depending dimensionless coefficient that allows the compari-
on local slopes, materials and runoff. son of flow conditions in channels, pipes and
Although they are commonly a significant around objects at different scales. William Froude
natural hazard, the environmental importance of (1810–1878) was a naval engineer and used the
mass movements lies in the fact that they are the coefficient to scale the behaviour of model ships.
means by which weathered debris can be moved The Froude number (F) is defined as:
from hillsides and cliffs into a river, on to a gla-
cier or into the sea to begin a more substantial U
F=
phase of transport. gD
Transport and deposition by fluids where U is the average velocity of the current; D is
The principle differences in the behaviour of the depth of the channel and g is the acceleration
fluids and gasses lie in their relative viscosities, due to gravity.
essentially how ‘stiff’ they are. Two general modes However, even in laminar flow a velocity gra-
of fluid flow are recognised, laminar flow and tur- dient is set up normal to the direction of flow.
bulent flow. This distinction was first recognised Fluid in contact with the bed surface is brought to
and demonstrated by Osborne Reynolds (1842– rest by friction and successive ‘layers’ above this
1912), a British physicist who released dye streams can be thought of as moving at increasing rates.
into long straight tubes full of flowing water. With a cohesionless granular bed, flow effectively
Reynolds described flow by reference to a dimen- extends into the bed surface. As the velocity of
sionless coefficient that has since become known the fluid over the bed increases, a point is reached
as the ‘Reynolds number’ (Re). when the stress applied to the surface is large
For sedimentary particles: enough to tear particles away. This threshold of
movement is the critical erosion velocity.
Ud
Re =
The physics of sediment movement
During fluid transport, the particles of sediment
where U is the velocity of the particle or flow; is (grains) are buoyed up by turbulence in the fluid
the density of the particle; is the viscosity of the that moves in response to gravity. The range of
fluid and d is the diameter of the pipe or grain. size of particles that can be carried, and the char-
At low Reynolds numbers, the threads of dye acteristics of movement of the particles, depend on
remain straight and the flow is therefore laminar. the relationship between the size and density of
As velocity increases the dye streams become the particles and the density and viscosity of the
highly disturbed, indicating that the flow is turbu- fluid, together with the rate of flow. In natural sys-
lent. For flow in tubes this transition occurs when tems the fluid may be either water or air. The fol-
the value of Re is about 2000. The important fea- lowing discussion relates largely to water, but the
ture here is that although the net movement of the principles of transport by air are essentially the
fluid is large and is still downstream, there are same (see the discussion of the Reynolds number).
now random secondary motions that at a fixed Particles placed in water will settle to the floor
point vary erratically in both direction and veloc- because of the force of gravity acting on them.
ity. These may provide lift for grains in erosion There are, however, two forces that oppose this
but are primarily responsible for the suspension of process. First there is frictional drag between the
smaller grains within the fluid. surface of the particles and the water molecules.
In open channels a second important coeffi- This depends largely on particle shape. Second,
cient, the ‘Froude number’ differentiates the flow there is the natural buoyancy of the particles,
conditions, referred to as ‘Lower’ and ‘Upper Flow expressed as the difference in density between the
GFE-03.qxd 6/24/05 2:48 PM Page 63
SEDIMENTOLOGY 63
particles and the water (or other fluid). The veloc- skeletons may have very elaborate shapes but
ity at which particles may be expected to settle also a relatively low density, reflecting the
through a fluid takes both of these forces into presence of internal pores rather than crys-
account and can be expressed by the ‘Stokes’ law talline calcite or aragonite. Finally, the fall
equation’. velocities of single grains are somewhat faster
than the hindered settling of large numbers of
similar grains in a sediment mixture because
Stokes’ law and sediment transport
they avoid grain-to-grain collisions.
Stokes related the erosion and transport of sedi-
2. Equation (1) assumes that the fluid is station-
ment particles to their size and density relative to
ary, whereas natural waters are rarely com-
the velocity and viscosity of the fluid around
pletely still.
them, as described in the equation:
2r 2 g Stokes’ Law can only be applied when the molec-

v (1) ular flow of the water around the particles is
9
smooth, or laminar (Figure 3.3a). In practice, at
where v is the settling velocity (m/s); g, the accel- velocities of only a few centimetres per second,
eration due to gravity; r, the radius of the particle water flow becomes turbulent and chaotic, and
(m); , the difference in density between the complex multidirectional eddies are superim-
particle and fluid (kg/m3) and , the dynamic vis- posed on the overall flow direction (Figure 3.3b).
cosity of the fluid (N s/m2 or Pa s). These oppose particle settling. In completely tur-
For most mineral grains the size of the particle bulent flow, the relationship between velocity and
has the greatest influence on the settling velocity, particle radius expressed by equation (2), is no
and as this is proportional to the square of the longer valid and changes to:
radius, equation (1) can be simplified to: v r1 / 2
v r 2 or v kr 2 (2)
Net flow direction
where k is a factor that embodies all of the other a)
factors in equation (1) except for particle radius.
However, the settling velocities of sediments in
natural environments such as lakes and rivers are
much lower than those calculated theoretically
from equation (1). There are several reasons for
this, but the two most important are: Laminar Flow
1. The Stokes’ law equation was derived for b)

perfectly spherical particles, as these offer
the least frictional resistance. Natural sedi-
mentary particles are not perfect spheres.
Although quartz grains are sometimes rea-
sonably spherical, grains of other minerals,
as a result of fundamental properties such as Turbulent Flow
cleavage, are more likely to have prolate or
bladed shapes. Particles of clay minerals Figure 3.3 Changes in state during fluid flow:
(A) laminar flow in which streamlines are broadly
present in mudstones and shales are flake-
parallel and movement around a particle is smooth with
like and so offer the maximum frictional resist- essentially linear flow lines and (B) turbulent flow
ance for their volume. Carbonate grains, with chaotic multidirectional eddies superimposed on
consisting of fragments of animal or plant the overall flow direction.
GFE-03.qxd 6/24/05 2:48 PM Page 64
As a result, the difference in settling velocities Very fine grained sediments carried in sus-
between relatively large and small particles is less pension by laminar flow (i.e. within the moving
than would be expected. Equation (1) indicates water column) form suspension deposits when the
that fine-grained particles are deposited more current decreases. When the flow is turbulent,
slowly than coarse ones. Consequently, when a coarse sand grains and even pebbles can be trans-
mixture of grain sizes enters moving water, coarse ported, but these are initially carried along the river
material is deposited first and finer material is or channel bed, and are not in suspension. This
carried down current. Coupled with reworking coarser sediment forms the bed load of the moving
this leads to a progressive size sorting of the orig- water. A much higher current velocity is required
inal mixture. Sorting, together with variations in to lift the bed load into suspension. Bed load move-
sediment loads and current velocities, leads to the ment takes place in two ways. Sand grains or peb-
deposition of laminated sediments. bles may simply be rolled downstream along the
Turbulent flow increases the ability of moving bed as the traction load, or they may bounce down-
water to sort sediment. This is because only the stream due to turbulence in the fluid and impact
coarser particles are able to reach the stream bed with larger particles on the river bed, a process
and smaller grains travel progressively further known as saltation. Typically, at moderate rates of
downstream before they are deposited. The rela- flow in water, sand grains saltate to a height of
tionship between the velocity of a current and the 2–3 mm before sinking back to the channel bed. If
sizes of particles that may be transported has been the current velocity decreases, such that part of the
determined experimentally and is shown, along bed load ceases to move, the resulting deposits are
with other data in Figure 3.4. The lower curve in referred to as traction deposits (Chapter 10).
the diagram represents the current velocities that A similar relationship between sediment
are required to keep grains of different sizes in movement and current velocity applies to trans-
motion, and divides the graph into a zone of transport in air. However, as turbulence is directly
port and a zone of deposition. related to the viscosity of the fluid and air has a
much lower viscosity than that of water, flow is
always turbulent. In addition, a grain of a given
1000
size has a much higher settling velocity in air than
Erosion and Transport in water because, as a result of the greater differ-
100
Erosion of sand ence in density, air provides less buoyancy. As a
and gravel begins
consequence, much greater wind speeds are
Current velocity cm s1
Erosion of consolidated Transport only

silts and clays begins as bed load required to hold or lift material into suspension
10 than are needed in water. Generally, only silt and
Deposition from
Erosion of unconsolidated bed load clay-sized grains (
150 m) form wind-blown
silts and clays begins
(aeolian) suspension deposits (Chapter 16). The
1.0
Transport in
Deposition sediment load carried in this way may be very
suspension only large and it has been calculated that dust storms in
the American Middle West may contain in the
0.1
0.001 0.01 0.1 1.0 10 100 order of 1000 tonnes per km3. Such storms may
Grain size (diameter) mm
be 2000–3000 km in length, with cross-sectional
Figure 3.4 Graph illustrating the relationship areas of 10 s or 100 s of square kilometres. In
between current velocity in a fluid and transport and 1977, a 24 hour storm in California with winds up
deposition of sedimentary particles. The graph is for to 300 kph is estimated to have removed nearly
fresh water and would be displaced downwards slightly 100 million tonnes of topsoil.
for salt water or upwards for air. Note the lack of ero- Saltation is a much more important method of
sion where finer-grained sediments have become cohe-
sive and that transport may continue after current sediment movement in wind than in water, espe-
velocities have fallen below those required for erosion cially for sand-sized grains. A typical sand grain of
and entrainment. 0.25 mm diameter saltating in air may rise steeply
GFE-03.qxd 6/24/05 2:48 PM Page 65
SEDIMENTOLOGY 65
to heights of 12–25 cm because of the low viscos- along a river bed when the current velocity
ity of air, before striking the ground about 10 cm reaches a value between about 25 and 75 cm/s,
downwind. On a large scale this creates a cloud of depending on conditions at the river bed. Where
sand grains flowing over the surface of windblown wind transport is concerned, wind speeds of the
sand dunes. Eye-witness reports of storms refer order of thirty times greater than those of water
to sand rising 1–2 m above the surface in this way. are required to set the same grain in motion. This
Although rolling of sand grains by the wind is not is a reflection of the greater contrast in densities
common, saltating grains may bombard those rest- incorporated in the Stokes equation. It seems to
ing on the ground so that they are effectively imply that wind is less effective as a transport
pushed along. This type of movement is referred to mechanism but this view neglects the size of wind
as surface creep, and should not to be confused systems that, as indicated, are commonly orders
with soil creep described earlier. However, in this of magnitude larger than those of rivers.
way, saltating grains can move grains with a diam- Wind deposits, loess, commonly form light
eter up to five or six times greater than their own. porous soils that may retain low densities (as low
Finally, as a consequence of the low viscosity, as 1.3) long after deposition. These carry particu-
impacts between grains in air are much greater lar risks in construction (Chapter 25). They are
than those in water with the consequence that sensitive to disturbance and particularly wetting
sand-sized grains more rapidly become well and may undergo sudden structural collapse. Such
rounded and acquire a ‘frosted’ surface texture. soils occur in about 30% of North America,
Paint and windscreens of vehicles, and structures including 500 000 km2 in lllinois and Indiana,
such as wooden poles and supporting cables, may and are common in arid and semi-arid areas
be at considerable risk from this sand-blast abra- in Wyoming, Colorado, Washington, Utah and
sion and pebbles exposed on the surface com- California. In China they are locally over 300 m
monly acquire planar facets forming ‘dreikanter’. thick. Spectacular collapse has resulted where
attempts have been made to provide irrigation
Erosion of sediments through the construction of canals, because on
It might be assumed that, once deposited, sedi- wetting the open fabric undergoes rapid hydro-
ment would automatically become buried and compaction.
compacted to form an irreversible addition to the For grains larger than 0.2 mm transported in
sedimentary record. This rarely happens, most of water there is an essentially linear relationship
it is soon eroded and reworked to undergo one or between the velocity required for erosion and
more further phases of transport before ‘final’ transport, and grain size (Figure 3.4). However,
deposition elsewhere. In rivers, transport may not for finer-grained particles, the resistance to
end until the sediment has reached the sea. To erosion not only ceases to decrease but actually
begin moving a grain requires a significantly increases progressively for consolidated sedi-
higher velocity than that at which it was deposited ments. This is because finer silts and clays show a
(See again Figure 3.4). This reflects the need to cohesive strength due to attractive forces between
overcome the friction between the grain and the the particles. Clay minerals in particular carry
channel bed. The velocities at which grains of surface charges and become linked together by
different sizes begin to move can be determined water dipoles. As a result they behave as plastic
experimentally, but not as precisely as the settling solids. Such sediments are referred to as ‘cohe-
velocities. A graph such as Figure 3.4 plotting sive’ and cohesion increases as the porosity of the
results, therefore, reflects a spread of data rather sediment decreases. A substantial increase in
than a narrowly constrained relationship. water velocity is required to lift the particle back
The area between the two upper curves on into suspension. Once a particle of silt 0.01 mm in
Figure 3.4 indicates the velocity ranges within diameter has been deposited, a velocity between
which sediment erosion will occur. A sand grain 1.4 and 10 cm/s is required to erode it, depending
2 mm in diameter might begin rolling or saltating on how dense the sediment has become. However,
GFE-03.qxd 6/24/05 2:48 PM Page 66
it requires the same current velocity (around Table 3.3 General terminology of beds and laminae.
500 cm/sec) to strip a consolidated surface of
1m Very thick bedding
grains of 0.001 mm diameter as it does to move
300 → 1000 mm Thick bedding
pebbles of 100 mm or so in diameter! 100 → 300 mm Medium bedding
Cohesive properties are important in the 30 → 100 mm Thin bedding
preservation of fine-grained sediments laid down 10 → 30 mm Very thin bedding
by rivers, in lakes or in coastal and offshore envir- 3 → 10 mm Thick lamination
onments on the continental shelf. They are also
3 mm Thin lamination
the reason why material transported into harbours
or similar enclosed areas by incoming tidal flow,
and deposited at slack water cannot be removed layers are only a few millimetres thick they are
by ebb currents. The resulting deposition can only referred to as laminae and the sediments or rocks
be remedied by dredging. The resistance to ero- are described as laminated. Layers thicker than
sion of consolidated clays is exploited in the ‘pud- about 1 cm are called beds (Table 3.3). Horizontal
dled’ clays used locally to protect river banks or planar laminae are generally characteristic of
from erosion. suspension deposits, but coarse horizontal layer-
The attractive forces that provide cohesion ing, planar bedding, also occurs in some traction
are dependent on water chemistry, and may be deposits. As current speeds increase, the sediment
destroyed in certain conditions. They may also be surface is built up into a series of bed forms that
responsible for deposition. In a weak electrolyte include the familiar ripple marks seen on beaches
such as sea water the particles stick together or and river bars swept by strong currents. (Bed
flocculate. Flocculation is responsible for much forms produced by wind action are described in
of the deposition in estuaries where, because flocs Chapter 16.)
have a relatively open structure, deposited muds The origins and structures of bed forms have
that remain submerged may initially retain porosi- been extensively studied in laboratory flume tank
ties as high as 70 to 80%. Slopes consisting of experiments. At low velocities flow is essentially
such materials are prone to failure and sub-marine laminar and there is no bed motion, giving a pla-
slumps have been reported on slopes of less than nar surface. As the flow rate increases grains begin
one degree. Silt-sized particles are also cohesive to move and there is a rapid transition from lami-
because of their small size in comparison to the nar to turbulent flow. Current ripples appear
strength of their surface charges. Silts and mix- almost instantaneously and generate a separation
tures of silts and clays may retain high porosity in the flow close to the bed (Figure 3.5; 3.6a). The
after deposition and during shallow burial and current adheres closely to the upstream (stoss)
may liquefy if subjected to vibration or shock. slope of the ripple, separating at the crest, with a
Some show dramatic changes in behaviour when rolling cylindrical eddy forming beneath, and reat-
subject to stress. For example, sediments that at taching on the next ripple downstream. The net
first appear solid and capable of bearing a load result is a decrease in the overall drag exerted on
may, when subject to vibration as when driving the bed by the flow. Sediment is typically eroded
piles, suddenly fluidify and lose any bearing on the stoss slope of the ripple and carried as bed
capacity. Such behaviour is described as load to the crest where grains are deposited as the
‘thixotropic’. flow ejects them into suspension. Grains settle
down the lee slope and periodically avalanche,
Sediments and sedimentary bed accumulating to form a series of parallel cross-
forms resulting from deposition in laminae. These are the dominant features pre-
flowing water served. However, where relatively large amounts
Suspension and traction deposition lead to charac- of coarser sediment are in suspension, erosion may
teristic types of layering in sediments that are pre- be prevented and sediment is deposited on the
served in sedimentary rocks. When the discrete stoss side as well, giving structures referred to as
GFE-03.qxd 6/24/05 2:48 PM Page 67
SEDIMENTOLOGY 67
Water surface
Shallow stoss slope Steep lee slope
Current direction
Flow separation,
deposition from bed-load
Sediment in suspension
Some deposition
Sediment moved in Bed load, Sediment spills from suspension
erosion of surface as avalanche
down face
Figure 3.5 Movement and deposition of sediment in a ripple. Solid arrows represent transport and deposition from
the bed load. Sediment is deposited as the flow separates at the ripple crest and may then avalanche down the lee
slope. Successive positions of the lee slope are indicated by inclined laminae formed as the ripple migrates down-
stream. Megaripple and dune forms (including those deposited by wind) have a similar structure and mechanism of
movement.
Figure 3.6 (A) Current ripples exposed on the foreshore at Wells-next-the-sea, Norfolk (UK) (B) Megaripples
with a wavelength of 8–10 m in a tidal channel at Wells-next-the-sea, Norfolk (UK) (C) Planar bedding an
parting lineation formed in the upper flow regime. Preserved in Jurassic sandstones at Cloughton Wyke,
Yorkshire (UK) (D) Standing waves generated in the upper flow regime. Channel on an intertidal beach,
Monifieth, Angus (UK).
GFE-03.qxd 6/24/05 2:48 PM Page 68
climbing ripples (see ‘Transport and deposition by of the ‘Upper Flow Regime’ conditions described
density flows’). by Froude. However, at higher velocities, the
As velocities increase, both the amplitude planar bed forms are disturbed, and a new wave
and the wavelength of the bed forms change form, about the same size as a dune, appears. This
to produce larger features variously described as is commonly accompanied by the appearance of
megaripples (Figure 3.6b), sand waves (Chapter standing waves (Figure 3.6d) on the water surface
16) and dunes (Figure 3.7). These migrate down- that periodically break upstream. The steeper
stream in the same way as ripples and are char- face of the bed form also faces upstream and
acterised by similar (but larger scale) flow bursts of deposition occur on the upstream face,
separation. They therefore generate large-scale although some downstream motion has also
sets of cross-laminae. However, as the current been observed. For this reason the structures
velocity continues to increase, ripples are swept are known as antidunes. Antidunes have been
away. Large quantities of sediment are now in observed in flume-tank experiments but are
suspension and are deposited as planar beds in a unlikely to be preserved in the sedimentary
carpet across the stream channel. The resulting record because, as the current decreases, planar
bed surface is commonly marked by linear sand beds are re-established. In any event they would
streaks (a parting lineation: Figure 3.6c) aligned be difficult to differentiate from downstream
parallel to the flow. In hydrodynamic terms the migrating dunes.
appearance of these planar beds heralds the onset The gradual downstream migration of sedi-
ment and bed forms under conditions where
sedimentation from the bed load exceeds rework-
Plane bed, no motion
ing, produces a carpet of material that may form
a prograding bar. Bars may also form in channels
Ripples (1-5cm)
during periods of low stage flow. The successive
Lower flow regime
accreting layers formed in these generate a cross-

stratification or cross-bedding in which individ-
ual layers are inclined to the main bedding planes
Megaripples, dunes and sand-waves (50cm->1m) that reflect the depositional surface (Figures 3.5,
3.7, 3.8, 3.9a). The assembly of layers in each unit
is known as a set and in a large-scale bar or dune
successive sets may be separated by erosion
surfaces. Laminae within a set may be planar or
Plane bed with primary current lineation
trough shaped, depending on the morphology of
the migrating bedforms (Figure 3.8). It is impor-
tant to note that features such as bars may accrete
Upper flow regime
laterally, and thus cross-bedding does not neces-

Antidunes (cm->1m) sarily face downstream. Commonly in bar accre-
tion there is a general downstream fining as well
as a fining upwards of the sediments involved.
Large-scale cross-beds, of the order of tens of
metres (Figure 3.9b), can be generated by the bar
accretion. Delta formation may also generate
cross-bedding on a similar scale but many deltas
Figure 3.7 The sequence of bed forms produced in are of kilometre scale and in these it may be dif-
response to increasing current velocity and their rela- ficult to identify relatively low-angle depositional
tionship to flow regimes and water movements. The
layering generated is shown schematically. Note the units.
appearance of in-phase water surface waves when So far, the precise current velocities that gen-
antidunes are formed. erate the various bed forms have not been stated.
GFE-03.qxd 6/24/05 2:48 PM Page 69
SEDIMENTOLOGY 69
Current direction ripples are restricted to sands finer than about

(a) 0.6 mm but do not form on sediments with a grain
size of less than about 0.1 mm. Sand waves and
dunes do not form in fine-grained sands, silts or
clays.
It is clearly impossible to define bed forms and
cross-stratification simultaneously in terms of the
(b) current velocity and water depth under which they
formed. It is therefore usual to describe them in
terms of the flow regime under which deposition
takes place. Planar beds, ripples, sand waves and
sand dunes form in the Lower Flow Regime, and
planar beds with sand streaks (parting lineation)
and antidunes form in the Upper Flow Regime
(Figure 3.10).
Cross-lamination or cross-bedding preserved
in sedimentary rocks may be used to determine
which way the current was flowing when the sed-
(c)
iments were originally deposited and may help to
define the orientation of a buried channel. This is
referred to as the palaeocurrent direction. It is
important to note, however, that depositional
systems vary in the nature of the record that they
Figure 3.8 Planar and trough cross-lamination preserve. A meandering stream, for example, with
formed by the lateral migration of ripples. In (A) linear a highly sinuous channel may preserve palaeocur-
ripples migrate as their stoss faces are eroded. By con-
rent directions that point in virtually any direc-
trast, in (B) similar structures beneath linguoid ripples
form a series of troughs visible in horizontal section tion, whereas in a braided stream or a more linear
and surfaces cut transverse to the flow direction. In channel they are more narrowly constrained to the
(C) climbing ripples are formed where large quantities mean transport direction. From the range of bed
of sand are carried in suspension and there is therefore forms present in a sequence of sediments, esti-
no erosion of stoss surfaces.
mates can be made about how flow conditions
varied with time.
This is because although there is an obvious and It may be difficult to recognise whether cross-
direct relationship between current velocity and stratification represents a cross-section of a bed
bed forms, the depth of the flowing water and the form parallel to the flow direction and errors may
mean grain size of the sediment forming the bed lead to the determination of incorrect palaeocur-
load are also significant. Water flowing in a deep rent directions. Figure 3.8 illustrates the three-
channel may cause no movement of the traction dimensional geometry produced by straight- and
deposits on the channel floor, whereas water flow- curve-crested ripples. It is clear that steeply
ing at the same speed in a shallow channel may dipping tabular cross-stratification may appear to
cause current ripples to form. Decreasing the be parallel laminated when viewed in a plane nor-
depth of flow while keeping the velocity constant mal to current flow. For this reason, it is essential
has the same effect as increasing the velocity of to identify the true direction of the depositional
the water at a constant depth. dip and at least two dimensions are typically
The effect of the mean grain size of the bed required, a single planar face in a quarry is never
load is shown in Figure 3.10. Planar beds arise by sufficient. Finally it is also necessary to return
deposition from the bed load of material coarser bedding to the horizontal (to unfold any folds)
than about 0.6 mm at low flow velocities. Current before the direction is determined. This can be
GFE-03.qxd 6/24/05 2:48 PM Page 70
Figure 3.9 (A) Large-scale trough cross-bedding in Quaternary aeolian carbonate sands, Sainte François,
Rodrigues, Mauritius (B) Large-scale cross-bedding in Jurassic carbonate sands, south of Dijon, France (C) Tidally
formed herringbone cross-lamination, south of Mombasa, Kenya (D) Flute casts on sole of turbiditic greywacke,
Bude, Cornwall (UK).
200 achieved by simple geometry using stereographic

Antidunes projections.
150
Mean Flow Velocity cm/s1
100
Upper Flow Regime Sedimentary bed forms resulting from
Plane Bed Dunes
80 oscillatory water movements
60
The movement of shallow sea water is controlled
Sand Waves by the ebb and flow of tidal currents and by wave
40 Current Ripples action. Both involve movement in more than one
Lower Flow Regime
30 Plane Bed
direction. If the currents in tidal flows are
approximately equal in strength, then the inclina-
No sediment
20 Movement tion of the cross-stratification generated will
0.1 0.2 0.3 0.4 0.6 0.8 1.0 1.5 2.0
reverse in direction on each tide. If sets reflecting
Mean Grain Size mm
opposing current directions are preserved, then a
distinctive herringbone pattern is produced
Figure 3.10 Graph illustrating the relationship (Figures 3.11 and 3.9c). However, opposing tidal
between mean flow velocity, mean grain size of the currents do not necessarily generate herringbone
sediment and the bed forms generated. Note that these
cross-stratification. Where one tide is stronger,
variables are not independent.
transport may leave only an erosion horizon, or
GFE-03.qxd 6/24/05 2:48 PM Page 71
SEDIMENTOLOGY 71
Herring-bone cross-bedding erosion surface
set
Current direction coset
Flood tide
cross-lamination produced by first flood tide
Ebb tide
cross-lamination produced by ebb tide lamination produced by first

flood tide partially eroded
Flood tide
cross-lamination produced cross-lamination produced

by first flood tide by second flood tide
Figure 3.11 Bed forms and sedimentary structures generated by tidal flow. Herringbone cross-bedding (indicat-
ing reversing tidal flow), and ripples generated by unidirectional, bi-directional and repeated unidirectional tidal cur-
rents. Note the asymmetry resulting where one tide is dominant.
reactivation surface, as a record of the reversal beds in response to wave action have more com-
(Figure 3.11). Unfortunately, reactivation sur- plex internal structures than those formed in
faces may also form where flow is unidirectional. response to unidirectional currents. Waves are
This occurs if the current velocity increases so produced by friction as wind blows over the surface
that erosion occurs but is insufficient to produce of a body of water. Within a wave, water particles
a change in the type of bed form generated. move with a circular (cylindrical) orbit with the
As a result of the greater variation in water direction of motion parallel to the direction of wave
movement the ripples that form on sea or lake motion as the wave-form passes (Figure 3.12a).
GFE-03.qxd 6/24/05 2:48 PM Page 72
The radius of the motion decreases with depth, water, but as the trough passes there is an offshore
such that when the depth is equal to half the wave- movement that is retarded by bed friction. The
length it can be considered negligible. Thus, in onshore movement tends to carry sediment land-
deep water surface waves have no direct effect on wards (up the beach as the wave breaks) whereas
the sea bed. When the water shallows, as the wave the offshore movement carries it seawards. This
approaches the shore, the orbit becomes flattened typically leads to the formation of symmetrical
to an ellipse due to friction with the sea bed ripples with complex cross-laminae facing in
(Figure 3.12b). The orbital paths mean that as the alternate directions. However, if the movement is
wave crest passes there is an onshore surge of more pronounced in one direction, asymmetrical
ripples may develop. When exposed on a bedding
plane in sedimentary rocks, wave-formed ripples
can usually be distinguished from current-formed
Deep water
ripples by their relative symmetry and by their
relatively straight and sharply defined crests.
Transport and deposition by

density flows
Density flows form a transition between mass
Shallowing slope gravity flows and fluid flow. In this transition, the
Wind initial movement of the sediment is due to gravity
but here the particles form a concentrated disper-
sion within the fluid. Once movement has begun,
it is the force of the moving fluid that keeps the
sediment in motion. Density flows may occur in
either air or water, but sub-aqueous flows are by
far the most important in terms of sediment
Structure of ripples
movement and erosion. In 1929 the Grand Banks
earthquake, off Newfoundland, triggered a gigan-
tic slump that devastated the continental slope
(Figure 3.13). It also produced a series of breaks
in sub-marine telephone cables but it was some
twenty years before it was realised that it was the
Figure 3.12 The orbital motion of water particles in slump rather than the earthquake that was respon-
waves in deep and shallow water. The particles remain sible for these. As it moved down the continental
effectively stationary whereas the wave form moves in slope the slump formed a concentrated dispersion
the direction of wind movement. In deep water (A) the
wave form is symmetrical, travelling in the direction of of sediment in sea water referred to as a turbidity
the motion of water particles at the top of circular current, moving down the slope and spreading
(cylindrical) orbits. The diameter of the orbits is equal across the abyssal plain (Chapter 22), decreasing
to wave height (trough to crest) at the surface but in velocity.
decreases with depth and where depth is equal to half Concentrated suspensions of sediment in water
the wavelength the motion become negligible. As the
water shallows (B) the counter flow at the base of behave as a fluid with greater density than the
the orbit is retarded by friction with the sea floor and water alone, and therefore, driven by gravity, are
as the seawards element of the motion slows the crest able to flow downslope beneath relatively sedi-
of the wave advances, becoming asymmetrical and ment-free water of normal density. The study of
ultimately breaking. At the sea bed orbits become simulated small-scale turbidity currents in labora-
progressively elliptical and are finally reduced to a
back-and-forth motion as a result of which the sedi- tory tanks has provided an understanding of this
ment forms symmetrical ripples (C) with laminae type of density flow. Each flow comprises a
facing both shoreward and seaward. ‘head’, a bulbous area with an overhanging nose,
GFE-03.qxd 6/24/05 2:48 PM Page 73
SEDIMENTOLOGY 73
Density currents are also known in air. The

most extensively studied form nuées ardentes,
180 m
violent surges of hot gases and incandescent ash
Newfoundland
ejected from certain volcanoes. In the eruption of
Mont Pelée in 1902 turbulent clouds with an inter-
nal temperature as high as 1075 C rolled down
slopes at speeds in excess of 160 kph. The burial of
North Pompei was probably the result of such a flow.
Large-scale dust-laden gravity currents have also
been recorded on the cold-fronts of weather sys-
Area of slumps tems (Chapter 23). By contrast, some forms of
and slides
snow avalanche are also effectively density flows.
The physics of sediment movement in a turbid-
ity current is complex; it is sufficient to say that a
Cables broken mass of sediment and water will begin to flow
59 min
and removed
down a sub-aqueous slope when the friction
183 min (73 kph) between the sediment–water mass and both the
Cables buried
slope and the overlying water, is overcome. The
but not broken
greater the density contrast between the turbidity
200 km current and the surrounding water, and the thicker
the flow, the faster it will move. The Grand Banks
earthquake broke twelve sub-marine cables in its
Figure 3.13 The inferred extent of the Grand Banks
path, and the precise timing of these breaks indi-
turbidity current of 1929. The epicentre of the earth-
quake and the area of the resulting slumps and turbid- cated that during its early stages the current veloc-
ity currents are shown together with the effects on ity must have been in the order of 100 kph. The
sub-marine telephone cables. Note the times of break- head of the flow is a region of devastating erosion,
age of the cables and the implied current speed calcu- with eddies generating characteristic scour fea-
lated for the flow at a specific cable given the distance
tures ( flute casts: Figure 3.9d) on the sediments it
travelled and time since the earthquake.
passes over. Larger pebbles or other objects
carried in the flow may form deep scratches (tool
and a generally thinner body. The head is a turbu- marks) on the sediment surface beneath.
lent chaotic mass in which water and sediment are Because many factors are involved in the initi-
mixed together. The nose is present because the ation and maintenance of turbidity currents, it is
bottom water into which it is intruding cannot be not yet possible to arrive at a comprehensive the-
deformed rapidly enough to be thrust upwards, oretical model. The driving force can be expressed
and so adheres to the bed, forcing the head (approximately) with reference to a unit of flow of
upwards. On the upper surface of the head, trans- fixed width and length:
verse vortices, billows, are generated by friction
Driving force xhB(2 1)gSin (N/m)
with the overlying (still stationary) water, forming
a wake. There is little mixing. Behind the head, where: x is the length of the element; hB its thick-
the flow moves in turbulent eddies and here dep- ness (or height above the base); 1 and 2 the den-
osition of the coarsest sediment begins at the sity of the medium and the current respectively; g
same time as the muddy sediment of the sea floor the acceleration due to gravity and the angle of
at the head is being eroded by the passage of slope.
the flow. It is believed that successive turbidity Although the initial velocity of the current is a
currents are responsible for gouging sub-marine function of the contrast in density between the
canyons, locally hundreds of metres deep, into the sediment-laden water mass and the surrounding
continental slopes. sea water, and of the thickness of the flow, the
GFE-03.qxd 6/24/05 2:48 PM Page 74
decrease in gradient towards the base of the con- because of the large volume of sediment still held
tinental slope leads to a rapid decrease in velocity in suspension the first structures to appear are
and results in deposition. The deposit formed is commonly parallel laminae with a parting lin-
referred to as a turbidite. However, deceleration eation and these are rapidly replaced by climbing
of the flow is not instantaneous, and deposition ripples.
therefore takes place in stages according to con- The ideal turbidite sequence (first outlined by
ditions within the flow and the settling velocities Arnold Bouma in 1962 and now referred to as a
of the various particles present. Typically, tur- Bouma sequence) ranges from a few centimetres
bidites exhibit graded bedding at the base, with to several metres in thickness. It is divided into
the coarsest particles in contact with the eroded five distinct units, conventionally designated A to E
sea bed, becoming progressively finer-grained (Figure 3.14), that can be interpreted in terms of
towards the top. However, because the flow the velocity of fluid flow. The thickness and
is dense and turbulent the sediment initially completeness of the Bouma sequence formed
deposited is commonly a poorly sorted muddy where the flow is at its maximum, decreases down
sand or greywacke. Once this period of rapid dep- current as the current velocity decreases and sed-
osition is over, the concentration of sediment iment is deposited. The scour and impact struc-
within the flow declines so that it begins to tures at the base of the sequence become smaller,
behave more like a conventional fluid. Deposition shallower and more localized. Frequent turbidity
then takes place from the bed load and changes in currents spilling from sub-marine canyons onto
bed forms reflect changes in the flow velocity (the the base of the continental slope build a sub-
sequence of bed forms is shown in Figure 3.14). marine fan.
Flow is still rapid (Upper Flow Regime), and The main types of transport and the varied sed-
imentary deposits that result from these are sum-
marised in Table 3.4. Transport by glacial action is
Clay Silt Sand Gravel treated fully in Chapter 13, and transport by wind
is discussed in Chapter 16.
Erosion surface
Pelagic sedimentation E 3.4 Sedimentary rocks

and turbidite muds
Laminated silts Lithification

and fine sands, D
lower flow regime The processes that transform loose unconsoli-
Climbing ripples, dated sediments into rocks are collectively referred
rapid deposition, C to as lithification. Lithification may be very rapid
lower flow regime
in geological terms, taking only a few years’ or
Plane bed, parallel
laminae. Upper B may be much slower. There are for example sands
flow regime
well over 200 Ma old that are still loose sedi-
Massive graded unit ments. The changes that take place are referred to
(greywacke), Upper A as diagenesis, and the two most important stages
flow regime, rapid
deposition. in this are compaction of the sediment as it con-
solidates under gravitational stress, and the
Erosion surface
cementation of individual grains.
Compaction. Freshly deposited sediments have
high porosities (the ratio of pore volume to total
Figure 3.14 Model Bouma sequence for a density volume). In freshly deposited sand the porosity
(turbidity) current illustrating the sequence of litholo- depends on a number of factors including the
gies and bed forms that may be expected from a single way in which grains are packed together, their
event. shape (their fabric: Figure 3.15), and whether the
GFE-03.qxd
6/24/05
Table 3.4 Summary of mechanisms of sediment transport and characteristics of the resulting deposits.
Mechanism Fluid flow Density flow Mass Sliding, deformation, melting

movement and refreezing of ice
2:48 PM
Method of Fluid flows downslope. Movement Movement of fluid initiated by density Grain inter- Material carried on or within
movement initiated by fluid current: material contrasts. flow turbulent, large volumes action driven flowing ice
carried as bedload or in suspension of sediment in suspension by gravity
Page 75
Agent of Water Air (wind) Turbidity currents Turbulent flow in Slumps and Brittle deformation and
transport (Traction (Saltation and surface (in water) air (Nuées ardentes) debris flows plastic flow of ice
saltation and creep but very large
suspension) volumes of sediment
in suspension)
Planar lami- Planar laminated Planar laminated Massive deposits, Slumps may Diamict deposits unsorted,
nated clays, clays and silts clays and silts at poorly bedded contain but may be locally layered
silts and fine (suspension load) top of sequence deformed by shear
grained (suspended load) bedding, debris
sands flows unsorted
(suspension or crudely
Nature of load) laminated
deposits
Cross-bedded Cross-bedded sands, Planar and cross-
and planar with steep foresets laninated sands
laminated (traction and
sands to saltation) Massive graded
gravels and poorly sorted
(traction and sands at base of
SEDIMENTOLOGY
saltation) sequence
(Turbulent
suspension)
75
GFE-03.qxd 6/24/05 2:48 PM Page 76
a) Packing and porosity quartz or similar silicate grains generally have a

low compressibility and their porosity after burial
may still be close to that at deposition even after
large stress increases. A few types of grains may be
able to deform plastically to reduce the total vol-
Cubic packing 47% Rhombohedral packing 26%
ume and some may fracture. Large stress increases
are concentrated at the points of contact between
b) Sorting grains and result in an increase in the solubility of
stressed areas, leading to pressure-dissolution.
Typically, however, this requires more than a kilo-
metre of burial. Pressure-dissolution results in an
increase in both the areas of contact and the
number of contacts between grains. Porosity
Well sorted Poorly sorted decreases and may ultimately approach zero. This
process alone may firmly bind the particles
c) Roundness
together to form a rock without the need for any
added cement. There is a linear relationship
between the bulk density of the rock and its
Angular Sub-angular- Well–rounded strength, reflecting the increasing size of the areas
sub-rounded of bonding between grains and the parallel
decreasing pore space.
Figure 3.15 The effects of grain packing and grain
The porosity of freshly deposited clay is much
size on porosity of a granular material: (A) spherical
grains of uniform size arranged in cubic packing have higher than that of sand. Because the clay mineral
a porosity of ⬃47%; in rhombohedral packing the same particles have a plate-like form and carry differ-
grains have a porosity of ⬃26%. In (B) a well-sorted ing charges on their faces and edges, they are able
sediment with a narrow range of grain sizes has a rela- to aggregate together to form loose box-like
tively high porosity whereas a poorly sorted sediment
structures, retaining porosities as high as 70–90%
with a wide range of grain sizes has a relatively low
porosity. (C) illustrates the effects of rounding, the when first deposited. As a result, clays are much
degree to which grain margins are worn. Note that this more compressible than sands and porosity
is distinct from sphericity, which describes the degree decreases rapidly with increasing stress, con-
of approximation to a sphere — an elongated grain can trolled almost entirely by burial. The large vol-
be as well rounded as a sphere.
umes of water that are expelled escape through
the overlying sediment where they play an impor-
tant role in later cementation. The final com-
sand also contains fine clay or silt particles. For paction of the clay to a rock requires a substantial
relatively well-sorted quartz sands, porosity on depth of burial and temperatures of 100 C or
deposition may be as high as 40%. By contrast, more, resulting in chemical changes in the clay
bioclastic sands, consisting of the skeletal remains minerals.
of a variety of organisms and made up of irregu- High porosity is an economically important
larly shaped grains may locally achieve porosities feature of sedimentary rocks that form natural
of 80–90%. reservoirs for the accumulation of oil, gas or
As successive layers of sediment are deposited, water, and porosity has a reciprocal relationship
burial causes compression and initially brings to strength. Overpressured formations are those in
about a reduction in porosity by rearranging grains, which porewaters are trapped, sealed in by overly-
the geological process known as compaction. In ing impermeable formations, with the result that
quartz sand with near spherical grains, porosity normal compaction cannot occur. Pumping of
may be reduced to around 30%, but values of this water or oil from these may lead to significant
order are exceptional. Sands, consisting mostly of surface subsidence as overburden pressures are
GFE-03.qxd 6/24/05 2:48 PM Page 77
SEDIMENTOLOGY 77
transferred from pore fluids to grains, resulting in and may form overgrowths on existing coarse
compaction. detrital grains or large numbers of new nuclei
Cementation. This involves the growth of new scattered at random on grain surfaces. In the latter
mineral crystals, referred to as cement, within the case crystals grow outwards from grain surfaces
pore spaces of sediment, and is common in both to fill pores, their relative sizes and shapes being
sandstones and limestones. The cement is most controlled by the space available to them. As a
often precipitated from porewaters expelled from result, crystals orientated parallel to grain sur-
inter-bedded or underlying clay sediments during faces are rapidly overgrown and remain small
compaction, but may also result from basin-wide whereas those that extend towards the centres of
fluid movements. The dissolved salts forming the pores are free to grow and become significantly
cement are commonly derived from water–rock larger.
interactions that occur during these large-scale Sands that are deposited in highly oxidizing,
movements but may also be generated locally by arid environments are commonly red coloured
late stage pressure-dissolution. The material dis- and this is why the ‘red beds’ found in the geo-
solved is reprecipitated as cement (Figure 3.16a), logical record are usually interpreted as indicating
Figure 3.16 (a) Thin section of quartz arenite with coarse calcite cement. Note that calcite crystals are so large
that they enclose many quartz grains that are approximately 0.25 mm diameter. Crossed polars. (b) Thin section of
lithic sandstone. A wide variety of grains, including lithic (rock) fragments are squeezed together by compaction
and boundaries have been modified by pressure-dissolution, there is no added cement. Grains are approximately
0.5 mm diameter. Plain polarized light. (c) Thin section of greywacke, note poor sorting, angular sand-sized grains
including quartz, feldspar and lithic fragments, in a dense muddy matrix. Largest grains are approximately 0.5 mm
diameter. Plain polarized light. (d) Thin section of bioclastic grainstone. Note fragments of a variety of fossils sur-
rounded by a granular calcite cement. Note the lack of compaction, implying early cementation. Grains are approx-
imately 0.5 mm diameter. Plain polarized light.
GFE-03.qxd 6/24/05 2:48 PM Page 78
a highly oxidizing atmosphere, and sometimes Table 3.5 The Udden-Wentworth scheme of grain-
even desert conditions. However, detailed investi- size classification of sediments. Compare with the
gations have shown that this interpretation must engineering classification given in Chapter 7.
be made with caution, as the sands in modern Class terms Grain size Sedimentary rock
deserts are not typically red. The colour in these (sediments) types
ancient rocks reflects the weathering of grains of
ferromagnesian minerals. These may be dissolved Boulders Conglomerate
256 mm
long after deposition — tens of thousands or pos- (rounded fragments
sibly millions of years — and form a red iron Cobbles or clasts), or breccia
64 mm (angular fragments
oxide stain that spreads over adjacent grain sur-
Pebbles or clasts)
faces. As little as 0.1% of iron oxide is enough to 4 mm
redden sediments. The red coloured surfaces may Granules
later be overgrown by quartz or other cements. It
is important to note that in the area where these 2 mm
Very coarse
processes have been studied in greatest detail, sand
intertidal and even marine sediments bordering 1 mm
the desert have also been stained red. Coarse sand
Although quartz cements are common, car- Sandstone
0.5 mm
bonates of calcium, magnesium and iron also Medium
form cements in sandstones. In all but a few very sand
0.25 mm
young marine limestones calcite is the dominant
Fine Sand
cement. The initial growth of cement within a
sediment binds the loose grains together, forming 0.125 mm
Very fine
a rock. Subsequent growth is largely responsible sand
for decreasing the porosity and increasing the
strength of the material. 0.0625 mm
Silt Siltstone Mudstone
Siliciclastic sediments and when non-
sedimentary rocks Clay 0.039 mm fissile, shale
Siliciclastic sedimentary rocks consist principally Claystone when fissile
of the silicate minerals that remain after weather-
ing; they are dominated by quartz. Sandstones are
defined in terms of their grain size (Table 3.5) and can be transported, small fluctuations in current
may be classified on the basis of the relative velocity and in the size and density of individual
proportions of quartz, feldspar, and lithic (rock) grains mean that sediments consist of a popula-
fragments (Figure 3.16b), providing information tion of grains rather than a single grain size. The
on the provenance, transport and depositional most important descriptor of this grain-size
history of the sediment. The common names are distribution is sorting which characterises the
illustrated in Figure 3.17. Sandstones with a car- range of particle sizes present in a sedimentary
bonate cement are described as calcareous (e.g. rock (Figure 3.15). Traditionally, grain size has
calcareous quartz arenite). Those in which a large been determined by sieving the sediment, but it
proportion of grains are carbonate fragments such can be measured more accurately by allowing
as bioclasts (see below) are described as calcare- grains to settle in a water column, or by direct
naceous. Both may weather in a similar manner to measurement. It should be noted that the sizes
limestones as the carbonate dissolves. attributed to grains by sieving bear little relation-
Although (as indicated in ‘Transport and dep- ship to their behaviour in fluids if they are of
osition by fluids’) there is a close correlation irregular shapes or show much variation in
between current speed and the size of grains that density, both grain shape and density may vary
GFE-03.qxd 6/24/05 2:48 PM Page 79
SEDIMENTOLOGY 79
important to appreciate that the degree of sorting

of sediment (or sedimentary rock) is also a func-
Mudstones
Quartz Quartz tion of the source rock from which it was derived.
arenite Erosion of a poorly cemented, well-sorted sand-
10%
Greywackes stone cannot produce a poorly sorted sediment,
(10-15%
mud matrix)
unless material is added from another source. The
more cycles of erosion and deposition a sediment
50% 50% suffers the more ‘well sorted’ (consisting of a nar-
Increasing rower range of grain sizes) it is likely to become.
mud content ‘Poorly sorted’ sediments (consisting of a wide
and commonly discontinuous range of grain sizes)
Arkose Lithic arenite
may be generated after deposition by mixing of
Feldspar 50% Lithic (rock) fragments originally separate grain-size populations by the
activities of burrowing animals (bioturbation) and
Figure 3.17 Illustration of the classification of clastic
(siliciclastic) sandstones based on the relative propor- by later infiltration of a finer-grained component,
tions of quartz, feldspar and rock (lithic) fragments a feature of some desert sands that would other-
(QFL). A variety of similar schemes are available but wise be well sorted. It should be noted that there
there is no agreement on precisely where boundaries is an important divergence in terminology that
should be drawn. Nevertheless, most sandstones can be
may lead to confusion. Many engineers refer to
plotted within the space defined by this geometrical
form. sediments comprising a broad range of sizes of
grains (commonly referred to as particles) as ‘well
graded’.
widely in carbonate sediments. Settling has two The fabric or texture of a rock is a reflection of
advantages. First the results reflect the real hydro- sorting and, with the exceptions noted, is a prod-
dynamic behaviour of the grains rather than an uct of the way in which the original sediment was
extrapolation based on their size alone. Second, it transported. It reflects the relationship between
measures the behaviour of a population (hindered coarser grains, which form the dominant compo-
settling) rather than that of individual grains. nent, and a finer-grained matrix that occupies
Irrespective of the method, a mean or average inter-grain spaces. In sediments where the domi-
grain size present in the sample can be calculated, nant grains form a self-supporting framework, the
and sorting represents the spread of grain sizes fabric is described as grain supported. Where the
about this mean size. larger grains are dispersed in a finer-grained
So far as depositional processes are con- matrix, with little or no inter-grain contact, rather
cerned, good sorting is achieved either where in the way currents are dispersed in a fruit-cake,
grains have undergone aqueous transport for long the fabric is said to show matrix-support. Matrix-
periods, or where sustained currents in the trans- support is common in sediments deposited by gla-
porting medium are selective with respect to the cial action, by mass movement and from density
size of grain that they are able to move as, for flows, where sorting is inefficient. Grain-support
example, in wind transport. Poorly sorted sedi- is common in suspension and traction deposits
ments are typically formed when a mass of mate- from fluid flow (wind and water) that are effi-
rial is transported and deposited rapidly without ciently sorted.
being reworked. This may occur when there is a The attrition between grains during transport
sudden reduction in the velocity of a current, or causes rounding. Sediments that have been
when the transporting medium is non-selective transported far enough for the grains to become
with respect to grain sizes, as in debris flows, tur- well rounded (Figure 3.15) and also efficiently
bidity currents and glacial transport. The typical separated so that the sediment is well sorted are
deposits of turbidity current transport are poorly described as texturally mature. During aqueous
sorted greywackes (Figure 3.16c). However, it is transport, inter-grain impact is reduced due to the
GFE-03.qxd 6/24/05 2:48 PM Page 80
cushioning effect of the relatively viscous water. Textural maturity is a sign of prolonged trans-
Attrition is therefore reduced although sorting port during one or many cycles in water or of
is moderately efficient. The effects of attrition aeolian transport. Mineralogical maturity may
decrease as grain size decreases. By contrast, result from a period of prolonged chemical
during wind transport inter-grain impact is great, weathering or substantial transport by water that
attrition is high and sorting is also very efficient. may include multiple sediment cycles. Glacial
Both water transported and aeolian (wind- diamicts, debris flows and sub-aerial density
blown) sands show increasing textural maturity deposits are commonly both texturally and min-
as they are transported further. However, glacial eralogically immature. Aeolian deposits may be
diamicts and deposits from mass wasting are texturally mature but mineralogically immature
characteristically texturally immature because due to the absence of significant chemical weath-
very little sorting or attrition is possible during ering in arid climates.
transport.
The mineral content of a sediment also indi- Limestones
cates ‘maturity’ because both physical and chem- Although limestones and dolomites (Ca–Mg car-
ical weathering continue during transport and bonate rocks) form only about 15% of the Earth’s
lead to the progressive disintegration and decom- exposed sediments, they have great economic
position of the less stable silicate minerals. These importance. Some 40% of the world’s oil reser-
processes continue long after burial. Thus, voirs are found in them. They are host rocks to
although feldspars are far more abundant than some of the largest lead–zinc deposits and are
quartz in igneous and metamorphic rocks, quartz closely associated with phosphorites and baux-
is one of the most abundant minerals in sedimen- ites. Many form important aquifers. They are the
tary rocks. This is because feldspars are chemi- key component in the manufacture of cement and
cally weathered more rapidly than quartz and are widely used in aggregate production. Because
decompose to form clay minerals. The presence of their solubility and the unusual landforms that
of feldspars in a water-lain sediment indicates develop on their surfaces they provide some of the
that the periods of both weathering and transport most intractable foundation problems in civil
must have been brief, with physical weathering engineering (Chapters 24 and 26).
dominant and deposition occurring close to the Among the soluble products of weathering,
source area. Limited weathering, rapid erosion that tend to end up in the sea, are Ca2 and HCO3
and rapid deposition commonly also result in the ions (Table 3.1). These are the raw materials from
formation of sands containing large numbers of which calcareous sediments, and eventually lime-
rock (lithic) fragments (Figure 3.17). These may stones, are formed. Surface sea water is supersat-
dominate all other grain types. It is important to urated with respect to both calcite and aragonite,
note that rocks resulting from the deposition of the common marine carbonate minerals. The
sediments containing these less stable compo- majority of both modern and ancient carbonate
nents weather more rapidly when used as build- sediments are biogenic, that is, they owe their
ing stones. origin to biological activity. A vast number of
With increasing intensity of chemical weath- microscopic planktonic organisms that float in
ering and longer aqueous transport, almost all the surface layers of the oceans and macroscopic
silicate minerals except quartz are decomposed, creatures living on the sea floor extract CaCO3
and so the end products are either quartz sand or from sea water to build their skeletons. These
clay-rich mud. These sediments are said to be include molluscs, echinoderms (sea-urchins
mineralogically mature and, by contrast, those and their relatives), corals and calcareous algae
that contain a high proportion of minerals such (Chapter 22). Skeletons may consist of the miner-
as feldspar, mica, ferromagnesian minerals or als aragonite or calcite. After the death of the
rock fragments are said to be mineralogically organisms they accumulate on the sea floor to
immature. form calcareous sediments. At the present day,
GFE-03.qxd 6/24/05 2:48 PM Page 81
SEDIMENTOLOGY 81
about 95% by area of marine calcareous sedi- muds are formed in part as the breakdown prod-
ments are accumulating in the ocean basins ucts of calcareous algae but also by direct precip-
beyond the continental shelves. The chalk of itation. In addition, towards the margins of the
southern England is an example of a carbonate banks, current-swept shoals are sites for the accu-
mud formed in this way. The remaining 5% mulation of oolites. These sand-sized grains are
by area of carbonate sediments are found in the made up of concentric layers of aragonite crystals
shallow waters on continental shelves or fringing that also form directly from sea water. Oolitic
volcanic islands and comprise a high proportion limestones are common in the rock record and
of macro-skeletal debris. many are used as building stones or form impor-
Regardless of where carbonates form, one tant aquifers. Direct precipitation of carbonate
overriding factor controls their accumulation as from sea water is also responsible for local cemen-
relatively pure carbonate sediments. This is the tation of the sea floor.
absence, or virtual absence, of any siliciclastic Limestones are classified on the basis of the
sediment. A high influx of siliciclastic material relative proportions of their principle compo-
will dilute the skeletal remains that are deposited nents, grains, and muddy matrix or cement,
so that mixed calcareous–siliciclastic sediments following Dunham (1962) as shown in Table 3.6.
are formed (these are generally referred to In addition, however, the dominant grain types
using terms such as ‘sandy limestone’). Muds in such as bioclasts (skeletal fragments of organ-
suspension may cloud the water and inhibit the isms) or ooids may be identified (Figure 3.16d).
growth of photosynthetic organisms including In the geological past, when sea level was
algae in reef-building corals. The most widely higher, shelf seas were far more extensive than
distributed carbonate sediments are found in the they are now, and so we find that shelf carbonates
tropics although there are occurrences even in make a very important contribution to the geolog-
polar regions. ical record. At present, extensive shelves occur off
Not all carbonates are biogenic. Because sea the north-east coast of Australia (the Great Barrier
water is supersaturated with respect to calcite and Reef, 2000 km long), off Honduras in the
aragonite a slight increase in temperature and or Caribbean, in the Arabian Gulf and Red Sea and
salinity may be enough to cause precipitation. The forming the Bahamas Bank in the western Atlantic
presence of Mg2 ions in sea water apparently and the Seychelles Bank in the western Indian
inhibits the nucleation of calcite and so it is the Ocean. In many locations the shallow-water areas
mineral aragonite that most commonly appears. generated by reef systems have provided the foun-
In the Bahamas at present this may take one of dations for land reclaimed for the construction of
two forms. In shallow lagoon areas, carbonate airports, harbour works and housing.
Table 3.6 Dunham’s (1962) classification of limestones based on the relative proportions of the dominant com-
ponents. Most limestones can be accommodated within this scheme with prefixes to indicate the dominant grain
type e.g. oolitic grainstone.
Grain supported Mud supported Components Depositional texture

(grains in contact) (muddy matrix) bound by obscured by diagenetic
organic growth change
during deposition
No mud present Sparse mud More than 10% Less than

(crystalline between grains grains 10% grains
cement)
Grainstone Packstone Wackestone Mudstone Boundstone Crystalline carbonate
GFE-03.qxd 6/24/05 2:48 PM Page 82
Chemical and organic sediments contains relatively large amounts of water. It is

Although generally beyond the scope of this book, unstable and recrystallises relatively rapidly to
the reader should be aware of a variety of chemi- microcrystalline quartz; cherts generally consist
cal and organic sediments, many of which are of of this. However, it is important to note that many
economic importance. Under the broad heading of limestones and dolomites also contain chert or
‘coals’ are two distinct groups. Those that formed flint nodules that form by a process of chemical
as land-based deposits of woody plant material, replacement of the original carbonate by silica
and accumulated where the trees grew and died, and are not biogenic. They may be decimetres
and those composed of algae, spores and frag- or even metres in diameter and have important
mented plant debris that were washed into lakes or implications in construction, and in aggregate and
the sea to accumulate. Like carbonates, these cement production.
accumulations require the absence of any other In addition to the organic sediments described
sediment. The gradual conversion of decaying above, a number of sedimentary rocks have an
vegetation to coal takes place with increasing tem- unequivocally chemical origin. As their name indi-
perature, typically resulting from an increasing cates, evaporites form where sea water or inland
depth of burial. Volatiles are lost and as a result drainage basins evaporate, and their presence in
the residue comes to consist of an increasing pro- the sedimentary record is therefore a good indica-
portion of carbon. During the formation of peat, tion of a former arid climate. The principal evap-
there is little alteration of plant material, but there orite minerals are gypsum (a hydrated form of
is a gradual progression from this to brown coals calcium sulphate, CaSO4 2H2O), anhydrite (the
and on to bituminous and anthracitic coals, hard, anhydrous form of calcium sulphate, CaSO4) and
black and shiny materials with few traces of the halite (NaCl), although potassium and magnesium
original plant structure. It is important to remem- sulphates and chlorides and a variety of other salts
ber that the gas methane is generated as a result of may also be present. In experiments, carbonates
the coal-forming process and may accumulate in form when sea water is evaporated to about
tunnels through coal-bearing rocks, potentially 50% of its original volume, gypsum at 20% and
forming an explosive hazard. halite at 10%. Evaporite minerals are an economic
Petroleum hydrocarbons (crude oil and gas) are resource but in some areas play a major role as
also formed by the alteration of organic matter impermeable seals to petroleum reservoirs.
buried in sediments. They originate in organic-rich Because of their high solubility they represent a
source rocks. Almost all mud rocks contain some significant hazard to construction even at low
organic matter (typically 2–3%) but source rocks concentrations. They may also play an important
such as oil shales may have as much as 10–12%. detrimental role in weathering.
Like coals, these mature with increasing tempera- Ironstones are sedimentary rocks containing
ture, with optimum conditions on average 2–3 km more than 15% iron that may be in the form of
beneath the surface but they retain a high hydro- iron oxides, carbonates, silicate or sulphides.
gen content. The dispersed droplets of oil and dis-
solved gas that result migrate upwards to be
trapped and accumulate in porous reservoir rocks. 3.5 Sedimentary facies and
Bedded cherts originate as biogenic siliceous environments
oozes on the sea floor. The oozes comprise the
siliceous skeletal remains of microscopic marine Ancient sedimentary environments are recon-
plankton, diatoms and radiolaria. These are nor- structed on the basis of the interpretation of fea-
mally found in deeper waters than calcareous tures of the individual rock units, of their
oozes, but some mixed siliceous–calcareous relationships with units above and below, and of
oozes also occur in shallow waters. These skeletal the way in they change laterally. Geologists use
remains consist of silica in the form of opal. This the term ‘facies’ to describe the individual units,
has a poorly defined crystal structure that but there are variations in meaning.
GFE-03.qxd 6/24/05 2:48 PM Page 83
SEDIMENTOLOGY 83
Strictly, ‘facies’ describes the specific set of fea- omissions, by bedding planes and by erosion sur-
tures that characterise a body of rock and is there- faces (Chapter 5).
fore wholly objective. Although it is now being As depositional models have become more
used to refer to both igneous and metamorphic refined, there has been a growing realisation that
rock associations, the term originated as a des- although our observations of the Recent and
criptor of sedimentary sequences. The features Quaternary strata are invaluable, it is unlikely that
that are considered can be observed in the field they indicate the relative importance of the varied
and laboratory. They include the overall geome- processes and depositional environments that have
try of the sediment body, the thickness, shape and existed through geological time (Chapters 5 and 6).
distribution of beds; their colour, mineral compo-
sition, texture, including grain size, shape and
sorting; sedimentary structures and fossil con- 3.6 Summary
tent. Two derivatives have appeared that conform
to this objective approach. Lithofacies refers Sedimentology is concerned with the manner
solely to the lithological features of the unit, in which sedimentary grains are produced, trans-
geometry, textures, and structures, while biofa- ported, deposited and buried to become sedimen-
cies is concerned with the nature and distribution tary rocks. It includes the study of sequences of
of the fossil flora and fauna present in the unit. sedimentary rocks aimed at interpreting the
However, other more subjective variations are nature of past environments and much of the
also in use and for some authors the term regret- remainder of this book outlines the characteristics
tably implies a specific origin, as in deltaic facies of such sequences.
or turbidite facies. Further discussion of the con-
cept of facies is included in Chapter 5 and in
Reading (1978). References
Facies models result from interpretations of
Blatt, H., Middleton, G. and Murray, R. (1980) Origin
both modern and ancient facies sequences, and of Sedimentary Rocks (2nd edn). Prentice-Hall,
for some it is the ancient record that is dominant. Englewood Cliffs, New Jersey, 782pp.
The ‘essence’ of the model is achieved by ‘boiling Dunham, R. J. (1962) Classification of carbonate rocks
away the local details, but distilling and concen- according to depositional texture. In Ham, W. E.
trating the important features they have in com- (ed.) Classification of Carbonate Rocks. American
Association of Petroleum Geologists, Memoir 1,
mon into a general summary’ (Walker, 1984, p.6). 108–121.
Such models may be specific to a particular Kukal, Z. (1971) Geology of Recent Sediments (trans-
sequence or generalised for an environment (see lated by Helena Zarubova). Academic Publishing
examples in Chapters 7, 8 or 15). House of the Czechoslovak Academy of Science
An implicit element in this distillation is the and Academic Press, London, 490pp.
Reading, H. G. (ed.) (1978) Sedimentary Environments
principle of uniformitarianism, that the geological and Facies. Blackwell Scientific, Oxford, 615pp.
record can be interpreted in terms of processes Walker, R. G. (1984) General introduction. In Walker
acting at the present time. Earth history has, for R. G. (ed.) Facies Models (2nd edn). Geoscience
the most part, unfolded in a gradual evolutionary Canada Reprint Series 1. Geological Association of
manner, and has not been dominated by a series of Canada, 1–10.
universal catastrophes, although there have been
significant rare events in geological history. It is
nevertheless important to realise that the individ- Further reading
ual units observed within a rock sequence are
The following texts provide more detailed intro-
records of specific events. While some of these
ductions to sedimentology.
were protracted in geological terms, many were of
short duration, measured in weeks, or even hours. Allen, J. R. L. (1985) Principles of Physical Sedi-
Most of geological time is represented by the mentology. George Allen & Unwin, London, 272pp.
GFE-03.qxd 6/24/05 2:48 PM Page 84
An accessible and well-illustrated introduction to Tucker, M. E. (2001) Sedimentary Petrology: An Intro-

the physics of sediment movement. duction to the Origins of Sedimentary Rocks (3rd
Collinson, J. D. and Thompson, B. D. (1982) Sedimentary edn) Blackwell Science, Oxford, 262pp.
Structures. George Allen & Unwin, London, 194pp. Scoffin, T. P. (1987) An Introduction to Carbonate
A well-illustrated introduction to the origin of bed- Sediments and Rocks. Blackie, Glasgow and London,
forms and sedimentary structures. 274pp.
Leeder, M. R. (1982) Sedimentology: Process and A clear and well-written introduction to carbonate
Product. HarperCollins Academic, London, 344pp. sediments and rocks (limestones and dolomites)
A wider-ranging account of sedimentary rocks and addressed only superficially in the texts above.
environments. This book provides both general descriptions of
Pettijohn, F. J., Potter, P. E. and Siever, R. (1972) Sand sedimentary environments and an introduction to
and Sandstone. Springer, Berlin and New York, sedimentary petrography.
618pp.
GFE-04.qxd 6/24/05 2:54 PM Page 85
4. Tectonics
Tony Waltham
4.1 Geological structures of the crust. The relatively brittle crust is therefore
broken into a number of plates (Figure 4.2),
All rocks have been subjected to some degree of which move relative to each other over the mantle
structural deformation, the results of which influ- cells. Plate boundaries are the zones of movement
ence the physical properties of the rock mass. The and deformation. There are three types.
only materials that have not been deformed Plates move sideways past each other along
are some of the younger sediments, whose low conservative boundaries. These have little out-
strengths mean that engineers classify them as ward significance, except for the development of
soils, rather than rocks. The small-scale effect of large earthquakes. The San Andreas Fault is part
the deformation is to create fractures and struc- of the Pacific/American plate boundary roughly
tural weaknesses that are the prime factors in along the Californian coast.
determining the engineering properties of in situ Plates move apart at divergent boundaries,
rock masses (Figure 4.1). After their deforma- which are also constructive, as new crust is cre-
tion, the upper parts of those rocks that constitute ated to add to the plates and fill the gap. Basaltic
the landmasses are partially removed by erosion. magma is fractionated from the mantle and rises
The large-scale effect of the rock deformation to form new oceanic floor of sheeted dykes, with
then becomes apparent by determining outcrop the excess creating submarine volcanoes. Only
patterns within the landscape, the shape and oceanic crust is created, and the Mid-Atlantic
distribution of some larger landforms and also Ridge is over the boundary where the American
the larger structural elements that remain beneath plate diverges from the African and European
the surface. plates.
Tectonic structures are by definition those that Plates move together at convergent boundaries,
originate by large-scale geological processes. They which are destructive, as crust must be destroyed
are ultimately the consequence of plate movements, to accommodate the convergence. Only oceanic
and the vast proportion of rock deformation takes plate can be totally destroyed when it is subducted
place within or adjacent to the orogenic belts that and melted into the mantle. Continental plate can
develop along convergent boundaries. Non-tectonic only be crumpled, shortened and therefore thick-
structures are smaller features that develop by sur- ened, to create an orogenic belt along a mountain
face processes, including unloading fractures, ice chain over the convergence. The Andes have
deformation (see Chapters 13 and 14), cambering formed where the American plates overrides the
and sliding (see Chapter 8). Pacific floor. The Himalayas are an even greater
mountain chain formed at the collision of the
Plate tectonics Indian and Eurasian plates where both are formed
All geological processes require energy, which of continental crust.
has its ultimate origin in the residual and radioac- The processes of orogenesis that occur within
tive heat energy of the Earth’s core. This creates an orogenic belt encompass nearly all the major
convection cells within the mantle by heating its geological processes (Figure 4.3). Most igneous
base. The rising currents roll over the top of the rocks (except oceanic basalt), sedimentary rocks,
cells, creating horizontal movements at the base metamorphic rocks and rock deformation owe
87
GFE-04.qxd 6/24/05 2:54 PM Page 86
Figure 4.1 Tectonic deformation of limestones exposed in a road cut about 5 m high in Greece. Both the sub-
horizontal reverse faults in the thick bed of limestone and the tight folding in some of the more thinly bedded rock
were formed by the same compressive earth movements.
Figure 4.2 Outline map of the major crustal plates and the active orogenic belts along the convergent boundaries.
GFE-04.qxd 6/24/05 2:54 PM Page 87
TECTONICS 87
Figure 4.3 Diagrammatic cross-section through convergent and divergent plate boundaries, indicating the relative
positions where major tectonic processes take place.
their origins to processes that are part of, or con- maximum and minimum values in boundary
sequent on, orogenesis at convergent plate bound- zones of plate convergence and divergence
aries. In addition, mountain uplift provides the respectively.
energy for rapid erosion and sediment transport. The state of stress within the ground is defined
It is therefore ultimately responsible for landscape in terms of the directions of maximum, interme-
evolution; geomorphological processes evolve diate and minimum stress, which are mutually
much more slowly on stable plates away from perpendicular. The scale of rock deformation
active convergent boundaries. depends on the differential stress, which is the dif-
ference between the maximum and minimum
Rock deformation stresses. The style of rock deformation depends
The behaviour of a rock or rock mass during on the confining stress, which is the value in the
tectonic deformation depends on the nature of the minimum direction. Under lower confining
rock, the ambient environment of temperature and stresses and at lower temperatures, brittle behav-
water pressure, and the degree of stress imposed iour leads to the development of fractures, both
upon it. These factors ultimately determine joints and faults. At higher confining stresses, at
whether a rock is folded or fractured, though rock higher temperatures and aided by higher pore
strength and the rate of deformation are also sig- water pressures, ductile behaviour allows plastic
nificant. The stresses imposed on rock are a func- deformation and the creation of fold structures.
tion of tectonic processes; they increase steadily Tectonic subsidence may carry sedimentary
when it sinks due to tectonic subsidence, and dra- rocks to depths of 5–8 km, where burial stress
matically when it is caught within zones of plate rises to 125–200 MPa. Clays are nearly always
convergence. Within the ground, vertical stress ductile, and therefore deform into fold structures,
increases by 25 MPa (and temperature increases though most clays also contain networks of
by about 20 °C) with each kilometre of burial. At micro-fractures. Sandstones and limestones may
depths of more than a few kilometres, horizontal also be ductile at these low confining stresses
stresses are generally less than vertical stresses, where pore water pressures are high and strain
but they vary by an order of magnitude between rates are low. Typical rates of strain within
GFE-04.qxd 6/24/05 2:54 PM Page 88
tectonic deformation are around 1% per 30 000 are a normal component of rock weakening
years. Far more rapid strain can be seen in the sag- within the weathering profile. In steep slopes and
ging of unsupported roof slabs in built structures. cliffs, the unloading fractures are steeply inclined
At depths of less than a few kilometres, horizon- and become potential slip surfaces for many of
tal tectonic compression deforms rocks upwards, the landslides and large rock falls that are major
towards the minimum stress. Open symmetrical components of slope degradation in moun-
folds are therefore the dominant structures in tainous terrains. Numerous massive rockfalls
sequences of sedimentary rocks. have occurred on the precipitous slopes of
Granite becomes ductile at confining stresses of Norway’s fjord country, and the most catastrophic
about 500 MPa and temperatures of around 400 °C, have been where tectonic joints, faults or foliation
conditions which pertain at depths of around 20 km sub-parallel to the cliffs have opened up due to
within the cores of orogenic belts over convergent this stress relief. A large-scale extension of this
plate boundaries. The same conditions account for process is seen in the Holocene ‘deglaciation
the main part of regional metamorphic processes. landslides’ of Iceland, which developed when ice
Metamorphic rocks are therefore characterised by support was lost from the walls of U-shaped val-
plastic deformation and complex folding. leys. On the small scale, rock bursts are failures
Faults are brittle failures that can develop at of the walls in deep mines, where the loss of con-
depths where locally high differential stress fining stress was due to excavation of the mine
exceeds rock strength. Failure is aided by high tunnel. These generally occur at depths
strain rates, most conspicuously imposed along of ⬎ 600 m where overburden stress concentrated
convergent boundaries. High confining stress in tunnel walls exceeds the tensile strength of the
means that frictional resistance on the fault plane is rock: dangerously explosive rock bursts occur in
high, so that strain energy accumulates in the rock strong rocks (unconfined compressive strength,
adjacent to the locked fault. Ultimately, the fault UCS ⬎ 140 MPa) where stress can accumulate
does slip with a sudden movement when the before failure; in weaker rocks, the unloading
strain energy is released as ground vibrations — effect is largely an increase in rock spalling.
expressed on the ground surface as earthquakes. All rocks have a stress history of multiple
High burial stresses mean that rocks are rarely phases, created by deep burial and denudation
subjected to absolute tensile stress, but effective exposure and by stages of their tectonic plate
tension is created in the direction of minimum evolution. They therefore have multiple suites of
stress. Normal faults and small tension gashes are deformation structures. Whether these are brittle
therefore created normal to the minimum stress in or ductile may depend on the smallest of variations
deeply buried rocks. Rifts and grabens subside in the ground environment; rock structures are not
where gravitational stress exceeds horizontal mutually exclusive. A ground engineer must antic-
stresses that are reduced by crustal divergence, ipate finding both folds and fractures in any rock
and the consequently low confining stress means mass under assessment.
that brittle failure occurs.
At shallower depths, brittle fracture increas- Fold structures and fold mountains
ingly dominates and joint systems become the An infinite variety of three-dimensional fold
normal mode of rock failure. Within about 100 m structures can be created by the intense plastic
of the ground surface, vertical stress is nearly deformation that is common in metamorphic
always less than horizontal stress, except in the rocks, though folds are generally more simple in
faces of high cliffs. Denudation and surface low- sedimentary rock sequences. Fold terminology
ering progressively reduce burial stress, and stress can be comparably complex, but may often be
relief fractures develop parallel to the exposed reduced to a few varieties of up-folded anticlines
surface as the ground relaxes upwards towards the and down-folded synclines (Figure 4.4).
minimum stress. In roughly level ground, this cre- The intensity and style of folding are also vari-
ates the sub-horizontal unloading fractures that able and are part of any full description of rock
GFE-04.qxd 6/24/05 2:54 PM Page 89
TECTONICS 89
Figure 4.4 Shapes and types of folds in a sedimentary rock sequence.
Figure 4.5 Structural relationships within a fold.
structure. Gentle folding, with dips ⬍ 20°, gener- account for many of the structural contrasts in
ally forms rounded structures, but more intense cover rocks at outcrops.
folding may be very angular. Both rounded and The axial plane of a fold is the plane that con-
angular folds may occur with wavelengths down tains the hinge-line axes on each bed. It defines the
to millimetre scales, but the largest folds, with geometry of the fold and is created normal to the
wavelengths on the kilometre scale tend to have maximum stress that caused the folding (Figure
more rounded profiles. 4.5). Therefore in low-grade metamorphism new
A single phase of rock deformation tends to micas grow parallel to the axial plane and create the
produce fold structures that are of roughly uni- cleavage in slates; these important planes of struc-
form intensity across many kilometres of terrain tural weakness are also known as ‘slatey cleavage’
due to the uniform stress distribution within rocks or ‘axial-plane cleavage’. Strong folding and
of comparable strength. A major exception is the squeezing may induce shear along the axial-plane
development of a monocline where near-surface cleavage and so create parallel folds in the more
rocks are draped over a fault step in the underly- deformable rocks; these contrast with the concen-
ing basement, while rocks on top of an intact tric folds produced by bending of less deformable
basement block remain relatively undeformed. sequences (Figure 4.4). Slatey cleavage is not the
Displacements within an unseen basement only rock fabric induced by metamorphism.
GFE-04.qxd 6/24/05 2:54 PM Page 90
Schistosity is defined by sub-parallel micas with rules of stratigraphy cease to apply (Figure 4.7).
grain size larger than in slate, but its pattern is Both the Alps and the Himalayas are heavily
generally more complex and variable than that of eroded fold mountains, structurally very distinct
cleavage, typically due to multiple phases of devel- from the stratimorphic fold mountains (see below).
opment. Foliation banding in gneiss may also be
complex, but is less important as a structural weak- Fractures, joints and faults
ness within the rock. Natural fractures occur in all rocks. The most
The folding that is produced by crustal conver- widespread are those induced by tectonic stress,
gence accommodates lateral shortening at the cost when the rock fails, normally along shear planes.
of vertical thickening. Most of the world’s moun- Joints are fractures with no measurable displace-
tain ranges and chains are essentially fold moun- ment. Faults are fractures along which there has
tains (notable exceptions are the volcanic chains, been significant movement, which often, but not
and the granite mountains, which are largely always, leaves telltale slickenside scratches on
remnants exposed by erosion of folded cover parts of the fault plane. Both may be formed by
rocks). The Himalayas are the prime example, typ- the same tectonic stress, where the strain is
ified by Dhaulagiri and Annapurna which are accommodated either by large movements on a
formed by a bed of carbonate over 2000 m thick few faults or by small movements on many joints.
stacked up to triple height by almost isoclinal Joints may be irregular in shape, but most are
recumbent folding (Figure 4.6). Plate convergence close to planar and they generally occur in sets
creates overfolds and recumbent folds where rocks or systems of sub-parallel fractures. Variable
override each other, and these develop into nappes imposed strains may be accommodated by three
in the cores of the orogenic belts. Nappes may have sets of joints that are roughly perpendicular to
moved more than 50 km over their basal thrust each other. This creates the simplest pattern of
planes, and stacked nappes create the famously joints, commonly found in sedimentary rocks,
complex layered structure of the Alps where the where one set is aligned on the bedding planes,
Figure 4.6 The recumbent folds that stack up a thick carbonate unit to form Dhaulagiri in the plate convergence
zone of the Nepal Himalayas. The thinning of the beds towards the south is only apparent in this view, as the South
Face of Dhaulagiri recedes far beyond the gneiss ridge in the foreground.
GFE-04.qxd 6/24/05 2:54 PM Page 91
TECTONICS 91
Figure 4.7 Structural complexity in the over-thrust nappes of the central Alps.
Figure 4.8 Slate exposed in the Nantlle Quarry, in North Wales, has seven systems of planar weaknesses, but only
four are recognisable in this view, where the ubiquitous cleavage is vertical and straight into the face.
with the other two sets across the beds. As these the Nantlle quarries of North Wales (Figure 4.8)
intersecting conjugate joints are produced in shear, contains five sets of joints in addition to the cleav-
at about 45° to the maximum stress, the angle age and the bedding. Each system of planar weak-
between them is bisected by the associated fold ness is recognised (and individually named) by
axes (Figure 4.5). In reality, most rock masses have the quarrymen — whose livelihoods depend on
been subjected to multiple phases of deformation extracting blocks of slate devoid of fractures other
and their joint patterns are more complex. Slate in than the ubiquitous cleavage.
GFE-04.qxd 6/24/05 2:54 PM Page 92
Figure 4.9 The three main types of fault. The tear fault is also known as a strike slip fault.
Joint density, or conversely joint spacing, may potential for future displacement. Broken rock,
vary considerably, and is always a key factor in shattered between the moving fault blocks, is
the engineering properties of a rock mass. A mean known as fault breccia (Figure 4.10), while fault
fracture spacing of ⬍ 200 mm defines a rock gouge is the same material ground to a fine paste.
mass of poor quality, with low bearing capacity Fault breccia can occupy zones many metres
and high permeability. Some granites may be across, and a single fault zone may have dozens of
truly massive, with fracture spacing of ⬎ 10 m, sub-parallel or braided faults, each with breccia,
and hence constitute very solid ground. Many or passing laterally into a single brecciated zone.
clays have few or no visible fractures, but their All this faulted, broken rock represents locally
properties at low water contents may be defined weak material that may constitute difficult ground
by networks of irregular and curved joints that are for an engineer; also most fault breccia has
spaced at ⬍ 10 mm in all directions. enhanced permeability. There is no direct link
Faults are significant structural breaks through between the scale of displacement and the extent
rock structures. In many cases they are spaced of brecciation on a fault, as the latter depends on
apart on a scale of kilometres, but structurally the confining stress at the time of movement.
complex rock units can have many smaller faults Fault gouge is rarely more than 300 mm thick but
that lie anywhere across a spectrum of rock frac- its presence makes subsequent sliding on the fault
tures from rare large faults to abundant small plane more easy and more likely.
joints. The scale of a fault’s movement is generally Fractures that constitute structural weakness in
expressed by its ‘throw’, its vertical component of rocks are not limited to tectonic faults and joints.
displacement. Faults are described as normal when Bedding planes are the dominant planar weak-
formed under tension, reverse when under com- nesses in sedimentary rocks. Most bedding planes
pression, or tear when in shear (Figure 4.9), were formed as breaks or changes in the original
though there are many other terms that can be used sedimentation and many are actually very thin
to describe fault movement. Normal and reverse layers or partings of clay or shale. With or without
faults may be recognised by the relationship the clay parting, bedding planes are structural
between their dip direction and downthrow side features and are therefore commonly developed
(Figure 4.9) but many vertical faults are not easily by tectonic stress, effectively into joints. Strong
classified. In a simple stress field, reverse faults lie concentric folding causes slip along the bedding
parallel to fold axes, and both are perpendicular planes, which therefore become faults with
to normal faults, but many structures are compli- slickensides normal to the fold axes. Cleavage,
cated by multiple stress phases. schistosity, columnar joints (formed by cooling
A fault zone is commonly a notable element of contraction) and unloading joints (formed by
weakness in a rock structure, regardless of any stress relief close to the surface) are all planar
GFE-04.qxd 6/24/05 2:54 PM Page 93
TECTONICS 93
landforms. Denudation rates are higher in areas of

higher relief and higher elevation, in the younger
mountain chains and in environments with higher
precipitation and runoff (Table 4.1). Rates of mean
denudation are less than those for valley incision,
but steep slopes have erosive losses that are locally
very high. Rock type accounts for the contrast
between slow denudation the ice-scoured base-
ment rocks of the St Lawrence basin and the rapid
erosion of the loess lands of the Huang He basin
(Table 4.1).
Localised crustal uplift is still occurring due to
isostatic rebound after the melting of thick
Pleistocene ice caps. The fastest current uplift
of this style is around 9 mm/year in the Gulf of
Bothnia, where coastlines are receding as a con-
sequence. Surface uplift and decline can also
occur on and around volcanoes as they inflate or
deflate (Chapter 23). Though the main signifi-
cance of these movements relates to eruption
hazards, Pozzuoli’s harbour in Italy was rendered
Figure 4.10 A fault breccia of broken rock constitut- partly unusable due to 2 m of volcanic inflation in
ing a zone of weak ground a metre wide. the late 1980s.
Tectonic movements do become significant
when they are localised on active faults, most of
structures that the engineer should regard as types which are in well-defined zones on or related to
of fracture and therefore potential weaknesses in plate boundaries. Frictional resistance, under the
a rock mass. confining stress of burial depth, ensures that
most faults move in a series of jerks, each of
Active earth movements which creates the radiating shock waves that are
Absolute rates of long-term tectonic movement earthquakes. Seismology (Bolt, 1999) and earth-
are so low that they are rarely of engineering quake engineering (Ambraseys, 1988) are major
significance. Mean relative lateral movements subjects with extensive literature. The prediction
of crustal plates are only 10–100 mm/year, and of earthquakes has proved almost impossible, to
vertical movements are generally an order of the extent that America and Japan have now vir-
magnitude less. The uplift of mountain chains, tually cut research funding. The more productive
currently by about 15 mm/year in the Himalayas earthquake research is currently into ground
and Karakoram, powers more active erosion behaviour and the means of engineering struc-
and river incision, and consequently creates steep tures to withstand vibration damage. The mini-
slopes that are unstable on engineering timescales mal earthquake death tolls of recent events in
(Chapter 8). Subsidence of sedimentary basins Japan and USA contrast the massive tolls in
is essential for thick sediment accumulation Asian regions of mainly adobe buildings, which
(Chapter 3) and can be significant to man-made indicates the value of properly designed and
structures where deltaic regions are subsiding built modern structures of steel and reinforced
partly due to crustal sag and partly due to ongoing concrete.
sediment compaction (Chapter 12). Infrastructure can be built successfully across
The balance between tectonic uplift and surface active faults if an allowance is made for movement.
denudation determines the long-term evolution of The Alaskan oil pipeline crosses the Denali Fault
GFE-04.qxd 6/24/05 2:54 PM Page 94
Table 4.1 Typical rates of surface lowering by erosion. Environment rates are
from various sources (largely after Burbank and Anderson, 2000). River basin
rates are estimated from sediment yields (after Milliman and Syvitski, 1992).
Valley incision rates are identified by drainage of cave systems.
Mean denudation rates in environments mm/year

Temperate hill country, mean relief 500 m 0.04
Temperate mountains, mean relief 2500 m 0.40
Old mountain chains at altitudes around 1500 m 0.08
Young mountain chains at altitudes around 1500 m 0.40
Himalayan region 0.4–0.5
Salt Range, Pakistan 1.5–2.0
Arabian desert 0.05
New Mexico semi-arid lands, USA 0.017
Active glacier basins 0.1–50.0
Landslide-prone slopes in New Zealand Alps 5.0–12.0
Mean rates in complete river basins
St Lawrence, Canada 0.001
Niger, Africa 0.012
Mississippi, USA 0.044
Amazon, Brazil 0.07
Huang He, China 0.52
Valley incision rates
Cumberland Plateau, USA, fluvial late Pleistocene mean 0.06
Yorkshire Dales, UK, fluvial and glacial late Pleistocene mean 0.12
Gunung Mulu, Sarawak, fluvial Pleistocene mean 0.19
Figure 4.11 The Alaskan oil pipeline crossing the active Denali Fault. The pipeline trestles rest on concrete beams
with steel caps and a teflon surface so that they can slide freely within the flexible limits of the pipe, while the
ground is displaced laterally beneath them.
GFE-04.qxd 6/24/05 2:54 PM Page 95
TECTONICS 95
ground, as opposed to the strength of unfractured,

intact rock. The mean fracture spacing is difficult
to estimate in poorly exposed rock without being
subjective. A convenient objective measure is
Rock Quality Designation (RQD), calculated
from the lengths of pieces of unbroken core
recovered from a borehole:
RQD ⫽ (⌺ lengths of cores each

⬎ 100 mm long) ⫻ 100
/ borehole length
A value of RQD ⬎ 70 indicates a generally

sound rock mass; RQD can be 0 for well-fractured
or thinly bedded materials.
The strength of a well-fractured rock mass may
be no more than 10% of the strength of the same
material almost devoid of fractures. This relation-
ship may be expressed as the Rock Mass Factor
(RMF), which can be correlated with mean
fracture spacing or with RQD (Figure 4.13). All
classifications of rock mass (Barton et al., 1974;
Figure 4.12 Sixth Street in Hollister, California, built Bienawski, 1973) are based largely on fracture
across a branch of the San Andreas Fault. Slow, smooth, spacing and other properties of the fractures.
creeping displacement along the fault deforms struc- A simplified guide to the safe bearing pressures
tures on its outcrop but shock waves typical of signifi-
cant earthquakes are not created. applied by foundation loading of engineered
structures is also based on fracture spacing,
besides the intact rock strength (Table 4.2).
Besides joints and true fractures, all other
on sliding foundations designed to tolerate 5 m of rock structures and rock fabrics influence the
lateral displacement (Figure 4.11). Some faults engineering properties of rock masses. Cleavage,
with smooth, sliding displacement do not create due to mica parallelism, may be barely visible in
destructive earthquakes but do cause structural a compact slate, but its orientation is critical to
damage directly over them (Figure 4.12). Vertical potential shear failure and can influence uncon-
movements on faults can be very significant in fined compressive strength (UCS) by a factor of
coastal regions. Large areas along Alaska’s south- five (Figure 4.13). Schistosity has a similar influ-
ern coast were transformed to wetlands by as much ence, except strengths of coarse-grained schists
as 2.5 m of subsidence that accompanied the 1964 never match those of compact slate. Foliation in
earthquake. Some offshore islands were uplifted by gneiss is due to mineral orientation and banding,
10 m and moved seaward by 14 m in the same but is less significant in this generally strong
earthquake, as the fault movement allowed the rock. Most sedimentary rocks are anisotropic to
stress relief and the unfolding of rocks that had some extent, even where they have no visible
been crumpled against the fault until it slipped. structure in between their major bedding planes,
which are always weak. The compressive
Engineering implications strength of apparently homogeneous Triassic
The density or spacing of fractures is the key sandstone from the English Midlands declines by
parameter in defining rock mass strength — the about 20% under loading that is oblique to the
strength of natural, disturbed rock within the bedding. The degree of parallelism within the
GFE-04.qxd 6/24/05 2:54 PM Page 96
Figure 4.13 The influence of fractures on rock properties: (a) the decline of the Rock Mass Factor (RMF) with
decrease of mean fracture spacing (and decline of RQD) and (b) the variation in unconfined compressive strength
(UCS) of slate with respect to orientation of its cleavage.
Table 4.2 Guideline values (expressed in MPa) of system, or even just one individual fracture, can
safe bearing pressure (SBP) for the engineering design determine the stability of a cut face, which may
of foundations, with respect to the fracturing of the have to cut back to the planar weakness, regardless
rock mass (indicated by RQD) and the unconfined of design concepts that were not related to rock
compressive strength of the intact rock (UCS).
structure.
Fracture spacing (mm) 60 200 600 The main impact of rock folding on engineer-
RQD 25 75 90 ing is the creation of dipping bedding planes,
UCS 100 MPa 4 8 12 which control most large landslides in rock
UCS 25 MPa 1 3 5 (Chapter 8). Nearly all slope failures in strong
UCS 10 MPa 0.2 1 2 rock are planar, or partially planar, and are there-
fore related to inherent rock structure. Bedding
planes in sedimentary rocks are critical to sta-
orientation of clay mineral particles accounts for bility where the dip is greater than the angle of
some of the variation in the shear strengths of friction — generally around 17⬚. Stability is
mudstones and clays, and is reflected in the more achieved in steeper dips by cohesion, which is
conspicuous contrast between massive mudstone difficult to determine in situ and is best estimated
and laminated shale. from local empirical data. Bedding plane irregu-
Patterns and orientation of systems of fractures larities are a component of fracture roughness,
and other structural elements greatly determine the which is critical to sliding resistance, but is diffi-
ease of excavation back to a clean face, and stable cult to quantify in the field. Fractures and bed-
faces are most easily engineered parallel to the ding planes that are slickensided or contain
dominant fracture system. Similarly, fracture significant fault gouge or clay parting are best
patterns define the extent of over-break in under- regarded as having nil cohesion; this matches the
ground excavations. A well-defined fracture concept of the drop to residual strength in
GFE-04.qxd 6/24/05 2:54 PM Page 97
TECTONICS 97
Figure 4.14 Profiles through escarpments formed in sedimentary rock sequences.
sheared soils. Rock structures that dip at angles formed where strong limestones or sandstones
of 20–30⬚ towards an exposed face are therefore are undercut and stripped bare by more rapid
the least stable; steeper structures are likely to differential erosion of weaker clays or shales
have failed already, and gentler structural dips (Figure 4.14). Most of England south and east of
are generally stable. the Pennine crest consists of scarplands, with the
Faults that can be seen to displace Holocene limestone Cotswolds and the chalk Downs as
sediments are classed as active and then define just the largest of a succession of escarpments.
construction practice in accordance with local A gently inclined dipping stratimorph is one ele-
building codes for seismic regions. Unbroken ment of an escarpment, except where the dip is
Holocene sediments over a fault outcrop define steeper than the naturally stable slope so that a
the fault as inactive. In such cases, and in all cases more symmetrical hogsback is formed. A scarp
in seismically quiet regions such as Britain, a face may form dramatic cliffs where it is effi-
fault may have little impact on engineering works. ciently undercut in a mountain environment, but
The exceptions are faults that may be reactivated it is generally degraded into a rounded step in
— notably by mining. All the deformation within a more mature landscape. The well-known grit-
a migrating subsidence wave over an active long- stone edges of the Derbyshire Peak District
wall mine panel (Chapter 12) may be accommo- (Figure 4.15) are all scarp faces above gentler
dated on any available fault plane, causing a shale foot slopes, but their rugged scars are all
ground step, or more commonly a narrow zone of
abnormally high subsidence damage. Now that
mining has ceased under many parts of Britain, a
major cause of fault reactivation is water table
rebound when mine pumping systems are finally
switched off; rising groundwater pressures reduce
effective stress across the fault planes, and
thereby permit renewed displacements in shear.
4.2 Tectonic landforms
Stratimorphs and escarpments

A stratimorph is defined as a component of the
landscape morphology that is an element of the
geological structure — generally a bedding plane
on top of a strong rock unit exhumed by erosive
stripping of weaker overlying rocks. Figure 4.15 A gritstone edge in the Derbyshire
Escarpments are the most widespread land- Pennines (UK), Curbar Edge, a natural scarp face that
forms in areas of dipping sedimentary rocks, has been modified by quarrying.
GFE-04.qxd 6/24/05 2:54 PM Page 98
the remains of old quarry faces that were cut into

much less rocky natural hill crests.
The most dramatic stratimorph is an anticlinal
mountain whose entire surface is the rounded arch
of a strong limestone or sandstone before it is
breached and reduced by erosion to leave a pair of
facing escarpments. The classic examples are in the
Zagros Mountains of southern Iran (Figure 4.16),
where soft clays and gypsum have been stripped off
the strong limestones to reveal the anticlines that
are the first pointers to deeply buried petroleum
reservoirs.
Plateaus, mesas and buttes

A large stratimorph in horizontal rocks forms a
plateau, of which South Africa’s Karoo Plateau and
Table Mountain are the perfect examples on differ-
ent scales. At many other sites strong beds form
only wide stratimorphic steps across hilly terrains.
Where these are horizontal they can be misinter- Figure 4.16 The classic stratimorph of Kuh-e-Kailan
preted as remnants of old erosion surfaces. The formed by an anticline of Asmari Limestone in the
wide limestone benches around Ingleborough, in Zagros Mountains of Iran (photo: Hunting).
the Yorkshire Pennines (UK), are both erosion
surfaces and stratimorphs (Figure 4.17); erosion
surfaces cutting across dipping rock sequences do these can only be a few metres high (Figure 4.19),
also exist, though they are less well defined, and and rarely survive in long-term landscape evolu-
there are also some sloping benches that are tion. Larger fault scarps are the product of
dipping stratimorphic surfaces. repeated movements, with some of the clearest
Arizona’s Monument Valley is the finest examples occurring along the margins of rift
demonstration of erosive dissection at a plateau valleys (see below).
margin to leave residual hills with profiles that Far more common are fault-line scarps. These
create one of the world’s most spectacular land- are steps in the landscape along the line of a fault,
scapes. Widely spaced vertical joints in the strong where the step is not the direct consequence of the
sandstone, 150 m thick, define vertical cliffs fault throw (though they are often described
above the slopes of the underlying shale– incorrectly as fault scarps). In most cases the
sandstone sequence. The larger remnants are landscape step has been formed by differential
known as mesas, but they are known as buttes surface lowering of rocks of contrasting resist-
when their width is reduced to less than their ance on either side of the fault; the lower side is
height (Figure 4.18); the thinnest remnants are on the weaker rock and is not necessarily the
variously described as towers, spires or pillars. down-throw side of the fault. In many cases, ero-
sion has left the scarp as a degraded slope that has
Fault scarps and fault-line scarps retreated from the line of the fault. The Craven
As faults are major breaks in bedrock structure, Faults have defined some major landscape fea-
they commonly define linear features within the tures in the Yorkshire Pennines (Figure 4.17).
landscape. The simplest forms are fault scarps — They mark the edge of the Craven Uplands with
steps in the landscape created by the fault 30 km of eroded hillsides that step down 200 m
movements. Some are formed during a single dis- to the south, though the fault throw is well over
placement with its associated earthquake, but 1000 m. Giggleswick Scar is a more cleanly
GFE-04.qxd 6/24/05 2:54 PM Page 99
TECTONICS 99
Figure 4.17 Outline map of the major landforms of Ingleborough and the Craven fault zone in the Yorkshire
Pennines (UK). The largest limestone benches lie where the 400 m erosion surface coincides with the stratimorph
on the top of the strong limestone. All three faults down-throw to the south.
Figure 4.18 The spectacular residual hills of Monument Valley, Arizona, with the small East Mitten Butte in the
near right, and the wide Sentinel Mesa in the left distance. Between them are the slender Stagecoach Buttes, with a
narrow unnamed pillar on the mesa’s right flank.
GFE-04.qxd 6/24/05 2:54 PM Page 100
influenced by small faults, but may excavate

major troughs where fault-guided fluvial valleys
are invaded by ice. The Great Glen of Scotland
cuts right through the Highlands along the line
of a massive tear fault that has been excavated
over millions of years; there has been no recent
movement on it. Wave action is the most selective
of erosion processes, and faults are commonly
marked by bays, inlets, fissures, caves and geos
within marine cliff lines.
Rift valleys, horst blocks,

fins and arches
Figure 4.19 A fault scarp created by surface dis- A graben is an elongated crustal block displaced
placement on the Hebgen Fault, in Montana; the asso-
downwards between two normal faults. A rift val-
ciated earthquake was of Richter magnitude 7.1.
ley, bounded by inward-facing fault-line scarps, is
created by recent graben sinking where the faults
defined fault-line scarp of strong limestone that effects have not yet been obliterated by surface
stands 100 m above a shale lowland; it is defined degradation. Long histories of graben sinking leave
by one clean fault within the wider fault zone. The faults with kilometres of throw, but rift valley
70 m high limestone cliff of Malham Cove is a scarps are normally much smaller (Figure 4.20).
fluvial feature that has retreated a kilometre from The Rhine Graben in western Germany has sub-
the fault; though its origins lie at the fault-line sided 4400 m in 45 million years, and has filled
scarp, it should not now be described as a fault with sediment so that the modern valley floor is
feature. only about 500 m below the marginal hills. The
The broken rock along many faults constitute normal faults are the product of regional tension
zones of weakness that are commonly picked out over a zone of deep crust that has been heated and
by erosion. Straight segments of river courses expanded laterally; commensurate vertical expan-
are common indicators of fault lines, as are lin- sion accounts for the highlands that flank most rift
ear lakes in shield terrains. Glaciers are little valleys.
Figure 4.20 Profile through the marginal faults of a graben beneath a rift valley.
GFE-04.qxd 6/24/05 2:54 PM Page 101
TECTONICS 101
More complex crustal tension creates fault in Utah’s Canyonlands, are elongated blocks
blocks that rise or subside relevant to each other bounded by normal faults that originally opened
along bounding normal faults. The Basin-and- in tension over a rising salt dome. Subsequently,
Range country of Nevada, USA, is the classic they have subsided behind sandstone slabs that
example, with uplifted horsts of Precambrian are sliding very slowly over the salt towards the
rocks forming mountains that rise above interven- canyon of the Colorado River. They are so fresh
ing basins floored with thick young sediments. that they now form rift valleys. Maintained
The tension structures are continued southwards tension in the ground causes joints and faults to
to Death Valley and Panamint Valley, both deep open. The famous Joint Trail leads right through a
rift valleys in the California desert. sandstone ridge in a single straight fissure 400 m
Small grabens are common in the head zones long, 40 m deep and only 1 m wide (Figure 4.21);
of landslides, where wedges of ground subside elsewhere, joints open beneath the soil cover and
behind the moving landslide block. The Grabens, cause localised linear collapses.
The same tension joints on the flanks of anti-
clinal salt domes in Utah have created the nar-
row ribs of rock known as fins in the strong
Entrada Sandstone. The faces of the fins have
then retreated by frost action, spalling and
undercutting as a part of long-term weathering,
and the numerous breaches of the fins have
enlarged into natural arches (Figure 4.22).
Patchy selective weathering is not uncommon,
but arches are rarely formed except in struc-
turally defined fins.
Engineering implications
The rock strength and denudational resistance of
a stratimorph makes it inherently stable, until it is
Figure 4.22 Double O Arch, weathered through a

Figure 4.21 The open tension fracture that carries narrow fin between open joints in the Entrada
the Joint Trail right through a sandstone ridge in Sandstone of Arches National Park, Utah. Visible
Canyonlands, Utah; there has been minimal addi- beyond and through the arch are fins within the same
tional opening of the fissure by weathering and face joint system, formed in tension over an elongated salt
retreat. dome.
GFE-04.qxd 6/24/05 2:54 PM Page 102
Figure 4.23 Valley side deformation by camber, gulls, joint dilation and plateau spreading.
undermined by erosive excavation of an adjacent geological structure in their smaller features and in
weaker rock. The world’s largest recognisable their landslides. Similarly, old-age terrains have
landslide was caused when a slab of dipping lime- minimal expression of lithology. The great shields
stone failed where support was partially removed of Precambrian metamorphic rocks have largely
by excavation of a river valley. In the Saidmarreh been reduced to lowlands with minimal relief,
slide, dated to 10 400 BP, 22 km3 of rock slipped though major faults can control much of the
off an anticlinal stratimorph in the Zagros drainage pattern. On Canada’s Laurentian Shield,
Mountains of Iran (Watson and Wright, 1969). faults can be traced for hundreds of kilometres by
Smaller landslide events are a normal component their lines of elongated lakes. In contrast, mature
of landscape evolution in mountainous areas of landscapes commonly reveal their bedrock geol-
folded rocks. ogy. Canada’s Front Ranges and the limestone
Plateau edges, scarp faces and fault scarps are mountains that fringe the Alps in south-eastern
characterised less by planar slides than by rockfalls France both have all their major landforms tightly
and toppling failures, which become a significant controlled by their geological structure.
engineering hazard where strong caprocks overlie Sedimentary rock sequences in mature low-
weak clays or shales. Squeezing of soft underlying lands are especially sensitive to selective erosion
clays causes ground flowage towards open slopes, as the weaker clays, shales and mudstones are
and thereby deforms and weakens them. Camber stripped away from the more resistant sandstones
folding, gull opening, lateral spreading and frac- and limestones. The results are scarplands or
ture dilation (Figure 4.23) are all precursors to terraced flatlands depending on the regional dips.
landslide development along plateau edges and Clays form broad valleys with the rich soils most
scarp faces of low-dipping escarpments. suited to arable farming. Sandstones provide the
dry sandy soils that support low-value heathland
and also in England the more extensive uncleared
4.3 Lithology and landforms woodlands; Sherwood Forest housed Robin Hood
because it was so extensive along the outcrop of
Differential erosion of sedimentary rocks the Triassic sandstones.
The lithology and structure of bedrock are para- Carbonate rocks almost invariably form positive
mount to geomorphological evolution and land- elements of a landscape. The older and stronger
scape development. Different rock types have limestones and dolomites are generally strong in
most expression within the landforms of mature their own right forming white scars and cliffs in
terrains (Figure 4.24), where they provide the often very spectacular landscapes (Chapter 24).
data for engineering geomorphological analysis Chalk is generally a weak rock, but still forms
and ground model construction (Fookes, 1997; positive landforms, notably the Downs and Wolds
Hutchinson, 2001). Rapid uplift of youthful moun- of England (Chapter 26). This is partly because its
tain ranges causes rapid and unselective erosion, outcrops are everywhere surrounded by clays, but
with steep and unstable slopes that only reflect also because its underground drainage reduces
GFE-04.qxd 6/24/05 2:54 PM Page 103
TECTONICS 103
Figure 4.24 Some influences of rock type and structure on landforms (after Fookes, 1997).
fluvial erosion of its surface. The distinctive short Salt and gypsum are both weak rocks. They are
grass of the chalk downlands is not a true function also highly soluble, and consequently can develop
of the bedrock, it is classic ‘sheepwalk’ where all some elements of karst terrains (Chapter 24). Salt
but grass is kept down by grazing sheep since their can only survive at the surface in a desert, as in
introduction by man (Figure 4.25). The natural Israel’s Negev, and is otherwise geomorphologi-
vegetation of English chalk is oak woodland. cally more significant with respect to its ground
Figure 4.25 Downland on chalk, with the typical short turf of ‘sheepwalk’, on the sides and floor of a deep dry
valley in the Yorkshire Wolds.
GFE-04.qxd 6/24/05 2:54 PM Page 104
Figure 4.26 Curved exfoliation joints in the granite of Half Dome above Yosemite Valley, California. The scale is
given by people on the cable ladder on the left.
subsidence (Chapter 12). Gypsum can form karstic rock can leave dykes as upstanding ridges or walls
uplands in the moderately dry regimes of central (Figure 4.28), while plugs and volcanic vent fills
Turkey, western Ukraine and northern Russia, but can form isolated high crags (Chapter 23). Sills
it is rapidly eroded and rarely reaches to outcrop in can form stratimorphs and escarpments similar
more temperate or wet environments. to those in a strong sedimentary rock, though
dolerites are distinguished by their dark colour
Landforms of igneous rocks and commonly by massive columnar jointing.
Granite is one of the most distinctive rocks in Slow underground cooling creates columns larger
terms of its landforms. Its high intact strength (typ- than those typical of lava flows. The Whin Sill is
ically with UCS of 200 MPa), massive structure a conspicuous feature of the northern Pennines
(giving it a high rock mass strength) and the large
size of typical batholiths, combine to make granite
a major positive feature in most landscapes. Large
domes are granite’s primary landform, best known
in California’s Yosemite Park. They are clearly
developed by progressive exfoliation and stripping
of curved sheets separated by relaxation (unload-
ing) joints parallel to the surface (Figure 4.26). The
massive vertical cliffs of granite country are
largely formed on faults or major joints exposed by
‘ice plucking’ in the plane of weakness. Half Dome,
at the head of the Yosemite Valley, has a clean ver-
tical face through the bisecting of what was once
an almost perfect dome (Figure 4.27). In periglacial
regimes, selective weathering of granite with vary-
ing joint densities leaves residual tors (Chapter 14).
All igneous rocks are strong, and the smaller
intrusions tend to create their own distinctive Figure 4.27 Half Dome, Yosemite, California, with
landforms when exhumed during denudation. the vertical face formed on a single major joint cut into
Selective removal of softer and weaker country the rounded dome defined by exfoliation sheet jointing.
GFE-04.qxd 6/24/05 2:54 PM Page 105
TECTONICS 105
Figure 4.28 An exhumed dyke forms a wall of rock 10–30 m high across the plains of New Mexico. It is a radial
feature from Ship Rock, itself a product of differential erosion around a sub-volcano plug that now stands 500 m
high (and is 5 km away in this view).
Table 4.3 Typical values for safe bearing pressures and safe slope angles on
different rock types.
Lithology, rock type Safe bearing pressure (SBP) Safe cut slope
(MPa) (degrees)
Strong igneous rocks 4–12 80–90

Strong metamorphic rocks 3–10 60–90
Stronger limestones 2–4 80–90
Weaker limestones 0.8–3 70–90
Chalk 0.5–1 45–80
Stronger sandstones 2–4 70–90
Weaker sandstones 0.7–2 50–70
Mudstone 0.7–2 40–70
Shale 0.4–1.5 25–50
Heavily fractured rock 0.3–1 20–40
Weathered rock and soil 0.01–0.4 15–25
(UK), forming inland cliffs (High Cup Nick), lites, but folding and erosion have obliterated the
coastal headlands (Bamburgh) and waterfalls original volcanic landforms.
(High Force), and it was utilised by Roman engi-
neers to enhance the impregnability of Hadrian’s Engineering implications
Wall, which was built largely on its escarpment. Rock type is the prime factor in determining ground
Youthful volcanic rocks create their own posi- conditions for engineering. Though fractures are crit-
tive landforms and environments (Chapter 23). ical to rock mass strengths (see above), most rocks
Older volcanic rocks are generally strong and also have characteristic styles of fracturing that allow
form mountain terrains. England’s and Wales’s broad concepts of engineering conditions to relate to
highest mountains, in the Lake District and the rock type. Table 4.3 shows the ranges of bearing
Snowdonia, are both formed of Ordovician rhyo- capacities and also stable angles for cut slopes for
GFE-04.qxd 6/24/05 2:54 PM Page 106
a variety of rocks. These figures are entirely based on nature of its ground conditions, the evolution of
empirical data, and any more accurate assessment of its landforms and the potential for the larger
ground conditions must be based on a specific inves- geohazards including earthquakes, major land-
tigation of the local rock structure. slides and regional subsidence.
Sedimentary rocks provide the greatest range of
difficult ground conditions. Clay is the only wide-
spread rock whose strength is so low that settle- References
ment under imposed load can be the limiting factor
Ambraseys, N. N. (1988) Engineering seismology.
in engineering design. The potential compaction of International Journal of Earthquake Engineering
clay is a function of its age, consolidation history, and Structural Dynamics 17, 1–106.
clay mineral content, silt or sand content, thickness Barton, N., Lien, R. and Lunde, J. (1974) Engineering
and drainage state. Pre-Miocene, over-consoli- classification of rock masses for tunnel design.
Rock Mechanics 6, 189–236.
dated, illite-rich, silty clays (that geologists would
Bieniawski, Z. T. (1973) Engineering classification of
define as a rock) cause minimal settlement, while jointed rock masses. Transactions of the South
Holocene, superficial, smectite-rich, silt-free clays African Institute of Civil Engineers 15, 335–343.
(that anyone would define as soil) can cause severe Bolt, B. A. (1999) Earthquakes. W. H. Freeman,
ground movements. Most of the science of soil New York.
Burbank, D. and Anderson, R. (2000) Tectonic
mechanics (Craig, 1997) is devoted to the engi-
Geomorphology. Blackwell Science: Oxford.
neering behaviour of clays. Peat is the only natural Craig, R. F. (1997) Soil Mechanics. Spon, London.
material with greater potential compaction than Fookes, P. G. (1997) Geology for engineers: the geo-
clay (Chapters 7 and 12). Naturally cavernous logical model, prediction and performance. Quart.
ground is almost restricted to the soluble lime- Journ. Eng. Geol. 30, 293–431.
Hutchinson, J. N. (2001) Reading the ground: mor-
stones, gypsum and salt (Chapters 12 and 19). The
phology and geology in site appraisal. Quart. Journ.
sandstones provide the better ground with fewer Eng. Geol. Hydrogeol. 34, 7–50.
problems for the engineer. Milliman, J. D. and Syvitski, J. P. M. (1992)
The intrusive igneous rocks are generally Geomorphic/tectonic control of sediment discharge
strong, though instabilities within them may to the ocean: the importance of small mountainous
rivers. Journal of Geology 100, 525–544.
derive from their jointing, notably exfoliation in
Watson, R. A. and Wright, H. E. (1969) The Saidmarreh
granites and columnar in dolerites. Perhaps the landslide, Iran. Geological Society of America
major influence on their fractures is with respect Special Paper 123, 115–139.
to their extraction and utilisation. Granite is typi-
cally favoured as a dimension stone by its widely
spaced fractures, while the higher joint densities Further reading
of most dolerites reduce quarrying costs in the
supply of hard aggregate. Burbank, D. and Anderson, R. (2000) Tectonic
Geomorphology. Blackwell Science, Oxford.
Hills, E. S. (1963) Elements of Structural Geology.
Wiley, New York.
4.4 Summary Kearney, P. and Vine, F. J. (1996) Global Tectonics.
Blackwell Science, Oxford.
The internal structures of rock masses are criti- Ollier, C. and Pain, C. (2000) The Origins of Mountains.
Routledge, London.
cal to the evolution of terrain morphology and Prentice, J. E. (1990) Geology of Construction Materials.
to all aspects of engineering geomorphology. Chapman & Hall, New York.
On the small scale, fractures determine the Ramsay, J. G. and Huber, M. I. (1967) The Techniques
strength of rock masses, with implications on of Modern Structural Geology, Volume 2: Folds and
their erosion resistance, slope stability and Fractures. Academic Press, New York.
Rikitake, T. (ed.) (1981) Current Research in Earthquake
bearing capacity for structures. On the larger Prediction. Reidel, Dortrecht.
scale, the tectonic history and the larger inherited Twiss, R. J. and Moores, E. M. (1992) Structural
structures of a terrain determine the overall Geology. W. H. Freeman, New York.
GFE-05.qxd 6/27/05 4:54 PM Page 107
5. Stratigraphy
Colin J. R. Braithwaite
5.1 Introduction and basic principles ‘law’ of superposition, and its application was
particularly important as it led to the development
Although the term ‘stratigraphy’ suggests simply of a relative time scale of events in the evolution
logging (recording) sequences of rock strata, of the Earth.
stratigraphy is more commonly seen as the study of James Hutton (1726–1797) in Edinburgh,
Earth History as interpreted from these sequences. realised that the sequences of layers were not
The methods used to determine the conditions of continuous. At Siccar Point on the Berwickshire
formation of a particular sedimentary unit are coast he saw steeply inclined (Silurian) turbidite
outlined in Chapter 3, but the three-dimensional layers overlain by gently dipping (Devonian)
ordering of these to create a conceptual (and fluvial sandstones and conglomerates. He recog-
predictive) model of the environmental system nised that the contact between the two sequences
depends on the exercise of stratigraphic methods. It represented a break in the stratigraphic record.
is the predictive capacity of such models that is of He referred to such breaks as unconformities
particular value to the engineer. and noted that they seemed to demand long peri-
Nicolaus Steno, in 1667, was the first to out- ods of time for their formation, time in which
line the basic principles relating to the ordering of folding, uplift and erosion could take place
sequences of rocks, which he stated in three before deposition was resumed. Subsequently
general laws pertaining to: others, principally in the Geological Survey in
Britain, used unconformities to subdivide the
stratigraphical sequence into units that were
1. original horizontality, which assumes that
identified as geological time periods: Cambrian,
regardless of their present attitude layered sed-
Ordovician, Silurian, Devonian and so forth.
imentary rocks were originally deposited as
(Table 5.1). These were generally named after
horizontal beds or strata
the areas in which the particular sequences were
2. lateral continuity of layers
first recognised (Table 5.2). The names defined
3. the law of superposition, which determines
relative ages and, for any area, provided and pro-
that layers can only be formed in sequence.
vide a guide to the relative induration of the
Inherent in the idea of wide distribution of
rock. Rocks of greater age are more likely to
similar layers of rocks is the realisation that in
have suffered greater burial and greater diage-
normally layered sequences, older beds are
netic or metamorphic change and are therefore
overlain by younger, on the assumption that
likely to be stronger. However, there are varia-
new layers cannot be deposited underneath
tions between different rock types of the same
existing deposits.
age and, on a global scale, the history of
particular areas may result in relatively young
These were important observations, predating sequences being significantly altered or rela-
by almost a hundred years acceptance of the fact tively old sequences remaining quite fresh rela-
that most of the rocks concerned were sedi- tive to the experience of rocks of the same age
mentary sequences laid down in the oceans. The elsewhere.
107
GFE-05.qxd 6/27/05 4:54 PM Page 108
Table 5.1 The stratigraphical column, showing the divisions of geological time and the relative ages of major events.
EON ERA Period Age Duration Orogenic Major Events
(Ma) (Ma) Phases
Quaternary Major glaciations of

Cenozoic (Pleistocene) 1.8 northern hemisphere
(Recent 1.8 Pyrenean First Hominids
life)
Tertiary 63.1 Age of mammals, birds
and flowering plants
65 Alpine- Major meteorite impact,
Laramide extinction of the Dinosaurs
Cretaceous 64.7
Indian and southern oceans
142 open
Mesozoic First birds, modern bony
(Middle Jurassic 63.8 fishes, rudist bivalves
life) Nevadan and flowering plants
205.7 Opening of North Atlantic
Phanerozoic (Evident life)
Ocean
Triassic 42.5 First Dinosaus
Hercynian- First Mammals
248.2 Appalachian Break-up of Gondwanaland
Permian 41.8 Mass extinction of rugose

corals, trilobites and many
290 others
Glaciation in the southern
Carboniferous 64 hemisphere
Bretonian- First reptiles
354 Acadian Last graptolites
Palaeozoic First insect and amphibians
(Ancient Devonian 63 First land-living animals
life) and plants
417
Silurian 26 First fish with jaws
443
Ordovician 52 First vertebrates (Jawless
495 fish
Cambrian 50 First graptolites
545 Cadomian First skeletal organisms
Proterozoic First soft-bodied animals
Precambrian Huronian ⬃590Ma forming tracks
(All rocks 2500 and trails
Archaean older Laurentian Increasing atmosphic O2 at
than 4000 1.7 Ga
Palaeozoic) Earliest bacteria 3.5 Ga
Priscoan 4600 Major cratering on the
Moon 4.2 Ga
GFE-05.qxd 6/27/05 4:54 PM Page 109
STRATIGRAPHY 109
Table 5.2 Derivations of names of the principle units in France was at this time arguing that the changes
of geological time. that could be seen in fossil faunas in the Paris
basin were a reflection of repeated catastrophes,
Cenozoic ⫽ recent life
Pleistocene ⫽ most newly formed each followed by a new creation. It was nearly
Tertiary ⫽ the third Era a hundred years before Charles Darwin (1809–
Neogene ⫽ later formed 1882) was able to demonstrate that they changed
Palaeogene ⫽ earlier formed in response to evolution. It is to William Smith’s
Mesozoic ⫽ middle life particular credit that he made careful observations
Cretaceous ⫽ Greek (Creta) for Chalk and used them to construct a conceptual frame-
Jurassic ⫽ from the French Jura Mountains work that allowed him to make predictions that
Triassic ⫽ from the three divisions in Germany were of significance to his work of construction
Palaeozoic ⫽ ancient life long before the existence of a supporting frame-
Permian ⫽ Permia, ancient kingdom of the Urals
work of evolutionary theory.
Carboniferous ⫽ coal-bearing
Pennsylvanian
(In North America)
Mississippian
Devonian ⫽ Devon 5.3 Bedding and time
Silurian ⫽ area of the Silures, an ancient Welsh tribe
Ordovician ⫽ area of the Ordovices, an ancient The layering (bedding) in a rock sequence is
Welsh tribe defined by bedding planes. These form as a result
Cambrian ⫽ Cambria, ancient name for Wales of pauses in deposition or intervals of erosion
Proterozoic ⫽ before life interrupting the original succession (Chapter 3),
Archaean ⫽ first or beginning and rock sequences rarely provide a complete
Priscoan ⫽ accretion of the Earth and Moon record of the time that has elapsed. The early
stratigraphers, following Steno’s principles,
measured individual rock units and sequences and
5.2 Correlation used the values obtained to calculate their relative
ages on the basis of their maximum thickness and
William Smith (1769–1839) a civil engineer and assumed rates of deposition. As no reliable esti-
surveyor, who worked on the construction of mates of rates of deposition were available there
roads and canals in England, is regarded as the were wide variations in the results, some of which
father of stratigraphic mapping. Using Steno’s suggested total ages for the Earth of only a few
principles he recognised that the same sequences million years. Even today there are considerable
of sedimentary rocks, containing the same fossils disparities in estimates of rates of both denuda-
and with the same physical characteristics, could tion and accumulation — commonly used to
found in different parts of Britain, and that their assess denudation (Chapter 3, Table 3.2) and in
distribution could be mapped to provide a predic- recent years our view of rates of accumulation has
tive framework for excavation, saving him time changed. We now believe that most of the time
and his clients money. His maps are recognised as represented by a rock or sediment sequence is
among the first of their kind in the world and reflected in the bedding planes (the divisions
stand comparison with maps of the same areas between the beds) whereas the beds themselves
produced by present-day geologists. At the core of represent events that took place over relatively
Smith’s observations was the realisation that he short periods of time that in some cases may
could use fossils to differentiate rocks that were be measured in hours or days. Derek Ager
otherwise similar on the basis of their relative age. (1923–1993) succinctly described this realisation
Such correlation was an objective exercise made in terms of his experience of National (military)
simply by recognising the fossils concerned, but it Service, ‘long periods of boredom punctuated by
formed the basis for what became the ‘Law of moments of blind panic’. In sedimentological
Faunal Succession’. Georges Cuvier (1769–1832) terms (Chapter 3) all depositional environments,
GFE-05.qxd 6/27/05 4:54 PM Page 110
apart perhaps from the pelagic realm of the deep 1896 and Ernest Rutherford’s (1906) attempts to
oceans, are places where the effects of commonly use this to measure the ages of minerals, in 1911
catastrophic events (storms or tidal surges, flood- Arthur Holmes, then working with R. J. Strutt
ing of rivers, volcanic eruptions, or extreme des- (who was the first to show that the Earth is not
iccation) have a greater potential for effect and for actually cooling as a result of heat loss), set out
preservation than those resulting from ‘normal’ the principles of radiometric dating. As methods
day-to-day processes. For this reason, although of analysis have improved, these have become
the time represented by an entire rock succession correspondingly sophisticated. Although correla-
may total millions of years, that recorded by indi- tion using fossils remains widespread, and is
vidual beds may account for only a tiny fraction cheap and practical, the age ranges of the various
of this. organisms used can now be established with some
It is important to note that although the general certainty, using radiometric dating of inter-
concept of accumulating sedimentary layers can bedded lavas or ashes, or minerals such as glau-
be applied to the entire geological record (accepting conite, phosphorite or carbonates that formed
the occasional intrusion or extrusion of igneous authigenically within the sediments. Table 5.1
rocks) it is not easily applied to Quaternary also includes the main events recorded in the rock
deposits on which most engineering works are record during the history of the Earth and the
founded. In these, as will be shown, sedimentary radiometric ages currently assigned to them.
units are seldom laterally continuous and the The rock units defined may be mapped on the
sequences of layers are most commonly charac- basis of their gross lithology or facies, represent-
terised by discontinuities. ing the sum of all the observable characteristics,
Estimates of the passage of time represented of rock type, geometry, sedimentary structures
by rock sequences and of the age of the Earth and fossil content. They may be correlated on the
have become reasonably accurate and reliable, but basis of containing similar fossil assemblages.
have not always been so. Much traditional stratig-
raphy, predicated on the concept of evolution has Radiometric dating
made little impact on this area, although Charles Where age dates are required it is important to
Lyell (1797–1875) attempted to calculate geolog- select material that contains an isotope or isotopes
ical times based on the supposed rate of change with an appropriate half-life. The half-lives deter-
of molluscan faunas. His figure of 240 Ma (mil- mine the useful range of the isotopes commonly
lion years) elapsed since the beginning of the used in dating Quaternary deposits.
Ordovician was plausible but incorrect, and even Potassium– 40
K→40Ar: 110 Ma (igneous and
now estimates of absolute age resting solely on argon metamorphic
supposed evolutionary rates would not be consid- rocks)
ered reliable. Uranium 234
U→230Th: 248 ka (corals,
For reasons indicated above, estimates based series speleothems
on accumulation rates fared no better. Some pro- and bone)
duced figures in excess of 1500 Ma others were as Radio- 14
C→14N: 5730 a (wood,
low as 3 Ma. At the turn of the 19th century Lord carbon charcoal,
Kelvin (William Thomson, 1824–1907) calcu- shells and
lated the supposed heat loss from a cooling Earth coral)
and concluded that only 20–40 Ma were neces- Lead 210
Pb→206Pb: 22.2 a (siliciclastic
sary for it to have reached its present temperature. sediments).
Kelvin was an influential figure and because his
estimates were based on objective mathematical However, this is only part of the story. There
calculations his conclusions dominated much has to be some degree of certainty that the pro-
geological thought. However, following the dis- portions of the original element and its isotope are
covery of radioactivity by Henri Becquerel in indeed a reflection of the time elapsed. If the rock
GFE-05.qxd 6/27/05 4:54 PM Page 111
STRATIGRAPHY 111
is heated, for example, elements may become concepts of facies relationships were founded and
mobile and the ‘clock’ reset, with the decay process underpins all palaeoenvironmental interpretations
restarting as the mineral cools below a critical of facies descriptions. Translated this states:
temperature, the blocking temperature, at which
diffusion ceases and the mineral grain becomes a
closed system. Different minerals have different The various deposits of the same facies area and,
similarly, the sum of the rocks of different facies
blocking temperatures that are independent of the
areas, were formed beside each other in space, but
isotopes they contain. in crustal profile we see them lying on top of each
other . . . It is a basic statement of far reaching
significance that only those facies and facies areas
5.4 Three kinds of stratigraphy can be superimposed, without a break, that can be
observed beside each other at the present time.
Lithostratigraphy Translation from Blatt et al. (1980).
Lithostratigraphy involves the recognition of
mappable rock bodies. Whereas lithology defines This observation, illustrated in Figure 5.1, is at
an individual rock type such as sandstone or lime- once very simple but profound and provides the
stone, lithostratigraphy refers to a facies. This geologist and the engineer with a valuable tool,
represents a body of rock that may include several predicting rock relationships in three dimensions.
lithologies but is identifiable on the basis of the It can be demonstrated by considering a common
association of lithology, texture, fossils and sedi- coastal environment — a tidal flat. Tidal flats typ-
mentary structures, together with the overall ically show lateral zones of sediment types (each
geometry of the sediment body, the shape, distri- with characteristic faunas and floras). A borehole
bution and thickness of beds. sunk at locality D would encounter the vertically
Defined in this way the facies reflects the sum changing sequence of sediments shown in Figure
of all the processes operating when the sediment 5.1 that can be related to the horizontal changes in
was deposited in a particular environment at a sediment type. The vertical sequence of superim-
particular place and time. It might represent, for posed sediment types occurs because the tidal
example, the products of part of a delta (Chapter flats are prograding, that is, building seawards.
20), a barrier beach, or a turbidite sequence. Moving back through time, there must have been
Facies form lithostratigraphic units that may vary a period when the zone now below the saltmarsh
in both time and space. A beach, for example, will above mean high tide level was below low mean
respond to follow a rising or falling sea level and tide level and covered by the sea. As the sedi-
change its position on a sloping shoreline. The ments prograded, the saltmarsh grew over the
direction in which it moves will depend on the muds of the former high tidal flats and these in
supply of sediment relative to the rate of sea level turn extended over mid-flat sediments. However,
change. Thus, the beach deposits will retain an not all rock sequences record a history of contin-
internal continuity but are to be found in different uous deposition. Where erosion has occurred,
places at different times. A unit that behaves in some sediment must be missing. If the missing
this way is described as diachronous (literally deposits represent a geologically long period of
through time), because, when viewed in three time so that the deposits of an entire environment
dimensions, it crosses time-defined surfaces. have disappeared, then Walther’s law no longer
Most facies cross time planes to some extent. applies.
It might appear from this that facies could vary Detailed description provides the starting point
at random but in fact they maintain a precise spatial for any facies study, irrespective of the objective or
relationship. This was first appreciated by Johannes the time available. Harms et al. (1982) suggest that
Walther (1860–1937), who in 1893–4 set out the once lithologies have been identified the next stage
‘Law of Correlation’, now more commonly known is the recognition of a series of ‘building blocks’
as Walther’s Law. This is the key idea on which the that simplify and summarise descriptive data. Such
GFE-05.qxd 6/27/05 4:54 PM Page 112
Salt Marsh
Mean high tide level

High-tidal flat (mud)
Mid-tidal flat (interbedded mud and sand)

D
C
Low-tidal flat (sand)
B
Mean low tide level A
A B C D
Unit of time
Lithological boundary
Figure 5.1 Illustration of Walther’s Law. Cross-section through the sediments of a tidal flat and boreholes A–D.
Note the vertical stacking (seen in cross-section) of the various lithologies as a result of progradation. The same
sequence is seen in the vertical succession at D (representing an extended period of time) as may be seen on the
present surface moving from A to D. Note that the horizontal lithological boundaries cross-cut the inclined time
planes as the slope progrades seawards. Thus, the same lithology occurs at different time intervals in successive
boreholes and is therefore diachronous.
blocks are defined as ‘homogeneous bodies of rock interpretation may not be correct! An incorrect
that differ from vertically and laterally adjacent interpretation can lead to false assumptions con-
bodies of rock by their physical, biological or chem- cerning the characteristics and distribution of the
ical characteristics’ that is, they represent a facies. deposit and the engineer should be wary of alleged
Although many facies sequences are subjectively descriptions offered in these terms.
defined, there are methods of producing statistical
summaries of relationships within a succession. Biostratigraphy
These methods are described in reviews by Harms Correlation by fossils offers a more reliable way
et al. (1982), Reading (1986) and Walker (1984). of obtaining the age equivalence of rocks than
Regrettably, whereas the term facies originated their lithological appearance, and fossils may be
with a holistic view of its components there are now used to define palaeontological ‘zones’ (chrons or
three extensions in its use. Lithofacies describe the units of time). The most useful ‘zonal fossils’ are
physical characteristics of a unit, lithology, bedding those that are abundant, easily recognisable, have
thickness, sedimentary structures and the like. the widest geographical distribution and, most
These are the characters of most value to the engi- importantly, have evolved rapidly so that a partic-
neer and are relatively easily mapped. Biofacies ular form lived for only a short geological time.
describe the characteristic fossil associations and However, organisms are environmentally selec-
are useful in refining the age and nature of the envi- tive and there are considerable difficulties in cor-
ronment but of less relevance to ground conditions. relating between contrasting environments, the
Neither of these is likely to cause any confusion so most dramatic contrast being between marine and
long as their point of reference is stated. However, non-marine systems. Few organisms are able to
the term facies is also used subjectively in the sense leave their remains in both. Nevertheless, rela-
of deltaic facies or turbidite facies. Such usage is tively good global correlations are available for
only acceptable in interpretation and should be most Phanerozoic sequences (Table 5.1) and define
avoided as a substitute for description because the a biostratigraphy. A zone may sometimes be
GFE-05.qxd 6/27/05 4:54 PM Page 113
STRATIGRAPHY 113
defined by the appearance and disappearance of a that can be seen with a reasonably high-powered
single species, an ‘index fossil’. More commonly, microscope. Apatite and zircon are the minerals
however, the distribution of fossil assemblages commonly affected in igneous rocks and the
provides a more sensitive subdivision of time. length and density of tracks (⬃10 ␮m) are meas-
ures of the time elapsed since the mineral formed.
Chronostratigraphy Tracks are annealed by heating, the time meas-
The chronostratigraphical subdivision of the rock ured is that since the last such heating event. For
record depends on the fact that some minerals apatite this is ⬃100 ⬚ C and for zircon ⬃250 ⬚ C.
contain naturally occurring elements that have The time since these relatively low temperatures
radioactive isotopes that decay to known daughter were achieved provides a useful record of the
products over specified time periods. cooling and erosion (unroofing) of igneous or
Ages can be obtained from minerals formed metamorphic bodies.
within rocks at the time of deposition, for example Luminescence dating depends on the fact that
crystals grown within cooling magmas, or clay any material that contains Uranium, Thorium or
minerals such as glauconite formed in marine sed- Potassium is continuously bombarded with ␣, ␤
iment. Individual volcanic lavas inter-bedded with and ␥ particles. The effect of this radiation is to
fossiliferous sedimentary sequences provide refer- produce local ionisation and this leads to trapping
ence time points and have allowed the construc- of metastable electrons within the atomic structure.
tion of an absolute global timescale (Table 5.1). These can be freed by heating. This causes a char-
In younger rocks and sediments, which are most acteristic emission of light (Thermoluminescence
commonly those of concern to the engineer, other or TL) from the mineral. It is quite separate from
methods are available. In the Pleistocene lime- the light emitted when any material is heated and
stones common on many tropical shores, 234U can only be measured once.
decays to 230Th with a half-life of 248 ka (thou- Optically stimulated luminescence (OSL) meas-
sands of years), while in more recent deposits, the ures the light emitted from the most light-sensitive
proportion of 14C provides an accurate measure up electron traps in minerals such a quartz and
to about 40 ka. Radiocarbon is particularly valuable feldspars. In this case trapping is a result of move-
in carbon-based materials such as wood that may ments of electrons in response to solar radiation. It
be found in young deposits formed on or close to is particularly valuable for determining how long
land, but is also present in marine carbonates and in sediment or rock surfaces have been exposed on the
cave deposits such as stalactites (speleothems). surface but is most effective from about 1–150 ka.
Some success has also been achieved using the Electron spin resonance (ESR) again measures
decay rate of amino acids in skeletal carbonate damage induced by ␣, ␤ and ␥ radiation from
grains. Indirect methods such as tree ring dating naturally occurring isotopes. Free electrons accu-
(dendrochronology) locally provide a very accumulate within the atomic structure of the mineral
rate recent time scale extending over some 2000 and high frequency electromagnetic radiation
years, while methods such as optically stimulated excites these to the point where their resonance can
luminescence (OSL), fission track analysis and be detected. The numbers of electrons is a direct
cosmogenic dating provide means of estimating the reflection of the age of the material. The method
age of eroded surfaces and rates of erosion. can be applied to teeth, mollusc shells and corals as
well as speleothem deposits and deep-sea cores.
Ages obtained range from 0–2000 ka.
5.5 Non-radiometric chronology The increased use of mass-spectrometers has
made cosmogenic dating more practical. 36Cl,
26
Fission-track dating is based on the principle that Al and 10Be are produced on exposed rock sur-
spontaneous fission occurs in minerals bearing faces as a result of damage caused by cosmic rays
238
U and that the energy released produces visible and can be used to estimate erosion rates and
damage (fission tracks) in surrounding minerals exposure histories.
GFE-05.qxd 6/27/05 4:54 PM Page 114
Magnetostratigraphy depends on the response timing of major offlap and onlap events has com-
of rocks and sediments to the Earth’s magnetic monly been the same worldwide, has led to the
field at the time when they were formed. Both construction of a global sea level curve (Figure 5.4)
small-scale secular variations and long-term field and the recognition of long-term cycles of sea level
reversals can be used to provide a time scale. change.
However, this is not absolute and the pattern of
change must be matched to sequences of events in
order to arrive at a time correlation. 5.7 The geological time scale
The earth formed about 4.6 ⫻ 109 years (Ga) ago

5.6 Seismic stratigraphy and meteorites found here, and the oldest rocks
recovered from the moon, have a similar age. The
Seismic reflection profiling, used universally oldest terrestrial rocks that have been discovered
in the oil industry to support exploratory drilling so far are continental crustal gneisses in West
programmes, has given birth to the technique Greenland that are between 3.7 and 3.8 Ga old.
of seismic stratigraphy, formalising the division Areas of very ancient rocks (⬎ 2.5 Ga) form the
of sedimentary successions on the basis of Archean Shields or cratons within continents
discontinuities. Seismic reflections are produced (Figure 5.5). Of particular importance in these is
by shock waves generated at the surface of the evidence of the first emergence of life on our
earth by explosions or mechanical vibrations. planet. The oldest fossils known are microscopic
These travel downwards and are reflected by spheres and filaments found in cherts from
boundaries such as bedding planes and unconfor- Swaziland that are around 3.5 Ga old. Over many
mities within the rock sequence where there are millions of years the photosynthetic activities of
relatively abrupt changes in rock density. Seismic primitive cyanobacteria and algae eventually led
reflections are typically only capable of resolving to the evolution of our oxygen-rich atmosphere.
units of a few metres thickness, but higher fre- The build-up of oxygen over 2.5 Ga years induced
quency investigations of shallow depths (for a massive precipitation of iron oxide as banded
example the seabed) may resolve layering of a iron formations, exploited as iron ores in Canada,
few centimetres and are increasingly used in engi- Australia and South America.
neering investigations. Discontinuities commonly
correspond with time-stratigraphic boundaries, Upper boundary
that is, interruptions in the sedimentary sequence
that represent pauses in deposition that occurred
during some specific interval of time.
Major conformable depositional packages Erosional truncation Toplap
(sequences) are often bounded above and below Lower boundary
by unconformities. The lower surfaces of these
packages may show relationships that are des-
cribed as onlap, downlap or concordant, whereas
upper surfaces may be truncational, or show toplap Onlap Downlap
or concordance (Figure 5.2). By correlating
sequences on a chronostratigraphic chart (Figure Baselap
5.3), where unconformities show up as major time Concordance
gaps, relative changes of ancient sea levels in a

region may be inferred. In the simplest case a rela-
tive rise in sea level is indicated by coastal onlap,
the magnitude of which can be used to infer the Figure 5.2 Stratal relationships recognised in seismic
amount of sea level change. The realisation that the stratigraphy (after Vail et al., 1977).
GFE-05.qxd 6/27/05 4:54 PM Page 115
STRATIGRAPHY 115
E
D 3
C
9
17 B 13
Coastal deposits 18
Marine deposits 21 A
500 25
Initial shelf edge
Depth (m)
Coastal aggradation
14 ⫽ age (Ma)
Downward shift
BCD... supersequence
0
0 25 Stratigraphic cross section
(km)
0 E
Geological time (Ma)
5
D
10 C
15 Hiatus B
20 A
25
Chronostratigraphic chart
500 400 300 200 100 0 ⫺100 Cycle Super-

cycle
0
E
Stillstand Coastal toplap
5 D
Geological time (Ma)
BCD
10 C
450
15 50 Fall B
Shelf edge
100
Base A
20 150
Ris A
e 100
25
Highstand Lowstand
Rise Relative change of sea level Fall
Regional chart of cycles of

relative changes of sea level
Figure 5.3 Correlation of stratigraphic sequence relationships, chronostratigraphic charts and charts of cycles of
relative sea level change (after Vail et al., 1977).
Although traces of soft-bodied animals are (Ma) ago (Table 5.1). The three eras of the
known from a few rocks around 1 Ga years old, Phanerozoic epoch (Table 5.1) are characterised
the first widespread and well-preserved shelled by the dominance of particular groups of fossils
fossils that can be used in dating and correlation and the boundaries between them correspond
only occur globally at about 600 ⫻ 106 years with major phases of worldwide extinction. They
GFE-05.qxd 6/27/05 4:54 PM Page 116
sections derived from this, (Figure 5.7) may be used

Period Relative change in sea level Time to generate a palaegeographical map (Figure 5.8),
1.0 0.5 0 (Ma) illustrating the regional depositional environment
0 during a specified time interval and the form of the
Tertiary
original sedimentary basin. In the hydrocarbon
Present
industry such models are used extensively to pro-
Cretaceous sea level 100 vide a predictive framework for exploration, com-
bining facies modelling with a structural geological
Jurassic appraisal of a region, but the same techniques are
200 also of value in large-scale hydrological modelling
Triassic of aquifers.
Permian
On a global scale, regional stratigraphical
300 reconstructions provide the basis for plate-tectonic
Carboniferous models. For example, subduction (Chapter 4)
adjacent to a trench margin results in under-
Devonian thrusting of successive units of sea-floor sedi-
400
ments as oceanic crust slips beneath the adjacent
Silurian
continental margin. As a result, a prism or wedge
Ordovician of sediment is accreted in which individual thrust
500 slices become progressively younger in the direc-
Cambrian
tion of the original oceanic abyssal plain. Thus, in
Precambrian Figure 5.9, slice A is younger than slice B. Within
individual slices, however, the beds form a nor-
mally stacked sequence and are younger from
Figure 5.4 Global time scale and changes in sea level bottom to top. Detailed mapping and strati-
(identified by Vail et al., 1977).
graphic appraisal of the terrain forming the
Southern Uplands of Scotland has revealed an
also broadly coincide with widespread mountain- assemblage of rocks of Silurian age that appar-
building and folding episodes (Table 5.1). ently shows these characteristics (Anderton et al.,
1979). However, no matter how elegant they may
appear to be, the reliability of such models ulti-
5.8 Correlating and interpreting mately depends on the accuracy of stratigraphic
sequences correlation.
Some rock sequences appear to be monoto-
It is impossible here to review the stratig- nous and are consequently difficult to subdivide.
raphic evolution of the world but the principles of However, a detailed examination can sometimes
correlation and interpretion of rock sequences can provide subdivisions in a variety of ways:
be demonstrated. The succession of mid-Jurassic
rocks in southern England consists of limestones, 1. macro-palaeontological
claystones and rarely, sandstones. The relative ages 2. micro-palaeontological
of these rocks are indicated by their fossil contents 3. mineralogical
and world correlation. Their interrelationships have 4. lithological.
been unravelled by methods that include mapping,
logging outcrops and boreholes, and seismic stratig- Macro-palaeontological zonation involves the
raphy. Correlation allows a thickness (isopach) map detailed collection and identification of macro-
to be constructed (Figure 5.6) that defines the form fossils and their comparison with known standard
and limits of accumulation of particular lithologies specimens. Where no macrofossils are available,
and time-stratigraphic units. Interpretative cross- either because the depositional environment was
GFE-05.qxd 6/27/05 4:54 PM Page 117
STRATIGRAPHY 117
Archaean cratons Late Proterozoic to early Mesozoic and Tertiary

> 2500 Ma Palaeozoic orogenic belts orogenic belts < 200 Ma
900–400 Ma
Early to middle
Proterozoic cratons Late Palaeozoic orogenic
and orogenic belts belts 400–200 Ma
2500–900 Ma
Figure 5.5 Generalised geological map of the world showing major tectonic provinces (after Anderton et al., 1979;
see also Chapters 4 and 9).
Edge of Faringdon London Platform

outcrop Southall
45 Sheerness
30
15 0
Westbury Strat A-1
30 Humbly Grove 30
15 Folkestone
Winchester Ashdown
0 Southampton
60 Brighton Hastings
Winterbourne Kingston 75
40 km
30 45 Isopachs (metres)
Kimmeridge 15
Line of section
Dorset Basin 0
Figure 5.6 Isopach (equal thickness) map of the Great Oolite Limestone Formation (Middle Jurassic) in south-east
England (after Sellwood et al., 1985).
unfavourable for their life or because burial con- microfossils may provide an alternative means of
ditions were unsuitable for their preservation, or zonation. The methods of identification, compar-
simply because an uncored borehole sequence ison and correlation are essentially the same as
provides only fragments in drill cuttings, then those used for macrofossils but as the fossils
GFE-05.qxd 6/27/05 4:54 PM Page 118
Winterbourne Southampton Winchester Humbly Strat-1 London

Kingston Grove Platform
56 km 20 km 26 km 20 km 30 km
Cornbrash Herriard member

100 m
Forest Marble
Hesters Copse Formation
Humbly Grove Member
Frome Clay Sandstone Unit
Passage Beds Dolomitic Unit
Hoddington member
Wackestone Unit
Fullers Earth
Figure 5.7 Lithostratigraphic correlation in the middle Jurassic Great Oolite Formation of south-east England
showing facies variation. The line of section is shown in Figure 5.5 (after Sellwood et al., 1985).
themselves are so small, large numbers may be

obtained from small samples. Foraminifera, ostra-
?
codes, pollen, spores and dinoflagellates are all
? ? London-Brabant
Massif
commonly employed.
Rhenish
Massif
Mineralogical variation in a sequence can be
Cornubia cryptic and may be of two kinds. The character-
istic detrital minerals of the sediment, or indeed
the composition of these minerals, may change
abruptly as a result of a change in the geometry
Oolite shoals B
of the basin and thus the origin, provenance, of
Paralic facies the grains. Such changes may be visible over a
Pelletal muds A wide area. However, mineralogical variation may
Open sea marls also reflect post-depositional processes. The dia-
Laminated muds
genetic degradation of feldspars in sandstone
may generate a rock rich in the clay mineral
kaolinite. During burial, and with rising temper-
ature, the kaolinite is progressively replaced by
the clay illite. Thick hydrocarbon-bearing reser-
Hardground voir sandstones are sometimes subdivided on the
Bioclastic sands basis of this mineral transition that, in the North
Foreshoal oolites Sea, occurs at a temperature of 120–150⬚ C. It is
Backshoal bioclastic oolites an irreversible transformation and may be recog-
nised even when rocks formerly buried into the
Figure 5.8 Generalised palaeogeography and facies illite zone have later been uplifted to shallower
distribution of the Middle Jurassic (Bathonian) of depths. However, as it is a reflection of both
southern Britain and north-east France, Belgium and depth of burial and local geothermal gradient it
the Netherlands (after Sellwood and Sladen, 1981; does not relate to conventional stratigraph-
Purser, 1969, and Dubois and Yapaudjian, 1979). The
section below illustrates the facies variation along the ical correlatives. The presence of fibrous illite
line of section A–B (after Sellwood, 1985). has serious implications for the reservoir qual-
ity of a sandstone because the fibres of illite
GFE-05.qxd 6/27/05 4:54 PM Page 119
STRATIGRAPHY 119
Outer high
Shelf
Slope
Inner slope
Trench
Abyssal Plain Trench
B
A
Abyssal
Figure 5.9 Model of a subducting margin. Note the successive under-thrusting of slices and the parallel rotation
of folds (after Anderton et al., 1979).
restrict pore-throats in the rock and thus reduce glacial–interglacial cycles, reflected in changes
permeability. in sea level of the order of 130 m, but in addition,
Lithological subdivision relies on the develop- superimposed on these there have been around
ment of easily recognisable ‘marker beds’. These twenty minor cycles. Thus, all Quaternary succes-
may be distinctive deposits such as volcanogenic sions are punctuated by erosion surfaces (Figure
tuff bands and bentonites, glauconitic layers, 5.10). The amount of incision that these represent
caliche (calcrete) profiles in alluvial sequences, is locally of the same order as the sea-level
or diagenetic nodular and concretionary beds. changes and may be greater than 100 m. As a
Provided that it is laterally extensive, any form result, wide variations in thickness are common
of marker unit may provide a datum upon which and particular units are missing locally. The
the geometry of stratal units can be modelled. In record is at best fragmented.
the North Sea area, for example, the base of the In ancient rocks the idea of a stratotype, a type
Eocene is marked by a regionally extensive vol- succession showing a typical facies succession
canic tuff, which is easily recognised in otherwise is widely applied. However, the variability of
featureless mudstones. the Quaternary means that such a system is
unworkable. Although units of different ages must
Problems in the Quaternary appear in chronological order, their local charac-
Quaternary sediments and rocks (Chapter 6), teristics may differ dramatically, reflecting deposi-
deposited in the last 2 million years, dominate the tion in contrasting environments. Depositional
problems encountered in engineering excavation. units may acquire names but these normally have
However, their stratigraphy and to some extent only local validity. The contrast can best be
their correlation differs in several important demonstrated by reference to high and low latitude
respects from those of older rocks. There are successions.
two reasons for this. The first is one of distribu- In northern Europe, and in high latitudes, the
tion. Typical marine deposits have remained Quaternary is generally unlithified. A thick over-
beneath the sea and are laterally extensive, but the consolidated diamict deposit, representing glacial
successions that concern us were mostly depos- debris overridden by a later ice surge, may form the
ited either sub-aerially or in the shallow marine base of a succession resting on a glacially gener-
margins where they have been subject to recent ated erosion surface. Laminated silts and clays,
changes in sea level. There have been four major that reflect accumulation in a glacially dammed
GFE-05.qxd 6/27/05 4:54 PM Page 120
Karst surface with local

Brown quartz intense dissolution
Red sandy sand
Sandy cavity-fill Halimeda sands
with terrestrial snails in channels laterite
Coral frame
Coral-bearing limestone
with areas of in situ frame
Pale quartz sands

with burrows Karst surfaces with
local palaeosols
Cross-bedded grainstones
Range from beach to aeolian
old dense limestone
Figure 5.10 Schematic section of the late Pleistocene succession of the Kenya coast illustrating the wide lateral
variation and punctuation of the succession by karst erosion surfaces (after Braithwaite, 1984).
lake may overlie it, and these may be followed in offer a convenient means of dividing successions.
turn by cross-bedded sands and gravels recording Hearty (1998) has identified as many as eight
the advance of outwash streams. These may be palaeosol-bounded sequences in succession in the
capped by unconsolidated diamicts containing Bahamas but as few as two may be present in any
sand and silt lenses that reflect meltout from a one place and they do not provide a means of cor-
later glacial advance. However, there will be con- relation. There are, however, many more methods
siderable lateral variability and it is unlikely that of absolute dating available. Uranium series
precisely the same succession will be identified dates, magneto-stratigraphy, amino-stratigraphy,
over distances of more than a few tens of metres. electron spin resonance (ESR) and stable isotope
In addition, successions elsewhere that appear stages (16O/18O ratios) defined both by reference
similar may be of different ages and not correlat- to deep ocean cores and ice cores from the Arctic
able. The time between depositional intervals is and Antarctic provide alternative means of deter-
generally too short to generate a satisfactory mining age, although diagenetic changes may
palaeontological timescale, although there render them unreliable.
have been some advances using teeth of rodents,
beetle carapaces and the jaws of chironomids Applications
(non-biting midges: Lowe and Walker, 1997) to Horswill and Horton (1976) provide an
define sequences of climatic change. The meth- illustration of the engineering application of
ods of dating such materials include radiocarbon micro-palaeontological zonation in a relatively
(for younger deposits) and pollen analysis. homogenous sequence. They investigated the
In the tropics the most common deposits are effects of cambering at the Empingham Dam
Quaternary limestones. In these, depositional site in the Gwash valley (English Midlands).
environments include coral reef related systems, Although the Upper Lias clays at the site may be
tidal sands, wind-blown deposits (aeolianites) and crudely divided using markers such as a silty
soils (Figure 5.10). Changes in sea level produce micaceous mudstone, an ammonite nodule bed, a
a rapid lithification and deep erosion, principally phosphatic nodule bed, and a pisolitic bed, the
by dissolution. Erosion surfaces and palaeosols thick intervening claystones are more closely
GFE-05.qxd 6/27/05 4:54 PM Page 121
STRATIGRAPHY 121
subdivided using assemblage zones defined published for sale. More typically the standard
by foraminifera and ostracods. This precision map is 1 : 50 000 (originally 1⬙ to the mile) and a
allowed the valleyward thinning of the sequence few special sheets at 1 : 25 000 are published for
to be more accurately defined, providing a better ‘classical’ areas of geology. For many areas two
understanding of the nature of the deformation. forms are available, referred to as solid and solid-
An illustration of the integration of geological, with-drift. ‘Solid’ maps show only the ‘solid’
macro-palaeontological and geophysical work geology and Quaternary deposits are therefore
in an engineering appraisal, is provided by omitted. The solid geology shows the distribution
Bruckshaw et al., 1961. This formed part of the of bedrock and is organised on the basis of age
site investigation for the Channel Tunnel Study but differentiates major lithostratigraphic units. It
Group and was concerned with the distribution of is important to realise that the names applied are
the Lower Chalk beneath the English Channel. often traditional ones and do not necessarily
It also provides a lesson in nomenclature. The describe the characteristics of the formations
Lower Chalk is easily characterised as a rock for- concerned. For example, the ‘Devonian Old Red
mation in engineering terms and the geological Sandstone’ may not be red and locally may
and engineering terminology closely correspond. include considerable thicknesses of conglomer-
By contrast, Dearman and Coffey (1981) attempted ate, shale and even volcanic rocks. The terminol-
to assess the engineering characteristics of the ogy applied: sandstone, limestone, shale etc. is a
late Permian Magnesian Limestone of north-east basic one and conveys nothing of the physical
England. They found a considerable diversity in attributes of the material or of its likely behaviour
terminology, if not of rock type, identified by a in engineering. The local or regional Guides pub-
variety of geologists that they properly described lished by the British Survey go some way to
as ‘bewildering’. Some of the reasons for this explaining variations and providing interpreta-
have been discussed and centre on the use of tions of the major features. However, not all of
subjective interpretative terms as descriptors. For these maps or guides are in print and it may be
engineering purposes the plethora of terms was necessary to go to libraries in order to obtain
reduced to only four main lithological types: them. The modern British Geological Survey
bedded limestone, reef limestone, concretionary offers consultancy on issues that require further
limestone and silty clay. These simple divisions clarification and in some areas will be able to
proved locally effective and were easily adapted offer advice on engineering characteristics, but
to include variously brecciated versions. However, most large companies now have their own engi-
the term ‘reef limestones’ is clearly a subjective neering geologists or earth scientists who can
interpretation and bodies of concretions are not interpret the information and draw attention to
bounded by time equivalent surfaces. It would critical issues.
therefore be unwise to attempt to apply the same In Britain so-called ‘drift’ maps, also on a
scheme elsewhere. scale of 1 : 50 000, include the distribution of
Quaternary alluvial, glacial or marine deposits.
These may be differentiated on the basis of age,
5.9 Geological maps the distribution of marine or alluvial terraces, or
different ages of diamict deposits. The areas cov-
Geological mapping forms part of the training of ered now include the continental shelf around
all geological earth scientists and geological maps Britain. Maps are also available for some areas on
are a fundamental tool in any preliminary site a scale of 1 : 25 000 that provide more detailed
investigation for engineering purposes. In Britain distributions of such deposits, their thicknesses
the Geological Survey has traditionally mapped at and engineering characteristics. In a very few
a scale of around 1 : 10 000 (originally 6⬙ to the areas specialised maps describe the distribution of
mile). Such maps are available for consultation hazards relating to flooding, subsidence, hydrology
at offices of the Survey, but relatively few are or other issues and indicate areas suitable for
GFE-05.qxd 6/27/05 4:54 PM Page 122
development or excluded for these and other sound the study of earth history through 3.8 Ga. The
geological reasons. principal applications are in the exploration for
In the rest of Europe, high quality geological hydrocarbons and minerals, but an accurate strati-
maps are available but are seldom at scales larger graphical interpretation of both bedrock and
than 1 : 50 000. Exceptions are found in Germany superficial deposits is essential to the engineer in
where excellent 1 : 25 000 maps are available site investigations for dams, tunnels and all major
for many areas and Switzerland where 30% of construction works.
the country is available at this scale. Many former Regional interpretations of strata depend on
European colonies and dependencies have limited correlation. This is best achieved regionally by
numbers of maps on scales from 1 : 50 000 or, fossils, and locally by the mapping of distinctive
more commonly, 1 : 200 000 or 1 : 500 000. Regre- marker beds such as volcanic tuffs. Age equiva-
ttably, large areas of the world have not been geo- lence in sequences dated by fossils may be corre-
logically surveyed in detail or have air photograph lated with an absolute timescale based on the
interpretations that may not provide adequate half-lives of radioactive isotopes in inter-bedded
‘ground truth’, and small-scale maps offer only igneous or sedimentary rocks. Recent develop-
the most general guidance for the engineer. ments have involved regional correlation by
The United States Geological Survey not only geophysical means, employing the technique of
publishes excellent standard geological maps, seismic stratigraphy. This has helped to refine
with and without superficial deposits, but also models of global sea level change over the past
specialist maps dealing with depth to bedrock, 600 Ma.
engineering characteristics and hydrology, flood
impacts and the distribution of landslides,
References
together with marine geology and compilations
illustrating the distributions of earthquakes, vol- Ager, D. V. (1981) The Nature of the Stratigraphical Record
canoes and even diagenetic characteristics. Useful (2nd edn). Macmillan, London.
Anderton, R., Bridges, P. H., Leeder, M. R. and Sellwood, B.
scales vary from 1 : 24 000 to 1 : 48 000 but com- (1979) A Dynamic Stratigraphy of the British Isles; a
pilations commonly cover State or larger areas. Study in Crustal Evolution. George Allen & Unwin,
The Survey’s web catalogue also includes publi- London.
Blatt, H., Midelleton, G. and Murray, R. (1980) Origin of
cations by the US Bureau of Mines and State Sedimentary Rocks (2nd edn). Prentice-Hall, Englewood
Surveys and independent organisations such as Cliffs, New Jersey, 782pp.
the influential American Association of Petroleum Braithwaite, C. J. R. (1984) Depositional history of the late
Pleistocene limestones of the Kenya coast. Journal of the
Geologists and the Society for Sedimentary Geological Society 141, 685–699.
Geology (SEPM). Bruckshaw, J. M., Goguel, J., Harding, H. J. B. and Malcor, R.
Information on the availability of maps of par- (1961) The work of the Channel Tunnel Study Group,
1958–1960. Proceedings of the Institution of Civil
ticular areas of interest can be obtained from the Engineers 18, 149–178.
Geological Surveys of the countries concerned or Dearman, W. R. and Coffey, J. R. (1981). An engineering zonal
from texts such as International Maps and Atlases Map of the Permian Limestones of NE England. Quarterly
Journal of Engineering Geology 14, 41–57.
in Print (edited by K. L. Winch). However, access Dubois, P. and Yapandjian, L. (1979) Jurassique moyen. In
to many regional catalogues is now also available Mégnien, C. (ed.) Synthèse Géologique du Bassin de
using the Internet and for the US, for example, Paris. JM1–JM4, Mem. BRGM.
Harms, J. C., Southard, J. B., Spearing, D. R. and Walker, R. G.
exhaustive lists can be available on the desk of the (1982) Structure and Sequence in Clastic Rocks. Society
enquirer within seconds. of Economic Paleontologists and Mineralogists (SEPM)
Short Course Notes 9 Tulsa, Oklahoma, 249pp.
Hearty, P. J. (1998) The geology of Eleuthera Island, Bahamas:
a Rosetta Stone of Quaternary stratigraphy and sea-level
5.10 Summary history. Quaternary Science Reviews 17, 333–355.
Horswill, P. and Horton, A. (1976) Cambering and valley
bulging in the Gwash Valley at Empingham, Rutland.
Stratigraphy is the description and interpretation Philosophical Transactions of the Royal Society, London A
of layered rock sequences and provides a basis for 283, 427–462.
GFE-05.qxd 6/27/05 4:54 PM Page 123
STRATIGRAPHY 123
Lowe, J. J. and Walker, M. J. C. (1997) Reconstructing Sellwood, B. W. and Sladen, C. P. (1981) Mesozoic and
Quaternary Environments (2nd edn). Prentice-Hall, London. Tertiary argillaceous units: distribution and compositions.
Purser, B. H. (1969) Syn-sedimentary marine lithification of Quarterly Journal of Engineering Geology 14, 263–
middle Jurassic limestones in the Paris Basin. Sedimentology 275.
12, 205–230. Vail, P. R., Mitchum, R. M. and Thompson, S. III. (1977)
Reading, H. G. (1986) Facies. In Reading, H. G. (ed.) Seismic Stratigraphy and global changes of sea level.
Sedimentary Environments and Facies (2nd edn). Blackwell, In Seismic Stratigraphy — Applications to Hydrocarbon
Oxford, 4–19, 615pp. Exploration. American Association of Petroleum Geologists,
Sellwood, B. W. (1985) Shallow-marine carbonate environ- Memoir 26, 49–212.
ments. Reading, H. G. (ed.) Sedimentary Environments and Walker, R. G. (1984) General introduction. In Walker, R. G.
Facies. Blackwell, Oxford, Chapter 10, 283–342. (ed.) Facies Models (2nd edn). Geoscience Canada
Sellwood, B. W., Scott, J., Mikkelsen, O. and Akroyd, P. (1985) Reprint Series 1. Geological Association of Canada, 1–9.
Stratigraphy and sedimentology of the Great Ooolite Winch, K. L. (ed.) (1976) International Maps and Atlases in
Group in the Humbly Grove oilfield, Hampshire. Marine Print (2nd edn). Bowker, London and New York. GeoPubs,
and Petroleum Geology 2, 44–55. 4 Glebe Crescent, Minehead. TA24 5SN UK.
GFE-06.qxd 6/27/05 4:30 PM Page 124
6. The Quaternary
Andrew Goudie
6.1 Introduction By the Pliocene, around 5 Ma ago, the degree

of cooling was such that a more temperate flora
Environmental changes can have a huge sig- was present in the North Atlantic region, and at
nificance for engineering projects. The degree, fre- 2.4 Ma ago glaciers started to develop in mid-
quency and abruptness of change, particularly over latitude areas and many of the world’s deserts
the last few millions of years (i.e. the Quaternary), came into being (Williams et al., 1998).
has been remarkable (Figure 6.1). Landscapes often
display the legacy of forms (e.g. relict landslides Cold Warm
(a)
that may be susceptible to reactivation) and materi-
0.4 deg.
Air temperature
als (e.g. sands and gravels that are possibly not

being replenished today) inherited from signifi- Year
1960 1920 1880
cantly different climatic conditions to those experi- (b)

Cold Warm
enced at present. There are some parts of the world,

1.5 deg.
termed paraglacial environments, where processes
and forms are still dominated by the effects of the 1900 1600 1300 1000
Year
glacial advances that happened over that period. (c)
Warm
Mid-latitude air temperature
The form of river systems and the nature of their

sediment stores may relate to those times. More- 10 deg.
Cold
over, as higher resolution environmental histories

are reconstructed from ocean, lake and ice cores, it 0 10 20 30
Years BP ⫻ 1033
x 10
becomes clear that in addition to the major glacials (d)
Warm
and interglacials of the Quaternary, there have been

10 deg.
a whole series of changes at timescales of the
Cold
decade, century and millennium. In terms of such

0 50 100 150
considerations as the return periods (i.e. probabil- 1033
⫻10
Global ice volume
Years BP x
(e)
Min
ity) of hazardous events, the present may be highly

atypical of the past and of the future. 5 1016 m3
Max
0 200 400 600 800

1033
⫻10
Years BP x
Varangian
6.2 The late Cainozoic

Cenozoic
Sturtian
Gnejsö
wanan
Gond-
(f)
Ordo-
vician
Late
Era
During the Tertiary, which started after the

Glacial
Cretaceous about 65 million years (Ma) ago, the
Non-glacial
world’s climate suffered the so-called Cainozoic 0 200 400 600 800 1000
climate decline (Goudie, 1992). Temperatures Years BP ⫻ 106
showed a general tendency, though not steady or

Figure 6.1 The variation in world climate over the
uninterrupted, to fall (Figure 6.2). Thus in the past billion years: (A) by decade; (B) by century; (C) by
North Atlantic region in the early Tertiary, condi- millennium; (D) by tens of millennium; (E) by hundreds
tions favoured a widespread, tropical moist forest. of millennium and (F) by era.
124
GFE-06.qxd 6/27/05 4:30 PM Page 125
THE QUATERNARY 125
Major ice sheet in Antarctica forms

Ice sheets in northern hemisphere
(a)
and mountain glaciers develop in
Cooler waters around Antarctica

Expansion of Antarctic ice sheet
Passage between Australia and

Small glaciers widespread in
‘Cenozoic decline’ begins

and formation of sea ice
Initial polar glaciation (?)

northern hemisphere
Antarctica opens
Antarctica
Early Eocene
Palaeocene
Pleistocene
Oligocene
Oligocene
Oligocene
Mid-early
Pliocene
Miocene
Miocene
Miocene
Mid-late
Eocene
Eocene
Eocene
Early
Early
Late
Late
Late
Mid
Mid
Mid
0 10 20 30 40 50 60
(b) ⫺2.0 20
Surface water temperature ( C)

⫺1.0
15
δ18O ( )
10
⫹1.0
5
⫹2.0
⫹3.0 0
281
279A
277
(c)
Warm
Palaeotemperature
10 C
Cold
0 10 20 30 40 50 60
106 years BP
Figure 6.2 The Cenozoic climate decline: (A) a generalised outline of significant events in the Cenozoic climate
decline; (B) Oxygen isotopic data and palaeotemperature indicated for planktonic foraminifera at three sub-antarctic
sites (277, 279A, and 281) and (C) temperature changes calculated from Oxygen isotope values of shells in the
North Sea (after Goudie, 1992).
In the Quaternary (which comprises the Pleis- wildly from values similar to, or slightly higher
tocene and the Holocene) the gradual and uneven than, today in interglacials to levels that were
progression towards cooler conditions, which sufficiently cold to treble the volume of ice
had characterised the earth during the Tertiary, sheets on land during the glacials (Figure 6.3).
gave way to extraordinary climatic instability Not only was the degree of change remark-
(Lowe and Walker, 1997). Temperatures oscillated able but so also, according to evidence from the
GFE-06.qxd 6/27/05 4:30 PM Page 126
150˚
180˚ 120˚
50˚ 50˚
150˚ Arctic 90˚E

Ocean
30˚N 30˚N
S
L
60˚
120˚
A
10˚ Atlantic Ocean 10˚
90˚W 60˚ 30˚ 0˚ 30˚
Figure 6.3 The possible maximum extent of glaciation in the Pleistocene in the northern hemisphere:
C ⫽ Cordilleran ice; L ⫽ Laurentide ice; S ⫽ Alpine ice (after Goudie, 1992).
sedimentary record retrieved from deep-sea approximately 150 years. This was not a finding
cores, was its frequency. In all there have been that received universal approval (see Frenzel and
about seventeen glacial–interglacial cycles in the Blundau, 1987).
last 1.6 million years. The cycles tend to be char- Ice cores, such as that from Vostok in Antarc-
acterised by a gradual build up of ice volume tica (Petit et al., 1999), provide a particularly fine
(over a period of c. 90 000 years), followed by a temporal resolution and contain large numbers of
dramatic glacial ‘termination’ in only about 8000 valuable environmental indicators (Figure 6.4).
years. Furthermore, over the three or so millions High frequency swings in isotopic and dust con-
of years during which humans have inhabited the tent and other indicators suggest that dramatic
earth, conditions such as those we experience oscillations have taken place in environmen-
today have been relatively short lived and atypi- tal conditions over quite short periods of time.
cal of the Quaternary as a whole (Wilson et al., The rapid temperature oscillations that have
2000). been identified from ice cores are known as
Dansgaard-Oeschger events. Dansgaard et al.
(1993), for example, documented no less than
6.3 Abrupt change twenty-four interstades (warmer phases) in the
last glacial period from an ice core from
One of the features of palaeoclimatic research in Greenland. However, high frequency shifts are
the past decade has been the realisation of just also known from the last interglacial (the Eemian
how abruptly climatic change can occur (Adams of Europe and the Ipswichian in Britain), but
et al., 1999). An indication of this came from a some workers fear that the ice core records could
deep peat core at Grande Pile in France (Woillard, have been corrupted by deformation within the
1979) where it was argued that the temperate ice. That said, there is also evidence of a sharp
forest of the Last Interglacial (Eemian), c. 120 000 Eemian cold phase in ocean cores (Maslin and
years ago (BP) was replaced by taiga within Tzedakis, 1996).
GFE-06.qxd 6/27/05 4:30 PM Page 127
THE QUATERNARY 127
Depth (m)
0 500 1,000 1,500 2,000 2,500 2,750 3,000 3,200 3,300
–420 110 kyr 390 kyr
–440
δD (‰)
a
–460
–0.5
(‰)
–480
0.0
atm
δ18O
0.5
b
1.0
11.24
0.0
5.4
Ice volume
0.5 c
11.3
11.1
9.1
5.3
5.5
7.1
5.1
8.5
7.5
7.3
9.3
1.0
100
Na (p.p.b.)
1.5 d 50
Dust (p.p.m.)
1.0
0
e
0.5
0.0
0 50,000 100,000 150,000 200,000 250,000 300,000 350,000 400,000
Age (yr BP)
Figure 6.4 Environmental change as revealed in the Vostok ice core. The values along line ‘c’ are marine isotope
stage (after Petit et al., 1999).
High frequency abrupt changes have also response to changes in the circulation of the
been identified from ocean cores (Oppo et al., ocean currents.
1998), where the observed saw-tooth patterns of Recent years have seen an increasing apprecia-
climatic variation have been named Bond cycles. tion of the importance of rapid changes over
Also within the ocean core sediment record there shorter timescales (Viles and Goudie, 2003),
are layers of sediment that are rich in dolomite including decadal and inter-decadal fluctuations
and limestone detritus but poor in foraminifera. such as El Niño Southern Oscillation (ENSO) and
Each layer is interpreted as being the result of the North Atlantic Oscillation (NAO) (Figure 6.5).
deposition by massive armadas of icebergs Such fluctuations can have a significant effect on
released from ice caps around the North Atlantic such geomorphological phenomena as debris flows
(Bond et al., 1992). The records of these iceberg (Grosjean et al., 1997), floods (Magilligan and
flotillas are termed Heinrich events (Andrews, Goldstein, 2001), wave climate (Wang and Surail,
1998) and it is evident that they represent cold 2001), lake levels (Birkett et al., 1999), glacial
stadials of short duration — less than 1000 years damming of drainage (Depetris and Pasquini,
(ka). The Younger Dryas event towards the end 2000), hurricane landfalls (Elsner and Kera, 1999),
of the Last Glacial (c. 12 000 years BP) was an valley-bottom incision (Bull, 1997) and dune reac-
example of a very short-lived stadial (no more tivation (Forman et al., 2001). Changes in the fre-
than a thousand years) that came and went quency and magnitude of such phenomena are of
within decades (Anderson, 1997), perhaps in great engineering concern, particularly in sensitive
GFE-06.qxd 6/27/05 4:30 PM Page 128
(a) 4
3
ARKOA NAO INDEX
2
1
0
-1
-2
-3
-4
1400 1450 1500 1550 1600 1650 1700 1750 1800 1850 1900 1950 2000
Year (December date)

(b) 6
2
L n - Sn
-2
-4
-6
1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000
Year
(c)
3.0 Correlation coefficient for annual data = 0.58
2.0
1.0
0.0
-1.0
NAO index 10 year lowpass filtered NAO index

-2.0 Annual mass balance 10 year lowpass filtered mass balance
1950 1960 1970 1980 1990 2000

Year
Figure 6.5 The North Atlantic Oscillation (NAO): (a) A 555-year reconstruction of the NAO based on tree-
ring analyses; (b) NAO index for December–March from 1864 to 1996 based on the difference in normalized mean
sea-level pressure between Lisbon and Iceland; (c) NAO index (1950–1999) and mass balance of Nigardsbreen
Glacier, Norway (1962–1998). (Modified from various sources in Viles and Goudie, 2003.)
GFE-06.qxd 6/27/05 4:30 PM Page 129
THE QUATERNARY 129
D C
A
A C
Figure 6.6 Glacial conditions in Europe: (A) the position of the present polar timberline in Europe; (B) the
position of the timberline during the maximum of the last (Würm) glacial; (C) the extent of north European
drift deposits that can be attributed to the Last Glacial; (D) the drift borders of the Riss-Saale and Mindel-Elster in
North Europe.
environments where critical thresholds may be more than about 3 per cent in the Southern.
crossed (Chapter 1). These are dealt with in detail Nonetheless, substantial ice sheets developed
elsewhere in this book. over Patagonia and New Zealand. The thickness
of the ice sheets may have exceeded 4 km, with
typical depths of 2 to 3 km.
6.4 The last glacial cycle Highly important changes also took place in
the oceans. During the present interglacial condi-
The last glacial cycle (the Devensian of Britain, tions of the Holocene, the Atlantic is at least
the Würm and Weichselian of Europe and the seasonally ice free as far north as 78 °N in the
Wisconsinian of the USA) reached its peak about Norwegian Sea. This condition reflects the bring-
18 000 years BP, with ice sheets extending over ing of warm water into this region by the Gulf
Scandinavia to the north German plain, over all Stream (North Atlantic Current). During the Last
but the south of Britain, and over North America Glacial Maximum, however, the oceanic polar
to 39 °N. To the south of the Scandinavian ice front probably lay at about 45 °N, and north of
sheet (Figure 6.6) was a tundra steppe underlain this latitude the ocean was mainly covered by sea
by permafrost, and forest was relatively sparse to ice during the winter.
the north of the Mediterranean. The Last Glacial The degree of temperature change that
had a major impact on ground conditions over occurred over land was particularly great in the
extensive areas and accounts for the present day vicinity of the great ice sheets. The presence of
inheritance of frost-shattered rock, sand and permafrost (permanently frozen subsoil) in south-
gravel resources, loess and unstable slopes. ern Britain (Ballantyne and Harris, 1994) sug-
Ice covered nearly one-third of the Earth’s land gests a temperature depression of the order of
area, but the additional ice-covered area was 15 °C. Mid-latitude areas probably underwent a
almost all in the Northern Hemisphere, with no lesser decline (perhaps 5–8 °C) though in areas
GFE-06.qxd 6/27/05 4:30 PM Page 130
subject to maritime air masses temperatures were widespread species-rich grasslands, scrub and
more likely to have been depressed by 4–5 °C. open woodlands, which grow on unleached, fertile
Periglacial conditions (i.e. conditions charac- soils with a still low humus content. In the third,
terised by snow and frost though not by glaci- or mesocratic phase, temperate deciduous forest
ation) caused many slopes in areas like southern and fertile, brown-earth soils develop under warm
England to be unstable, especially as a result of conditions, allowing the expansion of shade-
solifluction and landslides. Cambering and valley giving forest genera such as Quercus, Ulmus,
bulging were widespread phenomena. Relict Fraxinus and Corylus, followed by slower immi-
instability is widespread and is susceptible to grants such as Fagus and Carpinus. In the fourth
reactivation (Chapter 14). and last retrogressive phases, the telocratic phase,
Another feature of areas to the south of the great soil deterioration and climatic decline leads to the
northern-hemisphere ice sheets was the deposition development of open conifer-dominated woods,
of silt-sized aeolian material (loess) derived from ericaceous heaths and bogs growing on less fertile
deflation of outwash plains and moraines (Pye, humus-rich podzols and peats.
1987). Over vast areas (at least 1.6 ⫻ 106 km2 in
Europe) it blankets pre-existing relief, and reaches
thicknesses in excess of 200 m in Tajikistan, 6.6 Holocene
Uzbekistan and in China. Other important areas
include the Missouri–Mississippi plains, the North The maximum of the Last Glacial occurred at
German Plain and the Pampas of Argentina and around 18 000 years BP. Studies of the Oxygen
New Zealand. Loess presents its own particular isotope composition of deep-sea cores suggest
problems for engineering structures (Derbyshire that deglaciation started at around 15 000 to
et al., 2000; Chapter 25). 14 500 years BP in the North Atlantic and at
16 500 to 13 000 years BP in the Southern Ocean
(Bard et al., 1990). The years between the glacial
6.5 Interglacials maximum and the beginning of the Holocene are
usually termed the Late Glacial and they were
In general terms the Quaternary interglacials were marked by various minor stadials and interstadi-
short lived but appear to have been essentially als, but the character, identification and correla-
similar in their climate, fauna, flora and land- tion of the Late Glacial interstadials is a matter
forms to the present Holocene interglacial. One of which is still in need of clarification (Anderson,
their most important characteristics was that they 1997).
witnessed the rapid retreat and decay of the great The ending of the Last Glacial period, was,
ice sheets and saw the replacement of tundra con- however, not the end of substantial environmental
ditions by forest over the now temperate lands of change. Indeed, as the Holocene progressed
the Northern Hemisphere. At their peak they may the impact of climatic change was augmented
have been a degree or two warmer than now. as a cause of environmental fluctuation by the
The general sequence of vegetational and soil increasing role of humans (Roberts, 1998).
development during an interglacial has been Some portions of the Holocene may have been
described for north-west Europe by Birks (1986). slightly warmer than now and terms like ‘climatic
The first, or cryocratic phase, represents cold gla- optimum’ have been used to denote the existence
cial conditions, with sparse assemblages of pio- of a possible phase of mid-Holocene warmth,
neer plants growing on base-rich skeletal mineral when conditions may have been 1–2 °C warmer
soils under dry, continental conditions. In the sec- than present. There may also have been a Medieval
ond, or protocratic phase, there is the onset of Climatic Optimum between AD 750 and AD 1300.
interglacial conditions. Rising temperatures allow However, there have also been times which have
the base-loving, shade-intolerant herbs, shrubs been rather colder than today, as is made evident
and trees to migrate and expand quickly to form by phases of glacial re-advance (neoglaciations) in
GFE-06.qxd 6/27/05 4:30 PM Page 131
THE QUATERNARY 131
alpine valleys. The latest of these was ‘the Little Temperature depression and reduced glacial
Ice Age’ which peaked around AD 1700 and ended atmospheric CO2 levels also played a role in
towards the end of the nineteenth century (Grove, changing environmental conditions in low lati-
2004). tudes. The substantial degree of change in the alti-
Fluctuations of climate also occurred in lower tude of vegetation zones on tropical mountains
latitudes, and of especial importance for vege- during cold phases can be demonstrated by
tation and human activities was the early to detailed pollen analysis from lakes and swamps
mid-Holocene pluvial, which transformed the from tropical highlands. Vegetation zones and
Sahara from a hyper-arid region into a savanna geomorphological processes may have moved
(Ritchie et al., 1985). The hyper-arid belt more or through as much as 1700 m of altitude.
less disappeared for one or two millennia before A combination of temperature and precipita-
7000 BP. The northern limit of the Sahel shifted tion change had a dramatic impact on the nature
about 1000 km to the south around 18 000 BP and and extent of rainforests in Africa and South
about 600 km to the north at 6000 BP compared to America (Goudie, 1999). This is brought out by
the present. a consideration of pollen analyses undertaken in
Lake Bosumtwi in southern Ghana (Maley,
1989). At the time of the Last Glacial Maximum,
6.7 The Quaternary in low latitudes between 20 000 and 15 000 years BP, the lake had
a very low level and arboreal (tree) pollen per-
The events which led to the expansions and con- centages reached minimum values of between
tractions of the great ice sheets in middle and high 4 and 5 per cent. Trees were in effect replaced at
latitudes led to major changes in lower latitudes. that time by herbaceous plants. This compares
Periods of greater moisture (pluvials) were inter- with the situation since c. 8500 BP, when arbo-
spersed with periods of less moisture (inter- real pollen percentages have oscillated between
pluvials). The evidence for such changes is 75 and 85 per cent. The shifts in the position of
particularly evident on the margins of great the monsoon in the Last Glacial Maximum and
deserts, where dry phases saw the development the Holocene are shown in Figure 6.8. This
and advance of great sand seas, whereas in clearly had a significant impact on processes,
moister phases the dunes were stabilised by vege- especially erosion and instability (Chapters 8
tation and large lakes filled with water in areas and 11).
that had previously been salty wastes.
Some of the lakes that developed in pluvial
phases were enormous. One of the greatest con- 6.8 Sea-level changes
centrations of pluvial lakes developed in the
Basin and Range Province of the American In addition to the climatic and vegetational changes
Southwest. Between 100 and 120 depressions, which have been discussed so far, it is important to
formed by high-angle faulting, were occupied remember that the Quaternary has also been a time
wholly or in part by pluvial lakes during various of major changes in sea level (Cronin, 1999). Many
phases of the Pleistocene. factors influence sea level in any particular loca-
By contrast, in interpluvials large dune fields tion (Table 6.1), but among the most important are
expanded. Relict forms occur in areas where there glacial eustasy and glacio-isostasy. The former is
is now a well-developed vegetation cover and the worldwide change in sea level that results from
annual precipitation totals of around 800 mm the waxing and waning of the great ice sheets,
(Figure 6.7). The dunes probably formed when which took up or released fresh water into the
vegetation cover was much less capable of oceans. Thus since the last glaciation, the melting
inhibiting sand movement under annual precipita- of the ice caps has caused world average sea levels
tion totals that were less than 100 to 300 mm to rise by something between 100 and 150 m. The
(Goudie, 1992; 1999). Flandrian transgression, as this event is called
GFE-06.qxd 6/27/05 4:30 PM Page 132
(a) Today 150˚ 120˚ 90˚ 60˚ 30˚ 0˚ 30˚ 60˚ 90˚ 120˚ 150˚ 180˚
60˚
60˚
40˚ 40˚
20˚ 20˚
0 3,000km
0˚ 0˚
Equatorial
scale
20˚ 20˚
40˚ 40˚
150˚ 120˚ 90˚ 60˚ 30˚ 0˚ 30˚ 60˚ 90˚ 120˚ 150˚ 180˚
Ice-caps Sand dune areas
(b) 18 000 BP 150˚ 120˚ 90˚ 60˚ 30˚ 0˚ 30˚ 60˚ 90˚ 120˚ 150˚ 180˚
60˚
60˚
40˚ 40˚
20˚ 20˚
0˚ 0˚
20˚ 20˚
40˚ 40˚
150˚ 120˚ 90˚ 60˚ 30˚ 0˚ 30˚ 60˚ 90˚ 120˚ 150˚ 180˚
Figure 6.7 The extent of sand seas at c. 18 000 years ago (lower) in comparison with the situation today.
(Figure 6.9), flooded the continental shelves and marine and terrestrial sediments including poten-
created many of the flooded river valleys (rias, tially unstable clays and peats.
fjords etc.) and embayments of present-day coast- Sea level became stabilised at around its present
lines. Some of these inlets have since been filled in, height (give or take a few metres) at 6000 years BP.
partially or totally, by complex alternations of The significance of this cessation of sea-level rise
GFE-06.qxd 6/27/05 4:30 PM Page 133
THE QUATERNARY 133
0˚ 60˚ 120˚
AN
40˚
CE
IC O
AT LANT
N
SAHARA
EA
THAR
OC
C
IF I
PAC
INDIAN
0˚ OCEAN
Monsoon range:
Present Last Glacial Maximum Holocene optimum
Figure 6.8 The range of the monsoon between the Last Glacial Maximum and the Holocene (from Goudie, 1999).
0 Table 6.1 Factors in sea-level change.
Eustatic (worldwide) Local
Glacio-eustasy Glacio-isostasy
Infilling of basins Hydro-isostasy
Orogenic-eustasy Erosional and depositional
isostasy
50 Decantation Compaction of sediments
Depth (m)
(autocompaction)
Transfer from lakes Orogeny
to oceans
Expansion or contraction Epeirogeny
of water volume because Ice–water gravitational
of temperature change attraction
Juvenile water
100 Geoidal changes
0 5 10 15 modest amounts of replenishment, and thus may be

Age (1000 yrs) prone to serious decay if they are used as sources
of aggregates (Chapters 21 and 22).
Figure 6.9 The sea-level curve for Barbados, based on Glacio-isostasy results from the application
radio-carbon dated corals (Acropara palmata). The cir-
cles represent data from Barbados corrected for esti-
and release of pressure by glaciers on the earth’s
mated uplift. The crosses represent data from four other crust. As already noted, many of the great ice caps
Caribbean islands. The curve reflects the depth range of were very thick and caused crustal depression of
live corals while the broken line has been adjusted to the order of 200–300 m in certain parts of the
sea-level (after Fairbanks, 1989). high latitudes (Figure 6.10). Since the ice burden
has been released, the land has risen (and contin-
is that many of the shingle structures of the British ues to do so), but in areas more distant from the
coast (e.g. Chesil Beach, Dorset) resulted from the former ice sheets some compensatory depression
upward combing of sediments by a rising sea. They has occurred, leading to accelerated flooding of
are now largely relict features receiving relatively low-lying coastal areas.
GFE-06.qxd 6/27/05 4:30 PM Page 134
(a)
160
Metres of emergence
120
n
e
ia
id
nd
nt
ca
e
80 ur
os
La
nn
Fe
40
h
Britis
0
1200 1000 800 600 400 200 0 200 400
Kilometres from the centre of ice sheet
(b) (c)
0
70˚N 70˚N
00
1.
20 0.3
33000
000
1155200
2000
0
200 0
2.0 00 0.7
1000 0
1100000
0
1.0
500 5
1.
5 0.
0.3
3
1.
0 0.
1.5 0.5
0.7
2.0
2.5 0.9
1.0 0.3
2.0 0.7
0.9
3.5
3.0
1.3
1.1
5
2.
2.0
50˚N 50˚N
1.0
0.7
1.5 0.5
0.3
0 500 1000km 0 500 1000km

˚N ˚N
40 40
100˚W 80˚W 100˚W 80˚W
Figure 6.10 Glacio-isostatic adjustment: (A) cross-section through three Northern Hemisphere ice sheets, show-
ing the amounts of isostatic recovery over the past 7000 years; (B) Average rate of Holocene uplift in metres per 1000
years for northern and eastern Canada — the inset shows the thickness of the Laurentide ice sheet at 18 000 years
ago (contours in metres), and (C) present rate of uplift in metres per 100 years of northern and eastern Canada.
6.9 Summary time and many of them are not necessarily the
result of currently active processes. Even within
Over the last two million years the world has recent decades substantial changes have taken
seen many major environmental changes, and place in river discharges, lake dimensions,
has been in a state of constant flux. Many of the glacial snout positions etc. and these need to be
landforms and surface materials upon which considered when engineering schemes are being
human activities are conducted date from this planned.
GFE-06.qxd 6/27/05 4:30 PM Page 135
THE QUATERNARY 135
References Fairbanks, R. G. (1989) A 17 000-year glacio-eustatic

sea level record: influence of glacial melting rates
Adams, J., Maslin, M. and Thomas, E. (1999) Sudden on the Younger Dryas event and deep ocean circula-
climate transitions during the Quaternary. Progress tion. Nature 342, 637–42
in Physical Geography 23, 1–36. Forman, S. L., Oglesby, R. and Webb, R. S. (2001)
Anderson, D. (1997) Younger Dryas research and its Temporal and spatial patterns of Holocene dune
implications for understanding abrupt climatic activity on the Great Plains of North America:
change. Progress in Physical Geography 21, 30– megadroughts and climate links. Global and
249. Planetary Change 29, 1–29.
Andrews, J. T. (1998) Abrupt changes (Heinrich events) Frenzel, B. and Blundau, W. (1987) On the duration of
in late Quaternary North Atlantic marine environ- the interglacial to glacial transition at the end of the
ments. Journal of Quaternary Science 13, 3–16. Eemian Interglacial (Deep Sea Stage 5E): botanical
Ballantyne, C. K. and Harris, C. (1994) The periglacia- and sedimentological evidence. In Berger, W. H. and
tion of Great Britain. Cambridge University Press. Labeyrie, L. D. (eds) Abrupt Climatic Change. D.
Bard, E., Labergrue, L. D., Pichon, J. J., Labracherie, Reidel, Dordrecht and Boston. 151–162.
M., Arnold, M., Duprat, J., Moyes, J. and Duplessy, Goudie, A. S. (1992) Environmental Change (3rd
J. C. (1990) The last deglaciation in the southern edition). Oxford University Press.
and northern hemispheres. In Bleil, V. and Thiede, J. Goudie, A. S. (1999) The Ice Age in the Tropics. In
(eds) Geological History of the Polar Oceans: Slack, P. (ed.) Environments and Historical Change.
Arctic versus Antarctic. Kluwer, Dordrecht, 405– Oxford University Press, 10–32.
415. Grove, J. M. (2004) Little Ice Age: Ancient and Modern.
Birkett, C., Murtugudde, R. and Allan J. A. (1999) Routledge, London.
Indian ocean climate event brings floods to East Grosjean, M., Nunez, L., Castajena, I. and Messerli, B.
Africa’s lakes and the Sudd Marsh. Geophysical (1997) Mid-Holocene climate and culture changes
Research Letters 26, 1031–1034. in the Atacama Desert, Northern Chile. Quaternary
Birks, H. J. B. (1986) Quaternary biotic changes in Research 48, 239–246.
terrestrial and lacustrine environments, with particu- Lowe, J. J. and Walker, M. J. C. (1997) Reconstructing
lar reference to north-west Europe. In Bergland, B. E. Quaternary Environments (2nd edition). Longman,
(ed.) Handbook of Holocene Palaeoecology and Harlow.
Palaeohydrology. Wiley, Chichester, 3–65. Magilligan, F. J. and Goldstein, P. S. (2001) El Niño
Bond, G. and 13 collaborators (1992) Evidence for mas- floods and culture change: a late Holocene flood
sive discharges into the North Atlantic Ocean during history of the Rio Moquegua, Southern Peru.
the last glacial period. Nature 360, 245–249. Geology 29, 431–434.
Bull, W. B. (1997) Discontinuous ephemeral streams. Maley, J. (1989) Late Quaternary climatic changes in
Geomorphology 19, 227–276. the African rainforest: forest refugia and the major
Cronin, T. M. (1999) Principles of Palaeoclimatology. role of sea surface temperature variations. In Leinen,
Columbia University Press, New York. M. and Sarnthein, M. (eds) Palaeoclimatology and
Dansgaard, W. and 8 collaborators (1993) Evidence for Palaeometeorology. Reidel, Dordrecht, 585–616.
general instability of past climate from a 250 k yr Maslin, M. A. and Tzedakis, C. (1996) Sultry last inter-
ice-core record. Nature 364, 218–220. glacial gets sudden chill. Eos 77, 353–354.
Depetris, P. J. and Pasquini, A. I. (2000) The hydro- Oppo, D. W., McManus, J. F. and Cullen, J. L. (1998)
logical signal of the Perito Moreno Glacier Abrupt climate events 500 000 to 340 000 years ago:
damming of Lake Argentino (Southern Andean evidence from sub-polar North Atlantic sediments.
Patagonic): the connection to climate anomalies. Science 279, 1335–1338.
Global and Planetary Change 26, 367–374. Petit, R. J. and 18 collaborators (1999) Climate and
Derbyshire, E., Meng, X. M. and Dijkstra, T. A. (eds) atmospheric history of the past 420 000 years from
(2000) Landslides in the Thick Loess Terrain of the Vostok ice core, Antarctica. Nature 399, 429–436.
Northwest China. Wiley, Chichester. Pye, K. (1987) Aeolian dust and dust deposits.
Dickson, R. R., Osborn, T. J., Hurrell, J. W., Meincke, J., Academic Press, London.
Blindheim, J., Adlandsvik, B., Vinje, T., Alekseev, G. Ritchie, J. C., Gyles, C. H. and Haynes, C. B. (1985)
and Maslowski, W. (2000) The Arctic Ocean Sediment and pollen evidence for an early to mid-
response to the North Atlantic Oscillation. Journal Holocene humid period in the eastern Sahara.
of Climate 13, 2671–2696. Nature 314, 252–255.
Elsner, J. B. and Kera, A. B. (1999) Hurricanes of the Roberts, N. (1998) The Holocene. (2nd edition).
North Atlantic. Oxford University Press, New York. Blackwell, Oxford.
GFE-06.qxd 6/27/05 4:30 PM Page 136
Viles, H. A. and Goudie, A. S. (2003) Internannual, Routledge and the Open University, London and
decadal and multidecadal scale climatic variability New York.
and geomorphology. Earth-Science Reviews 61, Woillard, G. (1979) Abrupt end of the last interglacial
105–131. s.s. in north-east France. Nature 281, 558–562.
Wang, X. L. and Surail, V. R. (2001) Changes of
extreme wave heights in Northern Hemisphere
oceans and related atmospheric circulation regimes. Further reading
Journal of Climate 14, 2204–2221.
Williams, M. A. J., Dunkerly, D., De Dekker, P., Kershaw, Goudie, A. S. (2002) Great Warm Deserts of the World:
Landscapes and Evolution. Oxford University Press.
P. and Chappell, J. (1998) Quaternary Environments Ruddiman, W. F. (2001) Earth’s Climate: Past and
(2nd edition). Edward Arnold, London. Future. W. H. Freeman, New York.
Wilson, R. C. L., Drury, S. A. and Chapman, J. L. Wilson, R., Drury, S. A. and Chapman, J. L. (2000) The
(2000) The Great Ice Age. Climate Change and Life. Great Ice Age. Routledge, London.
GFE-07.qxd 6/24/05 3:01 PM Page 137
7. Engineering Behaviour of Soils and Rocks

George Milligan, Peter Fookes and Mark Lee
7.1 Introduction well recognised. Nevertheless, a fairly simple basic

framework for the behaviour of soil can be estab-
A brief introduction to the engineering properties lished, with natural variations being accounted for
of soils and rocks is appropriate in this book by modifications to or differences from this model.
for two reasons: the physical properties of soils For further reading, there are a number of excellent
and rocks, in particular strength and permeability, modern references dealing with soil mechanics and
affect many geomorphological processes, and the the classical problems of geotechnical engineering,
ground conditions encountered in many engi- such as Bolton (1991), Atkinson (1981; 1993) and
neering projects are the result of such processes. Powrie (1997).
Ground material properties, therefore, constitute The development of rock mechanics as the sci-
both input to and output from these processes. ence of rock engineering has been even more
For example, the ground conditions on a valley recent. Useful introductions to rock mechanics
side may be the result of past glacial or peri- and rock engineering are provided by Hudson
glacial actions; the current nature and state of (1989) and Waltham (2001).
those soils will then determine how and if the slope The main ideas to be covered in this chapter
will fail due to erosion of its toe by a present-day include the important differences between coarse-
river, or as a result of a cut into the hillside for a and fine-grained soils, the importance of plasticity
new road. in fine-grained soils, the concept of effective stress
No attempt will or should be made here to give in soil, the interrelation of density, volume change
extensive coverage of the sciences of soil and rock and strength, and some of the ways of assessing
mechanics or their application in geotechnical the properties of soil. Because geomorphology is
engineering. What can be attempted is to provide mainly concerned with relatively superficial
a general understanding of the range of material deposits, the main emphasis is on the behaviour of
behaviour and some of the key factors affecting soils. However, a brief discussion is also included
that behaviour. Soils and rocks are mechanically on the engineering behaviour of rock masses, in
complex, even in their simplest states, with particular the influence of discontinuities on near-
strength and stiffness related to their geological surface processes.
origin, past stress and strain history, current stress Many geomorphological processes and engin-
state and loading path. Natural ground conditions eering constructions involve issues of slope sta-
commonly comprise layers of different soil and bility (Chapter 8), so a brief discussion of the
rock types within each of which there will be vari- methods and some of the results of slope stability
ations in properties both spatially and in direction, analysis are included (7.3 Slope instability); for
i.e. the materials are neither homogeneous nor further information, reference should be made to
isotropic. Simons et al. (2001) or other specialist text on
Concepts in soil mechanics, and applications of slope stability. The particular problems or risks
geotechnical engineering, have both developed associated with different soil types resulting from
very rapidly in recent decades, and this has been geomorphological processes are summarised
the most active research area in civil engineering, (7.4 Geotechnical problems associated with dif-
so that the need for specialist practitioners is now ferent soil types). Typical values of engineering
137
GFE-07.qxd 6/24/05 3:01 PM Page 138
properties of a range of soil types are also Fine soils are those consisting primarily of par-
provided (7.5 Typical values of soil properties). ticles in the clay and silt size ranges. Clay particles
are produced mainly by chemical weathering of
7.2 Elements of behaviour parent rocks in situ, often followed by erosion,
transportation and deposition elsewhere, or by
Coarse and fine soils direct precipitation from water. The resulting clay
Soil particles exist in a range of sizes from less minerals are complex, but the most important are
than 1 m to more than 500 mm. The words clay, built up from basic crystalline units of silica and
silt, sand, gravel, cobbles and boulders are used in alumina to form rod or sheet structures, which may
British engineering practice in a standard way to combine in various ways to form different clay
describe particular ranges of particle sizes, as minerals. The three commonest clay minerals are
shown in Figure 7.1. They are also used, along kaolinite, illite and montmorillonite (a member of
with suitable qualifying terms, to describe natural the smectite family of clay minerals); some basic
soils in which the properties of a particular size properties are given in Table 7.1. Montmorillonite
range predominate. is particularly notable in that it is able to store
Coarse soils are those mainly in the sand size water within the particles between the crystal lay-
and upward; they consist predominantly of par- ers, which are much less strongly bonded together
ticles of strong crystalline minerals such as silica, than in other clay minerals. It is able to change
which have been broken down from parent rock volume by a much greater amount than other clay
by physical processes of weathering. The particles minerals, so that soil containing a high proportion
are of the same material as the parent rock and of montmorillonite tends to swell on wetting and
approximately equi-dimensional, but may be very contract on drying much more than other soils.
angular if recently broken, or smooth and rounded At the other extreme, the layers in kaolinite are
if they have suffered abrasion and polishing by fairly strongly bonded and particles of kaolinite
water or wind action. are relatively bulky for a clay.
0.002 0.006 0.02 0.06 0.2 0.6 2 6 20 60 200 mm
CLAY SILT SAND GRAVEL Cobbles Boulders
Fine Medium Coarse Fine Medium Coarse Fine Medium Coarse
Sedimentation analysis Sieve analysis
Limit of visibility
with naked eye
0.001 0.01 0.1 1 10 100 mm
CLAY
COLLOIDS
Thickness of clay mineral Sedimentation

sheets analysis
Limit of visibility with Limit of visibility

Diameter of electron microscope with optical
water molecule microscope
0.0001 0.001
mm
0.001 0.01 0.1 1 m
1 10
Figure 7.1 Range of particle sizes of soils.

GFE-07.qxd 6/24/05 3:01 PM Page 139
ENGINEERING BEHAVIOUR OF SOILS AND ROCKS 139
Table 7.1 Some basic properties of soil minerals.
Quartz Clays
Sand
Kaolinites Illites Montmorillonites
Particle diameter (m) ⬃ 1000 0.1⬃ 4 0.1–⬃2 0.01–⬃2

Particle thickness (Å) ⬃ 1000 500 50–300 10–100
Specific surface (m2/g) 0.002 5–20 80–120 700–800
Natural soils described as clays may contain as of approximately equi-dimensional particles. Thin
little as 10% clay-sized particles, and seldom more clay particles can pack so that the same volume
than 50%. This is because the clay-sized particles of solid material occupies very different total
have a profound effect on the soil behaviour, in two volumes; limits can be thought of as akin to a pack
main ways, both related to the pore size and pack- of cards (extremely densely packed), or the same
ing of the soil particles. Soil is a particulate mate- cards built into a card house (extremely loose
rial, and as such must have spaces or pores between packing). The properties of the soils will clearly
the solid grains; these are filled with fluid (usually be very different. The packing of particles is
water), gas (usually air), or a combination of both. described numerically by the porosity (usual
The presence of this pore fluid, and the ease with symbol n), which is the volume of voids divided
which it may move through the soil, are of funda- by the total volume; the void ratio (symbol e),
mental importance to the behaviour of soils. The which is the volume of voids divided by the volume
size of the pores is obviously related by geometry to of soil particles; or the specific volume (symbol v),
the size of the particles, and in clay soils will be which is the volume containing unit volume of
very small. In mixed soils the pore sizes may also solid material. They are related by:
be very small provided there is sufficient clay-sized
n e/(1 e), v (1 e)
material to fit into the spaces between the larger
particles. Such soils will be nearly impermeable. For coarse soils the value of e is generally in the
Also, very large capillary forces may be developed range 0.3 to 1.0, but for clays it may range from
in small pores, and the resulting suction will hold below 0.2 to above 9 as a result of variations in
the material together even when no external stresses the geological origin and stress history of the
are applied. This ability to hold together in a lump deposit (see Compression of soils).
and undergo substantial distortion without breaking If the voids are completely filled with air, the
up makes clay appear cohesive and plastic. Suction soil is dry; if completely filled with water it is
forces also exist in damp sand (hence sand castles), fully saturated. For intermediate states the soil is
but the effects are very limited and easily lost on the partially saturated, with the degree of saturation Sr
wetting or drying out of the sand. (volume of water over volume of voids) ranging
Silts often cause particular problems in identifi- from zero to unity. Fine soils are often fully satu-
cation and mechanical properties. Silt particles are rated in their natural state, except in arid condi-
similar to sand particles, and silt may exhibit many tions; the water content of the soil (w), which in
of the properties of fine sands, but the small parti- soil mechanics is defined as the mass of water
cle size results in silt having significant plasticity divided by the mass of soil particles, is then
and relatively low permeability. Most natural clay directly related to the void ratio by:
soils contain a substantial proportion of silt, but
w e/Gs
their properties are dominated by the clay fraction.
The different natures of fine and coarse par- where Gs is the specific gravity of the soil particles.
ticles greatly affect the possible particle packing The susceptibility of soil to erosion by water is
arrangements. There is only a limited range of dens- an important factor in many geomorphological
ity between the loosest and densest arrangements processes and also in the failure of many types of
GFE-07.qxd 6/24/05 3:01 PM Page 140
construction. Erosion may be due to impact of rain of material for the UK. Much information is
drops, overland flow, underground seepage, or available from databases and statutory registers
water flow in a river, sea or ocean. It is largely con- relating to possible geological and environmental
trolled by the particle size of the soil. Large parti- risks such as unstable ground, old mines or land
cles of gravel size or greater need large amounts of contaminated by past industrial use.
energy to lift and transport them. They can be dis- Detailed examination of the succession of soil
placed by rapid rivers in flood or storm waves, but strata will then be carried out using trial pits and
are quickly redeposited when water velocities boreholes; these will also be used for obtaining
reduce. At the other extreme, clay may be resistant samples of the soil for laboratory testing, and for
to erosion due to its innate cohesion, except in the undertaking in situ tests such as the standard pene-
case of dispersive clays. Sands and (especially) silts tration test (SPT) (Stroud, 1989), vane shear test
are the most susceptible to erosion, and silt in sus- or pressuremeter test. Methods of investigation
pension may be carried great distances before being are covered by British Standard 5930: Code of
deposited. Wind can also act as the agent of erosion; Practice for Site Investigations (BS, 1999), and
the particle sizes affected are more restricted to silt other national standards. There are also various
and fine sand but the size of a wind system may be methods of probing and in situ testing which may
such that very large amounts of material may be be done without boreholes, such as cone or self-
carried for very large distances. The importance of boring pressuremeter tests (Meigh, 1987; Mair and
erosion in geomorphology is addressed in Chapter Wood, 1987). In some cases geophysical methods
11, and to a lesser extent in Chapters 3 and 10. involving seismic, electrical, gamma radiation,
In engineering calculations soil is usually magnetic or gravimetric surveys undertaken from
treated as a continuum material, but its properties the surface or down boreholes can provide much
are closely related to its particle arrangement or additional information, but usually require inter-
microstructure. Its bulk properties may also be pretation by specialists. A good general introduc-
significantly affected by its ‘fabric’, such as fine tion to the purpose and methods of site
laminations of sand in a clay soil, cracks or fis- investigation is provided by Waltham (2001), while
sures, or inclusions of organic material (Rowe, Fookes (1997) provides an extensive discussion of
1972). An extreme example of an organic soil is site investigation as a means of establishing the
peat, which consists almost entirely of decayed basic geological model for a site. The organisation
vegetable matter. of a typical simple site investigation is summarised
Many real soils have complex structures and in Figure 7.2; for small projects a single stage of
their mechanical behaviour may depart signifi- field work is normal, but for large projects there
cantly from that of the ‘ideal’ soils discussed in this may be several stages, each building on the infor-
section. This should not be taken as a reason to mation from earlier work. For further coverage see
avoid attempting to apply quantitative methods of Clayton et al. (1995) and Fookes (1997). Fookes
analysis, but these should always be undertaken on et al. (2001) provide the basis for a total geological
the basis of a good understanding of the geological, history approach to site investigation.
hydrogeological and geomorphological origins of a The soils recovered are first described and
site, along with a clear picture of the resulting likely classified on the basis of visual and tactile prop-
variability in ground and groundwater conditions. erties; methods are given in British Standard 5930
(BS, 1999). Classification is then assisted by sim-
Identification and classification ple laboratory tests, details of which are given in
Investigation of a site will usually start with a British Standard 1377 (BS, 1991), Methods of Test
survey of existing information from geological for Soils for Civil Engineering Purposes. Particle
maps and memoirs, historical maps, aerial and size analyses, using sieving for coarse soils and
satellite photographs, records from adjacent con- sedimentation methods for fine soils, produce
struction, and other local knowledge. Perry and curves showing the distribution of particle sizes
West (1996) provide useful guidance on sources within the soil, typical examples being shown in
GFE-07.qxd 6/24/05 3:01 PM Page 141
Figure 7.2 Flow chart for a simple one-stage site investigation; large or complex projects may have several stages
of investigation (see Fookes, 1997).
Figure 7.3. For historical reasons these are plotted change in behaviour from that of a plastic solid to
as cumulative curves rather than frequency distri- a viscous liquid as the soil becomes wetter, and the
bution curves. plastic limit (PL) the change from a plastic to a
For fine soils, much useful information about brittle solid as the soil becomes drier. These water
the soil may be obtained by determining the natural contents are expressed as percentages, the LL and
water content as it exists in the ground and the PL to the nearest whole number; the difference
water contents corresponding to two arbitrarily between the LL and PL is called the plasticity
defined limits of consistency known as the index (PI). LL, PL and PI all tend to increase with
Atterberg limits. The liquid limit (LL) defines the the clay content (the percentage of clay-sized
GFE-07.qxd 6/24/05 3:01 PM Page 142
D E
B
F
G
C
Figure 7.3 Typical particle size distribution plots for various soils: (A) marine clay; (B) alluvial silty clay (estuarine);
(C) uniform silt (loess); (D) uniform fine sand (dune); (E) glacial till (boulder clay); (F) sandy gravel (coarse alluvium);
and (G) uniform (clean) gravel.
Table 7.2 Typical plasticity indexes and activities of soils from low to high plasticity and divides soils
clays. that are predominantly silty or organic from those
which have significant clay content. Detailed clas-
Plasticity index Activity
sification systems have been developed based on
Kaolinites 30 0.4–0.5 the particle size distributions and plasticity tests,
Illites 70 0.5–1.0 see Appendix A2.
Montmorillonites 200 1–7 These simple tests give an initial indication of
Silty clay 45–60 0.6 the likely engineering properties of fine soils.
Glacial clay 10–30 0.5–0.75 Natural water contents in the ground generally
Loess 4–9 0.5–0.75
range between somewhat below the PL to just
Weathered mudstone 17–35 0.56–0.70
above the LL. The former would indicate a stiff,
Marine clay 60–100 1.6–1.8
Organic estuarine clay 50–90 ⬃1.3 strong soil, the latter a very weak and compressible
Lateritic clay 20–45 0.3–1.15 soil. The relative consistency is described by the
London clay 35–60 0.75–1.25 liquidity index (LI) given by:
LI (w – PL)/PI
particles in the soil), and also depend on the nature In some countries a consistency index (CI) is
of the clay minerals. The activity of the clay is indi- used, which is given by:
cated by the ratio of plasticity index to clay
content; typical values for common clay minerals CI (LL – w)/PI
and some natural clays are given in Table 7.2.
and hence:
Classification of clay soils is assisted by the use
of the Casagrande plasticity chart which grades CI (1 – LI)
GFE-07.qxd 6/24/05 3:01 PM Page 143
However, it must be remembered that these Since stationary fluid cannot sustain shear stress,
index tests are carried out on ‘remoulded’ speci- total and effective shear stresses are the same, thus:
mens of the soil, in which all the natural soil fab-

ric has been destroyed. Thus while they can
provide a very useful baseline to likely engineer- Note that effective stresses cannot be measured, but
ing properties, the behaviour of the natural soils only determined from total stresses and pore pres-
in the field may be significantly affected by sures. However it is change in effective stress that
structural features. For example, stiff clays will results in deformation and failure of soil and hence
generally have a smaller mass strength than controls the behaviour. Similar concepts apply in
expected due to the presence of fissures, while unsaturated soils, but inevitably become more
soft clays may initially be stronger than expected complex and are still the subject of active research.
due to slight cementation between particle con- Soils generally fail in shear, in an essentially
tacts; soils of volcanic origin may have porous frictional manner, failure occurring when the
particles which hold water and give natural shear stress on some plane through the soil reaches
water contents and liquid limit values which are a certain proportion of the effective normal stress
not consistent with ‘normal behaviour’, while across the plane:
fine soils in a partially saturated state will appear
tan
much stronger than expected due to the effects of
capillary suction. The constant of proportionality () is commonly
expressed as the tangent of the angle of internal
friction of the soil (). Some heavily compressed
Effective stress and soil strength
or partially cemented soils exhibit shear resist-
A key to the understanding of soil behaviour is
ance even under zero normal stress; they have an
the concept of effective stress. The pore spaces in
effective cohesion (c), and the failure condition
soils are interconnected and fluid pressure may
becomes:
therefore be transmitted through the pore fluid; in
an unsaturated soil there will be both air and c tan
water pressures, which may be different. In sat-
This cohesion should not be confused with the
urated soils, fluid pressure will vary hydrostati-
undrained strength or cohesion (cu), which is
cally with depth below the groundwater level
discussed below.
(water table); above the water table the soil may
still be saturated but pore pressures will be nega-
Shear failure and critical states
tive relative to atmospheric pressure. Water may
The shear failure of soil may be considered in
flow through soil via the pore spaces, but addi-
terms of a direct shear test carried out in a simple
tional pressure differentials must exist to drive the
piece of laboratory equipment known as a shear
flow.
box (Figure 7.4). The test is normally carried out
The pore pressure has the effect of reducing
with a constant vertical stress from a load P and
the contact forces between the soil particles; it is
the soil specimen forced to undergo shear dis-
the summation of these forces over a unit area that
placement
x by a shear force Q. The soil is
is known as effective stress. Thus the total normal
confined horizontally, but free to expand or dilate
stress () across a plane in a body of soil will be
vertically, and the vertical expansion
y is
carried partly by pore pressure (u) and partly by
measured. Pore pressures are usually zero — either
effective stress () so that:
the soil is dry or testing is slow so that pore pres-
u sures are allowed to dissipate. Typical plots of
results are shown in Figure 7.4. If the soil is ini-
and tially dense, the shear force is seen to rise to a max-
imum, then drop back to a constant value. The peak
u in strength arises because the densely packed soil
GFE-07.qxd 6/24/05 3:01 PM Page 144
vary. On shearing, the dense soil will try to expand

and the vertical stress must increase to prevent it
from doing so; eventually a critical state will be
reached with the same initial density but a higher
stress. Conversely, the loose soil will tend to
contract and the vertical stress decrease.
A third type of test could have a constant ver-
tical applied stress, but the soil fully saturated,
with no water allowed to move in or out of the
specimen. Since both soil grains and water are
effectively incompressible, this specimen also
cannot change volume on shearing. To prevent
volume change in the dense soil, the effective
stress must increase, and this occurs by negative
pore pressure developing; in the loose soil, reduc-
Figure 7.4 Schematic of shear box apparatus for
measuring the shear strength of soils, with typical tion in volume will be prevented by positive pore
results plotted as: shear resistance (Q) against displace- pressure that reduces the effective stress.
ment (
x), and dilation (
y) against
x. In the first type of test the soil is sheared under
‘fully drained’ conditions, in the third under
grains have to be forced apart before the soil can ‘undrained’ conditions; note that either drained or
shear; this is seen as initial dilation of the specimen undrained strength may be the higher, depending
followed by further shearing at constant volume. on the initial conditions of the soil and also on
The peak of strength corresponds to the maximum the changes in stress imposed on the soil. In field
rate of dilation. If on the other hand the soil is ini- conditions, changes in pore pressures in coarse
tially loose, it will at first compress slightly on soils usually dissipate quickly due to the high
shearing, with the shear resistance increasing permeability of the soil, but fine soils will
monotonically, then shear further at constant vol- respond in an undrained manner initially and then
ume. For an important discussion of the relation revert to drained behaviour over a long period of
between strength and volume change see Bolton time. The period involved may be many years; for
(1986). example failures started to occur in cuttings for
The final state of density and shear strength railways in Weald Clay (a dense engineering soil)
under the same vertical stress will be the same for more than fifty years after their construction, as
both initially dense and initially loose states, and is pore pressures gradually re-established equilib-
known as a critical state. The angle of shearing rium conditions and drained failure states were
resistance in such states is referred to as the critical reached in the soil.
state or constant-volume value, and given the sym- The differences in behaviour of small speci-
bol cs, crit or cv. A series of such states at dif- mens of loose and dense soils have implications
ferent vertical stresses defines the critical state line for failure mechanisms in large masses of soil. In
in a plot of shear stress against vertical stress, or, dense soil, small local variations in stress or ini-
under more general stress conditions, of deviator tial density will allow parts of the soil to reach
stress against mean normal stress. Note that each peak strengths first. These parts will then start to
such state also has a void ratio or specific volume get weaker (post-peak) with further shear distor-
associated with it, so that the critical state line may tion, while the soil around is still pre-peak and
also be presented on a plot of void ratio against therefore getting stronger. Further deformation
mean normal stress (see Compression of soils). will tend to concentrate into these weakening
An alternative form of the test would be to pre- zones, which develop into relatively narrow fail-
vent any change in height (and therefore volume) ure ‘planes’ separating and allowing relative
of the specimen, but allow the vertical stress to movement between non-deforming blocks of soil.
GFE-07.qxd 6/24/05 3:01 PM Page 145
In loose soils the strain-softening behaviour in deform laterally and is loaded in a vertical direc-
shear is absent and there is less tendency for tion. The specimen is usually kept saturated, but
deformation to concentrate into such narrow water is allowed to move in or out of the specimen
zones. as it changes volume. The test is normally only
Critical states are reached as the final states in used for clays, as coarse soils are extremely stiff
tests on coarse soils and also with clays of low in one-dimensional compression. The test gives
plasticity. In clays of high plasticity, with a rela- results that may be applied directly to certain
tively high proportion of thin plate-shaped clay engineering problems, for instance the settlement
particles, the strength continues to drop, with of foundations on clay strata. It also elucidates
continuing shear in a thin failure zone causing what has happened to natural sedimentary soils as
particles to be broken or rotated to align along they have formed by deposition, been compressed
the failure plane. The friction angle may then fall by burial under further sediment or ice loading,
to very low values, known as residual values. and then unloaded due to melting of the ice or
Figure 7.5 gives the variation of critical state fric- erosion of overlying ground.
tion angle with the angularity of the grains for As the stress is increased the specimen grad-
coarse soils, while Figure 7.6 shows how the ually decreases in volume as water is squeezed
residual angle of friction varies with clay content out, a process known as consolidation. For most
for fine soils. (For further information on soil soils, the volume is found experimentally to be
strength see 7.4.) related approximately linearly to the logarithm of
Compression of soils 40˚

Just as the shear strength was discussed above in Sands
terms of a simple laboratory shear test, so the Quartz
random
orientation
30˚
compression of soil may be considered in terms Selset
Wiener
of a simple compression test in laboratory appar- Tegel
atus known as an oedometer (Figure 7.7). In this, Jackfield

Quartz
r
20˚ Jari
the soil specimen is confined so that it cannot London
clay
Oxford Walton’s
clay Wood Chlorite
Weser-Elbe Little
10˚
Belt Talc
Biotite
44 Quartz
Feldspar or feldspar quartz uniformly graded
0 20 40 60 80 100
mixtures Clay fraction (2) (%)
42
Well graded
40 Figure 7.6 Variation of the residual angle of friction

for fine soils with clay fraction, compared with charac-
38 teristic friction coefficients for some soil minerals.
36
cv
34
32
30
28
26
V Sub Sub
Ang Rounded Well rounded
ang ang round
0 0.2 0.4 0.6 0.8 1.0 Figure 7.7 Schematic of oedometer apparatus for
measuring the compressibility and swelling behaviour
Figure 7.5 Variation of the critical state angle of fric- of soils in one-dimensional compression (no lateral
tion for coarse soils with grading and particle shape. strain) under incremental loading.
GFE-07.qxd 6/24/05 3:01 PM Page 146
Figure 7.8 Compression and swelling behaviour of soils plotted as specific volume, void ratio or water content
against the effective vertical stress (logarithmic scale). Also shows typical stress paths for specimens subjected to
shearing under constant vertical stress or constant volume conditions.
the applied effective stress (Figure 7.8). If the this pressure to its current pressure is its over-
specimen has not previously experienced a higher consolidation ratio (OCR).
stress, the plot of volume against pressure is Many natural soils are heavily over-consoli-
known as the virgin compression line (VCL). The dated, having been compressed by considerable
slope of this line (the Compression Index Cc) depths of later sediment that has subsequently been
depends on the plasticity of the clay, an empirical eroded, or by great depths of ice which has subse-
relation being: quently melted. Over-consolidation can also be
caused by large suctions developed due to evapora-
Cc 0.007 (LL – 10)
tion from the surface or lowering of the groundwa-
If the specimen is unloaded, it increases in ter table after the soil has been formed. On the
volume but at a very much lower rate than during other hand, recently deposited alluvial soils may
virgin compression. On reloading, it returns still be normally consolidated, except perhaps for a
approximately along this ‘swelling line’ until the crust of over-consolidated material at the surface.
VCL is rejoined and the rate of compression In some situations, such as in the Gulf of Mexico,
increases again. Soil whose state lies on the VCL where clays are being deposited under water at a
(such as point D) is described as normally con- high rate, the dissipation of pore pressure may not
solidated; soil whose state lies below the VCL keep up with the rate of new deposition and the soil
(point A) is said to be over-consolidated. The may be under-consolidated.
highest pressure to which it has been subjected is The critical states (discussed in Shear failure
its pre-consolidation pressure, and the ratio of and critical states, above) in relation to soil strength
GFE-07.qxd 6/24/05 3:01 PM Page 147
Figure 7.9 Simple unified model for soil behaviour, showing stress paths for tests on normally consolidated and
over-consolidated soil specimens. Total stress paths are labelled T, effective stress paths E. The pore pressure U is
the difference in mean stress, positive or negative (suction), between total and effective stress.
tests may also be shown on this plot and are found The concept of OCR is less clear when applied
to give a line parallel to, but somewhat below, the to residual soils that have formed from the in situ
VCL. Note that a soil element starting at a state chemical weathering of rock. However the concept
below this critical state line (CSL), if tested under of a critical state line, which can be defined by
drained conditions at constant vertical stress, must testing, can still apply, and the initial state of the
increase in volume as it moves to the CSL during soil in relation to it will define its general engin-
failure (path AB); while if tested at constant vol- eering behaviour as discussed above.
ume (AC) it must develop increased vertical effec- Unified models of soil behaviour have been
tive stress by a reduction in pore pressure. This is developed based on these concepts. Essentially
consistent with the behaviour observed for ‘ini- they invoke the existence of a yield surface that
tially dense’ specimens; thus ‘dense’ soils are controls the failure of loose soils, a failure surface
essentially those starting at states below the CSL. that defines the peak strength of dense soils, and
Conversely, ‘loose’ soils are those starting at states elastic soil behaviour for soil states within these
above the CSL, so that failure under constant stress bounds (Figure 7.9). Elastic analyses may usefully
involves reduction in volume (DE), and failure at be used in many geotechnical problems when
constant volume requires a reduction in effective imposed stresses are significantly below those
stress and increase in pore pressure (DF). The for- needed to cause yield or failure; Poulos and Davis
mer are relatively heavily over-consolidated, the (1974) provide a compilation of relevant solutions.
latter either normally consolidated or lightly over- The basic models are subject to ongoing research
consolidated (such as point G). and development to incorporate more complex
GFE-07.qxd 6/24/05 3:01 PM Page 148
features of soil behaviour, including the response water can occur, the strength at failure depends
to cyclic loading, unsaturated soils, anisotropy, only on the initial water content or void ratio and
and non-linear and stress-path-dependent elastic is not affected by increases in confining stress,
behaviour. The primary purpose of these models is which are carried by increased pore pressure.
for incorporation into numerical analyses by finite The failure criterion then becomes simply:
element or finite difference methods, which allow
cu
calculation of ground deformations as well as fail-
ure conditions in complex engineering problems. where cu is the undrained strength or cohesion;
the alternative symbol su is sometimes preferred.
In situ stress conditions The undrained strength is measured in the labora-
Initial stress conditions in the ground may often tory by simple, quick and cheap tests, or may be
be of importance, particularly in the design of deduced from in situ tests. The failure condition
retaining structures or tunnels. The total vertical applies in terms of total stresses and allows
stress on an element in the ground may usually be simple calculations to be performed without
taken as being equal to the weight of the ground knowledge of pore pressure conditions, provided
above the element. The vertical effective stress is the problems involve loading or unloading over a
then the total stress less the pore water pressure time period that allows a negligible amount of
within the element. drainage to occur. Such analyses are, therefore,
The horizontal stress is more difficult to estab- usually only applicable to fine-grained soils of
lish, and may be smaller or greater than the verti- low permeability. In situations where the stresses
cal stress. The ratio between the vertical and are increasing, such as beneath a foundation on
horizontal effective stresses is the coefficient of clay, the short-term undrained analysis is usually
earth pressure at rest and given the symbol K0. For the most critical, since the soil will consolidate
normally consolidated soils the value of K0 is with time and become stronger. However when
given approximately by (1 – sin ). During the stresses are reducing, such as adjacent to an
unloading, horizontal stress reduces more slowly excavation, the long-term state will be more crit-
than vertical stress and the value of K0 therefore ical and the more stable undrained conditions will
increases with over-consolidation ratio; an only apply in the short term. This basic fact
approximate value of K0 for over-consolidated explains the large number of (often fatal) acci-
soils is given by: dents involving workers being buried in trenches
due to collapse of the sides.
K0 (1 sin ) √OCR. For normally consolidated soils the undrained
strength increases linearly with the vertical effect-
The state of initial stress in residual soils is dif- ive stress; the rate of increase is related to the
ficult to define, since the soil is produced by plasticity of the soil, an empirical relation due to
weathering processes from rock in which a wide Skempton (1957) being:
range of stresses may have been left by complex
geological processes. Weakening and softening in cu /v 0.11 0.0037 PI
the rock by weathering is likely to reduce hori- The undrained strength of over-consolidated clay
zontal stresses, but conversion of inert material to depends on the stress history of the soil element,
swelling clays may lead to an increase in lateral since it can be in equilibrium at different values of
stress. In addition, the inevitable heterogeneous water content under a particular vertical stress
nature of residual soils is likely to mean that ini- (Figure 7.9: points A and H). A useful empirical
tial stresses are also variable over short distances. relation (Ladd et al., 1977) gives:
(cu/p)over-consolidated (cu /p)normally
Undrained shear strength
consolidated OCR
0.8
When saturated soils are tested or loaded in the
ground under conditions in which no drainage of where p is the mean effective stress.
GFE-07.qxd 6/24/05 3:01 PM Page 149
Properties of rocks stresses between rock surfaces. Discontinuities

Rocks essentially differ from soils in that they also reduce the tensile strength to zero. The vari-
have significant bond strength, so that they can ous factors affecting the behaviour of discontinu-
sustain tensile stress and shear stress under zero ities in rock masses and quantitative methods of
normal stress. The unconfined compression describing them have been drawn up by the
strength of intact material can be very high, over International Society for Rock Mechanics (ISRM,
200 MPa in hard igneous rocks such as basalt or 1981); these are summarised in Appendix A3.
granite; while at the other end, the strengths of Various systems of classification have been
weak shale rocks merge with those of hard clays devised to relate rock mass properties to intact
at a strength around 1.0 MPa (Tables A3.2ac in strength; core recovery during drilling; fracture
Appendix A3). Young’s modulus may similarly spacing, orientation and condition, and ground-
vary from about 1 to 100 GPa. The initial stresses water conditions. These classifications have been
in the ground can be complex, reflecting the geo- related empirically to rock behaviour in various
logical processes that have formed the rock, from situations, such as for slope stability and tunnel
cooling of magma to severe compression and support requirements. A classification for geo-
folding. Horizontal stresses may be very substan- morphic purposes primarily related to slope
tially higher than vertical, and vary in different stability is given in Table A3.13 of Appendix A3.
horizontal orientations.
In many engineering situations the properties
of the rock mass are controlled less by the intact 7.3 Slope instability
rock strength than by the properties of the discon-
tinuities in the rock. These discontinuities (known Stability of soil slopes
as ‘defects’ in the USA and Australia) can range Many geomorphological processes involve failure
from major faults to closely spaced fractures. The of slopes, and slope stability is also a common
spacing, orientation and continuity of these fea- issue in engineering construction. Some applica-
tures must be considered in relation to the nature tions of the soil models introduced above to slope
and scale of the problem. For a large excavation in stability problems are considered in this section,
heavily fractured rock, the ground may essentially and a brief discussion of the stability of rock
be treated as dense coarse soil and calculations slopes. For further coverage of both topics, and of
undertaken using a continuum model of the their applications to slope engineering, see Simons
ground. On the other hand, rock falls into a shal- et al. (2001).
low tunnel will be controlled by a few sets of The stability of slopes may be introduced by
joints that allow discrete blocks to fall or slide consideration of ‘infinite’ slopes, slopes that are
into the open space. The effect of discontinuities sufficiently long that any element of the slope
on the stability of rock slopes is considered may be considered as typical and end effects are
briefly below (7.3 Slope instability). negligible. By considering such a typical element
Where failures are controlled by sliding along (Figure 7.10), it can be shown that:
discontinuities, it is the properties of the discont-
inuities rather than the intact rock that are import- F
tan  w
1
(
1 tan2 

)
ant. Discontinuities may be tightly closed, with tan  (1 tan tan ) 
 
clean rough surfaces, or may be open, or filled
with weaker material. Joints that are open or filled where F is the ratio of available to mobilised soil
with compressible material will obviously reduce strength on an assumed failure plane referred to as
the stiffness of the rock in compression as well as the factor of safety, is the saturated unit weight of
its strength in shear. Beneath the water table, the soil and w the unit weight of water. The slope
water pressures in connected fissures will increase angle is , and defines the direction of seepage
hydrostatically with depth, as in a soil, and the flow within the slope, with the water table at the
concept of effective stress will apply to contact ground surface. There are a number of interesting
GFE-07.qxd 6/24/05 3:01 PM Page 150
loosest and the changes needed to trigger a failure

are most likely to happen.
However, in the more realistic case where the
water table is below the surface, the calculated
factor of safety is found to get smaller the deeper
into the slope the failure surface is assumed to
occur. A similar result is obtained for short-term
(undrained) stability of an infinite slope in fine
soil. It then becomes unlikely that the assumption
of an infinite slope will be reasonable, and real
slips are more likely to occur along approximately
circular failure surfaces. This assumption is made
Figure 7.10 Typical element of an ‘infinite’ slope; the
stability of the slope is determined by considering the
in standard methods of analysing slope stability,
equilibrium of the element along the line of the slope in which a number of potential slip surfaces are
under gravitational forces, water pressures and friction considered and a search made (usually nowadays
forces. by computer) to find the most critical with the
lowest factor of safety.
special cases. For instance, if the slope is dry rather The short-term stability of slopes in clay of uni-
than saturated the expression simplifies to: form undrained strength is conveniently expressed
in terms of a stability number Ns, where:
tan
F Ns cu/F H
tan
or:
The same expression is reached for vertical per-
colation of water into the slope ( 90). The F cu/Ns H
slope is then just stable (F 1.0) when the slope where H is the height of the slope. Charts were
angle is equal to the angle of friction of the soil. produced by Taylor relating values of Ns and
This condition exists in many natural slopes, such slope angle for various values of the depth from
as in talus material, which are freely drained. the base of the slope to the strong layer (Y) (Figure
If the seepage flow is parallel to the slope 7.11). If Y is not restricted and is less than 53,
( ), the expression reduces to: then Ns has a constant value of 0.18. For other con-
tan  w  ditions reference should be made to the charts,
F 1 
tan   which are reproduced in many textbooks (Smith
and Smith, 1998; Whitlow, 1995; Powrie, 1997).
The maximum stable slope angle is approxi- Ns increases for 53 and is 0.25 for a vertical
mately halved. If seepage is out of the slope cut, and decreases for 53o when the depth of
( ), the stable slope angle is further reduced. slip is limited by a firm stratum (Figure 7.11).
Such conditions may well exist in thin soil cover General methods for analysing soil slopes,
on the sides of hills. Thus it is not uncommon for which may be used for undrained total stress analy-
natural slopes with gradients as low as 10–15 to sis or effective stress analysis including pore pres-
have a factor of safety close to unity; even minor sures, are based on dividing the slope into a
excavations in such slopes may trigger further number of (usually vertical) slices (Figure 7.12).
movement of the slope. The equilibrium conditions between these slices
In all these situations the factor of safety (F) is are considered, making simplifying assumptions
independent of the depth assumed for the failure that vary between different methods. Pore pres-
surface. Failure will tend to occur either on a sures within the slope are commonly expressed in
plane of weaker soil, perhaps on a shallow rock terms of the pore pressure ratio ru, which is the
surface or close to the surface where the soil is ratio between the pore pressure and the overburden
GFE-07.qxd 6/24/05 3:01 PM Page 151
(a) Parallel flow, no slope seepage

Water table
H
h

DIPPING STRATA
OR DRAWDOWN
w h
ru cos2 for 0.8 or (H–h) 3 m
H
(b) Horizontal flow, full slope seepage

Water table
H
Kh > Kv h
OR ARTESIAN
w h
ru for 0.8 or (H–h) 3 m
H
(c) Parabolic top full line
table
Water
H
Figure 7.11 Stability of a slope in fine soil in terms Seepage h
of total stresses (undrained behaviour): (A) definition of HOMOGENEOUS
terms and (B) Taylor’s chart showing variation of sta-
bility number with slope geometry.
w h
ru cos for 0.8 or (H–h) 3 m
H
Figure 7.13 Determination of approximate average

values for the pore pressure parameter ru in a slope for
different seepage conditions (after Mitchell, 1993).
pressure at any point. It will normally vary

throughout a slope but in some cases may be con-
sidered as approximately a constant for the slope
(Figure 7.13). In a homogeneous soil, the factor of
safety then varies linearly with ru, and charts have
been produced by Bishop and Morgenstern (1960)
giving stability coefficients m and n as functions of
Figure 7.12 General analysis for slope stability using c/ H, , and D, such that
the method of slices; the equilibrium of a series of slices F m n ru
bounded by a circular failure surface is considered,
making varying assumptions about the forces acting where D is the depth factor R/H (Figure 7.11).
between the slices. W is the weight of the slice; L are the These charts are also reproduced in many stand-
inter-slice forces; l is the length along the failure sur- ard text books (e.g. Smith and Smith, 1998;
face at the base of the slice; N is the normal reaction
force on the base of the slice; T is the tangential force Whitlow, 1995). In many cases they allow prelim-
on the base of the slice and is the inclination to the inary estimates to be made of the stability of a
horizontal of the base of the slice. slope and of its sensitivity to variations in the main
GFE-07.qxd 6/24/05 3:01 PM Page 152
parameters. For example, consider a 15 m high accuracy by assuming a failure mechanism involving
slope in London Clay, with assessed values of three sliding bocks and considering the equilibrium
and c of 25 and 7 kPa respectively, unit weight of the central block when acted on by gravity, normal
19 kN/m3, and a pore pressure regime approximat- and shear forces on the base of the block, a driving
ing to ru 0.3. It is required to determine what is force from the upper wedge and a resisting force
the steepest angle at which this slope will be stable, from the lower wedge.
the factor of safety when the slope angle is reduced
to 1 : 3, and the factor of safety for the latter slope Stability of rock slopes
if the effective cohesion is reduced to zero by stress Except in the case of very heavily fractured rock,
relief and weathering. In the first two cases failure of rock slopes will be dominated by
c/ H 0.025 approximately. On Bishop and discontinuities, with the failing mass sliding or
Morgenstern’s chart for parameter n, a series of falling as a block or collection of blocks: there are
dotted lines show that, provided the slope is steeper four basic mechanisms, as shown in Figure 7.15.
than about 1 in 4 (for 25), the failure mech- Planar and wedge failures depend on the shear
anism does not extend below the base of the slope strength along the discontinuities, toppling depends
(D H) and values of m and n for different slope purely on block geometry, while flexural toppling
angles are read off as in the following table: involves tensile failure in bending followed by top-
pling. The interrelation of sliding and toppling fail-
Slope angle 1:4 1:3.5 1:3 1:2.5 1:2 ures is shown in Figure 7.16 for the simple cases
Parameter m 2.39 2.13 1.87 1.61 1.36 with joint planes parallel to the slope.
Parameter n 2.15 1.91 1.69 1.38 1.29 More complex cases, including wedge failures,
Factor of 1.745 1.557 1.363 1.196 0.973 involve three-dimensional interactions between dis-
safety continuity and slope planes. Stereographic projec-
tions provide a useful method of representing
From these values it can be seen that a safety factor three-dimensional information about discontinu-
of 1.0 (just stable) is achieved with a slope of about ities in two dimensions and may also be used to
1 : 2.1 or 25.5; the factor of safety is increased to determine likely slope instabilities (see Hudson,
1.36 for a slope of 1 : 3 or 18. If c 0, the chart 1989 for an introduction and Hoek and Bray, 1981
for c/ H 0 gives m 1.40, n 1.55 and hence for further information). Limit equilibrium methods
the factor of safety is 0.935 and the slope becomes
unstable.
Slip plane Discontinuity
In many cases geological or other restraints Discontinuity Slip plane
t
may result in slope failures that approximate to
neither the infinite slope nor a circular slip (Figure
7.14). These may often be analysed with sufficient
or void
Toppling failure Flexural toppling
Figure 7.14 Examples of two cases where stability of
a slope may be analysed by considering the equilibrium Figure 7.15 The four basic failure mechanisms of rock
of three sliding blocks, where neither an infinite slope slopes, involving either sliding on one or two slip planes
nor circular failure surface analysis is appropriate. (discontinuities) or toppling (from Hudson, 1989).
GFE-07.qxd 6/24/05 3:01 PM Page 153
in Table 7.3. Detailed discussion of the processes

involved in the formation of the soils and the
resulting nature of the deposits is provided in
subsequent chapters.
Effects of climate and groundwater

conditions
The different climatic and groundwater regimes
affecting geomorphological processes have been
summarised in Chapter 2, and are discussed further
in relation to different specific environments in
later chapters. This section provides a brief sum-
mary of the main influences of climate and
groundwater conditions on the formation and
nature of soils in relation to their likely engineering
properties and the geotechnical problems that may
arise from them.
Where annual evapotranspiration exceeds
infiltration, water tables are generally low, con-
trolled by local stream levels and the presence of
more permeable strata. In areas without tree
cover, upward moisture movement and evapor-
ation from the ground surface may produce a
chemically active zone at ground level, leading to
cementation and the formation of duricrust.
Surface soils will be partially saturated, some-
times to great depth. Soils may have been sub-
jected to desiccation over a long period of time.
Plastic clays will then tend to heave on wetting or
when evaporation is prevented by sealing of the
ground surface, for example by construction of
new buildings; on the other hand, porous residual
Figure 7.16 The relationship between the stability or soils, weathered sands or loess may collapse. Soil
failure of a rock slope in sliding or toppling mode and suctions can facilitate construction of temporary
the geometry of a typical block (as defined by sets of
discontinuities) and the coefficient of friction at the steep slopes, excavations and shafts. However
base of the block () (from Hoek and Bray 1981). long-term protection of slopes from erosion by
vegetation may not be practical.
similar to those used for soil slopes may then be Where annual infiltration exceeds evapotrans-
used to analyse rock slopes and determine the rein- piration, groundwater flow will be downward from
forcement needed to stabilise the slope (Hoek and the surface, soils will be mainly saturated and the
Bray, 1981; Simons et al., 2001). water table is likely to be high in fine-grained
soils, even in slopes. Excavations and under-
7.4 Geotechnical problems associated ground works will generally suffer from water
with different soil types inflows, while dewatering may lead to settlements
due to the consolidation of fine soils or internal
Soil types according to origin erosion and the loss of fines from coarse soils.
Soil types and their related problems are pre- Erosion protection of slopes by vegetation is
sented below in relation to their origin, as set out usually practical.
GFE-07.qxd 6/24/05 3:01 PM Page 154
Table 7.3 Soil types according to origin.
Classification Formation Nature of deposit examples Chapters
Residual
Residual Formed in place by Type depends mainly on weathering process, less 15, 17,
chemical weathering on parent material. Pronounced structure, 18, 23
of parent rock or soil usually bonded, sometimes porous.
Heterogeneous and of variable depth.
Transitional with parent material
Evaporites Salts from saline water Carbonate oolites precipitated from sea water. 16
by precipitation or Gypsum commonly precipitated from
evaporation sulphate-rich desert playa lakes, often forms
cemented soil and hard sub-surface crust
Organic Formed in place by Wholly organic fibrous and amorphous peat.
growth and decay of Organic silts and clays
vegetation
Transported and deposited
Aeolian Transportation and Usually silt and fine sand with uniform grading. 16, 25
deposition by wind Loess often contains vertical cracks, joints
and root holes
Alluvial Transportation and Fine clay to coarse gravel. Coarse particles 10, 11,
deposition in water usually rounded. Soils usually sorted and 15, 16,
often show pronounced stratification 19
Colluvial Transport by gravity, Local in origin. Hillside creep, downwash and 14, 15,
freeze–thaw and solifluction deposits. Variable grading from 17, 26
water clay to fine gravel. Often termed ‘head’
Estuarine Transport and Fine clay to cobbles. Coarse particles usually 20
and coastal deposition by rounded. Soft clays and silts in estuaries; sand
water and wind in coastal dunes; sand, gravel and cobbles on
beaches. Carbonate deposits
Glacial Transportation and Tills and moraines, usually heterogeneous and 13, 19
deposition from can have a wide range of grading. Outwash
ice or meltwater material becomes finer with distance from
from ice meltwater source. Finer material usually
laminated and varved (glacial lake deposits)
Marine Transportation and Mainly clays and silts, some sand. Clays may be 22
deposition by very deep, and initially very soft, possibly
water; also aeolian under-consolidated or sensitive and liable to
material deposited liquefaction. Outwash material from rivers
in water becomes finer with distance from coast.
Biogenic materials may predominate in deep
water
Periglacial Heavily influenced Deep colluvial deposits. Ground contorted by 14
by freeze–thaw freeze–thaw. Parent soils and rocks fractured,
processes brecciated and cementing destroyed.
Cambering and valley bulging
Taluvial Transportation by Local in origin. Landslide debris, screes and 8, 13, 19
gravity coarse variety of colluvium. Heterogeneous.
Widely variable grading
GFE-07.qxd 6/24/05 3:01 PM Page 155
Table 7.3 (Continued).
Classification Formation Nature of deposit examples Chapters
Volcanic Ash and pumice Silt size particles with larger volcanic debris. 23
deposited during Highly angular particles, often vesicular.
volcanic eruptions Weathering often produces highly plastic clay
Man-made
Fill Deposited by man Loose dumped, hydraulically placed or 27
compacted. Old fills often heterogeneous and
may contain organic and toxic material, and
voids associated with human artefacts
Evapotranspiration and infiltration rates often soil. Development of the root system as a tree
vary seasonally and from year to year; they may grows will tend to cause shrinkage of clays and
also vary on a much shorter time scale, owing for settlement of shallow foundations. Conversely,
example to individual storms. Surface condi- clearance of vegetation leads to the swelling of
tions, water contents of fine-grained soils and clays, heave of shallow foundations and a loss of
their related engineering properties will vary soil strength in slopes, while decay of the root
accordingly. Conditions during construction may system removes the reinforcing effect; surface
differ significantly from those found during site erosion and slope instability may follow.
investigation. Seasonal variations may lead to complex
Seasonal variations will lead to heave and groundwater conditions, with different pressures
settlement of superficial clays and shallow foun- in permeable strata separated by less permeable
dations; the effects on new structures will depend strata. The permeability of superficial strata may
on the conditions prior to construction. The be strongly influenced by for example structure
undrained strength of superficial clays will vary and stress relief. Faults may form impermeable
and affect the behaviour of embankment founda- features, or provide zones of high permeability.
tions, and also the suitability of the soils as fill Adverse ground water flows may affect tunnels,
materials. Earthmoving of fine-grained soils shafts and pile bores, while dewatered excavations
becomes very difficult when rainfall exceeds may be affected by high inflows, base uplift or
evapotranspiration; seasonal variations will, ‘quicksand’ conditions. Temporary or permanent
therefore, control the season for earthmoving and dewatering may lead to ground settlement.
short-term variations will control working time. Artesian groundwater pressure is common
These variations depend on temperature, wind below the base of valleys in layered strata or
and rainfall duration, rather than simply rainfall below glacial till on the lower slopes of valleys,
amount. Problems are usually worst in temperate and may occur within alluvium near to slopes. It
climates and are often strongly influenced by ele- exacerbates all problems involving uplift below
vation of the site. At the other extreme, supplies structures and excavations, and causes severe
of water for treating fills may vary seasonally and problems with boreholes, pile bores and anchors
be in short supply during hot dry spells. Severe if the flow is significant.
cyclonic rainfall is a major cause of erosion and
failure of natural and man-made slopes. Common geotechnical problems
The presence of vegetation increases transpir- associated with different soil types
ation from soil and causes an increase in soil A checklist of the principal geotechnical prob-
suction and a decrease in water content; root sys- lems associated with different soil types is pro-
tems may provide structural reinforcement of the vided in Table 7.4. In discussing the properties
GFE-07.qxd
Table 7.4 Geotechnical problems associated with different soil types (adapted and extended from Fookes, 1997).
156
Soil type Investigation Properties/characteristics Applications
GEOMORPHOLOGY FOR ENGINEERS
6/24/05
Alluvial and other Typically strength decreases and compressibility
sedimentary soils increases with increasing fines content.
(These soils are
conveniently listed
3:01 PM
as below according
to dominant
particle size.)
Page 156
Gravels Difficult to investigate Densities vary widely according to deposition. Generally good foundation, full consolidation
by boring. Difficult to Strength high and compressibility low. occurs during construction. Large flows into
obtain representative Permeabilty variable, depending on grading excavations. Generally good fill material;
samples from below and packing; often very high and can only be single size gravels are self-compacting when
water table as fines determined by pumping tests. Can be deposited in water. However fills which are
are washed out during gap-graded with voids only partially filled ‘choked’ with silt and clay matrix material can
drilling and sampling. with fines, which may then migrate if be difficult to compact when wet, and may not
Thin layers of other hydraulic gradient increased. lose water readily under gravity drainage.
soils may not be
detected.
Sands and silts Often weakly bonded Properties usually improve with geological age. Generally good foundation, consolidation during
or with interlocking Densities vary widely according to deposition. construction. In excavations, beware of base
grains, but effects lost Loose sands and silts very susceptible to failures by piping, loss of soil through sheet
on disturbance and liquefaction during earthquakes, can develop pile clutches etc. Danger of soil and water
difficult to detect. Very flow slides, large settlements when subject to inflows into tunnels, high abrasion of
difficult to obtain vibration. Permeability moderate to high. Very tunnelling machines. Generally good fill
undisturbed samples, erodible, piping risks when subject to internal material, but dry single size sands and silts
ground below boreholes water flow. Surface erosion by water flow and have poor trafficking characteristics. Silty soils
often disturbed by water wind. Possibility of collapse of dry sands on prone to rapid deterioration and poor
inflow during drilling wetting, particularly if weathered. Silts are trafficking during wet weather, and may not
below the water table. subject to frost heave. lose water under drainage by gravity. Silty soils
SPT results may be low may ‘bounce’ when trafficked, creating difficul-
due to soil disturbance ties in forming graded surfaces prior to laying
or, in silt, high due to road bases etc., and increasing fuel
pore pressure effects. consumption of plant. If loose dumped, moist
GFE-07.qxd
Thin layers of other soil sands and silts may have low density, and
soil may not be detected. collapse may occur on inundation.
Alluvial clays Measured undrained Properties vary widely with mineralogy and For saturated clays, short-term undrained strength
6/24/05
(General properties: strength different grading (proportion of silt and sand). Drained governs stability in loading cases (foundations).
significant from different types strength decreases and compressibility Long-term drained strength governs unloading
differences between of test. increases with increasing clay content and cases (excavations). Little consolidation or
over-, normally- and plasticity. In clays with more than 30–40% swelling during construction except near
3:01 PM
under-consolidated platy clay minerals present, shearing produces drainage boundaries; post-construction consoli-
clays listed below.) discontinuities of low residual strength. dation and swelling occurs.
Permeability low unless sand or silt layers are
present. Strength likely to be anisotropic.
Page 157
Density and in situ strength depend on stress
history due to burial, desiccation etc.
Normally-consolidated Sensitive clays difficult Low undrained strength, high compressibility Low allowable loading pressures under structures
‘marine’ clays to sample without and secondary compression (creep), which and embankments. Large post-construction
(Current effective undue disturbance, increase with plasticity. Exposed surfaces settlements. Base heave and failure in strutted
stress is maximum need to use thin- usually over-consolidated by desiccation. excavations, and high strut loads. Down-drag
to which soil has walled piston on piles. Low strength and difficult working
been subjected.) sampling. In situ vane conditions for plant during excavation.
tests useful; allow
ENGINEERING BEHAVIOUR OF SOILS AND ROCKS

measurement of loss
of strength on
remoulding
(sensitivity). Cone
testing good for
locating thin layers of
silt and sand.
Over-consolidated Test results may be If effective stresses due to engineering work Mass strength of foundations often affected by
‘marine’ clays affected by sampling exceed pre-consolidation pressure, behaviour fissuring. Pre-existing shear surfaces, particu-
(Current effective disturbance, presence reverts to that of normally consolidated clay; larly in highly plastic clays, may control
stress is less than of fissures etc. In otherwise compressibility much less. stability of slopes. Probable high in situ
previous maximum — heavily over- Undrained strengths higher but difficult to horizontal stress in heavily over-consolidated
the pre-consolidation consolidated soils predict. Permeability may be controlled by clays, large horizontal movements during and
pressure) rotary coring better flow through fissures. after excavation, high lateral stresses on buried
than push-sampling. structures.
157
GFE-07.qxd
158
6/24/05
SPT tests often useful
back-up to laboratory
testing.
3:01 PM
Under-consolidated Very difficult due to Excess pore pressures still present. Very low Ground surface still settling. Stability of
‘marine’ clays very low strength. undrained strength and high compressibility excavated slopes may be controlled by
(Not yet fully com- relative to depth. undrained strength.
pressed under
Page 158
current stresses, due
to rapid deposition
or recent additional
loading)
Clay fill Bonding of parent clay usually destroyed by Traffickability and height of undrained slope
excavation and fill placing. Laminated and construction controlled by remoulded
varved clays usually mixed, and produce a undrained strength. Degree of saturation
clay of low permeability. Low-plasticity clays important. Strength and stability of fill slope
lose significant undrained strength with small usually governed by long-term conditions
increases in water content in wet weather. (drained strength plus pore pressures due to
Rutting of wet plastic clay fills under rainfall infiltration etc.). Deep-seated failure
trafficking causes shear surfaces that reduce may be governed by short-term undrained
bulk strength. Smooth drum rollers can cause strength. Clay haul road surfaces become too
horizontal surfaces of low strength. slippery for trafficking in rain. Early
compaction and profiling of clay important to
seal surface and reduce infiltration and
softening by rain. Drying of clay prior to
placement requires reliable warm
dry weather.
Colluvial soils Soils usually thin and Formation often involves shearing. Low strength Problems for embankment foundations on
subject to seasonal shears may be present, often continuous and sidelong ground and in excavations.
water content change. at base of soil, on slopes currently too flat for
Time of investigation slope movement. Properties may differ from
and of engineering those measured at time of site investigation
construction can be due to seasonal variations.
important.
GFE-07.qxd
Talluvial materials Landslides are often at limiting equilibrium, on Development of equilibrium slope in clays may
major shear surfaces. Strata disturbed in take many decades. Instability due to toe
massive landslides; open fissures and fracture erosion may continue long after erosion
6/24/05
zones formed. Soil more porous and wetter prevented. Excavation and filling likely to
than parent soil. Perched water table effects. initiate new movements. Screes may develop
avalanche/flow slide behaviour when disturbed.
Hot desert soils Generally have little or no fines. Usually granular, Soil may be highly erodible once thin protective
3:01 PM
uniformly graded and of low density when pavement removed or disturbed. Duricrusts or
wind blown or coastal. Often coarse, well- densely packed boulders in wadis may cause
graded and with angular particle shape when excavation difficulty. Aggregate may be in
deposited by ephemeral spate flow in wadis or short supply or contaminated by salts. Problems
Page 159
fans. Water table near to the surface leads to of sediment movement by water in sudden
precipitation of evaporite salts. storms and by wind. Deposits of wind-blown
sand liable to collapse on wetting. Salty ground
highly aggressive to structures and road
pavements.
Glacial soils Often variable and Often variable and heterogeneous, horizontally Problems with water-deposited soils as for
heterogeneous, and vertically, with complex groundwater alluvial soils. Tills generally good foundation
horizontally and conditions, including artesian. Density and material. Boulders cause problems in piling,
vertically. Original strength of tills depend mainly on method of tunnelling, excavation and filling. Drag

land forms obscured — deposition, not on stress history as for alluvial structures at base of tills on weak rocks etc.
buried valleys etc. soils. Grading curve may be almost straight line cause errors in rock level estimation, problems
High and low over wide range of particle size. Nature of fine with piles etc. Water flows into excavations and
permeability strata matrix material and presence of clay minerals stability of slopes in short term may be heavily
give complex control properties. Dense materials may be very influenced by layers or lenses of high
groundwater strong and stiff. Often contain local inclusions permeability. Problems with fills as for alluvial
conditions. Severe of water-deposited laminated sands silts and soils.
artesian pressure clays. Often have discontinuities as for alluvial
conditions often exist clays.
below tills which
mantle valley slopes.
In tills, boulders cause
problems with drilling.
Interface with rock
may be difficult to
determine.
159
GFE-07.qxd
160
6/24/05
Periglacial conditions Valley slopes will be Permafrost in active conditions. In relic conditions, Effects of cambering and valley bulging may
affected by strata past ground freezing is likely to have produced cause problems with foundations, excavations
disturbance and extensive colluvium deposits on slopes. Ground and tunnels.
shearing due to valley contorted by freeze–thaw features e.g.
3:01 PM
bulging and cambering. cryoturbation, frost wedges, pipes. Parent soils
Near-vertical fissures and rocks will have been fractured, brecciated
and gulls difficult to and un-cemented by ground freezing. Such
locate by conventional effects may be present generally or locally.
Page 160
drilling.
Organic soils Most methods of Highly compressible and subject to severe long- Very large settlements of foundations and
sampling and testing term creep. Often of very low unit weight. embankments underlain by organic soils,
not suitable for highly Methane gas may be present. requiring light weight fills or pre-compression
organic materials, by surcharging. Problems in slope stability due
particularly when to low passive resistance. Non-saturated peat
fibrous. deposits may float when flooded. Wastage of
Compressibility peats when exposed and subject to drying.
usually more important Usually difficult to run plant on and to handle
than strength, best as spoil.
measured in large
Rowe cells or, for
shallow deposits, by
large scale loading
tests — plate bearing,
skip test, trial
embankment etc.
Volcanic soils Layered and complex Properties significantly different from those of Low particle density makes earthwork fills very
deposition. Old sedimentary soils due to porosity and susceptible to erosion. May soften with
weathering surfaces crushability of silt and sand sized particles: compaction, are easily damaged by earth-
and residual soil often in situ moisture content higher than usual; moving machinery, leading to loss of
covered with fresh greater reduction in strength, but smaller trafficability and sometimes flow. Drying
deposits. Thin ash reduction in compressibility, with increase in produces non-reversible improvement, addition
layers, weathered to stress; no clear peak in compaction curve. of quicklime effective.
clay, within other Fine soils often of high plasticity (with
GFE-07.qxd
deposits. Investigation smectite, allophane), but strength higher and
by cone testing less dependent on plasticity than with
shows behaviour sedimentary soils.
6/24/05
for both loose
and dense soils
to be similar to that
for loose quartz sand,
3:01 PM
due to crushability of
soil grains.
Residual soils (see Difficult to obtain Wide ranging of grading, plasticity, mineralogy, Core stones within weathered profile cause
Fookes, 1997, ‘undisturbed’ samples etc., depending primarily on weathering problems in piling, can influence excavation
Page 161
Appendix 5, for without destroying processes. Grading often depends on methods in open cut. Relic low-strength
further details) structure. Core stones unweathered quartz particles present. discontinuities can cause slope instability if of
(unweathered rock) Recementation of soils may occur. ‘Black’ soils, critical extent and angle. Soils may be readily
within weathered usually formed with poor drainage have high eroded by water, causing severe gullying during
profile cause problems plasticity. Expansive soils with large volume heavy rain.
in drilling. May be changes on wetting and drying. ‘Red’ soils, Rapid consolidation during construction.
difficult to establish usually formed with good drainage, but with Porous soils existing in dry conditions may
rockhead, which varies pronounced structure due to weathering process. collapse on wetting. Porous soils with high
rapidly. Relic Engineering properties depend more on degree of saturation become weak with
discontinuities from structure than on grading, mineralogy etc. reworking and cause problems with operating

parent rock (often Usually behave as if bonded, with structure plant during excavation, forming roads etc.
containing iron and yielding at a certain stress level. Strength and Strength due to structure lost in fills, then
manganese salts) give compressibility depend on this yield as much as strength and compressibility depend on density
planes of low-drained on density. Mineralogy and properties can be achieved by compaction, which depends on
strength, very difficult changed by drying. Structure usually gives high water content at source. Compaction and loss
to locate during site in situ permeability. Porous soils with a high of structure usually gives substantial reduction
investigation. degree of saturation may be sensitive, giving in permeability. General properties of fill are
low undrained strength on remoulding and similar to those of alluvial clays of similar
destruction of structure; most likely in soils grading and mineralogy, but mineralogy often
from volcanic rocks. differs from that of sedimentary clays.
Man-made fills Old, non-engineered Fills will generally be loose and compressible. Major costs may be involved in rendering the site
fills likely to be hetero- Ground and groundwater may be toxic. environmentally acceptable for development.
geneous and diffi- Methane gas may be generated. Hydraulic fills Likely to be very poor foundation conditions,
cult to investigate. likely to be of low density, subject to requiring piles or ground improvement. Special
Site history is of great liquefaction if of appropriate particle size. precautions with boreholes and piling to
161
importance. Usually Difficult to characterise in engineering terms. prevent connecting polluted and clean aquifers.
GFE-07.qxd
162
6/24/05
investigated by trial Infilled quarries likely to have had steep sides,
pits, with simple giving rapid transition between natural ground
in situ tests or as for and deep fill material. Waste materials such
3:01 PM
organic soils. May as mine tailings and pulverised fuel ash may
contain toxic and be useful as engineered fills, particularly
organic materials, where of relatively low weight.
materials subject to
Page 162
decay, voids associated
with human artefacts,
remains of old con-
structions, sludge
lagoons from sewage,
agricultural and indus-
trial processes,
obscured by sub-
sequent filling. The
original topography is
obscured by tipping.
Compressible and
weak alluvial and
organic soils
associated with
streams and ponds may
be buried locally. Back
filling of old mineral
workings may involve
overburden and waste
material similar to
original geological
strata and may be
difficult to detect from
boring.
GFE-07.qxd 6/24/05 3:01 PM Page 163
and associated problems of clay soils in particu- in exposed excavation faces can cause internal
lar, consideration must always be given to the and surface erosion, swelling of clay layers and
influence of soil structure or fabric, which may rapid deterioration; similar effects can occur in
have a profound influence on the behaviour of the borings for piles and piers. Laminated clays are
soil and make it significantly different from that easily sampled in borings, but clay layers tend to
of a remoulded soil of the same mineralogy and swell immediately after sampling, taking in water
water content. Some aspects of the structure are from sand and silt layers. Incorrect water contents
fissures, joints and sheared surfaces; bonding; and undrained strengths will then be measured.
sensitivity; dispersion, and laminations or varves Smear of borehole walls by clay during drilling
of sand and silt with clay. influences the results of in situ permeability tests
Fissures, joints and sheared surfaces are fre- and reduces the efficiency of vertical drains.
quently present, particularly in over-consolidated
clays. Joints and fissures in low plasticity clays
may have coatings of higher-plasticity clay. Bulk 7.5 Typical values of soil
clay strength is reduced, particularly by pre- properties
sheared surfaces in plastic clays of low residual
strength. Conversely, permeability may be This section provides a selection of typical values
increased, particularly if the joints and fissures of some basic and derived properties of soils.
are opened by stress relief during excavation.
Some bonding, mainly due to cementation Specific gravity and density
between clay particles, is often present, varying The full range of specific gravity for soil minerals
from very slight to significant (forming weak is from about 2.3 for gypsum to around 5 for mag-
mudstone). Bonding is usually destroyed by netite. However for most of the commonest com-
weathering and also by mechanical disturbance. ponents of soil it falls within a narrow range. The
Bonding improves intact strength and reduces specific gravity of quartz, and hence of the par-
compressibility, but may lead to sensitivity, ticles of most coarse-grained soils, is 2.65 to 2.66.
or loss of undrained strength on remoulding The specific gravity of clay minerals is more vari-
and large settlements following yield. It may able, but typically in the range 2.55 to 2.75, while
allow fissures to remain open, giving higher bulk for calcareous sand it is between 2.64 and 2.73.
permeability and accelerated swelling and As a first approximation, a value of 2.65 is often
consolidation. assumed.
Sensitivity is usual in normally consolidated Some typical values of the bulk density of nat-
clays and can be severe in ‘quick’ clays, usually ural soils and compacted fills are given in Table 7.5.
due to bonding or to leaching of low-plasticity Densities of around 2000 to 2100 kg/m3 are typical
marine clay by fresh seepage water. It can also of dense soils, and 1700 to 1800 kg/m3 for loose
occur in residual soil. soils. Much lower densities are sometimes found in
Inter-particle attractive forces usually give highly organic soils and in soils with hollow or
clays resistance to erosion by water. In dispersive porous particles such as volcanic ash or pulverised
clays such forces do not exist, making them fuel ash (PFA).
highly erodible both in situ and if used as fills.
Laminated and varved clays are highly Permeability
anisotropic in strength and permeability, with Typical ranges of the permeability of soils are
drained strength along the layers controlled by the given in Table 7.6; note that the value can vary
most plastic layer present. High horizontal perme- over ten orders of magnitude. Permeability may
ability allows rapid consolidation or swelling, giv- be very significantly affected by the structure of
ing rapid strength gain under loading. However the soil. For example, it may be greatly increased
the short-term undrained strength cannot be relied by fissuring in otherwise very impermeable clay,
on for temporary works. Water from sand layers particularly around excavations or tunnels where
GFE-07.qxd 6/24/05 3:01 PM Page 164
Table 7.5 Typical values of density and moisture content (after Seed et al., 1962).
Material Natural densities BS/AASHTO compaction

Bulk density* Dry density Dry density Optimum
(kg/m3) (kg/m3) (kg/m3) m/c (%)
Sands and gravels: very loose 1700–1800 1300–1400 – –

loose 1800–1900 1400–1500 – –
medium dense 1900–2100 1500–1800 – –
dense 2000–2200 1700–2000 – –
very dense 2200–2300 2000–2200 – –
Poorly-graded sands 1700–1900 1300–1500 1500–1800 0–7
Well-graded sands 1800–2300 1400–2200 1700–2100 0–8
Well-graded sand/gravel mixtures 1900–2300 1500–2200 1800–2200 5–10
Clays: unconsolidated muds 1600–1700 900–1100 – –
soft, open-structured 1700–1900 1100–1400 – –
typical, normally consolidated 1800–2200 1300–1900 – –
tills (boulder clays) 2000–2400 1700–2200 – –
Compared sandy clays – – 1800–2200 15–30
Tropical red clays 1700–2100 1300–1800 1400–2100 20–40
* Assumes saturated or nearly saturated soil.
Table 7.6 Typical permeability values (after Casagrande and Fadum, 1940).
 Homogeneous Silts, fine sands, silty Clean sands, Clean

 clays below the sands, glacial till, sand and gravel gravels
 zone of stratified clays mixtures
Soil types  weathering
 Fissured and weathered
 clays and clays modified
 by the effects of vegetation
 1011 1010 109 108 107 106 105 104 103 102 101 1

 m/sec
 109 108 107 106 105 104 103 102 101 1 10 100

Coefficient 
of permeability  cm/sec
(log scale)  1010 109 108 107 106 105 104 103 102 101 1


 ft/sec
 Practically
 Impermeable Very low Low Medium High
Drainage
conditions { practically
impermeable Poor Good
Estimation of coefficient of permeability: for granular soils, the coefficient of permeability can be estimated
using Hazen’s formula:
k c1D210
where k is the coefficient of permeability in m/s, D10 is the effective particle size in mm, and c1 is a factor varying
between 100 and 150.
GFE-07.qxd 6/24/05 3:01 PM Page 165
Figure 7.17 Typical ranges of coefficient of permeability (k) for different types and conditions of rock.
the reduction in stress allows the fissures to open. Table 7.7 Typical values of shearing resistance of
It may also be highly anisotropic: old root holes cohesionless soils.
can provide vertical drainage paths in clay, so that
Material (degrees)
the bulk vertical permeability is greater than the
horizontal, while sand and silt laminations may Loose Dense
make the horizontal permeability much greater
Uniform sand, round grains 27 34
than the vertical. In heterogeneous materials like Well-graded sand,
glacial tills, lenses of sand and gravel provide angular grains 33 45
zones of high permeability, the consequences of Sandy gravels 35 50
which depend critically on the extent of the lenses Silty sand 27–33 30–34
and whether they are connected to a water source Inorganic silt 27–30 30–35
such as a river.
Except in porous rocks, such as sandstone and
chalk, the permeability of rock is largely con- SPTs are also useful for providing a measure of
trolled by the flow of water through joints and the undrained strength of over-consolidated clays,
fissures. Typical values for intact and fractured or using the relation:
weathered materials are shown in Figure 7.17.
cu f1 N
Strength
Ranges of values for unconfined compressive with f1 as given by the plot in Figure 7.21.
strength (UCS) of rocks and clays are given in
Tables A3.2a–c in Appendix A3. Note that for Compressibility and swelling
clays the UCS is double the undrained strength in potential
compression. The compressibility of clays under load and their
Typical values of the angle of shearing resist- tendency to shrink on drying or swell on wetting
ance () for coarse soils are given in Table 7.7 or unloading are both related to the plasticity of
and Figure 7.18, and the variation of with plas- the clay and its natural water content. Thus clays
ticity index for clays in Figure 7.19. Relations of high plasticity are more compressible than
between the value of , relative density and the those of similar strength but of low plasticity,
corrected blow count (N) from standard penetra- while any particular clay becomes less compress-
tion tests (SPTs) in sand are given Figure 7.20. ible as its void ratio decreases. Typical values for
GFE-07.qxd 6/24/05 3:01 PM Page 166
50
Angle of shearing resistance (deg.)
45
40
relative density 100% GP GW
SW
35 75%
ML
SM and SP material type
50%
30 25% range
0
25
20
1.2 1.4 1.6 1.8 2.0 2.2 2.4
Dry density (t/m3)
Figure 7.18 Variation of angle of shearing resistance with density and soil type: ML:silt of low plasticity; SM:
silty sand; SP: poorly graded (uniform) sand; SW: well-graded sand; GP: poorly graded gravel and GW: well-graded
gravel.
Figure 7.19 Variation of angle of shearing resistance (expressed as sin ) with the plasticity index of a soil (after
Kenney, 1959).
natural soils are given in Table 7.8. The coeffi- the potential for swelling being related to plastic-
cient of compressibility is the ratio of compres- ity index as shown in Table 7.9.
sive strain to applied stress for one-dimensional
compression (i.e. with no lateral strain). On dry- Bearing capacity
ing, wet plastic clays tend to shrink and crack; Typical values of allowable bearing pressure for
conversely dry clay which is wetted tends to swell, a range of soils and rocks are given in Table 7.10.
GFE-07.qxd 6/24/05 3:01 PM Page 167
Table 7.8 Typical compressibility values and descriptive terms used.
Type of clay Descriptive term Coefficient of compressibility, mv

(m2/MN) (ft2/ton)
Heavy over-consolidated tills, stiff weathered rocks Very low

(e.g. weathered mudstone) and hard clays compressibility 0.05 0.005
Tills, marls, very stiff clays and stiff tropical Low 0.05–0.1 0.005–0.01
red clays compressibility
Firm clays, glacial outwash clays, lake deposits, Medium 0.1–0.3 0.01–0.03
weathered marls, firm tills, normally consolidated compressibility
clays at depth and firm tropical red clays
Normally consolidated alluvial clays such as High 0.3–1.5 0.03–0.15
estuarine and delta deposits, and sensitive clays compressibility
Highly organic alluvial clays and peats Very high 1.5 0.15
V loose Dr% Table 7.9 Typical values of plasticity index and inher-
15 35 50 65 85 100 ent swelling capacity.
Loose Medium Dense Very dense Plasticity index (%) Inherent swelling capacity
0
0–15 Low
Peck, Hanson
10–35 Medium
and Thornburn (1953) 20–55 High
20
35 Very high
10
(N1) 60
40
8
N
C
OC
60
6
f1(kN/m2)
R
3
4
OC
80
R
2
10
0
28 30 32 34 36 38 40 42 44 46 0 10 20 30 40 50 60 70
Pl %
Angle of internal friction, (degrees)
Figure 7.20 Determination of the density and angle Figure 7.21 Plot showing the relation between the
of internal friction of coarse soils from SPT blow count factor f1 and the plasticity index (PI) of a fine soil; f1
(N); NC refers to normally consolidated soils, OCR is relates the undrained strength of the soil to the SPT
over-consolidation ratio (see Peck et al., 1953). blow count (N) (from Stroud and Butler 1975).
Note that for clays the allowable bearing pressure 7.6 Closing remarks
is about twice the undrained strength of the clay.
A good general reference for the design and It should always be remembered that the influence
construction of foundations is Tomlinson (2001). of water on the behaviour of soils is profound.
For more extensive information on foundations on Damaging effects may be due to erosion or scour
rock, see Wyllie (1999). either at the surface or underground, inflow of water
GFE-07.qxd 6/24/05 3:01 PM Page 168
Table 7.10 Typical values of allowable bearing pressures (after Carter, 1983).
Type of bearing material Allowable bearing

pressure (kN/m2)
 Massive hard crystalline igneous and metamorphic rocks 6000–10 000

 Massive hard crystalline limestones; thoroughly cemented 4000–6000
 sandstones and conglomerates
Rocks
 Unweathered schists and slates 3000–4000

 Hard shales and mudstones, moderate and weakly cemented
 sandstones; hard unweathered marl or chalk 1500–2500
 Weathered and broken bedrock; clayey shales and soft mudstones 800–1200
 Hard clays; cohesive strength 300 kN/m2 600

 Very stiff clays; cohesive strength 150–300 kN/m2 300–600
Clays
 Stiff clays; cohesive strength 75–150 kN/m2 150–300

 Firm clays; cohesive strength 35–75 kN/m 70–150
 Soft and very soft clays: cohesive strength 35 kN/m2 Negligible
 Very dense sands and gravels; SPT N-value 50 400


Sands
Dense sands and gravels; SPT N-value 30–50 300–400

 Medium dense sands and gravels; SPT N-value 10–30 100–300

 Loose sands and gravels; SPT N-value 5–10 50–100
The table above gives typical values of allowable bearing pressures for shallow spread foundations sub-
jected to vertical static loading. It is presumed that the founding levels is at about 1 m depth in soils and
that the ground surface is fairly level. Allowable bearing pressure is the pressure applied by the founda-
tions under design loads; the pressure is limited to a value such that the risk of collapse is minimal and
settlements are restricted to acceptable values.
into excavations and tunnels, softening of fill mate-

Bolton, M. D. (1986) The strength and dilatancy of
rials, initiation of collapse settlements due to inun- sands. Geotechnique 36(1), 65–78.
dation of sensitive soils, or increase in pore Bolton, M. D. (1991) A Guide to Soil Mechanics.
pressures leading to slope failures. On the other M. D. and K. Bolton, Cambridge.
hand, suctions in fine soils may provide useful BS (British Standards Institution) (1991) BS1377:
short-term stability of excavations and slopes, while Methods of Test for Soils for Civil Engineering
Purposes. BSI, London.
lack of water may also lead to problems, for exam- BS (British Standards Institution) (1999) BS5930: Code
ple of drying shrinkage, upward migration of harm- of Practice for Site Investigation. BSI, London.
ful salts, or difficulty in compaction. Proper Carter, M. (1983) Geotechnical Engineering Handbook.
understanding of the water regime, past present and Pentech Press, London.
future, affecting a project is often of prime impor- Casagrande, A. and Fadum, R. E. (1940) Notes on Soil
Testing for Engineering Purposes. Harvard University
tance, yet often not given the attention it deserves. Graduate School of Engineering Publication 268.
Clayton, C. R. I., Simons, N. E. and Matthews, M. C.
(1995) Site Investigation. Blackwell, Oxford.
References Fookes, P. G. (1997) Geology for engineers. Quarterly
Journal of Engng. Geol. 30, 293–424.
Atkinson, J. H. (1981) Foundations and Slopes. Fookes, P. G., Baynes, F., and Hutchison, J. (2001) Total
McGraw-Hill, London. geological history: a model approach to understand-
Atkinson, J. H. (1993) AN Introduction to the Mechanics ing site conditions. Ground Engineering. 34(3),
of Soils and Foundations. McGraw-Hill, London. 42–47.
Bishop, A. W. and Morgenstern, N. (1960) Stability Hoek, E. and Bray, J. W. (1981) Rock Slope Engineering
coefficients for earth slopes. Geotechnique 10, (Revised 3rd edn). Instn. of Mining and Metallurgy/
(4 December 1960). Spon, London.
GFE-07.qxd 6/24/05 3:01 PM Page 169
Hudson, J. A. (1989) Rock Mechanics Principles in Skempton, A. W. (1957) Discussion: the planning and
Engineerig Practice. CIRIA, London. design of new Hong Kong airport. Proc. Instn. Civ.
ISRM (International Society for Rock Mechanics) Engrs. 7, 305–307.
Brown, E. T. (ed.) (1981) Suggested Methods for the Smith, G. N. and Smith, I. G. N. (1998) Elements of
Quantitative Description of Discontinuities in Rock Soil Mechanics (7th edn). Blackwell, Oxford.
Masses. Pergamon Press, Oxford. Stroud, M. A. (1989) The standard penetration test – its
Kenney, K. C. (1959) Discussion. J. Soil Mech. and application and interpretation. Penetration Testing
Foundn. Div. ASCE 85, No.SM3, 67-69. in the UK. Thomas Telford, London, 29–49.
Ladd, C. C., Foote, R., Isihara, K., Schlosser, F. and Stroud, M. A. and Butler, F. G. (1975) The standard pen-
Poulos, H. G. (1977) Stress-deformation and strength etration test and the engineering properties of glacial
characteristics. IX Int. Conf. Soil Mech. and Foundn. materials. Proc. Symp. Engng. Behaviour of Glacial
Engng, Vol. 2, 421–494. Materials. University of Birmingham, 124–125.
Mair, R. J. and Wood, D. M. (1987) Pressuremeter Tomlinson, M. J. (2001) Foundation Design and
Testing, Methods and Interpretation. CIRIA, London. Construction. Pearson, Essex, UK.
Meigh, A. C. (1987) Cone Penetration Testing, Methods Waltham, A. C. (2001) Foundtions of Engineering
and Interpretation. CIRIA, London. Geology (2nd edn). Blackie, London.
Mitchell, J. K. (1993) Fundamental of Soil Behaviour Whitlow, R. (1995) Basic Soil Mechanics (3rd edn).
(2nd edn). Wiley, New York. Longman, Harlow.
Peck, R. B., Harson, W. E. and Thorburn, T. H. (1953) Wyllie, D. C. (1999) Foundations on Rock (2nd edn).
Foundation Engineering (2nd edn 1974). Spon, London.
Perry, J. and West, G. (1996) Sources of information for
site investigations in Britain. Transport Research
Laboratory (TRL) Report LR 192. Further reading
Poulos, H. G. and Davis, E. H. (1974) Elastic Solutions
for Soil and Rock Mechanics. Wiley, New York. Atkinson, J. H. (1981) Foundations and Slopes.
Powrie, W. (1997) Soil Mechanics: Concepts and McGraw-Hill, London.
Applications. Spon, London. Atkinson, J. H. (1993) An Introduction to the Mechanics
Rowe, P. W. (1972) The relevance of soil fabric to site of Soils and Foundations. McGraw-Hill, London.
investigation practice. Geotechnique 22 (2), 195– Bolton, M. D. (1991) A Guide to Soil Mechanics. M. D
300. and K. Bolton, Cambridge.
Seed, H. B., Woodward Jr, R. J. and Lundgren, R. (1962) Hudson, J. A. (1989) Rock Mechanics Principles in
Prediction of swelling potential for compacted clays. Engineering Practice. CIRIA, London.
J. ASCE 88, (SM3). Powrie, W. (1997) Soil Mechanics: Concepts and
Simons, N., Menzies, B. and Matthews, M. (2001) A Applications. Spon, London.
Short Course in Soil and Rock Slope Engineering. Waltham, A. C. (2001) Foundations of Engineering
Thomas Telford, London. Geology (2nd edn). Blackie, London.
GFE-07.qxd 6/24/05 3:01 PM Page 170
GFE-08.qxd 6/27/05 2:03 PM Page 171
Part II
Geomorphological Processes
GFE-08.qxd 6/27/05 2:03 PM Page 172
GFE-08.qxd 6/27/05 2:03 PM Page 173
8. Landslides
James S Griffiths
8.1 Introduction an important natural hazard that require investiga-

tion and mitigation. However, the scale of losses are
There are many different definitions of landslides considerably less than those associated with other
in the literature but the simplest and most useful is natural hazards such as floods, earthquakes, famine
that provided by Cruden (1991) for the UNESCO and epidemics. As a result, the development of
Working Party on World Landslide Inventory: techniques through research for anticipating land-
a landslide is the movement of a mass of rock, earth
slide occurrence is still in its infancy. Whilst there
or debris down a slope. is a good understanding of the processes and causes
of landsliding, our ability to anticipate where and
The term ‘landslide’ does not mean investigations when a landslide will occur is still very limited
are limited only to failures on land that move with compared to other forms of discrete ground move-
a sliding mechanism. With the increasing use of ment, such as faults. Also, within landslide zones
the seafloor for resource exploitation sub-marine there are complex patterns of compression and ten-
slope failures are an important area of study sion that make them difficult to manage, and there
in engineering geomorphology. In the literature is considerable uncertainty about the scale and
the terms ‘landslip’ or ‘mass movement’ are often nature of the landslide response to movement trig-
used rather than ‘landslide’. Whilst these terms gering events.
may be regarded as describing the same geomor- The landslide literature is vast, with the defin-
phological phenomena, detailed investigations of itive general work on landslides presently proba-
the movement of landslides have shown that slid- bly best represented by Turner and Schuster
ing along a basal separation or shear surface is one (1996). Almost every year there are landslide
of the main mechanisms involved, hence the term symposia and conferences, each generating a
‘landslide’ is preferred. However, the movement wide range of theoretical models and practical
mechanisms involved in landsliding are complex case studies (e.g. Bromhead et al., 2000). In addi-
and can include sliding, flowing and falling. tion there are numerous articles in the scientific
The losses and damage resulting from landslides literature on various aspects of landslide investi-
vary considerably around the world. Brabb (1991) gation (Hutchinson, 2001). For detailed landslide
estimated that during the early 1970s the average investigations reference should be made to this
annual number of deaths in the world caused by impressive wealth of material. However, in engi-
landslides was c. 600 but twenty years later the neering geomorphology the requirement is usu-
actual number of deaths had increased to several ally to go out to a site where there has been little
thousand. In the USA, it is estimated that landslides previous work and geomorphologists will have
annually result in over US$1.5 billion in losses and to rely on their own knowledge and skills to
25 to 50 deaths whilst in Japan the annual study the problem. The typical questions the
economic losses exceed US$1 billion (Smith, geomorphologists are asked in these situations
2001). Apart from these general figures there have are:
been some single very tragic events that have
resulted in appalling losses of life and examples of 1. What is the distribution of pre-existing land-
these are listed in Table 8.1. Clearly landslides are slides?
173
GFE-08.qxd
Table 8.1 Examples of high magnitude/low frequency landslide events in the latter part of the 20th century (after Schuster, 1995; Godt and Savage, 1999,
174
and Guadagno and Perriello Zampelli, 2000).
6/27/05
Year Location Name and Type* Volume (m3) Trigger Impact Comment
1962 Peru (Ancash) Huscaran 13 106 Unknown 4000–5000 killed; much Major debris avalanche from
Debris avalanche of Ranrahirca village Nevado Huscaran; average
destroyed velocity 170 km/hr
2:03 PM
1963 Italy Vaiont rock slide 250 106 Filling Vaiont c. 2000 killed; city of High velocity rock slide into
reservoir Longarone badly reservoir caused 100 m wave
damaged; total cost in to overtop the dam. Estimated
1963 US$ 200 million maximum velocity 50 m/s
Page 174
1964 Alaska The 1964 Unknown M9.4 Damages in 1964 cost Spreading failures caused major
Alaskan landslide earthquake US$180 million landslide damage in Anchorage,
(spreading failures) Valdez, Whittler & Seward
1966 Brazil (Rio de Rio de Janeiro Unknown Heavy rains ⬃1000 killed Many landslides around
Janeiro avalanches debris & Rio de Janeiro
mud flows
1966 Wales Aberfan Tip flow 1.1 105 Loose tipping 144 killed including Flowslide of loose tipped
slide (debris flow) on a spring 112 school children colliery waste, estimated
maximum velocity 8.8 m/s;
site of previous slides that
did not reach the village
1970 Peru Nevados Huascaran Unknown M7.7 18 000 killed; town of Rock/debris avalanche
rock/debris earthquake Yungay destroyed; from same peak as 1962;
avalanche Ranrahirca partially average velocity 280 km/hr
destroyed
1974 Peru Mayunmarca rock 1.6 109 Rainfall — Mayunmarca village Debris avalanche with average
slide – debris river erosion destroyed ⬃450 killed; velocity 140 km/hr;
avalanche failure of 150 m high land- dammed Maataro river
slide dam caused major
downstream flooding
1980 USA Mount St Helens 2.8 109 Eruption of World’s largest historic Began as rotational rock slide,
(Wash.) rotational rock slide Mt St Helens landslide; 5–10 killed — degraded to 23 km long debris
followed by debris most people evacuated; avalanche with average velocity
avalanche major destruction to 125 km/hr; surface remobilised
infrastructure in to 95 km long debris flow
1982 San Franciso 18 000 debris Unknown January 3–5 25 deaths and landslide
Bay Region flows rainstorm damage estimated
at $65 million
GFE-08.qxd
1983 USA (Utah) Thistle debris 21 106 Snowmelt and No deaths; destruction of
slide heavy rain infrastructure; dammed
Spanish Fork flooding
6/27/05
town of Thistle. Total losses
in 1983 US$400 million
1983 China Salasham 35 106 Unknown 237 killed; 4 villages Loess landslide
(Gansu) landslide buried; 2 reservoirs
filled with debris
2:03 PM
1985 Colombia Nevado del Ruiz Unknown Eruption of 4 town & villages Death toll unnecessarily
(Tolina) debris flows Navado destroyed; flow in valley large because hazard warnings
del Ruiz of Langunillas River killed not passed to residents
20 000 in city of Armero
Page 175
1986 Papua New Bairaman rock 200 106 M7.1 Village of Bairaman destroyed Debris avalanche formed 210 m
Guinea (East slide — debris earthquake by debris flow from breached high dam that impounded
New Britain) avalanche landslide dam; evacuation 50 million m3 lake; dam failed
prevented casualties; major causing 100 m high debris
environmental effects flow — flood downstream
1987 Ecuador Reventador 75110 106 M6.1 6.9 ⬃1000 killed; many kms of Landslides mainly in saturated
(Mapo) landslides (mainly earthquake oil pipeline & highway residual soils on steep slopes;
debris flows) destroyed; 1987 costs thousands of thin debris
US$1 billion flows in catchments
1994 Colombia Paez landslides Area 250 km2 M6.4 271 killed; 1700 missing; Thousands of thin, residual soil
(Cauca) (mainly debris earthquake 32 000 displaced; several slides on steep slope
flows) villages destroyed becoming debris flows
1997/8 San Francisco c. 300 landslide Largest High rainfall Landslide damage estimated Rainfall recorded at more than
Bay region events of all types individual at $158 million; only twice the annual average as
failure: one fatality a result of a Type 1 El
13 million m3 Niño Southern Oscillation
Mission Peak
Earthflow
1998 Campania, 129 separate Area of Very high daily 161 lives lost; towns of Rainfall return period calculated
southern initial slides – sliding and antecedent Quindici, Sarno, Bracigliano as 100 years; minimum flow
Italy predominantly 70 km2 rainfall and Siano devastated rates of 10–20 km/hr; failures
debris flows mainly in colluvium and
LANDSLIDES 175
weathered pyroclastic
material from Mt Vesuvius
* For a description of the different types of failure refer to Table 8.2.

GFE-08.qxd 6/27/05 2:03 PM Page 176
2. Are there typical landslide-prone strata or Commission Environment Programme on

sequences in the area? ‘Landslide Recognition’. Whilst this classification
3. What is the potential for first-time failures? has its detractors (see the discussion in
4. Where are landslides likely to occur? Hutchinson, 2001), it provides a relatively simple
5. Can an estimate be made of the probability of and easy to apply system that has gained wide-
landslide events, including the re-activation spread acceptance in engineering geomorphology.
of any pre-existing failures? The system identifies only five types of landslide
6. What are the likely consequences or impact and these are defined by their initial mechanism of
of any landslide events? movement: ‘fall’ (Figure 8.3), ‘topple’ (Figure 8.4),
7. If a pre-existing failure exists at the site, what ‘slide’ (Figure 8.5), ‘spread’ (Figure 8.6), and
was the cause, can it be stabilised or avoided? ‘flow’ (Figure 8.7) plus a ‘complex’ landslide
involving two or more of the failure mechanisms.
To investigate and understand landslides, it is Slides and flows are further subdivided based on
necessary to classify the landslide type, identify the the form of their initial failure surface. In addition,
mechanisms of movement and establish the causes the complex category does identify one specific
of the failure. Engineering geomorphology has a type of failure, the rock or debris avalanche, as this
key role in providing this information. Once this is a widely recognised and particularly devastating
understanding has been acquired it might then be form of landslide. As shown in Table 8.2, after the
possible to provide comments on the landslide sta- initial movement all landslides deform further as
bility, susceptibility, hazard and risk of an area or a the movement continues, until the failure stops at
particular site. Broader scale susceptibility, hazard a maximum runout distance.
and risk assessments provides end-users with a The classification proposed by Dikau et al.
product that has widespread application in devel- (1996) is based on that originally suggested by
opment planning and engineering feasibility stud- Varnes (1978). However, whilst the Varnes classi-
ies (Cruden and Fell, 1997; Lee and Jones, 2004). fication used the same main categories of fall,
Detailed geomorphological surveys of a single topple, slide, spread and flow, it differentiated
landslide can be used in conjunction with conven- between the failures occurring on rock, debris and
tional ground investigation techniques to provide soil. This is no longer deemed to be a useful
data that will feed directly in to the engineering additional complication as all the failure types
design process (Griffiths et al., 1995). can occur in all materials. Alternative systems
of classification, such as that proposed by
Hutchinson (1988), are of most value in detailed
8.2 Description geotechnical investigations of landslides and land-
slide research. For general purposes in engineering
In order to be able to collate landslide studies from geomorphology, therefore, the Dikau system is
a range of environments and geomorphological recommended.
situations it is necessary to adopt a standard Examples of the various landslide types that
method of describing the basic types of landslide, occurred in the latter part of the 20th Century and
their features and dimensions. The recommended their details are presented in Table 8.1. This illus-
approaches, as proposed by the European Com- trates that the rapidly moving rock and debris ava-
mission (Dikau et al., 1996) and UNESCO World lanches are the most devastating forms of failure,
Landslide Inventory (1990), are presented in in terms of deaths and economic losses.
Figures 8.1 and 8.2 and Tables 8.2, 8.3 and 8.4.
Landslide dimensions
Landslide types and mechanisms The UNESCO World Landslide Inventory (1993)
The classification of landslide types presented in proposed the standard system for describing
Figure 8.1 and Table 8.2 was proposed by Dikau the dimensions and features of a landslide
et al. (1996) in their report to the European which is presented in Figure 8.2 and Tables 8.3
GFE-08.qxd 6/27/05 2:03 PM Page 177
LANDSLIDES 177
1
20m
0m
4
3
3
2
20m
2
1
0m
20m
Extent of displaced material

Original ground surface
Trajectories of individual particles
0m
Figure 8.1 Landslide types from Dikau et al., 1996, p. A2.

Types of landslides:
(1) A fall starts with detachment of soil or rock from a steep slope along a surface on which little or no shear
displacement takes place. The material then descends largely through the air by falling, saltation or rolling.
(2) A topple is the forward rotation, out of the slope, of a mass of soil or rock about a point or axis below the centre
of gravity of the displaced mass.
(3) A slide is the downslope movement of a soil or rock mass occurring dominantly on surfaces of rupture or rela-
tively thin zones of intense shear strain.
(4) A spread is an extension of a cohesive soil or rock mass combined with a general subsidence of the fractured
mass of cohesive material into softer underlying material. The rupture surface is not a surface of intense shear.
Spreads may result from liquefaction or flow (and extrusion) of the softer material.
(5) A flow is a spatially continous movement in which surfaces of shear are short-lived, closely spaced and usually
preserved. The distribution of velocities in the displacing mass resembles that in a viscous fluid.
GFE-08.qxd 6/27/05 2:03 PM Page 178
landslide features 1 2 3 etc

defined in Table 8.3
landslide dimensions
Wd L etc.
defined in Table 8.4
1
note 13 - volume indicated by
14 2
note 16 - volume bounded by 3
2 17 20
4
note 17 - volume overlying 10
19
but underlying 20 5
W
note 18 - volume of 13 lying
r
L
W
above 20
d
15 Ld 20
ce
rfa
su
in al Lr
orig
Dr
6
Dd
7 10
11 13
8
12
9
travel angle ( ) is a measure of L from the horizontal
Figure 8.2 Dimensions of a landslide (based on the UNESCO World Landslides Inventory, 1990).
and 8.4. The landslide type shown in Figure 8.2 is Table 8.5 a method for classifying the pro-
a multiple rotational slide but the terminology is cesses that cause a landslide is presented (based
appropriate for all types of failure. In Table 8.4 on Brunsden, 1993). This identifies nine main
two methods are given for calculating the volume ‘external processes’ that change the situation in or
of displaced material in a landslide. The first is on a slope: weathering, erosion, subsidence, depos-
more suitable for rotational sides and some flows ition, seismic events, air fall, changes in water
and is based on a general shape that can be regime, complex ‘follow-on’ processes that occur
approximated by half an ellipsoid. The second after the initial failure and human interference.
method is suitable for failures controlled by dis- These result in a number of different effects to the
continuities such as planar slides, topples and slope that either reduce the strength of the mater-
falls. ials (e.g. decomposition of the rock by chemical
weathering) or increase the stress (e.g. erosion at
8.3 Landslide cause the toe of a slope changing its geometry).
Human interference as an external process cre-
Landslides occur when the strength of the mater- ates the same effects and changes to a slope as a
ials in a slope is exceeded by the stresses acting range of natural processes. For example, excav-
on the slope. The actual causes of the failure are ation at the toe of slope is essentially the same as
often complex and a slope is likely to have been natural fluvial erosion (Griffiths et al., 2004),
subject to many ‘causes’ over a long period of except that it is likely to take place in quite differ-
time before a single, possibly quite small, event ent locations. Likewise, one effect of flooding an
actually triggers a significant landslide movement area for a reservoir will be to raise the regional and
(triggering factors in relation to all geomorpho- local water table, possibly causing widespread
logical events are discussed in Chapter 1). These landsliding around the rim. If the reservoir is sub-
changes to the slope may be the result of both sequently drained, for whatever reason, this rapid
natural processes and human interference. In drawdown can create transient very high porewater
GFE-08.qxd 6/27/05 2:03 PM Page 179
LANDSLIDES 179
Table 8.2 Classification of landslides and landslide mechanisms (cf. Figure 8.1) (after Dikau et al., 1996).
Type Form of initial failure surface Subsequent deformation
Fall
Detachment form: a) Planar surface Free fall, may break up,
Pre-existing discontinuities b) Wedge (two intersecting joints) bounce, slide or flow down
or tension failure surfaces c) Stepped surface slopes. May involve fluidisation,
d) Vertical surface liquefaction, cohesionless
grain flow, heat generation
or other secondary effects
Topple
Detachment form: a) Single As above
pre-existing discontinuities b) Multiple
or tension failure surfaces
Slide
Rotational movement a) Single Toe area may deform in a
(failure surface essentially b) Multiple complex way. The ground
circular) c) Successive can bulge, the slide may creep
or even flow, it can override
Non-rotational compound a) Single existing failures. Failure
movement (non-circular b) Progressive might be retrogressive or
failure surface; may be c) Multistoried progressive. Graben often
listric or bi-planar) develops at the head of the
landslide. It may include
a toe failure of a different type
Translational movement (often a) Planar May develop complex runout
associated with discontinuity b) Stepped forms after disintegrating
controlled failures in bedded c) Wedge (see falls and flows)
or foliated rocks) d) Non-rotational
Spread
Lateral spreading of ductile or a) Soft layer beneath a hard rock Can develop sudden spreading
soft material that deforms b) Weak interstratified layer failures in quick clays when
c) Collapsing structure the slope opens up in blocks
and fissures followed by
liquefaction. Might be a slow
movement associated with
denudational unloading. Can be
represented by cambering and
valley bulging
Flow
Debris movement by flow a) Unconfined Flow involves complex runout
b) Channelled mechanisms. It may be
catastrophic in effect and it
may move in sheets or lobes.
The form of movement is a
function of the rheological
properties of the material
GFE-08.qxd 6/27/05 2:03 PM Page 180
Type Form of initial failure surface Subsequent deformation
Creep movement Failure surface rarely Creep may be a superficial

clearly defined gravity movement, seasonal
movements or it might represent
pre-failure and progressive
movements prior to a
larger scale failure
Rock flow (sometimes a) Single sided May be slow gravity creep or
referred to as sagging b) Double sided the early stages of larger scale
or Sackung). Usually c) Stepped movements that only show as
associated with mountain (failure surface may be rotational, bulging in the topography without
terrain or areas of rapid compound, listric, biplanar or a clearly defined toe deformation.
and deep incision intermittent) Where controlled by discontinuities
it may involve toppling
Complex
a) Movements involving two Dependent on the form of As described for the various
or more of the above failure as described above categories above
mechanisms (referred to
as compound when two
types of movement occur
currently)
b) Rock or debris avalanche Often initiated as fall/slide Complex long-runout mechanisms,
of rock and/or debris including fluidisation and
cohesionless grain flow
conditions causing quite extensive landslide activ- that indicate the presence of a landslide in the
ity. Similarly, blasting, or plant and machinery terrain. The book Landslide Recognition by Dikau
vibrations can replicate natural seismic activity. et al. (1996) provides the most comprehensive col-
The system presented in Table 8.5 provides a lection of landslide indicators as well as utilising the
useful method for understanding how the range of landslide type classification system recommended
natural and man-made causal processes can affect above. A simpler list of features, that can be
a slope. However, during the investigation of a observed both in the field and while using aerial
single landslide it will be necessary to identify the photographs, is provided in Table 8.7 (after Soeters
various causes and the triggering causal event. In and Van Westen, 1996). As with all aspects of geo-
such situations it is more appropriate to have a morphology, the ability to recognise landslide fea-
checklist of the possible causes and a useful tures, particularly when they are subtle or degraded,
example is presented in Table 8.6, from Cruden is a skill that develops with experience. This expe-
and Varnes (1996). rience can only be obtained through fieldwork.
8.4 Landslide recognition 8.5 Landslide activity and

behaviour
Armed with the appropriate terminology to
describe the features and dimensions of a landslide Whilst the general population tends to envisage a
and an understanding of the possible causes of fail- landslide as being a single, often catastrophic,
ure, it is necessary to identify the landscape features event that occurs rapidly, in geomorphological
GFE-08.qxd 6/27/05 2:03 PM Page 181
LANDSLIDES 181
Table 8.3 Landslide features in Figure 8.2. 17. Depleted mass: part of the displaced material that
overlies the rupture surface (10) but underlies the
1. Crown: the practically in situ material adjacent to original ground surface (20)
the highest parts of the main backscar scarp 18. Accumulation: the volume of the displaced
2. Main scarp: a steep surface on the undisturbed material (13) which overlies the original ground
ground at the upper edge of the landslide surface (20)
representing the backscar, caused by movement 19. Flank: undisplaced material adjacent to sides of
of the slide material (13) away from the surface of rupture
undisturbed ground; it is a visible part of the 20. Original ground surface: surface of slope that
surface of rupture (10) existed before the landslide took place
3. Top: the highest point of contact between the
displaced material (13) and the main scarp (2)
4. Head: the upper parts of the landslide along the
contact between the displaced material and the
Table 8.4 Landslide dimensions shown in Figure 8.2.
main scarp (2)
5. Minor scarp: a steep surface on the displaced Wd Width of displaced mass: maximum breadth of
material of the landslide, produced by differential displaced mass (13) perpendicular to Ld
movements within the sliding mass Wr Width of rupture surface: maximum width
6. Main body: the part of the displaced material of between flanks (19) of the landslide perpen-
the landslide that overlies the surface of rupture dicular to Lr
(10) between the main scarp (2) and the toe of Ld Length of displaced mass: minimum distance
the surface of rupture (11) from tip (8) to top (3)
7. Foot: the portion of the landslide that has moved Lr Length of surface of rupture: minimum distance
beyond the toe of the surface of rupture (11) and from the toe of surface of rupture (11) to crown
overlies the original ground surface (20) (1)
8. Tip: the point of the toe (9) furthest from the top Dd Depth of displaced mass: maximum depth of
of the landslide (3) displaced mass (13) measured perpendicular to a
9. Toe: the lower, usually curved, margin of the plane containing Wd and Ld
displaced material of a landslide, it is the most Dr Depth of surface of rupture: maximum depth of
distant from the main scarp (2) surface of rupture (10) below original ground
10. Surface of rupture: the surface that forms the surface (20) measured perpendicular to a plane
lower boundary of displaced material (13) below containing Wr and Lr
the original ground surface (20) L Total length: minimum distance from tip of
11. Toe of surface of rupture: the intersection landslide (8) to crown (1)
(sometimes buried) between the lower part of the Travel angle: difference in elevation between the
surface of rupture (10) of a landslide and the crown (1) and the tip (8) of the landslide divided
original ground surface (20) by L
12. Surface of separation: the part of the original
ground surface (20) overlain by the foot (7) of Landslide volume estimation
the landslide Approximate volume of the ground displaced by a
13. Displaced material: material displaced from its landslide (Volls) based on the failure being represented
original position on the slope by the movement of by half an ellipsoid:
the landslide
14. Zone of depletion: the area of the landslide
Volls (1/6)..Dr.Wr.Lr
within which the displaced material lies below
the original ground surface (20)
Approximate volume of the ground displaced by
15. Zone of accumulation: the area of the landslide
a landslide (Volls) based on the failure being planar
within which the displaced material lies above
in form and lateral limits determined by vertical
the original ground surface (20)
discontinuities:
16. Depletion: the volume bounded by the main
scarp (2), the depleted mass (17) and the original
ground surface (20) Volls Lr. Wr . Dr
GFE-08.qxd 6/27/05 2:03 PM Page 182
Figure 8.3 Rockfall associated with fluvial undercutting in a river cliff near Sorbas, south-east Spain.
Figure 8.4 Backtilted topple blocks on the outside of a meander bend near Sorbas, south-east Spain.
GFE-08.qxd 6/27/05 2:03 PM Page 183
LANDSLIDES 183
Figure 8.5 Small translational slide in middle ground, centre of the image, moved from right to left. On the western
flanks of Pen-y-Fan in the Brecon Beacons, south Wales.
investigations it is useful to distinguish between landslide activity classification system, as sug-

first-time failures at sites with no evidence of pre- gested in Table 8.8. This system divides landslides
vious movement, and pre-existing slides that into active, suspended, young dormant, mature
might be prone to re-activation. This is because dormant, old dormant and fossil forms. With
experience has shown that some landslides can decreasing activity the landslides become increas-
remain dormant for long periods of time and then ingly degraded and more difficult to identify in the
become active again. Other landslides, whilst landscape. However, even dormant and some fos-
active and continually moving, may only creep sil landslides that remain constrained within the
along at a few millimetres a year, despite being bounding shear surface will move when the resid-
long established features in the contemporary ual strength along the surface is exceeded. The
landscape. Landslides tend to have the typical fre- debris accumulation from these landslides will
quency–magnitude distribution associated with also be loose and poorly consolidated with a rela-
natural geomorphological processes where low tively low strength. These landslides, therefore,
magnitude events occur more often than high constitute an increased engineering risk and it is
magnitude events. However, it is not easy to apply necessary they be identified in any engineering
the concept of recurrence interval to landslides, as, geomorphological investigation.
for example, is possible with river flood or rainfall The other facet of landslide activity that has
events, because landslides vary both in location major engineering implications is the behaviour
and over time. This variation in landslide activity of the ground during movement, including its
is highly relevant to engineering, and for geomor- velocity. Figure 8.8 illustrates how a landslide
phological investigations it is useful to adopt a typically deforms the ground. In the crown and
GFE-08.qxd 6/27/05 2:03 PM Page 184
Figure 8.6 Tension cracks associated with lateral spreading of a limestone caprock overlying marls east of the
village of Los Perales, south-east Spain.
Figure 8.7 Removal of mudflow debris from a track west of Lyme Regis, Dorset (UK).
GFE-08.qxd 6/27/05 2:03 PM Page 185
LANDSLIDES 185
Table 8.5 Classification of the processes that cause landslides (based on Brunsden, 1993).
External process(es) Causal effects Description of typical Examples of specific

changes changes on slope
Weathering: physical, Changes in physical Changes in grading; Changes in: density,

chemical & biological and chemical properties; cation exchange; strength, permeability;
horizonation; changes cementation; formation stress, pore and cleft
in regolith thickness of weak discontinuities water pressure
or hard bands; increased
depth of low strength
materials
Erosion of material from Changes in slope Alterations to: relief; Changes in stress,
face or base of slope geometry; unloading slope height, length, permeability and
by fluvial, glacial and/or angle and aspect strength
coastal processes
Ground subsidence Undermining Mechanical eluviation Loss of support;
of fines; solution; loss consolidation; changes
of cement; leaching in porewater pressure;
seepage erosion; loss of strength
backsapping; piping
Deposition of material Loading; long term Alterations to: relief; Changes in stress,
to face or top of slope (drained) or short slope height, length, permeability, strength,
by fluvial, glacial or term (undrained) angle and aspect loading and
mass movement porewater pressure
processes
Seismic activity and Rapid and repeated Disturbance to Changes in stress;
general shocks and vertical and horizontal intergranular bonds; loss of strength;
vibrations displacements transient high high porewater
porewater pressures; pressures; potential
materials subject to for liquefaction
transient and repeated
periods of compression
and tension
Air fall of loess Mantling slopes with New slope created Changes in stress;
or tephra fines; adding fines to with well defined strength; water content
existing soils discontinuity boundary and water pressure
Water regime change Rising or falling Piping, floods, lake Excess porewater
groundwater; bursts; ‘wet’ years; pressures; changes in
development of perched intense precipitation; bulk density; reduction
water tables; saturation snow and ice melt; in effective shear
of surface; flooding rapid drawdown strength
Complex ‘follow-on’ Liquefaction; remoulding; Long runout landslides; Changes in effective
or runout processes fluidisation; ‘acoustic low values for ratio shear strength, water
after initial failure grain flow’ of initial failure volume distribution, bulk
to total failure volume; density and rheological
low angles of reach; low characteristics
breadth to length ratios
Human interference Excavation at toe of slope Same as natural erosion Same as natural erosion
Top loading of slopes Same as natural deposition Same as natural deposition
Flooding (e.g. leaking Same as natural water Same as natural water
services; reservoir regime change regime change
construction)
GFE-08.qxd 6/27/05 2:03 PM Page 186
Table 8.6 Checklist of landslide causes (after Cruden backscar areas there will be differential horizontal
and Varnes, 1996). and vertical movements with tension cracks and
scarps. Within the main body of the landslide
1. Geological a) weak materials
causes b) sensitive materials there will be areas in compression, secondary ten-
c) weathered materials sion scarps but generally an area of subsidence.
d) sheared materials Discrete shears or en échelon cracks may mark
e) jointed or fissured materials the lateral limits of the slide area. The toe area
f) adversely orientated mass will be an area of uplift or toe bulge. Material in
discontinuity (bedding, this area is likely to be under compression at the
schistosity etc.) base of the landslide debris whereas at the surface
g) adversely orientated structural there is likely to be cracking as the debris comes
discontinuity (fault, under tension. Knowing the state of the ground
unconformity etc.)
within a landslide is the key to understanding how
h) contrast in permeability
structures will respond to instability and how
i) contrast in stiffness (stiff,
dense material over plastic to design structures to accommodate landslide
material) movements.
2. Morphological a) tectonic or volcanic uplift With respect to landslide velocity, Cruden and
causes b) glacial rebound Varnes (1996) introduced a useful guide to the
c) fluvial erosion of slope toe destructive significance of landslides, repro-
d) wave erosion of slope toe duced in Table 8.9. This guide follows the style of
e) glacial erosion of slope toe the Modified Mercalli scale used to describe the
f) erosion of lateral margins local effects of earthquakes. The table illustrates
g) subterranean erosion that it is only the Velocity Class 7 landslides that
(solution; piping)
are truly catastrophic and that for all slower
h) deposition loading slope or
moving landslides, loss of life can be either mini-
its crest
i) vegetation removal (by forest mised or avoided. It also demonstrates that the
fire, drought) economic affects of landslides are far more wide-
3. Physical causes a) intense rainfall spread as it is only in the slowest moving failures
b) rapid snow melt in Velocity Classes 1 and 2 that structures can be
c) prolonged exceptional maintained. The concern with ‘creeping’ land-
precipitation slides in these slow and extremely slow moving
d) rapid drawdown (of floods velocity classes, however, is that they may be
and tides) creeping towards a threshold situation when a
e) earthquake larger, more rapid movement is triggered. This
f) volcanic eruption
emphasises the need to monitor any landslides
g) thawing
in the vicinity of both planned and existing
h) freeze-and-thaw weathering
i) shrink-and-swell weathering developments to ensure that movements are not
4. Human causes a) excavation of slope or its toe accelerating.
b) loading of slope or its crest
c) drawdown (or reservoirs)
d) deforestation 8.6 Landslide distribution and
e) irrigation occurrence
f) mining
g) artificial vibration The four main controls on landscape develop-
h) water leakage from utilities ment are climate, geology, relief and time. In
broad terms geomorphological studies have
shown that certain processes and landforms
are often specific to ‘morphoclimatic’ regions
GFE-08.qxd 6/27/05 2:03 PM Page 187
LANDSLIDES 187
Table 8.7 Landslide recognition (after Soeters and van Westen, 1996).
Terrain features Relation to slope instability Aerial photograph characteristics
Morphology
Concave/convex slopes Landslide niche and associated deposits Concave/convex anomalies
in stereo model
Step-like morphology Retrogressive sliding Step-like appearance of slope
Semi-circular backscarp Head of slide with outcrop Light-toned scarp, associated
and steps of failure plane with small, slightly
curved lineaments
Back-tilting of Rotational movement Oval or elongate depressions
slope facets of slide blocks with imperfect drainage
Hummocky and Micro-relief associated with shallow Coarse surface texture
irregular slope movements or small retrogressive contrasting with smooth
slide blocks surroundings
Infilled valleys with slight Mass movement deposit Anomaly in valley
convex floor where V-shaped of flow-type form morphology often with
valleys are normal lobate form and identifiable
flow pattern
Vegetation
Vegetational clearance on steep Absence of vegetation on head scarp Light-toned elongated areas
scarps coinciding with or on steps in main body of slide at crown of slide
morphological steps
Irregular linear clearances Slip surface of translational slides and Denuded areas showing light
along slope track of flows and avalanches tones often linear pattern in
direction of movement
Disrupted, disordered and partly Slide blocks and differential movements Irregular sometimes mottled
dead vegetation in main body grey tones
Differential vegetation associated Stagnated drainage on back-tilted blocks, Tonal differences displayed
with changing drainage seepage at frontal lobe and varying in pattern associated with
conditions conditions on main body morphological anomalies in
stereo model
Stretched and exposed roots Separation of blocks as a result of Rarely visible, might appear
downslope movements as disrupted vegetation
Drainage
Areas with stagnant drainage Landslide niche, back-tilting block and Tonal differences with darker
hummocky relief on main body tones associated with wetter
areas (shows up well in the
near infra-red)
Excessively drained areas Overbulging landslide body with Light-toned zones in association
differential vegetation and possibly with convex relief forms
soil erosion
Seepage and spring levels Springs along frontal lobe and at places Dark patches sometimes in
where failure surface crops out slightly curved pattern and
enhanced by differential
vegetation
Interruption of drainage lines Drainage anomaly caused by head scarp Drainage line abruptly broken
off on slope by steeper relief
Anomalous drainage pattern Streams curving around frontal lobe or Curved drainage pattern upstream
streams on both sides of main body with sedimentation or
of landslide meandering in asymmetric valley
GFE-08.qxd
Table 8.8 Activity classification for landslides (after Mather et al., 2003, adapted from Cruden and Varnes 1996; Keaton and DeGraft 1996).
188
Activity state Identification of causes Condition of main scarp Condition of lateral margins Internal morphology Estimated age
6/27/05
of movement (years)
Active — currently Causes of movement Sharp; unvegetated Sharp; unvegetated; Hummocky; fresh scarps; 100
moving (includes identifiable and active streams at edge reverse slopes; undrained (historic)
inactive landslides depressions; fresh
2:03 PM
that have been tension cracks
reactivated)
Suspended — moved Causes of movement Sharp; unvegetated Sharp; unvegetated; Hummocky; fresh scarps; 100
within the last identifiable and likely streams at edge reverse slopes; undrained (historic)
Page 188
12 months and to re-occur depressions; identifiable
likely to become tension cracks
active again
Dormant young Cause of movement still Relatively sharp; Relatively sharp; partially Hummocky; relatively sharp 100–5000
(inactive) identifiable and could partially vegetated vegetated; lateral streams and fresh scarps; reverse (Late Holocene)
re-occur fed by small tributaries slopes; undrained depressions;
flowing off the main body tension cracks closed and
of the slide vegetated but marked by
small depressions
Dormant mature Cause of movement still Smooth; vegetated Smooth vegetated; lateral Smooth, rolling topography; 5000–10 000
(inactive) — identifiable but not streams fed by tributaries disrupted and disjointed (Early Holocene)
described as likely to re-occur flowing off the main body internal drainage network.
abandoned by
Dikau et al. (1996) of the slide
Dormant old Causes of movement may Dissected; vegetated Vague lateral margins; Smooth, undulating topography; 10 000–100 000
(inactive) be inferred but associated no lateral drainage normal stream pattern (Late
or relict* with different climatic or Pleistocene)
geomorphological
conditions
Fossil (inactive) Causes of movement May not be identifiable; May not be identifiable; Fully integrated into the 100 000
or ancient* unknown but associated likely to be at least likely to be at least existing topography and
with different climatic and partially if not partially if not completely very little indication of the
geomorphological completely removed removed by erosion former landslide
conditions by erosion morphology remains
* Relict or fossil landslides may also be covered by subsequent deposits. Rib and Liang (1978) described these as ‘hidden’ or ‘buried’ landslides, and they
always have the potential to be exhumed and reactivated. This is the situation described by Schultz and Harper (1996) for the Late-Carboniferous palaeo-land-
slides reactivated by building excavations in Pennsylvannia, USA.
GFE-08.qxd 6/27/05 2:03 PM Page 189
LANDSLIDES 189
UPSLOPE PRINCIPAL SCARP
SECONDARY
SCARPS
ZONE OF DEPRESSION
OR SUBSIDENCE
DIAGONAL
TENSION-SHEAR
CRACKING:
EN ECHELON
CRACKS
ZONE OF BULGING
DOWNSLOPE
TENSION CRACK: POSSIBLY TOE

SHEAR DISPLACEMENT DOWNHILL MUD WAVE OR
TOE RIPPLE
TENSION CRACK ON TOE BULGE
SPRINGS, SEEPS
Figure 8.8 Landslide behaviour (after Sowers and Royster, 1978).
(Chapter 1). Within the morphoclimatic region Landslide-prone terrain

the local geology (structure, lithology, rock and The concept of morphoclimatic regions controlled
soil mass characteristics) and relief or topog- by climate can be used to identify the broad distri-
raphy controls the shape and nature of individual bution of different landslide types, and these
landforms. The occurrence and distribution of are presented in Table 8.10. Fookes (1997) took this
landslides, which are just one type of landform, broad scale terrain system concept further by
fit this general model of landscape development producing a sequence of climate controlled land-
and whilst individual landslides can be a site- scape models highlighting facets of geomorpho-
specific problem anywhere in the world, in some logical processes that affect engineering design.
situations landslides represent a major compon- In many of these models landsliding represents a
ent of the landscape. significant process that must be taken into account.
GFE-08.qxd 6/27/05 2:03 PM Page 190
Table 8.9 Destructive significance of landslides of different Velocity Classes (after Cruden and Varnes, 1996).
Velocity Description Velocity Typical Probable destructive significance

class (mm/s) velocity
7 Extremely rapid Catastrophe of major violence; buildings

destroyed by impact of displaced material;
many deaths; escape unlikely
5 103 5 m/s
6 Very rapid Some lives lost; velocity too great to
allow all persons to escape
5 101 3 m/minute
5 Rapid Escape evacuation possible; structures,
possessions and equipment destroyed
5 101 1.8 m/hour
4 Moderate Some temporary and insensitive structures
can be temporarily maintained
5 103 13 m/month
3 Slow Remedial construction can be undertaken
during movement; insensitive structures
can be maintained with frequent
maintenance work if total movement is
not too large during a particular
acceleration phase
5 105 1.6 m/year
2 Very slow Some permanent structures undamaged by
movement
5 107 16 mm/year
1 Extremely slow Imperceptible without instruments;
construction possible with precautions
A number of morphoclimatic regions can be of landslide processes that can be anticipated in

identified where landsliding is a very significant these regions.
process in landscape development. High moun- Hong Kong provides an example of the trop-
tainous areas (Figure 8.9) subject to seismic activ- ical wet–dry morphoclimatic region where land-
ity (Figure 8.10) and where rainfall is high stand slide occurrence is also significant. The average
out as being particularly landslide prone. An annual rainfall in Hong Kong is 2225 mm with a
examination of the major landslide events listed marked seasonal pattern and periodic typhoons.
in Table 8.1 supports this, the South American Hong Kong has been subject to extensive land-
Andean Mountains, for example, have been sub- slide research since a major failure occurred at
ject to a number of major landslide catastrophes Po Shan Road in 1972 killing 67 people (Li et al.,
during the latter part of the 20th Century. In con- 1998). The landscape model developed for Hong
trast, other mountainous areas where slopes are Kong arose out of a territory-wide terrain evalu-
gentle, rainfall is only moderate and which are no ation study carried out during the 1980s in order
longer seismically active are unlikely to be sub- to provide the basis for planning new building and
ject to extensive contemporary landsliding, infrastructure construction. The terrain model, pre-
although there may be relict features within the sented in Figure 8.11, was based on the creation of
landscape. For high mountainous terrain in Nepal, a detailed study of land forming processes illus-
Fookes (1997) produced the landscape model pre- trated in Figure 8.12 (Hansen, 1984). Subsequent
sented in Figure 19.3 which illustrates the range detailed studies have shown that the overall
GFE-08.qxd 6/27/05 2:03 PM Page 191
LANDSLIDES 191
Table 8.10 Landslide occurrence in the global morphoclimatic regions (after Chorley et al., 1984); see Table 2.1.
Morphoclimatic General geomorphic Morphological features Typical landslide types

region1 (typical processes
name)
Humid tropical Maximum chemical Low gradient rivers; Earthflows and rotational
(rainforest)2 weathering; episodic wide flat floodplains; slides in weathered
mass wasting; moderate steep slope arising debris
slope wash and fluvial abruptly from valleys,
processes; high solute stabilised by vegetation;
and suspended loads knife-edged ridges
in rivers
Tropical wet–dry High chemical and moderate Steep irregular slopes of Flows and slides in
(savannah)3 mechanical weathering; coarse debris; wide weathered debris; rock
moderate to high mass planation surfaces; isolated falls and topples from
wasting; seasonal inselbergs; some badlands exposed rock faces
sheetfloods; moderate
wind action
Arid (desert)2 Maximum thermal and salt Dunes, playas, pediments, Minor rockfalls; debris
weathering; rare episodic debris covered slopes, flows on alluvial fans
fluvial action (often relict); ephemeral stream channels
maximum wind action (some fossil), alluvial fans
Semi-arid Moderate chemical and Pediments backed by cliffs Wide range: rockfalls and
(Mediterranean)3 mechanical weathering; and talus slopes; inselbergs; larger scale translational
episodic moderate mass integrated ephemeral stream slides; debris flows on
wasting; maximum fluvial networks; badlands; alluvial fans; rotational slides in
erosion in ephemeral rivers fans; local dunes weathered debris
Dry continental Moderate rates for all forms Very similar to semi-arid Similar to semi-arid
(steppe)3 of weathering; moderate and tropical wet–dry but landscapes
mass wasting but seasonally with increased local
important; moderate to high effects of frost action,
fluvial activity; moderate e.g. angular talus slopes
wind action and fossil loess
Humid mid-latitude Full range of weathering Smooth soil covered slopes High rates of soil creep;
(temperate marine)3 processes at moderate rates; with rounded ridges and wide range of landslides
moderate fluvial processes; valleys; wide range of but often relict or only
evidence of fossil glacial grain sizes in alluvium slowly moving
activity
Periglacial (tundra)3 Maximum frost shattering Permafrost and seasonally Rockfalls; talus creep;
and mechanical weathering; frozen ground; screes; solifluction
maximum talus creep and solifluction sheets
gelifluction; moderate cryoplanation surfaces;
fluvial action in thaw outwash plains; patterned
season; relict glacial scour; ground; loess
moderate to high wind
action
Glacial (arctic)2 Maximum frost shattering Alpine topography; abrasion Rockfalls; talus creep
and mechanical weathering; surfaces; glacial and fluvio-
maximum glacial scour; glacial features; glaciers
maximum wind action and ice caps
GFE-08.qxd 6/27/05 2:03 PM Page 192
Morphoclimatic General geomorphic Morphological features Typical landslide types

region1 (typical processes
name)
Mountains4 Maximum mechanical and High mountains; deeply Maximum landslide

moderate chemical weathering incised valley with over- activity of all types
extensive mass high stream steepened slopes; braided
gradients and high sediment rivers with terraces
concentrations; often recent
or active glaciation
1
A morphoclimatic region is a large areal unit within which distinctive associations of geomorphological processes
operate and tend towards a morphoclimatic equilibrium wherein regional landforms reflect regional climates.
2
First order morphoclimatic region with non-seasonal processes, generally low average erosion rates, highly
infrequent and episodic erosive activity (including mass slope failure), and a tendency for the location of their cores
to persist in a certain latitude during periods of climate change.
3
Second order morphoclimatic region have processes that are more seasonal in operation, often with high rates of
erosion which may be episodic although is often consistent over long periods. Location and extent of region varies
considerably as climate changes. Landforms often left as relicts in other morphoclimatic regions when climate changes.
4
Azonal unit that crosses morphoclimatic regions.
frequency of landsliding within the 1000 km2 area

of Hong Kong is 325 per year (Evans, 1998).
In temperate regions such as the UK, whilst the
overall density of landsliding is relatively low
(Jones and Lee, 1994), there are a range of old
dormant periglacial mass movement features in
the landscape that were last active during the
Late-Glacial period (c. 10 000 BP). These features
continue to pose a threat to engineering design
unless correctly identified and fully investigated.
An example of a terrain model highlighting land-
slide activity in this type of landscape is provided
by Croot and Griffiths (2001) for the south-west
of England (Figure 8.13).
Another type of ‘terrain’ found throughout the
world that is particularly landslide prone,
although it has no clearly defined relationship
with the climate, is an eroding coastline (Figures
8.14 and 8.15). Landslides are a major component
in the process of coastal cliff retreat. Not all the
landslides that occur on the coast are caused
solely by sea erosion at the base of cliffs, many
are also the result of weathering or changes to the
water regime in the slope. Lee and Clark (2002)
presents a classification of coastal cliffs that treats
Figure 8.9 Talus cone associated with large-scale rock them as open sediment systems characterised by
slide, Rocky Mountains, British Columbia. inputs, throughputs and outputs of material. The
GFE-08.qxd 6/27/05 2:03 PM Page 193
LANDSLIDES 193
Figure 8.10 The Carrasco translational landslide near Los Molinos, south-east Spain.
cliff system is coupled with the foreshore system failed mass. Debris from the cliff may only
(supply and removal of debris; undercutting etc) reach the foreshore after a sequence of events
and the cliff top (supply of water; inland spread of involving landslide reactivation.
tension cracks etc). Lee and Clark (2002) recog- 2. Composite systems: comprising a partly
nises a range of types of cliff system on the basis coupled sequence of contrasting simple sub-
of the throughput and storage of sediment within systems. The output from one system may
the system (Figure 8.16): not necessarily form input for the next, for
example where material from the upper unit
1. Simple cliff systems: comprising a single falls directly onto the foreshore.
sequence of sediment inputs, from falls or 3. Complex systems: comprising strongly linked
slides, and outputs with limited storage. A sequences of sub-systems, each with their own
distinction is made between cliffs prone to inputs and outputs of sediment. The output
falls and topples and those shaped by simple from one sub-system forms the input for the
landslides. The former is characterised by next. Such systems are often characterised by a
limited storage of sediment within the cliff high level of adjustment between process and
system, with material from the cliff top and form, with complex feedback mechanisms.
face reaching the foreshore in a single event. 4. Relict systems: comprising sequences of pre-
By contrast simple landslide systems com- existing landslide units which are being
prise a single sequence of inputs and outputs gradually reactivated and exhumed by the
with variable amounts of storage within the progressive retreat of the current sea cliff.
GFE-08.qxd 6/27/05 2:04 PM Page 194
Figure 8.11 Landform model for Hong Kong (after Hansen, 1984).
GFE-08.qxd 6/27/05 2:04 PM Page 195
LANDSLIDES 195
Figure 8.12 Land forming processes in Hong Kong (after Styles and Hansen, 1989).
bench development on
in situ Upper Greensand
hollow formed
by spring head
larger, ancient
sapping
landslides
recent
Ancient head deposits
shallow
landslide Clay with flints
solifluction Uppernd
lobes Greensa
incised
tributory bluff Lias
river stream
terrace
terrace
series of degraded
bluff Beds
cut-off Rhaetic landslide blocks
lake forming benches
udstone Group
floodplain Mercia M thin solifluction ancient landslide block
sheet partially buried by later
solifluction deposits
alluvium & head deposits
Recent clay with flints
terrace gravels
River Axe Upper 'Foxmould' -weakly cemented
Greensand sandstone with chert bands
alluvium Lias Interbedded limestones,
shales and mudstones
Rhaetic Beds Interbedded shales
& limestones,
Mercia Mudstone Carbonate mudstones
Group
Figure 8.13 Schematic representation of the landscape in the Axminster area, south-west England (after Croot and
Griffiths, 2001) reprinted with the permission of the Geological Society of London.
GFE-08.qxd 6/27/05 2:04 PM Page 196
Figure 8.14 Mudslides and rotational failure in the West Dorset coast (UK).
Figure 8.15 Major fall of chalk, Alum Bay, Isle of Wight (UK).
GFE-08.qxd 6/27/05 2:04 PM Page 197
LANDSLIDES 197
Figure 8.16 Classification of coastal cliffs (after Lee and Clark, 2002).
GFE-08.qxd 6/27/05 2:04 PM Page 198
For all types of morphoclimatic region, or indication of landslide susceptibility, hazard and
where non-climatically controlled landscapes are risk. For example, the work in Hong Kong
identified such as coasts and mountains, it is (Evans, 1998) established that the highest density
important for the engineering geomorphologist to of landsliding was associated with the weathered
develop landscape models (as exemplified by zone of particular rock types, notably tuffs and
Figures 8.11, 8.13 and 8.16). These allow the engi- lavas, and slope angles between 35° and 40°.
neering geomorphologist to identify potentially Where possible a schematic representation of the
hazardous situations in the field or when using landslide situation can be a very useful tool to
remote sensing data. For landslide investigations illustrate the nature of the landslide problem
the model establishes the typical combination of to clients and non-specialists. An excellent exam-
landscape elements that are likely to result in ‘mass ple of a landslide summary diagram for the
movement’. Thus when undertaking any form of Undercliff landslide on the Isle of Wight, by
geomorphological survey where a terrain model is Geomorphological Services Ltd (Lee and Moore,
available, the geomorphologist is provided with a 1991), is presented in Figure 8.17. This falls into
graphical ‘checklist’ of landscape features. Such the ‘complex system’ category within the Lee
models are also an effective way of illustrating the and Clark (2002) classification of coastal cliffs
scale of the natural hazards in a region, including (Figure 8.16). Similarly, where landslides might
landslides, to less experienced staff and clients. result from the injudicious use of the landscape
This whole subject of developing geological and by humans, a synoptic model can be created that
geomorphological models for engineering is highlights some of the potential landslide sites.
discussed in detail in Fookes et al. (2000). Figure 8.18 illustrates how human activity
affected slopes in South Wales, an area of exten-
Landslide situations sive mining.
In any general study of the distribution of land-
slides for engineering purposes within a given
morphoclimatic region it is necessary to identify 8.7 Landslides: susceptibility,
typical landslide situations that are a function of hazard and risk
the materials, the groundwater conditions, the
particular geological structure or lithology, or Landslide studies in engineering geomorphology
the geomorphological circumstances. Whilst can be very site specific but the track record in
the morphoclimatic landscape models provide the subject has primarily been in the investigation
general indicators of landslide incidence, the of large areas, often for highways (Brunsden
occurrence of individual landslides in specific et al., 1975; TRL, 1997). Whilst detailed surveys
locations or situations is controlled by the local using large-scale geomorphological mapping
conditions. techniques typically at scales of 1 : 5000 or larger
When a single specific landslide is examined are most suited to site-specific investigations,
the geotechnical properties of the materials and when large areas have to be covered at smaller
the hydrogeological conditions become the criti- scales for engineering feasibility studies or plan-
cal factors. For example the shape of the rockfall ning purposes other forms of analysis have to be
in Figure 8.3 is controlled by the discontinuity adopted. Mapping at scales smaller than
patterns in the sandstones that form the river 1 : 25 000 is ideal for the development of general
cliff, although the primary cause was the result of landscape models (see above) and falls generally
undercutting by the river increasing the stress on within the broad category of ‘land surface evalu-
the slope and exceeding the shear strength of the ation’ (Griffiths, 2001). However, over the past
materials. In an engineering geomorphological decade in landslide studies increasing use is
study, the identification of these landslide ‘situa- being made of the techniques associated with
tions’ will allow both the creation of a distribu- hazard and risk assessment both at a site-specific
tion map of existing landslides and provide an and regional level. The definitions of the terms
GFE-08.qxd 6/27/05 2:04 PM Page 199
LANDSLIDES 199
Figure 8.17 Types of contemporary movement in the Undercliff (after Lee and Moore 1991; Moore et al., 1995).
Figure 8.18 Effects of human activity on the stability of South Wales slopes (from Halcrow, 1989, Rhonda
Landslip Potential Assessment Summary Report, for Department of the Environmental and Welsh Office).
GFE-08.qxd 6/27/05 2:04 PM Page 200
Table 8.11 Definitions in risk assessment and management adapted for landslide studies (based on the Royal
Society, 1992).
Landslide susceptibility: slopes, materials and geomorpho-logical situations where landslides are likely to occur.
Landslide hazard: the probability that a landslide event of given magnitude will occur within a given time
period.
Landslide risk: a combination of the probability, or frequency, of occurrence of a defined landslide
hazard and the magnitude of the consequences of the occurrence.
Environmental risk: a measure of the potential threats to the environment which combines the probability
that events will cause or lead to degradation of the environment, and the severity of
that degradation.
Risk criteria: a qualitative and quantitative statement of the acceptable standard of risk with which
the assessed risk needs to be compared.
Landslide risk the integrated analysis of the landslide risks inherent in a region and their
assessment: significance.
Risk quantification: the estimation of a given risk by a statistical and/or analytical modelling process.
Risk evaluation: the appraisal of the significance of a given quantitative (or, when appropriate,
qualitative) measure of risk.
Societal risk: the relation between frequency of occurrence and the number of people in a given
population suffering from a specified level of harm from the specified hazards,
including landslides.
Risk management: the process whereby decisions are made to accept a known or assessed risk and/or the
implementation of actions to reduce the consequences or probability of occurrence.
used in these types of investigation are presented nearly 27 000 landslides. Given this quantity and
in Table 8.11. quality of data a genuine attempt to evaluate the
Landslide susceptibility studies specifically hazard is possible (Evans, 1998). Many previous
attempt to identify the geological and geomorpho- landslide ‘hazard’ studies only really established
logical situations where landslides would most susceptibility or used qualitative statements about
likely be encountered (see above). These analyses the level of ‘hazard’.
can be based on the known distribution of An alternative to hazard analyses having to
landslides and the factors controlling them, or the complete a historical record of landsliding in
distribution of the typical controls on landsliding an area is to establish the relationship between
(e.g. soft clays underlying permeable hard cap- landslide activity and causal factors for which the
rocks). The outcome from these studies will nor- frequency is known. Typically these might mean
mally be landslide inventory maps and databases. investigating the relationship between rainfall
Such analyses are clearly ideal for showing where (intensity, antecedent conditions, totals etc.) or
landslides are likely to occur but they do not seismic activity and landsliding. Normally these
address the question of ‘when’. studies would indicate that landslide conditions
To establish when a landslide might occur (i.e. prevail once certain thresholds are reached but
to define the landslide hazard) it is necessary to they would not necessarily be able to identify
know the landslide frequency. To establish fre- where specific failures could be expected. One
quency requires a record of the historical occur- non-quantitative example of this approach was
rence of landslides in a defined area. One of the presented by TRL (1997). As shown in Figure
few places in the world where this has been pos- 8.19, a combination of maps on slope angle,
sible is Hong Kong where a landslide database physiography aspect and land use distribution was
has been developed that contains the records of compiled in order to produce an ordinal rating of
GFE-08.qxd
0 1 2 0 1 2 0 1 2
kms kms kms
S
F
6/27/05
F S
A A
A A
A
S
F
A
F ROCK Sedimentaries 1.2 3
S
S
A
A TYPE Phyllite 0.8 2
S
S Quartzite 0.3 1
2:04 PM
S
F
S SLOPE North 0.7 1
A S ASPECT South 2.0 2.5
A
S
A
East 2.2 3
West 0.7 1
F
Northwest 0.6 0
Rivers Northeast 0.9 2
Page 201
INCISED FLANK:
Landslides
(differentiated by Steep (>35°) lower valley sides Southeast 0.9 2
mechanism on
original map)
Southwest 0.5 0
STEEP TRIBUTARY:
Moderate (25-35°) valley sides
Quaternary Alluvium Phyllite
OTHER: Shallow (0-25°) spur
Bare alluvium S Scrubland PHYSIO- Incised flank 2.9 3
Quartzite Sedimentaries top or depositional slopes GRAPHY Steep tributary 1.1 2
A Agriculture F Forest
Other 0.3 1
Landslide Distribution and Slope Physiographic
Underlying Geology Classification Land Use Distribution LAND Scrub 1.0 N
USE Agriculture 1.1 O
Forest 0.9 T
0 1 2 0 1 2 0 1 2
SLOPE 0 - 15 0.7 S
kms kms kms
ANGLE 16 - 30 1.0 I
5 31 - 45 1.3 G
6
4 4 5 46 - 60 0.8 N
3
5
8 5
6 > 60 0.4 I
2 7
5
3 8 F
8
4 5 4 3
6
4
4 CHANNEL Stream rank: I
7 6
2 8
6
1
PROXIMITY First Order 0.9 C
8 4 5 6 7 5
7
Second Order 1.1 A
7
6
2 4 1
8 1 8 Third Order 1.5 N
4 7
2 5 4 6 Fourth Order 0.9 T
6 2 6
7 Fifth Order 1.0
2 1 6
1 7 4 Build up of hazard rank for the illustrated
catchment. Hazard ranks for each factor
4 8 1 category are summed for every terrain unit
6
0 - 15
2 4 6
Most stable and assigned to one of three hazard
16 - 30 classes. 3 is the most unstable condition.
31 - 45 Moderate
46 - 60 Flat N.B. The expected number of landslides (E)
Least stable
is calculated on the basis of the percentage
>60
1 North 3 East 5 South 7 West Stacking/climbing study area coverage of each factor
section category multiplied by the total number
2 Northeast 4 Southeast 6 Southwest 8 Northwest Chosen corridor Alluvium
of observed landslides (O) in the study
LANDSLIDES 201
area. Thus, higher O/E ratios indicate a
Slope Angle Distribution Slope Aspect Distribution Summary Landslide greater occurrence of instability than would
Hazard Zones be expected from a random distribution.
Figure 8.19 Landslide hazard mapping for route alignment through an unstable river basin in east Nepal (with acknowledgement to the Transportation
Research Laboratory (TRL), 1997).
GFE-08.qxd 6/27/05 2:04 PM Page 202
landslide hazard for an area in east Nepal. Another hazard for a town in Papua New Guinea based on
example of this type of analysis is provided by the geomorphological mapping, slope classification,
study of the Undercliff landslide at Ventnor on the a hazard classification and previous runout routes
Isle of Wight, Figures 8.20 and 8.21 (Lee et al., (Figure 8.22).
1991a; 1991b). In Figure 8.20 a geomorphologi- The ultimate stage in more regional landslide
cal map of the Undercliff is presented that illus- studies is to establish the landslide risk. This is
trates the complexity of the landslide system. This the relationship between the landslide hazard
was used as the basis for an investigation of the and the human environment. Within the human
relationship between landslide movements and environment it is necessary to establish what is at
rainfall to establish the ‘sensitivity’ of the system risk (people, infrastructure, economic activity),
to rainfall input of different recurrence intervals how vulnerable the components are, and what the
(Figure 8.21). This work identified the compo- consequence would be of any landslide activity
nents of the landslide system that were likely to be (Lee and Jones, 2004). Cruden and Fell (1997)
active even during moderate rainfall conditions, provide a ‘state of the art’ review in this complex
and this information was used to help plan both and rapidly developing field. To date it is rare
new developments and remedial works in the that engineering geomorphological studies have
Undercliff. been able to complete a full landslide risk assess-
The Undercliff example raises another issue in ment except for very specific sites (e.g Lyme
landslide hazard and risk studies, which is the Regis, Dorset (UK), Lee et al., 2000; Hong Kong
potential influence of climatic change on landslide building development, Hardingham et al., 1998).
susceptibility and, hence, hazard. As shown in the To assist in the process of risk assessment the
Undercliff, landslide frequency is closely related problem can be broken down into separate com-
to antecedent rainfall, and the trigger for a major ponents and the probability of occurrence of
failure is often an extreme climatic event, perhaps different events happening can be calculated.
after a long period of gradually increasing suscep- One technique suggested by Wu et al. (1996) to
tibility due to loss of strength through weathering. achieve this is the use of an ‘event tree’, and an
However, current projections of climate change example from the Lyme Regis site in Dorset by
may significantly alter the landslide hazard Lee et al. (2000) is presented in Figure 8.23.
around the globe. For example rising sea levels As part of the risk assessment it is necessary to
and increases in winter storms and rainfall may include an evaluation of the consequences of a
accelerate erosion rates of soft rock cliffs along landslide occurring, either in terms of building
the south coast of England. Thus even where damage or potential for loss of life. Indeed, it is pos-
research has managed to establish the levels of sible to develop consequence models that use a
landslide hazard, future projections will need to rational framework to take into account the key fac-
take into account the potential changes in climatic tors such as runout distance, location and type of
conditions. structures affected, and the spatial and temporal dis-
Another component in landslide hazard studies tribution of the population at risk from a landslide
is calculating how far the landslide is likely to event. This approach was used by Bunce et al.
travel should it occur. Various models of landslide (1997) to back-calculate the risk of the fatalities that
runout have been developed (e.g. Hungr, 1995) occurred on a highway in British Columbia,
but this is a subject that requires further research. Canada. To date the technique appears to have
At present a useful guide to the length of runout mainly been used in hindsight, however the
of landslides in a region is to investigate the travel approach clearly has merit. In Table 8.12 an
angle (Figure 8.2). This can provide an indication ‘Infrastructure Damage Classification’ system is
of the maximum travel distance of landslides of presented (after Geomorphological Services Ltd.,
different types and volumes that can be expected 1991) that could be used in consequence models.
(see Corominas, 1986). TRL (1997) provides an The classification can be used both for surveying
illustration of an attempt to establish the runout existing building and road damage or for estimating
GFE-08.qxd 6/27/05 2:04 PM Page 203
LANDSLIDES 203
Figure 8.20 Geomorphological maps of the Ventnor Undercliff, Isle of Wight (UK), after Lee and Moore (1991).
GFE-08.qxd 6/27/05 2:04 PM Page 204
Figure 8.21 Return period rainfall and susceptibility of landslide systems, Ventnor, Isle of Wight (UK) (Moore
et al., 1995; Lee et al., 1998).
35 35
10
30 25 111 40
45 H
75
109 35
40
149
38 93
20 75 50 30
144 8
35 118
83 74 25
40 45
113
82
45 72 40
35
43 74
45 30 115
40
30
73 H
40 116 14
40 40 ?
40
7
40
69
30 27 40
59 30 71
70%
68 67 16
62
66 30
50% 30 114 119
30
72
49 38/220
30
20 67 15
15
30 15 69 15
20
30 15
45 10 20
47 65
60 68
63
64 5.10 5
5
30 10
70
5 62
61
5 30
60 20 71
46 20
? ?
10
60
5m Bund
5m Bund
4.5m high
Township Township Township Township Township
a) Geomorphology b) Slope classification c) Hazard classification d) Previous landslide runout e) Potential landslide runout
Gradual convex change in slope Darai limestone High hazard (refers to c) only) 50 Failure number
Marked convex break in slope Limestone talus Rockslide
Gradual concave change in slope Dip of strata (dip angle/dip direction) Rock fall/rock avalanche
Marked concave break in slope F Likely fault structure Low hazard Rock roll/rock bounce
Cliff in rock 1 Denotes field note Backscar of failure Predicted runout of failure
Steep slope unit 432 Slope unit based on classification Mass movements H High slope hazard derived from hazard analysis
30 30
Slope angle in degrees Average slope angle in each slope Base of hillside Recommended additional bunding
Erosion scar or active slope ravelling Pnyang mudstone Spring line
Permanent stream Talus and/or boulder colluvium (locally derived)
Spring Mudstone colluvium
Figure 8.22 Landslide hazard and runout mapping in Papua New Guinea (with acknowledgement to the
Transportation Research Laboratory (TRL), 1997).
GFE-08.qxd 6/27/05 2:04 PM Page 205
LANDSLIDES 205
Figure 8.23 Event tree for Lyme Regis, Dorset (UK) (after Lee et al., 2000).
the likely damage resulting for a landslide event of site investigation of the type typically required in
given magnitude and frequency. landslide studies is provided in Figure 8.24 (Fookes,
It should be recognised that as databases 1997). Whilst engineering geomorphologists
improve and geographical information systems should and do become involved with the ground
become more sophisticated, risk assessment will investigation stage, in this Chapter the emphasis is
become part of a routine engineering geomorpho- on the earlier stages of a landslide investigation
logical appraisal for landslide areas (Hearn and where the geomorphologist is primarily using a
Griffiths, 2001). combination of desk studies, remote sensing inter-
pretation and field mapping. Table 8.13 (after
Sowers and Royster, 1978) provides a checklist of
8.8 Planning and undertaking a all the data the geomorphologists should be collat-
landslide investigation ing in their investigation. Initial interpretation of
remote sensing images and photographs, by the
A full-scale site investigation of a landslide would identification of terrain features listed in Table 8.7,
involve desk studies, including aerial photograph provides an invaluable database before going out
interpretation, field reconnaissance, detailed map- into the field. This is the case whether it is the inves-
ping and a ground investigation involving geo- tigation of a single landslide or a 200 km highway
physics, drilling, pitting, instrumentation, sampling through mountainous terrain. However, no matter
and a materials’ testing programme. This would fol- how good the remote sensing interpretation, the
low the long established procedures for site investi- main part of any engineering geomorphological
gation presently covered in the UK by BS 5930: investigation must be in the field. The techniques
1999 (BSI, 1999). The full scope of a preliminary for data collection will be dependent on the scale of
GFE-08.qxd 6/27/05 2:04 PM Page 206
Table 8.12 Infrastructure Damage Classification times the emphasis must not just be on giving
(after Lee and Moore, 1991). geomorphological detail but providing the informa-
tion the client requires for engineering or planning
Damage rating Description
purposes.
Negligible Hairline cracks to roads; pavements
and structures with no appreciable
vertical displacement or separation 8.9 Landslide stabilisation
Slight Occasional cracks; distortion,
separation or vertical displacement Remediation: design options
apparent; small fragments of debris Where a landslide could affect any proposed new
may occasionally fall onto roads development it will be necessary to undertake some
and structures causing light form of remediation, in the form of stabilisation or
damage; repairs are not urgent preventative measures. This generally involves
Moderate Widespread cracks; settlement may some or all of the following (Holtz and Schuster,
cause slight tilt to walls and
1996):
fractures to structural members
and service pipes 1. avoiding the problem
Serious Extensive cracking; settlement may 2. reducing the driving forces
cause open cracks and considerable
3. increasing resisting forces by either applying
distortion to structures; walls out
of plumb and the road surface may
an external force or increasing the internal
be affected by subsidence; parts of shear strength.
roads and structures may be covered
These nature and extent of the various meas-
with landslide debris from above;
repairs urgent to safeguard the ures covered by these three broad categories are
future use of roads and structures summarised in Table 8.14. The selection of the
Severe Extensive cracking; settlement may appropriate engineering measures is crucial to
cause rotation or slewing of road; effective slope stabilisation. The scope of works
gross distortion to roads and may vary in their applicability according to the
structures; repairs will require size and mechanism of failure, the soil and
partial or complete rebuilding groundwater conditions and financial constraints.
and may not be feasible; severe The manner in which remedial measures are
movements likely to lead to the implemented is also important to the success of
abandonment of the site or area
the works as, for example, incorrect placement of
fill and drainage can lead to a reduction in the
factor of safety. It must also be emphasised that
the task and the requirements of the client. the effective stabilisation of most landslides can
Normally some form of map output would be generally only be achieved following a detailed
expected, and the accuracy of these will be totally ground investigation and stability analysis. This
dependent on the quality of the base maps. If a large will have considerable resource implications and,
area is to be covered, serious consideration should if part of a larger development programme, it may
be given to producing a synoptic terrain model delay the project. Whilst these aspects of design
(Fookes et al., 2000). The mapping techniques used are quite rightly mainly the concern of the
should draw upon the best practices of geological geotechnical engineers, engineering geomorphol-
(Barnes, 2004), engineering geological (Dearman, ogists must have an understanding of the scope
1991; Griffiths, 2002) and geomorphological map- and nature of the design process if they are going
ping (Gardner and Dackombe, 1983; Lee, 2001). to provide an effective input.
Where data are available and the client requires The manner in which geomorphology can be
such output, it might be possible to carry out land- included in this design process can be illustrated
slide susceptibility, hazard and risk surveys. At all with two UK case studies.
GFE-08.qxd 6/27/05 2:04 PM Page 207
LANDSLIDES 207
Figure 8.24 Organisation of a typical preliminary site investigation (after Fookes, 1997).
GFE-08.qxd 6/27/05 2:04 PM Page 208
Table 8.13 Checklist for planning a landslide investigation (after Sowers and Royster, 1978).
1. Topography a) contour map (land forms; morphology; anomalous patterns)

b) surface drainage
c) slope profiles (correlation with geology; correlation with contour maps)
d) topographic changes (rate of change over time; correlation with weather,
groundwater, seismic activity)
2. Geology a) geological formations at the site (stratigraphy; superficial deposits;
landslide prone sequences; minerals subject to alteration)
b) structure, i.e. three dimensional geometry (stratification; folding; dip and
strike of all discontinuities including bedding, foliation, joints; faults;
shear zones)
c) weathering (nature and depth)
d) development of the general geological model for the area
3. Groundwater and surface a) piezometric levels (normal; perched; artesian)
water b) variations in pieometric level (correlate with weather; seismic activity;
slope changes)
c) groundwater chemistry (dissolved salts and gases; changes in radioactive
gases)
d) nature and extent of surface water (streams; springs; seepage; ponds;
lakes; tidal reaches)
e) surface indications of subsurface water (seepage zones; areas of poor
drainage; changes in vegetation)
f ) effects of human activity (groundwater utilisation, impoundment,
restriction and recharge; surface water abstraction and recharge; changes
to infiltration rate and capacity, e.g. by covering in concrete or tarmac)
4. History of slope changes a) natural processes (long term geological change; erosion;
past movements; submergence and emergence)
b) human activity (earthworks; changes in surface and groundwater; changes
in surface cover, including paving and deforestation; flooding; rapid
reservoir drawdown)
c) rate of movement (visual accounts; topographic evidence; remote sensing
evidence; instrumentation results; review of historical maps and reports)
d) correlation of movement with other factors (surface and groundwater;
seismic activity; weather; human activity, including earthworks and
vibrations)
e) development of the site-specific landslide ground model incorporating
geology, geomorphology, anthropology and the range of external
processes (Tables 8.5 and 8.6).
5. Weather a) precipitation (type; continuous through to annual rates)
b) temperature (diurnal means and ranges; cumulative degree-day deficit
(freezing index); thaws)
c) barometric changes
6. Vibration a) seismicity (events magnitude, frequency and duration; microseismic
intensity and changes)
b) human-induced (blasting; transport; heavy machinery)
GFE-08.qxd
Table 8.14 Approaches to landslide remediation (after Holtz and Schuster, 1996).
Method Procedure Best application Limitations Remarks
6/27/05
Avoid problem 1. Relocate facility Alternative site off the None if identified during planning Detailed studies of proposed relocation
area of landsliding phase; large cost implications should ensure alternative site is an
if planning and design is improvement
complete or if construction
2:04 PM
has already started
2. Completely or Where small volumes of Can be costly to control Detailed analyses of the stability of the
partially remove excavation are involved excavations; unlikely to be the site are required; depth of excavation
unstable materials best option for large landslides; must be sufficient to ensure firm
Page 209
may not be feasible because foundations
of human factors such as
Rights of Way
3. Install bridge On sidelong slopes with May be costly and needs to Analyses must be performed on
relatively shallow slope provide adequate support anticipated loadings over the long term
failures capacity to withstand the as well as the structural capability
lateral pressures resulting
from a moving landslide
Reduce driving 1. Change exact During preliminary design For highways it will affect
forces location site, phase of project alignment adjacent to landslide
alignment or grade area; for all projects it may
increase land-take requirements
2. Drain surface Appropriate for any scheme Will only correct surface Slope vegetation should be considered
both as part of design and infiltration or seepage to in all cases
in remedial measures surface infiltration
3. Drain subsurface On any slope where lowering Not effective when sliding Stability analyses should include
of groundwater table will mass is impervious consideration of seepage forces
increase slope stability
4. Reduce weight At any existing or potential May require use of lightweight Very careful stability analyses required
landslide materials that may be costly or to ensure excavation or placement of
unavailable; excavation waste lightweight fill will improve stability
LANDSLIDES 209
may create problems; may require
access to land beyond site area
Increase resisting 1. Use buttress and At an existing landslide in May not be effective on deep- Where space is limited reinforced walls
forces by applying counterweight fills; combination with other seated landslides; must be and slopes with geotextiles maybe
external force toe berms methods placed on firm foundations valid alternative
GFE-08.qxd
Table 8.14 (continued).
210
Method Procedure Best application Limitations Remarks
6/27/05
2. Use structural To prevent movement Will not withstand large Stability analyses must incorporate
systems before excavation; where deformations; must be soil–structure interaction
space is limited founded well below the
sliding surface
2:04 PM
3. Install anchors Where space is limited Foundation materials must be In situ strengths of soils and rocks must
(appropriate for both able to resist shear forces by be known; economics dependent on
soils and rocks) anchor tension; can inhibit anchor capacity, depth and frequency
future development on
Page 210
adjacent land
Increase resisting 1. Drain subsurface At any landslide where Requires expertise to install
forces by water table is above the and ensure long-term
increasing shear surface effective operation
internal strength 2. Use geosynthetic For embankments and fill Geosynthetic reinforcement Need to consider stresses imposed on
reinforcement in slopes on steep sidelong must be durable in the reinforcement during construction
backfill ground; in landslide long term
reconstruction
3. Install in situ As temporary structures in Requires long-term durability Design methods still undergoing
reinforcement stiff soils; for small scale of nails, anchors, micropiles development
failures on rock slopes and masonry dentition
4. Use biotechnical On soil slopes of modest Climate: may require Design is by trial and error linked to
stabilisation heights irrigation in dry seasons; local experience
longevity of selected plants
5. Treat chemically Where sliding surface is well- May be reversible; long-term Field installation must evaluate possible
defined and soil reacts effectiveness has not been long-term environmental effects
positively to treatment proved; environmental
stability unknown
6. Use electro-osmosis To relieve excess porewater Only a short-term construction Used when nothing else works; possible
pressures and increase solution as requires application for emergency stabilisation
shear strength at required constant diect current power of landslides
construction rate supply and maintenance
7. Treat thermally To reduce sensitivity of clay Requires expensive and carefully Methods are experimental and costly
soils to action of water designed systems to artificially
dry or freeze subsoils
GFE-08.qxd 6/27/05 2:05 PM Page 211
LANDSLIDES 211
Figure 8.25 Plan and cross-section of the Taren landslide south Wales (after Kelly and Martin, 1985).
A470 crossing of the Taren landslide rises 220 m between toe and crest, and involves
(after Kelly and Martin, 1985; Cobb, 2000) nearly 8 million m3 of material. In the 19th Century
Just south of the village of Aberfan in South Wales a railway and canal had been constructed across the
lies the late-glacial landslide of Taren. During the foot of the landslide but the A470 required far
1980s, Stage IV of the A470 trunk road was con- more extensive earthworks. The investigations for
structed along the upper valley of the River Taff the road in the area of the landslide involved
and this had to cross the Taren landslide and bypass detailed engineering geomorphological mapping
the village of Aberfan. Because of the tragedy in and comprehensive ground investigations over a
Aberfan in 1966 when 144 people, mainly chil- ten-year period with extensive monitoring of
dren, were killed by a flowside failure of a mine- ground movements and porewater conditions.
waste tip (Bishop et al., 1969), there was sensitivity These studies produced a general model (Figure
to any form of construction that might cause land- 8.25) that divided the landslide into three zones
slide movement. The Taren landslide is 640 m long, where the main movements had taken place (upper,
GFE-08.qxd 6/27/05 2:05 PM Page 212
Figure 8.26 Construction works for the Channel Tunnel terminal with the Castle Hill landslide in the foreground,
Folkestone, Kent (UK).
middle and lower zone), and two adjacent areas landslide at Castle Hill near Folkestone in Kent
(the NW block and the northern wing) that had (Figure 8.26). The failure had formed in
been slightly displaced. Inclinometer readings Cretacaeous Lower Chalk overlying Gault Clay.
showed that the main landslide was still moving at Engineering geomorphological mapping of the
3–4 mm/year during the 1970s predominantly landslide (Figure 8.27) showed it had a multiple
along seat-earths in a minor carboniferous coal rotational form along a basal shear surface in
sequence called the Cefn Glas. In order to achieve the Gault Clay. There was inter-digitation of the
a suitable level of safety for the construction of the toe with periglacial coombe-rock deposits, and
road, two techniques were employed: weighting subsequent secondary shallower degradational
through the placement of a major embankment on failures within the main complex (Figure 8.28).
the toe of the landslide, and drainage of the central Inclinometer readings and historic map studies
zone of the landslide via an adit constructed under indicated the landslide had been moving at
the basal shear surface. Construction was com- 1– 2 mm/year for the previous 200 years. The
pleted in 1984 and no movements have been Channel Tunnel alignment required the two
recorded in the landslide since. separate railway tunnels to be driven through the
slide mass. In order to achieve this a combination
UK Channel Tunnel portal at Castle Hill of stabilisation measures and a strictly controlled
(after Griffiths et al., 1995; Varley et al., construction programme was required. A 100 000
1996) tonne toe-weighting berm was placed on the foot
The UK portal to the Channel Tunnel enters of the landslide and three drainage adits were
the ground through a late to early post-glacial driven into the slide to lower groundwater levels.
GFE-08.qxd 6/27/05 2:05 PM Page 213
LANDSLIDES 213
reservoir
10
120
110
reservoir
0
0
13
80
levelled
ground
90
levelled
ground quarry
d
70
Roa
70
Hill
Castle
spoil
12
dump 0
landslide
accumulation 11
0
terrace θ 1 10
0
60
quarry
90
landslide
spoil
accumulation dump 80
terrace θ 2
n in
tio 70
s ec 5
f
e o ure
lin fig
50
60
B
0 100m
LEGEND contour interval 10 metres
Chalk plateau Backscars of secondary

degradational landslides
Chalk scarp slope
Undulating ground associated with
secondary landslide movements
Backscar of the main Valleyside slopes developed in the
Castle Hill landslide Chalk Marl and Gault Clay with
Edge of main landslide variable hillwash cover
movement where identified
Coombe Rock depositional fan
Landslide accumulation zones;
terraces θ 1 and θ 2
Made ground
Front scarps to main landslide
accumulation zones Steep cut and fill slopes in made
ground
head Landslide blocks displaced by
main landslide movement,
in places disrupted by secondary Back face of disused quarries
landslide movements
scarp
Figure 8.27 Geomorphological map of the Castle Hill landslide (after Griffiths et al, 1995).
GFE-08.qxd 6/27/05 2:05 PM Page 214
LEGEND
Disturbed material, shears or shear zones
recorded in boreholes
Type 1 shear zone (associated with lateral
expansion and extrusion)
Type 2 shears -main landslide failure
-liable to be a shear zone
. . Type 3 shears -primarily associated with
.. ..
solifluction activity
NE . Type 4 shears -secondary shallow landslide SW
movements
11b /1-1a
140 Borehole with base shown 140

main
A landslide
backscar
Fill
120 120
bench x x Chalk Marl /Gault boundary
11b /2b
11b /3b
11b /3
100 bench backscar of 100
?
11b /4
secondary
landslide
? road
80 quarry 80
spoil dump
?
. upper
?
m. AOD
m. AOD
CHALK MARL terrace

? lower
60 terrace 60
x x x x x x x
x x x ? . .
.. .. B
? .. ..
40 GAULT 40
20 20
scarp backscarp multiple rotational /block disruption zone main landslide Coombe Rock accumulation
accumulation zone terrace- possibly interdigitated
0 with landslide debris 0
-10 -10
50 100 150 200 250 300 350 metres
NO VERTICAL EXAGGERATION
Figure 8.28 Cross-section of the Castle Hill landslide Folkestone, Kent (UK) (after Griffiths et al., 1995) reprinted
with the permission of the Royal Geographical Society.
A carefully phased ‘top-down’ construction the landscape and provide data on natural land-
approach was adopted to minimise changes in slides through analyses of the terrain and field
overburden pressure, and ground movements observation. Such studies must precede any form
were closely monitored throughout construction. of detailed ground investigation if it is to be
Whilst increased movements were recorded designed in a cost-effective manner, and by iden-
during construction over the period 1989–90, on tifying potential landslide problems at an early
completion of the works the landslide stabilised stage in a project it may be possible to alter sites
and no further activity has been recorded. or change alignments in order to avoid major
problem areas. The engineering geomorphologist
8.10 Conclusion is also able to provide the spatial context within
which the subsequent site investigation results
As a landscape feature, landslides in either natu- can be interpreted. In addition engineering
ral or artificial slopes are a well-recognised engi- geomorphology is now developing the tools for
neering problem. Geotechnical engineers have landslide hazard and risk assessment that will be
developed sophisticated methods for modelling and able to be incorporated into planning and feasibil-
analysing the stability of slopes as part of normal ity studies for all types of development project.
earthworks design. These analyses are predomi- Landslides are a geomorphological process of
nantly based on detailed ground investigations landscape development. To deal with problems
and testing of materials. In contrast, as a conse- arising from slope failures, geomorphologists
quence of their training, engineering geomorpholo- have expertise and experience of immense value
gists are more able to take a broader-scale view of and importance to the construction industry.
GFE-08.qxd 6/27/05 2:05 PM Page 215
LANDSLIDES 215
References Dearman, W. R. (1991) Engineering Geological

Mapping. Butterworth-Heinemann, Oxford, 387pp.
Barnes, J. W. (2004) Basic Geological Mapping (4th Dikau, R., Brunsden, D., Schrott, L. and Ibsen, M.-L.
edition). Wiley, Chichester, 184pp. (1996) Landslide Recognition. Wiley, Chichester,
Bishop, A. W., Hutchinson, J. N., Penman. A. D. M. and 251pp.
Evans, H. E. (1969) Geotechnical investigation into Evans, N. C. (1998) The natural terrain landslide study.
the causes and circumstances of the disaster of 21 In Li, K. S. Kay, J. N. and Ho, K. K. S. (eds) Slope
October 1966. In A Selection of Technical Reports Engineering in Hong Kong. Balkema, Rotterdam,
Submitted in the Aberfan Tribunal. Welsh Office, 137–143.
HMSO, London, 1–82. Fookes, P. G. (1997) Geology for engineers: the geo-
Brabb, E. E. (1991) The world landslide problem. logical model, prediction and performance.
Episodes 14, 52–61. Quarterly Journal of Engineering Geology 30,
Bromhead, E., Dixon, N. and Ibsen, M.-L. (eds) (2000) 293–424.
Landslides in Research, Theory and Practice. Fookes, P. G., Baynes, F. and Hutchinson, J. N. (2000)
Thomas Telford, 1684pp. Total geological history: a model approach to
Brunsden, D. (1993) Mass movement: the research understanding site conditions. Proceedings of
frontier and beyond: a geomorphological approach. GeoEng2000, Melbourne 1, 370–460.
Geomorphology 7, 85–128. Gardner, V. and Dackombe, R. (1983) Geomorphological
Brunsden, D., Doornkamp, J. C., Fookes, P. G., Jones, Field Manual. Allen & Unwin, London, 254pp.
D. K. C. and Kelly, J. M. H. (1975) Large-scale Geomorphological Services Ltd (1991) Coastal landslip
geomorphological mapping for highway engineer- potential, Isle of Wight Undercliff, Ventnor. Technical
ing. Quarterly Journal of Engineering Geology 8, Report to the Department of the Environment,
227–253. Research Contract PECD 7/1/272, 68pp. ˆ
BSI (British Standards Institute) (1999) BS 5930: Code Godt, J. W. and Savage, W. Z. (1999) El Nino
of Practice for Site Investigations. London. 1997–98: direct costs of damaging landslides in the
Bunce, C. M., Cruden, D. M. and Morgensten, N. R. San Francisco Bay Region. In Griffiths, J. S.,
(1997) The assessment of the hazard of rock fall on Stokes, M. R. and Thomas, R. G. (eds) Landslides.
a highway. Canadian Geotechnical Journal 34, Balkema, Rotterdam, 162pp.
344–356. Griffiths, J. S. (ed.) (2001) Land Surface Evaluation
Chorley, R. J., Schumm, S. A. and Sugden, D. E. (1984) in Engineering Practice. Geological Society Engin-
Geomorphology. Methuen, London, 605pp. eering Geology Special Publication No.18, 249pp.
Cobb, A. E. (2000) Taren landslide: road construction Griffiths, J. S., Brunsden, D., Lee, E. M. and
across a landslide. In Siddle, H. J., Bromhead, E. N. Jones, D. K. C. (1995) Geomorphological investiga-
and Bassett, M. G. Landslides and Landslide Mana- tions for the Channel Tunnel terminal and portal.
gement in South Wales. National Museum & Galleries The Geographical Journal 161, 275–284.
of Wales, Geological Series No.18, Cardiff, 97–98. Guadagno, F. M. and Perriello Zampelli, S. (2000)
Cooke, W. H. (1882) Hereford. John Murray, London. Triggering mechanisms of the landslides that inun-
Corominas, J. (1986) The angle of reach as a mobility dated Sarno, Quindici, Siano and Bracigliano (S.
index for small and large landslides. Canadian Italy) on May 5–6, 1998. In Bromhead, E., Dixon, N.
Geotechnical Journal 33, 260–271. and Ibsen, M.-L. (eds) (2000) Landslides in Research,
Croot, D. and Griffiths, J. S. (2001) Examples of Theory and Practice. Thomas Telford, 671–676.
periglacial activity in south-west England and their Hansen, A. (1984) Engineering geomorphology: the
engineering geological significance. Quarterly application of an evolutionary model of Hong Kong
Journal of Engineering Geology & Hydrogeology terrain. Zeitschrift für Geomorphologie, supplemen-
34, 269–282. tary volume 51, 39–50.
Cruden, D. M. (1991) A simple definition of a land- Hardingham, A. D., Ho, K. K. S., Smallwood, A. R. H.
slide. Bulletin of the International Association of and Ditchfield, C. S. (1998) Quantitative risk
Engineering Geology 43, 27–29. assessment of landslides – a case study from Hong
Cruden, D. and Varnes, D. J. (1996) Landslide types Kong. In Li, K. S., Kay, J. N. and Ho, K. K. S. (eds)
and processes. Chapter 3 in Turner, A. K. and Slope Engineering in Hong Kong. Balkema,
Schuster, R. L. (eds) Landslides Investigation and Rotterdam, 145–152.
Mitigation. Special report 247 of the Transportation Hearn, G. J. and Griffiths, J. S. (2001) Landslide haz-
Research Board, National Research Council, ard assessment. In Griffiths, J. S. (ed.) Land Surface
National Academy Press, Washington DC. Evaluation in Engineering Practice. Geological
Cruden, D. and Fell, R. (eds) (1997) Landslide Risk Society Engineering Geology Special Publication
Assessment. Balkema, Rotterdam, 371pp. No.18, 43–52.
GFE-08.qxd 6/27/05 2:05 PM Page 216
Holtz, R. D. and Schuster, R. L. (1996) Stabilisation of Lee, E. M., Brunsden, D. and Sellwood, M. (2000)
soil slopes. Chapter 17 in Turner, A. K. and Schuster Quantitative risk assessment of coastal landslide
R. L. (eds) Landslides Investigation and Mitigation. problems, Lyme Regis, UK. In Bromhead, E.,
Special report 247 of the Transportation Research Dixon, N. and Ibsen, M. L. (eds) Landslides in
Board, National Research Council, National Research, Theory and Practice. Thomas Telford,
Academy Press, Washington DC. London, 899–904.
Hungr, O. (1995) A model for the runout analysis of Mather, A. E., Griffiths, J. S. and Stokes, M. R. (2003)
rapid flow slides, debris flows and avalanches. Anatomy of a ‘fossil’ landslide from the Pleistocene
Canadian Geotechnical Journal 32, 610–623. of SE Spain. Geomorphology 50, 135–149.
Hutchinson, J. N. (1988) General report: morphological Moore, R., Lee, E. M., and Clark, A. R. (1995) The
and geotechnical parameters of landslides in rela- Undercliff of the Isle of Wight: a review of ground
tion to geology and geohydrology. Proceedings of behaviour. South Wight Borough Council.
the 5th International Symposium on Landslides, Rib, H. T. and Liang, T. (1978) Recognition and
Lausanne 1, 3–35. identification. Chapter 3 in Schuster, R. L. and
Hutchinson, J. N. (2001) Reading the ground: morphol- Krizek, R. J. (eds) Landslides – Analysis and Control.
ogy and geology in site appraisal. Quarterly Journal National Academy of Sciences, Washington DC,
of Engineering Geology and Hydrogeology 34, 5–50. 33–79.
Jones, D. K. C. and Lee, E. M. (1994) Landsliding in Royal Society (1992) Risk: Analysis, Perception and
Great Britain. HMSO, London. Management. Report of the Royal Society Study
Keaton, J. R. and DeGraff, J. V. (1996) Surface obser- Group, The Royal Society, London.
vation and geologic mapping. In Turner A. K. and Schultz, Ch. H. and Harper, J. A. (1996) Pittsburgh
Schuster R. L. (eds.) Landslides: Investigation and red beds cause renewed landsliding after a c.
Mitigation. Transportation Research Board, Special 310 Ma pause. In Chacon, J., Irigaray, C. and
Report 247, National Research Council, National Fernandez, T. (eds) Landslides. Balkema,
Academy Press, Washington DC, 178–230. Rotterdam, 189–196.
Kelly, J. M. H. and Martin, P. L. (1985) Construction Schuster, R. L. (1995) The 25 most catastrophic land-
works on or near landslides. In Morgan, C. (ed.) slides in the 20th century. In Chacon, J., Irigaray, C.
Proceedings of the Symposium on Landslides in South and Fernandez, T. (eds) Landslide. Balkema,
Wales. Polytechnic of Wales, Pontypridd, 85–102. Rotterdam, 53–62.
Li, K. S., Kay, J. N. and Ho, K. K. S. (1998) Slope Eng- Styles, K. A. and Hansen, A. (1989) Geotechnical area
ineering in Hong Kong. Balkema: Rotterdam, 342pp. studies programme: Territory of Hong Kong. GASP
Lee, E. M. (2001) Geomorphological mapping. Report XII, Geotechnical Control Office, Civil
InÿGriffiths, J. S. (ed.), Land Surface Evaluation in Engineering Services Department, Government of
Engineering Practice. Geological Society Engineer- Hong Kong, 346pp.
ing Geology Special Publication No.18, 53–56. Smith, K. (2001) Environmental Hazards: Assessing
Lee, E. M. and Moore, R. (1991) Coastal landslip Risk and Reducing Disaster (3rd edition). Routledge,
potential assessment: Isle of Wight Undercliff, London, 392pp.
Ventnor. Department of the Environment. Soeters, R. and van Westen, C. J. (1996) Slope
Lee, E. M. and Clark, A. R. (2002) The Investigation instability recognition. Chapter 8 in Turner, A. K.
and Management of Soft Rock Cliffs. Thomas and Schuster, R. L. (eds) Landslides Investigation
Telford, London, 392pp. and Mitigation. Special report 247 of the
Lee, E. M. and Jones, D. K. C. (2004) Landslide Risk Transportation Research Board, National Research
Assessment. Thomas Telford, London, 454pp. Council, National Academy Press, Washington DC.
Lee, E. M., Doornkamp, J. C., Brunsden, D. and Sowers, G. F. and Royster, D. L. (1978) Field inves-
Noton, N. H. (1991a) Ground movement in tigation. Chapter 4 in Schuster, R. L. and Krizek,
Ventnor, Isle of Wight. Report for the Department R. J. (eds) Landslides – Analysis and Control.
of the Environment. National Academy of Sciences, Washington DC,
Lee, E. M., Moore, R., Brunsden, D. and Siddle, H. J. 80–111.
(1991b) The assessment of ground behaviour at TRL (Transportation Research Laboratory) (1997)
Ventnor, Isle of Wight. In Chandler, R. J. (ed.) Slope Principles of low cost road engineering in moun-
Stability: Engineering Developments and tainous regions. Overseas Road Note 16. TRL,
Applications. Thomas Telford, London, 219–225. Crowthorne, 149pp.
Lee, E.M., Moore, R., and McInnes, R. G. (1998) Turner, A. K. and Schuster, R. L. (eds) (1996)
Assessment of the probability of landslides reacti- Landslides Investigation and Mitigation. Special
vation: Isle of Wight Undercliff, UK. In Moore D. report 247 of the Transportation Research Board,
and Hungr O. (eds.) Engineering Geology: The View National Research Council, National Academy
from the Pacific Rim, 1315–1321. Press, Washington DC.
GFE-08.qxd 6/27/05 2:05 PM Page 217
LANDSLIDES 217
UNESCO World Landslide Inventory (1990) A sug- Turner, A. K. and Schuster, R. L. (eds) Landslides
gested method for reporting a landslide. Bulletin of Investigation and Mitigation. Special report 247 of the
the International Association of Engineering Transportation Research Board, National Research
Geology 41, 5–16. Council, National Academy Press, Washington DC.
UNESCO World Landslide Inventory (1993) A sug-
gested method for describing the activity of a
landslide. Bulletin of the International Association Acknowledgements
of Engineering Geology 47, 53–57.
Varnes, D. J. (1978) Slope movement types and pro- The author wishes to thank Prof. Peter Fookes for
cesses. Chapter 2 In Schuster, R. L. and Krizek, R. permission to use Figure 8.24; Dr Mark Lee for
J. (eds) Landslides – Analysis and Control. National
Academy of Sciences, Washington DC, 11–33. permission to use Figures 8.16, 8.17, 8.20, 8.21
Varley, P. M., Warren, C. D. and Avgherinos, P. (1996) and 8.23; the Geological Society of London for
Castle Hill west landslip. In Harris, C. S., Hart, M. permission to use Figure 8.14; the Transportation
B., Varley, P. M. and Warren, C. D. Engineering Research Laboratory for permission to use Figures
Geology of the Channel Tunnel. Thomas Telford, 8.19 and 8.22, and the Royal Geographical Society
London, 295–309.
Wu, T. H., Tang, W. H. and Einstein, H. H. (1996) for permission to use Figures 8.27 and 8.28. The
Landslide hazard and risk assessment. Chapter 6 in photographs have all been supplied by the author.
GFE-09.qxd 6/27/05 4:49 PM Page 218
9. Active Tectonic Environments and

Seismic Hazards
James V. Hengesh and William R. Lettis
9.1 Introduction Mw 7.72001 earthquakes that occurred in north-

western India (Hengesh and Lettis, 2002), and the
Geological hazards are greatest along plate M ⬃ 8 1755 Lisbon earthquake that occurred in
boundary regions of the world due to high rates Portugal (Johnston, 1996).
of seismic and volcanic activity and dynamic It should be noted that:
geological and geomorphological processes
(Figure 9.1). Although some major cities such as 1. maximum earthquake magnitudes in stable
San Francisco, California or Wellington, New continental regions are of similar size to maxi-
Zealand are built along active plate boundaries, mum earthquake magnitudes in plate boundary
geological hazards in these areas are being miti- regions (excluding subduction zones)
gated through hazard mapping programs, and 2. larger areas of stable continental regions are
implementation and enforcement of geological affected by earthquake strong ground shaking
hazards characterisation and seismic design in than areas of plate boundary regions for com-
building code provisions. However, vast popula- parable sized earthquakes
tions in cities and rural areas of developing coun- 3. recent studies have identified short earthquake
tries live along highly active plate boundaries recurrence intervals in some stable continen-
and lack hazard information, building codes, or tal regions (Kelson et al., 1996; Tuttle et al.,
code enforcement to adequately mitigate risk 2002a).
(e.g. Indonesia, the Mediterranean, Asia, Latin
and South America). Earthquake ‘Risk’ for engineered
Geological hazards also are important consid- structures and systems
erations within stable continental regions Recent earthquakes in Turkey, Taiwan and India in
(Johnston, 1996; Bakun and McGarr, 2002). 1999 and 2001 dramatically illustrate the damag-
However, the processes that produce earthquakes, ing effects of strong ground shaking, permanent
i.e. strain accumulation and release, are less well ground deformation and liquefaction on struc-
understood in stable continental regions than tures, buildings, and lifelines, as well as the nega-
plate boundary regions (Kenner and Segall, tive impact on the local economy (Figure 9.2;
2000), and therefore, there is greater uncertainty EERI 2000, 2001, 2002). As in many previous
in estimates of hazard and risk in these areas. earthquakes, these three events produced damage
Despite uncertainties in our understanding of the as a result of surface fault rupture, strong ground
earthquake generation process in stable continen- shaking, amplified ground shaking, liquefaction,
tal regions, a number of factors indicate that high and ground settlement.
levels of risk exist for some parts of these areas. The term ‘risk’ is commonly used within
Historical examples of large magnitude earth- earthquake-related studies, but often has no clear
quakes within stable continental regions include definition. In this chapter the term ‘risk’ has a very
the three M ⬃ 7.3 to 8.1 events that occurred in specific definition. Risk describes the losses that
the central United States within a three month may occur during an earthquake. Losses can be
period between December 1811 and February measured in a variety of ways, such as: (1) direct
1812 (Tuttle et al., 2002a), the M ⬃ 8 1819 and economic losses resulting from damaged buildings
218
GFE-09.qxd 6/27/05 4:49 PM Page 219
ACTIVE TECTONIC ENVIRONMENTS AND SEISMIC HAZARDS 219
90ºN
60ºN
30ºN
0º
30ºS
60ºS
DEPTH OF FOCUS
0 – 70 km
70 – 300 km Lamant-Doharty Earth observatory
300 – 700 km of Columbia University from Harvard CMT Data
90ºS
90ºE 120ºE 150ºE 180º 150ºW 120ºW 90ºW 60ºW 30ºW 0º 30ºE 60ºE 90ºE
Figure 9.1 Map showing global seismicity that defines major plate boundaries. Note the lack of seismicity within
the continental interiors compared to seimicity along plate boundaries. Black circles represent shallow seismicity,
light gray clusters represent intermediate depth seismicity. Map shows events ⬎Mw 5.5 for period 1977 to 1992.
or other facilities; (2) indirect losses resulting from area, and the distribution of people and economic
business interruption and economic impacts; networks (McGuire, 2004).
(3) injuries and fatalities; and (4) impact on the
environment. Types of seismic hazards
Risk, as described above, is actually the prod- Earthquake-related geological hazards include:
uct of three important parameters including: geo-
logical hazard, structural vulnerability, and 1. primary surface fault rupture i.e. the sudden
socio-economic factors. This can be thought of displacement of adjacent crustal blocks along
conceptually as: a fault which may cause permanent displace-
ment of the ground surface (Figure 9.2)
Risk ⫽ Hazard ⫻ Vulnerability ⫻ Consequence 2. strong ground shaking
3. liquefaction and related ground deformation
where the term ‘hazard’ describes parameters such (Figure 9.2)
as level of ground shaking, liquefaction potential, 4. seismically induced slope failure
site response, surface fault rupture, and slope 5. tsunami.
instability. The term vulnerability describes the
susceptibility of buildings or lifeline systems to Strong ground shaking and tsunami are seismic
damage from the effects of earthquakes, and con- waves generated from the primary fault rupture
sequence reflects factors such as damage, death/ and ground displacement, and are measured in
injury and economic performance. Therefore, terms of their frequency, amplitude, velocity, par-
assessing ‘risk’ requires information on geological ticle displacement, and acceleration. Liquefaction
hazards, vulnerability of structures in a specific and seismically induced slope failure, in turn, are
GFE-09.qxd 6/27/05 4:49 PM Page 220
of techniques often are used to integrate

these observations and develop interpretations
that can be translated into engineering terms
for use in seismic hazard assessments during
the feasibility, siting, design, or construction
phases of major projects (Gutenberg and
Richter, 1954; Cornell, 1968; Molnar, 1979;
Schwartz and Coppersmith, 1984; Youngs and
Coppersmith, 1985; Weznouski, 1988; Krinitsky
and Slemmons, 1990; SSHAC, 1995 and Yeats
et al., 1997).
Fault characteristics most commonly required
for engineering projects include:
1. location of fault or tectonic province

2. fault type, i.e. strike-slip, reverse, normal or
combination
3. fault length and seismogenic width (vertically
through the crust)
4. width of a fault zone on ground surface
5. earthquake magnitude
6. coseismic displacement
Figure 9.2 Photographs of damage from the 1999 7. fault slip rate
Turkey earthquake. (A) Collapse of apartment buildings 8. earthquake recurrence interval
due to surface fault rupture and near-field strong 9. recency of fault movement.
ground shaking. Note approximate 5-metre offset of
concrete wall. (B) 500 metres of waterfront property
catastrophically submerged due to liquefaction induced For seismic hazard analysis projects this infor-
lateral spreading by translation/rotation. Ismit Bay, mation is used to develop a seismic source model,
Turkey. Photographs by James Hengesh. which may include a combination of individual
fault segments (referred to as line sources or pla-
nar sources if dipping), seismotectonic provinces
produced as seismic waves propagate through (areal source zones), and point sources. Each line,
soils (Figure 9.2). plane, point, or areal source zone in a seismic
source model must be assigned coordinates, maxi-
mum earthquake magnitude, and earthquake
9.2 Earthquake occurrence and recurrence model.
magnitude scaling
Earthquake mechanics
Identifying and characterising active faults is Large damaging earthquakes occur along plate
essential both for the development of earthquake boundaries and within continental interiors
design ground motions and the evaluation of (Figure 9.1). Plate boundary earthquakes occur
site-specific surface fault rupture hazards. due to the concentration of stress along the edges
Seismic geologists attempt to link observations of crustal-scale plates, or blocks, that define
such as location and type of tectonic landforms, planes of weakness known as faults. The stress is
locations of instrumentally recorded earthquakes, caused when adjacent crustal-scale plates or
and geodetic measurements of crustal plate blocks have different rates and directions of rela-
motion with processes occurring deep within the tive motion (Figure 9.3). Stress is accommo-
earth’s crust that generate earthquakes. A number dated by elastic strain of large crustal blocks, and
GFE-09.qxd 6/27/05 4:49 PM Page 221
Figure 9.3 Continental extrusion in the eastern Mediterranean. The Aegean-Anatolia block is escaping westward
from the Arabia-Eurasia collision zone toward the Hellenic subduction zone. Current velocity vectors relative to
Eurasia in mm/yr (arrows), using GPS (Global Positioning System) and SLR (satellite laser ranging) are from
Reilinger et al. (1977). In the Aegean, the westward propagation of the North Anatolian fault is associated with
localised and rapid transtension (Armijo et al., 1996). CR, Corinth Rift; NAT, North Aegean Trough; NAF North
Anatolian fault; K, Karliova triple junction; EAF, East Anatolian fault; DSF, Dead Sea fault. Modified from Hubert-
Ferrari et al. (2002).
distributed brittle faulting along the margins of the accumulation of stress (measured as seismic
blocks. moment) along three alternative fault segments of
An earthquake occurs when the dynamic stress the Aleutian subduction zone (e.g. single-block,
exceeds the static frictional stress of the fault double-block, of triple-block ruptures). Once the
plane, and the two adjacent plates slip with a dis- static stress (expressed as characteristic seismic
crete displacement across a fault plane (Figure moment) of the fault plane is exceeded, coseismic
9.4). As discussed below, the size of an earth- displacement occurs and produces an earthquake,
quake is related to the amount of stress released the seismic moment is released as energy, and the
along the displacement, (␴o ⫺ ␴f on the lower process of accumulating stress repeats itself.
diagram of Figure 9.4). The rate of earthquake Earthquakes within plate interiors also are due
occurrence is related to the rate that stress accu- to the concentration of stress along fault planes,
mulates across the plate boundary, and is propor- however, the model of strain accumulation and
tional to the velocity differential between the two release differs dramatically. In plate interiors,
plates. Therefore, because earthquake occurrence earthquakes may occur from residual stress fol-
rates are related to the plate velocity differential lowing prior tectonic episodes, or due to the
across a fault plane, many faults in plate bound- mechanical concentration of broad regional stress
ary regions follow a time dependent earthquake on localized zones of weakness, or due to non-
occurrence model of strain accumulation and tectonic processes such as isostatic loading or
release (Kanamori and Brodsky, 2004). Figure 9.5 unloading (EPRI, 1994; Kenner and Segall,
illustrates the time dependent earthquake occur- 2000). The concentration and release of strain
rence model for a portion of the Aleutian subduc- energy within plate interiors may not follow a
tion zone that produced the 1965 Mw 8.7 Rat simple time-dependent model and as such, there
Island earthquake. The model shows a constant is greater uncertainty in the location, magnitude
GFE-09.qxd 6/27/05 4:49 PM Page 222
and frequency of these events. Strain rates for

plate interiors may be several orders of magnitude
lower than for plate boundaries; however, areas
such as the New Madrid Seismic Zone in the
central United States (e.g. Tuttle et al., 2002a;
Obermeier et al., 2001; Kenner and Segall, 2000)
have earthquake recurrence intervals that are
similar to those of active plate boundary regions
during the last 2000 years or more. This points out
a major source of uncertainty as geologists grap-
ple to understand the earthquake occurrence
process in stable continental regions.
Earthquake magnitude
Because it is difficult to directly measure the
stress released during an earthquake, the amount
of coseismic displacement, slip area, and rigidity
of crustal blocks are used to estimate the energy
released (or moment) during an earthquake (Hanks
and Kanamori, 1979; Kanamori and Brodsky,
2004). The moment magnitude scale is now the
most commonly used measure of earthquake
magnitude in earthquake engineering studies.
Whereas previous earthquake magnitude scales
measure a particular type of ground motion (e.g.
surface wave (Ms) or body wave (Mb) magni-
tudes), the moment magnitude (Mw) scale is a
physical measure of the size of an earthquake
(Kanamori, 1983). The energy released or moment
(Mo) of an earthquake is defined by:
Mo ⫽ AD␮
where (A) ⫽ rupture area, (D) ⫽ displacement,

and (␮) ⫽ crustal rigidity. The moment magni-
tude of an earthquake (Hanks and Kanamori,
1979; Kanamori and Brodsky, 2004) is related to
seismic moment by the equation:
Figure 9.4 This figure illustrates the response of the
earth’s crust to tectonic forces. Equal, but oppositely Mw ⫽ log Mo ⫺ 16.1/1.5
directed forces act tangent to the light and dark shaded
planes on the boundaries of the slab. The forces cause
stress, which is defined as the magnitude of the force of the fault plane, ␴o. When an earthquake occurs, plates
divided by the area of the planes. As stress occurs within on opposite sides of the fault experience a relative dis-
the slab, the dark shaded plane is displaced a distance y placement (D) over an area (A) (middle diagram). As
relative to the light colored plane, and the distance displacement (D) continues, the frictional stress drops
between the two planes is x; strain of the slab is defined from the original static frictional stress ␴o to a lower
as y/x (upper diagram). An earthquake occurs when kinetic stress ␴f, shown on the lower diagram. Figure
stress on a fault plane exceeds the static frictional stress modified from Kanamori and Brodsky, 2004.
GFE-09.qxd 6/27/05 4:49 PM Page 223
Figure 9.5 Hypothetical rupture model for the Aleutian Subduction Zone near Rat Islands, Alaska. Model illus-
trates the process of stress accumulation and release along the subduction zone plate interface. The top of each
triangle represents the characteristic moment for each characteristic segment rupture. The sloping line represents the
constant rate of stress accumulation, or seismic moment rate, across the Aleutian subduction zone. Once the
accumulated moment reaches the characteristic moment, the strength of the fault plane is exceeded and a fault
displacement and earthquake occur. During a particular earthquake cycle, the characteristic event may involve a
single-block, double-block, or triple-block rupture. Because the characteristic moment varies by the size of the
rupture segment, the recurrence time (T) varies for each segment rupture.
The moment magnitude scale is preferred for

engineering studies because it is stable through-
out the entire range of earthquake magnitudes,
whereas other magnitude scales are appropriate
for a limited range of earthquake magnitudes, or
particular geographic regions. Furthermore, the
moment magnitude scale is used in empirical
relationships among fault rupture dimensions
and magnitude (Wells and Coppersmith, 1994),
and most modern ground motion attenuation Figure 9.6 Simple shear associated with strike-slip
equations (SRL, 1997; EPRI, 2003). faulting produces preferred orientation of fractures,
faults, and folds, as well as extensional and contrac-
tional landforms. Figure modified from Burbank and
Anderson, 2001. Examples of fault types shown on
Figure 9.7.
9.3 Types of fault displacement
The coseismic displacement depicted in Figure region undergoing simple shear. The repeated
9.4 can occur along different types of faults (e.g. occurrence of these displacements and long term
strike-slip, normal, reverse, or subduction zone) response of the earth’s crust to regional stress pat-
depending on the orientation of the fault with terns, climatic conditions and surficial geological
respect to the regional stress field. Figures 9.6 and processes also forms characteristic landforms,
9.7 show the different types of faults and the ori- such as linear valleys, faceted mountain fronts, or
entation in which they form with respect to a fold and thrust belts (Figure 9.8). Principals of
GFE-09.qxd 6/27/05 4:49 PM Page 224
fault. These faults are important sources of

large damaging earthquakes in many trans-
current or transform plate boundary settings,
examples of which include the San Andreas,
North Anatolian, Great Sumatra, and Alpine
fault systems.
Strike-slip faults are very high angle to verti-
cal structures and as such form very straight
fault traces along the ground surface. As shown
on Figure 9.8, strike-slip faults form a number of
characteristic landforms due to the lateral trans-
lation of adjacent crustal blocks on opposite
sides of the fault. These may include offset or
deflected drainages, offset ridges, sag ponds,
shutter ridges, and/or fault scarps (Figure 9.8).
Due to high slip rates, faults such as the San
Andreas often form pronounced linear valleys
(Figure 9.9).
Complexities along the trace of a strike-slip
fault produce secondary features that vary from
metre to kilometre scale. Releasing bends or
stepovers form pull-apart basins due to localised
extension within the fault zone (Figure 9.10a).
During the 1999 Turkey earthquake subsidence
within pull-apart basins caused dramatic coastal
subsidence and extensive damage (Figure 9.11).
Restraining bends or stepovers form localised
uplift, folding and/or faulting due to contraction
within the fault zone (Figure 9.10b). Perhaps
one of the best examples of a restraining bend is
Figure 9.7 Types of faults and illustrations of fault
slip. The preferred orientations of each fault type with the ‘Big Bend’ along the San Andreas Fault in
respect to the regional stress direction are shown on southern California (Figure 9.12; Yeats et al.,
Figure 9.6. 1994).
Earthquakes along strike-slip faults form a rela-
tively narrow zone of permanent ground deforma-
tectonic geomorphology are used to evaluate tion (Figure 9.13). Primary surface fault ruptures
landforms, locate faults, and characterise poten- on strike-slip faults cause a zone of ground crack-
tial earthquake magnitude and recurrence intervals ing, bulging and tearing of near surface materials
(e.g. Yeats et al., 1997; Burbank and Anderson, that is commonly 5- to 15-metres wide.
2001).
Normal faults
Strike-slip faults Normal faults have dip-slip displacement in
A strike-slip fault is a fault along which most which the hanging wall has moved downward rel-
movement is parallel to the fault strike (Figures ative to the footwall (Figure 9.7). These faults are
9.6 and 9.7; Bates and Jackson, 1987). Strike- important sources of large damaging earthquakes
slip faults are either right-lateral (dextral) or left in continental rift zones and extensional tectonic
lateral (sinistral) depending on the relative provinces. Examples of continental rift zones
motion of the block on the opposite side of the include the Rio Grande New Mexico, East African,
GFE-09.qxd 6/27/05 4:49 PM Page 225
Figure 9.8 Examples of landforms associated with active strike-slip faulting. Modified from Wesson et al. (1975).
Figure 9.9a Sag pond and side-hill bench along North Anatolian fault. Photograph courtesy of Professor Aykut
Barka.
and Taupo New Zealand rift zones; examples of Normal faults are moderate to steeply dip-
extensional tectonic provinces include the Rhine ping structures and, as such, fault traces typically
Graben, Basin and Range province, and Tibetan form irregular and discontinuous traces across
Plateau rift zone. the ground surface (Figure 9.14a). Vertical
GFE-09.qxd 6/27/05 4:49 PM Page 226
Figure 9.9b Linear valley formed along the Gualala segment of the San Andreas fault. From Wallace, R.E., ed. (1990).
glacial moraines, linear mountain fronts, triangular

facets, and/or fault scarps (Figure 9.15).
Complexities along the traces of normal faults
form secondary features that vary from metre to
kilometre scale, and include relay ramps, transfer
faults, synthetic and antithetic faults, and grabens
(Figure 9.14). During the 1956 Dixie Valley/
Fairview Peak, 1959 Hebgen Lake, 1915 Pleasant
Valley, and 1983 Borah Peak earthquakes,
ground rupture occurred along multiple fault
Figure 9.10 (A) Formation of a pull apart basin at an splays across zones 5 to 10 kilometres wide
extensional bend on a right lateral strike-slip fault. (B) (Figure 9.16; Wheeler, 1989). During some of
Formation of a contractional duplex at a restraining these earthquakes surface fault ruptures occurred
bend on a right lateral strike-slip fault.
on range-bounding faults on opposite sides of a
basin and had opposite directions of displace-
displacement of adjacent crustal blocks on opposite ment (e.g. east-side-down vs. west-side-down).
sides of the fault affects erosion of mountain fronts Earthquakes along normal faults form a rela-
and drainage basin evolution. Landforms associated tively wide zone of permanent ground defor-
with normal faults may include offset geomorphic mation (Figures 9.16 and 9.17). Primary surface
surfaces such as alluvial fans, fluvial terraces or fault rupture on normal faults typically forms
GFE-09.qxd 6/27/05 4:49 PM Page 227
Figure 9.11 (A) Map showing extensional step-over and related area of tectonic subsidence and coastal submer-
gence. (B) Tectonic subsidence in extensional step-over during the 1999 earthquake, Ismit Bay, Turkey. Note sub-
merged buildings and apartment complexes in foreground. Photograph by William Lettis.
zones of ground cracking, bulging and tearing of surface. Reverse faults form a number of charac-
near-surface materials that are up to 100-metres teristic landforms due to the vertical displacement
wide (Figure 9.17). Individual fault splays may be of the hanging wall above the footwall, including
up to 10-km apart and secondary deformation fault scarps, folds, and terrace or alluvial fan
may occur within approximately one kilometre of sequences (Figure 9.18).
these fault splays. Complexities along the traces of reverse faults
form secondary features that vary from 10’s of
Reverse faults metres to kilometre scale. Secondary features
Reverse faults are dip-slip faults in which the such as lateral ramps or tear faults can produce a
hanging wall block has moved up relative to the strike-slip sense of deformation within the
footwall block (Figure 9.7). In general, thrust reverse fault system where the strike of the fault
faults place older rocks or sediments over younger approximates the slip direction, or where a
rocks or sediments, resulting in repetition of reverse fault terminates or steps between fault
stratigraphy in cross-section above the fault. splays. The 1999 Mw 7.6 Chi-Chi Taiwan reverse
Reverse faults are moderately to shallowly dipping earthquake produced four primary types of
structures and, as such, fault traces typically form ground deformation, including primary fault
a sinuous, discontinuous trace across the ground rupture, secondary ground deformation, rotated
GFE-09.qxd 6/27/05 4:49 PM Page 228
Figure 9.12 The ‘Big Bend’ in the San Andreas fault in southern California forms a zone of contractional defor-
mation that has caused uplift of the Transverse Ranges. The Transverse Ranges (shaded) are generally bounded by
reverse faults and prominent scarps. Modified from Anderson (1971).
fold limb, and hingeline deformation (EERI, cracking, bulging and tearing of near surface mater-
2001; Kelson et al., 2001; Figures 9.19 to 9.23). ials that are up to 1 kilometre wide (Figures 9.20,
Up to 13 metres of permanent ground deforma- 9.21 and 9.22). Secondary deformation such as
tion occurred along the fault rupture. Ground back-thrusts resulting from surface rupture on
deformation resulted from a combination of pri- reverse faults may occur several kilometres from
mary tectonic displacement and secondary the main fault trace.
‘plowing’ of the ground surface by the fault tip,
as well as internal deformation of the hanging- Subduction-related submergence and
wall block. The amount of total ground deforma- emergence
tion can be 2 to 3 times greater than the tectonic Subduction zones produce the world’s largest
displacement due to the plowing effects of the earthquakes, including the Mw 9.5 1960 Chile
fault tip (Figure 9.23). earthquake, the Mw 9.3 2004 Sumatra earthquake
Earthquakes along reverse faults form a rela- and the Mw 9.2 1964 Alaska earthquake (Plafker,
tively broad zone of permanent ground deforma- 1969; Ruff and Kanamori, 1983; Thatcher, 1990).
tion. Primary surface fault ruptures on reverse However, these earthquakes do not produce pri-
faults can form zones of secondary folding, ground mary surface fault rupture because the slip surface
GFE-09.qxd 6/27/05 4:49 PM Page 229
Figure 9.13 Surface fault rupture produced during the 1999 Turkey earthquake along North Anatolian
fault. Lateral displacement in this area was 4.7 metres. The width of ground deformation was typically
less than 10 metres. This form of ground deformation is referred to as a mole track. Photographs by James
Hengesh.
GFE-09.qxd 6/27/05 4:49 PM Page 230
Figure 9.14 (A) Schematic fault geometry in the East African Rift System. (B) Schematic of cross section of normal
faults in an extensional regime. Modified from Burbank and Anderson (2001).
on the plate interface occurs at depths of 10- to and d depict tectonic subsidence and submergence
50-km. The process of strain accumulation and during the 1964 Alaska Earthquake, and the previ-
release along subduction zones does, however, ous earthquakes along the 1964 rupture zone.
produce a combination of recognizable coastal Subduction zone earthquakes can produce coseis-
uplift and subsidence both coseismically and inter- mic land level changes of up to 10 metres, and
seismically (Plafker, 1969; Figures 9.24 and 9.25). interseismic land level changes that may be locally
Figures 9.25a and b show examples of reef and greater than the coseismic land level changes, and
beach ridges that emerged coseismically during in the opposite direction (Plafker and Rubin,
the 1855 Wellington, New Zealand and 1964 1992). Where uplift or subsidence occurs along
Alaska earthquakes, respectively. Figures 9.25c the intertidal zone, depositional environments can
GFE-09.qxd 6/27/05 4:49 PM Page 231
Figure 9.15 Photographs of normal faults along the Sierra Nevada range front fault zone, near Bishop California.
Photograph A shows fault scarp along offset debris flow fan. Photograph B shows offset crests (arrows) of approxi-
mate 20 000 year old Tioga glacial moraine (inner) and 114 000 year old Tahoe glacial moraine (outer). Photographs
by James Hengesh.
GFE-09.qxd 6/27/05 4:49 PM Page 232
9.4 Seismic source

characterisation
Three primary geological questions need to be

considered in earthquake engineering studies:
where are earthquakes likely to occur; how big
an event needs to be addressed; how often will
these events occur? Each of these questions need
to be addressed both for probabilistic and deter-
ministic seismic hazard analyses (PSHA and
DSHA, respectively) and for evaluation of sur-
face fault rupture, liquefaction, and seismically
induced slope failure. The questions typically are
considered by conducting a series of investigations
that often include one or more of the following
steps:
1. data compilation and review

2. analysis of regional tectonic setting
3. analysis of geological, seismological, geo-
physical and geodetic data
4. aerial and field reconnaissance
5. analysis of local geological setting
6. geological and geomorphic mapping of
Figure 9.16 Sketch maps of normal fault surface rup- specific faults
tures in the Basin and Range Province, western United 7. subsurface investigations such as paleoseis-
States. Maps A through D show historic rupture. Map E mic trenching, boring programs
shows the late Quaternary traces of the Wasatch fault 8. laboratory and dating studies
zone. Modified from Wheeler, 1989.
9. development of parameters for specific
project requirements
change from subarial to intertidal conditions. As 10. treatment of uncertainty.
shown on Figure 9.25 (photograph D), these
environmental changes can be used to identify Location––Where are earthquakes
and date paleoseismic events along subduction likely to occur?
zones. Identifying the location and geometry of seismo-
Earthquakes along subduction zones produce genic faults (or seismic source zones) is accom-
uplift and subsidence over large areas. The 1964 plished through interpretation of regional
Alaska earthquake caused land-level changes in an geology, seismic activity, geodetic data, and
area over 140 000 km2 (Figure 9.24; Plafker, Quaternary geology and geomorphology. The
1969). Secondary faults above the plate interface regional geologic, seismologic and geodetic data
zone also can produce large surface displace- provide general information on the location, style
ments. For example, eight metres of secondary and rate of crustal deformation. However, identi-
fault rupture occurred along the Patton Bay fault fying and characterising specific faults and seis-
during the 1964 Alaska earthquake (Plafker, mic source zones commonly requires detailed
1969), and 12.5-metres of secondary displacement assessment of the Quaternary geology, geomor-
occurred along the Wairarapa fault during the phology and seismology of a study region (Yeats
1855 earthquake near Wellington, New Zealand et al., 1997). In order to identify the locations of
(Grapes and Wellman, 1988). active faults, it is necessary to have deposits or
GFE-09.qxd 6/27/05 4:49 PM Page 233
Figure 9.17 Graben formed in the fault rupture of the 1983 Mw 7.3 Borah Peak, Idaho earthquake. The zone of
ground deformation is approximately 30 metres wide at the intersection of the road. Photograph from Gori and
Hayes, 1992.
landforms of Pleistocene or Holocene age strain gauge, which can be dated to document the
(Hanson et al., 1999). location, rate of activity, and whether or not the
Although different techniques are used to char- most recent movement of a fault is older or
acterise strike-slip, normal, and reverse faults, the younger than the activity threshold applicable to
overriding principles are the same. Investigations the project (e.g. 10 000 or 50 000 years before
are conducted to identify where repeated moder- present).
ate to large magnitude earthquakes have produced A stepwise process typically is followed to
permanent ground deformation, affected the dis- identify and map faults. Following literature
tribution of surficial deposits, and modified review and analysis of regional geological, seis-
the landscape in Quaternary time. The primary mological, and geodetic data, these steps include:
objective of geologic and geomorphologic field
investigations is to identify landforms such as 1. Aerial photograph interpretation. Figure 9.26
fault scarps, drainage systems, glacial or volcanic provides an example of an interpreted aerial
deposits, alluvial fans, and fluvial or marine photograph and illustrates the types of fea-
terraces, which indicate that normal landscape tures that can be recognised. The photograph
evolution processes have been purturbated by tec- is from the eastern side of the Sierra Nevada,
tonic activity. These landforms provide a spatial near Bishop, California. This area was blan-
datum (e.g. a surface or lineation) that serves as a keted by the Bishop Tuff, an ignimbrite flow,
GFE-09.qxd 6/27/05 4:49 PM Page 234
Figure 9.18 Uplifted and dissected alluvial fans along the eastern Andes range-front, Mendoza, Argentina. Note
the offset alluvial fan complexes and fault scarps (arrows) on A; and, deep stream incision of alluvial fan complexes
shown on B. Stream incision indicates a change in stream base-level due to uplift of the hanging wall (typically the
mountain front). Photographs by James Hengesh.
approximately 700 000 years ago. The tuff related to climatic changes during late
forms a plateau surface which can be used as Quaternary glacial and interglacial periods. In
a strain gauge. The tuff has been incised by the addition to these geomorphic surfaces, scarps,
Owens River and a smaller stream. A series of offset drainages, and tonal lineaments are
terraces (Qt1⫺3) are inset into the tuff surface. mapped.
On the lower left of the photo, an alluvial fan Based on published information on the
(Qa1) has eroded and removed the terraces. geology and seismicity, field reconnaissance
The terraces and alluvial fan surfaces are and interpretation of the aerial photographs,
GFE-09.qxd 6/27/05 4:49 PM Page 235
Figure 9.19 Photograph and schematic cross section of the primary fault rupture produced during the 1999 Mw
7.6 Chi-Chi, Taiwan earthquake (Kelson et al., 2001). Collapse of the overriding tip of the hanging wall block forms
a wedge of material above the ground surface. Note the difference in the angle of the fault scarp where track mater-
ial prevented collapse and in the field where collapse of the scarp occurred.
multiple lines of evidence suggest that the and formed the Qt1–3 terraces. The 1986
Tungsten Hills fault is a capable tectonic fea- Chelfant Valley earthquakes (M5.7, 6.3, 5.4)
ture. The Tungsten Hills fault offsets and forms occurred along the eastern projection of
a scarp on the 700 000 year old Bishop Tuff the Tungsten Hills fault. Therefore, this fault is
surface, offsets the Owens River Gorge, and recently active, offsets mid-Pleistocene deposits,
forms lineaments across the alluvial fan sur- offsets the mid- to late Pleistocene Owens River
face. Uplift on the southeastern side of the Gorge, and forms a series of late Quaternary
fault may have changed the stream base-level terraces. Beause of the fault’s geological and
GFE-09.qxd 6/27/05 4:49 PM Page 236
potential earthquake magnitude and recur-

rence interval.
2. Surface Geological and Geomorphological
Mapping. Geological and geomorphological
mapping provides detailed field verification of
the distribution of Quaternary features such as
marine or fluvial terraces, volcanic or lacaus-
trine deposits, soil types, landslides, and ero-
sional and/or depositional surfaces, such as
stream terraces and alluvial fans. The example
of an interpreted aerial photograph, illustrated
on Figure 9.26, shows the variety of features
that can be mapped in a small area. Maps show-
ing the distribution of Quaternary deposits are
particularly useful for defining areas undergo-
ing active deformation or alluviation. The scale
and level of detail shown on geologic and geo-
morphic maps is dependent on the level of docu-
mentation and geotechnical requirements of a
project.
Quaternary deposits and geomorphic sur-
faces are mapped on the basis of several strati-
graphic, geomorphic, and pedologic criteria,
including: (1) topographic position in a
sequence of inset deposits or surfaces; (2) rela-
tive degree of surface modification (e.g. ero-
sional dissection, mima-mound relief, etc.);
(3) relative degree of soil-profile development
or other surface-weathering phenomena;
(4) superposition of deposits separated by ero-
sional unconformities and/or buried soils;
Figure 9.20 Photograph and schematic cross section of (5) relative or numerical age of individual
secondary deformation produced during the 1999 Mw 7.6 deposits; (6) physical continuity and lateral
Chi-Chi, Taiwan earthquake (Kelson et al., 2001). Three correlation with other stratigraphic units;
types of secondary deformation were observed: tensional (7) distinctive lithology or mineralogy; and
faults in the fault propagation fold; back-thrusts; and
(8) textural or lithologic uniqueness such as
complex deformation involving collapsed fault scarps,
back thrusts, and penetrative fracturing or cracking. inclusion of distinctive volcanic ash, lacustrine
sediments, or exotic clast lithologies.
3. Soil profile development. The analysis of
geomorphic expression, long-term history of soil-profile development is an effective means
movement, and location it was considered in to assess the relative age of Quaternary
developing seismic design criteria for the deposits and features and is an integral part of
Pleasant Valley dam. Aerial photograph inter- Quaternary geologic mapping (Sowers et al.,
pretation of the area provided a basis to locate 1998; Birkeland, 1984). For example, the use
the Tungsten Hills fault and to identify geo- of soil-profile development can be a time- and
morphic features that could be used to estimate cost-effective approach to evaluate relative
the Quaternary fault slip rate, fault length, and ages of related surfaces, or to correlate a
GFE-09.qxd 6/27/05 4:50 PM Page 237
Figure 9.21 Photograph and schematic cross section of a rotated fold limb, produced during the 1999 Mw 7.6
Chi-Chi, Taiwan earthquake (modified from Kelson et al., 2001). The block between the primary fault trace and
the secondary fault trace is rotated in a counterclockwise direction.
surface of unknown age to one of known age relative chronology for landforms, sometimes
(Sowers et al., 1998). referred to as a soil chronosequence. The esti-
Soil profile development typically is analysed mated ages of landforms across a fault zone
in shallow soil pits on Quaternary surfaces, such can then be used to locate and evaluate the
as glacial moraines, alluvial fans, stream ter- activity of a fault, and to estimate a fault slip
races, floodplains, and landslide deposits. Soil rate. These data are used in characterising
characteristics that typically are assessed faults for seismic hazard assessment.
include: depth of profile development; soil tex- Buried soils represent former periods of
ture; clay and silt accumulation and orientation landscape stability/instability, and can be
(flows); salt accumulation (e.g. carbonate); soil indicators of past climatic, tectonic and
structure; soil color; and degree of horizona- depositional conditions. Buried soils can be
tion (Sowers et al., 1998; Birkeland, 1984). recognised in natural exposures, exploratory
Comparison of the relative degree of soil-profile excavations, and boreholes. Buried soils in
development between like surfaces with similar outcrops typically have many diagnostic features
parent material leads to the development of a similar to those of surface soils (e.g. soil
GFE-09.qxd 6/27/05 4:50 PM Page 238
4. Geomorphic analysis. Geomorphic tech-

niques are an integral part of Quaternary
geologic mapping and provide useful infor-
mation on the timing, rates, and locations of
ground deformation. Geomorphic analyses
commonly involve inferring the amount of
fault displacement during individual earth-
quakes, and over specified lengths of time,
from measurements of landform deforma-
tion. Geomorphic features such as stream
channels (thalwegs), stream terraces, marine
terraces, and glacial moraines commonly
form well preserved datum’s from which to
assess fault slip rate, recency of activity, and
the direction and amount of displacement
during an earthquake.
Fluvial terraces, marine terraces, or glacial
landforms are examples of geomorphic fea-
tures that can be used as strain gauges to evalu-
ate location, timing, amount and rates of
deformation where these types of features
overlie a fault (Figure 9.27). For example,
streams form at a relatively constant gradient;
however, where crossed by faults, the position
and gradient of stream thalwegs and terraces
may be affected by fault displacement, tilting
or folding. Terraces on the upthown side of a
fault may be higher in elevation than terraces
of equivalent age on the downthrown side of
the fault (Figure 9.27). Increasing scarp
height on progressively older terrace surfaces
indicates recurrent fault moment as shown
by terraces Qt1 and Qt2 on Figure 9.27b.
Measurements of scarp height across these
surfaces and estimates of surface ages based
on soil development or other properties may
Figure 9.22 Photographs and schematic cross sec- yield data on fault slip rate. The degree of post-
tion of ‘hinge-line’ fault produced during the 1999 Mw earthquake scarp modification also can be an
7.6 Chi-Chi, Taiwan earthquake (modified from indicator of the age of an event (Hanks, 1998).
Kelson et al., 2001). A zone of cracking and ground
deformation occurs above the fault-bend, at depth. Terraces that cross a fault and are not
deformed provide a useful datum to confirm
the inactivity of a fault.
horizonation, soil structure, color). In a rapidly
aggrading floodplain or alluvial-basin deposi- Magnitude––How large an event
tional environment, buried organic- rich soils needs to be considered?
commonly are preserved and useful for estab- Once an active fault is located, the maximum earth-
lishing the stratigraphic correlation and deposi- quake that the fault is capable of producing in the
tional history of alluvial sequences. current tectonic environment needs to be estimated.
GFE-09.qxd 6/27/05 4:50 PM Page 239
Figure 9.23 Photograph and topographic profile of scarp produced during the 1999 Mw 7.6 Chi-Chi, Taiwan
earthquake. Note that the total permanent ground deformation is approximately 8 metres, but only 2 metres of the
displacement is due to tectonic offset. The scarp has been amplified due to the ‘plowing’ effect of the fault tip. The
photograph is of an originally horizontal rice paddy. From Kelson et al., 2001.
For engineering purposes the maximum earthquake requires evaluation of surficial geology and
magnitude (Mmax) is usually expressed in terms of geomorphology to define the fault geometry and
its moment magnitude (Mw). segmentation points (Yeats et al., 1997; McCalpin
The Mmax is commonly estimated by identify- 1996; Schwartz and Sibson, 1989). Investigations
ing the length, area or displacement of a particu- following the 1999 Turkey earthquake provided
lar fault segment, and deriving the characteristic valuable new information regarding the role of
magnitude from empirical relationships (e.g. fault geometry (e.g. fault stepovers) in the termi-
Figure 9.28; Schwartz and Coppersmith, 1984; nation of fault rupture (Lettis et al., 2000).
Wells and Coppersmith, 1994, Hanks and Bakun, Detailed mapping of the North Anatolian fault
2002). Defining the length of a fault segment to showed a correlation between width of a fault
be used in estimating earthquake magnitude stepover and the ability of the stepover to arrest
GFE-09.qxd 6/27/05 4:50 PM Page 240
Figure 9.24 Area of surface deformation associated with the 1964 Mw 9.2 Great Alaska earthquake. The earth-
quake caused up to 30 feet of uplift and 8 feet of subsidence. Modified from Plafker, 1969.
GFE-09.qxd 6/27/05 4:50 PM Page 241
Figure 9.25 Examples of emergent and submergent land forms produced during subduction zone earthquakes:
(A) Raised wave-cut platform in Wellington Harbor formed during the 1855 Wellington, New Zealand earthquake.
(B) Storm berm and driftwood lines raised approximately 14 feet during Mw 9.2 Great Alaska Earthquake (Kayak
Island). Upper driftwood line represents the pre-1964 earthquake shoreline. (C) Spruce forest killed due to tectonic
subsidence and salt-water innundation during 1964 Alaska earthquake. (D) Spruce stump and subareal peat layer
buried by intertidal mud, Hinchinbrook Island, Alaska. Tree killed due to tectonic subsidence and submergence, as
in Photo C, approximately 1336 to 1390 years ago. Photographs by James Hengesh.
propagation of the fault rupture (Lettis et al., underscore the importance of conducting careful
2002). For example, the 5-km wide Karamursel geological mapping and analysis to identify the
stepover was sufficient to arrest 4 to 5.5 metres of geometry of fault segmentation points when esti-
fault displacement near the Hersek peninsula, mating maximum earthquake magnitudes.
while 3- to 5-metres of fault displacement rup- Maximum earthquake magnitude can be diffi-
tured through the 2-km wide Gulcuk and Sapanca cult to estimate in areas characterised by low
stepovers (Figure 9.29). Data from the 1999 seismic activity or where faults are obscured
Turkey earthquake combined with data from other by youthful surficial geological deposits. For
strike slip earthquakes were used to develop a example, a sequence of three M ⬃ 7.3 to 8.1
preliminary empirical relationship between sur- earthquakes occurred along the New Madrid
face rupture displacement and fault stepover Seismic Zone in the Mississippi Embayment
width. The graph on Figure 9.29 provides an ini- province of the central United States in
tial constraint on the size of a fault stepover that 1811–1812. However, no surface expression of
can be considered a segmentation point in esti- the faults that caused two of these events has
mating fault rupture length and earthquake mag- been recognised; therefore, empirical relations
nitude for strike slip faults (Lettis et al., 2002). between fault length, area and magnitude cannot
Observations from the 1999 Turkey earthquake be readily used to estimate the future Mmax. This
GFE-09.qxd 6/27/05 4:50 PM Page 242
Figure 9.26 Interpreted aerial photograph of Tungsten Hills fault, near Bishop, California (US Bureau of Land
Management (date 10.02.77), series CA01-77, no. 4-28-7, scale 1:24 000). The interpreted photograph identifies mid-
Quaternary volcanic deposits, late Quaternary fluvial terrace and alluvial fan surfaces, as well as scarps, offset drainages,
and tonal lineaments. Lower photograph is an oblique southeastward view of same area in upper photograph. Note ⬃300
metre offset of post 700 ka Owens River Gorge across Tungsten Hills fault. Lower photograph by James Hengesh.
historical earthquake sequence represents the 1811–1812 earthquakes, can be used to estimate
expected future maximum earthquake magnitude the maximum magnitude of future earthquakes in
for the New Madrid Seismic Zone. Therefore, this region. The Mmax of the New Madrid Seismic
historical seismicity, e.g. the magnitude of the Zone has been estimated by Hough et al. (2000)
GFE-09.qxd 6/27/05 4:50 PM Page 243
Figure 9.27 (A) Idealised block diagrams illustrating the use of stream terraces to evaluate recurrent fault activity.
Sequence A shows the development of two river terraces that are subsequently faulted. Sequence B shows the devel-
opment of two terraces separated by a faulting event and followed by a second faulting event. The fault scarp is pro-
gressively larger across the older terrace surface (modified from McCalpin, 1996). (B) Longitudinal profile of stream
terraces (Qt1–Qt4) along Stevens Creek across Monte Vista and Cascade faults, Santa Cruz Mountains, California
(Hitchcock and Kelson, 1999).
and Bakun and Hopper (2004) using historical damage reports to estimate the moment magni-
intensity data from the 1811–1812 earthquake tude of the earthquakes.
sequence. These approaches use Modified In areas such as the New Madrid Seismic Zone
Mercalli Intensity (MMI) data from historical where fault exposures are not available, several
GFE-09.qxd 6/27/05 4:50 PM Page 244
sand blow deposits, and the geotechnical proper-

ties of subsurface soils can all be used to estimate
earthquake magnitudes. Table 9.1 summarises the
data requirements and outputs for a number of
these techniques.
There are considerable uncertainties in the
use of paleoliquefaction features to derive earth-
quake magnitude. Some of these uncertainties
may include variation in groundwater elevations,
site response effects, and crustal attenuation
properties (Tuttle, 2001). However, in areas such
as the Mississippi Embayment where large mag-
nitude earthquakes are known to have occurred,
but little data are available regarding fault
dimensions, use of liquefaction fields can pro-
vide valuable constraints on the size of past
earthquakes (Tuttle et al., 2002a; 2002b;
Figure 9.28 Relation between earthquake magnitude Obermeier et al., 2001)
and rupture area (after Wells and Coppersmith, 1994).
Earthquake recurrence — how often will
additional approaches are available to estimate the events occur?
magnitude of an earthquake. If paleoliquefaction Earthquake recurrence interval is a measure of the
features are present, the areal distribution of frequency of occurrence of an earthquake with a
liquefaction features, size and characteristics of given magnitude. Earthquake recurrence intervals
Table 9.1 Summary of methods used to estimate magnitude or ground motion acceleration from paleoliquefaction
features.
Method Input Data Required Output Reference
Magnitude- Maximum distance from Magnitude Ambraseys, 1988; Obermeier et al., 2001;
bound source to liquefaction field Pond, 1996
Cyclic Stress Geotechnical borings Magnitude Hwang et al., 1995; Seed and Idriss
(blow counts), ground 1971; Seed et al., 1983; Whitman,
motion attenuation 1971; Youd et al. 2001; Idriss and
models Boulanger, 2004; Cetin et al., 2004
Energy Geotechnical data, Back calculated PGA Berrill and Davis, 1985
Hypocenter data for assumed magnitude
Ishihara Thickness of cap, Back calculated PGA; Obermeier et al., 2001; Ishihara, 1985;
thickness of source thickness of fractured Youd and Garris, 1995
sand cap if PGA is known
Energy Based Geotechnical borings Energy and stress Dobry et al., 1982; Green, 2001; Law
Approach (blow counts) conditions at site et al., 1990; Obermeier et al., 2001;
Pond, 1996
Comparison Modern earthquake Magnitude Tuttle et al., 2002a
to Modern analogs, size and spatial
Earthquakes distribution, with known
magnitudes
Arias Intensity Ground motion, field Magnitude, source Tuttle et al., 2001; Arias, 1970; Kayen
penetration tests distance and Mitchell, 1997, Obermeier et al., 2001
GFE-09.qxd 6/27/05 4:50 PM Page 245
Figure 9.29 (A) Comparison of pull-apart basins along the 17 August 1999 Izmit earthquake rupture. (i) Karamursel
step-over; (ii) Golcuk step-over; (iii) Lake Sapanca step-over; (iv) Eften Lake step-over. Modified from Lettis et al.,
2002. (B) Relationship between average surface displacement or maximum displacement entering the step-over
(which is largest) and average width of the step-over. Modified from Lettis et al., 2002.
usually are estimated using one of three general 3. Seismological Approach – statistical analysis
approaches: of an earthquake catalogue to develop an earth-
quake magnitude-frequency distribution.
1. Paleoseismic Approach – direct observation
and dating of past earthquakes from offset geo- Paleoseismic Approach The paleoseismic
logical units, paleoliquefaction features, or approach offers the benefit of directly describing
coseismically produced landforms. and dating evidence of past earthquakes. This
2. Slip Rate Approach – quotient of fault dis- typically is accomplished through subsurface
placement and fault slip rate. investigations, which provide the most definitive
GFE-09.qxd 6/27/05 4:50 PM Page 246
information on fault location, fault behaviour, fea-

tures produced during paleoliquefaction, and the
timing of past earthquakes. Subsurface investiga-
tions include exploratory trenching, large- and
small-diameter boreholes, and geophysical profil-
ing. Exploratory trenching is the most commonly
used, and generally most successful, method for
assessing paleoseismic activity. Sites for trenching
should be carefully chosen following initial geo-
logic observation and mapping. Preferred sites
include those with datable late Quaternary deposits
across the fault trace or paleoliquefaction feature,
minimal erosion, and continuous deposition to
provide a complete record of earthquake activity.
Periods of erosion and/or non-deposition may
destroy of fail to preserve evidence of earthquakes.
Fault rupture associated with past earthquakes
produces a range of features that can be recog-
nised in trenches to evaluate the presence and tim-
ing of individual paleoseismic events. As shown
on Figure 9.30, these include:
1. deposits offset by a fault

2. abrupt upward termination of a fault strand
3. deposits and surfaces deformed by folding,
tilting or warping
4. lithologic variations across a fault
5. variation in thickness of stratigraphic units or
soils across a fault
6. deposits directly related to faulting, including
scarp-derived colluvium, impounded or
ponded alluvium (e.g. sag pond deposits), or Figure 9.30 Schematic diagram illustrating the princi-
fissure-fill deposits pal stratigraphic and structural criteria used to identify the
occurrence and timing of past earthquakes. Dated sam-
7. transformed deposits including liquefied sed- ples at locations A predate the earthquake. Dated samples
iments and sediment sheared by faulting at locations B post-date the earthquake. Dated samples at
8. systematic and abrupt or stepped increases in C are not helpful for deciphering chronology of past
displacement downsection earthquakes because the sample may pre-date or post-
9. intruded material such as fissure fills, fault date the earthquake depending on geologic context.
Dated samples at locations A and B, therefore, constrain
gouge, or liquefied clastic deposits the timing of the earthquake. Where evidence of multiple
10. open fissures or pockets along a fault plane; earthquakes is present, locations A1 pre-date the earlier
and fault planes exposed at the surface (i.e. event and locations A2 post-date the earlier event and pre-
along scarp). date the later event. Modified from Lettis and Kelson,
1998.
Figures 9.31a through 9.31d illustrate the develop-
ment of multiple colluvial wedges by recurrent across the fault, producing a colluvial wedge
surface fault rupture. Each fault rupture surface (Figure 9.31b). Given sufficient time between sur-
produces a scarp with a steep ‘free face’ (Figure face faulting events, weathering and soil develop-
9.31a). Sediment eroded from the free face is shed ment will occur on the upper surface of the
GFE-09.qxd 6/27/05 4:50 PM Page 247
Figure 9.31 Formation of colluvial wedges from recurrent surface faulting on a single normal fault (modified
from Schwarz and Coppersmith 1984).
colluvial wedge (Figure 9.31c). The degree of and describe the full range of features observed,
weathering and soil development provides an indi- and special attention should be given to the tex-
cation of the length of time between earthquakes. tures and orientations of clasts in trench walls and
If a second faulting event occurs (Figure 9.31d) floor. Figure 9.32 provides an example of the type
the first colluvial wedge will be displaced, and the of detail that should be included in a trench log.
renewed scarp will give rise to a second colluvial This detail allows recognition of subtle folding,
wedge. The use of colluvial wedge stratigraphy, warping, or tilting, or structural fabrics such as
combined with the other features illustrated on oriented clasts, that might not be readily observed
Figure 9.30, enables paleoseismologists to deci- if only major stratigraphic units are logged.
pher an ‘event-stratigraphy’ of past earthquakes. An event stratigraphy may be dated by
When logging a trench it is critical to document geochronologic methods (Sowers et al., 1998). The
GFE-09.qxd 6/27/05 4:50 PM Page 248
Figure 9.32 Log of trench from Salt Lake fault, California. Note how the detail of the log shows tectonic fabric
of sand and gravel deposits. The fabric results from displacement along fault plane and rotation, tilting or alignment
of clasts or layers. Exposure shows bedrock, fluvial stratigraphic units, and soil horizonisation.
method used is dependent on the type of feature occurred within the time interval represented by
being dated, the approximate age of the feature the samples. Care should be taken in describing
(different methods are applicable for different time the stratigraphy at a liquefaction site to deter-
periods and sample materials, and the presence of mine the geological and geotechnical properties
suitable material for dating e.g. charcoal, volcanic of the soils, the geological context of the site,
ash, etc.) in a clear relationship to the fault related and to provide descriptions that could be used in
feature. Generally, the identification and dating of interpreting whether one or more events are rep-
two or more of these features is required to docu- resented in the exposure.
ment a paleoseismic event and to determine the Establishing the recurrence of large magnitude
relative timing between past earthquakes. subduction zone earthquakes involves the dating
In areas such as the central and eastern of either submergent or emergent landforms that
United States where earthquakes have occurred were produced by coseismic crustal deformation
without producing surface fault rupture, pale- of coastal areas above the plate interface part of
oliquefaction features can be used to date the the subduction zone complex. Although subduc-
occurrence of past earthquakes (Obermeier et tion zone earthquakes do not produce primary
al., 2001; Tuttle et al., 2002a). Figure 9.33 illus- surface fault rupture, areas affected by these
trates the form of a typical paleo-sand vent and earthquakes experience both coseismic and inter-
the approach to dating the earthquake history seismic land-level changes (Figure 9.34; Plafker
using paleoliquefaction features. Samples and Rubin, 1992). Where land-level changes
should be collected from within the vented sand occur along the coast, rapid environmental
material or below the vented horizon to provide changes (e.g. from subareal to intertidal or visa
a maximum limiting age of the event, and sam- versa) take place (Figures 9.24, 9.25 and 9.34).
ples should be collected from above the vented The deposits representing the time of change of
sand horizon to provide a minimum limiting age distinct depositional environments can be dated
of the event. The age of the event horizon is then and serve as a proxy to bracket the time of the
bracked and the event can be inferred to have earthquake.
GFE-09.qxd 6/27/05 4:50 PM Page 249
Figure 9.33 Schematic vertical section showing where liquefaction related sand dykes and sills form preferen-
tially. Length of sketch represents 10 to 100 metres. Height represents 3 to 5 metres. Such severe sill development
is typically accompanied by venting of sand to the surface. Triangles indicate position of samples that could be col-
lected and dated to bracket the age of the liquefaction event. Modified from Obermeier, et al., 2001.
Figure 9.34 Simplified displacement history and stratigraphy for the younger part of the Copper River delta
sequence. Column shows stratigraphy and ages of the last three coseismic uplift events. Graph shows an idealised
tectonic displacement history for the last two pre-1964 events to illustrate the ‘yo-yo’ effect of coseismic uplift
(heavy arrows) alternating with larger interseismic subsidence (dashed lines). Also shown is the vertical displace-
ment path of the ground surface (dotted line), the sum of the tectonic displacements plus interseismic sediment accu-
mulation. The amount of uplift for the pre-1964 events is arbitrarily assumed to be comparable to the 2 m uplift that
occurred during the 1964 Alaska earthquake, the interseismic subsidence rate shown is about 5.5 mm/yr, and eusta-
tic sea level rise is taken as ⫹1mm/yr. Modified from Plafker and Rubin, 1992.
GFE-09.qxd 6/27/05 4:50 PM Page 250
Figure 9.25d shows an example of an inter- studies (Figures 9.26, 9.27, and 9.31), tectonic
tidal marsh with evidence of sudden changes observations, or geodetically-based crustal velo-
from intertidal to subareal environments. The city surveys (Figure 9.3).
site shown on this photograph has estuarine silt The geomorphic feature identified on the aer-
overlain by a subareal peat, which is overlain by ial photographs in Figure 9.26 provide an example
a second estuarine silt. The lower silt, dated of the types of features that can be used to esti-
1390 to 1720 2␴ calibrated years before present mated fault slip rate. In this example, the Owens
(cal. ybp), was coseismically uplifted out of the River gorge is offset approximately 300 metres
tidal range and colonised as a spruce forest. The across the Tungsten Hills fault. Assuming a maxi-
overlying subareal peat is dated 1060–1336 2␴ mum age of 700 000 years for the Owens River
cal ybp. These dates represent the minimum and gorge, based on the age of the Bishop tuff surface,
maximum limiting ages, which bracket the time the amount of offset and age of the gorge yield
of the earthquake. The forest subsequently was a minimum fault slip rate of approximately
submerged interseismically and was again 0.4 mm/yr.
uplifted during the 1964 Alaska earthquake. The Although the fault slip rate can be estimated
current vegetated surface represents the modern based on geomorphic data, no direct information
analog of the environment that formed the is available regarding the amount of displace-
buried peat. An intervening event dated approx- ment per event. In this case, the displacement
imately 700 ybp was not preserved at this site, can be estimated using empirical relationships
which highlights one source of uncertainty that between fault length and earthquake magnitude,
must be considered in estimating earthquake and earthquake magnitude and displacement
recurrence in marsh deposits (i.e. lack of preser- (Wells and Coppersmith, 1994). Based on the
vation). 16-km length of the Tungsten Hills fault, the
Slip Rate Approach The slip rate approach for maximum earthquake magnitude is estimated to
estimating earthquake recurrence is based on the be Mw 6.5 using the empirical relationship
assumption that the earthquake occurrence between fault length and earthquake magnitude
process for many major faults follows a time- (Wells and Coppersmith, 1994). The displace-
dependent model of strain accumulation and ment for a Mw 6.5 earthquake is estimated to be
release (Kanamori and Brodsky, 2004; Figure 9.5), 0.3 metres using the empirical relationship
and that these faults produce characteristic earth- between displacement and magnitude for all
quake magnitudes with characteristic displace- fault types (Wells and Coppersmith, 1994).
ments. This is referred to as the characteristic Therefore, the estimated recurrence interval for
earthquake model (Schwartz and Coppersmith, a Mw 6.5 earthquake on the Tungsten Hills fault,
1984). Using the slip rate approach to estimate calculated by dividing 0.3 metres (300 mm) by
earthquake recurrence requires knowledge of both 0.4 mm/yr, is 750 years.
amount of coseismic displacement and fault slip The assumptions made using this approach can
rate. Given these parameters the earthquake recur- significantly affect the estimated earthquake mag-
rence interval (Tr) is: nitude, fault slip rate and earthquake recurrence
parameters. Therefore, care must be used in esti-
Tr ⫽ Displacement/Fault Slip rate.
mating the data (e.g. age of Owens River Gorge,
The amount of coseismic displacement can be amount of displacement across the Tungsten Hills
determined directly from an event stratigraphy fault, length of fault) used in calculating these
established from paleoseismic investigations parameters. A range of values can be developed to
(described above), or estimated using an empir- capture the uncertainties in the data. The range of
ical relationship between earthquake magnitude displacement and slip rate estimates would then be
and fault displacement (Wells and Coppersmith, used to compute a range of recurrence estimates
1994). The fault slip rate is determined using geo- for input to a probabilistic assessment of seismic
logical data from paleoseismic or geomorphic hazard for this fault.
GFE-09.qxd 6/27/05 4:50 PM Page 251
Seismological Approach The seismological

approach uses the instrumental and historical (i.e.
pre-instrumental) record of earthquakes to estimate
earthquake recurrence. An earthquake catalogue
for a region is statistically analysed to develop a
magnitude-frequency distribution (Gutenberg and
Richter, 1954). The occurrence of earthquakes
follows the form:
Log10 N(m) ⫽ a ⫺ b(m)
where N(m) is the cumulative number of earth-

quakes greater than magnitude (m), and the a- and
b-values are constants.
The recurrence curve developed using only
seismological data is referred to as the exponen-
tial earthquake recurrence model and is repre-
sented by the upper part of the graph on Figure
9.35. The exponential recurrence model ade-
quately predicts the rate of earthquake occur- Figure 9.35 Diagrammatic characteristic earthquake
rence for large regions, but underestimates the recurrence relationship for an individual fault or fault
segment. Modified from Schwartz and Coppersmith,
occurrence of large events for individual fault
1984.
sources.
The standard of practice for developing an
earthquake recurrence curve for individual faults 9.5 Probabilistic seismic hazard
involves combining the exponential and charac- assessment and treatment of
teristic earthquake recurrence models (Figure uncertainty
9.35). This approach uses the seismicity record to
describe the recurrence of small to moderate The location, magnitude, and rates of activity
events, and the geological record to describe for seismic source zones provide the primary
the recurrence of large ‘characteristic’ events inputs for a probabilistic seismic hazard assess-
(Schwartz and Coppersmith, 1984; Youngs and ment (PSHA). The probabilistic seismic hazard
Coppersmith, 1985). The recurrence of a ‘charac- methodology is now the preferred approach for
teristic earthquake’ determined from the geologi- developing design ground motions for most criti-
cal data is combined with the seismological cal facilities, and is the basis for seismic design
earthquake recurrence curve to develop the char- provisions in international building codes. The
acteristic earthquake recurrence relationship input parameters typically are described in terms
(Figure 9.35). The characteristic event is best of a seismic source model (e.g. Figure 9.36).
determined from large magnitude historical earth- The objective of a PSHA is to estimate the
quakes or paleoseismic data. However, the magni- probability that different levels of ground motion
tude and recurrence interval of the characteristic may occur at a site during some future period of
event can also be estimated using the indirect time (e.g. 1, 50, 100, 500, 1000 years). The prob-
approach, described in the Slip Rate Approach abilistic seismic hazard is computed by (1) develop-
section above. When using this indirect approach, ing a seismic source model for a region that
a range of values should be considered to ade- describes each significant seismic source in terms
quately capture the uncertainty in the estimates of of maximum earthquake magnitude, and earth-
maximum earthquake magnitude and recurrence quake recurrence, (2) identifying the appropriate
interval. ground-motion attenuation equations for the
GFE-09.qxd 6/27/05 4:50 PM Page 252
Figure 9.36 Illustration of major steps used in conducting a probabilistic seismic hazard assessment.
region, and (3) calculating the contribution to the estimated. However, the input parameters to the
hazard from each model source for a given period logic tree must still be developed. Two approaches
of time (Figure 9.36). The results are typically dis- have traditionally been used: (1) expert elicitation
played in terms of a hazard curve, which gives the to develop input to a logic tree (SSHAC, 1995);
annual probability of a ground motion being and (2) analysis by an individual or team to
exceeded (e.g. 10⫺2, 10⫺3, 10⫺4 annual probability develop the input parameters to the logic tree.
of exceedence; Figure 9.36). The hazard calcula- Expert elicitation involves convening a panel of
tion explicitly considers the randomness of ground experts on a range of issues related to the charac-
motions for a given magnitude and source-to-site terisation of seismic sources, seismological char-
distance (aleatory variability). In addition to the acteristics of a region, ground motion attenuation
aleatory variability there are also scientific uncer- characteristics, and calculation of hazard. Each of
tainties due to choice of models (e.g. earth- these experts is asked to develop the inputs to the
quake recurrence models) and limitations of data source models and the results are added to a logic
(epistemic uncertainties). Identifying and quanti- tree for eventual hazard analysis. This process can
fying both aleatory and epistemic uncertainties be a time and cost intensive approach. The second
has become an integral part of conducting PSHA approach is more efficient and more widely
(SSHAC, 1995). employed, where a selected team estimates the
When completing a PSHA for critical facilities range of parameters for input to the logic tree,
such as nuclear reactor sites, LNG terminals, rather than expert opinions being individually
dams, and deep-water platforms, it is necessary to elicited from a group of experts.
consider the uncertainties (e.g. SSHAC, 1995; The results of a PSHA typically are expressed
McGuire, 2004). Logic trees are used to explicitly in terms of rock ground motions accelerations.
treat multiple interpretations and to track the deci- However, many sites are located on soft rock or
sion process used in computing the hazard for a soil and a site-specific site-response analysis is
site (Figure 9.37). Use of the logic tree approach conducted to modify the rock ground motions for
allows multiple parameters and multiple models to the specific subsurface conditions at the site.
be considered in the hazard assessment, and thus Figure 9.38 is a schematic illustration of the effect
the uncertainty in the hazard computation to be of soil response on rock ground motions. The site
GFE-09.qxd 6/27/05 4:50 PM Page 253
Figure 9.37 Logic tree showing parametres used in fault characterisation.
response due to soil conditions can have the effect displacements of up to 10 meters tore apart
of dampening short period ground motions and buildings, sheared piles, severed pipelines, com-
amplifying long-period ground motions compared pressed or collapsed bridges, and caused general
to rock conditions. destruction. Overall, liquefaction induced ground-
failures caused severe damage to tens of square
kilometers of Niigata and its environs (Hamada
9.6 Liquefaction and related et al., 1986).
ground deformation The 1964 Mw 9.2 Great Alaska Earthquake
triggered large liquefaction induced flow failures
The 1964 Niigata, Japan earthquake (M 7.5) that demolished port facilities in Valdez, Seward
caused widespread and spectacular building and Whittier and carried large parts of those towns
and ground damage as a consequence of soil into the sea (Plafker, 1969). Earthquake shaking
liquefaction. Loss of soil bearing strength caused and flow failures generated seiches in surrounding
buildings to settle and tilt, including several 4-story bays, some of which over-ran coastal areas, caus-
apartment blocks in the Kawagishicho complex ing additional damage and many deaths. The
that were otherwise structurally undamaged by earthquake also triggered numerous lateral spread
shaking effects of the earthquake. Widely publi- failures that severely damaged the Alaska highway
cised photographs of the tilted buildings dramati- and railway systems, and damaged 266 bridges,
cally focused world attention on the liquefaction many beyond repair. That destruction prevented
process and associated hazards. During the Niigata use of much of the highway and rail systems for
earthquake, liquefaction-induced lateral-spreading months following the earthquake.
GFE-09.qxd 6/27/05 4:50 PM Page 254
Causes and mechanisms

Perhaps the most widely accepted definitions of
terms related to seismic liquefaction are those
recommended by the American Society of Civil
Engineers (ASCE Committee, 1978):
Liquefaction The act or process of trans-
forming cohesionless soils from a solid state to a
liquefied state as a consequence of increased pore
pressure and reduced effective stress.
Several general comments can be made regard-
ing the liquefaction process:
1. Liquefaction usually is associated with, and ini-
tiated by, strong shaking during earthquakes,
which causes certain soils (mainly cohesionless,
uniformly-graded fine sands and coarse silts) to
densify. Rearrangement of soil particles causes
increases in intergranular pore water pressure
with resulting decrease in shear strength. The
liquefaction process always produces a transient
loss of shear resistance, but not always a longer-
term loss of shear strength.
2. Liquefaction is most likely to occur: 1) in satur-
ated, relatively uniform, cohesionless, fine
sands, silty sands, or coarse silts of low relative
density (loose); 2) generally above depths from
15 to 20-metres; and 3) in areas where the water
table is within 5 metres of the ground surface.
Although liquefaction effects are normally
observed only in loose soils, dense sands and
silts may show initial liquefaction (pore pressure
Figure 9.38 Illustration of ground shaking amplifica- buildup) effects, but these are rapidly inhibited
tion process. (A) Ground motions from an earthquake by the dilatancy characteristics of such soils.
are transmitted through rock and enter a soil column.
(B) Properties of the soil column modify the rock Characteristic forms and
motion and can cause amplification or dampening of ground behaviour
ground motion, depending on the frequency of motion. Liquefaction damage during earthquakes usually
(C) After transmission through soil column, the dur-
ation, frequency content, and amplitude of ground is related to lateral spreads or settlement in flat,
shaking is modified from rock input ground motions. low-lying areas (Seed, 1968; Youd and Perkins,
1978; Seed and Idriss, 1982; and Youd 1991). The
The 1999 Turkey and 2001 India earthquakes four main types of ground failures and typical
also caused widespread damage from liquefaction ground damage effects due to liquefaction are
and related ground damage (Figure 9.39). The summarised in Table 9.2.
Turkey earthquake caused catastrophic submer-
Assessment of liquefaction potential
gence of coastal development (Figure 9.2b) and the
The assessment of liquefaction potential involves
India earthquake caused liquefaction of approxi-
three steps:
mately 15 000 km2 and contributed to the failure of
many small and moderate-sized dams (Hengesh 1. Evaluating liquefaction susceptibility. This
and Lettis, 2002; Tuttle et al., 2002b). involves identification of areas or layers that
GFE-09.qxd 6/27/05 4:50 PM Page 255
Figure 9.39 (A) Example of lateral spread failure, from 2001 Mw 7.7 Bhuj India earthquake; (B) Sand boils asso-
ciated with extensional cracking in lateral spread failure, Bhuj earthquake; (C) Example of bearing capacity failure,
1999 Turkey earthquake. Photographs A and B by James Hengesh, C from EERI, 2000.
GFE-09.qxd 6/27/05 4:50 PM Page 256
Table 9.2 Liquefaction failure modes and their characteristic ground damage effects (after Tinsley et al., 1985).
Liquefaction failure mode Typical ground damage and effects
1. Lateral Spreads Small to large lateral displacements of surficial blocks of sediments, on

gentle slopes (⬍ 3⬚). Movements, commonly of several metres to tens of
metres, usually are toward a free face, particularly in incised stream
channels, canals, or open cuts. Lateral spreads are particularly damaging to
pipelines, bridges, and structures with shallow foundations, especially on
flood plains adjacent to river channels.
2. Flow Failures Flow failures, commonly the most catastrophic mode of liquefaction failure,
usually are developed on slopes greater than 3⬚, with movements ranging
from tens of metres to several kilometres, at very rapid velocities. Such
flows involve great volumes of material, and are highly damaging to
structures located on them or in their paths.
3. Ground Oscillation Oscillation occurs when liquefaction is triggered at depth, or within confined
liquefied layers. May produce visible ground oscillation waves, ground
settlements, opening and closing of fissures, and ejections of sand and
water from cracks and fissures (sand ‘boils’). Subsurface structures (pipes,
tanks, etc.) may be damaged from this phenomenon, but damage typically
is less than from lateral spread or flow failures.
4. Loss of Bearing Strength Strength loss caused by liquefaction can cause ground collapse and settlements.
Structures may settle and topple, and buried structures (pipelines, septic tanks,
etc.) may float to the surface. Spreading and collapse of embankment fills may
occur from liquefaction of foundation soils.
have the physical characteristics of liquefiable coarse silts in a loose state, at depths less than
soil. 20 metres, and where the groundwater level is
2. Assessing the opportunity for strong ground within about 5 metres of the ground surface.
shaking. This involves identifying seismic However, liquefaction can occur in other less
sources that are capable of generating moderate susceptible deposits at higher levels of ground
to large magnitude earthquakes and estimating shaking, or from long duration earthquakes.
the occurrence of ground shaking that is strong
enough to generate liquefaction in susceptible Geomorphological characterisation of
materials. liquefaction susceptibility and related
3. Assessing liquefaction potential. This is ground failure
completed by comparing the distribution of Liquefaction does not occur randomly, but is
susceptible deposits and the opportunity for restricted to areas with a narrow range of geologic
strong ground shaking to identify the areas in and hydrologic characteristics that can be identi-
which the level of ground shaking is greater fied and mapped based on established Quaternary
than the liquefaction triggering threshold of mapping techniques. Most liquefaction occurs in
susceptible deposits. Liquefaction opportunity areas of poorly engineered hydraulic fills and in
maps typically are produced for specified fluvial deposits less than 1500 years old. Mapping
probability levels of ground motion. liquefaction susceptibility (e.g. Figures 9.40 and
9.41) involves:
Liquefaction potential is site-dependent
because certain soils are more susceptible to lique- 1. Quaternary geological mapping. This provides
faction than others. Liquefaction can occur at low an important first step in assessing liquefac-
levels of ground shaking (e.g. 0.10 percent gravity tion susceptibility mapping (Hengesh and
[g]) in saturated, relatively uniform fine sands or Bachhuber, 2005; Pyke, 2001). The distribution
GFE-09.qxd 6/27/05 4:50 PM Page 257
Figure 9.40 Decision tree used to evaluate relative liquefaction hazards.
of Quaternary deposits at the surface and in the provide lithologic and engineering data that are
shallow subsurface can be mapped to differen- useful for assessing liquefaction susceptibility.
tiate units based on age, texture, and deposi- The soil characteristics from SPT, Cone
tional environment. Quaternary geological Penetrometer Test (CPT), and laboratory data
mapping for liquefaction susceptibility analy- can be compiled for each Quaternary geo-
ses is based on: (1) interpretation of existing logical map unit in a study area. These data are
geologic and soil maps; (2) compilation and used to develop quantitative estimates of lique-
interpretation of geotechnical borings; (3) inter- faction triggering thresholds or relative sus-
pretation of topographic maps and aerial pho- ceptibility ratings for each map unit (Seed and
tography, in particular pre-development 1:35 000 Harder, 1990; Pyke, 2001).
scale or larger photography; (4) construction of 3. Depth to groundwater maps. Depth to ground-
local Quaternary stratigraphic columns, includ- water is a significant factor governing liquefac-
ing age estimates, and a correlation chart; tion hazard. Saturation reduces the normal
(5) construction of geologic cross sections; and effective stress acting on loose, sandy sediments.
(6) field reconnaissance. This condition, particularly in the upper 20
2. Interpreting geotechnical and geophysical metres of the ground surface, increases the like-
data. Geotechnical boring logs and water well lihood of liquefaction and resulting ground fail-
logs from previous explorations in an area can ure (Youd, 1973). Because groundwater levels
GFE-09.qxd 6/27/05 4:50 PM Page 258
Figure 9.41 Data layers used in creating liquefaction susceptibility map.
may vary due to seasonal variations and historic 1983; 1985; Seed and Harder, 1990; Youd,
groundwater use, the highest reasonable water 1997; Idriss and Boulanger, 2004; Cetin et al..
levels should be considered in the liquefaction 2004). In the absence of detailed subsurface
susceptibility analysis. data, liquefaction susceptibility ratings can be
4. Liquefaction susceptibility maps. Figure 9.40 assigned by using decision tree analysis (Figure
illustrates an approach used in assigning relative 9.40). Keying liquefaction susceptibility rat-
liquefaction susceptibility hazard levels. This ings to estimated threshold triggering ground
flow chart illustrates a process that systemati- motion values allows evaluation of changes in
cally assesses degrees of liquefaction suscepti- liquefaction potential for various earthquake
bility, and was developed in collaboration with scenarios or probabilistic motions that might
the State of California Geological Survey. be used for risk assessment. In addition, the
Quantification of liquefaction susceptibil- liquefaction susceptibility ratings incorporate
ity typically is performed for the uppermost parameters that may change with time, includ-
20 metres of sediment based on the Seed ing depth to groundwater and, therefore, can be
‘simplified procedure’ and subsequent revi- modified in the future to reflect new geotechni-
sions (Seed and Idriss, 1971; 1982; Seed et al., cal borings or other additional data.
GFE-09.qxd 6/27/05 4:50 PM Page 259
9.7 Implications for engineering Berrill, J. B. and Davis, R. O. (1985) Energy dissipation
and seismic liquefaction of sands; revised model.
Soils and Foundations 25(2), 106–118.
Geological and geomorphological investigations
Birkeland, P. W. (1984) Soils and Geomorphology,
are carried out for engineering projects in order to Oxford University Press.
identify and characterise the location, severity Bonilla, M. G. (1970) Surface faulting and related
and frequency of seismic hazards including pri- effects, in, Earthquake Engineering, R. L. Wiegel
mary surface fault rupture, strong ground shak- (ed.) Prentice-Hall, p. 47–74.
Burbank, D. W. and Anderson, R. S. (2001) Tectonic
ing, liquefaction, as well as seismically induced
Geomorphology, Blackwell Science, 274p.
slope failure and tsunami. This information is, in Cetin, K. O., Seed, R. B., Der Kiureghian, A., Tokimatsu,
turn, used to evaluate site suitability, develop K., Harder, L. F., Keyen, R. E. and Moss, R. E. S.
engineering design criteria, assess the probability (2004) Standard penetration test-based probabilistic
of an event exceeding the design criteria for a par- and deterministic assessment of seismic soil lique-
faction potential. Journal of Geotechnical and
ticular facility, and to evaluate the residual risk to
Geoenvironmental Engineering 130(12), 1314–1340.
a facility from the geological conditions at a site Cornell, C. A. (1968) Engineering seismic risk analysis.
or region. Due to the inherent variability in geo- Bulletin of the Seismological Society of America 58,
logical processes, and uncertainties in analytical 1583–1606.
approaches, the treatment of uncertainty has Dobry, R., Ladd, R. S., Yokel, F. Y., Chung, R. M. and
Powell, D. (1982) Prediction of pore water pressure
become a fundamental part of geological hazards
build-up and liquefaction of sands during earth-
investigations. quakes by the cyclic strain method. NBS Building
Science Series 138, U.S. Dept. of Commerce, 152 p.
Earthquake Engineering Research Institute (2000)
References Reconnaissance Report on Kocaeli, Turkey,
Earthquake of August 17, 1999. Youd, T. L., Bardet,
Ambraseys, N. N. (1988) Engineering seismology. J. P. and Bray J. D. (editors) Supplement A to Vol. 16.
Earthquake Engineering and Structural Dynamics, Earthquake Engineering Research Institute (2001)
16 985–1006. Chi-Chi, Taiwan Earthquake of September 21, 1999
Anderson, D. L. (1971) The San Andreas fault. Reconnaissance Report, Uzarski, J. and Arnold, C.
Scientific American 225(5), 52–68. (eds) Supplement A to Vol. 17, Earthquake Spectra.
Arias, A. (1970) A measure of earthquake intensity. In Earthquake Engineering Research Institute (2002) The
Hansen, R. J. (ed) Seismic design for nuclear power 2001 Bhuj, India Earthquake Reconnaissance
plants. MIT Press, Cambridge, Massachusetts. Report, Jam, S. K., Lettis, W. R. Murty C. V. R.,
Armijo, R., Meyer, B., King, G. C. P., Rigo, A. and Bardet, J. P. (eds) Supplement A to Vol. 18,
Papanastassiou, D. (1996) Quaternary evolution of Earthquake Spectra.
the Corinth Rift and its implications for the late Electric Power Research Institute (1994) The earthquakes
Cenozoic evolution of the Aegean. International of stable continental regions; Volume 1: Assessment
Journal of Geophysics 126, 11–53. of large earthquake potential. EPRT Technical
ASCE Committee (1978) Definition of terms related to Report, TR-102261-V1.
liquefaction: Report submitted by the Committee on Electric Power Research institute (2003) Ground motion
Soil Dynamics of the Geotechnical Engineering attenuation.
Division of the American Society of Civil Engineers Electric Power Research Institute (2003) (Rev. 1)
(ASCE). Journal of the Geotechnical Engineering 1008910 EUS ground motion project-model devel-
Division, 104, No GT9. opment and results.
Bakun, W. H. and Hopper, M. G. (2004) Magnitudes Gori, P. L. and Hayes, W. W. (1992) Assessment of
and locations of the 1811–1812 New Madrid, regional earthquake hazards and risk along the
Missouri, and the 1886 Charleston, South Carolina Wasatch Front, Utah. U.S. Geological Survey
earthquakes. Bulletin of the Seismological Society of Professional Paper, 1500-A-J.
America 94(1), 64–75. Grapes, R. and Wellman, H. (1988) The Wairarapa fault.
Bakun, W. H. and McGarr, A. (2002) Differences in Victoria University of Wellington, New Zealand, p.54.
attenuation among the stable continental regions. Green, R. A. (2001) Energy-based evaluation and reme-
Geophysical Research Letters 29(23), 2121. diation of liquefiable soils. Ph.D. Thesis, Civil
Bates, R. L. and Jackson, J. A. (1987) Glossary of Engineering Dept., Virginia Tech, Blacksburg,
Geology (3rd edition), American Geological Virginia, 397p. (http://scholar.lib.vt.edu/theses/
Institute, Alexandria, Virginia. available/etd-08132001-170900/)
GFE-09.qxd 6/27/05 4:50 PM Page 260
Gutenberg, B. and Richter, C. F. (1954) Seismicity of the Ishihara, K. (1985) Stability of natural deposits during
earth. Princeton University Press, 440p. earthquakes. Proceedings of 11th International
Hamada, M., Yasuda, S., Isoyama, R. and Emoto, K. Conference on Soil Mechanics and Foundation
(1986) Study of Liquefaction induced permanent Engineering, San Francisco, California 1, 321–376.
ground displacements. Association for the Johnston, A. C. (1996) Seismic moment assessment of
Development of Earthquake Prediction, Japan, 87pp. earthquakes in stable continental regions — III.
Hanks, T. C. (1998) The age of scarp-like landforms New Madrid 1811-1812, Charleston 1886, and
from diffusion-equation analysis, In Sowers J. M., Lisbon 1755. Geophysical Journal International
Noller, J. S., and Lettis, W. R. (eds) Dating and 126, 314–344.
Earthquakes: Review of Quaternary geochronology Kanamori, H. (1983) Magnitude scale and quantifi-
and its application to Paleoseismo1ogy NUREG/ cation of earthquakes. Tectonophysics 93, 185–199.
CR-5562, U.S. Nuclear Regulatory Commission. Kanamori, H. and Brodsky, E. E. (2004), The physics of
Hanks, T. C. and Bakun, W. H. (2002) A bilinear source- earthquakes. Reports on Progress in Physics,
scaling model for M-log A observations of continen- Institute of Physics Publishing 76, 1429–1496. PII:
tal earthquakes. Bulletin of the Seismological Society S0034-4885(04)25227–7.
of America 92(5). Kayen, R. E. and Mitchell, J. K. (1996) Arias intensity
Hanks, T. C. and Kanamori, H. (1979) A moment magni- assessment of liquefaction test sites on the east side
tude scale. Journal of Geophysical Research 84, of San Francisco Bay affected by the Loma Prieta,
2348–2350. California, earthquake of 17 October 1989. Natural
Hanson, K., Kelson, K. I., Angell, M. A. and Lettis, W. R. Hazards 16(2–3), 243–265.
(1999) Techniques for identifying faults and their Kelson, K. I., Kang, K.-H., Page, W. D., Lee, C.-T. and
origins, NUREG/CR-5503, U.S. Nuclear Regulatory Cluff, L. S. (2001) Representative styles of deforma-
Commission. tion along the Chelungpu fault from the 1999 Chi-Chi
Hengesh, J. V. and Bachhuber, J. L. (2005) Liquefaction (Taiwan) earthquake: Geomorphic characteristics and
susceptibility zonation map of San Juan, Puerto responses of man-made structures. Bulletin of
Rico. In Mann, P. (ed) Active tectonics and seismic Seismological Society America 91(5), 930–952.
hazards of Puerto Rico, Virgin Islands and offshore Kelson, K. I., Simpson, G. D., Van Arsdale, R. B.,
areas. Geological Society of America Special Paper Haraden, C. C. and Lettis, W. R. (1996) Multiple
385, 249–262. Late Holocene Earthquakes along the Reelfoot
Hengesh, J. V. and Lettis, W. R. (2002) Tectonic setting Fault, Central New Madrid Seismic Zone. Journal of
and geological effects of the Mw 7.7 2001 Bhuj Geophysical Research 101(83), 6151–6170.
Earthquake. Earthquake Spectra A(18), 19–28. Kenner, S. J. and Segall, P. (2000) A mechanical model
Hitchcock, C. S. and Kelson, K. I. (1999) Growth of late for intraplate earthquakes: Application to the
Quaternary folds in southwest Santa Clara Valley, New Madrid Seismic Zone. Science 289, 2329–2332.
San Francisco Bay area, California: Implications of Krinitsky, E. L. and Slemmons, D. B. (1990)
‘triggered slip’ for seismic hazard and earthquake Neotectonics in earthquake evaluation. Reviews in
recurrence. Geology 26(5), 391–394. Engineering Geology, Vol. VIII.
Hough, S. E., Armbruster, J. G., Seeber, L. and Law, K. T., Cao, Y. L. and He, G. N. (1990) An energy
Hough, J. F. (2000) On the Modified Mercalli inten- approach for assessing seismic liquefaction potential.
sities and magnitudes of the 18 11–12 New Madrid Canadian Geotechnical Journal 27(3), 320–329.
earthquakes. Journal of Geophysical Research 105, Lettis, W. R. and Kelson, K. I. (1998) Applying
23 839–23 864. geochronology in paleoseismology. In Sowers, J. M.,
Hubert-Ferrari, A., Armijo, R., King, G., Meyer, B. and Noller, J. S. and Lettis, W. R. (eds) Dating and
Barka, A. (2002) Morphology, displacement, and slip Earthquakes: Reviews of Quaternary Geochronology
rates along the North Anatolian Fault, Turkey, 2002. and its application to Paleoseismology. US Nuclear
Journal of Geophysical Research 107(B10), 2235. Regulatory Commission, Washington DC, NUREG/
Hwang, J. H., Chang, C. T. and Chen, C. H. (1995) Study CR 5562: 3–1 to 3–26.
on stress reduction factor rd for liquefaction analysis. Lettis, W. R., Bachhuber, J., Witter, R., Barka, A.,
Proceedings of 1st International Conference on Bray, J., Page, W., Swan, F., Altunel, E., Bardet, J. P.,
Earthquake Geotechnical Engineering, Tokyo, Japan Boulanger, R., Brankman, C., Cakir, Z.,
1, 617–622. Christofferson, S., Cluff, L., Dawson, T., Fumal, T.,
Idriss, I. M. and Boulanger, R. W. (2004) Semi- Guneysu, A. C., Hengesh, J., Kaya, A., Langridge,
empirical procedures for evaluating liquefaction R., Rathje, E., and Stenner, H. (2000) Surface fault
potential during earthquakes (invited paper). 3rd rupture, in Kocaeli, Turkey, Earthquake of August
International Conference on Earthquake Geotechnical 17, 1999, Reconnaissance Report (Supplement A to
Engineering, University of California, Berkeley, vol. 16), in T. L. Youd, J. P. Bardet and J. D. Bray
California, USA. (Editors), Chapter 2, 11–53.
GFE-09.qxd 6/27/05 4:50 PM Page 261
Lettis, W., Bachhuber J., Witter, R., Brankman, C., Schwartz, D. P. and Sibson, R. H. (1989) Fault seg-
Randolph, C. E., Barka, A., Page, W. D. and Kaya, A. mentation and controls of rupture initiation and
(2002) Influence of releasing stcpovers on termination. U.S. Geol. Surv. Open File Report
surface fault rupture and fault segmentation: 89–315.
Examples from the 17 August 1999 Izmit earthquake Seed, H. B. (1968) Landslides during earthquakes
on the north Anatolian fault, Turkey. Bulletin of the due to liquefaction. Journal of Geotechnical
Seismological Society of America 92(1), 19–42. Engineering 94(5), 1055–1122.
McCalpin, J. P. (1996) Paleoseismology. Academic Seed, R. B. and Harder, L. F. (1990) SPT-based analysis
Press, 588p. of cyclic pore pressure generation and undrained
McGuire, R. K. (2004) Seismic hazard and risk analy- residual strength, in Duncan, M. J., editor, H. Bolton
sis. Earthquake Engineering Research Institute, Seed Memorial Symposium Proceedings, May,
Monograph No. 010, 240p. 1990: BiTech Publishers, Vancouver, B. C., Canada,
Molnar, P. (1979) Earthquake recurrence intervals and 351–376.
plate tectonics. Bulletin of the Seismological Society Seed, H. B. and Idriss, I. M. (1971) Simplified proce-
of America 70, 1321–1335. dure for evaluating soil liquefaction potential.
Obermeicr, S. F., Pond, E. C., and Olson, S. M., with Proceeding of the American Society of Civil
contributions by Green, R. A., Stark, T. D., and Engineers, Journal of the Soil Mechanics and
Mitchell, J. K. (2001) Paleoliquefaction studies in Foundations Division, 93, no. SM9, 1249–1273.
continental settings: Geologic and geotechnical Seed, H. B. and Idriss, I. M. (1982) Ground Motions and
factors in interpretations and back-analysis. U.S. Soil Liquefaction During Earthquakes. Earthquake
Geological Survey Open-File Report 0 1–29. Engineering Research Institute Monograph, Berkeley,
Plafker, G. (1969) Tectonics of the March 27, 1964, California, 134pp.
Alaska earthquake. U.S. Geol. Surv. Professional Seed, H. B., Idriss, I. M. and Arango, I. (1983)
Paper 543-I, 74p. Evaluation of liquefaction potential using field
Plafker, G. and Rubin, M. (1992) “Yo-yo” tectonics performance data. Journal of Geotechnical
above the eastern Aleutian subduction zone: coseis- Engineering 109(3), 458–482.
mic uplift alternating with even larger interseismic Seed, H. B., Tokimatsu, K., Harder, L. F. and
submergence. Proceedings of the Wadati Conference Chung, R. M. (1985) Influence of SPT procedures
on great subduction earthquakes, Geophysical in soil liquefaction resistance evaluations. Journal
Institute, Univ. Alaska, Fairbanks. of Geotechnical Engineering 111(12), 1425–
Pond, E. C. (1996) Seismic parameters for the central 1445.
United States based on paleoliquefaction evidence in Senior Seismic Hazard Analysis Committee (SSHAC)
the Wabash Valley. PhD Thesis, Virginia Tech, Civil (1995) Recommendations for Probabilistic Seismic
Engineering Dept., Blacksburg, Virginia, 583p. Hazard Analysis: Guidance on Uncertainty and
Pyke, R. (2001) Discussion of “liquefaction resistance Use of Experts, Lawrence Livermore National
of soils: Summary report from the 1996 NCEER and Laboratory, UCRL-lD-122l60, NUREG/CR-6372.
1998 NCEER/NSF Workshops on evaluation of liq- Sowers J. M., Noller, J. S. and Lettis, W. R. (1998)
uefaction resistence of soils”, Jour. of Geotech. And Dating and Earthquakes: Review of Quaternary
Geoenviron. Eng. Robertson, P. K., and Wride, C. E., geochronology and its application to Paleoseis-
1997, Cyclic liquefaction and its evaluation based on mology; NUREG/CR5562, U.S. Nuclear Regulatory
SPT and CPT: Seismic Short Course on Evaluation Commission.
and Mitigation of Earthquake Induced Liquefaction Seismological Research Letters (1997) Special volume
Hazards, NCEER Workshop, San Francisco. on ground motin attenuation equations, Seismological
Reilinger, R. E., McClusky, S. C., Oral, M. B., King, Society of America, 68(1).
R. W., Toksoz, M. N., Barka, A. A., Kinik, I., Lenk, O. Tinsley, J. C., Youd, T. L., Perkins, D. M. and Chen,
and Sanli, I. (1997) Global Positioning System A. T. F. (1985) Evaluating Liquefaction Potential. In
measurements of the present day crustal movements Ziony, J. I. (ed.) Evaluating earthquake hazards in
in the Arabia-Africa-Eurasia plate collision. Journal the Los Angeles Region -an earth-science perspective.
of Geophysical Research 102, 9983–9999. United States Geological Survey Professional Paper
Ruff L. and Kanamori, H. (1983) The rupture process 1360, 263–316.
and asperity distribution of three great earthquakes Thatcher, W. (1990) Order and diversity in the modes of
from long-period diffracted P-waves. Phys. Earth circum-Pacific earthquake recurrence. Journal of
Planet. Interior 31, 203–230. Geophysical Research 95B3, 2609–2623.
Schwartz, D. P. and Coppersmith, K. J. (1984) Fault Tuttle, M. P (2001) The use of liquefaction features in
behaviour and characteristic earthquakes: examples paleoseismology: Lessons learned in the New
from the Wasatch and San Andreas faults. Journal of Madrid seismic zone, central United States. Journal
Geophysical Research 89, 5681–5698. of Seismology 5, 361–380.
GFE-09.qxd 6/27/05 4:50 PM Page 262
Tuttle, M. P., Schweig, E. S., Sims, J. D., Lafferty, Yeh H., Liu, P. L. -F. and Synolakis, C. E. (1997) Long
R. H., Wolf, L. W. and Haynes, M. L. (2002a) The Wave Runup Models. World Scientific, Singapore,
earthquake potential of the New Madrid seismic 405pp.
zone. Bulletin of the Seismological Society of Youd, T. L. and Garris, C. T. (1995) Liquefaction-
America 92(6), 2080–2089. induced ground-surface disruption. Journal of
Tuttle, M. P., Hengesh, J. V., Tucker, K. B., Lettis, W., Geotechnical Engineering 121(11), 805–809.
Deaton, S. L. and Frost, J. D. (2002b), Observations Youd, I. M., Idriss, L. M., Andrus, R. D., Arango, I.,
and comparisons of liquefaction features and Castro, G., Christian, J. T., Dobry, R., Liam, F.,
related effects induced by the Bhuj earthquake. Harder, W. D., Hynes, L. F., Ishihara, M. E. , Koester,
Earthquake Spectra A(18), 79–100. K., Liao, J. P., Marcuson III, S. S. C. W. F., Martin, G.
Wallace, R. E. (1990) The San Andreas Fault System. US R., Mitchell, J. K., Moriwaki, Y., Power, M. S.,
Geological Survey Professional Paper 1515, 283p. Robertson, P. K., Seed, R. B., and Stokoe III, K. H.
Wells, D. L. and Coppersmith, K. J. (1994) New empir- (2001) Liquefaction resistence of soils: Summary
ical relationships among magnitude, rupture length, report from the 1996 and 1998 NCEERINSF work-
rupture width, rupture area, and surface displace- shops on evaluation of liquefaction resistence of soils.
ment. Bulletin of the Seismological Society of Journal of Geotechnical and Geoenvironmental
America 84(4), 974–1002. Engineering 127(10), 817–833.
Wesson, R. L., Helley, E. J., Lajoie, K. R. and Wentworth, Youd, T. L. (1973) Liquefaction, flow and associated
C. M. (1975) Faults and future earthquakes. In ground failure. US Geological Survey, Circular 688.
Borcherdt, R. D. (ed) Studies for seismic zonation of Youd, T. L. (1991) Mapping of earthquake-induced
the San Francisco Bay region. US Geological Survey liquefaction for seismic zonation. Proceedings, Fourth
Professional Paper, 941-A, A5–30. International Conference on Seismic Zonation,
Weznouski, S. G. (1988) Seismological and structural Earthquake Engineering Research Institute, Stanford,
evolution of strike-slip faults. Nature, 335(6188), 1, 111–147.
340–343. Youd, T. L. and Perkins, D. M. (1978) Mapping lique-
Wheeler, R. L. (1989) Persistent segment boundaries on faction-induced ground failure potential. Journal of
basin and range normal faults. In Schwartz D. P. and Geotechnical Engineering, American Society of
Sibson R. H. (eds) Fault Segmentation and Controls Engineers 104(4), 433–446.
of Rupture Initiation and Termination, U.S. Geol. Youd, T. L. (1997) Updates on the simplified procedure:
Surv. Open File Report 89-3 15, 432–444. An overview of NCEER workshop in Salt Lake City
Whitman, R.V. (1971) Resistance of soil to liquefac- on liquefaction resistance of soils. Seismic Short
tion and settlement. Soils and Foundations 11(4), Course on Evaluation and Mitigation of Earthquake
59–68. Induced Liquefaction Hazards, NCEER Workshop,
Yeats, R. S., Huftile, G. J. and Stiff, L. J. (1994) Late San Francisco, CA, March, 1997.
Cenozoic tectonics of the cast Ventura basin, Youngs, R. R. and Coppersmith, K. J. (1985)
Transverse Ranges, California. AAPG Bull., 78, Implications of fault slip rates and earthquake
1040–1074. recurrence models to probabilistic seismic hazard
Yeats, R. S., Sieh, K. and Allen, C. R. (1997) The geo- estimates, Bulletin of Seismological Society of
logy of earthquakes. Oxford University Press, 568p. America 75(4), 939–964.
GFE-10.qxd 6/24/05 3:44 PM Page 263
10. Rivers
Mark Lee
10.1 Introduction can be seen, in part, as reflecting the need to

maintain continuity of water and sediment trans-
River systems are open channels that carry water port. Among the key features of a basin are the
and sediment. The overall system comprises three drainage network (Figure 10.2) — which often
broad zones, each with a characteristic form and reflects a combination of climatic (Figure 10.3)
function (Figure 10.1): and geological (Figure 10.4) controls — and the
stream ordering (Figure 10.5).
1. Zone 1: the drainage basin or water and sedi-
ment source area (i.e. mountain or hillslope
catchments) 10.2 River channel changes
2. Zone 2: the sediment transport zone (i.e. the
main river channel) Between around 1865 and 1915 many broad valleys
3. Zone 3: the sediment store or deposition zone and plains in the arid and semi-arid south-west
(i.e. the river floodplain, delta or estuary). USA underwent significant changes, developing
trench-like gulleys with rectangular cross sections
This sub-division should only be viewed as illus- and near-vertical side-walls (arroyos). For example,
trating how the dominant processes in different Kaneb Creek in south Utah developed a 25 m deep
parts of the system are interlinked into a water and and 80–120 m wide channel between 1880–1914.
sediment cascade, from source areas to sinks; sedi- At the time the changes were blamed on land use
ments are, in reality, supplied (i.e. by erosion), change, especially over-grazing, and the resultant
transported and deposited in all of the Zones. increase in runoff. However, the importance of cli-
The drainage basin (catchment or watershed) mate change and high magnitude, low frequency
is the fundamental unit for studying river chan- events has been recognised as a factor in triggering
nels and hydrology. River channel adjustments periods of entrenchment (Figure 10.6). In south
Utah, arroyos such as Kaneb Creek developed
during a period of increased flood magnitude
Upstream controls that had been preceded by over 350 years without
(climate, tectonics, ZONE 1 (production) significant floods.
land-use.) Drainage Basin
In 1982 an intense rainstorm over the Howgill
Fells, UK triggered widespread gulley erosion,
delivering large volumes of sediment to the stream
ZONE 2 (transfer)
channels. This in turn prompted localised switches
from meandering to braided channels. Since the
ZONE 3 (deposition) event some channels have changed back to mean-
Downstream
controls dering forms as the sediment supply has dimin-
(baselevel, ished (Harvey, 1986). Similar changes have
tectonics.)
followed mining activities. In the northern Pennines
Figure 10.1 An idealised river system (from Schumm, and mid-Wales, UK the dumping of mine waste
1977). in streams during the eighteenth and nineteenth
263
GFE-10.qxd 6/24/05 3:44 PM Page 264
Figure 10.2 Types of drainage networks (after Mitchell, 1973).
centuries led to the switch from meandering to Some rivers may undergo major course
braided channels with valley floor aggradation. changes. The Huang He (Yellow River) in China,
On some streams, incision and reversal to a mean- for example, has had at least twenty changes during
dering channel has taken place since the end of the last 4000 years or so. Some of these changes
waste dumping (Macklin and Lewin, 1997). have been very substantial. The decline of the
GFE-10.qxd 6/24/05 3:44 PM Page 265
RIVERS 265
civilisation of Mohenjo Daro is believed to have them high and dry. Doornkamp (1982) describes
coincided with course changes of the River Indus. how the Kosi River, which supplies irrigation water
The ‘Old Testament’ city states such as Ur and to the Ganges plains shifted course away from the
Uruk declined when the Euphrates moved and left entrance to the feeder canals in the early 1970s,
leaving the irrigation system short of water.
Changes in both location and depth of chan-
nels can be of particular importance in relation
1
400
yr
10ºC 20ºC
to river crossings. Major river training may be
2
20
Sediment yield, t km
300
necessary to keep a migrating river flowing
200
? under a bridge (e.g. the Jamuna River crossing in
100
Bangladesh), while increase in scour at bridge piers
Drainage density, km km2
15
0
0 500 1000 1500 can lead to bridge collapse. Examples include
Mean annual precipitation, mm
the loss of a section of railway embankment at
10 Dalguise and structural damage to Fortevoit rail
bridge during the 1993 Perth floods in Scotland; in
MAXIMUM October 1987 four people died when channel
5 scour and erosion around the piers of a railway
bridge at Glanrhyd, Dyfed in Wales led to the
MINIMUM bridge collapsing under the weight of a train
0 (Lee, 1995).
0 500 1000 1500 2000 2500
Mean annual precipitation, mm
Significant channel changes can take place as
a result of water impoundment by reservoirs (e.g.
Figure 10.3 Variation of drainage density with climate: Petts, 1984). Generally dams decrease peak flows
mean annual precipitation (after Gregory, 1976). downstream and impound practically all sediment.
Gabbro, North Carolina Granite, British Guiana Granite, South Dakota
Shale, Utah Shale, South Dakota Phyllite, Alabama

One Mile
Scale
Figure 10.4 Drainage densities on different rock types (from Ray and Fischer, 1960).
GFE-10.qxd 6/24/05 3:44 PM Page 266
channels developed gravel bars and became

more braided. River channel engineering has had
notable upstream and downstream effects. Such
channelling, which generally involves attempts to
control the channel form at a site, can initiate
instability not only in the improved channel reach
but also upstream and downstream as the channel
adjusts to a new state of equilibrium.
Gravel extraction increases the channel gradient
through a reach and increases bed roughness, lead-
ing to enhanced bed scour and gravel/cobble trans-
port. If the rate of removal is greater than the rate
of supply from upstream, channel down-cutting
will occur. This incision is self-enhancing because
flow becomes more confined to an entrenched
channel with higher flow velocity and bed shear
Figure 10.5 Approaches to stream ordering (after stresses. Entrenchment may continue long after
Gregory and Walling, 1973).
extraction ceases, as the river channel continues to
adjust to these changes.
Channels tend to reduce in size downstream but These examples serve to illustrate how signifi-
also tend to incise because of the lack of sedi- cant river metamorphosis (i.e. modification of
ment. Most of the adjustment seems to be by the river channel form) can occur over ‘engineer-
cross-section form and slope but even along one ing time’. Changes in discharge or sediment sup-
river there may be both increases and decreases in ply, together with the calibre of bed or bank
channel capacity. Channel degradation and scour materials, vegetation or slope angle, can lead to
will persist until the reduction of channel slope adjustments of channel form (Table 10.1). This is
reduces the flow velocity below the threshold for because of the close links between different parts
sediment transport. The timescales of adjustment of the channel and sediment erosion, transport
to river impoundment also vary but five years and deposition processes (Figures 10.7a and b).
may be required before any channel response is However, as system thresholds need to be
observed. Continuing channel changes have been exceeded, not all changes will lead to sustained
reported more than 50 years after dam construc- river metamorphosis.
tion, and stability in terms of sediment transport
and channel form may take 200 years.
Urbanisation can lead to channel modifica- 10.3 Channel form: regime theory
tions downstream following an increase in the
size of peak flows and a decrease in sediment sup- The ability to respond to changes in the controlling
ply (e.g. Knight, 1979; Hollis and Luckett, 1976). factors depends on the resistance (i.e. erodibility)
Among the most common changes are bed and of the bed and bank materials. Channels that have
bank erosion, loss of riverbank trees and under- been formed in strong bedrock reflect geological
mining of structures. Following the development control (i.e. lithology and structure) and can be
of Cumbernauld New Town in Strathclyde, UK described as confined in that they are not generally
some streams had enlarged through the urban area prone to significant channel change. In contrast,
(Roberts, 1989). Locally there were instances of channels formed in erodible sediments can be des-
extensive erosion and bank collapse; up to 10 m of cribed as unconfined (or ‘alluvial’) channels.
vertical incision through glacial till was observed Under stable climatic conditions, alluvial
in gullies draining industrial areas, and some channel geometry can be considered to be in
GFE-10.qxd 6/24/05 3:44 PM Page 267
RIVERS 267
Random frequency-magnitude variations Secular climatic changes Human land-use changes
‘More Humid’ ‘More Arid’

Incd. Incd. Decd. Decd.
Morpholog- Biological Climatic precip. precip. freq. Ditches Bridges
precip. Logging Roads
ical and or intensit- of Culti- levees and
and or and Grazing and
decd. ies light vation canals embank-
incd. fire trails
temp. rain galleries ments
temp.
Reduced vegetation cover Local removal of

over drainage basin vegetation on
valley floor
Reduced Weakened soil Alteration

infiltration structure over of valley
capacity drainage basin bottom soils
Unusually Increased More flashy Increased sediment

great runoff runoff contribution from
runoff from slopes from slopes drainage basin
Masking valley
Inherent dis- Localized flows Lateral shifts Steepening of
bottoms with
adjustments in and erosion in channel valley-side
erodible
long profiles in drainage net position slopes
sediments
Higher peak discharge through Concentration of flows

valley bottoms than through valley bottom
previously experienced
Local increases in Increased Reduced vegetational

slope of valley hydraulic radius (depth) of retardance
floors flows along valley floors in valley floors
Increased erosiveness (velocity) of flows through valley bottom Increased erodibility of valley floor materials
Arroyo formation
Figure 10.6 A model of arroyo formation (after Cooke and Reeves, 1976).
Table 10.1 Channel changes in response to changes in discharge and sediment load (after Schumm, 1977).
Change River bed morphology Change River bed morphology
Qs Qw Aggradation, channel instability, Qw Qs Incision, channel instability, wider

wider and shallower channel and deeper channel
Qs Qw Incision, channel instability, Qw Qs Aggradation, channel instability,
narrower and deeper channel narrower and shallower channel
Qs Qw Aggradation Qw Qs Processes decreased in intensity
Qs Qw Processes increased in intensity Qw Qs Incision, channel instability, wider
channel
Note: Qs: sediment discharge; Qw: water discharge; : increase; : decrease; : remains constant.
GFE-10.qxd 6/24/05 3:44 PM Page 268
Figure 10.7a A feedback model of channel change

(after Richards and Lane, 1997).
Figure 10.7b A summary of the alluvial channel system, highlighting the interlinkages between variables
(redrawn from Richards, 1982).
equilibrium (i.e. in ‘regime’) with the prevail- d c Q f; v k Qm, as determined by Leopold

ing flows, along with slope gradient, sediment and Maddock (1953) for rivers in mid-west USA
type, bank vegetation and valley constrictions. and elsewhere (Table 10.2; Figure 10.8). These
Channel cross-sectional form adjusts by erosion relationships have formed the basis for predicting
and deposition to accommodate the sediment stable channel forms.
load and all but the highest flows. As discharge These hydraulic geometry equations highlight
(Q) increases so width (w), mean depth (d ) and the tendency for mean velocity to increase down-
velocity (v) all increase. These adjustments can stream. Although the long-profile gradient nor-
be described by power functions: w a Qb; mally decreases, velocity may increase because of
GFE-10.qxd 6/24/05 3:44 PM Page 269
RIVERS 269
Table 10.2 A selection of downstream hydraulic

geometry relations (from Knighton, 1984).
Location Discharge b f m
Midwest USA Qm 0.50 0.40 0.10

Pennsylvania, USA Q50 0.34 0.45 0.32
Q15 0.38 0.42 0.32
Qb 0.42 0.45 0.05
Appalachians, USA Q2.33 0.55 0.36 0.09
Cheshire, UK Q50 0.46 0.16 0.38
Q15 0.54 0.23 0.23
Idaho, USA Qb 0.54 0.34 0.12
Gravel bed rivers, UK Qb 0.45 0.40 0.15
Discharge (Q) defined as mean annual flood (Q2.33);
bankfull (Qb); mean annual (Qm); and the one equalled
or exceeded 50% (Q50); 15% (Q15).
Figure 10.8 Relationships between bankfull discharge and various channel variables (redrawn from Knighton, 1984).
the increased cross-section efficiency (i.e. less ‘bankfull discharge’ i.e. the flow that, over time,
friction) and decreased bed roughness if the bed yields the maximum bed and bank erosion, sedi-
materials become finer downstream. ment transport and deposition. Relatively high
The dominant discharge that controls the chan- frequency flows are thus more significant than
nel geometry is generally assumed to be the rare floods in terms of their cumulative impact on
GFE-10.qxd 6/24/05 3:44 PM Page 270
channel form, hence the assumption that a regime possible that they are relict forms, adjusted to
equilibrium form is adjusted to the bankfull suit previous climatic conditions. Although the
discharge. forms may be stable under current conditions
The frequency of bankfull discharge varies they might never have evolved under them. This
between rivers and can range from 1–32 years. has been described as ‘passive disequilibrium’
However, the importance of bankfull discharge (Ferguson, 1981).
varies between river environments (Wolman and
Miller, 1960). It is the dominant discharge where
the entrainment threshold of the bed sediments is 10.4 Planform types
low (i.e. in sand/mud bed rivers). The more vari-
able the flow, as in semi-arid and arid environ- River channel form reflects the interaction
ments, the greater the effectiveness of rare flash between the discharge and sediment load with the
floods. materials through which the river flows and the
Many changes initiated by large rare floods vegetation along its course. These interactions
may be subsequently removed by the action of are expressed in terms of a characteristic cross-
smaller, more frequent floods. The duration of section, long-profile and planform. The planform
this ‘healing’ process is termed the recovery time, is often of greatest interest in erosion hazard
which varies between river environments. In studies. Despite the great variety of flow regimes,
deserts, the forms may be very persistent and sediment loads, bed and bank materials and vege-
channels will not undergo significant modifica- tation types, there are a limited number of plan-
tion until the next major flood passes through. form types: straight, meandering, braided and
However, in temperate rivers the recovery time anastomosing channels (Figures 10.9a and b).
may be relatively short. Straight alluvial channels are not common.
Stream power is a useful guide to the erosive The channel bed tends to have a sequence of regu-
potential of a river and is defined by: larly spaced riffles (shallows) and pools (deeps),
generated by non-uniform flows that spiral down-
Stream power (W/m2) gQS
stream in three-dimensional circulatory cells.
where is the fluid density; g is the acceleration These cells generate both vertical and lateral vari-
due to gravity; Q is the discharge; S is the channel ations in flow velocity, with deep scour adjacent
slope. to one bank and a shoaling bar by the opposite
The areas most vulnerable to channel change bank. These features alternate from one side of
in a catchment often coincide with the middle the channel to the other downstream and lead to a
reaches, where the river emerges from the upland tendency for meandering (Figure 10.10). Straight
zone. Here, rivers are sufficiently steep to have channels tend to exhibit very slow lateral migra-
the potential for erosion, while entering softer tion and planform evolution.
rocks or alluvial areas where the beds and banks Two broad groups of meandering channel can
are more readily eroded. be recognised. Passive meanders occur where a
The recognition that climate and flood fre- river does not have sufficient power to adjust its
quency has been, and will be, variable over channel through bed scour and bank erosion, with
‘engineering time’ puts severe constraints on the the course tending to be diverted across the flood-
assumption of equilibrium channel forms devel- plain by hard-points or interlocking spurs. Active
oped under static conditions. Indeed, many rivers meanders are associated with alluvial channels
appear to adjust their form more frequently and that undergo almost constant readjustment in
rapidly than had previously been appreciated. It response to variations in discharge and sediment
is probably more realistic to assume that an allu- load. Although most actively meandering rivers
vial channel form is transient or unstable unless comprise a complex pattern of bends of varying
it can be demonstrated otherwise (Macklin and sizes and intervening straights, there appears
Lewin, 1997). Where stable channels exist it is to be a good relationship between the meander
GFE-10.qxd 6/24/05 3:44 PM Page 271
RIVERS 271
Figure 10.9a Classification of channel pattern, based

on sinuosity and degree of division (after Thorne, 1997).
Figure 10.10 Formation of asymmetric pools, bars

and riffle crossings (from Thorne, 1997).
Figure 10.9b Classification of channel pattern, based

on sediment load and system stability (from Thorne,
1997).
wavelength (Figure 10.11) and other channel

parameters, notably:
L 12.34 w – (Richards, 1982)
Figure 10.11 Meander planform: terminology (from
L 54.3 Qb0.5 (Dury, 1955) Thorne, 1997), note that the wavelength is the complete
width of the schematic planform shown (i.e. 2L).
where L is the meander wavelength, w – is the mean
channel width and Qb is the bankfull discharge.
It should be appreciated, however, that outcrops of is a common feature of many alluvial channels,
stronger materials (e.g. clay plugs) within the rather than the maintenance of an equilibrium
floodplain sediments can have a major influence form. Eventually cut-offs occur (Figure 10.12),
on meander form and pattern. creating oxbow lakes that will infill over time.
Meanders may migrate downstream, although Some meandering rivers appear to experience con-
this process can be impeded when less erod- tinuous change and increased instability over time
ible materials are encountered in the riverbanks. (Figure 10.13) i.e. change is inherent and that
Continued growth and elongation of meanders ongoing change with little stability might occur.
GFE-10.qxd 6/24/05 3:44 PM Page 272
Figure 10.13 Channel changes on the River Bollin,

Cheshire, UK (from Hooke, 1991).
Figure 10.12 Examples of meander change (redrawn

from Hooke, 1991).
Braided channels tend to be associated with Anastomosed channels are associated with
high-powered rivers where there is an abundance low-energy rivers. The typical planform is of two
of relatively coarse material moving as bed-load, or more highly sinuous channels separated by
and with weak banks. Typically there is a large, semi-permanent vegetated islands at a simi-
sequence of sub-channels (anabranches) sepa- lar elevation to the surrounding floodplain (i.e. at
rated by braid bars that are flooded at bankfull bankfull stage the flow remains multi-thread).
discharge (i.e. at bankfull stage the flow is in a These tend to be relatively stable channels. Rates
single channel). This type of planform can resem- of bank erosion, bend migration and planform
ble a string of beads with relatively long, wide change are characteristically small.
multi-thread island reaches separated by short, These four major planform types represent a
narrower single-thread reaches (nodes). Braided continuum of forms. Whilst it is tempting to seek
rivers are susceptible to rapid bank erosion and thresholds between these forms and to use these
channel change (avulsion). Figure 10.14 illus- as a measure of the potential instability of a chan-
trates how the position of bars within an island nel, in reality a range of transient patterns occurs.
reach can evolve through time. Thorne (1997) offers some practical advice to
GFE-10.qxd 6/24/05 3:44 PM Page 273
RIVERS 273
Figure 10.14 Evolution of gravel bars in a braided channel (redrawn from Church and Jones, 1982).
help determine whether a river is likely to be help determine past trends. However, great care is
vulnerable to major planform change. If a river is needed in their use because of the potential accu-
susceptible to switching between meandering and racy or reliability problems (e.g. Carr, 1962; Hooke
braiding this should be apparent in the current and Kain, 1982). Table 10.4 provides a guide
and abandoned channel forms. If a channel dis- to some field indicators of channel instability.
plays only typical meandering characteristics, However, caution is needed before simply extrapo-
then it is probably safe to assume it is not prone lating these trends into the future.
to braiding.
Figure 10.15 and Table 10.3 present a useful
system for classifying channel forms. 10.5 Bank erosion and instability
Figure 10.16 presents an indication of the pat-
terns of dynamic morphological adjustment that Bank retreat generally involves a combination of
may occur in alluvial channels. Analysis of histori- flow erosion and mass movement followed by the
cal maps, photographs and satellite imagery can removal of the debris from the bank toe by
GFE-10.qxd 6/24/05 3:44 PM Page 274
Figure 10.15a Rosgen’s classification of rivers (from Rosgen, 1994).
posts or automatic systems (e.g. the Photo-

Electronic Erosion Pin; Lawler, 1992).
A range of typical bank failure types is pre-
sented in Figure 10.17. Bank stability can be
analysed using conventional stability analysis
techniques (e.g. Bromhead, 1986). Alternatively, a
useful tool is the dimensionless stability chart
developed by Osman and Thorne (1988) for rota-
tional and slab-type failures in cohesive materials
(Figure 10.18).
Figure 10.15b Stream types (from Rosgen, 1994).
10.6 Bed scour

running water. Bank erosion rates can be deter-
mined from a comparison of maps and charts, Many serious bank erosion problems are usually
aerial photographs and satellite images of differ- related to the meander planform. Erosion often
ent dates, repeated surveying along established occurs at the outer banks in a bend, as a result of
baselines or through monitoring with erosion both elevated near-bank velocities and deep bed
GFE-10.qxd
6/24/05
Table 10.3 A summary of the criteria used in the Rosgen channel classification system (after Rosgen, 1994).
Stream type Description Entrenchment ratio W/D ratio Sinuosity Slope Landform/soils/features
3:44 PM
Aa Very steep, deeply entrenched, 1.4 12 1.0–1.1 0.10 Very high relief. Erosion, bedrock or
debris transport streams. deposition features; debris flow
potential. Deeply entrenched streams.
Page 275
Vertical steps with deep scour pools,
waterfalls.
A Steep, entrenched, cascading, 1.4 12 1.0–1.2 0.04–0.10 High relief. Erosive or depositional
step/pool streams. High energy/ and bedrock forms. Entrenched and
debris transport associated with confined streams with cascading
depositional soils. Very stable if reaches. Frequently spaced deep
bedrock or bedrock dominated pools associated step-pool bed
channel. morphology.
B Moderately entrenched, moderate 1.4–2.2 12 1.2 0.02–0.039 Moderate relief, colluvial deposition
gradient, riffle-dominated and/or residual soils. Moderate
channel with infrequently spaced entrenchment and width/depth
pools. Very stable plan and ratio. Narrow, gently sloping
profile. Stable banks. valleys. Rapids predominate with
occasional pools.
C Low gradient, meandering, point 2.2 12 1.4 0.02 Broad valleys with terraces, in
bar, riffle-pool, alluvial channels association with floodplains, alluvial
with broad well defined soils. Slightly entrenched with well
floodplains. defined meandering channel.
Riffle-pool bed morphology.
D Braided channel with longitudinal N/A 40 N/A 0.04 Broad valleys with alluvial and colluvial
and transverse bars. Very wide fans. Glacial debris and depositional
channel with eroding banks. features. Active lateral adjustment,
with abundance of sediment supply.
RIVERS
275
GFE-10.qxd
276
6/24/05
3:44 PM
DA Anastomosing narrow and deep 4.0 40 Variable 0.005 Broad, low gradient valleys with fine
with expansive well vegetated alluvium and/or lacustrine soils.
floodplain and associated Anastomosed, fine deposition with
wetlands. Very gentle relief well vegetated bars that are laterally
Page 276
with highly variable sinuosities. stable with broad wetland floodplains.
Stable streambanks.
E Low gradient, meandering, riffle- 2.2 12 1.5 0.02 Broad valley/meadows. Alluvial
pool stream with low width/depth materials with floodplain. Highly
ratio and little deposition. Very sinuous with stable, well vegetated
efficient and stable. High banks. Riffle-pool morphology with
meander width ratio . very low width/depth ratio.
F Entrenched meandering riffle-pool 1.4 12 1.4 0.02 Entrenched in highly weathered
channel on low gradients with material. Gentle gradients, with a
high width/depth ratio. high width/depth ratio. Meandering,
laterally unstable with high bank
erosion rates. Riffle-pool
morphology.
G Entrenched gulley step-pool and 1.4 12 1.2 0.02–0.039 Gulley, step-pool morphology with
low width/depth ratio on moderate slope and low width/depth
moderate gradients. ratio. Narrow valleys or deeply
incised in alluvial or colluvial
materials i.e. fans, deltas. Unstable
with grade control problems and
high bank erosion rates
Note that ‘entrenchment’ is the ratio of channel width to bank height and ‘W/D ratio’ is the ratio of the channel width to depth.
GFE-10.qxd 6/24/05 3:44 PM Page 277
RIVERS 277
S. 'STABLE'
D. 'DEPOSITIONAL' M. 'LATERAL MIGRATION' E. 'ENLARGING'
No observable indication of
Consistent decrease in channel Migration of most bends. Consistent increase in channel
morphological adjustment in
width and/or depth cross-sectional dimensions preserved width and/or depth by erosion
progress
m. 'lateral migration'
d. 'depositional' initiation of alternating bank e. 'enlarging'
Selective deposition creating erosion in straightened channels initiation of continuous erosion, often
reduced width channel of erosion of only sharpest at channel toe
bends
R. 'RECOVERING'
U. 'UNDERCUTTING'
C. 'COMPOUND' Development of a sinuous
Continuous erosion and
Aggradation of channel bed with channel within straightened
migration of full width channel,
erosion of channel banks channels, including selective
coarse inner bank deposits
erosion of outer banks
Figure 10.16 Channel classification, based on types of change (from Downs, 1995).
scour adjacent to the bank. The magnitude of the 10.7 Floodplain accretion
near-bed velocity increases and the additional
scour depth due to the bend effects are related Floodplains accumulate by two processes:
to the shape and, particularly, to the curvature of
1. Vertical accretion from over-bank flow, pro-
the bend.
ducing thin laminae of fine sediments
Potential scour depth can be estimated from
2. Lateral accretion from the migration of
the maximum flow depth during flood flows,
channels across a floodplain. This results in a
using a simple empirical approach (Neill, 1973).
fining upwards sequence of gravels and sands,
A multiplication factor is applied to the maximum
capped by finer sediments.
flow depth as shown in Table 10.5.
Floodplain accretion rates in temperate rivers
Scour depth (Mean depth Multiplication generally vary between fractions of mm per year
Factor) Mean depth to a few cm per year (Reid and Frostick, 1994).
However, the potential for faster rates should be
The estimates are potential scour depths. In many considered along rivers prone to flash floods or
channels this predicted depth may not be realised debris flows. In Bijou Creek, Kansas, USA, 1 m
as there may be only a thin veneer of loose sedi- of sediment was deposited in a flash flood
ment above bedrock. (McKee et al., 1967). At Shishkat in the Hunza
GFE-10.qxd 6/24/05 3:44 PM Page 278
Table 10.4 Field indicators of channel instability (from Sear and Newson, 1992; Newson et al., 1997).
Uplands Upland margins/Piedmont Lowlands
Evidence of erosion Perched boulder berms Terraces Old channels

Terraces Old channels Undermined structures
Old channels Narrow/deep channels Exposed tree roots
Old slope features Undermined structures Narrow/deep channel
Undermined structures Exposed tree roots Bank failures
Exposed tree roots Narrow/deep channel Deep gravel exposed
Narrow/deep channel Bank failures
Bank failures both banks Armoured/compacted bed
Armoured/compacted bed Deep gravel exposed
Gravel exposed in banks
Evidence of aggradation Buried structures Buried structures Buried structures
Buried soils Buried soils Buried soils
Large uncompacted bars Large uncompacted bars Large silt/clay banks
Eroded banks at shallows Eroded banks at shallows Eroded banks at shallows
Contracting bridge space Contracting bridge space Contracting bridge space
Deep fine sediment in bank Deep fine sediment in bank Deep fine sediment in
bank
Many unvegetated bars Many unvegetated bars Many unvegetated bars
Evidence of stability Vegetated bars and banks Vegetated bars and banks Vegetated bars and banks
Compacted weed-covered Compacted weed-covered Weed-covered bed
bed bed
Bank erosion rare Bank erosion rare Bank erosion rare
Old structures in position Old structures in position Old structures in position
Valley, Pakistan, over 100 m of sediment was lead to excessive settlements of foundations and
deposited by a single debris flow event in 1973 embankments. Alternatively they may be filled
(Brunsden, 1996). with coarse material and constitute high-perme-
Figure 10.19 illustrates a range of typical ability channels incised into low-permeability
floodplain landforms. Levees develop on channel rocks or soils and cause problems with dewatering
margins as a result of fall-out of the coarser com- of excavations or inflows into tunnels. They may
ponents of the suspended load. There is a rapid have been buried under later floodplain sediment
decline in grain size, sediment thickness and, and not be apparent at the surface. Buried river
hence, elevation away from the channel. Splays channels may be bigger, deeper and at lower lev-
can develop around localised breaches at low els than would be expected from present-day river
points along the levees. Away from the levees the systems.
over-bank deposits are predominantly fine silts
and clays. Ground conditions are likely to be very
variable both horizontally and vertically over rela- 10.8 River terraces: phases of
tively short distances in floodplain environments, aggradation and incision
especially where river channels have been contin-
ually changing. River terraces are relatively flat surfaces within
Old river channels can present problems for a river valley, separated from the active flood-
new developments. For example they may have plain by a scarp slope (terrace riser). These ter-
become filled with soft organic silts and clays and races are the product of channel incision into
GFE-10.qxd
6/24/05
3:44 PM
Page 279
Figure 10.17 Bank failure modes (redrawn from Hey et al., 1991).
RIVERS
279
GFE-10.qxd 6/24/05 3:44 PM Page 280
Figure 10.18 A simple method for assessing bank stability: dimensionless stability chart for rotational and
slab-type failures in cohesive materials (redrawn from Osman and Thorne, 1988).
GFE-10.qxd 6/24/05 3:44 PM Page 281
RIVERS 281
Table 10.5 Estimated potential scour depth (after Neill, 1973).
Location Multiplying Factor Scour Depth (10m Flow)
Straight channel reach 1.25 2.5

As above with mobile bed dunes 1.50 5.0
Moderate channel bend 1.50 5.0
Severe channel bend 1.75 7.5
Right-angle abrupt channel bend 2.00 10.0
Figure 10.19 Typical floodplain morphology (from

Reid and Frostick, 1994).
Figure 10.20 Paired and unpaired river terraces (from
Sparks, 1972; Thornbury, 1969).
the floodplain. Paired terraces occur where inci- the product of how the system seeks to adjust to a
sion is rapid compared with the rate of lateral new equilibrium form (i.e. complex behaviour;
channel migration; unpaired terraces tend to Chapter 1).
develop along rapidly migrating channels
(Figure 10.20).
River incision and terrace formation are gener- 10.9 Floods
ally the product of changes in channel gradient
(e.g. base level change caused by neotectonics or Floods occur when water levels rise to overflow
sea-level change), sediment load (e.g. as a result land not normally submerged; this can occur in a
of land use change) or discharge (e.g. as a result variety of settings from ‘dry valleys’ to river
of climate change). However, terrace formation floodplains and coastal lowlands. Although floods
may also be an inherent feature of the river system can result from a variety of factors (Chapter 2
behaviour and not necessarily the product of Figure 2.11), the most common causes are rainfall
external factors. For example, a stream channel and snowmelt. Amongst the most important fac-
may experience repeated phases of aggradation tors include the following regime characteristics:
and incision under a relatively constant climate The catchment area affects the total volume of
and sediment supply; valley floor aggradation stream-flow generated by a catchment-wide event
occurs until the channel gradient exceeds a crit- (i.e. the larger the catchment the greater the
ical angle, initiating a phase of rapid bed erosion. potential rainfall input).
Figure 10.21 illustrates how a single impulse Slope characteristics influence the amount of
(base level fall) can set in motion a pattern of runoff produced by an event (i.e. slope angle,
episodic incision and aggradation. In this model, bedrock geology, soil type, land use and vegeta-
valley infills and terrace development are simply tion cover).
GFE-10.qxd 6/24/05 3:44 PM Page 282
Figure 10.22 A river flood hydrograph (from Ward,

1978).
By contrast, artificial drainage, such as field

drains, helps speed up the movement of water
towards the channel network.
Channel characteristics influence the ability of
the channel to carry a flood flow. Channel cap-
acity is not constant; deposition of eroded sedi-
ments can significantly reduce the channel depth
and cross-section. Entrapment of debris behind
structures may cause ‘backing-up’ behind these
temporary dams, and lead to over-topping of the
banks.
Antecedent conditions determine the amount
of the catchment that is saturated prior to a rain-
storm or snowmelt event and, hence, the amount
Figure 10.21 Schematic representation of the com- of runoff that is generated. The river or stream
plex behaviour of a river system to a change in base
level (from Summerfield, 1991). level prior to an event is also critical as this will
influence whether the channel system can carry
the additional runoff.
Network characteristics influence the speed at The hydrogeological characteristics of a river
which water is transmitted through the channel flood can be described with reference to the flood
system. In general, the time of rise of floodwaters hydrograph, the continuous trace of discharge over
after a rainstorm or snow melt event will be deter- time (Figure 10.22). Flooding does not begin
mined by a range of factors related to the nature immediately at the onset of rain, as the initial
of the drainage network itself. Dendritic networks increase in discharge is contained within the chan-
(Figure 10.2) tend to produce a marked concen- nel system. The rate of water level rise, the magni-
tration of flow in the lower catchment as floodwa- tude of the peak flow and duration of flooding are
ters are delivered down the major tributaries at a central in defining the nature of a flood event and
similar speed; trellised networks tend to produce its impact; these attributes can vary significantly
a more muted response. The number and size of from ‘flashy’ streams which can have a high peak
lakes, reservoirs or other storage areas can be a and short duration and ‘sluggish’ streams which
significant factor in reducing the size of a flood. may have a relatively low peak and long duration.
GFE-10.qxd 6/24/05 3:44 PM Page 283
RIVERS 283
In other types of flooding, climate is only Table 10.6 A range of exponents that can be used for
partly or indirectly responsible. In many estuaries estimating suspended sediment yield in different river
flooding is often caused by the ponding back of environments (from Reid and Frostick, 1994).
high river discharges by rising tides. Country Environment Exponent a Exponent b
Kenya Arid 2570 0.512

10.10 Sediment transport Mexico Arid 100 0.700
Israel Arid 4217 0.159
Along many rivers, sediment has to be regularly Austria Temperate,
removed to maintain navigable channels or control humid 0.004 2.200
potential flooding problems (sediment accu- USA Temperate,
mulation reduces channel capacity and, hence, sub-humid 0.01 1.600
increases flood risk). In order to address the sig- USA Temperate,
nificance of sedimentation problems and identify humid 40 2.500
appropriate solutions (e.g. sediment traps or nar- Germany Temperate,
rowing of the cross-section) it is helpful to develop humid 31 1.391
a sediment budget for the catchment or section of
interest. In addition to identifying potential sedi-
Muller (1948) equation and the Bagnold (1980)
ment source areas (e.g. eroding hillsides and chan-
equation. The Meyer-Peter and Muller (1948)
nel banks) it is useful to estimate the volumes of
equation is:
sediment that can be carried into the reach (i.e. the
sediment transport rates). s
Rivers transport sediment in a variety of ways: qb
In solution: soluble materials may be dissolved s
3 2
in water and flushed into rivers, especially during  ( K b Kg )2 3 Y S 0.047{( s ) }Dm 
the initial stages of a storm. In terms of the total  
amount of material removed by rivers, the solute  (0.25 )( g)1 3 
component can be high.
In suspension: suspended sediments are trans- where qb is the specific bed load flux rate in sub-
ported in the flowing body of the water where they aerial mass terms (kg/s/m); is the specific
are supported by turbulent eddies in the flow. The weight of the fluid; s is the specific weight of the
finest materials (clays) may be evenly distributed sediment; Kb the bed roughness as u/Y2/3S1/2
through the water column, whereas coarser mater- (where u is the mean water velocity); Kg specifies
ials may be concentrated near the bed. grain roughness as 26/D901/6 (D90 is the ninetieth
Bed load: material is moved along the stream percentile of the bed surface grain size distribu-
bed. tion); Y is water depth; S is the slope of the water
For river management the suspended and bed surface; Dm is the effective diameter of the sur-
loads are the most significant. Suspended sediment face bed sediments defined as 兺Di Pi/100 (where
concentrations may be estimated from a sediment Di and Pi are the average diameter and percentage
rating curve (C) of the form: fraction by weight of the ith size fraction — often
C a Qb approximates D64 in gravel bed rivers); g is the
A range of values of the exponents a and b for acceleration due to gravity.
different types of river are presented in Table 10.6 The Bagnold (1980) equation is:
(Q is the discharge).
3 2
Estimates of bed load can be made by s  o 
using one of a number of empirical or semi- ib . ibr .  
s  ( o )r 
theoretical equations that have been developed.
Amongst the most useful are the Meyer-Peter and (Y Yr )2 3 ⋅ ( D Dr )1 2
GFE-10.qxd 6/24/05 3:44 PM Page 284
where ib is the specific bed load flux; is the The conservation need is for the continuance of
stream power (W/m2) ( gQS as per secion the processes not for preservation of forms cre-
10.3); o 5.75 (0.04( s ))3/2(g/)1/2D3/2 ated. Channelling of mobile river reaches com-
log10(12Y/D) (W/m2); D D50 for unimodal bed monly involves direct destruction and removal of
sediment size distributions (m); ibr 0.01 kg/s/m plants and habitats, particularly the loss of pools
and is an arbitrary reference value for bed load and riffles, vertical eroding banks and sinuous
flux rate; ( o)r 0.5 (W/m2) and is a refer- courses. Changes in fish populations can result
ence value for excess specific stream power; from the loss of natural rifle-pool sequences that
Yr 0.1 m; Dr 0.0011 m. provide a variety of low flow conditions suitable
These equations give an indication of potential as cover for both fish and the organisms on which
sediment transport rather than actual. The avail- they feed (Brookes, 1988). Fish require sheltered
ability of sediment often limits the transport rate. water in fast flowing rivers, these conditions may
Bed armouring can also affect the transport rate, be absent where a meandering stream has been
especially in gravel bed rivers where a coarse sur- artificially straightened. Changes to the channel
face layer often protects underlying finer sedi- width and depth may create unsuitable habitats or
ments, preventing its entrainment and removal. In present topographical difficulties for fish migra-
such cases rivers may experience two phases of tion. Clearance of bank-side vegetation may
sediment transport. Phase I involves a ‘through- destroy valuable cover for fish.
put’ load with little or no bed disturbance. Once
the armour is disrupted (e.g. by higher velocity
flows) the finer material can be rapidly removed 10.12 Soft engineering and
during Phase II transport. This provides an indi- river restoration
cation of the accelerated bed erosion that can
follow inadvertent disruption of the bed armour Often river engineering works are an integral
(e.g. during pipeline construction). part of flood defence and channel stabilisation
measures. It is now recognised that so-called
‘soft engineering’ designs (that work with natural
10.11 The significance for processes rather than attempting to impose solu-
conservation tions) are likely to be the most cost-effective
solutions, requiring less maintenance and min-
River beds, banks and floodplains are all import- imising the environmental impact. Soft engineer-
ant in supporting particular species or communi- ing solutions can also help restore or rehabilitate
ties of plants and animals throughout or during reaches that have been damaged by previous
parts of their life cycle. For example, some fish channelling works.
require shallow water and gravels, as found on rif- Regime equations can be used to predict the
fles, in which to spawn but commonly live in the geometry of stable channels, from discharge,
deeper water at other times. sediment load, bed and bank materials and valley
Natural channel changes can be vital to the slope (section 10.3). For gravel bed rivers, Hey
maintenance of both channel and floodplain sites. and Thorne’s (1986) regime equations offer a
Floods tend to maintain ecological diversity by method for predicting stable channels.
removing vegetation which may have clogged a
channel, by recreating bare vertical banks which
have become overgrown, or by forming new riffles 10.13 Summary: the catchment
and pools (Lewis and Williams, 1984). In addition, approach to river engineering
maintenance of natural vegetation and ecological
diversity can help maintain water quality through Although it is tempting to treat river channel
natural processes of purification. change hazards as site-specific problems, this can
GFE-10.qxd 6/24/05 3:44 PM Page 285
RIVERS 285
References
Bagnold, R. A. (1980) An empirical correlation of
bedload transport rates in flumes and natural rivers.
Proceedings of the Royal Society, London A, 372,
453–473.
Bromhead, E. N. (1986) The Stability of Slopes. Surrey
University Press.
Brookes, A. B. (1988) Channelized Rivers: Perspective
for Environmental Management. Wiley, Chichester.
Brunsden, D. (1996) Geomorphological events and
landform change. Zeitschrift für Geomorphologie
40, 273–288.
Carr, A. P. (1962) Cartographic error and historical
accuracy. Geography 47, 135–144.
Church, M. and Jones, D. (1982) Channel bars in
gravel bed rivers. In Hey, R. D., Bathurst, J. C. and
Thorne, C. R. (eds) Gravel-bed Rivers. Wiley,
Chichester, 291–338.
Cooke, R. U. and Reeves, R. W. (1976) Arroyos and
Environmental Changes in the American Southwest.
Oxford University Press.
Doornkamp, J. C. (1982) The physical basis for plan-
ning in the Third World. Third World Planning
Review 4, 11–31; 111–128; 213–246.
Downs, J. (1995) Estimating the probability of river
channel adjustment. Earth Surface Processes and
Landforms 20, 687–705.
Dury, G. H. (1955) Bed width and wave-length in
meandering valleys. Nature 176, 31.
Figure 10.23 Typical ‘knock-on’ effects within a Ferguson, R. I. (1981) Channel forms and channel
drainage basin (redrawn from Sear and Newson, 1992). changes. In Lewin, J. (ed.) British Rivers. George
Allen & Unwin, London, 90–125.
Gregory, K. J. (1976) Drainage networks and climate. In
Derbyshire, E. (ed.) Geomorphology and Climate.
lead to frequent and costly maintenance. A more Wiley, Chichester, 289–315.
efficient approach is to view the site problems Gregory, K. J. and Walling, D. (1973) Drainage Basin
as a symptom of broader-scale processes that Form and Process. Arnold, London.
operate throughout a catchment. Many problems Harvey, A. M. (1986) Geomorphic effects of a 100 year
storm in the Howgill Fells, Northwest England.
can be related to the way water and sediments are Zeitschrift für Geomorphologie 30, 71–91.
transported from supply areas towards the coast Hey, R. D. and Thorne, C. R. (1986) Stable channels
and held in, or released from, temporary stores with mobile gravel beds. Journal of Hydraulic
such as bars or spreads on the river bed. Figure Engineering, American Socity of Civil Engineers
10.23 highlights the linkages between different 112 (8), 671–689.
Hey, R. D., Heritage, G. L., Tovey, N. K., Boar, R. R.,
parts of a river system and illustrates the knock- Grant, N. and Turner, R. K. (1991) Streambank
on effects of a large sediment input in the uplands. Protection in England and Wales. NRA, London R &
Within the catchment framework, engineering D Note 22.
geomorphology can be used to identify the causes Hollis, G. E. and Luckett, J. K. (1976) The response of
and significance of site specific problems; iden- natural streams to urbanisation: two cast studies
from south-east England. Journal of Hydrology 30,
tify the potential impacts (upstream and down- 351–363.
stream) of proposed works and identify suitable Hooke, J. M. (1991) Non-linearity in River Meander
mitigation measures, including channel restoration Development. University of Portsmouth Working
operations. Paper No. 19.
GFE-10.qxd 6/24/05 3:44 PM Page 286
Hooke, J. M. and Kain, R. J. P. (1982) Historical Osman, A. M. and Thorne, C. R. (1988) Riverbank sta-
Change in the Physical Environment: a Guide to bility analysis I: Theory. Proceedings of the
Sources and Techniques. Butterworths, Sevenoaks. American Society of Civil Engineers, Journal of
Knight, C. R. (1979) Urbanisation and natural stream Hydraulic Engineering 117 (8), 1091–1092.
channel morphology: the case of two English new Petts, G. E. (1984) Impounded Rivers. Wiley, London.
towns. In Hollis, G. E. (ed.) Man’s impact on the Ray, R. G. and Fischer, W. A. (1960) Quantitative
hydrological cycle in the United Kingdom. Geobooks, photography — a geologic research tool. Photogram-
Norwich, 181–198. metric Engineering 25, 143–150.
Knighton, D. (1984) Fluvial Forms and Processes. Reid, I. and Frostick, L. E. (1994) Fluvial sediment
Arnold, London. transport and deposition. In Pye, K. (ed.) Sediment
Lawler, D. M. (1992) Design and installation of a Transport and Depositional Processes. Blackwell,
novel automatic erosion monitoring system. Earth Oxford, 89–155.
Surface Processes and Landforms 17, 455–463. Richards, K. (1982) Rivers: Form and Process in
Lee, E. M. (1995) The Occurrence and Significance of Alluvial Channels. Methuen, London.
Erosion, Deposition and Flooding in Great Britain. Richards, K. S. and Lane, S. N. (1997) Prediction
HMSO, London. of morphological change in unstable channels. In
Leopold, L. B. and Maddock, T. (1953) The Hydraulic Thorne, C. R., Hey, R. D. and Newson, M. D. (eds)
Geometry of Stream Channels and some Physio- Applied Fluvial Geomorphology for River Engin-
graphic Implications. United States Geological eering and Management. Wiley, London, 269–292.
Survey Professional Paper 252. Roberts, C. R. (1989) Flood frequency and urban-
Lewis, G. and Williams, G. (eds) (1984) Rivers and induced channel changes: some British examples.
Wildlife Handbook. RSPB, Lincoln. In Beven, K. and Carling, P. A. (eds) Floods:
Macklin, M. G. and Lewin, J. (1997) Channel, floodplain Hydrological, Sedimentological and Geomorpho-
and drainage basin response to environmental change. logical Implications. Wiley, London, 57–82.
In Thorne, C. R., Hey, R. D. and Newson, M. D. (eds) Rosgen, D. L. (1994) A classification of natural rivers.
Applied Fluvial Geomorphology for River Engineer- Catena 22, 169–199.
ing and Management. Wiley, London, 15–45. Schumm, S. A. (1977) The Fluvial System. Wiley,
McKee, E. D., Crosby, E. J. and Berryhill, H. L. (1967) New York.
Flood deposits of Bijou Creek, Colorado, June 1965. Sear, D. A. and Newson, M. D. (1992) Sediment and
Journal of Sedimentary Petrology 37, 829–851. Gravel Transportation in Rivers including the use of
Meyer-Peter, E. and Muller, R. (1948) Formulas of bed- Gravel Traps. NRA Project Report 232/1/T.
load transport. International Association for Sparks, B. W. (1972) Geomorphology. Longman,
Hydraulic Structures Research, Report of Second London.
Meeting, Stockholm, 39–64. Summerfield, M. A. (1991) Global Geomorphology.
Mitchell, C. (1973) Terrain Evaluation. Longman, Longman, Harlow.
London. Thornbury, W. D. (1969) Principles of Geomorphology.
Neill, C. R. (1973) Guide to Bridge Hydraulics. Roads Wiley, New York and London.
and Transportation Association of Canada, University Thorne, C. R. (1997) Channel types and morphological
of Toronto Press. classification. In Thorne, C. R., Hey, R. D. and
Newson, M. D., Hey, R. D., Bathurst, J. C., Brookes, A., Newson, M. D. (eds) Applied Fluvial Geomorphology
Carling, P. A., Petts, G. E. and Sear, D. A. (1997) for River Engineering and Management. Wiley,
Case studies in the application of geomorphology to London, 175–222.
river management. In Thorne, C. R., Hey, R. D. and Ward, R. (1978) Floods. Macmillan, London.
Newson, M. D. (eds) Applied Fluvial Geomorph- Wolman, M. G. and Miller, J. P. (1960) Magnitude and
ology for River Engineering and Management. frequency of forces in geomorphic processes.
Wiley, London, 311–363. Journal of Geology 68, 54–74.
GFE-11.qxd 6/27/05 2:06 PM Page 287
11. Soil Erosion

Mark Lee and John Charman
11.1 Introduction the disaggregation of minerals into a soil com-

prising a collection of discrete particles. Rainfall
Soil erosion is a three-stage process comprising quantity, duration and intensity influence the rate
detachment of material, its transport by water or, of erosion in which disaggregated particles are
less frequently, wind and the subsequent deposition detached and transported.
when sufficient energy is no longer available to The highest rates of erosion are often associated
transport the material. The main factors involved with deforestation or during the construction phase
in the process are summarised in Figure 11.1. The of development. Sediment yields during construc-
severity of erosion depends on the quantity of tion can be up to 2000 times greater than those
material supplied by detachment and the ability of from undisturbed forested areas (Table 11.1a). For
the running water or wind to carry it. As described example, in Maryland USA, Wolman and Schick
in the following sections, erosion can be viewed as (1967) demonstrated that sediment yields during
a function of the power of water or wind (erosivity) construction reached 55 000 t/km2 per year,
and the resistance of the material (erodibility). whereas neighbouring undisturbed areas yielded
Recognition of the relative significance of these less than 400 t/km2 per year.
factors provides the basis for developing appro- Erosion, by water or wind, presents a variety of
priate soil erosion control or soil conservation problems for engineers and environmental man-
strategies. agers. It can severely limit the long-term sustain-
Climate is an important control on the soil able use of agricultural land and hinder policies
erosion process (Chapter 2). Temperature, both that seek to deliver greater, more reliable crop
seasonal and daily, together with rainfall, influences yields. The Ethiopian Highlands, for example,
the rate and type of weathering. Mechanical have lost around 1 Mt of topsoil per year as a
weathering may cause breakage of rock into more result of overuse. In El Salvador, 77% of the land
closely fractured components while chemical area is severely eroded. The ‘dust bowl’ years of
weathering causes decomposition of the rock and the 1930s in southern USA are perhaps the most
Table 11.1a Representative rates of erosion in the

Climate Geology USA (from US Environmental Protection Agency,
1973).
Land Use Sediment Yield Relative to

Rainfall Soil Type t/km2 per year Forestry ⫽ 1
Forest 8.5 1
Grassland 85 10
Vegetation Soil Erosion Abandoned surface 850 100
mines
Cropland 1 700 200
Figure 11.1 The main factors involved in the process Construction sites 17 000 2000
of soil erosion by water (from Selby, 1987).
287
GFE-11.qxd 6/27/05 2:06 PM Page 288
Table 11.1b Annual rates of erosion in selected coun- catches between 1960 and 1978. Sedimentation in
tries (sediment yield t/km2 per year) (from Morgan, watercourses leads to reduced channel capacity
1995). and, hence, increased flood risk. High sediment
Natural Cultivated Bare Soil loads may also trigger significant alluvial river
channel changes (Chapter 10).
China 10–200 15 000–20 000 28 000–36 000 Construction and post-construction problems
USA 3–300 500–17 000 400–900 associated with accelerated sediment yields off
Australia 0–6400 10–15 000 4400–8700 bare ground and subsequent reinstatement diffi-
Nigeria 50–100 10–3500 300–15 000 culties can be major issues for the environmental
India 50–500 30–4000 1000–18 500
impact of engineering projects such as roads,
Ethiopia 100–500 800–4200 500–7000
railways or pipelines.
famous example of disastrous wind erosion, 11.2 Types of erosion features

damaging 9 Mha and generating large dust storms
(Worster, 1979). Figure 11.2 presents a simple model that high-
The ‘on-site’ problems of erosion include the lights the dominant slope processes that generally
encroachment of gully-heads onto an alignment operate on different units within a slope profile.
or site, erosion loss of land or exposure of Soil erosion tends to be concentrated in units 5
services. ‘Off-site’ impacts, include blocked roads and 6. A variety of distinctive erosion types and
and ditches and damage to property. In the UK, forms can be recognised, including:
over thirty separate damaging erosion events Sheet erosion (inter-rill erosion): the washing
(so-called ‘mudfloods’) were reported in the of surface soil from hillslopes.
South Downs between 1976 and 1990 affecting Pedestal erosion: where stones or tree roots
around 200 houses, with the scale of problems protect erodible soil from rain-splash erosion.
ranging from inundation by soil-laden water to Although these features generally develop slowly
damage to outbuildings and gardens (Boardman, over many years, they may also occur on agricul-
1990). Estimates of costs associated with the tural land in response to intense storms.
events suggest a minimum figure of £836,000 Rills: small, linear channels that have been cut
over a 15-year period since 1976. into a slope as a result of the concentration of
Erosion can cause the loss of water storage overland flow. Rills often occur as a dense net-
capacity as soil removed from upland areas is work of parallel channels, occasionally feeding a
deposited in reservoirs. For example, the capacity master rill. Although they tend to be ephemeral
of the San Gabriel Dam, one of the Los Angeles features they can have considerable erosive power.
County Drainage Area reservoirs declined from Once rills have formed they may migrate upslope
65 Mm3 in 1938 to 55 Mm3 by 1980. During this by headcut retreat. Maximum movement occurs
period some 22 Mm3 of sediment had been removed when the depth of water flow is about equal to the
at a cost of $20M (1981 prices, Bruington, 1982). particle diameter, so that as the water becomes
A decline in water quality, involving the trans- concentrated into rills so its ability to carry larger
fer of agricultural chemicals to watercourses, dis- particles increases. Thus, still at a small scale, the
coloration and increased cost of treatment, due to aggregated particles become at risk and the
high sediment yields from eroding catchments process self perpetuates as the water/sediment
may also create problems for fisheries. Of particu- mixture scours the bottoms and sides of the rills,
lar concern is the possible impact of increased erodes the head of the channel and causes mass
sediment loads on salmon and trout rivers. For slumping from the over-steepened head and sides.
example, Drakeford (1979) correlated the reduc- The amount of soil detached is in proportion to
tion in catches of salmonoid fish in the River the square of the velocity, whereas the sediment
Fleet, Scotland, with the expansion of forestry in transport potential increases in proportion to the
the catchment, noting a 90% decline in sea trout fifth power of the velocity.
GFE-11.qxd 6/27/05 2:06 PM Page 289
SOIL EROSION 289
Figure 11.2 A simple slope profile model relating slope units to dominant processes (from Dalrymple et al., 1968).
Piping: subsurface tunnels or pipes can develop occurrence of a weak clay horizon or gulley heads
where through-flow is concentrated along erodible and landslide backscars (Figure 11.3). Piping can
soil layers (Jones, 1981), often forming a separate be prevalent in semi-arid regions where the soils
subsurface drainage system that can cut across contain swelling clays that crack on drying. Pipes
topographic divides. There is a number of com- often collapse and initiate gullying. Collapses
mon settings that are prone to piping; a permeable under roads or buildings can cause significant
soil horizon below an impermeable horizon, the problems.
GFE-11.qxd 6/27/05 2:06 PM Page 290
Figure 11.3 Settings prone to soil piping: (a) cracking and a permeable horizon below an impermeable horizon;
(b) an horizon of dispersible clay and (c) a gully head (from Selby, 1982).
associated with an acceleration in the degradation

of an area, in response to increased runoff (e.g. as
a result of changing patterns of rainfall intensity
or land use change), or a change in slope gradient
(e.g. undercutting by a stream or pipe collapse).
Gullies are most common in weak materials
(e.g. loess, alluvium, colluvium, gravels and
deeply-weathered soils).
The characteristic stepped profile can take two
forms. Knickpoints occur where there is a channel
both upstream and downstream, and are related to
the upslope limit of the current phase of erosive
activity (e.g. in response to a change in base level
or climate) or a change in lithology. Gully-heads
occur where there is a switch from un-channelled
hill-slope to channelled flow.
Gully development may involve the random
convergence of flow around vegetation; local
disturbance of the ground surface which reduces
its resistance to shear (e.g. grazing or burrowing)
and the through-flow or groundwater seepage
Figure 11.4 Stages in gully development (after
Leopold et al., 1964). (i.e. spring sapping). This process requires
large flows and is most efficient in unconsoli-
dated layered sediments where the limiting
Gullying: from a practical viewpoint gullies factor is the transport capacity of the spring-fed
are rills that are so large and well established that flows (Thornes, 1994). Gully development
they cannot be crossed by traffic (Hudson, 1995). may also involve mass movement of soil into a
They are steep-sided, deeply incised features with gully from the sides and head, which have been
characteristic steep headcuts and a stepped long- over-steepened by the scouring effect of the
profile (Figure 11.4). These features are generally channel flow.
GFE-11.qxd 6/27/05 2:06 PM Page 291
SOIL EROSION 291
Figure 11.5 Composite of badlands lithogeomorphic characteristics. (1) Headcutting by valley-bottom gully in
alluvial fill with accompanying slump and collapse by basal sapping. (2) Pipe-induced collapse initiates discontin-
uous gully. (3) Surface drainage into sinks and pipe shafts controlled by structural discontinuities. (4) Valley-side
gullies fed by convergent rill flow. (5) Percoline-controlled large-scale piping generates slope collapsed rills. (7)
Gully heads expanding into undissected surface. (8) Early stage of meso-scale pipe development. (9) Mature meso-
scale piping initiating slope collapse, large gulies and tributary valley formation. (10) Multiple cut-and-fill gully
deposits in bedrock-floored valley. (11) Hoodoos formed by dissection of resistant caprock (from Campbell 1997).
There may be successive cycles of gully activ- 2. Valley-side gullies, formed where runoff is con-
ity. Initiation is often accompanied by rapid inci- centrated on hillslopes, subsurface pipes col-
sion, channel widening and gully-head retreat. lapse or landslides create elongated scars. The
Over time the channel gradient is reduced and the typical form comprises a broad arcuate head
gully may stabilise as the channel infills and the and a narrow outlet channel downslope. They
flanks degrade and re-vegetate. The next phase of tend to occur in swarms across a hillside or
activity may be triggered by climate or land use entire landscape.
change. In many instances, different sections of
a gully may experience stability or activity at Badlands: badland terrain has a characteristic
different times. For example, the head may be assemblage of steep slopes dissected by rills
actively retreating while the lower section is and gullies, narrow elongated ridgelines, high
stabilising through active deposition. drainage density, thin bare soils, rapid erosion and
Two main types of gully occur: shallow landsliding (Figure 11.5; Howard, 1994;
Campbell, 1997). In many badlands the steep
1. Valley-floor gullies, including ephemeral fea- eroding slopes abut almost flat alluvial surfaces.
tures to more deeply incised channels (e.g. Among the best known examples are the Mancos
entrenched arroyos and wadis). These features Shale badlands in the Henry Mountains, Utah
occur where runoff is concentrated in alluvial (Howard, 1994), the Brule Formation shale bad-
valley floors. lands of South Dakota (Schumm, 1956) and the
GFE-11.qxd 6/27/05 2:06 PM Page 292
Figure 11.6 Slope and valley development in badland terrain by combination piping, collapse and gullying. (A)
Three-stage meso-scale evolution of slope and valley information from initial piping (1) through pipe-induced col-
lapse at (2), to final valley from at (3), where hydraulic gradients are no longer steep enough to generate extensive
large pipe systems. (B) Macrovalley development from the early deep incision of trunk streams at (1) and extensive
pipe formation with steep topographic and hydraulic gradients, through (2) and (3) gradually diminishing as bad-
lands extend into upper surface with large pipe collapse triggering mass movements, to the stage at (4), where pip-
ing ceases to play the major role in badland valley evolution because of reduced hydraulic and topographic gradients
(from Campbell, 1997).
Loess Plateau badlands in China (Chen, 1983; voids. The side-slopes may be at the angle of repose
Chapter 25). of dry weathered debris. Consequently these slopes
These erodible terrains can produce extremely are almost permanently on the verge of failure when
high sediment yields. The Loess Plateau of wetted.
China, for example, generates around 38 000 t/ Variations in form range from smooth rounded
km2 per year (Chen, 1983). The estimated annual ridge crests and convex side-slopes to pinnacle
yield from the Cheyenne River Badlands of badlands with knife-edge divides, reflecting
South Dakota is 13 500 t/km2 (Hadley and changes in lithology and climate. Pinnacle bad-
Schumm, 1961). The Alberta Badlands were lands develop where mudrocks occur beneath a
shown to erode at up to 83 mm/year, with an aver- caprock which erodes very slowly compared with
age annual rate of around 38 mm/year (Campbell, the surrounding slopes. Eventually the caprock is
1981). lost and the underlying rocks erode rapidly, devel-
Badlands are generally associated with weak, oping fluted forms because of rapid rill erosion on
impermeable, smectite-rich mudrocks in high-relief the steep slopes.
areas with semi-arid climates. Often the surface Figure 11.6 illustrates the development of bad-
exhibits narrow polygonal cracking or an irregular, land valley slopes through a combination of pip-
loose ‘popcorn’ texture with large intervening ing, collapse and gullying.
GFE-11.qxd 6/27/05 2:06 PM Page 293
SOIL EROSION 293
For tropical rainfall, Hudson (1965) developed a

revised equation:
127.5
KE ⫽ 29.8 ⫺
I
These equations show that at rainfall intensities

greater than 75 mm/hour, kinetic energy levels off
at around 29 J/m2 per millimetre of rainfall.
A number of indices have been developed to
express the relationship between kinetic energy
and soil loss, including EI30, the product of the
kinetic energy of a storm and the maximum 30
minute rainfall intensity (Wischmeier and Smith,
1958) and KE ⬎ 25 (Hudson, 1965). A worked
example is presented in Table 11.2. To estimate
soil loss from high intensity rainfall, a maximum
value of 28.3J/m2/mm for all rains above
76.2mm/hour and a maximum value of 63.5mm/
Figure 11.7a The relationship between particle size hour for I30 (the maximum 30-minute intensity)
and raindrop erosion.
can be used. For temperate latitudes, Morgan
(1980) has suggested that the Hudson KE ⬎ 25
index can be modified by using a lower threshold
value such as KE ⬎ 10.
11.3 Detachment: erosivity of rainfall Raindrop impact may, however, lead to soil
compaction and the development of a surface
The impact of raindrops is widely regarded to be crust a few millimetres thick. Surface crusts may
the most significant detachment agent, although limit the rate of detachment but, by reducing the
weathering, running water and wind can also infiltration rate, will also promote greater surface
loosen the soil so that it may be removed. Silts runoff and, hence, greater potential for particle
and fine sand particles tend to be the most prone transport elsewhere on a slope.
to detachment –– clay particles are more resistant
because of adhesive or chemical bonding (Figure
11.7a). 11.4 Transport: running water
The rate of detachment (Dr) varies with the
instantaneous kinetic energy of rainfall: Surface water flow (runoff ) occurs when the
rainfall intensity during a storm exceeds the rate of
Dr ⬀ I a S c
infitration (infiltration capacity) into the soil or
where I is the rainfall intensity (mm/hour) and S when the soil is saturated. Runoff can occur
is the slope (m/m). The exponents a and c are as shallow flows across a slope (sheet flow or
defined as follows: a ⫽ 2.0 ⫺ (0.01 ⫻ % clay) overland flow) or as channelled flows in rills,
and c is in the range 0.2–0.3, depending on the gullies and streams. Saturated through-flow can
grading. also be important in generating concentrated dis-
A general relationship between rainfall intensity charge adjacent to streams or gully-heads or where
(I) and kinetic energy of the rain (KE) (i.e. the ero- soils are thin or impermeable (Kirkby 1969;
sivity of rainfall) is (Wischmeier and Smith, 1958): Figures 11.8a and b).
The amount of runoff from a catchment can be
KE (J/m2per mm rainfall) ⫽ 11.87 ⫹ 8.73 log I estimated in a number of ways, depending on the
GFE-11.qxd 6/27/05 2:06 PM Page 294
Figure 11.7b The relationship between particle size, fluid and impact threshold velocities, characteristic modes of
aeolian transport, and resulting grading of aeolian sand formations (after Mabbutt, 1977).
data availability and the degree of precision run-off curve 90, the depth of run-off is
required. For example, the US Soil Conservation 72.5 mm.
Service Method was developed in the USA, but is 4 Determine the time of concentration (t ⫽
applicable in other settings: 0.02L0.77 S⫺0.385. L is the maximum length of
flow (m) S is the average stream gradient
1 Determine the design storm event: in this (m/m) for a steep gradient (0.1), 1000 m long
example 100 mm/hour. catchment; the time of concentration is 4.5
2 Determine the run-off curve number (Table hours.
11.3): for soil group C and paved roads the 5 Determine the unit peak discharge expected
run-off curve is 90. from the design storm (Table 11.4): for a time
3 Determine the run-off during the design of concentration of 4.5 hours, the unit dis-
storm (Figure 11.9): for 100 mm rainfall and charge is 0.058m3/s.
GFE-11.qxd 6/27/05 2:06 PM Page 295
SOIL EROSION 295
Table 11.2 Calculation of rainfall erosivity (from Morgan, 1995).
Time from Rainfall Intensity (I) Kinetic energy Total kinetic energy
start (min) (mm) (mm/hour) (J/m2 per mm rainfall) (KE) (Col. 2 ⫻ Col. 4) (J/m2)
0–14 1.52 6.08 8.83 13.42

15–29 14.22 56.88 27.56 391.90
30–44 26.16 104.64 28.58 747.65
45–59 31.50 126.00 28.79 906.89
60–74 8.38 33.52 26.00 217.88
75–89 0.25 1.00
127.5
Kinetic energy ( KE ) ⫽ 29.8 ⫺ (Hudson, 1965).
I
Wischmeier Index (EI30)
1. Maximum 30-minute rainfall ⫽ 26.6 ⫹ 31.50 mm
⫽ 57.66 mm
2. Maximum 30-minute intensity (I30) ⫽ 57.66 ⫻ 2
⫽ 115.32 mm/h
3. Total kinetic energy ⫽ total of col 5
⫽ 2277.74 J/m2
4. EI30 ⫽ 2277.74 ⫻ 115.32
⫽ 262 668.98 J mm/m2 hour
Hudson Index (KE ⬎ 25)
1. Total kinetic energy (rainfall ⬎ 25 mm/h)
⫽ Total Column 5, lines 2, 3, 4 and 5 only
⫽ 2264.32 J/m2
A specific nomograph has been developed for

identifying appropriate curve numbers for dry-
land conditions (Figure 11.10), taking account of
different plant covers. The key limitation of the
method is that the choice of curve number is crit-
ical, but remains largely subjective.
The velocity of flow is critical in initiating soil
erosion, as the flow must attain a threshold value
before entrainment of particles begins. Once this
critical condition is reached (it varies with particle
size: Figure 3.4), the entrainment rate is dependent
Figure 11.8a Mechanisms of delivery of rainfall to on the shear velocity of the flow and the discharge.
a stream channel from a hillslope: (1) overland flow (2) The greater the turbulence, the greater the ero-
through-flow (3) groundwater flow (from Freeze, 1978). sive power generated by the flow. The velocity (v)
of fully turbulent flow can be determined using
6 Convert the unit peak discharge to the actual Manning’s equation:
discharge.
Peak discharge ⫽ Unit discharge ⫻ catchment r 2 Ⲑ 3 S1Ⲑ 2
v⫽
area (km2) ⫻ run-off depth n
(mm)
⫽ 0.058 ⫻ 1 ⫻ 72.5 where r is the hydraulic radius, S is the slope (m/
⫽ 4.205 m3/s m) and n is the Manning roughness coefficient.
GFE-11.qxd 6/27/05 2:06 PM Page 296
Figure 11.8b A schematic runoff model (from Selby, 1982).
Tables A5.1 to A5.3 (Appendix A5) provide a 11.5 Transport: erosivity of wind
guide for Manning’s n for different plant covers
and stream bed types. For flows that are not fully Wind erosion commences when air pressure on a
turbulent (Reynold’s number, Re, ⬍ 2000; see loose, dry soil surface overcomes the force of
Chapter 3), the following equations can be used: gravity acting on the particles. Particles are
moved through the air by saltation (bouncing),
v ⬀ r1.7 S 0.95 Re ⫽ 250
surface creep or suspension within the flow.
v⬀r S 0.95 0.7
Re ⫽ 500 The relationship between soil particle size and the
v ⬀ r0.5 S 0.4 Re ⫽ 1000 wind velocity required for entrainment is shown
in Figure 11.7b; the fluid threshold is the velocity
The Reynold’s number (Re) is an index of the needed to initiate saltation, the impact threshold
turbulence of flow. is that needed to initiate entrainment by the
The transport capacity of flow (Tf) can be impact of saltating particles. However, most soil
determined using either empirical or theoretical surfaces comprise aggregates (collections of soil
sediment transport equations: particles held in a single structure such as a clod)
Tf ⬀ Q5/3 S5/3 (Meyer and Wischmeier, 1969) or crusts that limit the potential for wind erosion.
As wind erosion can selectively remove silt and
Tf ⫽ 0.0085 Q1.75 S1.625 D84–1.11 (Carson and
fine sand particles there can be a tendency for the
Kirkby, 1972)
remaining material (the lag deposit) to armour
where Q is the discharge or flow rate, S is the slope the ground surface, protecting it from further
gradient and D84 is the 84th percentile grain size. erosion unless or until it is disrupted.
GFE-11.qxd 6/27/05 2:06 PM Page 297
SOIL EROSION 297
Table 11.3 Runoff curve numbers for use in the Soil be expanded for the total wind erosivity (EW)
Conservation Service Method for estimating runoff over all vectors:
(for use in Figure 11.9).
15 n
Land Use Hydrologic Soil Group EWj ⫽ ∑ ∑ Vt 3ij fij
Condition (see below) j⫽0 i⫽1
A B C D
Range or Poor 68 79 86 89
where vectors j ⫽ 0 to 15 represent the principal
Pasture Fair 49 69 79 84 compass directions beginning with East: j ⫽ 0
Good 39 61 74 80 and working anticlockwise ENE ⫽ 1 etc.
Meadow Good 30 58 71 78 The eroded material is either sand sized or
Woodland Poor 45 66 77 83 dusts (⬍ 0.8 mm diameter). Whereas sand is
Fair 36 60 73 79 moved mainly in saltation close to the ground sur-
Good 25 55 70 77 face, dust can be carried in suspension, often at
Dirt Roads 72 84 87 89 great heights and for considerable distances.
Paved Roads 74 84 90 92 Figure 11.11 illustrates the concept of sediment
Hydrologic Conditions:
sources, transport pathways and permanent or
Poor: heavily grazed or plant cover ⬍ 50% or regularly temporary stores for both sand and dust move-
burned ment in a desert environment (Chapter 16).
Fair: moderately grazed or plant cover 50–75% or not The rate of soil movement (Q) can be des-
regularly burned cribed by:
Good: lightly grazed, plant cover ⬎ 75%.
Q (g/cm per second) ⫽ a (De)1/2 ␳/g (V)3
Soil Groups: where V is the drag velocity above the eroding
A: high infiltration capacity––sands, gravels, deeply surface; De is the average equivalent diameter of
weathered, well-drained
soil particles moved by the wind; ␳/g is the mass
B: moderate infiltration capacity, moderately to deeply
weathered, moderately to well-drained, moderately
density of air; a is a coefficient that varies with
fine to moderately coarse texture soil type (adapted from Chepil and Woodruff,
C: low infiltration capacity, moderately fine to fine tex- 1963). The quantity of material removed from a
ture, usually with a horizon that impedes drainage given area (X) is:
D: very low infiltration capacity, swelling clays, soils
with permanent high water tables, soils with clay X (tons/acre) ⫽ a (V )5
lenses, shallow soils over impervious materials.
11.6 Resistance to erosion: soil

A simple wind erosivity index is based on erodibility and plant cover
velocity and duration of the wind, rather than the
kinetic energy approach used for assessing rain- The soil erodibility is a key factor in control-
fall erosivity (Skidmore and Woodruff, 1968): ling the erosion process. A range of conditions
can be recognised, as illustrated in Figure 11.12.
n Materials may be completely undetachable and
EWj ⫽ ∑ Vt 3ij fij the rate of erosion will be weathering limited. In
i⫽1 other circumstances, the soil is readily detachable,
but the rate of erosion will be limited by the trans-
where EWj is the wind erosivity value for vector j, port capacity of the wind or water. The erodibility
Vt is the mean velocity in the ith speed group for of soil varies with the grading (soil texture), sta-
vector j above a threshold velocity (taken as bility of the aggregates and clods, shear strength,
19 km/hour) and fij is the duration of the wind for infiltration capacity and the organic/chemical
vector j in the ith speed group. The equation can content (Chapter 7).
GFE-11.qxd 6/27/05 2:06 PM Page 298
Figure 11.9 Nomograph for estimating runoff from rainfall. Runoff curve numbers represent hydrologic and soil
conditions as outlined in Table 11.3. (redrawn from US Soil Conservation Service, 1972).
Table 11.4 Unit Peak Discharges*
Time of Peak Time of Peak Time of Peak

Concentration Discharge Concentration Discharge Concentration Discharge
(hours) (m3/s) (hours) (m3/s) (hours) (m3/s)
0.1 0.337 1.0 0.158 4.0 0.063

0.2 0.300 1.5 0.120 5.0 0.054
0.3 0.271 2.0 0.100 10.0 0.034
0.4 0.246 2.5 0.086 20.0 0.021
0.5 0.226 3.0 0.076 24.0 0.019
*Discharge rate in m3/s from a run-off depth of 1 mm and a discharge area of 1 km2.
The size and density of particles above about Successively smaller sizes below 0.1 mm tend to
0.1 mm in diameter govern the initial resistance to require higher forces to displace and transport
displacement by wind or rain-splash erosion and them. For these reasons the soils most susceptible
their susceptibility to transportation in running to erosion are silts and fine sands. Other soils that
water. Coarser grained particles also form a soil are particularly susceptible to erosion are dry
with high porosity, which encourages infiltration organic soils, some residual soils such as those
so that in short duration storms runoff may be min- derived from volcanic ash, and lightweight fill
imised. However, if particles below this size exhibit materials such as pulverised fuel ash (PFA). This is
plasticity this provides inter-particle cohesion. mainly because the soil particles are of low density
GFE-11.qxd 6/27/05 2:07 PM Page 299
SOIL EROSION 299
by the heavy line in Figure 11.13, for a soil with

the following properties:
Proportion of sand and silt ⫽ 65%

Proportion of sand ⫽ 5%
Amount of organic matter ⫽ 2.8%
Structural class ⫽ 2
Permeability class ⫽ 4
From the nomograph, the resulting erodibility

index (K ) is 0.30.
Vegetation cover can reduce soil erosion by
providing a protective layer or buffer against rain-
splash (by absorbing raindrop energy), runoff (by
dissipating the energy of running water through
increased roughness) or wind erosion (by reducing
the shear velocity of wind by imparting rough-
ness). It also modifies the moisture content of the
soil and thus its shear strength. Mechanically,
vegetation increases the strength and competence
of the soil in which it is growing and therefore
contributes to its stability, for example it decreases
pore water pressure and increases soil suction
because of its own water requirement. Vegetation
also improves soil structure and porosity through
enrichment with organic material and protects the
soil from trampling by humans and animals.
Undisturbed forest is effective in controlling
erosion because the tree canopy intercepts rainfall
and reduces its energy. Drops from the canopy are
absorbed in the leaf litter and thence into a porous
Figure 11.10 Nomograph for estimating runoff curve
numbers in drylands for use in the US Soil
soil surface. Once the forest is disturbed by tree
Conservation Service method (after Jencsok, 1968). removal or grazing, the gaps in tree cover remove
the erosion protection.
Plants vary in their effectiveness, depending
on the height and continuity of the canopy, stage
due to their material content or form — fibrous, of growth and amount of bare earth exposed to
porous or hollow — as well as having particles erosion. For adequate protection around 70%
falling in the critical size range. cover is necessary, but reasonable protection can
The most widely used erodibility index is the be achieved with over 40% (Morgan, 1995).
K value that represents the soil loss per unit of Agricultural land is more susceptible to ero-
EI30 (as measured on a standard bare soil plot, sion under some crops than others because of
22 m long and on a 5o slope). Nomographs have different planting dates in respect of rainfall, dif-
been developed to assist the estimation of K ferent row spacing, growth rates and, therefore,
(Figure 11.13; Wischmeier et al., 1971; Landon, effective ground protection against rain-splash.
1984), although there may be problems when it is Certain management practices associated with
necessary to extrapolate beyond the nomograph particular crops may encourage erosion: in the
values. The procedure for estimating K is illustrated UK the frequency of vehicle traffic producing
GFE-11.qxd 6/27/05 2:07 PM Page 300
Figure 11.11 Sources and stores for sand and dust (from Jones et al., 1986).
Bare rock Coarse fragments on a Water-stable soil Crusted fine-grained Cohesionless

rocky substrate aggregates soil grained soil
Type of Slope Material
Increasing detactability
Weathering- limited Transport- limited
Figure 11.12 The detachability continuum for soils (from Parsons, 1988).
compaction is a problem on sugar beet and veg- The interaction between the various land man-
etable crops, as is the rolling of seedbeds on win- agement and physical factors results in risk peri-
ter cereals and maize. There are also frequently ods associated with different crops defined by the
occurring associations of crops and soils, for bare ground associated with the growing of a par-
example sugar beet with sandy soils, which influ- ticular crop and rainfall. Each of the six periods of
ence erosion rates. There is general agreement erosion risk identified by Boardman (1991) for
that the reported increase in significant lowland the UK is associated with a different land use or
erosion events in temperate areas is due to the management practice:
adoption of winter cereals and the consequent
expansion of the area left bare in autumn and 1. in late summer and early autumn in fields
winter (e.g. Evans and Cook, 1986). sown with oil seed rape or grass ley
GFE-11.qxd 6/27/05 2:07 PM Page 301
SOIL EROSION 301
Figure 11.13 Nomograph for estimating the erodibility factor K (from Wischmeier et al., 1971; Landon, 1984).
2. a short period in autumn when land has been hedgerows increases the erosivity of the wind by
cultivated before drilling winter cereals reducing the number of windbreaks.
3. land drilled with winter cereals the period of
risk depends on rainfall distribution, drilling
date and growth rate of the crop––in some 11.7 Erosion hazard assessment
years crop growth may be insufficient to
inhibit erosion until April, in others the period A range of approaches is available for assessing
of risk may extend only to December the erosion hazard within an area, ranging from
4. land ploughed and cultivated over winter and simple measures derived from readily available cli-
spring before sowing spring cereals matic data (these consider erosion hazard in terms
5. a short period of risk in spring on land drilling of the erosivity and not soil erodibility) to more
with spring cereals, although rapid crop detailed mapping-based methods that consider
growth can limit the risk period field evidence of erosion. Widely used methods
6. in May and June for land planted with late are detailed below.
spring crops such as maize. Erosivity indices: maps of erosivity can be
prepared using a rainfall index R(J/m2), from the
A similar situation can be identified in forestry rainfall energy E(J) and maximum 30-minute
practice, with the risk period corresponding to the rainfall intensity I30(mm/hr):
5–10 year interval between ploughing and the
establishment of an effective tree cover or at the R ⫽ EI30 /1000
end of clear-felling and timber extraction.
Soils tend to be at greatest risk of wind erosion Note that this produces an R value in metric units
when they are left bare with a fine, smooth sur- (Figure 11.14). Where rainfall intensity data is not
face in spring and early summer. Removal of available, erosivity can be estimated from the mean
GFE-11.qxd 6/27/05 2:07 PM Page 302
Figure 11.14 Mean annual values of the rainfall erosion index R (107 J/ha) (from Wischmeier and Smith, 1978).
annual rainfall. For example, in Malaysia, Morgan Ib (low relief, E ⫽ 27.12 (p2/P) ⫺ 475.4
(1974) defined the mean annual erosivity (EVa) as tropical,
follows: subtropical,
EVa ⫽ 9.28 P ⫺ 8838.15 semi-arid climate)
II (high relief, E ⫽ 52.49 (p2/P) ⫺ 513.21
Rainfall aggressiveness: Fournier’s (1960) p2/P humid climate)
index is a measure of the concentration of precipit- III (high relief, E ⫽ 91.78 (p2/P) ⫺ 737.62
ation (P) into a single month and, hence, is a meas- semi-arid climate)
ure of rainfall intensity (p ⫽ the highest mean
monthly precipitation, P ⫽ mean annual precipita- Sediment yield tends to be greatest in tropical
tion). The mean annual sediment yield Qs (g/m2) areas with seasonally concentrated rainfall
from a catchment was defined as: (Figure 11.15).
Land classification: a wide variety of app-
log Qs ⫽ 2.65 log p2/P roaches have been developed to describe soils or
⫹ 0.46 (log H) (tan S) ⫺ 1.56 land in terms of the limitations for agriculture or
other land uses. Many of these methods use the
where H is the mean altitude (m) and S the mean
degree of erosion as a factor in the classification
slope (degrees). An alternative approach is to
system (e.g. Table 11.5; Figure 11.16).
identify which relief and climate type the drainage
Soil erosion survey: field mapping or aerial
basin falls into and calculate p2/P, as follows:
photograph interpretation of evidence of soil ero-
Ia (low relief, E ⫽ 6.14 (p2/P) ⫺ 49.78 sion activity (e.g. sheet wash, rills, gullies) can
temperate climate) provide a measure of erosion hazard across
GFE-11.qxd 6/27/05 2:07 PM Page 303
SOIL EROSION 303
Figure 11.15 World distribution of suspended sediment yield (from Fournier, 1960).
Table 11.5 Land capability classes (US system, modified from Stallings, 1957; Morgan, 1995).
Class Characteristics
I Deep, productive soils easily worked, on nearly level ground; not subject to overland flow; no or slight
risk of damage when cultivated; use of fertilizers and lime, cover crops, crop rotations required to
maintain soil fertility and soil structure.
II Productive soils on gentle slopes; moderate depth; subject to occasional overland flow; may require
drainage; moderate risk of damage when cultivated; use of crop rotations, water-control systems or
special tillage to control erosion.
III Soils of moderate fertility on moderately steep slopes, subject to more severe erosion; subject to severe
risk of damage but can be used for crops provided plant cover is maintained; hay or other sod crops
should be grown instead of row crops.
IV Good soils on steep slopes, subject to severe erosion; very severe risk of damage but may be cultivated
if handled with great care; keep in hay or pasture but a grain crop may be grown once in 5 or 6 years.
V Land too wet or stony for cultivation but of nearly level slope; subject to only slight erosion if properly
managed; should be used for pasture or forestry but grazing should be regulated to prevent plant cover
from being destroyed.
VI Shallow soils on steep slopes; use for grazing and forestry; grazing should be regulated to preserve
plant cover; if plant cover is destroyed, use should be restricted until cover is re-established.
VII Steep, rough, eroded land with shallow soils; also includes droughty or swampy land; severe risk of
damage even when used for pasture or forestry; strict grazing or forest management must be applied.
VIII Very rough land; not suitable for woodland or grazing; reserve for wildlife, recreation or watershed
conservation.
GFE-11.qxd 6/27/05 2:07 PM Page 304
Figure 11.16 Urban capability classification for soil erosion control, Australia (from Hannam and Hicks, 1980).
different terrain units. A simple coding system for which uses Lettau’s formula:
assessing the severity of erosion is included as
Q(V ⫺ ␷t)V 2 ⭈ t
Table 11.6.
Potential sand drift calculation: this can be where Q is the proportional amount of sand drift,
calculated using Fryberger’s (1979) method V is the average wind velocity at 10 m, ␷t is the
GFE-11.qxd 6/27/05 2:07 PM Page 305
SOIL EROSION 305
Table 11.6 Coding system for soil erosion appraisal (from Morgan, 1995).
Code Indicators
0 No exposure of tree roots; no surface crusting; no splash pedestals; over 70% plant cover
(ground and canopy).
0.5 Slight exposure of tree roots; slight crusting of the surface; no splash pedestals; soil level slightly higher
on upslope or windward sides of plants and boulders; 30–70% plant cover.
1 Exposure of tree roots, formation of splash pedestals, soil mounds protected by vegetation, all to depths
of 1–10 mm; slight surface crusting; 30–70% plant cover.
2 Tree root exposure, splash pedestals and soil mounds to depths of 1–5 cm; crusting of the surface;
30–70% plant cover.
3 Tree root exposure, splash pedestals and soils mounds to depths of 5–10 cm; 2–5 mm thickness of
surface crust; grass muddied by wash and turned downslope; splays of coarse material due to wash
and wind; less than 30% cover.
4 Tree root exposure, splash pedestals and soil mounds to depths of 5–10 cm; splays of coarse material;
rills up to 8 cm deep; bare soil.
5 Gullies; rills over 8 cm deep; blow-outs and dunes; bare soil.
Figure 11.17a Sand drift potential (redrawn from

Jones et al., 1986).
Figure 11.17b Sand rose and dust rose for airport site
in Saudi Arabia, showing percentage of mean annual
drift potential of winds from 16 sectors (redrawn from
Jones et al., 1986).
impact threshold wind velocity (e.g. 12 knots), way by Jones et al. (1986), using Fryberger’s
and t is the duration of wind). Figure 11.17a tables. Sand rose diagrams can be developed to
shows the distribution of sand drift potential by illustrate the dominance of particular wind direc-
month for a site in Saudi Arabia, developed in this tions (Figure 11.17b). It should be noted that the
GFE-11.qxd 6/27/05 2:07 PM Page 306
Figure 11.18a Dust hazard assessment, Saudi Figure 11.18b Dust hazard assessment, Saudi
Arabia: dust source significance map (redrawn from Arabia: potential dust source areas (redrawn from Jones
Jones et al., 1986). et al., 1986).
approach estimates potential rather than actual

sand drift. The potential drift rates can only
be realised if there is an unlimited supply of
sand––actual drift rates can be considerably
lower. Geomorphological mapping should be
used to identify sediment source areas and trans-
port pathways.
Dust hazard assessment: Jones et al. (1986)
describe a method for assessing dust hazard
around a town in Saudi Arabia. The method relates
a dust transport rose (Figure 11.17b) to the outline
of the urban area, producing overlapping rays of
differing width. The values for each segment of
the dust rose were then applied to the respective
rays and the values summed for those areas where
the rays overlap (Figure 11.18a). The higher the
value the greater the likelihood that dust will be
blown into the town. A geomorphological map of
potential dust sources (Figure 11.18b) and the Figure 11.18c Dust hazard assessment, Saudi Arabia:
dust hazard map, prepared by multiplying the values on
dust drift assessment (Figure 11.18a) were com- the dust source significance map by 1, 2 or 3 depending
bined to produce a hazard score for the area on whether areas were estimated to be of ‘low’, ‘medium’
around the town (Figure 11.18c). or ‘high’ drift potential (redrawn from Jones et al., 1986).
GFE-11.qxd 6/27/05 2:07 PM Page 307
SOIL EROSION 307
11.8 The Universal Soil Loss C is the crop management factor and repre-
Equation sents the ratio of soil loss under a given crop to
that from bare soil (see Table A5.4, Appendix A5).
The most important equation for predicting gen- P is the erosion control factor. With no erosion
eral soil loss from rain-splash and runoff is control in place P ⫽ 1.0 (see Table A5.5,
the Universal Soil Loss Equation (USLE) Appendix A5).
(Wischmeier and Smith, 1978): Table A5.6 (Appendix A5) provides a worked
example of the use of the USLE for predicting
E ⫽ R ⫻ K ⫻ LS ⫻ C ⫻ P mean annual soil loss (i.e. long-term erosion
rates and not loss associated with single storms).
where E ⫽ average annual soil loss (t/ha) and R is It is a widely used tool for soil conservation
the rainfall erosivity factor based on the mean planning, but caution is needed when applying it
EI30 (see above). For R to be in metric units: outside its research base, i.e. eastern and central
USA, on slopes less than 7o.
R ⫽ EI30 /1000 For a summary of other soil loss prediction
models see Morgan (1995).
K is the soil erodibility index, i.e. the soil loss per
unit of EI30 (as measured on a standard bare soil
plot, 22 m long and on a 5o slope; see Figure 11.9 Erosion control methods and
11.13). LS is the combined slope length (L) and
materials
slope steepness (S) factor and can be derived
from Figure 11.19 or the following equation: A variety of methods have been used to control
surface erosion in different settings, with varying
LS ⫽ (x/22.13)n (0.065 ⫹ 0.045s ⫹ 0.0065s2)
degrees of success. Although many of the meth-
where x is the slope length (m); s is the slope gradi- ods originated in soil conservation practice for
ent in percent, the exponent n varies with slope agriculture and forestry projects, they have
steepness: for 3o slopes, n ⫽ 0.4; for 2o slopes, become increasingly used in civil engineering,
n ⫽ 0.2 for; ⬍ 1o slopes, n ⫽ 0.1; for slopes, ⬎6o especially as a way of reducing visual impact and
n ⫽ 0.6. enhancing the environment.
The following sections can only draw attention
to the more successful methods; more details can
be found in a range of manuals and guidebooks,
including FAO (1996), Schiechtl and Stern (1996,
1997), Gray and Leiser (1982), Coppin and
Richards (1990), Barker (1995) and Hudson
(1995). It should be appreciated that each soil ero-
sion problem will tend to be unique, because of
the great range of slope forms and processes and
the inherent variability of the soils and bedrock
materials. Solutions, therefore, need to reflect site
conditions and cannot be provided ‘off-the-shelf’.
An essential starting point is to understand
the erosion processes operating at the site.
Sympathetic design and construction; an under-
standing of the relative risks, a mechanism for
observation and monitoring of the development,
Figure 11.19 Relationship of the USLE slope factor and a plan for future maintenance and mitigation
(LS) to gradient and length of slope (from Hudson, 1995). of problems are all necessary.
GFE-11.qxd 6/27/05 2:07 PM Page 308
Wind erosion widely used methods include the use of wind-

Approaches to mitigating wind erosion either breaks and land management practice.
aim to reduce the force of the wind or to improve Windbreaks: placing a barrier across the path of
the ground surface characteristics so that parti- the wind reduces velocity at the ground surface
cle movement is restricted. There are four gen- both in front of and behind the barrier, and reduces
eral methods of approach (Figure 11.20): (1) the field length. Some measured reductions for
establish and maintain vegetation and organic average tree shelter belts are provided in Table
residues (2) produce or bring to the surface non- 11.7. Barriers may be relatively permanent live
erodible aggregates or clods (3) reduce field vegetation structures (hedges or lines of trees) or
width (exposure) along the prevailing wind- they may be constructed of artificial materials
erosion direction or (4) roughen the land surface. such as geotextiles, stakes or palm fronds.
All of these methods are encompassed in good Windbreaks should be set as closely as pos-
land husbandry practice (FAO, 1996). The more sible at right angles to the dominant wind erosion
Figure 11.20 Approaches to management of wind erosion.

GFE-11.qxd 6/27/05 2:07 PM Page 309
SOIL EROSION 309
force. The degree of protection is determined by Land management practice: protecting the
the spacing between barriers and by their width, surface from attack and trapping moving particles
height and porosity. In approximate terms wind can be achieved by keeping the surface covered
velocity is reduced to about five to ten times the throughout the year. Planting ‘cover’ crops to pro-
windbreak height on the windward side and about tect the surface in windy seasons when they occur
ten to thirty times the windbreak height on the outside the main crop growing period is an effec-
leeward side. Other factors that should be consid- tive and cheap method which may produce another
ered are that the ends of barriers tend to cause useful crop or provide an effective green manure
funnelling and local increases in velocity and or mulch. Crops of differing type can be mixed so
therefore fewer longer barriers are preferable to a that the differing heights, or rates of germination
greater number of shorter ones. Barriers that are and growth, increase surface roughness or provide
semi-permeable are also preferable to those pro- strips of vegetation that protect intervening strips
viding a complete obstacle to the wind, which can of still-bare soil. Table 11.8 illustrates typical widths
cause eddying, turbulence and local increases in of vegetated strip required for different soil types and
velocity. It should also be remembered that in wind direction.
periods when wind speed is particularly high even The management of crop residue and stubble
reductions effected by windbreaks might not be can also be significant, since these also trap moving
sufficient to prevent particle transport. particles, provide a rough surface and contribute
organic matter to the soil. Again relationships exist
Table 11.7 Effect of barriers in reducing wind velocity between stubble height, width of the stubble strip
(after FAO, 1996). and the type of stubble. Ploughing creates a rough
Percentage reduction Distance from barrier surface and can contribute to preventing soil
in velocity (multiples of height) erosion particularly if the ridges and furrows are
created at right angles to the prevailing winds. Care
60–80 0 is needed to ensure that the choice of equipment is
20 20 suited to the soil type, particularly if erosion
0 30–40 prevention is of major concern.
Table 11.8 Strip dimensions for the control of wind erosion (from Chepil and Woodruff, 1963).
Soil class Width (m) of strips
Wind at right Wind deviating 20° Wind deviating 45°

angles from a right angle from a right angle
Sand 6.1 5.5 4.3

Loamy sand 7.6 6.7 5.5
Granulated clay 24.4 22.9 16.5
Sand loam 30.5 28.0 21.3
Silty clay 45.7 42.7 33.5
Loam 76.2 71.6 51.8
Silt loam 85.4 79.3 57.9
Clay loam 106.7 99.1 76.2
Note: the table shows average width of strips required to control wind erosion equally on
different soil classes and for different wind directions, for conditions of negligible surface rough-
ness, average soil cloddiness, no crop residue, 300 mm high erosion resistant stubble to wind-
ward, 64.4 km/h wind at 15.24 m height and a tolerable maximum rate of soil flow of 203.2 kg
per 5 m width per hour.
GFE-11.qxd 6/27/05 2:07 PM Page 310
Conditioning the soil by increasing its cohesion Table 11.9 A guide to contour spacing on sloping
with the addition of organic matter, mulching to ground.
retain its moisture or even irrigating to keep the
Slope angle Contour spacing (m)
surface moist all help to resist erosion. Moisture
retention may merely involve a change in the tim- Per cent Degrees
ing of ploughing in relation to seeding. A relatively ⬍6 ⬍4 100
new technique is the conditioning of soil by the 8 5 60
spraying of artificial additives. 10 6 30
12 7 25
⬎ 12 ⬎7 20
Rain and sheet erosion
Good land management is also the key to control-
ling the loss of soil from rainfall and sheet flow.
When land is under active production careful con- velocities increase rapidly over short distances
sideration needs to be given to land use, crop man- and the ridges can be easily breached. A solution
agement, tillage methods and the application of is to use the ridges as a drainage control by slop-
manures and fertilisers. If the land is fallow then ing them obliquely down the slope at a very shal-
the establishment, re-establishment or maintenance low angle to encourage water behind them to flow
of vegetation cover is important. Alternatively, across the slope to a collection and distribution
reductions in runoff velocity can be achieved by system.
dividing land into small plots or benching to reduce The effectiveness of this method can be
slope steepness and the introduction drainage extended by using the ridges in conjunction with
ditches and sediment traps can conserve soil cover. a drainage channel, or by using a geotextile separ-
In addition specific measures may be necessary to ator to prevent the soil from being carried into the
address particular problem areas. Such measures drain. Thus the soil is preserved while the water is
may include contouring, strip cropping, terracing, drained away through the system.
and the construction of drainage measures or struc-
tures. The control of runoff and its effects on Gully erosion
nearby watercourses is now an important aspect of Methods to protect the gully-head from further
major engineering schemes. erosion involve either the reduction of the volume
In contour farming, rows are orientated and velocity of flow into the gully or the direct
across the slope and thus act as a barrier to the protection of the gully-head from erosion due to
down-slope flow of water. It is most effective on excessive flow. If the gully is in a state of active
shallow slopes. It becomes difficult to operate development then the source of the flow and its
machinery on steeper slopes, because it needs to velocity should be calculated to determine if
work across the slope to create ruts that act as reductions are achievable. Reducing the volume
small dams. Contour farming reduces runoff of flow may be achievable by modifying farming
and, therefore soil erosion. Generally, as the practices on the slopes above the gully using the
slope becomes steeper the contour strips need to techniques described earlier. If water flow has
be closer together (Table 11.9). recently been diverted, for example during the
Contour ridging and ridge drains are used construction of a new road without attention to
to produce specific ridge features, rather than accommodating the pre-existing drainage regime,
relying on the cross-slope texture produced by the preferable step is to establish a suitable
contour farming, which significantly improves drainage system to accommodate the additional
the ability of the system to reduce flow velocities. flow.
The ridges are simple water control structures Artificial methods may be required to protect
that act to dam the flow and provide a temporary the gully-head and the measures adopted depend
storage until infiltration can occur. They are less on the size and slope of the gully and on the typi-
effective as slopes become steeper because flow cal maximum flows (Figure 11.21). If the duration
GFE-11.qxd 6/27/05 2:07 PM Page 311
SOIL EROSION 311
Figure 11.21 Methods for protection of the gully head.
of a potential event can be estimated and the chan- For low flow regimes it may be possible to
nel geometry is known then flow velocity can be check erosion by establishing vegetation. Grasses
calculated from standard open-channel hydraulic and legumes are effective in providing soil bind-
relationships and an appropriate structure ing, and bamboo with its hardy stems and foliage
selected. is effective in diffusing flow. A distinct gully-head
GFE-11.qxd 6/27/05 2:07 PM Page 312
feature points to the fact that at least moderate

flow has occurred in the past.
In moderate flow regimes a simple structure
may be needed to provide an erosion resistant
gully-head. Brushwood bundles can be laid and
pegged in the gully-head. Alternatively a rubble
bank can be constructed in the gully head, using
large stones. These stones must be of sufficient
individual size to resist potential detachment and
transportation during peak flow. The main prin-
ciples to follow are that the flow of water should not
be impeded by the structure, otherwise flow will
be diverted around or behind and under the struc-
ture. Ideally, the structure should also help to
dissipate the energy of the water.
In areas of very high flows gabion structures
may be necessary. Masonry structures are not rec-
ommended because they are impermeable and
resist flow and the mortar inevitably disintegrates
after a few years. They are also rigid and crack as
erosion develops in front of the apron. Gabions
are highly permeable and tend to break up and
dissipate the flow and they are also flexible. They
should have a long aprons so that the energy is
dispersed along the length of the structure; gabion
aprons deform to accommodate the erosion at
their toe. Figure 11.22 Grass components in gully scour
Any structure should mimic the slope profile protection.
at the head of the gully so that flow continues
unimpeded onto the structure.
Prevention of channel scour can be achieved in
low flow regimes by a grass lining to channels
(Figure 11.22). The sward reduces the flow
velocity at the soil surface by interfering with the
flow, and when it is deflected under higher veloc-
ity flow it provides a protective cover to the soil.
The litter layer also provides protection to the soil
surface and the roots provide mechanical stability
to the soil particles, anchoring the soil into the
underlying subsoil. The use of geomeshes or
geomats can further improve the stability of
grass-lined channels. Indications of the scale of
improvement are illustrated in Figure 11.23,
which shows the limiting velocities that can be
withstood by various grass or reinforced grass
covers. Figure 11.23 Limiting velocities for various grass
and reinforced grass covers.
In moderate flow regimes live branches reduce
erosion by initially providing a vegetative cover
GFE-11.qxd 6/27/05 2:07 PM Page 313
SOIL EROSION 313
Figure 11.24 Gully protection using live branches.
over the gully floor, which reduces velocity. As is laid, staggered down the gully and covered in
root development takes place this provides a bind- turn by a soil layer. The process is repeated until
ing to the gully floor and sides which continues the required area is covered. Initially, the live
the protection even during dormancy. A layer of branches must be held in place until the roots
branches is laid in a herringbone pattern over develop sufficiently to provide resistance to
the gully floor and extending to the gully sides flow. Cross-poles can be used at approximately
(Figure 11.24). The branches are then covered by 2 m intervals. They are placed over the live
a soil layer, ensuring that the tips of the branches branches and embedded into the gully sides to at
are left uncovered. A further layer of branches least 0.5 m.
GFE-11.qxd 6/27/05 2:07 PM Page 314
Figure 11.25 Methods of gully scour protection.
Check dams are constructed along the length of eventually overtops. The water drops its sediment
gullies in order to decrease the gradient of the gully load and the sediment accumulates until it reaches
floor (Figure 11.25). Check dams slow the water the top of the check dam. The result is a shallower
flow because they create a small reservoir that gradient along the length of the gully over which
GFE-11.qxd 6/27/05 2:07 PM Page 315
SOIL EROSION 315
Figure 11.26 Dimensioning and spacing of check dams.
the check dams have been constructed. If a greater

separation is employed sediment will not accumu-
late to the necessary extent and erosion will work
back to undermine the next dam upstream.
Eventually successive dams will be undermined
until the gully-head protective works are destroyed.
Check dams can be made with vegetation,
rock-fill, timber, dry-stone masonry and gabions.
As they hinder water flow, extreme care is needed
in their design to ensure that they do not cause
such an obstruction as to promote increased
erosion of the side banks, or cause the gully flow
to divert around the check dam.
There are two main rules for the siting of
check dams: Figure 11.27 Orientation of check dam structures.
Check dams must be perpendicualr to gully banks.
1. The top of each dam should be at or just
below the base of the next dam up-gully. The
maximum gradient between the top of the
through which they have been constructed, the
dam and the base of the next dam up-gully
velocity will increase further downstream and
should be 3% (Figure 11.26).
may cause enhanced erosion in that area. Ideally,
2. Dams should be positioned so that they are
the natural gully gradient below the lowest check
perpendicular to the flow (Figure 11.27) — if
dam should be equal to or less than the gradient
they are not, they divert the flow to one side
between the top of the lowest check dam and the
of the gully and cause erosion in the adjacent
base of the next check dam up-gully. If this is
bank.
not the case, erosion will occur immediately
It should be remembered that as check dams below the lowest check dam and eventually
effectively decrease the velocity over the length undermine it.
GFE-11.qxd 6/27/05 2:07 PM Page 316
11.10 Summary Coppin, N. J. and Richards, I. G. (1990) Use of

Vegetation in Civil Engineering. CIRIA (Construc-
tion Industry Research and Information Association)
Soil erosion by water or wind can cause a range of
Report, Butterworth, London.
problems for engineers and environmental man- Dalrymple, J. B., Blong, R. J. and Conacher, A. J.
agers, from loss of productive soils to blocked infra- (1968) A hypothetical nine-unit landsurface model.
structure routes, loss of reservoir storage capacity Zeitschrift für Geomorphologie 12, 60–76.
and environmental degradation. Gullying and the Drakeford, T. (1979) Report of Survey of the Afforested
Spawning Grounds of the Fleet Catchment. Forestry
development of badland terrain can develop over
Commission.
engineering timescales and, hence, can present a Evans, R. and Cook, S. (1986) Soil erosion in Britain.
significant problem for some projects. Often ser- SEESOIL 3, 28–58.
ious problems arise because of a combination of FAO (Food and Agriculture Organization) (1996)
changes to land use or management practice and the Methods and Materials in Soil Conservation. FAO
Soils Bulletin 70, FAO, Rome.
occurrence of low–medium frequency triggering
Fournier, F. (1960) Climat et Érosion: La Relation Entre
events––heavy rainfall or high wind speeds. Once L’érosion du Sol par L’eau et Les Précipitations
the nature and cause of the problems have been Atmosphériques. Presses Universitaines de France,
established, simple but effective measures can be Paris.
implemented to reduce or control erosion. Freeze, R. A. (1978) Mathematical modelling of hill-
slope hydrology. In Kirkby, M. J. (ed.) Hillslope
Hydrology. Wiley, Chichester, 177–225.
References Fryberger, S. G. (1979) Dune forms and wind regimes.
In McKee, E. D. (ed.) A Study of Global Sand Seas.
Barker, D. H. (ed.) (1995) Vegetation and Slopes: US Geological Survey Professional Paper 1052,
Stabilisation, Protection and Ecology. Thomas 137–169.
Telford, London. Gray, D. H. and Leiser, A. J. (1982) Biotechnical Slope
Boardman, J. (1990) Soil erosion on the South Downs: Protection and Erosion Control. Van Nostrand
a review. In Boardman, J., Foster, I. D. L. and Reinhold, New York.
Dearing, J.A. (eds) Soil Erosion on Agricultural Hadley, R. F. and Schumm, S. A. (1961) Sediment
Land. John Wiley and Sons, Chichester, 87–105. sources and drainage basin characteristics in upper
Boardman, J. (1991) Land use, rainfall and erosion risk Cheyenne River basin. US Geological Survey Water
on the South Downs. Soil Use and Management 7, Supply Paper 1531B, 137–198.
34–38. Hannam, I. D. and Hicks, R. W. (1980) Soil conserva-
Bruington, A. E. (1982) Fire-loosened sediment mention and urban land use planning. Journal of the Soil
aces the city. In Proceedings of the Symposium Conservation Service NSW 36, 134–145.
on Dynamics and Management of Mediterranean- Howard, A. (1994) Badlands. In Abrahams, A. A. and
type ecosystems, June 22–26, 1981, San Diego, Parsons, A. J. (eds), Geomorphology of Desert
California. USDAFS General Technical Report Environments. Chapman and Hall, London, 213–
PSW-58, 40–422. 242.
Campbell, I. A. (1981) Spatial and temporal variations in Hudson, N. W. (1965) The Influence of Rainfall on the
erosion measurements. In Symposium on Erosion Mechanics of Soil Erosion with Particular
and Sediment Transport Measurements. Proceedings Reference to Southern Rhodesia. Unpublished MSc.
of the Florence Symposium International Association thesis, University of Cape Town.
of Hydrological Sciences 133, 447–456. Hudson, N. W. (1995) Soil Conservation. Batsford,
Campbell, I. A. (1997) Badlands and badland gullies. London.
In Thomas, D. S. (ed.) Arid Zone Geomorphology: Jencsok, E. I. (1968) Hydrologic Design for Highway
Process, Form and Change in Drylands. Wiley, Drainage in Arizona. Arizona Highway Department,
Chichester, 261–291. Phoenix.
Carson, M. A. and Kirkby, M. J. (1972) Hillslope Form Jones, D. K. C., Cooke, R. U. and Warren, A. (1986)
and Process. Cambridge University Press Geomorphological investigation, for engineering
Chen, Y. (1983) A preliminary analysis of the processes of purposes, of blowing sand and dust hazard.
sediment production in a small catchment on the Loess Quarterly Journal of Engineering Geology 19,
Plateau. Geographical Research (China) 2, 35–47. 251–270.
Chepil, W. S. and Woodruff, N. P. (1963) The physics of Jones, J. A. A. (1981) The Nature of Soil Piping — a
wind erosion and its control. Advances in Agronomy Review of Research. British Geomorphological
15, 211–302. Research Group Monograph, Geo Books, Norwich.
GFE-11.qxd 6/27/05 2:07 PM Page 317
SOIL EROSION 317
Kirkby, M. J. (1969) Infiltration, throughflow and over- Selby, M. J. (1987) Slopes and weathering. In Gregory,
land flow: and erosion by water on hillslopes. In K. J. and Walling D. E. (eds), Human Activity
Chorley, R. J. (ed.) Water Earth and Man, Metheun, and Environmental Processes. Wiley, Chichester,
London, 215–238. 183–205.
Landon, J. R. (ed.) (1984) Booker Tropical Soil Manual. Skidmore, E. L. and Woodruff, N. P. (1968) Wind ero-
Longman, Harlow. sion forces in the United States and their use in pre-
Leopold, L. B., Wolman, M. G. and Miller, J. P. (1964). dicting soil loss. USDA Agricultural Research
Fluvial Processes in Geomorphology. Freeman, Service Handbook 346.
San Francisco. Stallings, J. H. (1957) Soil Conservation. Prentice Hall
Mabbutt, J. A. (1977) Desert Landforms. MIT Press, Englewood Cliffs NJ.
Cambridge, Mass. Thornes, J. B. (1994) Channel processes and forms. In
Meyer, L. D. and Wischmeier, W. H. (1969) Abrahams, A. A Parsons A. J. (eds), Geomorphology
Mathematical simulation of the process of soil of Desert Environments. Chapman and Hall,
erosion by water. Trans. Am. Soc. Agr. Engnr. 12, London, 288–318.
754–758, 762. US Environmental Protection Agency (1973) Methods
Morgan, R. P. C. (1974) Estimating regional variations for Identifying and Evaluating the Nature and
in soil erosion hazard in Peninsular Malaysia. Extent of Nonpoint Sources of Pollutants. US
Malaysia Nat. J. 28, 94–106. Department of Agriculture, Washington DC.
Morgan, R. P. C. (1980) Soil Erosion. Longman, London. US Soil Conservation Service (1972) National
Morgan, R. P. C. (1995) Soil Erosion and Conservation. Engineering Handbook, Section 4: Hydrology. US
Longman, London. Department of Agriculture, Washington DC.
Parsons, A. J. (1988) Hillslope Form. Routledge, London Wischmeier, W. H. and Smith, D. D. (1958) Rainfall
and New York. energy and its relationship to soil loss. Trans. Am.
Schiechtl, H. M. and Stern, R. (1996) (English transla- Geophys. Un. 39, 285–291.
tion) Ground Bioengineering Techniques for Slope Wischmeier, W. H., and Smith, D. D. (1978) Predicting
Protection and Erosion Control. Blackwell Science, rainfall erosion losses. USDA Agricultural Research
Oxford. Service Handbook 537.
Schiechtl, H. M. and Stern, R. (1997) (English trans- Wischmeier,W. H., Johnson, C. B. and Cross, B. V. (1971)
lation) Water Bioengineering Techniques for A soil erodibility nomograph for farmland and con-
Watercourse Bank and Shoreline Protection. struction sites. J. Soil and Water Conservation 26,
Blackwell Science, Oxford. 189–193.
Schumm, S. A. (1956) The role of creep and rainwash Wolman, M. G. and Schick, A. P. (1967) Effects of con-
on the retreat of badland slopes. American Journal struction on fluvial sediment, urban and suburban
of Science 254, 693–706. areas of Maryland. Water Resources Research 3,
Selby, M. J. (1982) Hillslope Materials and Processes. 451–464.
Oxford University Press. Worster, D. (1979) The Dust Bowl. OUP, New York.
GFE-12.qxd 6/28/05 3:43 PM Page 318
12. Subsidence
Tony Waltham
12.1 Subsidence environments open cavities create subsidence that extends

over only small areas, but it may involve cata-
Downward movement of the ground, or ground strophic collapse within those areas.
subsidence, can only occur where the subsurface 3. Tectonic subsidence, where plastic rock moves
conditions are such that space exists for the ground away at great depth from beneath a site.
to move into. Most of the Earth’s surface is on Crustal deformation can cause regional subsi-
stable rock, where subsidence cannot occur. But dence (12.11 Tectonic subsidence), and magma
there are specific environments where subsidence movement can cause volcanic deflation of
is a significant geohazard, and these fall into four smaller areas.
distinct categories that may be summarised as: 4. Landslides, where the head zone subsides and
the toe moves outwards. This geohazard is a
1. Porous and deformable rocks and soils that are function of slope profiles (Chapter 23) and is
therefore compressible. Compaction and con- not true ground subsidence.
sequent surface subsidence occur as the rock or
soil restructures with declining pore space, nor- It is possible for a single site to have more
mally accompanied by abstraction or expulsion than one subsidence hazard. A thick soil over a
of interstitial groundwater. The main material karst limestone may develop both compaction
involved is clay (12.2 Subsidence on clay), but subsidence and sinkhole events, and piping fail-
the process can apply in peat (12.3 Subsidence ures may further complicate an understanding of
on peat) and silt (12.4 Hydrocompaction of col- the processes, while mining collapse and tec-
lapsing soils), as well as in most types of artifi- tonic movements can be superimposed on any
cial landfill and made ground, in permafrost geomorphological environment. However, a con-
when ground ice is melted (12.5 Subsidence on structive engineering approach to remediation of
permafrost), and in some sands when vibrated any subsiding ground is best achieved by model-
by earthquakes (12.11 Tectonic subsidence). ling the primary processes — which must fall
Subsidence may be local, due to structural into one or more of the above four categories.
loading, or may be regional, due to changes in When regional subsidences are separated out
groundwater conditions. from localised ground failures, each of the latter
2. Rocks that contain large cavities, into which can occur in only a few geological environments.
ground can fall, when rock forming the roof col- It is also significant that most subsidence is
lapses or deforms. Natural caves occur mainly caused, induced or exacerbated by man’s own
in limestone or gypsum where subsidence is activities; very little ground subsidence is
therefore a frequent geohazard (Chapter 24) and entirely natural.
in basalt lava flows (Chapter 23). Subsidence
does occur on salt, where few caves survive to
maturity (12.7 Salt subsidence), but natural 12.2 Subsidence on clay
caves are not significant in other rocks. Mined
cavities can be left in almost any rock type The special properties of clay minerals, particu-
(12.10 Subsidence due to mining). All forms of larly their water retention or expulsion and their
318
GFE-12.qxd 6/28/05 3:43 PM Page 319
SUBSIDENCE 319
low strength, account for the singular plastic are well known (Waltham, 2002). Clays gain a
deformation properties of the clay soils and part of their internal support (and therefore their
rocks, which are therefore the most widespread load-bearing capacity) from their pore water
cause of destructive ground subsidence. These pressure. If this pressure is reduced, as water is
clay materials include those like the Paleogene removed, the clays compact. The natural process
London Clay that are viewed as rocks by a geol- is for water to be squeezed out during very slow
ogist (because they are within old lithified consolidation and lithification. Artificial removal
sequences) and are treated as soils by an engi- of the water, by abstraction pumping, induces far
neer (because they can be excavated without more rapid compaction. The subsidence hazard
pre-breaking). Their plasticity is derived from lies in alluvial sequences with alternating beds of
the water that is loosely bonded to the chemi- poorly consolidated sand and clay beneath large
cally complex clay mineral particles (Chapter 3). cities. Cheap and convenient water supplies are
They may therefore compact (and cause subsi- pumped from the productive sand aquifers, but
dence) when the water is squeezed out under these are almost incompressible, so negligible
load or in response to water abstraction. In subsidence is induced. But the inevitable con-
either case, the loss of water causes consolida- sequence is equilibrium of pore water pressures
tion, as increased grain-to-grain contact creates in the sands and the clays, when the clays then
increased strength. Subsidence is the precursor compact.
of consolidation, and may be either regional or The amount of subsidence is directly propor-
localised. tional to the groundwater head decline, but is also
a function of the age and mineralogy of the clay.
Clay subsidence due to fluid For equal head decline and clay thickness, greater
abstraction subsidence occurs on the younger clays (which
Worldwide, this ranks as the most widespread and are therefore less consolidated by self-weight)
most destructive subsidence process. Entire cities and on those with higher contents of the unstable
have subsided; Venice (Figure 12.1), Shanghai, smectite. The clay minerals kaolinite and illite are
Mexico City and Bangkok are just examples that both more stable with far less removable water
Figure 12.1 Boardwalks over one of the piazzas of Venice, which have now subsided below the level of most winter
high tides.
GFE-12.qxd 6/28/05 3:43 PM Page 320
Table 12.1 Comparison of subsidence parameters for seven sites on clay.

Site Subsided Maximum Clay Head Specific Compressibility Smectite Age
area (km2) subsidence thickness decline subsidence (⫻10⫺3m/m/m) content (%)
(m) (m) (m) (m/m)
London 300 0.35 60 100 0.0035 0.06 0 Eocene

Savannah 200 0.19 50 48 0.004 0.08 60 Miocene
Venice 150 0.12 130 9 0.013 0.10 10 Recent
Houston 6000 2.3 150 90 0.025 0.21 50 Recent
Bangkok 600 1.2 200 30 0.05* 0.5* 40 Recent
Santa Clara 650 5.3 145 49 0.11 0.74 70 Recent
Mexico City 220 9.0 50 55 0.16 32.0 80 Recent
* Specific subsidence and compressibility values for Bangkok are for the upper part of the clay sequence only.
Subsidence rates on clay may therefore be pre-

dicted for any site, except that local geological
factors preclude accurate predictions without a
database of recorded movements.
The role of pressure decline within overpumped
aquifers was first recognised in the Santa Clara
Valley of California (Figure 12.2). Subsidence was
virtually stopped by water table recovery during
1935–45, largely due to higher rainfalls, but
restarted as soon as water levels declined past
their previous minimum, when pumping increased
in drier years. The only engineering response to
prevent or reduce subsidence is to facilitate
groundwater recovery. Reduction of abstractions
Figure 12.2 Correlation of aquifer head decline and
ground subsidence in the Santa Clara Valley, California. is critical, and aquifer recharge can be employed
where seasonal water excesses can be injected
via existing abstraction wells. Both were applied
than smectite, which is formed primarily by at Santa Clara after 1965, and subsidence has
weathering of volcanic rocks in wet tropical envir- stopped. Similarly, the subsidence of many other
onments, and therefore imposes some geological cities (notably Tokyo, Osaka, Shanghai and
and climatic control on the scale of clay subsid- Houston) has been controlled, but Bangkok contin-
ence. These factors combine to influence the ues to subside while a new water supply system is
compressibility of a clay (measured as metres of developed to replace the thousands of shallow
subsidence per metre of head decline per metre of wells in the city.
clay thickness). Table 12.1 demonstrates the roles Injection of surface water into an aquifer may
of smectite and age, even though the data are gen- cause substantial head recovery, but only stops
eralised because there is rarely a clearly defined the subsidence. Reversal of subsidence, known as
base to the clays that are compressed due to rebound, due to re-inflation of the aquifer is min-
groundwater abstraction that is mainly at shallow imal. Reversible elastic deformation of sand
depths. The compressibility of the Bangkok clay aquifers are recorded through seasonal fluctuations
may appear high because it is based on data from of water table levels, but the amounts are only a
a short period of rapid subsidence, whereas data tiny fraction of the largely inelastic compression of
at the other sites covers a longer term, during the interbedded clays (Poland, 1984). Ground
which stability is approached due to consolidation. rebound is therefore normally only 1–6% of the
GFE-12.qxd 6/28/05 3:43 PM Page 321
SUBSIDENCE 321
Figure 12.3 Causes and effects of the subsidence of Venice over the last 100 years. The bar graph shows, for each
5–year period, the numbers of flood events when high tides reach more than 600 mm above the level that initiates
flooding in the lowest part of Piazza San Marco.
amount of previous subsidence. This did give rise two-thirds of the movement was due to crushing
to claims that Venice had started to rise instead of of the feldspar grains at contact points once
subside soon after groundwater pumping was hydrostatic support was lost in the reservoir sand-
severely reduced by legislation. The claims were stones at depths of 600–1200 m. Secondary oil
however short-lived, as rebound soon stopped, production by water injection then stopped the
while world sea levels continue to rise so that the subsidence, and caused about 6% rebound. Chalk
winter flood events continue to be ever more fre- reservoir rock in the North Sea’s Ekofisk oilfield
quent (Figure 12.3). was similarly crushed, necessitating raising of the
Groundwater withdrawals can be matched by production platform during its abstraction lifetime.
extraction of petroleum, natural gas and steam as Natural gas exploitation has caused comparable
causes of major ground subsidence, where the subsidence over reservoirs in Japan, Holland and
loss of hydraulic support is the common factor, Italy. Depletion of geothermal steam has caused
though it may also affect rocks other than clay. over 4 m of ground subsidence above parts of the
Oilfield subsidence includes the classic case at Wairakei field in New Zealand. This movement
Wilmington, which caused the Long Beach har- developed after pressure decline in reservoirs of
bour area of Los Angeles to subside by nearly 9 m porous rhyolitic pyroclastics; geothermal fields
in the period 1935–1960. Though part of this was elsewhere have suffered no ground subsidence
due to compression of inter-bedded clays, about over reservoirs of stronger rocks.
GFE-12.qxd 6/28/05 3:43 PM Page 322
Table 12.2 Safe distances for houses to stand away and removal of any closer is often the only way to
from various species of trees to avoid subsidence dam- eliminate building subsidence. Movement may
age by root water abstraction (extracted from ISE, also be the consequence of site drainage to allow
2000). deep excavations, and in such cases installation of
Species Mature Recorded Safe a cut-off wall may be the only way to protect adja-
height influence distance cent structures. Climatic effects impact on larger
(m) (m) (m) areas, and the famously dry summers of 1976 and
1989 in the UK started a wave of subsidence dam-
Willow 24 18 40 age to older houses on foundations so shallow that
Poplar 28 20 35
clay beneath them suffered first-time shrinkage in
Oak 24 18 3
the new regime of drier climates.
Elm 25 19 30
Cherry 17 6 11
Loading compaction of clay
All soils and rocks compact under load, but the
At shallow depths, where stronger rocks are movement is only more than negligible on weak
stable, shrinkage due to water loss only causes soils, and notably on clay. The normal engineer-
subsidence in clays. This is well known when ing response is to distribute loads vertically or
trees extract excessive amounts of soil water dur- horizontally by foundation designs that include
ing times of drought. The amount of subsidence is rafts and friction piles respectively. Acceptable
related to potential shrinkage of the soil, but all stresses on clay soils have been established by
clays may be affected, including those in the decades of soils engineering practice (Table
weathering profile of mudstones that are stable at 12.3). These values are determined by the clay
depth. Oaks, poplar and willow are the most mineralogy and consolidation history, but are
powerful tree species, and flowering cherries have always determined by soil properties measured
a bad reputation because they are often planted by routine testing for each site, as geological
close to houses. All trees should be placed away classifications cannot adequately define the vari-
from houses by a safe distance, which is usually ations in soil properties. Precise loading limits
rather more than the tree’s height (Table 12.2), are calculated with respect to water content and
Table 12.3 Acceptable Bearing Pressures (ABPs) (or Presumed Bearing Values) for foundations with static loading
on cohesive clay soils, where long-term settlement is ⬍ 50 mm; acceptable values are also modified, upwards or
downwards, with respect to foundation width and shape.
Description Example in Britain Shear strength CPT (MN/m2) Liquidity Acceptable Bearing
(kN/m2) Index Pressure (kN/m2)
Soft soil Alluvium 20–40 0.3–0.5 ⬎0.5 ⬍50

Firm soil Drift at depth 40–75 0.5–1.0 0.2–0.5 75–100
Stiff soil Weathered London Clay 75–150 1–4 ⫺0.1–0.2 150–250
Very stiff soil Fresh London Clay 150–300 2–8 ⫺0.4–⫺0.1 300–500
Hard soil Gault Clay ⬎300 ⬎4 ⬍ ⫺0.4 600
Weak rock Carboniferous shale 500–1000 400–750
Strong rock Silurian mudstone ⬎2000 1000–2000
Note: the comparable values for rock are only approximate and their ABPs are lower because fracture spacing influ-
ences bearing parameters in rocks.
CPT ⫽ end resistance in a Cone Penetration Test (which cannot be applied to rock).
Liquidity Index ⫽ (water content ⫺ plastic limit)/plasticity index, and is a measure of soil strength at a given
water content.
The mudstone is only strong in the context of the clay rocks.
GFE-12.qxd 6/28/05 3:43 PM Page 323
SUBSIDENCE 323
the Mohr-Coulomb parameters of internal fric-

tion and cohesion (Chapter 3; Craig, 1997).
Structural loading may easily cause shear failure
of a soil, and safe limits are defined by a suitable
factor of safety (normally 3) in relation to ulti-
mate failure stress. In practice, the acceptable
stress limit is normally determined by the accept-
ability of the induced settlement.
Exceeding the safe and acceptable loading may
lead to gross settlement, which is even more seri-
ous when it is differential (Figure 12.4). The ulti-
mate example is the Leaning Tower of Pisa, in Italy
(Figure 12.5). Its foundations were hopelessly
small, and a detail of lateral variation within the
ground instigated differential subsidence; this then
became self-enhancing by differential loading, as
the centre of gravity of the tall narrow tower moved
with the tilt. The remedial engineering was a
Figure 12.5 Differential subsidence and its remedia-

tion at the Leaning Tower of Pisa. The diagram above
shows the ground profile and remedial methods to sta-
bilise the tower, and the lower graph shows the rates of
subsidence since the tower was built.
classic, with a temporary surcharge followed by the

drilled removal of soil beneath the foundations,
both on the north side, to induce a tilt towards the
north, counter to the long-term southerly tilt
(Burland, 1997).
Extreme subsidence problems are created
where a soft clay is both loaded and de-watered,
and the prime example is provided by Mexico
City (Figueroa Vega, 1984). The downtown area is
built on an old lake bed of thick, young and very
soft, volcanic-derived, smectite-rich clays, and
these are inter-bedded with sands that have been
over pumped for water supply for many decades.
Subsidence is therefore twofold:
Figure 12.4 Differential subsidence of an old build-
ing that stands on short timber piles into the soft young 1. The clays are so soft that all buildings subside
clay that underlies Amsterdam. under self-load. The Palace of Fine Arts, built on
GFE-12.qxd 6/28/05 3:43 PM Page 324
Figure 12.6 Diagrammatic profile through Mexico City with, from the left, the subsided Palace of Fine Arts, the
successful Latin American Tower, a ‘perched’ building on deep foundations, and an old well casing that now protrudes
above street level (because it is founded below the clays that it has helped to de-water by abstraction from the sands).
a massive raft, has subsided by 3 m. Movements Tower is successfully founded on piles to a shallow
are exacerbated by earthquakes that cause par- sand, so that loading compaction of the underlying
tial liquefaction of some of the soils. clay matches de-watering compaction of the higher
2. The entire city centre has subsided due to the clay — and the tower’s entrance remains at street
groundwater abstraction. This has reached a level (Figure 12.6).
maximum of about 9 m. Subsidence rates are
now reduced by pumping controls.
12.3 Subsidence on peat
The engineering response has been the evolution of
survivable building designs. All new structures are Formed entirely of partially decayed plant material,
placed on raft foundations that can tolerate settle- peat is the weakest natural soil that exists. Blanket
ments reaching up to 1 m. Modern high-rise build- peats are spread over wet upland, but the main peat
ings have deep piles to sand horizons, and many areas are the extensive fenlands, distinguished by
also have deep basements that act as compensated their totally flat landscapes and black soils. Peat
foundations by reducing effective loads. However, has a bearing capacity that is effectively nil, until it
deep-piled foundations leave a stable building has been drained or consolidated by surcharge, and
standing higher than surrounding ground that has even then it is stable only under very small loads.
subsided due to de-watering. The Latin American Excessive subsidence is almost inevitable on peat.
GFE-12.qxd 6/28/05 3:44 PM Page 325
SUBSIDENCE 325
The organic structure of peat varies from fibrous to peat are rarely more than about 10 m thick. On
granular and amorphous, but its engineering prop- many construction sites the response to peat is
erties relate largely to the local burial and drainage therefore its entire removal. An alternative is pla-
history. Water contents are typically 500–2000%, cing end-bearing piles through to the base of the
but may be as low as 100% from above the water peat, as is required by state law for all buildings in
table. Undrained, peat acts as a liquid with negligi- the peatlands of Florida. Options on thick peat
ble strength, but after drainage and structural load- include floating structures on footings of poly-
ing a safe bearing pressure of 70 k Pa may be styrene, sawdust, brushwood, timber corduroy or
applied to the partially consolidated peat. baled peat (which are stable when depressed
Subsidence on peat is induced by either load- below the water table). Raft foundations (maybe
ing or drainage, and is therefore similar to clay with compensating basements) can support build-
except that the compressibility of peat is far ings on drained peat, and highway construction is
greater. Subsidence under load means that few successful after surcharging of peat.
structures can be placed on peat without special Drainage causes immediate subsidence of peat.
treatment. Most hill or moor peat is only a few This is widely recorded, but nowhere better than by
metres thick, while lake-fills, muskeg and fenland the Holme Post in the English Fens (Hutchinson,
Figure 12.7 A sequence of images that show the emergence of the Holme Post from its drained and subsiding peat.
The 1850 image is a reconstruction, and the 1932 image is from Fowler (1933).
GFE-12.qxd 6/28/05 3:44 PM Page 326
1980). As the Holme Fen was the last to be drained 20–30% of head decline on subsequent phases of
(to permit useful agriculture) subsidence was antic- renewed drainage.
ipated and the post was placed in the peat but was After the initial drainage and compaction, sub-
founded on the stable clay beneath (Waltham, sidence on peat continues due to wastage — loss
2000). Through multiple phases of land drainage, to the atmosphere of the peat left above the water
the post emerged from the subsiding peat (Figure table and thereby exposed to microbial oxidation.
12.7). Each pumped fall of the water table causes The subsidence rate depends on the type of peat,
almost immediate subsidence on the peat (Figure and increases with greater depth to the water table
12.8). Across sites worldwide, this subsidence on and with higher soil temperatures (Figure 12.9).
peat (that was originally 5–10 m thick) is about Annual surface lowering is in the range of
50–65% of the head decline on first drainage and 5–100 mm and is typically 1–7% of the depth to
the water table. The effect of wastage subsidence
is that agricultural land inevitably becomes
too low and too wet, thus requiring renewed drain-
age, causing renewed subsidence. Rivers have to
be maintained between high banks across the
subsidence bowl, land drainage water has to be
pumped up into the elevated rivers, and buildings
on piles progressively rise from the peat requiring
increasing flights of steps up to their entrances. All
these features can be seen in the English Fens.
12.4 Hydrocompaction of
collapsing soils
Sediments prone to internal structural collapse

Figure 12.8 The record of 150 years of fenland when water is added to them are generally known
drainage and ground subsidence on the peat, that is
provided by the Holme Post.
as collapsing soils. Weak clay bonds between the
grains of loosely packed silt sediments are broken
by the introduction of water; the soil then densi-
fies by repacking under self-load or imposed load,
in the process known as hydrocompaction.
Potentially collapsible soils are wind-deposited
loess and some alluvial silts that were rapidly
deposited and then desiccated in large basin fans.
Hydrocompaction can promote ground subsid-
ence of up to 5 m over wide areas, and this repre-
sents up to 10% of the original thickness of the
collapsible soil. This only occurs in semi-arid
regions where the soils have not been previously
wetted, and collapse potential is normally limited
at depth by the water table or by prior consolida-
tion. Subsidence occurs most commonly where
water is added to the soils artificially, notably by
Figure 12.9 Peat wastage rates related to mean
irrigation schemes in semi-desert terrains. The
temperatures and depths to the water table (after rapid progress of hydrocompaction can be seen
Stephens and Stewart, 1977). both in laboratory consolidation tests where water
GFE-12.qxd 6/28/05 3:44 PM Page 327
SUBSIDENCE 327
Figure 12.10 Hydrocompaction of collapsible silt soils when water is added to them, recorded in laboratory
consolidation tests (from Waltham, 1989) and in field tests in the Central Valley of California, which reveal the time
delay of subsidence on thicker soils (after Lofgren, 1969).
is added during loading, and also in field tests 12.5 Subsidence on permafrost
where surface water is allowed to infiltrate the
soils (Figure 12.10). Collapse only occurs in soils Continuous, permanently frozen ground exists
with initial porosities that are ⬎ 40% (Dudley, where mean air temperatures are lower than
1970), and hydrocompaction is at a maximum in ⫺8 °C, and discontinuous permafrost occurs
soils with 12% clay (that provides the metastable where mean air temperatures are between ⫺1 °C
bonds between silt particles). The amounts of and ⫺8 °C. Permafrost may be only 20 m thick
potential subsidence can be approximately esti- where it is discontinuous, but can reach to more
mated from the laboratory tests, which can also than 500 m deep in the coldest regions. Below the
identify those soils that may compact by more permafrost, the talik zone is kept unfrozen by
than 10% when also subject to engineering loads. geothermal heat (Chapter 14).
The San Joaquin Valley, the southern part of The top 1–3 m of the ground is known as the
the Central Valley of California, is a prime site of active layer as it thaws and re-freezes annually, cre-
massive ground subsidence by hydrocompaction ating the huge unstable bogs that characterise the
due to the introduction of irrigation water (as well Arctic in summer. Gravels that retain grain-to-grain
as including even larger areas that have subsided contact are thaw-stable, as are all rocks. Clays
due to water abstraction from underlying clays). and silts are solid when frozen, but are very weak
Rolling profiles are acceptable in the newly irri- when wet and saturated. Sands are thaw-stable with
gated fields, but the subsidence has a serious low ice contents, but some permafrost contains over
impact on the canals that carry the irrigation 60% ice, and this must collapse when it is thawed.
water into and through the area (Prokopovich, Structures can be founded over permafrost,
1986). Pre-construction flooding over a period of but subsidence is inevitable if they cause thawing
two years has successfully instigated complete of any soils that obtain their strength largely
hydrocompaction to leave zones of subsided but or wholly from their interstitial ice. Dawson
stable ground. Selective soil wetting by sprinkle City grew out of the Klondike goldfields in
irrigation, and mechanical compaction, can the Canadian Yukon, and its original houses were
improve shallow or thin collapsible soils, but may built directly on thaw-stable gravels that overlie
leave dormant the potential for a subsequent frozen clay soils. Downward heat loss from the
phase of hydrocompaction. warm houses thawed the unstable clays, and every
GFE-12.qxd 6/28/05 3:44 PM Page 328
Figure 12.11 Houses subsided on permafrost thawed by their own heat in Dawson, Canada.
one of the original timber buildings subsided and

tilted (Figure 12.11). The key to subsidence pre-
vention is conservation of the permafrost.
Stable roads are built on thick gravel pads
that provide the insulation beneath the bare
road surface. The roadbed for Alaska’s Dalton
Highway was built by end-tipping gravel directly
onto the natural surface. The vegetation and
organic muck were never stripped out, as they
provide the natural insulation and removing
them would have let construction traffic sink
into an unfrozen quagmire (as happened on
the original military Alaska Highway). A layer
of expanded polystyrene 70–150 mm thick can
preserve more fragile ground and reduce the
required thickness of gravel.
Low-rise houses can rest on wooden blocks,
preferably 1 m above the ground, with open access
beneath for cold winter air. Larger buildings rest
on thick gravel pads, like those for roads. Heated
buildings often require extra cooling of the gravel
by air ducts aligned with prevailing winter winds;
some have ammonia-filled refrigeration tubes.
Figure 12.12 Utilidors and buildings on piles in
Longyearban, Svalbard. The tunnels in the foreground
Alternatively, structures are built on adfreeze piles
take the elevated utilidors through a gravel bank sup- which are sunk into stable ground ice below the
porting a road, and the pile-supported buildings have active layer: oversize pile holes are drilled before a
free-air underspace closed only by slatted boarding. pile is inserted and the hole is then filled with
GFE-12.qxd 6/28/05 3:44 PM Page 329
SUBSIDENCE 329
slurry that freezes after a few days or weeks to cre- rockhead can cause its almost total removal within
ate a stable bearing pile. Water and sewer lines are the weathering zone, leaving a complex breccia of
installed above ground in utilidors (Figure 12.12); collapsed insoluble inter-beds, remnant gypsum
the Trans-Alaska Oil Pipeline works on the same blocks, slumped soils and isolated open voids.
principle of keeping warm pipe contents away These collapse breccias provide very unstable
from fragile permafrost, though with the addition ground, that is unmatched in limestone karst, and
of refrigerated piles for added protection. sites in England have required extensive grouting
and ground improvement for safe construction on
them (Cooper and Saunders, 2002).
12.6 Subsidence on limestone
and gypsum
12.7 Salt subsidence
The vast majority of the world’s natural cavities
are in limestone and their existence provides Salt, as halite or rock salt, is totally and rapidly
the potential for ground subsidence wherever soluble in natural waters. Significant voids and
limestone is at outcrop, just below outcrop or total removal of salt beds can develop within a
lying beneath a soil cover. Collapses of rock few years, orders of magnitude faster than min-
over caves are very rare events, and voids within eral losses in limestone and gypsum, and salt is so
the limestone present only a very small but soluble that it cannot survive at outcrop except in
unpredictable risk to engineering works. A much desert climates. Most salt bodies occur within
more widespread hazard is the development of clay sequences where they remain dry because
subsidence sinkholes in soil covers that are groundwater cannot reach them. The major subsid-
washed into caves or fissures in underlying lime- ence process is dissolution at a rockhead beneath
stone. Karst terrains have a very distinctive geo- thick soil and drift covers that contain flowing
morphology of closed depressions, isolated hills groundwater.
and bare rock outcrops, that are fully considered In lowland sites like the English Cheshire Plain,
in Chapter 24 –– this covers the processes, haz- rockhead dissolution of salt is limited by the layer-
ards and remediation of both cave collapses and ing of the drift groundwater where heavier brine
subsidence sinkholes (Waltham et al., 2005). (saturated with salt) lies beneath the fresh water
Gypsum is the world’s second most cavernous that is capable of dissolving the salt. Where salt
rock, and can develop extensive karst landscapes beds are dissolved at a rockhead, insoluble residues
that have their own subsidence hazards (Chapter and the typically abundant inter-bedded clays are
24). However, gypsum is much more rapidly dis- left as a residual dissolution breccia (Figure 12.13).
solved in natural environments than is limestone This further restricts groundwater flow, but
and can therefore cause gentle subsidence of large dissolution does continue on the top of the salt
areas of ground by extensive dissolution at its rock- beneath (known in Cheshire as the ‘wet rockhead’).
head (or at a buried contact with an aquifer). Dissolution rates and subsidence rates are a func-
Quarries in gypsum in England and Germany have tion of groundwater flow patterns, and overall sub-
provided cut sections through subsidence bowls sidence is generally ⬎ 0.1 mm/year. Continuing
that are hundreds of metres across and only a few through a large part of the Pleistocene, this has cre-
metres deep. A comparable subsidence bowl has ated subsidence hollows that became lakes as they
deformed a terrace of houses in Ripon, UK, within sank below the water table; these are now known as
their lifetime of less than 200 years. It is unclear ‘meres’, many of which are over 1000 m across,
how much of this subsidence is due to active dis- and are diagnostic of the Cheshire-style of lowland
solution of the gypsum rockhead, and how much is salt karst. These natural movements are generally
due to compression of peat soils that have accumu- too slow to record, except where tilting of new
lated in an older subsidence bowl that had formed houses show that they do continue today. The most
more slowly. The rapid dissolution of gypsum at its severe movements occur in linear subsidences
GFE-12.qxd 6/28/05 3:44 PM Page 330
Figure 12.13 Diagrammatic profile through zones of active subsidence and stable ground over salt beds in the
Cheshire Plain (UK). Subsidence can occur anywhere where the salt beds are in contact with fresh groundwater in
the drift sediments — along the zone known colloquially as ‘wet rockhead’. Under a protective cap of dry mudstone,
the salt is stable.
Figure 12.14 The linear subsidence occupied by Moston Flash in the Cheshire salt field (UK). The road
required repeated raising when subsidence was accelerated by wild brining, but has become almost stable since
the local brine pumping ceased.
(typically 1–5 km long, 100–400 m wide and replace the abstracted brine. Traditional wild brin-
1–10 m deep) that appear as valleys except that ing targets the brine streams, and thereby causes
they continue over interfluves. These can form over the linear subsidences above them to deepen by
rockhead outcrops of thicker beds of pure salt and 100 mm/year or more, ultimately creating new rib-
‘brine streams’ that are zones of enhanced ground- bon lakes, known as ‘flashes’, within them (Figure
water flow just above rockhead (Waltham, 1989). 12.14). Catastrophic subsidences were caused by
All salt subsidence is vastly accelerated by small mines beneath wet rockhead, which were
any brine-pumping operations that draw in new abandoned when unsafe, allowed to flood, then
supplies of chemically aggressive freshwater to brine-pumped (in the ‘bastard-brining’ operations)
GFE-12.qxd 6/28/05 3:44 PM Page 331
SUBSIDENCE 331
Figure 12.15 A house in Northwich, UK, seriously subsided over an old mine that collapsed during bastard brin-
ing. It survived intact because the brickwork formed only panels within a strong timber frame (from an old postcard).
until the remnant pillars were dissolved by new wild-brining operations, but the almost complete
inputs of freshwater. Major collapses caused sub- end of wild brining and its associated major sub-
sidence lakes to form overnight — the true, rap- sidence made the Board obsolete. The risk of any
idly formed flashes. one house now subsiding on the salt is statistically
All the severe salt subsidence was induced very small, and is sensibly carried by appropriate
by the mining and brining. Bastard brining was insurance surcharges.
banned in Cheshire in the 1930s (but continued in
Thailand until the 1980s). Since then, wild brin-
ing has steadily been replaced by stable controlled 12.8 Volcanic subsidence
brining from caverns in deep dry rock, where
Cheshire’s one pillar-and-stall salt mine is also Active volcanoes offer a unique suite of subsid-
dry and stable. The engineering response in the ence processes whereby liquid rock can flow
salt fields has been to create solid structures that away from beneath particular sites. On the small
can be jacked up after they have subsided. Timber scale, liquid lava can drain out from the cores of
frames, steel frames or concrete rafts have been lava flows. Drainage from a lava pond normally
used in the Cheshire salt fields for 100 years causes immediate subsidence of its unsupported
(Figure 12.15), but simple reinforced strip foot- solid crust, but this is a process of eruption activ-
ings are only regarded as an optional precaution ity, with no influence on engineering. In contrast,
for today’s minimal movements. drainage from a single tube leaves a lava cave
Jacking and repair costs for any houses were with a thin rock roof that is liable to be stable in
covered by the Cheshire Brine Compensation the short term but offers the long-term threat of its
Board, which was funded from a levy on all potential collapse (Chapter 23).
GFE-12.qxd 6/28/05 3:44 PM Page 332
Figure 12.16 Subsidence and uplift of the Temple of Serapis, over the active magma chamber of Campi Flegri,
near Naples, Italy. Monte Nuovo erupted from a new vent just 3 km away in 1538, after a period of rapid uplift.
On the large scale, entire volcanoes and their 3. Caves in chalk are rarely more than a metre
immediate surrounds, can deflate due to migration across and the main subsidence hazard is the
or retreat of magma from beneath them. Deflation development of sinkholes when soil is washed
of Italy’s Campi Flegri volcanic centre, just west of down narrow fissures (Chapter 24).
Naples, caused inundation of the coastal Temple 4. Caves in gypsum are widespread. They are
of Serapis, which subsided by over 10 m in 1200 mostly smaller than caves in limestone but they
years (Figure 12.16). This is, however, one case form and expand by dissolution more rapidly
where ground subsidence is welcome, as the reverse and collapse more readily in the weaker rock.
uplift is due to volcanic inflation that normally 5. Caves in salt occur only in arid environments
precedes an even more destructive eruption. and are mostly small.
6. Caves in quartzite are formed by very slow dis-
12.9 Implications for engineering on solution of silica and only grow to significant
naturally cavernous ground size in very ancient and long-exposed craton
terrains. Some collapse cavities in quartzite in
The major engineering problem posed by natural Venezuela are over 100 m across, but they are
ground cavities is that their locations are generally stable, incredibly old and in high, remote
unpredictable and those with no open entrance can mountains.
be very difficult or very expensive to find prior 7. Caves in granite are mainly complex systems of
to construction activity (Culshaw and Waltham, narrow fissures and breakdown rooms, and are
1987). Fortunately, natural caves only occur in a not a significant feature of most granite terrains.
limited number of rock types and their main char- 8. Caves in soil are known as ‘pipes’. They
acterstics may be summarised as follows: develop headwards along seepage lines as the
fine and then the coarse fractions of the soil
1. Caves in limestone are the most abundant. are progressively washed out (Figure 12.17).
They may be many tens of metres across, but Piping failures are common in soils that are
most are at depths that make them irrelevant to washed into broken drains and sewers, and
surface subsidence. The scale of cave develop- thereby exacerbate subsidence of houses on
ment in limestone, and therefore of potential shrinkable clays (where the clay movement
subsidence, increases in the hot and wet cli- caused an initial drain break).
matic regions (Chapter 24).
2. Caves in dolomite are essentially the same as The only other natural caves are sea caves and
those in limestone but are less common. river cliff notches which all have conspicuous
GFE-12.qxd 6/28/05 3:44 PM Page 333
SUBSIDENCE 333
Figure 12.17 Subsidence due to collapse of piping cavities in a terrace of fine-grained soils in southern Tunisia,
where uncontrolled drainage developed pipes to point outlets on a lower terrace.
entrances and present no hidden geohazard; ten-

sion fissures (also known as ‘tectonic caves’,
‘gulls’ and ‘windypits’) that develop in the head
zones of landslides (Chapter 8); caves in clastic
rocks that have soluble calcite cements and/or
clasts; isolated fissure caves that have opened in
tension zones or due to glacial drag, and caves in
glacier ice (that are irrelevant to foundation engi-
neering). All these are rare but, along with mined
cavities, can occur in any rock.
In any of these geological conditions where
caves could occur, practising engineers must
consider the subsidence hazard that they present.
The engineering risk must be evaluated for each
site with respect to the impact of a ground col-
lapse and the chance of an unknown cave under-
lying the structure. Even in mature karst terrains,
the statistical chance of encountering a cave at a
critical location under a single structure is very
small, but its potential impact may be disastrous.
A highway across karst is more likely to
encounter a cave, but it probably creates minimal
engineering inconvenience, and many long road
excavations reveal no caves at all.
Each site in potentially cavernous ground must
be treated on its own merits. Engineers have the Figure 12.18 A heavily reinforced raft for structural
option of two responses. They may find and reme- foundations in cavernous pinnacle karst on limestone at
diate every significant cavity. This may demand Shilin, China.
GFE-12.qxd 6/28/05 3:44 PM Page 334
Figure 12.19 Collapse sinkholes in the gypsum karst of central Turkey: (A) a very large old sinkhole with no sign
of recent collapse and (B) the small-scale progressive collapse active within a corner of a nearby sinkhole whose
total area equals that of the large old feature.
very expensive, close-spaced grids of boreholes to extended ground beams, reinforced rafts (Figure
eliminate every possibility, as geophysical searches 12.18) and sacrificial piers may be employed as
cannot be totally reliable (Chapter 24). A rock arch appropriate. Though collapse sinkholes and
over a natural cave can develop a stable span that is collapse-modified caverns up to hundreds of
at least twice the width of its thickness in sound metres across are known in terrains of limestone,
rock (thereby excluding soils prone to piping and gypsum and quartzite, no surface collapse more
most chalks). Under-drilling to prove between 3 and than about 20 m across has ever been seen or can be
5 m of sound rock is therefore sufficient to ensure recognised as a single past event (Figure 12.19).
the integrity of foundations in almost any area of The larger features have developed by progressive,
cavernous ground (Waltham et al., 2005). However, multiple collapses over geological timescales. This
every site must be assessed for its own conditions. implies a worst-case scenario for sensible design of
Alternatively, engineers may design structures bridging structures. In most terrains, the largest
safely to span any unseen ground cavity that potential collapse is smaller. The exception is
could possibly collapse. Reinforced strip footings, provided by large subsidence sinkholes that can
GFE-12.qxd 6/28/05 3:44 PM Page 335
SUBSIDENCE 335
develop almost instantaneously by adequately thick shapes of orebodies and have evolved around the
soils slumping into cavernous limestone below. world and over the centuries, but their subsidence
Excavating a tunnel through cavernous ground hazards fall into four groups — longwall mining,
involves an unavoidable element of risk, which block caving (and stope mining), old pillar-and-
can only be evaluated by reference to local stall workings and old shafts.
records on known caves and voids revealed in Longwall mining is a total extraction method.
boreholes. Exploration ahead of the face is best It removes all of a thin roughly horizontal ore
by long advance probes, and large voids can often bed, on a usually mechanised face, that retreats
be grouted or treated before the tunnel reaches and lets the ground collapse behind its travelling
them, though tunnel diversion may be required zone of temporary support (Figure 12.20).
round large unstable caverns (Milanovic, 2000). Ground subsidence is inevitable, immediate and
The benefits of flexibility make drill-and-blast broadly predictable (Waltham, 1989), but is nor-
preferable to tunnel boring machines (TBMs) on mally completed within about two years of its
many projects in cavernous ground. undermining, when complete closure of the mine
workings removes most of the potential for fur-
12.10 Subsidence due to mining ther surface decline. Longwall mining has been
widespread in coal workings since about 1930,
The extraction of minerals for economic benefit and its ground distortion in active coalfields is
becomes a geomorphological process when man generally referred to as ‘mining subsidence’
is accepted as part of the natural world, and the (Figure 12.21).
resultant subsidence is certainly of major concern Block caving is also a total extraction method. It
to engineers. Also, it cannot be ignored as it is one is applied in massive orebodies where the ore is
of the most widespread and destructive surface blasted and then tapped off from beneath into
processes, especially in the developed world. underlying transport tunnels while the walls above
There is immense variation in underground min- are allowed to collapse into a mass of broken rock
ing techniques that have been applied in different that subsides as more ore is removed from below.
Figure 12.20 Diagrammatic profile through an advancing longwall mine face, showing the components of movement
within the migrating wave of ground subsidence. Note the scale distortion: a coal seam is typically 1–2 m thick, and is
generally worked at depths of 200–400 m.
GFE-12.qxd 6/28/05 3:44 PM Page 336
roof failures are followed by upward void migration

and crown hole failures at the surface; both types of
subsidence can occur over mines long abandoned.
Most thin, sub-horizontal ore was worked by this
method, including most coal before about 1930.
Mine shafts were essential features above and
adjacent to all underground mines. Even where the
actual mining creates no subsidence threat, aban-
doned shafts create serious hazards especially
where their walls are unstable or they lie unseen
beneath inadequate covers. A shaft failure creates
only a small sinkhole-type subsidence, but this can
totally destroy any house or structure inadvertently
placed on a ground cover that hides it.
Engineering implications of old and

modern mining
The subsidence hazard from modern mining should
be predictable and controllable as long as the mine
is properly designed and managed. Subsidence over
longwall coal mines involves ground compression,
extension and tilt besides ultimate subsidence, and
all parameters are predictable on the basis of both
theoretical deformation and long-term empirical
Figure 12.21 ‘Very severe’ mining subsidence dam-
records (Whittaker and Reddish, 1989) except
age to a house on the Nottinghamshire coalfield (UK),
where ground strain has damage the brickwork. where geological factors localise and distort move-
ments along faults and joints. The scale of move-
Total removal of the ore leaves no support for the ment increases over thicker seams, and also over
ground above, and surface subsidence may be mas- wider workings closer to the surface. Ground strain
sive, commonly lowering the floors of the aban- is the most destructive component and it controls
doned open-cut pits that worked steeply dipping the classification of subsidence damage used in
orebodies before extraction was forced to continue coalfields but is also applicable more widely
underground at greater depths. A smaller version of (Table 12.4). Strain is typically 1–10 mm/m; values
block caving is the stope working of narrow and at the lower end of the range are normal for modern
steeply inclined vein deposits, which is very wide- deep mines, so damage to houses is minimal,
spread in mining fields all across the world. though bridges and larger structures may still be
Pillar-and-stall mining is the main method of significantly affected. Vertical subsidence is typi-
partial extraction. Ore is excavated from grids or cally 200–1500 mm over a seam 1–2 m thick, and is
networks of stalls (tunnels) while intervening often most significant with respect to drainage in
blocks of ore are left in place as pillars that provide lowland areas where levees and bridges may require
roof support. Dimensions and patterns of working raising, while ground tilt develops in the marginal
vary enormously, and there is no ground subsidence zones of the subsidence bowls (Figure 12.23). With
over pillar-and-stall mines that are designed to be the decline of deep mining in Britain, most work-
stable with respect to the rock mass strengths of the ings are now kept under agricultural land where
pillars and roofs. However, many old mines were subsidence damage is minimised.
left seriously unstable, notably through ‘robbing the Block caving and vein stoping produce such
pillars’, until collapse was imminent (Figure 12.22). large zones of unstable and cavernous ground that
Pillar failures cause major ‘areal subsidence’, while land above them is generally regarded as sterile
GFE-12.qxd 6/28/05 3:44 PM Page 337
SUBSIDENCE 337
Figure 12.22 Narrow pillars of limestone left in an old pillar-and-stall stone mine in Dorset (UK). Though extraction
was about 90%, this mine is stable because it has only a shallow cover resting on strong pillars, but pillars of coal of
similar proportions are liable to collapse causing ‘areal subsidence’.
Table 12.4 Classifications of subsidence damage (after NCB, 1975 and ISE, 2000). The NCB system is related
primarily to ground strains over longwall mines, while the ISE system relates to the crack damage in houses regard-
less of cause.
Classification Class descriptions Crack Length Causative ground strain

terms width change (mm/m)
(mm) (mm)
NCB (1975) ISE (2000) House Bridge
20 m across 100 m long
0 Hairline Negligible cracks that can ⬍ 0.1

be covered
1 Very slight Fine Barely noticeable, small 0.1–1.0
plaster cracks ⬍ 30 1.0 0.2
2 Slight Moderate Slight internal cracks, doors/ 1–5 30–60 2.5 0.5
windows may stick
3 Appreciable Serious Slight external fractures, 5–15 60–120 5.0 1.0
service pipes may
fracture
4 Severe Severe Floors slope and walls 15–25 120–180 7.5 1.5
lean, door frames are
distorted, extensive
repairs needed
5 Very severe Very severe Floors slope badly, walls ⬎ 25 ⬎ 180 10.0 2.0
bulge, floors and beams
lose bearing, partial or
complete rebuilding
GFE-12.qxd 6/28/05 3:44 PM Page 338
Figure 12.23 Tilt of a building in the West Midlands coalfield (UK), which is unusually severe as it stands over
the edge of very thick workings and also on the edge of a zone of soft valley sediment fill.
Figure 12.24 Diagrammatic profile through a shallow old coal mine, with subsidence hazards created by an unsta-
ble hillside adit, crown holes and an ‘areal subsidence’ over multiple pillar failure, while deeper workings with wide
pillars are stable except where an old shaft is inadequately capped.
both during and after active mining. On dipping soon after the mining is abandoned but failures
orebodies, the footwall is normally solid, but the can occur 100 years or more later. The main
hanging wall may be unstable for considerable danger lies in shallow mines (Figure 12.24). Roof
distances from the outcrop. Roads can bridge failures are unlikely to migrate more than about
narrow-vein workings on reinforced ground slabs, 30 m, so crown holes are rare over deeper mines.
but large-scale remediation of old workings is Thin pillars in shallow mines may be critically over-
rarely cost-effective. loaded by new structures placed on the surface,
Old pillar-and-stall mines can constitute a whereas the wider pillars left to support thicker
major subsidence hazard. Many failures occur cover in deep mines are less influenced by
GFE-12.qxd 6/28/05 3:44 PM Page 339
SUBSIDENCE 339
engineering works. Guideline values for the cover

depth required to render a mine of no general
concern vary from 3 m in intact strong rock to
30 m in thinly bedded weak rock (Waltham and
Swift, 2004); local conditions and experience are
always critical to engineering design and practice.
Treatment of shallow old mines may demand
expensive grout injection through close grids of
bored holes, though more specifically tailored
methods can be applied at many sites (Littlejohn,
1979; Waltham, 1989).
Old mine shafts are only rendered safe for
construction over them by an appropriate combin-
ation of filling, grouting and/or capping by over-
sized concrete slabs (NCB, 1982). Most old shafts
were inadequately treated before they were
covered and forgotten, and buildings subsequently
placed on them can suffer, many years later,
subsidence that ranges from distressing to cata-
strophic (Figure 12.25). The main difficulty with
concealed old shafts is locating them when records
of them are imprecise or non-existent. Geophysica
l surveys can be useful (McCann et al., 1987) and
ever-improving techniques mean that any engineer
faced with the problem is well advised to contact
Figure 12.25 Catastrophic subsidence of a house over
a specialist subcontractor to investigate optimum a forgotten mine shaft in the West Midlands (UK), where
approaches, which are usually site-specific. unconsolidated fill had settled within the shaft and also
run into the workings below (the shaft was already being
12.11 Tectonic subsidence backfilled when the photograph was taken).
True tectonic subsidence occurs in two-plate envir- largely due to global warming that has been
onments. The first is where the marginal zones of continuous for about the last 500 years. The
plates are being subducted close to convergent combination has made the Thames Barrier neces-
boundaries — the Nagoya region of eastern Japan sary to combat tidal flooding of London where
is subsiding by 10 mm/year and is among the few there has been no risk in the past.
places where this crustal warping is not restricted Major deltaic basins are also sites of very slow
to ocean floors. The second is where the stretch- but relentless subsidence that is due to a combin-
ing of a plate causes its necking and thinning ation of factors. Soft clay sediments, with typical
and therefore its surface lowering, where mantle porosities of about 55% at depths down to tens of
flow beneath maintains isostatic balance — the metres, are consolidated into mudstones with
North Sea is a prime example of this. While these porosities of about 20% at depths around 3000 m
processes affect very large areas, subsidence rates (Skempton, 1970), and this compaction causes sub-
are mostly very low: the London area is subsiding sidence. Crustal sag of the basin floor matches sed-
by more than 2 mm/year due to thinning and sink- iment accumulation in continuously active deltas to
ing of the North Sea Basin. Such rates of subsid- provide the second component of subsidence.
ence have no impact inland, but may be critical at Under the Po Valley of northern Italy, about 2400 m
coastal sites, especially when combined with the of Quaternary sediment has partially consolidated
current rise in world sea-levels (at 1–2 mm/year), into a sequence about 2000 m thick. A mean
GFE-12.qxd 6/28/05 3:44 PM Page 340
Figure 12.26 The drowned forest of Portage after the rapid ground subsidence associated with the 1964 earthquake
in Alaska.
deposition rate of 1.2 mm/year has been accommo- Accelerated subsidence may be accompanied
dated by 0.2 mm/year of compaction and 1.0 mm/ by earthquakes when crustal deformation is trans-
year of crustal sag. Venice lies towards the margin ferred to fault displacement. Large areas in Alaska,
of this deltaic basin, where the long-term mean sub- between Prince William Sound and the Chugach
sidence is 0.4 mm/year. This is due to both com- Mountains, subsided by up to 2.5 m during the
paction and sag, and is the natural, uncontrollable 1964 earthquake, causing permanent drowning of
component of Venice’s continued subsidence coastal forests (Figure 12.26). Both the earthquake
(Figure 12.3). In the centre of the Mississippi delta, and the subsidence were results of relaxation of the
surface subsidence is about 8 mm/year, due to the American plate that had been buckled and over-
combination of sediment compaction and crustal stressed over the subducting Pacific plate (Figure
sag. The latter occurs under the sediment load and 12.27). Tectonic subsidence is a common result of
there is no scope for subduction down-warping of earthquakes, though usually on smaller scales than
the crust, as may be a contributory factor at oro- that in Alaska. It is uncontrollable, and is no more
genic sites including the Ganges and Po valleys. predictable than are the earthquakes themselves,
Crustal subsidence also occurs under the loads except that subsidence may be anticipated if any
imposed by ice caps and reservoirs. Scandinavia pre-event uplift has been monitored.
temporarily subsided by up to 200 m under its Subsidence is more widespread as a secondary
Pleistocene icecaps. Sea levels were lower at the effect of earthquakes, due to liquefaction of
same time, but isostatic rebound by the landmass unconsolidated sand soils. During the period of
was slower than sea level recovery when ice melted earthquake vibration, sands may temporarily
at the end of each glacial stage, leaving successions behave as a liquid and offer greatly reduced sup-
of raised beaches, marine cliff lines and the famous port to structures, which therefore subside into
Norwegian strand-flat in the coastal region profiles. them, as happened famously to the Kawagishi-cho
Very large masses of water impounded by high buildings in Niigata, Japan, in 1964. The vibra-
dams provide the only cases of crustal subsidence tions also cause densification of loose sands by
that are due to man. Lake Mead, behind the Hoover their improved grain packing, while inter-granular
Dam on the Colorado River, caused a maximum water is expelled through sand eruptions, and this
down-warping of about 200 mm; as at other sites, causes the ground surface to subside permanently.
this appeared to be crustal sag as it occurred on Soil liquefaction only occurs in sands of fine and
strong rocks with very little potential compaction. uniform grain size that are poorly consolidated
GFE-12.qxd 6/28/05 3:44 PM Page 341
SUBSIDENCE 341
Figure 12.27 Diagrammatic profile through the subsidence zone that was created during the 1964 earthquake over
the plate boundary subduction zone of Alaska.
(recognised by low SPT N-values on a standard References

penetration test) and lie below the water table. Dry,
dense or poorly sorted sands do not liquefy, and Burland, J. (1997) Propping up Pisa. Royal Academy of
Engineering, London, 20pp.
mapping the soil parameters allows the definition
Cooper, A. H. and Saunders, J. M. (2002) Road and
of subsidence hazard zones for future earthquakes bridge construction across gypsum karst in England.
(Tinsley et al., 1985). In high-value urban areas, Engineering Geology 65, 217–223.
improved land drainage may then be an effective Craig, R. F. (1997) Soil Mechanics. Spon, London,
protective measure against subsidence. 486pp.
Culshaw, M. G. and Waltham, A. C. (1987) Natural and
artificial cavities as ground engineering hazards.
12.12 Summary Quart. Journ. Eng. Geol. 20, 139–150.
Dudley, J. H. (1970) Review of collapsing soils.
Proc. Amer. Soc. Civil. Engrs. 96 (SM3) 925–
Ground subsidence must be recognised as a 947.
significant geomorphic process and as a potential Figueroa Vega, G. E. (1984) Land subsidence case
geohazard in a range of specific geological environ- history: Mexico. Studies and Reports in Hydrology
ments dictated mainly by rock type. The feature 40, UNESCO, Paris, 217–232.
Fowler, G. (1933) Shrinkage of the peat-covered fen-
common to nearly all subsidence processes, with
lands. Geographical Journal 81, 149–150.
the notable exception of tectonic subsidence, is that Hutchinson, J. N. (1980) The record of peat wastage in
they are largely initiated and exacerbated by man’s the East Anglia fenlands at Holme Post, 1846–1978
own activities. The two most widespread ‘natural’ AD. Journal of Ecology 68, 229–249.
processes of clay compaction and subsidence sink- ISE (2000) Subsidence of Low-rise Buildings. Institution
of Structural Engineers, London, 126pp.
hole development over limestone occur when the
Littlejohn, G. S. (1979) Consolidation of old coal work-
truly natural processes of groundwater movement ings. Ground Engineering 12 (4), 15–21.
and drainage are disturbed, unbalanced or acceler- Lofgren, B. E. (1969) Land subsidence due to the appli-
ated by man. Subsidence on mined ground is cation of water. Geol. Soc. Amer. Reviews in
entirely due to man’s actions, and most other subsi- Engineering Geology 2, 271–303.
McCann, D. M., Jackson, P. D. and Culshaw, M. G.
dence events, on salt, peat, silt and permafrost,
(1987) The use of geophysical surveying methods in
follow a history of local engineering activity. The the detection of natural caves and mineshafts.
implication is that subsidence is a geohazard Quart. Journ. Eng. Geol. 20, 59–73.
created by man, and is therefore avoidable or con- Milanovic, P. (2000) Geological Engineering in Karst.
trollable by man. It is a classic example of ‘man as Zebra, Belgrade, 345pp.
NCB (1975) Subsidence Engineers’ Handbook. National
a geomorphological agent’, and should be in the
Coal Board, London, 111pp.
mind of every engineer embarking on a ground NCB (1982) The Treatment of Disused Mine Shafts and
project. Adits. National Coal Board, London, 88pp.
GFE-12.qxd 6/28/05 3:44 PM Page 342
Poland, J. F. (ed.) (1984) Guidebook to studies of land Waltham, A. C. and Swift, G. M. (2004) Bearing capac-
subsidence due to groundwater withdrawal. Studies ity of rock over mined cavities in Nottingham.
and Reports in Hydrology 40, UNESCO, Paris, Engineering Geology 75, 15-31
323pp. Waltham, T., Bell, F. and Culshaw, M. (2005)
Prokopovich, N. P. (1986) Origin and treatment of hydro- Sinkholes and Subsidence. Springer-Verlag, Berlin
compaction on the San Joaquin Valley, CA, USA. Int. and New York, 382pp.
Assoc. Hydrol. Sciences Publ. 151, 537–546. Whittaker, B. N. and Reddish, D. J. (1989) Subsidence.
Skempton, A. W. (1970) The consolidation of clays by Elsevier, Oxford, 528pp.
gravitational compaction. Quarterly Journal of
Geological Society 125, 373–411.
Stephens, J. C. and Stewart, E. H. (1977) Effect of Further reading
climate on organic soil subsidence. Int. Assoc.
Hydrol. Sciences Publ. 121, 647–655. Carbognin, L. and Gatto, P. (1986) An overview of the
Tinsley, J. C., Youd, T. L., Perkins, D. M. and Chen, subsidence of Venice. Int. Ass. Hydrol. Sci. Publ.
A. T. F. (1985) Evaluating liquefaction potential. 151, 321–328.
US Geological Survey Professional Paper 1360, Holzer, T. L. (1991) Nontectonic subsidence. Geological
263–315. Society of America Centennial Special Volume 3,
Waltham, A. C. (1989) Ground Subsidence. Blackie, 219–232.
Glasgow, 202pp. Stephens, J. C., Allen, L. H. and Chen, E. (1984)
Waltham, T. (2000) Peat subsidence at the Holme Post. Organic soil subsidence. Geol. Soc. Am. Reviews in
Mercian Geologist 15, 49–51. Eng. Geol. 6, 107–122.
Waltham, T. (2002) Sinking cities. Geology Today 18, Waltham, T. (2001) Foundations of Engineering Geology,
95–99. Second Edition. Spon, London, 92pp.
GFE-13.qxd 6/24/05 3:57 PM Page 343
Part III
Environments and Landscapes
GFE-13.qxd 6/24/05 3:57 PM Page 344
GFE-13.qxd 6/24/05 3:57 PM Page 345
13. Glacial Environments

Lewis A. Owen and Edward Derbyshire
13.1 Introduction essentially controls the distribution, dynamics and

morphology of glaciers. As such, in polar and
At present, glaciers cover about 10% of the Earth’s sub-polar regions glaciers may extend to sea level,
surface and create some of the most complex envir- while in tropical and equatorial regions they are
onments on the planet. During the last few million usually restricted to altitudes above ⬃ 4000 m
years the area covered by glaciers has fluctuated above sea-level. In high latitude regions, the
between times of more extensive glaciation glacial environment is commonly bordered by
(glacials) when glaciers covered as much as one- tundra (a mixture of bare ground with varying
third of the Earth’s surface, and times of more lim- degrees of lichen, moss and low bush cover),
ited glaciation (interglacials) similar to today. giving way distally to tundra woodland, and then
Thus, glaciers exert an important influence upon to circumpolar taiga (coniferous forest), each of
most high-latitude and high-altitude environ- these regions being progressively less affected by
ments, both now and in the past. Furthermore, gla- perennially frozen ground, or permafrost, and a
ciers affect adjacent non-glaciated regions by wide range of periglacial processes (Chapter 14).
controlling the nature of stream systems, lakes, In mid-latitude, tropical and equatorial regions,
coastal environments and wind systems. The glacier-bordering environments may include
importance of glaciation in mid- and high-latitude alpine meadows and forests. Topography is another
regions may be gauged by the dominant influence strong influence on glaciation. For example, in
on the landscape exerted by the extensive cover of deeply dissected terrain, glaciers tend to be con-
glacial and proglacial sediments and landforms fined to the valleys while, in low-relief landscapes,
readily seen on maps of surface geology (Eyles they may spread out to form ice sheets.
and Dearman, 1981; Eyles et al., 1983a). Most The morphological form of glaciers is highly
engineering activities in such regions are, there- variable, ranging from small cirque glaciers to
fore, bound to have to deal with sediments and extensive ice sheets (Figures 13.1 and 13.2). The
landforms arising directly or indirectly from type of glacier ice also varies geographically. In
glaciation. It is important to note also that the need high-latitude areas, basal ice may be below the
for engineers to understand the dynamics of gla- pressure melting point (cold-based ice); this
cial systems will increase as populations rise in diminishes the glacier’s ability to slide on its bed
such regions in response to continuing economic with consequent reduction in erosion of the bed
development. The overview of the nature and vari- and entrainment of debris. In contrast, the basal
ability of glacial environments, glacial dynamics ice in mid- and low-latitude glaciers is above the
and glacial sediments and landforms presented pressure melting point, the meltwater released
here is designed to provide a groundwork of at the glacier bed allowing the glacier to slide
understanding for engineers working in glaciated and readily erode the substrate, and entrain
or formerly glaciated terrains. and deposit sediment. The last few million years
Glaciated environments are encountered in have been characterised by major oscillations in
all latitudes (Broecker and Denton, 1990). How- climate that have varied on decadal to millennial
ever, latitude and altitude are fundamental in influ- timescales. As a consequence, glaciers have fluc-
encing the regional and local climate that, in turn, tuated in response to variations in temperature
345
GFE-13.qxd 6/24/05 3:57 PM Page 346
Figure 13.1 (A) The current distribution of major glacier types by morphology, within the areas affected by
Pleistocene glaciation in the northern hemisphere (after Sugden and John, 1976); (B) The morphology of glaciers
(after Sugden and John, 1976).
GFE-13.qxd 6/24/05 3:57 PM Page 347
GLACIAL ENVIRONMENTS 347
Figure 13.2a Series of large nivation hollows, with Figure 13.2c Two valley glaciers, with a cirque gla-
well-developed cirque-like headwalls, facing north cier between (centre, top), hanging above Wright Dry
(left) along a sub-summit ridge in the Sierra de Peña Valley. Wright Valley was formerly occupied by a tran-
Labra, north-western Spain. section glacier that drained the Antarctic ice sheet.
when referring only to the physical characteristics

of a facies with no reference to the origin of the
deposit. A genetic facies provides reference to the
specific mode of formation. Facies can be grouped
into associations and successions and can be used
to help characterise a particular depositional envir-
onment (Walker, 1992). Study of contemporary
environments allows these facies associations and
successions to be compared with ancient sedi-
ments. Thus facies models facilitate the interpret-
ation of sedimentary deposits. On a larger scale, the
facies associations may be grouped together into a
Figure 13.2b Small cirque glacier adjacent to Wright depositional system or landsystem model (Boulton
Dry Valley, eastern Antarctica. The area of the glacier and Eyles, 1979; Eyles, 1983; Evans, 2003). Most
between the fresh snow and the boulder-topped glacial geologists rely on facies and landsystem
moraine ridge (seen above the tents) is mantled with models to help them interpret ancient sediments
aeolian sand in this hyper-arid glacial environment.
and landforms. This approach has proved useful in
providing a framework for predicting the spatial
and precipitation. This has further complicated and temporal variability of glacial landforms and
the glacial environment, resulting in a wide var- sediments that, in turn, can be used to help quan-
iety of patterns and types of glacially eroded and tify the variability of materials likely to be encoun-
depositional landforms and suites of sediments. tered during an engineering project (Fookes,
In order to aid understanding of the complex- 1997). The main glacial landsystems are described
ities of glacial environments, glacial geologists in this chapter. However, because a full appreci-
have resorted to developing facies and landsystem ation of the dynamics and products of the glacial
models for the various types of glacial settings. system requires some understanding of the general
A facies is an individual sedimentary deposit hav- physics of glaciers, glacial erosion, entrainment of
ing a distinct combination of properties that dis- debris and the main depositional processes and
tinguish it from its neighboring sediments (Reading, characteristics of glacial sediments, these topics
1986; Walker, 1992). The term lithofacies is used are considered first.
GFE-13.qxd 6/24/05 3:57 PM Page 348
13.2 Physics of glaciers total accumulation and gross ablation (the sum of
melting and sublimation) on a glacier. Mass bal-
Glaciers tend to form in regions where snow ance may be measured at a particular point, over
layers persist year after year. The transformation an area or for the glacier as whole, and the meas-
of perennial snow into glacier ice is dependent on ured period of time may vary. Measurement of the
the temperature and precipitation regimes within annual mass balance of a whole glacier provides
a region. The successive accumulation of snow an indication of whether a glacier will expand or
layers eventually leads to transformation of the contract. A mass balance that is persistently neg-
snow into glacier ice as the volume of air-filled ative (ablation ⬎ accumulation) will result in gla-
pores is reduced and the material increases in cier contraction. Conversely, when the mass
density from ⬃ 50–200 kg/m3 to 830–910 kg/m3 balance is persistently positive (ablation ⬍ accu-
(Paterson, 1994). It is usual for glacial ice to form mulation) the glacier will expand. A zero mass
initially within small depressions, known as balance (ablation ⫽ accumulation) results in a
nivation hollows, where snow has preferentially stable condition with no substantial advance or
accumulated, or on the tops of high mountains or retreat of the glacier. Figure 13.3a–c shows the
plateaus (Figure 13.2a). If the ice continues to distribution of accumulation and ablation zones,
accumulate as a consequence of climate change, and ice-flow characteristics for different types of
small cirque glaciers (Figure 13.2b) will eventu- glaciers.
ally form from the perennial snow patches in
nivation hollows, and ice caps will develop from
accumulating ice on mountain peaks and
plateaus. If climate change is sufficiently severe
these glaciers may enlarge to form extensive val-
ley glacier systems or ice sheets (Figure 13.2c).
The snow accumulation rate and atmospheric
temperature of a region determine the glacial ice
temperature. This, together with the thickness
of the glacier, the geothermal heat flux and
the frictional heat generated by ice flow deter-
mines whether the basal ice of the glacier is above
or below the pressure melting point (warm- or
cold-based glaciers). Under the influence of
gravity, glacier ice moves downslope. This move-
ment occurs mainly by creep involving plastic
deformation within or between individual ice
crystals, added to which movement may occur by
fracturing when brittle deformation occurs
(Weertman, 1983; Alley, 1992; Paterson, 1994).
In warm-based glaciers, movement may also be
accommodated by deformation in rocks or sedi-
ment beneath the glacier sole (Boulton and Jones,
1979; Boulton and Hindmarsh, 1987) and basal
sliding where slippage occurs over a water layer
(Echelmeyer and Wang, 1987).
The velocity of ice flow is a function of the
Figure 13.3 Models of (A) ice sheet (B) ice shelf and
shape of the glacial basin, the temperature of the (C) valley glacier, showing the distribution of accumu-
ice and the mass balance characteristics. Mass lation and ablation and related flow characteristics
balance is a measure of the difference between (after Sugden and John, 1976).
GFE-13.qxd 6/24/05 3:57 PM Page 349
Figure 13.4 Major factors, associated variables and processes responsible for terrestrial glaciogenic deposition
(after Whiteman, 1996).
The physics of glacier movement and mass may result when glaciers retreat and leave behind
balance is complex because of the highly variable undercut or unsupported slopes that may fail as a
nature of the geology and topography, climate and consequence of stress release (Bjerrum and
hydrology within any one region. This makes Jorstad, 1968; Ballantyne and Benn, 1996).
modelling and predicting the dynamics of glacial The debris produced by subglacial erosion may
systems difficult. Nevertheless, a substantial become incorporated into the glacial ice at or near
amount of glaciological research on a wide range the base of the glacier when subglacial meltwaters
of glacier types has resulted in notable advances refreeze to form regelation ice (Figures 13.4 and
in predicting glacier movement. 13.5). Debris entrained within the ice is referred to
as englacial debris. If such entrainment does not
occur, it may form a subglacial debris layer
13.3 Glacial erosion and entrainment beneath the glacier; even so, entrainment may
of sediment eventually occur owing to subsequent deformation
of the sediment with the advance of a glacier.
Erosion, deposition and deformation of rock and Streams beneath glaciers may erode subglacial
sediment occur as a consequence of the continu- debris and redeposit it within or beyond the glacial
ous movement of a glacier, and glacier advances system. Rock fall and avalanche debris deposited
or retreats (Figure 13.4). Glacial erosion occurs as on the surface of the glacier may also become
a direct consequence of melting and refreezing of entrained within the ice by movement down open
glacial ice on the underlying substrate, abrasion by crevasses, or by incorporation by progressive bur-
particles carried in the ice and subglacial meltwa- ial under new snowfalls. Some of the surface
ter stream processes (Alley et al., 1997). Glacially (supraglacial) debris will be transported to the
eroded forms vary in scale from single mm- and glacier’s terminus without being incorporated into
cm-size striae and friction cracks, m-size bed- the ice. During its journey, supraglacial streams, as
forms such as channels, depressions, roche mou- well as mass movement and aeolian processes may
tonnées (glacially eroded streamlined bedrock rework the supraglacial debris (Owen et al., 2003).
knolls), to large scale forms such as cirque basins, When glaciers advance they may override sedi-
glacial troughs and valleys that dominate the land- ments deposited earlier at points near or adjacent
scape (Benn and Evans, 1998). Further denudation to the ice margin. Such proglacial sediments are
GFE-13.qxd 6/24/05 3:57 PM Page 350
from valley-side weathering

SUBAERIAL INPUT
from older till deposited on
valley-side slopes lacial aeolian transport
rag aeolian and
sup glaciofluvial
supraglacial fluvial transport intercalations
flow slide
SUPRAGLACIAL meltout
wa flow slide DEPOSITED
te ENGLACIAL meltout
r- TILL SUITE
or flow meltout
b
ne BASAL lodgement glaciofluvial
se d intercalations
from abraded and i m e n ts englacial & subglacial fluvial transport
plucked glacier bed
from overridden and SUBGLACIAL INPUT
re-incorporated till and
glaciofluvial sediments
Figure 13.5 Schematic diagram representing glacial sediment systems (after Derbyshire et al., 1979).
⫹9 ⫹8 ⫹7 ⫹6 ⫹5 ⫹4 ⫹3 ⫹2 ⫹1 0 ⫺1 ⫺2 ⫺3 ⫺4 ⫺5 ⫺6 ⫺7
100 100
90 90
80 80
70 70
PER CENT FINER
60 60
50 50
40 Iceland. Fluted lodgement tills (various sources) 40
France, La Dombes region. Drumlinized

30 lodgement till of penultimate glaciation. 30
(Billard & Derbyshire 1985)
20 Pakistan Karakoram, Hunza valley. 20
Lodgement tills. (Li et al., 1984)
10 United Kingdom, North Wales. 10

Drumlinized lodgement tills. (Carter, 1983)
0 0
0.002 0.006 0.01 0.02 0.06 0.1 0.2 0.6 1 2 6 10 20 60 100 200mm
FINE MEDIUM COARSE FINE MEDIUM COARSE FINE MEDIUM COARSE
(B.S) CLAY PEBBLES
SILT SAND GRAVEL BOULDERS
U.S.D.A U.S.D.A. SILT U.S.D.A. SAND U.S.D.A. GRAVELS
Figure 13.6 Particle size distribution envelopes for selected subglacial tills from low plateau (Iceland), piedmont
(France), low maritime mountains (North Wales, UK) and high continental mountains (Karakoram Mountains,
northern Pakistan).
deformed by the overlying ice and may become grain size range, being an important member of
entrained within the glacier. the group of multi-modal particle size sediments
The sediment produced by glacial erosion and commonly referred to as diamicts or diamictons.
transport characteristically displays a very broad Dreimanis and Vagners (1971) showed that, for
GFE-13.qxd 6/24/05 3:57 PM Page 351
a particular source rock, there is a limiting parti- deformation till. These tills are usually fine
cle size to which rock will be ground down. They grained, have high bulk densities and are gener-
referred to this as the terminal grade, meaning ally over-consolidated. Comminution till, a term
that, irrespective of the amount of time involved rarely used today, is also a product of subglacial
or the distance over which the glacial sediment is deformation; composed of crushed and powdered
transported, the particles cannot be ground down local bedrock, it has moderately high bulk densi-
below a distinctive terminal diameter. The larger ties and shear strengths (Elson, 1989), but may be
particles, however, become progressively smaller difficult to distinguish from deformation and
as the processing time and the distance of glacial lodgement tills. When rocks and sediments are
transport increase. Figure 13.6 shows the range of subglacially sheared, the resulting mélange is
variability of particle size distribution curves referred to as a glaciotectonite (Banham, 1977).
(expressed as envelopes) for subglacial tills from The process of meltout is one of direct depos-
selected environments. Benn and Ballantyne ition of sediments by melting of stagnant or very
(1993; 1994) showed how particle shape is indica- slowly moving debris-rich ice. This may occur
tive of the mode of transport and deposition. below, within, or above the glacier and sediments
Reworking of sediments by ice-marginal or deposited by these processes are known as sub-
proglacial processes following deposition may glacial, englacial and supraglacial meltout tills,
continue to modify the size distribution and shape respectively. During the meltout process, the sedi-
characteristics of a sediment. Despite these general ment mass may fail resulting in debris slides or
relationships, it may prove very difficult to pre- flows, materials sometimes referred to as flow tills.
dict the characteristic properties of a given till. However, the term debris flows is more appropriate
because it refers to the true mode of deposition
(Lawson, 1981). When debris-rich ice sublimates,
13.4 Depositional processes and the resultant sublimation till is formed and may
classification of tills inherit the foliated structure of the ice (Shaw, 1977).
These tills usually have low bulk densities, are
Entrained sediment will eventually be deposited poorly indurated and have coarser particular size
beneath, adjacent to or beyond the glacier margin. distributions than deformation and lodgement tills.
Sediment deposited directly by ice, without any The sedimentary characteristics of each type
appreciable meltwater action, is referred to as till of till are shown in Table 13.1. However, given the
and the landforms so produced are known as complexity of the multiple processes summarised
moraines. Numerous attempts have been made to above, the origin of a till cannot always be deter-
classify tills (Goldthwait and Matsch, 1989; mined unequivocally. Describing tills in terms of
Trenter, 1999), but the generally accepted classi- their lithologies provides a more objective
fications are based on the mode of deposition. Till approach. Miall (1978) and Eyles et al. (1983b)
may be deposited by lodgement, meltout, slide, provide a convenient framework for describing
flow, sublimation and deformation. The interrela- the lithologies of glaciogenic sediments, known
tionships between the main factors and param- as lithofacies coding (Table 13.2).
eters influencing subglacial erosion and deposition Different types of till may be deposited on top
are shown in Figure 13.7. of each other. For example, it is common for
Lodgement occurs when debris in the basal supraglacial till to be deposited upon deformation
layers of a sliding glacier is plastered on to a rigid till. Furthermore, other sediments may be inter-
or semi-rigid bed (Dreimanis, 1989). Sediment bedded into a glacial deposit such as glaciofluvial
accumulated by the lodgement processes is and glaciolacustrine sediments that may be laid
referred to as lodgement till. However, when down subglacially, englacially or supraglacially.
subsequently overridden, lodgement till behaves These sediments may be overridden by a glacier
as a deforming bed (Murray, 1997; Hindmarsh, and incorporated into a deformation till. The
1997) beneath a glacier, and is then referred to as different till types have different geotechnical
GFE-13.qxd 6/24/05 3:57 PM Page 352
A amount and
behaviour of
thermal
subglacial
(hydrological)
meltwater
ice-base regime
subgiacial
total melting
velocity conditions
of ice body
erosive
relief of capacty
ice-base/bedrock
interface conditions
for removal of
loose material
intensity
of ice-sheet
erosion
subglacial
local ice pressure
hydrology pattern velocity conditions
(channel and of internal
crevasse system) deformation
behaviour
of ice on
load
ice-bedrock
interface
ice thickness Value of

(zone of ice sheet) tangential
pressure bedrock sensitivity
lithology to erosion
B
thermal local
ice-bed fluvioglacial
subglacial
regime processes
thermal
conditions
ice-bed basal ice meltout

topography velocity processes
subglacial
stress
fluctuations
subgiacial debris depositional lodgement
hydrology rheology processes processes
systems
ice-bed flow
ice thickness interface(s) processes
conditions
deformation
debris and/or
availability ´
melange
processes
Figure 13.7 Models showing the interrelationship between main parameters influencing subglacial processes and
(A) erosion and (B) deposition (after Menzies and Shilts, 1996).
GFE-13.qxd 6/24/05 3:57 PM Page 353
Table 13.1 Some characteristics and properties of different types of till.
Till type Genesis Particle size Mesofabric Microfabric Bulk Shear

distribution density strength
Glaciotectonite Subglacially Poorly sorted Moderate Microshears Moderate Low

sheared
sediment
and bedrock
Comminution Subglacially Poorly sorted Moderate Microshears Moderate Moderate
crushed and skewed
powdered towards fine
local bedrock
Lodgement Subglacially Poorly sorted Strong Microshears High High
plastered glacial up-valley
debris on a rigid dip
or semi-rigid bed
Deformation Subglacially Poorly sorted Strong Microshears Moderate Moderate
deformed to high to high
glacial sediment
Meltout Glacial sediment Poorly sorted May be Microshears Low Low
deposited directly stratified
from melting ice
Sublimation Glacial sediment Poorly sorted Preserves Microshears Low Low
deposited directly ice foliation
from sublimated ice
Flow till* Sediment deposited Poorly sorted Downslope Few Low Low
off the ice by debris
flow processes
* This is classed as a debris flow rather than a till, but it is included here for comparison.
properties that are essentially a function of the Europe during the glaciations of the Quaternary
depositional environment. In particular, the plas- (Figure 13.1a). As a result, characteristic suites of
ticity and particle size of tills are critically depend- glacial landforms and sediments dominate exten-
ent on the mode of deposition and the nature of the sive areas of these landscapes. Studies of the ice
terrain over which the glacier or ice sheet was sheets in Greenland and Antarctica have provided
active (Trenter, 1999). Predicting the vertical and some understanding of the nature of ice-sheet
lateral variability of the geotechnical properties is dynamics, although it is important to recognise
extremely difficult when a glacial deposit consists that these modern ice masses are not perfectly
of several types of till, as may be seen from the analogous to the larger Quaternary ice sheets
study of modern ice margins (Figure 13.8). that formerly covered much of the northern
hemisphere.
The variability in the environments produced
13.5 Contemporary glacial by the Quaternary ice sheets is illustrated in the
environments: landsystem landsystems and lithofacies models shown in
models Figures 13.9 to 13.12. Many of these landforms
were produced subglacially beneath several
Ice sheets kilometres of glacial ice. Subglacial deposition
Ice sheets extended over much of the high and and deformation produce elongated streamlined
mid-latitude regions of North America and landforms (Figures 13.9 and 13.10). These range
GFE-13.qxd 6/24/05 3:57 PM Page 354
Table 13.2 Simplified lithofacies codes (adapted in size from a few metres wide and long (flutes)
from Miall, 1978 and Eyles et al., 1983b). to tens of metres across and high, and hundreds
of metres long (drumlins) to many hundreds of
Code Description
metres across and tens of metres high and kilo-
Diamictons Very poorly sorted admixture of wide metres long (megaflutes). Their internal structure
particle size range and composition can be complex, comprising
Dmm Matrix-supported, massive
tills, glaciofluvial and glaciolacustrine sediments.
Dcm Clast-supported, massive
Dcs Clast-supported stratified
Sinuous ridges (eskers), tens of metres high and
Dms Matrix-supported stratified many kilometres in length represent subglacial
Dml Matrix-supported laminated glaciofluvial channels that were infilled with sedi-
- - - (c) Evidence of current reworking ments and left behind as the ice sheets melted.
- - - (r) Evidence of resedimentation The internal sediments and structures within
- - - (s) Sheared
these landforms are highly complex, often includ-
Boulders Particles ⬎256 mm (b-axis) ing a whole range of lithologies, bedding types
Bms Matrix-supported, massive
and deformation structures. Stagnating ice left
Bmg Matrix-supported, graded
Bcm Clast-supported, massive during glacial retreat also often gives rise to a
Bcg Matrix-supported, graded suite of complex landforms and sediment types
Gravels Particles of 8–256 mm that include hummocky moraines, kettle holes
Gms Matrix-supported, massive and kame terraces (Figure 13.11). Large
Gm Clast-supported, massive proglacial lakes are also associated with ice
Gh Horizontally bedded sheets. Again, the sediments and landforms in
Gt Trough cross-bedded these lacustrine environments may be very
Gp Planar cross-bedded
complex. Large supraglacial lakes may develop,
Granules Particles 2–8 mm
particularly when the ice has retreated into moun-
GRcl Massive with clay laminae
GRch Massive and infilling channels tainous terrain. A landsystems model for such a
GRh Horizontally bedded supraglacial lake is shown in Figure 13.12. The
GRt Trough cross-bedded landforms along former ice margins may also be
GRcu Upward coarsening further complicated by deformation during gla-
GRfu Upward fining cial advances. The advancing glacier may over-
Sands Particles of 0.063–2 mm ride or push previously deposited sediment so as
St Medium to very coarse and trough to form folded and faulted strata. Predicting the
cross-bedded
Sp Medium to very coarse and planar
internal structure of these landforms is extremely
cross-bedded difficult because of their complexity. In the first
Sm Massive stages of analysis, therefore, each individual land-
Sr Ripple cross-laminated form should be considered as a unique entity.
Sh Very fine to very coarse horizontally/
plane bedded or low angle cross- Valley glaciers
lamination
Sc Steeply dipping planar cross-bedding
Valley glacier landsystems are highly variable and
Sd Deformed bedding may include glaciers with limited debris cover,
Suc Upward coarsening debris-mantled glaciers, rock glaciers and hanging
Suf Upward fining glaciers (Benn et al., 2003). Boulton and Eyles
Srg Graded cross-laminations (1979) provide a model to describe a valley glacier
Silts & clays Particles of ⬍0.063 mm landsystem for a glacier with moderate amounts of
Fl Fine lamination often with minor fine debris cover. However, this model cannot be fully
sand and very small ripples
applied to valley glaciers having substantial
Fm Massive
Flv Fine lamination with rhymites or varves amounts of surface debris. Such debris-mantled
Frg Graded and climbing ripple glaciers produce large lateral-frontal ‘dump’
cross-laminations moraines, as well as ice-contact fans and ramps
GFE-13.qxd 6/24/05 3:57 PM Page 355
Figure 13.8 Complex interrelationship between sedimentary facies at the terminus of the 59 km long Batura
Glacier, Karakoram Mountains, Northern Pakistan. ‘GFL’ indicates glaciofluvial and ‘GL’ glaciolacustrine facies.
The scale bars indicate metres. After Li et al. (1984).
(Small, 1983; Owen and Derbyshire, 1989; 1993; marine sedimentation. A schematic figure showing
Owen, 1994). These glacial environments are the major glaciomarine enviroments is presented in
characterised by the incorporation of ice-marginal, Figure 13.13. Here, sediment is transferred from
supraglacial, subglacial, proglacial, periglacial and supraglacial and subglacial pathways into the
paraglacial sediments and landforms, recording the marine environment by density currents, mass
juxtaposition and migration of very different movements and icebergs. These sediments may be
depositional environments. When a rock glacier deformed by grounded and advancing ice. Study of
becomes progressively buried by debris, it may glacial processes within glaciomarine environments
eventually form a glacial rock glacier comprising is clearly difficult because of their inaccessibility.
a core of deforming ice beneath thick debris (Benn However, the geometry and structure of offshore
and Evans, 1998). The resultant form is a steep glaciomarine sediments are somewhat easier to
sided tongue-like or series of lobate masses with study than terrestrial deposits because geotechnical
ridges, furrows of coarse debris that advance techniques such as seismic reflection profiling can
downslope as a consequence of internal deforma- be more readily applied in the marine realm. Two
tion of the ice at its core. Hanging glaciers occur main glaciomarine settings can be recognised
in steep terrain in which the glacier tongues are (Hambrey, 1994; Trenter, 1999). The first is within
perched in hanging fashion high above valley fjords where sediments are supplied from glaciers
floors. Benn et al. (2003) provide models showing in the form of ice-contact deposits, glaciofluvial
the details of glaciated valley landsystems, and deposits and ice-rafted debris. Clastic and biogenic
highlight the importance of the strong influence of marine sediments are supplied together with fluvial,
topography on glacier morphology, sediment mass movement and aeolian deposits from the
transport paths and depositional basins, as well as terrestrial environment. The second setting com-
the importance of debris from supraglacial sources prises continental-shelf and open-ocean environ-
in the glacial sediment budget. ments where sedimentation is dominated by
grounded ice-margins, floating ice-tongues, ice-
Glaciomarine environments shelves and marine processes. Sediments are
The glaciomarine environment is one of the supplied mainly from subglacial sources when
most complex, involving subglacial, subaerial and debris is exposed at grounding lines with minor
GFE-13.qxd 6/24/05 3:57 PM Page 356
Figure 13.9 (A) Subglacial landsystem in an area of hard substrate (after Eyles, 1983). (1) abraded and stream-
lined rock knobs; (2) Basal debris; (3) lodgement till on low-relief rock surface; (4) lee-side cavity fill; (5) basal
melt-out till; (6) debris melting out at ice surface and dumped by gravity on freshly exposed subglacial surface;
(7) subglacial esker with gravel core; (8) hummocky or kettled outwash surface produced by the melt-out of ice
buried by outwash fans; (9) the proglacial stream carries subglacial abrasion products. (B) The subglacial landsys-
tem in an area with low-relief limestone terrain and where multiple stacked till units have been deposited during
a single glaciation (after Eyles, 1983; Benn and Evans, 1998). This subglacial landsystem is reworked by:
(I) hummocky kame-and-kettle topography; (II) outwash cut into the till surface and comprising stratified sands and
GFE-13.qxd 6/24/05 3:57 PM Page 357
Figure 13.10 Idealised zones of glacial landforms and sediments on the prairies of North Dakota. (A) Proglacial
suite; (B) supraglacial suite; (C) transitional (submarginal) suite; (D) subglacial suite. (a) Proglacial lake; (b)
proglacial meltwater channel; (c) subglacial meltwater channel or tunnel valley; (d) ice-walled lake plain; (e) esker;
(f) transverse thrust moraines cupola hills; (g) prairie mound; (h) flutings; (i) transverse recessional moraines; (j)
hummocky moraine; (k) isolated kames. (Modified from Clayton and Moran, 1974).
inputs from supraglacial debris and debris trans- this limit, the sediments are not easily recognisable
ferred from the ice to the marine environment, by as glaciogenic in origin and may comprise lami-
way of ice shelves, ice cliffs, outlet glaciers and nated sediment with only occasional dropstones,
ice streams. At the terrestrial-marine boundary, a and debris flow dominated turbidites.
distinct set of ice-marginal landforms and deposits
develop that are easily recognisable as glaciogenic Glaciofluvial
in origin. These include stratified diamict (waterlain Water drains through the glacial system in
till), as well as laminated sediments both with and three main ways: supraglacially, englacially and
without ice-rafted debris, depending on proximity subglacially. This glaciofluvial water is usually
to the ice front when deposition took place. Beyond concentrated within channels that are thermally
gravels; (III) esker deposited during ice wastage and therefore not truncated by subglacial tills like other channel
fills in the subglacial landsystem. The base of the subglacial landsystem is characterised by: (1) striated rock head;
(2) buried channel/valley with a fill of subglacial sands and gravels and till; (3) glacitectonised rockhead, with rock
rafts and boulder pavements; (4) lowermost till, comprising local lithologies, which thickens in the lee sides of rock
protuberances as lee-side cavity fills; (5) cold-water karst. The sediments of the subglacial landsystem are charac-
terised by: (a) predominantly preferentially aligned, faceted clasts; (b) crude shear lamination produced by the
smearing of soft lithologies (deformation till/glaciotectonite); (c) slickensided edding planes (fissility) produced by
glacitectonic shear; (d) stratified gravels, sands and clays deposited in subglacial cavities, pipes or canals and trun-
cated by overlying till (the base of each till unit may be fluted) — they constitute lenses which are elongated in the
direction of ice flow and typically internally disturbed by folding and faulting due to post-depositional deformation
by glacier/till overriding; (e) folded and sheared-off channel fill; (f) diapiric intrusion of till squeezed up into sub-
glacial cavity; (g) vertical joints produced by post-depositional pedogenic processes; (h) drumlinised surface of
upper till sheet; (i) inter-drumlin depressions filled with post-glacial solifluction debris and peat. The horizontal
scale may range from 10 m to 10 km and the vertical scale may be 10 cm to 100 m high.
GFE-13.qxd 6/24/05 3:57 PM Page 358
eroded into the ice, commonly opening and clos-

ing in response to the stresses associated with
flowing ice. This makes difficult the prediction of
the points where meltwaters will exit from the
glacier, a fundamental problem in reconstructing
the glacial hydrology. Some subglacial channels,
however, may be eroded into basal till or bedrock
to form deep channels that may fill with glacioflu-
vial sediment. The waters that eventually exit the
glacier are usually characterised by highly variable
discharges that fluctuate diurnally and seasonally
in response to variations in the glacier’s mass bal-
ance. Glaciofluvial waters usually have high sedi-
ment loads, typically containing abundant silt,
which is transported and held in suspension, as
well as boulders and gravel in the traction load.
Much of this sediment is carried to the proglacial
environment where it is deposited in braided chan-
nel systems that may form extensive outwash fans
(tens of kilometres in extent), referred to as san-
durs. Some glaciofluvial sediment, however, may
be deposited within the glacier as metre-size chan-
nel fills associated with subglacial tills. Such
deformed channel-fill material sometimes forms
part of deformation tills. When glaciofluvial sedi-
ment accumulates so as to fill large subglacial or
englacial ice channels, long (100s to 1000s of
metres) sinuous ridges of coarse gravels and boul-
ders with steep cross bedding may be preserved as
the ice retreats. These distinctive landforms,
known as eskers, are often an important aggregate
resource. Glaciofluvial sediments may also be
deposited on or along the margins of the ice and
they may form impressive terraces when the ice
eventually melts. The bedding within these accu-
mulations may be deformed and slumped when ice
within or beneath melts out, leading to bowl-
shaped, collapse depressions called kettle holes.
Figure 13.11 Evolution of glacier karst on an ablating Numerous lithofacies models have been produced
debris-mantled glacier. (A) Young stage; (B) mature stage;
(C) old stage. (1) Debris bands in glacier ice; (2) ice- to describe the range of sediment types and associ-
cored ridge; (3) depositional trough; (4) back-wasting ations found in glaciofluvial terrains, a good
slope of exposed ice; (5) debris flow; (6) enlargement of review being that by Maizels (1996).
crevasse by ice melt; (7) subglacial conduit; (8) sink
holes; (9) collapsed tunnel roof; (10) enlargement of sink- Glaciolacustrine
hole by ice melt and collapse; (11) lake with back-
wasting margins; (12) dead ice; (13) ice-free hummocky Several different types of lakes are commonly
terrain; (14) deposits of supraglacial environments; associated with glacial environments, occurring
(15) lakes; (16) deposits of subglacial environments. in subglacial, supraglacial and proglacial settings.
(Redrawn from Krüger, 1994). Supraglacial lakes are often found on glaciers that
GFE-13.qxd 6/24/05 3:57 PM Page 359
Figure 13.12 Model for sedimentation in a supraglacial lake system in an area of moderate to high relief, based
on the stratigraphy of the Fraser River valley, British Columbia, Canada. (From Eyles et al., 1987. Reproduced by
permission of Blackwell Ltd.)
are in retreat. These may drain catastrophically result in abundant load structures and slumping.
as the ice thins and as channels open within the All these lakes and the resultant sediments are
glacier. Subglacial lakes are common under ice intimately associated with glaciofluvial systems;
sheets, surging glaciers, and where glaciers over- in fact, extreme lateral and vertical variation may
lie volcanic centres such as in Iceland. These be observed within glaciolacustrine successions.
lakes may also drain catastrophically as new Ashley (1996) provides a summary of the main
drainage conduits open and lead, through the environments and lithofacies associated with
glacier, to the ice margin. The high discharges glaciolacustrine environments.
result in the resedimentation of large quantities
of glaciogenic sediment to form flood deposits Glacioaeolian
in the proglacial environment. Proglacial lakes Strong winds are associated with glacial environ-
form within moraines and behind moraine ments because of the high-pressure systems that
ridges, within ice-marginal valleys, and within develop over glaciated regions. Severe cooling of
bedrock depressions (Figure 13.14). In all these air above a large valley glacier may give rise to air
lakes a complex assemblage of sediments may drainage closely constrained by high and steep val-
be deposited including fan deltas, turbidites, ice- ley walls. Such katabatic winds sometimes reach
rafted debris and laminated deposits. Deform- very high velocities and thus are capable of remov-
ation structures are common within many of these ing surface sands and silts well beyond the glacier
sediments, particularly if buried ice or ice adjacent snout. Thus, the finer fractions of the abundant
to the lake melts out. This causes slumping and debris produced by glacial abrasion and cryogenic
faulting. Furthermore, sediments of contrasting (freeze–thaw) weathering contribute to aeolian sed-
densities, deposited on top of each other, frequently imentation at a range of scales from local sand and
GFE-13.qxd 6/24/05 3:57 PM Page 360
Figure 13.13 Schematic figure illustrating the major environments for glaciomarine sedimentation in a cool temper-
ate climate, based on the present north-north-west Gulf of Alaska (after Powell and Domack, 1996). (1) Tidewater ter-
minus; (2) side drainage with gravel pocket beach and talus fan on fjord floor; (3) bergstone mud on glacial fjord floor;
(4) sandur plain in marine outwash fjord; (5) sediment gravity flow channels in marine outwash mud; (6) entrance sill
capped with morainal banks and grounding-line fan; (7) continuation of outwash mud; (8) rocky shore with gravel
pocket beaches; (9) tectonically uplifted banks of older rocks exposed by winnowing on the shelf; (10) extensive tidal
mudflat in estuary; (11) recurved and cuspate spits from alongshore transport; (12) marine outwash mud on the conti-
nental shelf; (13) relict sand deposits; (14) thick temperate rain forest of spruce/hemlock with muskeg swamps;
(15) raised marine terrace; (16) aeolian dunes fed from spit at wave-dominated delta; (17) large slide/slump areas on
the continental shelf; (18) ice-dammed lake with sublacustrine grounding-line fan and lake laminites with iceberg-
rafted debris; (19) terminal moraine and gravelly beach; (20) sandur/delta system with small moraines and gravelly
beach; (21) modern littoral sand; (22) bank on continental shelf of relict glacier-marginal sediment, currently being
winnowed; (23) tidewater terminus just at sea level with short-headed stream, fan deltas; (24) river-dominated estuary
with stunted barrier island system offshore; (25) sea valley on continental shelf being infilled with marine outwash mud.
silt drapes on moraines to extensive mantles of by aeolian processes, have deflated surfaces, and on
regional extent (typically grading with distance which are found ventifacts and sand dunes. The sec-
from the front of an ice sheet from sands to pro- ond setting is generally beyond, but adjacent to the
gressively finer silts known as loess: Chapter 25). Pleistocene glacial limit, and is characterised by
Derbyshire and Owen (1996) describe these extensive sheets of sands often referred to as cover-
processes and highlight the polygenetic nature of sands. The third setting is often a considerable dis-
such glacioaeolian sediments. Three main glacioae- tance from the Pleistocene glacial limit and is
olian settings exist. The first occurs adjacent to con- characterised by massive loess deposits. Many of
temporary glaciers and within the approximate the silts that comprise loess are thought to derive
maximum limit of Pleistocene glaciation. This is from glacial flour (silts produced by glacial grind-
characterised by landforms that have been eroded ing and carried and deposited by glaciofluvial
GFE-13.qxd 6/24/05 3:57 PM Page 361
as a suite of hydrological changes arising from

movement and storage of glacial meltwaters.
In formerly glaciated terrain, factors of major
significance include the bulk properties of the
principal sedimentary facies (glaciogenic, glacio-
fluvial, glaciolacustrine, glaciomarine, glacioa-
eolian), the degree of alteration by processes
including weathering and mass movements, the
hydrological characteristics of the deposits, and
the erosional and depositional landforms.
Glacial advance and retreat, shifting

Figure 13.14 Partially floating tongue of Breidamer- meltwater river regimes, and glacial lake
kurjökull, an outlet glacier of the Vatnajökull ice sheet,
south-east Iceland, showing large ice-front crevasses, outburst floods: hazards and mitigation
calving, and the resulting icebergs in the proglacial lake Tufnell (1984) regarded the three main types of
Jökulsárlon. glacial hazards as expanding glaciers, glacial
lake outburst floods, and ice that breaks off from
glacier snouts and avalanches into valleys,
streams) and then blown considerable distances destroying all artificial objects in its path.
beyond the ice margins. The loess covering the The advance and retreat of glaciers often con-
plains of Northern Europe and the northern USA stitutes a substantial hazard, especially in moun-
are thought to be largely of this origin. However, the tain regions with low, but locally concentrated
silt making up the loess in some other regions of the human populations. Advancing and retreating gla-
world, such as North Africa, may be derived from ciers may directly cause important landscape
non-glacial sources. The great loess deposits of changes by re-shaping mountain slopes, deposit-
northern China, although undoubtedly related to the ing large morainic accumulations, and diverting
waxing and waning of the great Pleistocene ice and damming natural drainage. Indirect effects of
sheets, are a dryland periphery accumulation of ice advance and retreat include important changes
complex origin. in the local and regional climate and ecology,
which, in turn, have a bearing on hydrology and
slope stability. Changes in the position of ice
13.6 Engineering in glacial fronts may cause several types of direct and sub-
environments stantial damage to important human structures
including highways (Derbyshire et al., 2001),
From an engineering point of view, it is use- settlements, hydro-electric plants, irrigation chan-
ful to distinguish between glacierised terrain nels and fields. Well-known examples include the
(those occupied by active glaciers) and glaciated advance of the Mer de Glace in Chamoix,
landscapes (those bearing the marks of past Switzerland, between 1644 and 1919 (threatening
glaciation). The problems posed by the presence the village of Les Bois) and the advance of the
of glaciers differ in certain obvious ways from Pasu Glacier in the Hunza valley (Karakoram
those arising from engineering works on glacially Mountains of northern Pakistan) that threatened
eroded rock surfaces and the associated and often Pasu village in 1913 (Figure 13.15; Tufnell, 1984).
diverse range of ancient glacial sediments. In the Fluctuating meltwater rivers on glacial forelands
case of terrain close to the margins of glaciers and can cause considerable damage to constructions
ice sheets, the dynamic behaviour of glaciers in such as bridges, roads and cultivated fields. Many
response to several forcing factors, notably historical changes of this type have been docu-
climatic variations, results in relatively short-term mented for the Icelandic glaciers including, for
mass motion (glacier advance and retreat), as well example, the foreland of Hoffellsjökull Glacier
GFE-13.qxd 6/24/05 3:57 PM Page 362
Figure 13.16a (Walder and Costa, 1996). Glaciers

carrying large volumes of supraglacial debris, such
as those in parts of the Himalaya (e.g. Owen et al.,
2003), pose a particular threat to human commu-
nities because such debris is usually coarse and so
may be breached only a short time after formation
of the lake, the process sometimes being cata-
strophic. The Pokhara basin in Nepal was cata-
strophically flooded by a glacier lake outburst
flood through a breached ice-cored moraine dam
about 450 years ago, the resulting lake attaining an
area of 10 km2 (Yamada, 1993). Rapid release
of about 5 million m3 of rock debris into the
proglacial Damenlahai Lake, Tibet, in 1964 trig-
gered a breach of the end moraine, the resulting
flood being estimated at over 3500 million m3; the
debris carried by this flood buried twelve houses
and over 2 km of the Sichuan–Tibet highway
(Derbyshire, 1990). Richardson and Reynolds
(2000) have recently provided a schematic dia-
gram of a hazardous moraine-dammed glacial lake
based on examples from Bhutan (Figure13.16b;
Table 13.3). Glaciers with a thick debris cover
often develop supraglacial ponds and small lakes
(Figure 13.17); in exposing bare ice cliffs by
enhanced melting, these supraglacial water bodies
Figure 13.15 The debris covered snout of the Pasu
coalesce, leading to rapid growth of ice marginal
Glacier (upper right centre), Karakoram Mountains of lakes (Kirkbride, 1993; Benn et al., 2000). Large
northern Pakistan, in July 1988. The proglacial lake, fed glaciers not in equilibrium with the present climate
by a subglacial stream (see exit tunnel: right centre) is may constitute a heightened risk if the current
held back by a complex set of end-moraine ridges warming trend continues (Naito, 2000). Methods
(centre) left in 1981 as the glacier retreated. The moraine
complex is being eroded by the lake-outlet stream (left
of hazard assessment relevant to such situations
centre). The village of Pasu is 2 km downslope of this site. have been applied relatively rarely in major moun-
tain ranges such as the Himalaya, however.
Mitigation measures, including progressive lower-
(Tufnell, 1984). The destructive effect of shifting ing of lake levels by siphoning, have met with
meltwater channels also occurs on a smaller scale, some success in Nepal and Peru (Reynolds, 1998).
though often with a major impact on routeways and
settlements. For example, the main outlet stream of Building on glacial deposits: foundation
the Ghulkin Glacier, Karakoram Mountains, problems both onshore and offshore
breached the Karakoram Highway in the summer of Any construction on glacial deposits requires a
1980, burying it under a new gravel fan c. 0.5 km consideration of the geotechnical properties,
across (Derbyshire and Miller, 1981; Jones et al., hydrology and variability of the sediments. This is
1983). particularly important for large structures such as
Glaciers and their moraines frequently provide dams, oil and gas platforms and industrial com-
temporary meltwater dams in mountain valleys. plexes. Derbyshire (1992) provided examples of
Two styles of glacier-dammed lakes and their some of the geomorphological and engineer-
associated breaching mechanisms are shown in ing considerations encountered during the site
GFE-13.qxd 6/24/05 3:57 PM Page 363
Figure 13.16 (A) Two styles of ice-dammed lakes and associated breach mechanisms (after Walder and Costa, 1996).
Figure 13.16 (B) Schematic diagram of a hazardous moraine-dammed glacial lake (after Richardson and Reynolds,
2000). The geologic successions include (1) bedrock; (2) receding glaciers; (3) stacked subglacial tills and
(4) supraglacial tills representing multiple glacial advances; (5) complex inter-bedded glaciofluvial sediments,
supraglacial tills and gravity flow diamicts in terminal and lateral moraine ridges; (6) stagnant glacier ice with; (7) basal
meltout till; (8) valley-side fan deposits; (9) hummocky moraine resulting from melt of ice-cores (thermokarst), and
(10) lake sediments. Factors contributing to the hazard risk include (a) large lake volume, (b) narrow and high moraine
dam, (c) stagnant glacier ice within the dam, (d) limited freeboard between the lake level and crest of the moraine ridge.
Potential outburst flood tiggers include avalanche displacement waves from (A) calving glaciers (B) hanging glaciers,
and (C) rock falls; (D) settlement or piping within the dam (due to progressive seepage or seismic activity); (E) melting
ice-core; (F) catastrophic glacial drainage into the lake from sub and englacial channels or supraglacial lakes.
GFE-13.qxd 6/24/05 3:57 PM Page 364
Table 13.3 Types of glacial and glacially related hazards (adapted from Richardson and Reynolds, 2000).
Category Hazard event Description Time scale
Glacier hazards Avalanche Slide or fall of large mass Minutes

of snow, ice and/or rock
Glacier outburst Catastrophic discharge of Hours
water under pressure from a glacier
Jökulhlaup Glacier outburst associated with Hours–days
sub-glacial volcanic activity
Glacier surge Rapid increase in rate of glacier flow Months–years
Glacier fluctuations Variations in ice front positions due to Years–decades
climatic change, etc.
Glacier hazards Glacier lake outburst Catastrophic outburst from a Hours
as above, plus: Floods (GLOFs) proglacial lake, typically moraine
dammed
Débâcle Outburst from a proglacial lake (French) Hours
Related hazards Lahars Catastrophic debris flow Hours
associated with volcanic activity
and snow fields
Water resource Water supply shortages, particularly Decades
problems during low flow conditions,
associated with wasting glaciers
and climate change etc.
exploration stages of the Brenig dam site on the

drumlins in the Denbigh Moors in North Wales
and the Kielder dam site on tills along the North
Tyne River, in northern England. He emphasised
the variability of sediment types and geotechnical
properties within the proposed dam sites.
Furthermore, he showed that microstructure of
tills may greatly influence the mechanical proper-
ties and loading characteristics (e.g. Boulton and
Paul 1976; McGown and Derbyshire 1977).
The geotechnical properties of glacial sediments
are usually a function of the depositional environ-
Figure 13.17 Aerial view northwards across the ter- ment (Table 13.1). Farrell et al. (1995) showed that
minal zone of the Hooker Glacier, the south island of weathering may also alter the geotechnical proper-
New Zealand, in 1974. The debris-covered glacier ties of tills (Figure 13.18). In particular, weathering
shows evidence of a complex hydrology. This includes
several supraglacial lakes (each sealed off from the
increases the fines content and the plasticity of till,
next, as indicated by the range of water levels); ice- reducing its density and shear strength. The typical
contact (proglacial) lake (foreground); supraglacial ranges of geotechnical properties for different tills
streams, englacial streams (lower left of photograph); are illustrated in Tables 13.4 and 13.5 and Figure
subglacial and proglacial streams (right of photograph). 13.19. Hird et al. (1991) and Lehane and Faulkner
Parallel lateral moraine series on both side slopes of the
glacier (top and bottom of photograph) show varying
(1998) provide useful studies of the stiffness and
degrees of eroded incision with associated debris strength characteristics of tills.
cone/fan development, one fan constraining the limits Depositional environment and pre- and
of the proglacial lake (lower right of photograph). post-depositional process history are important
GFE-13.qxd 6/24/05 3:57 PM Page 365
Figure 13.18 Variation in the geotechnical properties through (A) a lodgement till from Northumberland, England
(after Eyles and Sladen, 1981) and (B) a weathered till (after Farrell et al., 1995). See Table 13.5 for details of the weath-
ering zones.
influences on the geotechnical properties of These include: the misidentification of rockhead;

glacial sediments. Many glacially deposited sedi- the presence of large boulders; the presence of
ment types have a multimodal or bimodal grain glaciolacustrine laminated silty clays; water-
size distribution that includes fines (notably silts bearing silts and sands and water-bearing bedrock.
but often with some clay grade material: Figure These features may increase the risk of slope fail-
13.6). These fines, whether making up the matrix ure and subsidence and they can often substantially
of tills or as relatively uniform fine-grained beds increase the cost of protective measures to assure
such as glaciolacustrine deposits, transfer stresses the safety and longevity of the construction or pro-
applied by glacial and other sources of loading, a ject being undertaken. Furthermore, Trenter (1999)
process often intricately preserved in their micro- highlighted the following characteristics of glacial
fabric properties. Examples of some common sediments that influence both foundation design
microfabric types are shown in Figures 13.20, and construction:
13.21 and 13.22.
Trenter (1999) emphasised several prob- (1) spatial variations in soil type, exhibited as
lems encountered when excavating in till for an changes in strength, compressibility and con-
engineering development or construction project. solidation properties
GFE-13.qxd 6/24/05 3:57 PM Page 366
Table 13.4 Strength of tills of north Norfolk and Table 13.5 Typical geotechnical properties for
Holderness (UK) (from Bell, 1993). Northumberland (UK) lodgement tills (after Eyles and
Sladen, 1981). See Figure 13.8b.
Unconfined compressive
strength (kPa) Property Weathering Zones
Intact Remoulded Sensitivity I III & IV
North Norfolk Bulk density (kg/m3) 2150–2300 1900–2200
1. Hunstanton till (Holkham) Natural moisture 10–15 12–25
Max. 184 164 1.22 (L) content (%)
Min. 152 128 1.18 (L) Liquid limit (%) 25–40 35–60
Mean 158 134 1.19 (L) Plastic limit (%) 12–20 15–25
2. Marly drift (Weybourne) Plasticity index 0–20 15–40
Max. 120 94 1.49 (L) Liquidity index ⫺0.20 III ⫺0.15 to
Min. 104 70 1.28 (L) to ⫺0.05 ⫹0.05
Mean 110 81 1.34 (L) IV 0 to ⫹30
3. Contorted drift (Trimingham) Grading of fine (⬍ 2 mm) fraction
Max. 180 168 1.67 (L) % clay 20–35 30–50
Min. 124 76 1.08 (L) % silt 30–40 30–50
Mean 160 136 1.23 (L) % sand 30–50 10–25
4. Cromer till (Happisburgh) Average activity 0.64 0.68
Max. 224 188 1.19 (L) c (kPa) 0–15 0–25
Min. 154 140 1.10 (L) ⌽ (degrees) 32–37 27–35
Mean 176 156 1.13 (L) ⌽r (degrees) 30–32 15–32
Holderness c ⫽ cohesion.
1. Hessle till (Dimlington, Hornsea) ⌽ ⫽ angle of internal friction.
Max. 138 116 1.31 (L) ⌽r ⫽ angle of internal friction (residual).
Min. 96 74 1.10 (L)
Mean 106 96 1.19 (L)
These difficulties arise from:
2. Withernsea till (Dimlington)
Max. 172 148 1.18 (L) (1) the presence of several different sediment
Min. 140 122 1.15 (L) types in the tunnel face
Mean 160 136 1.16 (L) (2) variations in the thickness of sediments and
3. Skipsea till (Dimlington) the depth to rockhead
Max. 194 168 1.15 (L) (3) the possibility that nests of cobbles and boul-
Min. 182 154 1.08 (L) ders may be present
Mean 186 164 1.13 (L) (4) varying groundwater conditions, including
4. Basement till (Dimlington) artesian pressures.
Max. 212 168 1.27 (L)
Min. 163 140 1.19 (L) When undertaking any substantial engineering
Mean 186 156 1.21 (L)
project in glaciated or formerly glaciated terrain, it
L ⫽ low sensitivity. is advisable to undertake a comprehensive site
investigation. Boone et al. (1998) emphasised that
(2) variation in depth to rockhead, with conse- one of the major problems in underground excav-
quential variation in thickness of overlying ations and piling in tills is the prediction of the
compressible soils likely size and frequency of the occurrence of boul-
(3) the presence of groundwater, the spatial vari- ders on the basis of the very limited sampling of
ation in sediment type, and the variation to the ground that takes place in a conventional site
rockhead (and hence compressible thickness). investigation. Preliminary investigations should,
GFE-13.qxd 6/24/05 3:57 PM Page 367
ral exposures, and cable percussion or rotary core

boring methods. The different sedimentary units
should be sampled once the variability of sediment
types has been determined. Sampling methods
include tube sampling, rotary methods and bulk
and block sampling. The extracted samples can be
used to determine the geotechnical properties of
the sediments including bulk density, moisture
content, particle size distribution, liquid and plastic
limits, compressive strength, coefficient of consol-
idation, and shear strength. More comprehensive
data profiles include the addition of in situ tests,
which are particularly valuable because they are
more representative of true site conditions. They
may include standard and cone penetration tests,
permeability tests and pressure-metre tests.
Fookes (1997) has stressed the importance of
using systematic site investigation to construct
landsystem models. Not only does this provide a
means of minimising uncertainties concerning
ground conditions in engineering sites, it is likely
to yield improved basic data and enhanced design
quality. Given the state of knowledge of glacial
landsystems, landsystem models are capable of
further development, according to Fookes, by
organising information on a site as a geological
environment matrix, consisting of tabulated geo-
logical data susceptible to computer manipulation
from which to derive predictive models of ground
conditions. However, a number of steps have yet
to be taken to render such data readily applicable
as part of a standard approach.
Water resources and pollution in

glacial sediments
An understanding of the nature of Quaternary sed-
iments is important for assessing the potential for
groundwater resources and predicting the nature of
groundwater pollution. Most aquifers within
Figure 13.19 (A) Some bulk properties of subglacial
and supraglacial tills (after Lutenegger et al., 1983). Quaternary glacial deposits occur within proglacial
(B) Effects of fissuring on shear strength (after McGown sands and gravels. The geometry of these aquifers
et al., 1974). is a function of the depositional setting and the
landforms. Buried glaciofluvial channels and
eskers, for example, form substantial elongated
therefore, include strata definition to characterise aquifers. The lithology and structure of the glacial
the vertical and lateral variation in sediment type. deposit affects its ability to transmit and store water
This should include geomorphic mapping the and pollutants. Daly (1992) provided a useful clas-
excavation of trial pits or the examination of natu- sification of pollution vulnerability rating for
GFE-13.qxd 6/24/05 3:57 PM Page 368
Figure 13.20 (A) Mildly over-consolidated clay-rich Upper Pleistocene lodgement till from Cowden, Holderness,
north-east England, showing compact fabric (bulk density c. 2100 kg/m3). Ice moved from left to right. Field of view
is 300 ␮m wide. (B) Over-consolidated clay-rich till of the penultimate Pleistocene glaciation at Boulton Moor, near
Derby, northern England. The dense matrix fabric is cut by shear systems, rising from left to right (in the same sense
as the glacier thrust).
GFE-13.qxd 6/24/05 3:57 PM Page 369
Figure 13.21 (A) Pleistocene subglacial meltout till of the penultimate European glaciation, from Pouilleux,
Rhône valley, south-east France. Relatively high voids ratio is evident from the loose fabric of this coarse-grained
matrix. The coarser silts and sand grains are often coated with fine clays as a result of meltwater eluviation.
(B) Modern meltout till from beneath the margin of Blåisen, southern Norway, released from the glacier in 1965.
The loose and fragile fabric is a product of observed meltwater eluviation. Some of the silt grains show a capping
of very fine clay grade material, attributed to the same process. Scale bar is c. 50 ␮m.
GFE-13.qxd 6/24/05 3:57 PM Page 370
A D
E
B E
Figure 13.22 Micrographs of two British Pleistocene glaciolacustrine sediments. (A) Sub-horizontal fabric in silts
rich in micas from a depth of c. 47 m within a finely laminated deposit. (B) Very uniform grain size of apparently
massive silts with few, barely perceptible laminae: depth 42 m. (C) and (D) Enlarged views of silts shown in
B, revealing an intimate mixture of medium and fine silts, scattered throughout. Fairly compact fabric, with variable
grain anisotropy owing to abundant clusters of edge-to-face fabric elements in the medium silts. (E) Sand grain
‘dropstone’, showing evidence of disruption of the silt fabric (indicated by vertically aligned fabric zone above and
to the right of the sand grain). Width of photograph is 1 cm. (F) Dropstone consisting of a small pod of till matrix,
disrupting the laminae in ice-proximal lacustrine silts: depth 12 m.
Micrographs A–E are from a core section from Pentre, Shropshire, England. Micrograph F is from a core at Jarrow,
northern England.
GFE-13.qxd 6/24/05 3:57 PM Page 371
Table 13.6 Vulnerability rating for typical Irish geohydrological settings (after Daly, 1992).
Vulnerability Rating Geohydrological setting
Extreme 1. Outcropping bedrock aquifers (particularly karst limestone) or where

overlain by shallow (⬍ 3 m*) subsoil
2. Sand and gravel aquifers with a shallow (⬍ 3 m*) unsaturated zone
3. Areas near karst features such as sink holes
High 1. Bedrock — major, minor and poor aquifers and non-aquifers — overlain by 3 m ⫹ sand
and gravel or 3–10 m sandy till or 3–5 m low permeability clayey tills or clays
2. Unconfined sand and gravel aquifer with 3 m⫹ unsaturated zone
Moderate 1. Bedrock — major, minor and poor aquifers and non-aquifers — overlain
by 10 m ⫹ sandy till or 5–10 m clayey till, clay or peat
2. Sand and gravel aquifers overlain by 10 m ⫹ sandy till or 5–10 m clayey
till, clay or peat
Low 1. Confined bedrock aquifers overlain by 10 m ⫹ clayey till or clay
or low permeability bedrock such as shale
2. Non-aquifers and poor aquifers overlain by 10 m ⫹ clayey till or clay
3. Confined gravel aquifers overlain by 10 m ⫹ clayey till or clay
* Note: less than 1 m subsoil or unsaturated zone beneath a development rather than 3 m could be the cut-off depth
for the ‘extreme’ rating. However, taking a thickness of 3 m rather than 1 m is regarded as more practical and useful
for the following reasons:
1. The base of many developments — septic tank systems or farmyard effluence holding tanks for instance — are
1–3 m below ground level.
2. In preparing a vulnerability map the general rather than the site-specific situation must be taken into account.
3. A 3 m cut-off depth allows for lateral variations and often provides a safety margin. Obviously if the base of a
potentially polluting development is ⬎3 m deep, the rating classification may be affected.
Quaternary sediments in typical Irish hydrological lithofacies models described above, therefore, pro-
settings that can be applied to other glaciated envi- vide a working framework designed to aid in the
ronments (Table 13.6). He emphasised the need for efficient and effective exploration and exploitation
detailed, sedimentological and geotechnical studies of glacial sediments for aggregate.
of Quaternary sediments in arriving at a compre-
hensive assessment of the nature of groundwater Mass movements
resources and vulnerability to pollution. Mass movement is particularly common in
glaciated terrain because of the abundance of
Aggregate exploitation unconsolidated sediment and the generally steep
Glacial and proglacial sediments can provide the slopes. Even in glaciated landscapes where slopes
engineer with a useful source of aggregates. The are gentle, such as the glaciated plains of North
highly variable nature of these sediments offers an America, mass movement is common particularly
opportunity to exploit a variety of sediment types. because sensitive clays (a common component of
An understanding of the landform and sediment sediments such as those deposited in glaciolacus-
associations is useful for efficiently exploiting trine environments) frequently play a major role
these resources, including predicting the likely in initiating landsliding. In a study of 8835 inland
extent, geometry and distribution of the sedi- landslides in the UK, 20% were found to lie in
ment. Particularly important aggregate resources tills, with most (32%) being complex (involving
include eskers, sandurs (outwash plains) and more than one mechanism), followed closely
kames (dominated by ice contact glaciofluvial and behind by debris flows (28%) and planar slides
glaciolacustrine sediments). The landsystems and (26%) (DoE, 1994; Trenter, 1999). In glaciated
GFE-13.qxd 6/24/05 3:57 PM Page 372
terrain, mass movements may be both natural and glacial terrain as unique, at least in the earlier
human-induced (Vaughan and Walbancke, 1975; stages of site investigation. Nevertheless, a basic
Chandler, 1984). Natural mass movements may knowledge of the types and sediments associated
occur as the sediments and landforms are forming with each of the landsystems described can pro-
or after glaciation as the landscape readjusts to the vide the investigator with a valuable framework
new environmental conditions. Eisbacher and with which to build up an understanding of the
Clague (1984) provided a useful overview of reasons for the considerable diversity found in so
glacier-related mass movements in high moun- many glacial terrains.
tains and suggested means of mitigating these
hazards. Human-induced mass movements are
particularly common when excavations are made
References
in glacial sediments, as occurs during highway
construction, for example. In a study of the haz- Alley, R. B. (1992) Flow-law hypotheses for ice-sheet
ards along the Karakoram Highway in Northern modeling. Journal of Glaciology 35, 108–139.
Pakistan, Derbyshire et al. (2001) have recently Alley, R. B., Cuffey, K. M., Evenson, E. B., Strasser, J. C.,
Lawson, D. E. and Larson, G. J. (1997) How glaciers
illustrated the variability of mass movements in entrain and transport basal sediment: physical con-
glaciated terrain. straints. Quaternary Science Reviews 16, 1017–1038.
Trenter (1999) suggests that remediation Ashley, G. M. (1996) Glaciolacustrine environments.
strategies for landslides in glaciated terrains In Menzies, J. (ed.) Modern Glacial Environ-
should include adequate drainage and the modifi- ments: Processes, Dynamics and Sediments, Vol. 1.
Butterworth-Heinemann, Oxford, 417–444.
cation of slope profiles, as well as retaining and Ballantyne, C. K. and Benn, D. I. (1996) Paraglacial
restraining structures. He stresses that the methods slope adjustment during recent deglaciation: implica-
available for use in glaciated terrain do not differ tions for slope evolution in formerly glaciated terrain.
in principle from those available for application to In Brooks, S. and Anderson, M. G. (eds) Advances in
landslides in non-glaciated regions. However, Hillslope Processes. Wiley, Chichester, 1173–1195.
Banham, P. H. (1977) Glacitectonites in till stratigraphy.
when selecting an approach on a glaciated site, Boreas 6, 101–105.
due account should be taken of the spatial variation Bell, F. G. (1993) Engineering Geology. Blackwell
in the substrate and the depth to rockhead, the Scientific Publications, Oxford, 359pp.
presence of glaciofluvial and glaciolacustrine Benn, D. I. and Ballantyne, C. K. (1993) The description
deposits, and other local factors such as climate, and representation of particle shape. Earth Surface
Processes and Landforms 18, 665–672.
vegetation and land use. Benn, D. I. and Ballantyne, C. K. (1994) Reconstructing
the transport history of glacigenic sediments: a new
approach based on the co-variance of clast shape
13.7 Conclusions: recommendations indices. Sedimentary Geology 91, 215–227.
Benn, D. I. and Evans, D. J. A. (1998) Glaciers and
to engineers working in glacial Glaciation. Arnold, London, 734pp.
or glaciated environments Benn, D. I., Wiseman, S. and Warren, C. R. (2000)
Rapid growth of a supraglacial lake, Ngozumpa
When undertaking any engineering activity Glacier, Khumbu Himal, Nepal. In Debris-covered
on glaciated or formerly glaciated terrain it is glaciers. IAHS (International Association of
Hydrological Sciences) Publication 264: 177–185.
important to understand the complex distribution Benn, D. I., Kirkbride, M. P., Owen, L. A. and Brazier, V.
and diverse properties of glacial sediments and (2003) Glaciated valley landsystems. In Evans, D. J.
landforms. The above descriptions of glacial envir- (ed.) Glacial Landsystems. Arnold, London, 372–406.
onments and sediments illustrate some of the typ- Billard, A. and Derbyshire, E. (1985) Pleistocene strati-
ical characteristics and the range of variability graphy and morphogenesis of La Dombes: an alter-
native hypothesis. Bulletin de l’Association francaise
found within glacial terrain. However, predicting pour l’etude du Quaternaire 22, 85–96
the nature of the substrate and processes within Bjerrum, L. and Jorstad, F. (1968) Stability of rock
such landscapes is notoriously difficult. Thus, it is slopes in Norway. Norwegian Geotechnical Institute
advisable to consider each engineering site within Publication 79, 1–11.
GFE-13.qxd 6/24/05 3:57 PM Page 373
Boone, S. J., Westland, J., Busbridge, J. R. and Garrod, B. Derbyshire, E., Gregory, K. J. and Hails, J. R. (1979)
(1998) Prediction of boulder obstructions. Negro Jr Geomorphological Processes. Dawson, Westview,
and Ferreira (eds) Tunnels and Metropolises. 312pp.
Balkema, Rotterdam, Vol. 2, 817–822. Dreimanis, A. (1989) Tills, their genetic terminology and
Boulton, G. S. and Eyles, N. (1979) Sedimentation by classification. In Goldthwait, R. P. and Matsch, C. L.
valley glaciers: a model and genetic classification. In (eds) Genetic Classification of Glacigenic Deposits.
Schluchter, C. (ed.) Moraines and Varves. Balkema, Balkema, Rotterdam, 17–84.
Rotterdam, 11–23. Dreimanis, A. and Vagners, U. J. (1971) Biomodal
Boulton, G. S. and Hindmarsh, R. C. A. (1987) distribution of rock and mineral fragments in
Sediment deformation beneath glaciers: rheology basal tills. In Goldthwait, R. P. (ed.) Till: A Sym-
and sedimentological consequences. Journal of posium. Ohio State University Press, Columbus,
Geophysical Research 92, (B9), 9059–9082. 237–250.
Boulton, G. S. and Jones, A. S. (1979) Stability of tem- Echelmeyer, K. and Wang, Z. (1987) Direct observa-
perate ice caps and ice sheets resting on beds of tion of basal sliding and deformation of basal drift
deformable sediment. Journal of Glaciology 24, at sub-freezing temperatures. Journal of Glaciology
29–43. 33, 83–98.
Boulton, G. S. and Paul, M. A. (1976) The influence of Eisbacher, G. H. and Clague, J. J. (1984) Destructive
genetic processes on some geotechnical properties mass movements in high mountains, hazard and
of glacial tills. Quarterly Journal of Engineering management. Geological Survey of Canada Paper
Geology 9, 159–194. 84–16.
Broecker, W. S. and Denton, G. H. (1990) What drives Elson, J. A. (1989) Comment on glacitectonite,
glacial cycles? Scientific American January, 43–50. deformation till, and comminution till. In:
Carter, T. G. (1983) Site investigation and engineering Goldthwait, R. P. and Matsch, C. L. (eds) Genetic
characterization of glacial and glaciolacustrine Classification of Glacigenic Deposits. Balkema,
materials. Unpublished PhD, University of Surrey. Rotterdam, 85–88.
Chandler, R. J. (1984) Recent European experience of Evans, D. J. (2004) Glacial Landsystems. Arnold,
landslides in over-consolidated clays and soft rocks. London, 535pp.
International Symposium on Landslides 4, Toronto, Eyles, N. (1983) Glacial Geology. Pergamon, Oxford,
Sept. 1984. Vol. 1. 409pp.
Clayton, L. and Moran, S. R. (1974) A glacial process- Eyles, N. and Dearman, W. R. (1981) A glacial terrain
form model. In Coates, D. R. (ed.) Glacial Geomor- map of Britain for engineering purposes. Bulletin
phology. State University of New York, USA, 89–119. of the International Association of Engineering
Daly, D. (1992) Quaternary deposits and groundwater Geology 24, 173–184.
pollution. Quaternary Proceedings 2, 79–89. Eyles, N. and Sladen, J. A. (1981) Stratigraphy and
Department of the Environment (DoE) (1994) geotechnical properties of weathered lodgement till
Landsliding in Great Britain. Her Majesty’s Station- in Northumberland, England. Quarterly Journal of
ery Office, London. Engineering Geology 14, 129–142.
Derbyshire, E. (1990) Environment: understanding and Eyles, N., Dearman, W. R. and Douglas, T. D. (1983a)
transforming the physical environment. In Cannon, T. The distribution of glacial landsystems in Britain
and Jenkins, A. (eds) The Geography of Contem- and North America. In Eyles, N. (ed.) Glacial
porary China. Routledge, London and New York, Geology. Pergamon, Oxford, 213–228.
80–101. Eyles, N., Eyles, C. H. and Miall, A. D. (1983b)
Derbyshire, E. (1992) Engineering in Quaternary Lithofacies types and vertical profile models, an
sediments: case studies from Western Europe alternative approach to the description and environ-
and Eastern Asia. Quaternary International 2, 33–48. mental interpretation of glacial diamict and diamic-
Derbyshire, E. and Miller, K. J. (1981) Highway tite sequences. Sedimentology 30, 393–410.
beneath the Ghulkin. Geographical Magazine 53, Eyles, N., Clark, B. M. and Clague, J. J. (1987) Coarse-
626–635. grained sediment–gravity flow facies in a large
Derbyshire, E. and Owen, L. A. (1996) Glacioaeolian supraglacial lake. Sedimentology 34, 193–216.
processes, sediments and landforms. In Menzies, J. Farrell, E. R., Coxon, P., Doff, D. H. and Pried’homme, L.
(ed.) Modern Glacial Environments: sediments Forms (1995) The genesis of the brown boulder clay of
and Techniques, Vol. 2. Butterworth-Heinemann, Dublin. Quarterly Journal of Engineering Geology
Oxford, 213–238. 28, 143–152.
Derbyshire, E., Fort, M. and Owen, L. A. (2001) Fookes, P. G. (1997) Geology for engineers: the
Geomorphological hazards along the Karakoram geological model, prediction and performance.
Highway: Khunjerab Pass to the Gilgit River, Quarterly Journal of Engineering Geology 30,
Northernmost Pakistan. Erdkunde, Band 55, 49–71. 293–424.
GFE-13.qxd 6/24/05 3:57 PM Page 374
Goldthwait, R. P. and Matsch, C. L. (eds) (1989) Genetic Menzies, J. and Shilts, W. W. (1996) Subglacial
Classification of Glacigenic Deposits. Balkema, environments. In Menzies, J. (ed.) Modern
Rotterdam. Glacial Environments: Processes, Dynamics and
Hambrey, M. J. (1994) Glacial Environments. UCL Sediments, Vol. 1. Butterworth-Heinemann,
Press, London, 296pp. Oxford, 15–136.
Hindmarsh, R. (1997) Deforming beds: viscous and Miall, A. D. (1978) Lithofacies types and vertical profile
plastic scales of deformation. Quaternary Science models in braided river deposits: a summary. In
Reviews 16, 1039–1056. Miall, A. D. (ed.) Fluvial Sedimentology. Canadian
Hird, C., Powell, J. J. M. and Yung, P. C. Y. (1991) Society of Petroleum Geologists, Memoir 5, 597–604.
Investigations of the stiffness of a glacial clay till. Murray, T. (1997) Assessing the paradigm shift:
Proc. 10th ECSMFE, Firenze 1, 107–110. deformable glacier beds. Quaternary Science Reviews
Jones, D. K. C., Brunsden, D. and Goudie, A. S. (1983) 16, 995–1016.
A preliminary geomorphological assessment of part Naito, N., Nakawo, M., Kadota, T. and Raymond, C. F.
of the Karakoram Highway. Quarterly Journal of (2000) Numerical simulation of recent shrinkage of
Engineering Geology 16, 331–355. Khumbu Glacier, Nepal Himalayas. In Debris-
Kirkbride, M. P. (1993) The temporal significance of covered Glaciers. IAMS (International Association
transitions from melting to calving termini at gla- of Hydrological Sciences) publication 264, 245–
ciers in the central Southern Alps of New Zealand. 254.
The Holocene 3, 232–240. Owen, L. A. (1994) Glacial and non-glacial diamictons in
Krüger, J. (1994) Glacial processes, sediments, land- the Karakoram Mountains and Western Himalayas. In
forms and stratigraphy in the terminus region of Warren, W. P. and Croot, D. G. (eds) The Formation
Myrdalsjökull, south Iceland. Folia Geographica and Deformation of Glacial Deposits. Balkema,
Danica 21, 1–233. Rotterdam, 9–28.
Lawson, D. (1981) Sedimentological Characteristics Owen, L. A. and Derbyshire, E. (1989) The Karakoram
and Classification of Depositional Processes and glacial depositional system. Zeitschrift für Geo-
Deposits in the Glacial Environment. Cold Regions morphologie Supp.-Bd. 76, 33–73.
Research and Engineering Laboratory, Report Owen, L. A. and Derbyshire, E. (1993) Quaternary and
81–27, Hanover, New Hampshire. Holocene intermontane basin sedimentation in the
Lehane, B and Faulkner, A. (1998) Stiffness and Karakoram Mountains. In Shroder, J.F. (ed.)
strength characteristics of a hard lodgement till. In Himalaya to the Sea. Routledge, London, 108–131.
Evangelista, A. and Picarelli, L. (eds) The Geotech- Owen, L. A., Derbyshire, E. and Scott, C. H. (2003)
nics of Hard Soils–Soft Rocks. Balkema, Rotterdam, Contemporary sediment production and transfer in
637–646. high-altitude glaciers. Sedimentary Geology 155,
Li, J. J., Derbyshire, E., Street-Perrott, F. A., Xu, S. Y. 13–36.
and Waters, R. S. (1984) Glacial and paraglacial Paterson, W. S. B. (1994) The Physics of Glaciers, 3rd
pediments of the Hunza Valley, north-west edition. Pergamon, Oxford, 480pp.
Karakoram, Pakistan: a preliminary analysis. In Powell, R. and Domack, E. (1996) Modern glacio-
Miller, K. J. (ed.) The International Karakoram marine environments. In Menzies, J. (ed.) Modern
Project, Vol. 2, Cambridge University Press, Glacial Environments: Processes, Dynamics and
Cambridge, 496–535. Sediments, Vol. 1, 445–486.
Lutenegger, A. T., Kemmis, T. J. and Hailberg, G. R. Reading, H. G. (1986) Facies. In Reading, H. G. (ed.)
(1983) Origin and properties of glacial till and Sedimentary environments and Facies, 2nd edition.
diamictons. In Geological Environments and Soil Blackwell, Oxford, 4–19.
Properties. Special Publication American Society Reynolds, J. M. (1998) High-altitude glacial lake
of Civil Engineers, Geotechnical Engineering hazard assessment and mitigation: a Himalayan
Division, 310–410. perspective. In Maund, J. G. and Edleston, M. (eds)
Maizels, J. (1996) Sediments and landforms of modern Geohazards in Engineering Geology. Geological
proglacial terrestrial environments. In Menzies, J. Society, London, Engineering Geology Special
(ed.) Modern Glacial Environments: Processes, Publications 15, 25–34.
Dynamics and Sediments, Vol. 1. Butterworth- Richardson, S. D. and Reynolds, J. M. (2000) An
Heinemann, Oxford, 365–416. overview of glacial hazards in the Himalayas.
McGown, A. and Derbyshire, E. (1977) Genetic influ- Quaternary International 65/66, 31–47.
ences on the properties of tills. Quarterly Journal of Shaw, J. (1997) Till deposited in arid polar environ-
Engineering Geology 10, 389–410. ments. Canadian Journal of Earth Science 14,
McGown, A., Sali, A. and Radwan, A. M. (1974) 1239–1245.
Fissure patterns and slope failures in boulder Small, R. J. (1983) Lateral moraines of Glacier De
clay at Hurlford, Ayrshire. Quaternary Journal of Tsidjiore Nouve: form, development and implica-
Engineering Geology 7, 1–26. tions. Journal of Glaciology 29, 250–259.
GFE-13.qxd 6/24/05 3:57 PM Page 375
Sugden, D. E. and John, B. S. (1976) Glaciers and Walker, R. G. (1992) Facies, facies models and mod-
Landscape. Arnold, London. ern stratigraphic concepts. In Walker, R. G. and
Trenter, N. A. (1999) Engineering in Glacial Tills. James, N. P. (eds) Facies Models: Response to Sea-
Construction Industry Research and Information level Change. Geological Association of Canada,
Association (CIRIA) Report C504, London, 259pp. Toronto, 1–14.
Tufnell, L. (1984) Glacial Hazards. Longman, London, Weertman, J. (1983) Creep deformation of ice. Annual
97pp. Review of Earth and Planetary Sciences 11,
Vaughan, P. R. and Walbancke, H. J. (1975) The stability 215–240.
of cut and fill slopes in boulder clay. Proc. Symp. Whiteman, C. A. (1996) Processes of terrestrial
Engineering Behaviour of Glacial Materials. deposition. In Menzies, J. (ed.) Modern Glacial
Midland Soil Mechanics and Foundation Engineering Environments: Processes, Dynamics and Sediments,
Society, Birmingham, UK, 209–219. Vol. 1. Butterworth-Heinemann, Oxford, 293–
Walder, J. S. and Costa, J. E. (1996) Outburst floods 308.
from glacier-dammed lakes: the effect of mode of Yamada, T. (1993) Glacier Lakes and their Outburst
lake drainage on flood magnitude. Earth Surface Floods in the Nepal Himalaya. Water and Energy
Processes and Landforms 21, 701–723. Commission Secretariat, Kathmandu, Nepal.
GFE-14.qxd 6/24/05 3:59 PM Page 376
14. Periglacial Forms and Processes

H. Jesse Walker
14.1 Introduction helping geomorphologists and engineers in their

interpretation of the fossil forms encountered in
In 1909 Lozínski (Jahn, 1975; 1985) introduced areas formerly subjected to the same types of
the term ‘periglacial’ in order to provide a way of processes.
designating those areas affected by the cold condi- Emphasis is placed on the Arctic and its unique
tions that bordered Pleistocene ice sheets. Today environmental conditions of which, from the
the term is used in reference to a much broader standpoint of engineering, permafrost is of major
area than either etymology or original concept sug- importance (Ferrians et al., 1969).
gest. It is now generally accepted that periglacial
environments possess temperature regimes that
include alternations between terrestrial freezing 14.2 Permafrost distribution
and thawing. Some 35% of the earth’s land surface
is affected by frost action (cryogenic) processes Present-day permafrost, which underlies some
(Williams and Smith, 1989). Because such tem- 20% of the world’s land area, is concentrated in the
perature regimes occur at high elevations, as well Northern Hemisphere and the ice-free portions of
as at high latitudes, periglacial processes also Antarctica. Although permafrost has been region-
operate in alpine areas of low latitudes even alised in a variety of ways, it is now usually dis-
including those straddling the Equator. At times in played under four categories: sub-sea permafrost,
the past, conditions that favour the development of continuous permafrost, discontinuous permafrost,
periglacial forms were more extensive than at pres- and alpine permafrost (Figure 14.1). Continuous
ent and, therefore, their remnants are often encoun- and sub-sea permafrost are presently limited to
tered in areas that today enjoy more temperate high latitudes; alpine permafrost, on the other
conditions. The inheritance of periglacial features hand, is found in both middle and low latitudes.
can have significant implications for engineering The lower elevation of alpine permafrost gradually
works. increases with decreasing latitude. Non-polar
Because both high latitude and high altitude alpine permafrost is affected by a variety of factors
environments have generally remained beyond that differ from those prevailing at high latitudes.
major human concern, scientific research on their Harris (1988) notes, for example, that
landforms has been neglected and engineer-
insulation and reradiation are particularly great at
ing experience has been limited. However, the
high altitudes . . . cold air drainage and inversions
realisation of the strategic importance of some
are very important, wind speeds are much higher,
periglacial areas and the recent push to exploit the differences in aspect are accentuated, . . . [and] num-
mineral wealth they possess are providing an bers of freeze–thaw cycles in the ground are much
increasing number of opportunities for geomor- higher.
phological and other research and for the devel-
opment of appropriate engineering principles by During parts of the Pleistocene, permafrost
scientists and engineers. This research is not only extended much further towards the equator
leading to a better understanding of present-day than at present. By combining the present-day
periglacial processes and environments but also is and Quaternary distributions of glacial ice and
376
GFE-14.qxd 6/24/05 3:59 PM Page 377
PERIGLACIAL FORMS AND PROCESSES 377
Figure 14.1 Permafrost in the Northern Hemisphere (after Péwé, 1983).
permafrost (Figure 14.2), a notion of the area in having only developed since the last glaciation.
which periglacial processes have operated at In the Arctic the areas with the oldest (and deep-
times in the past and are presently operating est) permafrost are those that were not glaciated
can be obtained. Over much of the Northern during the Pleistocene, such as the North Slope
Hemisphere, permafrost is relatively young, of Alaska and Eastern Siberia.
GFE-14.qxd 6/24/05 3:59 PM Page 378
Figure 14.2 Permafrost and glacial cover in the Northern Hemisphere through time (modified from Nekrasov, 1983).
14.3 Periglacial processes water freezes in its cracks and pores. The rate of
fracture and the size and shape of the end product
The permafrost landscape is an amalgam of depend on the type and character of rock as well
forms, some of which are created by processes as the intensity, frequency, and rate of freezing
that are uniquely periglacial; others are the prod- and the quantity and availability of water (e.g.
uct of processes more commonly associated with subsurface aquifers and melting snow). The end
other parts of the world. However, even they are product of frost wedging ranges in size from large
usually modified to some extent within the perma- blocks (that are likely to be tabular in shape when
frost environment. Forms created by freezing and derived from slate or schist) to silt (Figure 14.3).
thawing, mass wasting, nivation, wind, water, and Frost wedging leads to the production of talus
biota are all in evidence. which accumulates at the base of rocky cliffs.
The importance of freeze–thaw oscillations in When associated with a permanent snowfield it
geomorphology derives especially from the becomes a protalus rampart (see Figure 14.4).
unique property changes of water as it freezes and In extremely arid periglacial zones exfoliation
thaws. Solidification on freezing leads to the for- is common and may be more important than frost
mation of ground ice and expansion of the soil. wedging; in the better vegetated zones (e.g. wet
Liquefaction and contraction on thawing lead to tundra) chemical and biochemical weathering
solifluction and subsidence. also occur. However, unless they are derived from
mudstones, most periglacial soils contain min-
Frost action imal amounts of clay, partly because of the relative
Frost action (the process of freezing and thawing) unimportance of chemical weathering.
includes such specific actions as frost wedging Frost heaving is the displacement of rock (soil)
(shattering, riving, scaling, splitting)*, frost heav- particles or other included materials by ice as it
ing, frost creeping, frost sorting, nivation, and forms in a freezing soil. Although there is 9%
solifluction (gelifluction). expansion of pore water within the soil on freezing,
Frost wedging, considered by many to be the the most important aspect in frost heaving is the
main physical weathering process in cold regions production of segregation ice that forms as
(Washburn, 1985), is the fracturing of rock as unfrozen water flows to the freezing front and
freezes into lenses and flakes of various thick-
* Terms in parentheses are frequently found in the literature nesses. Although displacement caused by segre-
and are basically synonymous with those used here. gated ice is normally upward it may also be lateral
GFE-14.qxd 6/24/05 3:59 PM Page 379
hummocks (thufur) and for the displacement

upwards of buried rocks and other objects (includ-
ing power poles and house pilings). It is especially
common in fine-grained soils when moisture con-
tent is high. In clayey soils, on the other hand, dif-
ferential freezing is accompanied by deflocculation
and particle disaggregation. Upon thawing such
soils are virtually cohesionless and flow easily.
Frost heave and ice segregation cause some of the
most severe problems facing engineers in the
Arctic. They have to be considered in the construc-
tion of buildings, road beds, and pipelines.
Figure 14.3 Angular debris produced by frost shat- Frost sorting and frost creeping are two other
tering and ice wedging of limestone in the English
Pennines. Photograph courtesy of Tony Waltham
results of the freezing and thawing of soils. Size-
Geophotos. sorting can lead to the formation of various types
of patterned ground. Frost creeping results when
(sometimes referred to as frost thrusting). Frost frost-expanded soil thaws and settles downslope
heaving, which is irregular (except possibly in from where freezing occurred (Harris, 1986).
completely homogeneous material), is responsible Nivation (snow-patch erosion), which
for the production of surface forms such as earth occurs mainly because of the frost wedging that
Figure 14.4 Schematic of Late Devensian periglacial features in Scotland. Adapted from Ballantyne and Harris, 1994.
GFE-14.qxd 6/24/05 3:59 PM Page 380
accompanies freezing and thawing at the base and hazards they represent a challenge to the periglacial
edge of snowbanks, also includes chemical action engineer.
(Thorn, 1988). Snow has a high carbon dioxide As frozen soil thaws, the liquid water that does
content and is chemically active especially in not run off accumulates increasing the pore pres-
limestone terrain (Higginbottom and Fookes, sure in the soil which in turn causes a reduction in
1970). The frequency and intensity of nivation soil strength often leading to soil flow. On the
depend not only on rock type but also on the other hand, the runoff of melt water, combined
thickness and duration of snow cover. Snow, with settling of the soil particles as the ice melts,
being a good insulator, tends to reduce the num- results in thaw consolidation which affects the sta-
ber and intensity of freeze–thaw cycles and there- bility of the soil layer (Williams and Smith, 1989).
fore the effectiveness of mechanical weathering. Solifluction was sufficiently widespread dur-
ing the Pleistocene that it affected all of the
Solifluction (gelifluction) British Isles where its effects today are relic.
It has been stated that the ‘. . . predominant form Hutchinson (1991) writes that:
of periglacial mass movement in both the present
Arctic and, in relic form, in Britain’ is periglacial From an engineering viewpoint, relic solifluction
solifluction (Hutchinson, 1991, p. 285). In the materials may be regarded as having been loosened
most recent review of solifluction (Matsuoka, by frost heave, to some extent reconsolidated during
2001), it is noted that the term has yet to be and after thawing, and displaced downslope. For
defined unequivocally (p. 108). Many definitions foundation purposes, they are likely to have a lower
of the term have occurred since it was first pro- bearing capacity and a higher compressibility than
the underlying, parent stratum.
posed in 1906 by Anderson. Matsuoka (2001)
states that solifluction, as understood today,
Frost cracking
. . . involves several components: (1) needle ice Frost cracking, (ground cracking, thermal contrac-
creep and diurnal frost creep originating from diur-
tion cracking) occurs as frozen ground becomes
nal freeze–thaw action; (2) annual frost creep,
gelifluction and plug-like flow originating from colder and contracts. The rate at which the tem-
annual freeze–thaw action; and (3) retrograde move- perature drops appears to be more important than
ment caused by soil cohesion. the actual temperature at the time of cracking
(Lachenbruch, 1962). Such thermal contraction is
Although not limited to areas with frozen similar in many respects to the contraction that
ground, solifluction is most active there, because, occurs with desiccation, some chemical reactions,
even on low slopes, the frozen subsurface ensures and phase changes. The contraction polygons that
a saturated upper layer by preventing the water may be produced are highly variable in size and
from melting snow and thawing surface soil shape, a variability that is controlled by the rheo-
from percolating downward. This concentration logical nature of the cracking material as well as
also may lead to excess pore pressures in the temperature conditions. In periglacial areas, ther-
thawing soil which promotes slope movements mal-contraction cracks range up to several mil-
(Lewkowicz, 1988). Rates are affected by degree limetres in width, several metres in depth, and
of slope, soil texture, depth of thaw, water con- several tens of metres in separation and serve as
tent, and vegetation composition. voids into which melt water, snow, and sediment
In those permafrost zones where the ter- can collect.
rain is sloping, active-layer failure is common.
Although most solifluction is slow, relatively Other processes
rapid failures can also occur. Known as skin flows Although the interrelated processes (ground
(skin slides), detachment slides, and mudbursts, cracking, mass wasting, nivation, etc.) described
they appear to be caused by over-saturation at above are those that are most distinctly periglacial,
the permafrost table (Barsch, 1993). As potential other, more familiar types of processes such as
GFE-14.qxd 6/24/05 3:59 PM Page 381
Figure 14.5 Riverbank erosion along the Nechelik channel of the Colville River.
wind and running water are also active in when the ground surface is still frozen, water
periglacial regions. accumulates between the many hummocks that
Wind is regionally highly variable in periglacial characterise much of the surface. Soon thereafter,
environments; its effectiveness depends upon however, it flows across the surface and concen-
many conditions other than the velocity, direction, trates in streams. The flooding caused by melting
and duration of the wind itself. During much of snow accounts for most of the annual discharge in
the year (usually that part of the year when wind many streams in periglacial areas. For example,
velocities are highest) most of the surface in the the Colville River (the largest river on the North
Arctic is protected with a snow cover, albeit usu- Slope of Alaska) carries about half of its total
ally thin. On irregular surfaces snow accumulates annual flow during a 3–4 week period (Arnborg
in drifts, some of which may last through much of et al., 1966). It is during such flooding that much
the summer. Even on level surfaces snow ridges of the fluvial activity in periglacial areas occurs.
(sastrugi) form on the wind-packed snow surface. As Clark (1988) notes, it includes ‘water and sed-
Snow, with a hardness that may equal that of iment inputs, discharge and sediment transport,
quartz grains at low winter temperatures, is easily and the resulting hydrological and geomorpho-
moved and becomes an effective erosive tool. logical outputs’ (p. 415). Stream erosion may be
During snow-free periods, wind erosion, transport, extreme (Figure 14.5). Relatively warm water
and deposition are common (Walker, 1967). River flowing against snow-covered and frozen banks is
bars, outwash plains, and other non-vegetated areas responsible for thermal erosion (Figure 14.6),
are sources of much sediment and help account for which in turn may lead to block collapse (Figure
the wide distribution of loess in both present-day 14.7; Walker and Arnborg, 1966). Similar results
and former periglacial areas (Chapter 25). are found along the shorelines of lakes, lagoons,
Running water, a phenomenon of the frost-free and the open coast.
period of the year in the Arctic, is especially Ice jams are the frequent cause of flooding on
abundant in the snow-melt season. It originates arctic rivers. For example, in May 2001, ice jams on
from melting snow, thawing ground, and rainfall. the Lena River downstream from Yakutsk threat-
During the early stages of the snow-melt season, ened to flood that city with water that accumulated
GFE-14.qxd 6/24/05 3:59 PM Page 382
problems in periglacial environments. It is, as

stated by Williams (1995), the ‘ . . . formation
(aggradation) and disappearance (degradation) of
permafrost that are geotechnically the most impor-
tant consequences of temperature change’ (p. 348).
Engineering experience in coping with surface
ice during freeze-up and break-up and in the
period of time between them in temperate lati-
tudes has been substantial. Many of the publica-
tions of CRREL (Cold Regions Research and
Environmental Laboratory — see reference list
for web site) treat these topics in detail.
Figure 14.6 Thermoerosional undercutting of frozen
stream bank with subsequent collapse, on the Arctic Ground ice
Plain of Alaska. Photograph courtesy of Tony Waltham Ground ice, a frost phenomenon, is nearly ubiqui-
Geophotos. tous in periglacial areas where water is present. It
may ‘occur as thin fibers, grains, veinlets, large
vertical wedges, horizontal sheets or in irregular
masses of all sizes, shapes and colour . . . it may
be prismatic or granular, contain numerous air
bubbles which are either oriented or without
orientation and have horizontal or vertical planes’
(Hamelin and Cook, 1967, p. 17).
A number of classifications of ground ice have
been proposed. Two of the most widely referenced
(and modified) are those of Shumskiy and Vtyurin
(1966) and Mackay (1972) (Figure 14.8). In add-
ition to those that form in situ (pore ice, ice wedges,
Figure 14.7 Block collapse along ice wedges into a pingo ice) there are those ice masses that form
small river channel on the Arctic Plain, Alaska. above the surface and are buried (e.g. snow, icings,
Photograph courtesy of Tony Waltham Geophotos. and glacial, river, lake, and sea ice). Although such
buried ice may be significant locally, it generally is
upstream from the ice dam. Explosives dropped on of small areal extent.
the ice jams from helicopters breached the ice and The most widely distributed of ground-ice
reduced flooding. The 2001 ice-jam flooding, the types is pore ice (interstitial ice, ice cement)
worst in a century, was aggravated by exceptionally which bonds soil grains together. As long as there
thick ice, heavier than usual snowfall, and an is no excess water involved, this type of soil is
unusually warm spring (Kriner, 2001). subjected to minimal disturbance with phase
change. However, phase changes do bring about
impermeability and strengthening with freezing
14.4 Ground ice and permafrost and reduction of cohesiveness with thaw (Williams
and Smith, 1989).
Although temperature variations are a requisite for When moisture is present in the surface layers
periglacial processes, it is the presence of ice at the of the soil, freezing may produce nearly pure crys-
surface and in the ground that renders these tals that grow perpendicular to the surface. These
processes especially meaningful. Freeze and thaw, crystals, known as needle ice (pipkrake), are capa-
whether in river, lake, or sea, or within the ground, ble of heaving materials above the ground surface.
are responsible for most of the engineering They can develop in soils with textures ranging
GFE-14.qxd 6/24/05 3:59 PM Page 383
Figure 14.8 Classification of underground ice (modified from Mackay, 1972).
from clay to coarse sand although they are best conspicuous than either of them, and is usually
developed in silt. If that surface is generally level, more pure and granular. When the injected water
soil stirring will be the major result; if, on the other freezes in a tabular form it becomes sill ice; when
hand, the ground is sloping (even if less than 5⬚), it occurs as the core of a large mound it is pingo
the lifted material will settle downslope as the nee- ice. Small mounds with injected ice are known as
dle ice melts resulting in one type of frost creep. frost mounds.
Segregated ice (taber ice) develops in supersatu- Vein ice and wedge ice form in the contraction
rated soils. It is ‘epigenetic’ when it predates the cracks that develop in frozen ground usually after
surrounding sediments and ‘aggradational’ when rapid temperature drops in winter. Because these
forming in conjunction with the material enclosing cracks fill with any material that drops into them,
it. This enclosing material may be organic or min- veins and wedges vary in composition. However,
eral and it may accumulate slowly as with soil creep the most common ingredient appears to be the
or more rapidly during river flooding, mudflows, or water that fills them during snow-melt season. In
wind storms. Both types of segregated ice occur as Antarctica, where the quantity of liquid water is
lenses, some of which grow to be several metres minimal, the filling material is almost always
thick. Such growth is responsible for ground heave. fine-grained sediment which produces sand
Upon melting the excess water that characterises wedges instead of ice wedges.
these lenses is lost and subsidence occurs. An ice vein is the first step in the formation of
Intrusion ice (injection ice) develops where wedge ice which develops when sequential crack-
water is forced between sediment layers. Although ing occurs along the vein (Figure 14.9). Ice wedges
less important quantitatively than pore or segre- may grow to be several metres wide at their tops
gated ice, it produces surface forms that are more and occupy as much as half of the upper several
GFE-14.qxd 6/24/05 3:59 PM Page 384
Figure 14.9 Ice wedge formation (after Lachenbruch,

1962).
Figure 14.10 Ice wedge exposed in the Fox Tunnel
metres of ground in many areas in the Arctic. In cut into frozen silts near Fairbanks, Alaska; the ice
appears dark in this photograph as it absorbs more light
North America ice wedges are usually less than 8 m
from the camera flash. Photograph courtesy of Tony
deep and 3 m wide; in Siberia, on the other hand, Waltham Geophotos.
some wedges are much larger (Harry, 1988). Such
sizeable growth occurs under extremely cold con-
ditions, in fine-textured and peaty soils (Figure downward. The definition has nothing to do with
14.10), and in locations with aggrading alluvial the presence of water. If water is absent, as in solid
surfaces (Harry, 1988). Although wedge ice can rock, the material is called dry permafrost. When
develop in sandy soils (even sand dunes) it is gen- water is present, it is wet permafrost, the type that
erally a narrow vertical band of ice and seldom is of major concern to engineers because thawing
takes the classical wedge shape. causes subsidence and may cause erosion. There
are situations where water retains its liquid form
Permafrost and the active layer even though temperature may render the term
The optimum expression of periglacial forms and permafrost appropriate. One example is sub-sea
processes occurs in those areas where the subsur- permafrost where saline water fills the pores of the
face is perennially in a frozen state. The general sediment; another is in clay soil when the freezing
term for such a condition is permafrost. It is temperature is depressed below 0 ⬚C.
defined as being a naturally occurring earth mater- The upper surface of permafrost (i.e. the level
ial in which the temperature has been below 0 ⬚C below which the temperature remains below 0 ⬚C
for two or more years (Muller, 1947). It is initiated even during summer) is known as the permafrost
when a ground layer freezes in winter, but does not table. Above this table is the active layer, the layer
thaw in summer, and grows as freezing progresses that freezes in winter and thaws in summer.
GFE-14.qxd 6/24/05 3:59 PM Page 385
Figure 14.11 Permafrost characteristics from north to south across Canada. Adapted from Ballantyne and Harris, 1994.
Normally the bottom of the active layer is coex-

tensive with the permafrost table. However, during
some seasons or in some locations there is a zone
between these two where the temperature remains
above 0 ⬚C. Such unfrozen layers are called taliks.
Although taliks may be surrounded by permafrost
some are open at the top as for example beneath
lakes or river channels that are sufficiently deep
not to freeze to the bottom (Figure 14.11).
Permafrost extends to depths of several hun-
dred metres in the high Arctic. Depths of more
than 600 m in North America and nearly 1500 m
in Siberia have been recorded. Permafrost thick- Figure 14.12 Surface features and the distribution of
permafrost (after Lachenbruch, 1968).
ness is controlled by mean surface temperature
and geothermal heat flow. Low values for each
favour permafrost. Surface temperatures, and because they possess mean annual bottom tem-
therefore permafrost growth, vary with surface peratures that are higher than those of the sur-
character (Figure 14.12). Water bodies (lakes, rounding sub-aerial ground surface. In the case of
rivers, lagoons, the ocean) are major modifiers of the ocean, the same general characteristics prevail
regional permafrost development. Water bodies even though sea water freezes at temperatures
that are more than about 2 m deep do not freeze to below 0 ⬚C (⫺1.9 ⬚C at a salinity of 35‰).
the bottom and, even though surface ice on them Other factors influencing the average ground
usually lasts more than half a year, the water temperature include snow cover (thickness, dur-
beneath the ice cover ensures that bottom temper- ation), micro-relief, vegetation type, and soil
atures remain above freezing. In some cases, as characteristics (French, 1988). These factors have
when a river meanders across the land surface, a major influence on the development of the
thaw occurs beneath the river and a thaw-bulb active layer (Figures 14.13 and 14.14). This layer
develops. Where deep lakes and rivers are suffi- ranges in thickness from a few centimetres under
ciently large, the unfrozen zone (talik) may dense vegetal mats to more than 2 m in non-vege-
extend completely through the regional perma- tated, coarse textured, southerly exposed sand
frost (Figures 14.11 and 14.12). Even shallow dunes. Thickness may vary greatly over short dis-
lakes and streams (i.e. those that freeze to the bot- tances but basically follows the contours of the
tom in winter) influence permafrost development surface. In the zone of continuous permafrost, the
GFE-14.qxd 6/24/05 3:59 PM Page 386
active layer will normally be at or near saturation and low gradient slopes suppress horizontal
throughout the summer. Sand dunes and alluvial groundwater flow.
fans are often exceptions because of the permea-
bility of their coarse grained texture. Although
14.5 Periglacial landforms
the total annual precipitation is low in the Arctic,
high water content is possible in the active layer Periglacial landforms are highly varied in size,
because the permafrost table prohibits percolation shape, and distribution. Those most typical of
periglacial regions are found in zones of perma-
frost. Many, such as solifluction stripes and ice-
wedge polygons, are relatively small features but
because they are highly repetitive in occurrence,
they often extend over large areas (Walker, 1983).
Still others, such as pingos, are fewer in number
and often isolated.
Altiplanation terraces, block fields, and

rock glaciers
Altiplanation terraces (cryoplanation terraces) are
benches formed by frost action on hillsides. Formed
in bedrock, these terraces usually support a thin
layer of frost-shattered rock. They often occur in
steps and may encircle individual hill crests. In con-
trast, block fields (Felsenmeere) are thick accumu-
lations of frost-shattered rocks of different sizes and
shapes. There is relatively little movement of indi-
vidual blocks as most block fields tend to be level
or have low gradients. Rock glaciers consist of
Figure 14.13 Active layer thickness in various masses of poorly sorted angular blocks that are con-
periglacial terrains (after Mackay, 1970, in Washburn, solidated by interstitial ice. Nonetheless, controlled
1980). by slope and ice content, movement, often several
Figure 14.14 Active layer development in sand under different types of vegetation, Colville River delta.
GFE-14.qxd 6/24/05 3:59 PM Page 387
metres per year, is characteristic of rock glaciers among them. The first three types are most com-
during their active stage (Barsch, 1988). Upon mon on low-gradient surfaces, the latter two are
descending relatively steep slopes they may related to slopes.
progress across more gently sloping terrains. Rock Circles, which may be sorted or unsorted
glaciers have shapes that usually range from elon- (non-sorted) and are widespread in periglacial
gated tongues to gently curving lobes. These are environments, occur singly or, more commonly, in
features that are especially common in mountains groups, and are generally small (⬍ 3 m in diam-
such as those in Scotland (Figure 14.4). eter). There are different types of circles, e.g.
among sorted circles some have stone borders that
Patterned ground reach to the permafrost table whereas others have
Although patterned ground is not restricted to borders that are surface features only. Nets have
periglacial regions, it reaches its optimum devel- patterns that are intermediate between those of
opment there, especially in zones of continuous circles and polygons.
permafrost. Patterned ground is one of the most Polygons (Figure 14.15), the most common of
distinctive features to be formed in freeze–thaw patterned ground types, range from a few centime-
climates. tres to over 100 m in diameter. As in the case of cir-
The surface forms in patterned ground are cles and nets, the degree of sorting is diagnostic.
highly varied in shape, size, composition, degree Sorted polygons, which may be more than 10 m
of sorting, and rate of development. Basic pat- across, have borders comparable to those of sorted
terns include circles, nets, polygons, steps, and circles and do not need permafrost to form.
stripes. Although the shapes are often geometric- Unsorted polygons, which are also wide-
ally representative, there are frequent gradations spread, may be composed of peat, silt, sand, and
Figure 14.15 Sorted stone polygons in Alaska. Photograph courtesy of Tony Waltham Geophotos.
GFE-14.qxd 6/24/05 3:59 PM Page 388
even gravel. The borders between unsorted poly- Lachenbruch (1962) divided ice-wedge poly-
gons are often cracks which contain water and gons into groups based on the way ground cracking
support a type of vegetation that differs from occurs. When most of the cracks are at right angles,
that of the centres. The most distinctive type of they are labelled orthogonal; when most are about
unsorted polygon in periglacial regions is the 120⬚, non-orthogonal. Orthogonal polygons, which
ice-wedge (tundra) polygon (Figure 14.16). It typically form in inhomogeneous material, may be
only forms in areas with permafrost and is repet- random or oriented. The formation of oriented sys-
itive across thousands of square kilometres in tems is favoured by gradual recession of water
the Arctic. Counterparts in ice-free Antarctica from a surface as occurs during river meandering
are the polygons that result from sand-wedge and lake drainage.
formation. Steps (lobate soils) and stripes (solifluction
Ice-wedge polygons are usually classified as stripes) occur on sloping terrain as the result of the
either low-centred or high-centred depending on gravitational migration of material downslope. If
the relative heights of their borders and centres. the downslope border of the step is composed of
Low-centred polygons are indicative of actively coarser material than occurs upslope, sorting has
growing ice-wedges, usually contain water in occurred. In contrast, unsorted steps are bordered
their centres during summer, and serve as ideal by vegetation. In essence, stripes have the same
locations for peat formation. High-centred poly- distinctions as steps and differ from them mainly
gons reflect degradation and often have water in in that their orientation is down the slope (Figure
their troughs. In both types, surface irregularity is 14.17) instead of across it. Both usually occur in
great although total relief is low. groups although they can form singly.
Figure 14.16 Ice-wedge polygons in old lake sediments in the Mackenzie Delta, Canada. Photograph courtesty of
Tony Waltham Geophotos.
GFE-14.qxd 6/24/05 3:59 PM Page 389
Figure 14.17 Solifluction stripes created by slope Figure 14.18 Inverted relief formed as polygonal ice
movement of a sorted soil in permafrost, on Disko wedges melt in the summer at Prudhoe Bay, Alaska.
Island, western Greenland. Photograph courtesy of Photograph courtesy of Tony Waltham Geophotos.
Tony Waltham Geophotos.
Frost mounds and pingos Thermokarst is considered by many to be one

The increase in volume that occurs with the freez- of the ‘most dynamic process suites modifying
ing of water is responsible for the formation of arctic and subarctic landscapes’ (Harry, 1988,
rather symmetrical mounds some of which reach p. 132). Although the development of thermokarst
heights of more than 50 m. Such large forms are occurs naturally, it can also be initiated and aggra-
known as pingos. Although there are different vated by human activities. It provides periglacial
ways in which pingos form, the majority develop engineers with some of their more serious problems
when a talik (as would be present beneath a small in arctic development. Any future warming of the
lake) begins to freeze from all sides. Progressive climate that may occur will aggravate thermokarst
expansion within the subsoil forces surface layers development.
upward. Smaller features ( frost mounds, palsas)
form in shallow water bodies such as low-centred Other landforms
polygons. Although frost mounds, palsas, and Within periglacial regions, river valleys, deltas,
pingos are perennial features there are others, sand dunes, glaciated terrain, beaches, and other
including icing mounds, icing blisters, and frost more familiar landforms are also present. However,
blisters, that are seasonal (Pollard and van even these features are affected by periglacial
Everdingen, 1992). processes and most have many of the forms dis-
cussed above superimposed on them. For exam-
Thermokarst and associated melt forms ple, the Colville River delta, Alaska, is a typical
If the mounds described above are considered to be delta in most respects but also associated with its
positive forms (i.e. heave) then the development of bars, flats, dunes, distributaries, and lakes are
thermokarst is negative (i.e. subsidence). When the such periglacial forms as ice wedges, ice-wedge
ground ice within permafrost thaws, volume is polygons, frost mounds, and pingos.
decreased and the surface subsides. Such subsid-
ence often leads to the creation of thaw lakes, some 14.6 Periglacial soils
of which are oriented and which abound in per-
mafrost regions. Melting ice-wedge junctions can Soils formed in periglacial environments are like
lead to the formation of a distinctive drainage those that develop in other environments in that
pattern known as beaded drainage. The melting their morphological, physical, and chemical charac-
of ice-wedges creates conspicuous troughs (Figure teristics are dependent upon a number of factors
14.18) bringing about an inversion of relief. including lithology, climate, hydrology, topography,
GFE-14.qxd 6/24/05 4:00 PM Page 390
biota, and time. In periglacial areas the relative Contrasting with these ahumic (i.e. without
importance of these factors varies greatly. In those humus) and arid zones are those periglacial areas
regions with little or no biotic component (such as where biota flourish and liquid water (even if for
is common in Antarctica, the high Arctic, and at only a part of the year) is abundant at the surface
high elevations), soil development is almost exclu- and in the active layer. Gentle slopes and impeded
sively the result of physico-chemical processes near groundwater circulation, both of which occur
the surface (Tedrow and Ugolini, 1966). Such areas over much of the Arctic, aggravate waterlogging,
also tend to be extremely arid. The soils are usually and lead to ‘gleization’ (development of a sticky,
coarse grained. compact, blue-grey layer) and to the production
of tundra and bog soils (Figure 14.19).
Vegetation affects the development of soils
through the insulation it provides, through root
binding, and through its slow decay rate (Washburn,
1980). Peat, which is extensive over much of the
Arctic, has a lower thermal conductivity than min-
eral soil and thus different freezing characteristics.
One of the major variables in periglacial soil
development is time. In areas only recently
exposed because of glacial retreat and rebound
(isostatic uplift), bare bedrock or freshly exposed
deposits are little modified. Further, in areas of
deposition (deltas, river valleys, sand dunes, loess
deposits), the upward movement of the perma-
frost table limits the length of time during which
soil development can occur.
Soil morphology is complicated by the ‘soil-
destroying’ processes associated with frost action
including solifluction (Figure 14.17) because,
according to Tedrow (1977), these processes
Figure 14.19 Tundra soil profile. Dashed line operate too fast for soil morphology to adjust
approximates permafrost table (after Tedrow, 1966). (Figure 14.20).
Figure 14.20 Cross-section of tundra soil illustrating complicated structure (after Tedrow, 1977).
GFE-14.qxd 6/24/05 4:00 PM Page 391
Figure 14.21 Soils and related patterned ground features (after Tedrow, 1977 as compiled from Brown, 1966).
Soil types in periglacial environments often permafrost, it is also responsible for the lengthy
reflect the type of patterned ground with which period of the year during which rivers, lakes, and
they are associated. Brown (1966), for example, the sea are covered with ice, a cover that is usually
found for one valley in arctic Alaska that there topped by snow as is the surrounding land. Snow
were no fewer than ten soil types associated with cover, over most of the Arctic, is so complete that
thirteen types of patterned ground (Figure 14.21). at some locations it is difficult to distinguish
Two of the most important characteristics of soil visually between land and sea during winter.
in the Arctic is that it is frozen — seasonally, in Arctic rivers are highly variable in length, dis-
the case of the active layer — and that frozen soil charge, and dates of freeze-up and break-up. In the
is usually stronger than either ice or unfrozen soil periglacial areas of the Northern Hemisphere,
(Williams and Smith, 1989). rivers are of two basic types: (1) those that origi-
nate outside the zone of continuous permafrost,
and (2) those whose basins are confined to it.
14.7 River ice, lake ice, and sea ice Although flow characteristics vary among the
rivers that flow into the Arctic Ocean, break-up and
One of the major components of the periglacial the flooding that accompanies it is critical in all of
regions of the Northern Hemisphere is water, in them. The timing and intensity of break-up affects
the liquid state for part of the year and frozen many human activities in the river’s floodplains
(at least on the surface) for the rest of the year. and deltas. Much of the annual erosion and deposi-
Lakes and rivers, many of which enter the tion by rivers occurs in the relatively short period
periglacial environment from more southerly lati- of break-up and the flooding that accompanies and
tudes, are abundant and their shorelines, along follows it.
with those of the Arctic Ocean and its bordering Sea ice, which dominates the Arctic Ocean
seas, serve as the most important locations for (even in summer), is shore-fast along the coastline
human settlements. Just as the temperature regime for varying lengths of time ranging from only a
is responsible for the formation and distribution of few months to a full year; along some coasts, for
GFE-14.qxd 6/24/05 4:00 PM Page 392
Figure 14.22 The Chukchi and Beaufort Seas showing the length of fetch during September 1986.
some years. It is during break-up and freeze-up displayed in a variety of forms such as involu-
and the time in between (i.e. the relatively ice-free tions, solifluction, valley bulging, and camber-
summer period) that most of the shoreline is ing in Britain, peat deposits in former pingo
impacted. Although ice is an eroding, depositing, craters in The Netherlands, ice-wedge casts
and transporting agent, especially during periods in Poland, and crop (ice-wedge) polygons in
of freeze-up and break-up, it is the waves that Denmark (Figure 14.23). Of these numerous
develop during storms that are the major coastal forms and processes, Hutchinson (1991) writes
modifiers. The most adverse conditions occur that in Britain the
when the ice front retreats some distance from the
shore thereby providing a long fetch over which reactivation of relic clayey solifluction mantles by
storm waves can develop (Figure 14.22). At such ill-advised earthworks probably constitutes at pres-
times beach and bluff erosion may be severe ent the most frequent and costly type of failure . . .
endangering coastal settlements. To date few having a periglacial origin.
coastal engineering projects (such as sea walls and
beach renourishment) have been undertaken to Among the many micro-relief forms present in
prevent or control coastal erosion (Walker, 1991). many parts of north-western Europe are the rem-
nants of ice-wedge polygons. They often become
conspicuous because of differences in the rates at
14.8 Engineering and fossil which crops grow. Aerial photographs of these
periglacial forms patterns have been made in Denmark, Sweden,
Germany, and England (Christensen, 1978).
Even in those vast areas no longer subjected Several hundred locations of fossil ice wedges in
to periglacial conditions (Figure 14.2), surface Denmark have been identified (Figure 14.23). They
and subsurface forms and soil characteristics tend to be of two types: those where the wedges are
frequently reflect former periglacial processes filled with fine-grained materials and those with
and forms. Evidence of such activity is well sediments coarser than the surrounding material
GFE-14.qxd 6/24/05 4:00 PM Page 393
Figure 14.24 Illustrations of the relative effect of ice-

wedge casts on cereal growth at maturity (July–August).
Adapted from Christensen (1978, p. 259).
Figure 14.23 Crop polygons above fossil ice-wedge

polygons in barley field, Denmark (after Christensen,
1978).
(Christensen, 1978). These differences are reflected

in variations of growth rates which may be negative
or positive (Figure 14.24).
The occurrence of ice wedges, involutions, frost
mounds, and pingos in the fossil form are major
concerns in engineering because they may repre-
sent a ‘. . . sudden and usually unexpected replace-
ment of one material by another, often with inferior
geotechnical properties’ (Higginbottom and
Fookes, 1970, p. 95). They (1970) also note that the
problem is especially important in connection with Figure 14.25 Fossil wedge from central Poland
(Jahn, 1975).
shallow foundations, roads and runways. A dia-
gram by Alfred Jahn (Figure 14.25) illustrates the
conditions one might encounter vis-à-vis ice in two ways:
wedges during construction in a former permafrost Firstly, the trench was subject to local flooding by
environment. Morgan (1971) notes that in the water running from the wedges. Secondly, the walls
English Midlands fossil wedges affected trenches of the trench collapsed, either because the wedges
GFE-14.qxd 6/24/05 4:00 PM Page 394
were running parallel to, but behind, the trench wall Although agriculture and mining are affected
or because the sand and gravel at the base of the till by periglacial conditions, more important from the
sheet was washed out. standpoint of engineering in the north is the influ-
ence of these conditions on the construction and
Although such forms as fossil ice wedges may
maintenance of the many material structures
be unexpected, other fossil forms, including some
(buildings, roads, sewers, etc.) that comprise mod-
pingo remnants which are common in northern
ern culture. Engineering problems are especially
Europe, are rather conspicuous and thus can be
acute where poorly drained, fine-grained sediment
avoided by engineering works (Hutchinson, 1991).
predominates. Péwé (1998, p.758), noting these
Cambering and valley bulging are mass move-
difficulties, wrote that ‘engineering problems in
ments that may have a periglacial connection.
the north are of four fundamental types:
Cambering (the downward displacement of strata)
and valley bulging (the upward displacement of
strata) are considered as related phenomena and 1. those involving thawing of ice-rich permafrost
are both responsible for a number of engineering and subsequent subsidence of the surface
problems. One of the major aspects associated under unheated structures such as roads and
with cambering are the gulls (filled fractures airfields
between blocks) that can cause leakage affecting 2. those involving subsidence under heated
foundation stability (Petley, 1991). structures
Present over much of former permafrost areas 3. those resulting from frost action, generally
are vast deposits of loess which is considered intensified by poor drainage caused by
by Eyles and Paul as ‘the most widespread of permafrost
periglacial sediments’ (as reported in Culshaw 4. those involved only with the temperature of
et al., 1991, p. 14). Loess (Chapter 25) presents permafrost that causes buried sewer, water, and
problems to the engineer primarily because of oil lines to freeze’.
their tendency to collapse when wetted (Culshaw
et al., 1991).
Because of low population density, roads and
airfields are relatively few in number in the Arctic
14.9 Engineering and present-day and railways are generally restricted to the sub-
permafrost environments Arctic. Nonetheless, virtually all undergo frost
heaving in winter and subsidence in summer.
The challenges that face engineers in present-day Maintenance is a major and almost continuous
permafrost environments include those found problem. The major challenge is the establishment
elsewhere plus many that are unique; solutions are of a bed that prevents permafrost degradation and
equally as diverse and complex (Morgenstern, thaw settlement (Williams, 1995). Such fills may
1981). Nearly every human activity in the Arctic have to be as much as 2–3 m thick to ensure
vis-à-vis the environment is affected by periglacial sub-fill permafrost stability. Insulation materials
conditions. Agriculture, even though generally including synthetic fabrics (Figure 14.26) help
limited to zones of discontinuous permafrost may, reduce the thickness of fill needed (Johnston,
nonetheless, be affected. Large masses of residual 1981). Another major concern in the construc-
ice in the ground when disturbed (e.g. by clearing tion of roads, railways, and airfields is drainage.
the natural vegetation cover) is likely to result in Because natural drainage is poor any disruption
thermokarst topography and field abandonment. of the surface can aggravate ponding and perma-
In the case of mining, permafrost is usually a frost thaw.
negative factor for open pit operations. However, Buildings, because they are heated, have an
permafrost, because of its stability (as long as it intensified degradational effect on the underlying
remains in the frozen state), can be a positive permafrost. A variety of techniques has been
factor in underground operations. devised to assist in preserving the integrity of
GFE-14.qxd 6/24/05 4:00 PM Page 395
Figure 14.26 Oil tank built on a gravel pad in Inuvik, northern Canada; the tube vents within the gravel allow cold
air to blow through in winter, ensuring the conservation of permafrost. Photograph courtesy of Tony Waltham
Geophotos.
permafrost and therefore of the buildings. One of and may vary in size up to those that can be
the earliest, other than thick gravel pads, was to walked through.
build on pilings thus preventing heat from the
building being transferred to the ground beneath.
Of course pilings are subject to frost jacking 14.10 Oil industry
which can damage bridge pilings as well as those
used in buildings or other structures. A more The relatively recent expansion of the oil industry
recent technique for maintaining the integrity of in the Arctic — especially since the establishment
the permafrost has been the utilisation of thermal of the Prudhoe Bay fields in the late 1960s — has
piles (thermosyphons, thermotubes) which are led to much research on exploration, drilling,
passive heat transfer devices that operate when storage, and distribution of oil and gas in areas
the air temperature is lower than that of the with permafrost. The challenges posed in the
ground. development of the resource involves those con-
The problems of supplying utilities to the cerned with preserving the integrity of the envi-
residents of centres of population in the Arctic ronment including not only land, water, and air
and sub-Arctic are difficult and expensive but also flora and fauna.
undertakings. The common practice is the use Many of the problems facing the construction of
of utilidors (Figure 14.27) which contain fuel roads, airports, and buildings are aggravated in
distribution lines, electrical cables, sewer lines, hydrocarbon exploitation (Ferrians et al., 1969).
and hot water pipes. They may be buried in or These difficulties are possibly best illustrated in
beneath the active layer or placed above ground connection with the 1285 km long, 1.22 m diameter
GFE-14.qxd 6/24/05 4:00 PM Page 396
Figure 14.27 Utilidors linking houses raised clear of the ground on pile foundations to conserve the permafrost
in Inuvik, a modern Inuit town in northern Canada. Photograph courtesy of Tony Waltham Geophotos.
pipeline that delivers oil across Alaska from 14.11 Summary and future
Prudhoe Bay to the Gulf of Alaska. The line crosses challenges
many different types of landscape including those
with continuous permafrost, buried ice wedges, The periglacial landscape presents the engineer
fine to coarse-grained sediments, bogs and swamps, with unique and serious problems: problems that
and a variety of vegetation types. Because the oil is can be solved only by the use of special tech-
hot (70–80 ⬚C), thawing of permafrost would occur niques, procedures, and materials. Most of these
if the pipeline was buried. Thus about half of the problems are related to a thermal regime that
Alaskan pipeline is above ground. Frost jacking of includes alternating freeze and thaw. When freeze
the supporting pilings is prevented by the use of and thaw cycles occur in areas of poorly drained
thermosyphons (Figure 14.28). fine-grained sediments, periglacial processes are
Exploitation of petroleum is being pur- especially effective and surface forms such as pat-
sued offshore as well as onshore. Additional prob- terned ground, pingos, and solifluction lobes are
lems associated with exploration, drilling, and common. Changes in this thermal regime, which
production involve the construction of platforms can be affected by many of man’s activities, may
and offshore islands that can withstand sea ice lead to subsidence and frost-heave (Committee on
and wave action. Further, in addition to these Permafrost, 1983). Most of the challenges facing
oceanographic hazards, sub-sea permafrost the periglacial and permafrost engineer are
(Figure 14.1) must be considered in drilling and related to such changes and they are likely to
the placement of pipelines. increase in number and intensity in the near
GFE-14.qxd 6/24/05 4:00 PM Page 397
Figure 14.28 The trans-Alaskan pipeline with thermosyphons. Photograph courtesy of Tony Waltham
Geophotos.
future. General Circulation Models (GCMs) indi- References

cate that not only will atmospheric temperature
rise around the globe but that it is likely to rise Anderson, J. G. (1906) Solifluction, a component of
subaerial denudation, Journal of Geology 14,
two to four times as much in the Arctic as else-
91–112.
where. High latitude periglacial regions, because Arnborg, L., Walker, H. and Peippo, J. (1966)
of the unique characteristics discussed above, are Water discharge in the Colville River Alaska,
especially sensitive to changes in temperature and 1962. Geografiska Annaler, Series A 48, 195–
other climatic variables and are likely to be 210.
Ballantyne, C. K. and Harris, C. (1994) The
altered more than most other regions (Maxwell,
Periglaciation of Great Britain. Cambridge
1992). There are already indications that such University Press, Cambridge.
changes are occurring in parts of the Arctic. In Barsch, D. (1988) Rockglaciers. In Clark, M. J. (ed.)
some locations, for example, the active layer is Advances in Periglacial Geomorphology. John
thickening, permafrost temperatures are rising, Wiley & Sons, Chichester, 69–90.
Barsch, D. (1993) Periglacial geomorphology in the
thermokarst development is accelerating, sea ice
21st century. In Vitek, J. D. and Giardino, J. R. (eds)
is thinning, snow fall amounts are changing as are Geomorphology: The Research Frontier and
albedos (reflected sunlight), and the dates of Beyond. Elsevier, Amsterdam, 141–163.
freeze-up are later and break-up earlier. All such Brown, J. (1966) Soils of the Okpilak River Region,
changes will tax the ingenuity and expertise of Alaska. CRREL Research Report 188, US Army,
Hanover.
those engineers called upon to work in the area.
GFE-14.qxd 6/24/05 4:00 PM Page 398
Christensen, I. (1978) Waterstress conditions in cereals Kriner, S. (2001) Ice bombings bring reprieve from
used in recognizing fossil ice-wedge polygonal pat- Siberian flood. http://www.redcross.org/news/in/
terns in Denmark and northern Germany. In flood/010522siberia.html
Proceedings, Third International Permafrost Lachenbruch, A. (1962) Mechanics of Thermal
Conference. National Academy of Sciences, Contraction Cracks and Ice-wedge Polygons in
Washington, DC, 254–261. Permafrost. Geological Society of America, Special
Clark, M. J. (1988) Periglacial hydrology. In Clark, M. J. Publication No. 70, New York.
(ed.) Advances in Periglacial Geomorphology. Lachenbruch, A. (1968) Permafrost. In Fairbridge, R.
John Wiley & Sons, Chichester, 415–462. (ed.) The Encyclopedia of Geomorphology.
Committee on Permafrost (1983) Permafrost Research: Reinhold, New York, 833–839.
an Assessment of Future Needs. National Academy Lewkowicz, A. G. (1988) Slope processes. In:
Press, Washington, DC. Clark, M. J. (ed.) Advances in Periglacial
CRREL. Cold Regions Research and Environmental Geomorphology. John Wiley & Sons, Chichester,
Laboratory. http://www. crrel.usace.army.mil/ 325–368.
Culshaw, M. G., Cripps, J. C., Bell, F. G. and Mackay, J. R. (1970) Disturbances to the tundra and
Moon, C. F. (1991) Engineering geology of forest tundra of the western arctic. Canadian
Quaternary soils: 1. Processes and properties. In Geotechnical Journal 7, 420–432.
Forster, A., Culshaw, M. G., Cripps, J. C., Little, J. A. Mackay, J. R. (1972) The world of underground ice.
and Moon, C. F. (eds). Quaternary Engineering Annals of the Association of American Geographers
Geology. The Geological Society, London, 3–38. 62, 1–22.
French, H. M. (1988) Active layer processes. In Matsuoka, N. (2001) Solifluction rates, processes and
Clark, M. J. (ed.) Advances in Periglacial Geomorph- landforms: a global review. Earth Science Reviews
ology. John Wiley & Sons, Chichester, 151–177. 55 (1-2), 107–134.
Ferrians, O. J., Jr, Kachadoorian, R. and Greene, G. W. Maxwell, B. (1992) Arctic climate: potential for change
(1969) Permafrost and Related Engineering under global warming. In Chapin, F. III, Jefferies,
Problems in Alaska. Geological Survey Professional R., Reynolds, J., Shaver, G., Sroboda, J. and Chu, E.
Paper 67, Washington, D C. (eds) Arctic Ecosystems in a Changing Climate: An
Hamelin, I. and Cook, F. (1967) Illustrated Glossary of Ecophysiological Perspective. Academic Press, Inc.,
Periglacial Phenomena. Laval University Press, New York, 11–34.
Quebec. Morgan, A. V. (1971) Engineering problems caused
Harris, S. A. (1986) The Permafrost Environment. by fossil permafrost features in the English
Barnes and Noble, Totowa. Midlands. Quarterly Journal Engineering Geology
Harris, S. A. (1988) The Alpine periglacial zone. In 4, 111–114.
Clark, M. J. (ed.) Advances in Periglacial Geomor- Morgenstern, N. R. (1981) Geotechnical engineering
phology. John Wiley & Sons, Chichester, 369–413. and frontier resource development. Geotechnique
Harry, D. G. (1988) Ground ice and permafrost. In: 31 (3), 303–366.
Clark, M. J. (ed.) Advances in Periglacial Muller, S. W. (1947) Permafrost or Perennially Frozen
Geomorphology. John Wiley & Sons, Chichester, Ground and Related Engineering Problems.
113–149. Edwards Brothers, Ann Arbor.
Higginbottom, I. and Fookes, P. (1970) Engineering Nekrasov, I. (1983) Dynamics of the cryolithozone in
aspects of periglacial features in Britain. Quarterly the northern hemisphere during the Pleistocene.
Journal Engineering Geology 3, 85–117. In Proceedings, Fourth International Permafrost
Hutchinson, J. N. (1991) Periglacial and slope pro- Conference. National Academy of Sciences,
cesses. In Forster, A., Culshaw, M. G., Cripps, J. C., Washington, DC, 903–906.
Little, J. A. and Moon, C. F. (eds) Quaternary Petley, D. J. (1991) Report on Session 2b. In
Engineering Geology. London, The Geological Forster, A., Culshaw, M. G., Cripps, J. C., Little, J. A.
Society, 283–331. and Moon, C. F. (eds) Quaternary Engineering
Jahn, A. (1975) Problems of the Periglacial Zone. Geology. The Geological Society, London, 409–414.
Polish Scientific Publishers, Warsaw. Péwé, T. L. (1983) Alpine permafrost in the contiguous
Jahn, A. (1985) Experimental observations of United States: a review. Arctic and Alpine Research
periglacial processes in the Arctic. In Church, M. 15, 145–156.
and Slaymaker, O. (eds) Field and Theory, Lectures Péwé, T. L. (1998) Permafrost. Encyclopedia Britannica
in Geocryology. Univ. of British Columbia Press, 20, 752–759.
Vancouver, 17–35. Pollard, W. H. and van Everdingen, R. O. (1992)
Johnston, G. H. (ed.) (1981) Permafrost: Engineering Formation of seasonal ice bodies. In Dixon, J. C. and
Design and Construction. John Wiley & Sons, Abrahams, A. D. (eds) Periglacial Geomorphology.
Chichester. John Wiley & Sons, Chichester, 281–304.
GFE-14.qxd 6/24/05 4:00 PM Page 399
Shumskiy, P. and Vtyurin, B. (1966) Underground ice. Williams, P. J. and Smith, M. W. (1989) The Frozen
In Proceedings, First International Permafrost Earth. Cambridge University Press, Cambridge.
Conference. National Academy of Sciences,
Washington, DC, 108–113.
Tedrow, J. (1966) Arctic soils. In Proceedings, First
International Permafrost Conference. National Further reading
Academy of Sciences, Washington, DC, 10–55.
Tedrow, J. (1977) Soils of the Polar Landscapes. Church, M. and Slaymaker, O. (eds) (1985) Field and
Rutgers University Press, New Brunswick. Theory: Lectures in Geocryology. University of
Tedrow, J. and Ugolini, F. (1966) Antarctic soils. In British Columbia Press, Vancouver.
Tedrow, J. (ed.) Antarctic Soils and Soil Forming Davis, N. (2001) Permafrost: A Guide to Frozen Ground
Processes. National Academy of Sciences, in Transition. University of Alaska Press, Fairbanks.
Washington, DC, 161–177. Dixon, J. C. and Abrahams, A. D. (eds) (1992) Periglacial
Thorn, C. (1988) Nivation: a geomorphic chimera. In Geomorphology. John Wiley & Sons, Chichester.
Clark, M. J. (ed.) Advances in Periglacial Geomorph- French, H. M. (1996) The Periglacial Environment.
ology. John Wiley & Sons, Chichester, 3–31. Second Edition. Addiser Wesley Longman, London
Walker, H. J. (1967) Riverbank dunes in the Colville and New York.
delta, Alaska. Coastal Studies Bulletin 1, 7–14. John, B. and Sugden, D. (1975) Coastal geomorphology
Walker, H. J. (1983) E Pluribus Unum: small landforms of high latitudes. Progress in Geography 7, 53–132.
and the arctic landscape. In Gardner, R. and King, C. (ed.) (1976) Periglacial Processes. Benchmark
Scoging, H. (ed.) Mega-Geomorphology. Clarendon, Papers in Geology 27. Dowden, Hutchinson & Ross,
Oxford, 39–55. Inc., Stroudsburg, Pennsylvania.
Walker, H. J. (1991) Bluff erosion at Barrow and Linell, K. and Johnston, G. (1973) Engineering design
Wainwright, Arctic Alaska. Zeitschrift für Geomorph- and construction in permafrost regions: a review. In
ologie, Supplementband 81, 53–61. Proceedings, Second International Permafrost
Walker, H. J. and Arnborg, L. (1966) Permafrost and ice- Conference. National Academy of Sciences,
wedge effect on riverbank erosion. In Proceedings, Washington, DC, 553–575.
First International Permafrost Conference, NAS- Linell, K. and Tedrow, J. (1981) Soil and Permafrost
NRC, Publ. 1287, National Academy of Science, Surveys in the Arctic. Clarendon, Oxford.
Washington, DC, 164–171. Péwé, T. L. (ed.) (1969) The Periglacial Environment:
Washburn, A. L. (1980) Geocryology. Survey of Past and Present. McGill-Queens University,
Periglacial Processes and Environments. Edward Montreal.
Arnold, London. Williams, P. J. (1986) Pipelines & Permafrost. Carleton
Washburn, A. L. (1985) Periglacial problems. In University Press, Ontario.
Church, M. and Slaymaker, O. (eds) Field and Woo, M-K. and Gregor, O. J. (eds) (1991) Arctic
Theory, Lectures in Geocryology. Univ. of British Environment: Past, Present & Future. McMaster
Columbia Press, Vancouver. University, Hamilton.
Williams, P. J. (1995) Permafrost and climate change: Walker, H. J. (1973) The morphology of the North Slope.
geotechnical implications. Philosophical Trans- In Britton, M. (ed.) Alaskan Arctic Tundra. Arctic
actions 352 (1699), 347–358. Institute of North America, Washington, DC, 49–92.
GFE-15.qxd 6/24/05 4:04 PM Page 400
15. Temperate Environments

Kenneth J. Gregory
Temperate landscapes have often been visualised infiltration rate, or vegetation or land use cover
as those that remain when other, more distinctive density. The temperate domain can be introduced
world environments have been identified. The by the deposits, sediments and the landscape
types of temperate area (Figure 15.1) are listed (15.1 Deposits, sediments and the landscape), the
with process characteristics, engineering problems movement of water through the landscape system
and geohazards (Table 15.1). These areas include (15.2 Water movement), leading to the signifi-
some of the most densely populated parts of the cance of human impact for engineering (15.3
earth’s surface, have presented great challenges Human activity), and thence to the key elements
for engineering, and have stimulated research for an engineering geomorphology approach
which has given much of the foundation for (15.4 Summary).
geomorphological understanding. They are the
areas which inspired the normal cycle of erosion,
the importance of the drainage basin, and the 15.1 Deposits, sediments and
basis for contemporary hydrological understand- the landscape
ing of runoff generation.
However, with hindsight, for temperate areas Thin deposits over the surface of many temperate
to provide the norm against which other areas areas are often relict from former climatic or envi-
should be studied is inappropriate because the ronmental conditions. It is therefore difficult for the
rates of operation of present processes (rivers, engineer to deduce how extensive a particular
Chapter 10; coasts, Chapter 21; slopes, Chapter deposit is and how much variability occurs within it
11) are not typical; processes and environments because, as the deposits originated under condi-
have been greatly affected by human activity tions which no longer exist, there is no simple basis
especially through removal of the original vege- for estimation. The range of deposits encountered
tation, and temperate landscapes contain much in temperate landscapes includes not only weather-
evidence of landscape change during former cli- ing, colluvial and fluvial materials originating from
mates, so that the majority of landscape features contemporary conditions, but also marine, aeolian,
and their deposits do not now relate to contempo- glacial, periglacial, and tropical deposits, which can
rary environmental processes. The range of tem- still be significant amounts of material although
perate environments (Figure 15.1) can be thought they originated during earlier phases of landscape
of as the domain of rain and rivers operating on development. The major problems associated with
landscapes which often contain the record of a the interpretation of each type of deposit are indi-
variety of past environmental processes. Process cated in Table 15.2.
domains have been identified (Thornes, 1979; It is often necessary to have some indication
1983) according to the spatial distribution of work of the lateral and vertical extent of a deposit;
done by several processes, often typified by some whereas some deposits blanket the landscape, as
particular environmental parameter such as rain- a glacial deposit may completely infill a former
fall intensity. Domains therefore represent equi- valley with little or no indication on the surface
librium relationships between processes as related (Figure 15.2), in other cases the deposits are
to the controlling parameters such as climate, associated with particular features or distinctive
400
GFE-15.qxd 6/24/05 4:04 PM Page 401
TEMPERATE ENVIRONMENTS 401
Figure 15.1 Temperate landscape systems. Based upon Tricart and Cailleux (1972). Details for each system
including possible geohazards are given in Table 15.1.
landforms. Anomalous channels or valleys temperate environments can be envisaged (Brown,

which are steep-sided and flat-floored and which 1991) namely:
at present contain a small stream or no drainage
at all, were the lines along which glacial melt-
1. First order: caused by major cyclic, and proba-
water flowed. Such channels commonly have
bly Milankovitch mechanism, variations in cli-
glacial till or periglacial deposits on their floor
mate involving large fluctuations in the water
and perhaps also on their sides, and the floor
balance. Such changes included changes from
may have been infilled by peat (Figure 15.3).
glacial to interglacial climate conditions which
When building construction takes place in areas
occurred more than twenty times during the
where such channels occur the different proper-
Quaternary.
ties of each deposit in the sequence dictate that
2. Second order: caused by shorter climatic vari-
site location must be planned very carefully and
ations such as interstadials with a periodicity
that engineering structures must often be
of either 40 000 or 25 000 years.
anchored to the underlying bedrock. Deposits
3. Third order: due to climate or geomorphology
over the landscape may be changed significantly
(e.g. earth movement in Japan, Turkey or New
by various types of human activity as sum-
Zealand) or human activity (e.g. deforesta-
marised together with attendant problems that
tion).
may arise in Table 15.3.
4. Fourth order: effects of individual years or
events that may persist for many years and
Climatic changes may be important thresholds to new states. The
Four orders of change each having had a signifi- effects of a major flood could have long-term
cant effect on the deposits and landforms of effects upon the landscape or lead to a threshold
being crossed in order to move to a new state.
GFE-15.qxd
Table 15.1 Subdivisions of the temperate environment (Based upon Tricart,1957; details of zones described in Tricart and Cailleux, 1972). Under natural con-
402
ditions zones 1–4 would be forested but have been substantially transformed by human activity. Zones 1 and 5 are transitional to other zones.

Zone (see Climate Processes and features Major hazards which Problems relevant to engineering
6/24/05
Figure 15.1) may occur (using list in
Alexander, 1999)
(1) Forest on Severe winters, Permanently frozen ground Wildfires Modification of surface affects thermal
4:04 PM
Quaternary may be associated beneath land surface may be Frost or ice storm regime and can lead to thermokarst
permafrost with periglacial continuous or discontinuous, Snowstorm features with surface collapse
zone is residual from the Quaternary Subsidence
and not forming at present Windstorm
Page 402
(2) Forested zone Maritime without Chemical erosion limited by Accelerated erosion Landsliding on devegetated slopes
of middle severe winters. moderate temperatures, some Avalanches Many ancient deposits over landscape
latitudes No large seasonal frost action but penetration Soil heave and collapse Flooding may increase downstream of
variations in rarely reaches bedrock. High Floods vegetation changes
temperature or angle slopes can be stable Landslides
humidity where still covered by forest Coastal erosion
(3) Forested zone With severe winters Heavy showers and snowmelt Drought Landslides when vegetation removed
of middle latitudes and seasonally can produce higher streamflow Severe thunderstorms Downstream flooding increases when
distributed rates than in zone 2, mechanical Hailstorms vegetation changed and other catchment
precipitation processes more important as frost Snowstorms characteristics altered
penetration is great and can reach
bedrock. Chemical erosion limited by
winter frost but slopes can be steep,
up to 20o to 35o, and rectilinear
(4) Mediterranean Seasonal precipitation, Alternation of wet and dry Soil erosion Sheet erosion where vegetation removed
forested zone of mild winters, warm/ conditions induce major Floods, high spatial and Increased flooding downstream, and gully
middle latitudes hot summers. Frost influence such as an incidence temporal variability development may occur
uncommon at low of landslides. Seasonal stream High sediment yields
elevations flow regime can give high along rivers
seasonal discharges which elevate Earthquakes
course debris and rapid dissection Volcanic eruptions
and gullying where vegetation Landslides
removed or degraded
(5) Subdesert Summer rainstorms, Transitional to temperate deserts Drought Deflation encouraged by removal of
steppes and dry cold, severe with some frost action in winter. Tornadoes vegetation
prairies winters Wind action, occasional sheet Soil erosion Gullying where land ploughed
wash and gullying
GFE-15.qxd 6/24/05 4:04 PM Page 403
Table 15.2 Types of deposit which may occur in temperate areas.
Type of deposit developing Characteristics Associated problems

in temperate areas at present
but also in the past
Weathering and colluvial Medium grain size, not usually deep, Some lateral variation due to
may have developed on top of fossil subsurface drainage
deposits, podzolized soils on surface
Fluvial Range of grain sizes present and often Small changes in river position
incorporate remnants of fossil can release new exposures of
deposits — frequent lateral and vertical relict sediments
changes
Marine Range of grain sizes present, may Deposits may also occur above
incorporate fossil deposits from cliff present sea level marking former
deposits or from off-shore sediments shorelines
Aeolian Fine silts common in zone 5 and also in May be mantle over surface and
other areas where vegetation removed not related to deposits of very
and then field boundaries removed. different character below
Under former conditions had loess
or water-sorted loess (brick earth)
cover
Relict from former conditions
Glacial Till deposits, lake clays, gravels Distribution and character not
and sands easily deduced; rapid variations
in thickness
Periglacial Angular scree deposits, unsorted Distribution localised but
slope deposits, fine wind-blown character reflects locally
deposits available rock types — may be on
slopes which are relict and
unstable when modified
Tropical Clays which are remnants of deep Occur localised, often on plateau
weathering sites, to considerable depth in
pockets
Figure 15.2 Section across an infilled channel palaeomeander at Czmoniec, Poland (redrawn from Kozarski, 1983)
showing how there can be great differences between the form of the land surface (here virtually flat) and the form of
the subsurface profile.
GFE-15.qxd 6/24/05 4:04 PM Page 404
Figure 15.3 Significance of deposits in relation to road construction. In (a) the early stages of motorway embank-
ment construction above an infilled glacial drainage channel at Walton’s Wood Staffordshire are indicated. Failure
of the embankment necessitated reconstructions and remedial measures as shown in (b). After Chandler (1977) and
based upon Early and Skempton (1972).
Table 15.3 Examples of problems following changes in deposits.
Change due to: Consequences
Weathering Progressive reduction in shear strength may give slope failure in

natural or manmade slopes
Drainage Contraction of deposits can give substantial lowering of surface
elevation, and so increase flooding. Desiccation cracks can provide
location for slumps or for gully development
Groundwater Groundwater rise, for example due to creation of reservoir, can
changes increase pore pressures and increase possibility of slope failure
Loading Uneven amounts of settlement because of variations in deposits and
in layers of deposits
Piping Localisation of infiltration on surface by land use practices can lead
to increase of subsurface piping and this can subsequently lead to
collapse and gully development
Expansive sediments Swelling especially of soils with appreciable montmorillorite clay
and rocks content can occur and with a slight increase of moisture content can
have differential effect on surface
Temperate areas are now covered by several domains, with sediments from past systems still
types of deposit, many of which originated under retained in the landscape environment. It has been
past process regimes. Similarly the landscape fea- suggested (Büdel, 1982; Lewin, 1987) that 95% of
tures and landforms owe their origins to the impact the temperate landscape may be composed of
of a succession of climate changes. Deposits and landforms inherited from pre-Holocene condi-
landforms of temperate landscapes were once tions. Such inherited landscapes include features
parts of one of the succession of dynamic systems produced by glacial processes, others which include
that were arrested leaving material stored over the evidence of periglacial processes (Chapter 14) and
landscape. Analysis of this succession of changes some reminiscent of contemporary loess lands. As
has been undertaken using landscape stability and the climate and the associated erosion systems in
instability, identification of critical thresholds, and the past changed from glacial to periglacial to tem-
linkages between different sediment stores. From perate and back to glacial again, so a considerable
a sediment point of view the temperate landscape range of deposits and landform features can occur
is the result of the intersection of different in a single area (Figure 15.4).
GFE-15.qxd 6/24/05 4:04 PM Page 405
Figure 15.4 Fluctuations of world landscape zones in the late Tertiary and the Quaternary. The zone of former
valley formation and the subtropical zone which are the principal temperate areas have varied in position during the
glacial and interglacial phases of the Quaternary. This diagram based on Büdel (1982) presents a very simplified
picture which is now known to be much more complex and composed of many oxygen isotope stages (Chapter 6).
Such changes, which occurred in a time

sequence, were accompanied by changes in sea
level, which in turn had an influence on land-
forms because valley deepening was associated
with low sea levels leaving former valley floors as
river terraces. The reverse development, aggrada-
tion, a building up of alluvium and alluvial
deposits, accompanied higher sea levels of the
past. Therefore the temperate landscape may be Figure 15.5 Generalised section across an upland
thought of as broadly composed of three major valley in the temperate zone (compare with Figure
15.6). Three elements in the cross-section are (1) pla-
components (Figures 15.5 and 15.6) as listed nation surface remnants which are the remnants of old
below. land surfaces often produced in the Pliocene or earlier;
(2) valley development which occurred mainly during
the Quaternary and (3) Holocene erosion during the
1. Plateau areas are where many features were last 10 000 years often confined to areas closest to
initially produced by pre-Quaternary erosion. rivers and streams.
These are the remnants of landscapes produced
under Tertiary conditions, which may have
resulted in planation surfaces generated under and some changes in slope or river dynamics
tropical or subtropical conditions, later affected may have been concentrated in short periods
by erosional and depositional processes of gla- of time.
cial or periglacial conditions.
2. Major valleys are the product of the erosional Viewed along the line of major valleys, this
and depositional phases of the last two sequence has culminated in the stages illustrated
million years of the Quaternary, in which in Figure 15.7. The present temperate landscape
the alternation of glaciations and inter- must therefore be viewed against the background
glacials was accompanied by major sea level of up to 10 million years of landscape develop-
fluctuations. ment under the successive influence of a sequence
3. The detailed development of the environments of contrasting erosion systems so that a profile
along the present river courses and valley floors across a major temperate valley (Figure 15.8) can
are the products of the last 10 000 years of the reflect many of the stages of Quaternary develop-
Holocene, when the climate changed from ment (Chapter 6).
glacial to interglacial. The impact of human A composite section across the valley of
activity has become increasingly pronounced, the Wislowka in southern Poland is shown in
GFE-15.qxd 6/24/05 4:04 PM Page 406
Figure 15.6 Headwaters of River Exe basin, Somerset, UK. The snow-covered plateau surfaces are remnants of
land surfaces produced more than 2 million years ago. The valley sides, covered by woodland and plantations, are
the result of dissection during the Quaternary.
Figure 15.7 Diagrammatic representation of stages of Quaternary valley development. These stages occurred as
valley development took place as shown in Figure 15.5. Diagram based on Butzer (1976) after Dury (1970). Phases of
erosion and downcutting occurred when sea levels were relatively low and deposition occurred when sea levels were
higher. In addition to the influence of sea level, changes of climate also affected river regime and sediment transport.
Figure 15.8. The deposits in that section have valleys have evolved during the Quaternary
accumulated in the last 40 000 years and many of (Figure 15.5). Such alluvial chronologies are
them during the last 11 000 years. In temperate based largely upon the deposits and dateable
areas the alluvial chronology may be reconstructed material which have survived and five major
as the stages of development through which the influences upon the alluvial chronology in any
GFE-15.qxd 6/24/05 4:04 PM Page 407
Figure 15.8 Detailed valley cross-section across Wisloka valley, southern Poland showing the complex sequence
of deposits produced during the last 6000 years (after Starkel, 1983).
area may be recognised. First is the response to way in which the past evolution of temperate
changes of sea level, as indicated in Figure 15.7, landscapes has been interpreted in that a more ret-
as there will be phases of deposition of river and rospective approach has been employed using
marine deposits, and of erosion of such deposits. knowledge of contemporary process domains to
Secondly, there are effects of changes of climate extrapolate back into the past. In this way palaeo-
which will be indicated by the character of the hydrology, which has been defined as ‘the science
ground slope and the fluvial deposits. Thirdly, of the waters of the earth, their composition, dis-
there will be influences from other areas, when, tribution and movement on ancient landscapes
for example, ice-transported deposits or wind from the beginning of the first rainfall to the
blown sediments are added to the alluvial beginning of continuous hydrological records’
sequence. Fourthly, in tectonically active areas (Gregory, 1983, p. 10; 1996, p. 2), has utilised an
like Japan, Turkey or New Zealand, there may be approach retrospective from current processes.
earth movements which displace the sediments. Reconstruction of environmental change
Finally, the influence of human activity may have (Gregory, 2000) in temperate landscapes has
affected not only the amount of particular come to depend upon an increasingly multidisci-
deposits but also the type of sediment produced. plinary approach and one which endeavours to
Such deposits have been dated in a variety of understand the way in which contemporary
ways including reference to archaeological evi- process–response systems are at variance with
dence and by use of radionuclides (e.g. Stokes those of the past as a consequence of human
and Walling, 2003). The course of valley develop- activity.
ment in a temperate landscape can result in some
or all of these five major contributions being Sediment transfer and model hillside
blended together to make up cross-profiles such Although many deposits in the temperate land-
as the one illustrated in Figure 15.8. scape may be inherited from past environmental
Advances in the techniques available for the conditions, two types of changes of deposit can
interpretation of the environment and the dating occur under contemporary conditions. First,
of its evolutionary stages have enabled a more weathering processes (Chapter 2) can lead to a
refined chronology to be determined for the reduction in material strength and hence to slope
Cainozoic; this has been accompanied by an failure. Secondly, transfers of sediments in the
enhanced understanding of the way in which landscape can be achieved either by gravitational
processes operate in temperate landscapes and of processes operating on hillslopes or by fluid
the ways in which those processes differ over the transport.
temperate environment and have changed in The temperate hillslope can be regarded as
recent decades especially as a result of human composed of any combination of nine hypotheti-
activity. A change has therefore occurred in the cal landsurface units (Dalrymple et al., 1968) and
GFE-15.qxd 6/24/05 4:04 PM Page 408
Table 15.4 Units of the hypothetical nine-unit landsurface model (Figure 15.9) based upon
Dalrymple et al. (1968).
Unit Predominant present geomorphic processes
(1) Interfluve Pedogenetic processes associated with vertical subsurface

soil water movement
(2) Seepage slope Mechanical and chemical eluviation by lateral subsurface
water movement
(3) Convex creep slope Soil creep, terracette formation
(4) Fall face Fall; slide; chemical and physical weathering
(5) Transportational midslope Transportation of material by mass movement (flow, slide,
slump, creep); terracette formation; surface and subsurface
water action
(6) Colluvial footslope Redeposition of material by mass movement and some
surface wash. Fan formation. Transportation of material;
creep; subsurface water action
(7) Alluvial toeslope Alluvial deposition; processes resulting from subsurface
water movement
(8) Channel wall Corrosion, slumping, fall
(9) Channel bed Transportation of material downvalley by surface water
action; periodic aggradation and corrosion
each of these has certain properties and therefore

a certain significance in relation to engineering
development (Table 15.4). This hypothetical
model (Figure 15.9), has been developed in rela-
tion to soil profile development and is particularly
related to the mode of water flow. In addition it is
necessary to envisage the ways in which several
types of mass movement may take place and also
to estimate where and when a particular type of
movement will occur. The locations may be envis-
aged from the scars of previous mass movements
but in general the incidence of specific failures
may arise as a consequence of the reduction of the Figure 15.9 The hypothetical nine-unit landsurface
material shear strength occurring as a result of a model (after Dalrymple et al., 1968). Details of the
number of significant variables (Table 15.5). In units are given in Table 15.4.
some areas such as California the sequence of
intense rainstorms can produce mass movements
together with material carried as solutes or
in slopes over the landscape (Figure 15.10) and
dissolved load (Chapter 10).
slope instability is also found in areas where
recent uplift has occurred in Mediterranean areas
and New Zealand. Location and storage of sediment
Sediment is also transferred in the temperate Although in temperate areas there are no very high
landscape by water flow, over the slopes as sheet rates of erosion, unless there is profound distur-
flow, and along river courses, where it is trans- bance by human activity, two aspects of sediment
ported as bedload rolled or jumped along the transport are particularly relevant to engineering.
bed, or as suspended load, carried in suspension First is the fact that in many temperate areas the
GFE-15.qxd 6/24/05 4:04 PM Page 409
Table 15.5 Variables which may cause reduction in shear strength of materials and
land-surface.
High intensity precipitation The temperate landscape is adjusted to events which

occur on average every few years so that precipitation
falls which are rare (e.g. 1 in 100 years) may reduce
material shear strengths in specific locations
Vegetation removal Increases amount of moisture infiltrating and so can
increase porewater pressures and decrease shear
strengths
Removal of root network removes binding influence
upon material
Ploughing In grassland areas breaks down the plant––soil
structure where a mat is provided by the perennial
grasses
(see also Tables 15.3, 15.8 for rate of groundwater changes, drainage, weathering).
Figure 15.10 Coast ranges of California showing extent of landslide scars produced by heavy rainfall
January–March 1980. The rainstorms succeeded a wet autumn so that debris slides occurred extensively over the
landscape and disrupted the chapparal (Mediterranean type) vegetation.
GFE-15.qxd 6/24/05 4:04 PM Page 410
sources of sediment are restricted so that the of aquatic ecosystems, for example in the United
movement of sediment through the temperate sys- States, emphasise the protection and restoration
tem is less than would be expected and takes place of stream side riparian forests. However the issue
at sub-capacity levels. If unlimited amounts of of whether the banks of streams and rivers are
sediment were to be available then it would be pos- better left to revert to forest or not is complicated
sible to deduce amounts of transport more easily because there is a myriad of interrelated variables
from values of water discharge or of water veloc- involved which can include sediment supply, size
ity using hydraulic equations. However, in temper- and lithology; magnitude and frequency of water
ate areas most sediment sources occur near stream discharge; nature of overbank materials; presence
channels so that as little as 10% of a drainage and type of vegetation on the banks; types of
basin will be contributing to sediment transfer. channel and its history of disturbance; as well as
Erosion rates in temperate areas tend to be fairly the drainage basin context (Montgomery, 1997) —
low unless the erosion regime is very seasonal and all affecting whether streambank trees can stabilise
characterised by rainfall intensities greater than or erode the banks on which they grow.
25 mm per hour. If stream or river discharge is A second feature affecting sediment transfer
reduced, for example as a result of the construc- in temperate areas is the considerable amount of
tion of a dam or reservoir, then the river channel storage in the system. Not only does this apply
may be armoured by the selective concentration of on the long time scale in the sense of the various
a coarse layer on the channel bed which reduces forms of deposition referred to earlier (climatic
the availability of sediment to the stream channel. changes), but on a much shorter timescale
Conversely if the discharge is increased new sedi- of decades, there can be storage on slopes, in
ment sources may be eroded, and wherever the flood plains, and along valley floors (e.g.
landsurface is modified by human activity then the Figure 15.11) which can all be released if
areas most likely to influence the sediment supply the hydrological system is changed substantially.
must be treated with great care. This applies In many temperate areas floodplains store the
particularly to the riparian vegetation along sediment that was released when the vegetation,
water courses which should be removed only if usually forest, over the catchment areas was
absolutely necessary; removal means that protec- cleared by human activity which could have
tive engineering measures are then required. This begun some 4000 years ago. In the Severn and
is one reason why strategies for the conservation Wye catchments it was suggested (Brown, 1987)
Figure 15.11 Sediment storage for Coon Creek, Wisconsin (after Trimble, 1983). Numbers are annual averages of
sediment supply in 103 Mg year and show the importance of different sediment sources and also compare the stor-
age and supply before conservation measures (1853–1938) with the subsequent situation (1938–1975). The storage
for 1975–1993 is reduced significantly (Trimble, 1999).
GFE-15.qxd 6/24/05 4:04 PM Page 411
that one might expect much of the Severn’s flood- Table 15.6 Changes in sediment storage due to
plain sediment, rapidly deposited around 2000– specific structures.
3000 years ago, to have been remobilised by
Type of structure Effect on sediment storage
increased fluvial action due to increases in flood
magnitude and frequency and decreased slope ero- River dam with Sediment trapped upstream of
sion. In fact, the channelization and regulation of reservoir dam; accelerated erosion
British lowland channels prevented this occurring downstream of dam.
by artificially lengthening the residence time of May have deltaic accumulation
sediment in floodplain storage. Embanking had of sediment where rivers enter
reservoir
also led to a reduction in flooding and in flood-
Channelization, Sediment sources covered.
plain accumulation rates, but to an increase in
flood prevention Low rates of sediment
within-channel sedimentation especially in lower schemes, urban transport until sources become
reaches. The quantities of stored sediment in low- drainage systems available downstream. In-
land basins such as that of the Severn suggested channel storage on riffles and
that human activity altered the relationships bars also precluded
between the erosion, transport and storage Slope control Render sediment unavailable to
processes during the mid and late Holocene. measures by surface runoff
Consequently as much sediment went into storage grading
as has left the Severn and Wye basins during the or drainage
Holocene. The sediment storage component is
therefore not spatially uniform and in particular it
is likely that the majority of hillslope erosion went reaches of the river channel have silted and many
into proximal colluvial storage. The complexity of sections of the river bed have risen. The ‘sus-
alluvial sediment accretion is illustrated by the pended river’ means that the river bed in the
agricultural Coon Creek basin, Wisconsin where lower reaches is generally 3–5 m higher than the
sediment budgets with conservation measures floodplain behind the levees and in some sections
1938–1975 were a fraction of those 1853–1938 the height above the floodplain is 10 m (Liu
(Figure 15.11) and in 1975–1993 were only about Changming, 2000). Major problems in the basin
6% of the rate that occurred in the 1930s (Trimble, include severe soil erosion in the middle reaches,
1999). Adjustments in sediment storage that arise flood hazard in the lower reaches, degradation of
from specific structures are indicated in Table ecosystems throughout the basin as a result of
15.6, together with indications of the increase in deforestation, grassland degradation, desertifica-
sediment production that can arise from human tion, saline and alkalisation, and water pollution
activity. throughout the basin for which a management
Some parts of the temperate zone have been strategy has been developed. The river has ceased
characterised by conditions in which sediment flowing in its lower reaches on twenty-one occa-
supply has been greatly increased, usually as a sions in the period from 1972 to 1998 and in the
consequence of human activity. Soil erosion over 1990s the river’s lower reaches dried up every
the catchment area of the Huang He (Yellow year for increasingly longer periods. Problems
River) basin of China is indicated in the saying arising from the drying up of the river include
‘once the skin is gone where can the hair grow?’ increased flood hazard (as a result of the sus-
It has been responsible for the largest sediment pended channel), serious ecological effects in the
load of any world river, carrying nearly 10% of delta area, increased saline intrusions, and water
all the sediment transported to the oceans from supply problems.
the surface of the globe (Walling, 1981) and Deforestation in the Mediterranean basin pro-
averaging 1.69 billion tonnes annual sediment duced accelerated erosion which has been reflected
production 1919–1996 (Liu Changming, 2000). in a series of stages recorded in the alluvial
As a result of the high sediment load the lower chronology.
GFE-15.qxd 6/24/05 4:04 PM Page 412
15.2 Water movement be fed by groundwater or delayed or base flow

during dry periods; of intermittent streams that
Temperate areas are classically regarded as the flow only when the water table is seasonally high,
domain of rain and rivers. As problems have arisen and of ephemeral streams that will flow only dur-
in relation to some parts of the transfer of water ing particularly heavy rainstorms. Thus under nat-
over the land surface, it has been discovered that an ural conditions water will flow along a network of
engineering solution, necessary to solve a particu- stream channels which is constantly changing in
lar problem in one area, may often create a further relation to the hydrograph expressing variations of
problem elsewhere. Thus a flood prevention discharge along a river against time (Figure
scheme involving channelization in one area may 15.13). The drainage network will be most exten-
lead to increased flooding downstream. Such expe- sive and composed of perennial, intermittent and
rience has influenced the way in which engineering ephemeral streams just before times of peak
geomorphology has evolved (15.4 Summary).
Drainage basins and water transfer

Studies within temperate areas are often based
upon the drainage basin (Figure 15.9) or catch-
ment which is the unit bounded by a watershed or
divide within which precipitation is collected and
conveyed to rivers (Gregory and Walling, 1973).
Such a unit is necessary to analyse the rain and
river budget of temperate areas. The unit can be
used for the calculation of sediment yield from an
area or for the solution of the water balance equa-
tion which relates runoff (R in mm), precipitation
(P), losses by evapotranspiration (Et) and changes
in surface soil or groundwater storage (⌬S) in
the form:
R ⫽ P ⫺ Et ⫾ ⌬S
This indicates why the drainage basin is often

used as the basis for the management and plan-
ning of temperate areas.
Water transferred through drainage basins in
temperate areas, as part of the hydrological cycle,
follows a series of possible routes (Figure 15.12)
and three aspects of this series of routes are poten-
tially important. First is the relative importance of
quick flow and delayed flow in the basin. The
amount of delayed flow is important in relation
to groundwater recharge and therefore to the pro-
vision of ground water supplies. Human activity
tends to increase the amount of quick flow.
Second is the fact that in a temperate landscape
the network of stream channels changes in short Figure 15.12 Routes taken by water in the land-based
periods of time. A network is composed of peren- part of the hydrological system (after Derbyshire et al.,
nial stream channels that flow all the time and will 1979).
GFE-15.qxd 6/24/05 4:04 PM Page 413
Table 15.7 Examples of velocities of water flow in

drainage basins. Based upon data collected in Table 2.2
of Derbyshire et al. (1979).
Type of flow Velocity of water flow
Overland flow 3–15 cm/s on slope of 0.40

Less than 0.1 cm/s on low
slopes with thick vegetation
cover
Vertical Less than 7.5 cm/day in Georgia
percolation
Saturated
throughflow 20 cm/h (0.2–37.2 cm/h saturated
hydraulic conductivity values
collected from various field
Figure 15.13 The relationship of a discharge hydro- measurements)
graph to the extent of the stream network. As the Throughflow 80 cm/day in B horizon in
drainage network extends with more streams flowing so Somerset; 50 cm/day in B/C
the rate of water flow through the drainage basin takes
horizon, 10–20 cm/s in Central
place more rapidly, affecting the shape of the discharge
hydrograph. Wales
Stream channel Average 45 cm/s
flow
stream flow, whereas when the discharge is low it
will be the less extensive network of permanent
streams which will contribute to river flow. These which altogether work with the river rather than
three modes of water flow through a drainage against it, and endeavour to imitate the natural
basin are the end members of a continuum. At the features of river channels rather than to eliminate
head of the ephemeral streams there may be them from the drainage basin (Figure 15.14).
natural pipes which can be up to 0.5 m in diame-
ter and there will also be flow through the soil, Wood debris in channels
either matrix flow or diffuse flow. Although there Prior to the influence of human activity many
is a continuum of water transfers through the temperate areas were forested, and the forest
basin, ranging from water flow through soil pores influenced the way in which environmental
at one extreme, to open channel flow in a large processes operated. A characteristic feature was
river channel at the other, there is a contrast in the that not only did the forest inhibit mass movement
velocity of channel flow and other types of water on slopes but it also contributed debris to the
flow (Table 15.7). If the characteristics of the stream and river channels. In the headwaters of
flow routes are changed inadvertently, it is possi- drainage basins accumulations of coarse woody
ble that water will move through the drainage debris were much more commonly encountered
basin more rapidly, so that an increased incidence than at the present time. Such debris accumula-
of flooding may be an inevitable consequence tions may occupy less than 2% of stream channel
further downstream. length, but can be responsible for half the total
A third aspect of the variety of flow routes is flow resistance, can account for 4% of the vertical
that they are often extended, modified or changed drop in a channel long profile, and for 70% of the
in temperate areas, and such changes can induce sediment stored in the channel. Such debris dams
flooding as illustrated in Table 15.8. Particular can be very densely distributed with average spac-
concern has arisen from the modification of river ing as little as every 2.8 m and so they signifi-
courses by channelization. Measures of stream ren- cantly affect stream channel processes. In central
ovation or stream restoration have been proposed Europe the great phase of deforestation (Darby,
GFE-15.qxd 6/24/05 4:04 PM Page 414
Table 15.8 Significance of changes of water flow routes. Left-hand column shows factors affecting
the flow velocities indicated in Table 15.7.
Increase due to: Consequences include:
Increased overland flow velocities Possible increased erosion of surface.

due to reduction of surface Runoff production time reduced and possible greater
roughness for example by removal peak discharges downstream with more frequent
of vegetation or of surface flooding
irregularities
Land drainage whereby channel Decreased time of travel of water through drainage
flow replaces overland flow basin which can give higher flood peaks and increased
flooding downstream
Channelization of river by decreased Flooding may be increased and erosion may be
roughness and decreased flow accelerated downstream
resistance can give higher stream
velocities
New areas of impervious surface Higher runoff rates, and possible localised erosion of
such as roads allowing surface sediment adjacent to roads
flow where infiltration occurred
previously
Figure 15.14 Effects of channelization (after Coates, 1982). The natural meandering stream (a) contrasts with
the straightened engineered channel (b) and subsequently (c) the channelized stream begins to adjust by becoming
sinuous with developing point bars.
1956) that occurred for 200 years after 1050 AD despite deforestation, although it is estimated that
must have had a significant influence upon the in North America and Europe more than 80% of
supply of wood and upon the dynamic influence riparian corridors have disappeared in the last 200
that it had on rivers. Riparian woodland persisted years (Hupp, 1999).
GFE-15.qxd 6/24/05 4:04 PM Page 415
15.3 Human activity
References have already been made to deforesta-

tion which began in the Mesolithic (c. 9000 years
ago) and the Neolithic (5000 years ago). Large
tracts of land were deforested in Britain before the
Romans arrived in the first century BC, and there
was a great wave of deforestation in western and
central Europe in mediaeval times. In mediter-
ranean lands the natural vegetation was mixed
evergreen and deciduous forest of oaks, pine,
beech and cedars; deforestation began as early as Figure 15.15 Response of a system to disruption
4600 years ago. Temperate North America which (after Graf, 1977). After a disruption the reaction time
was wooded from the Atlantic coast as far west as is when the system accommodates the effect of the dis-
ruption and is followed by the relaxation time which
the Mississippi when the first Europeans arrived,
sees the development of a new equilibrium (B). The
lost more woodland in the following 200 years y axis can be any parameter such as stream channel
than Europe had lost in the previous 2000. size, density of stream channels, or nature of channel
Deforestation, followed by conversion of land to pattern.
grazing and arable land, or use for urbanisation
affects much of the temperate zone and such Table 15.9 Reasons for delays in landscape adjustment.
changes influence the nature and rate of operation
of geomorphological processes. The conse- In Figure 15.15 there is a delay between the disturbance
and the response of the landscape. That delay can be
quences for water and sediment movement are
due to:
most pronounced in cities (Chapter 27), but
throughout temperate areas there have been many gradual changes in see Tables 15.3, 15.5.
effects which have become apparent only long materials
after the changes due to human activity. It has gradual changes in for example where a series
been proposed that the rate of change of a land- landscape of events or a certain
amount of change is needed
scape system is analogous to the decay of radioac-
before adjustment will
tive isotopes so that a rate law of landscape
occur.
change may be envisaged (Graf, 1977) whereby incidence of specific storm events of a particular
there may be a reaction time before a change events size may be necessary to
begins, and then a relaxation time which is neces- initiate change or floods of
sary before the effects of the change are fully a particular size along a
incorporated into the system (Figure 15.15). In river valley.
most parts of the temperate zone the energy avail-
able to environmental processes is less than in
other world areas, but the equivalent of a great engineering solutions to problems, often to pre-
increase in energy can be released by human vent problems arising. Mass movement, slope
activity although the effects may not always be instability and river flooding are examples of
immediately apparent for the reasons listed in major problems that have arisen in many areas
Table 15.9. and engineering solutions have focused on struc-
tures to modify processes, as in the case of slopes
The need for engineering prone to landslip, and structures to preclude
geomorphology process events affecting human activities as exem-
It is the intersection of extensive human activity plified by river channelization and flood protec-
with the various characteristics of temperate tion schemes. Particularly important was the
landscapes that has prompted the need for analysis of time series of processes to establish
GFE-15.qxd 6/24/05 4:04 PM Page 416
the likely occurrence of hazards of a particular floodplain, should now be restored. Such an
frequency and the design of structures to guard approach would significantly alter the way in
against an event of a particular frequency such which river channels are restored (e.g. Figure
as the 100 year flood. Also important has 15.14) and temperate landscapes are managed.
been the investigation of landscape sensitivity This example shows how a spectrum of
(Downs and Gregory, 1995) because it is the issues needs to be considered in environmental
areas near the threshold between stability and management. There are areas where hard engi-
instability that are most sensitive and so require neering provides the only possible solution but
the most careful management. Whereas initial there are others where softer approaches are fea-
solutions to landscape problems of mass move- sible. Issues which have to be considered in the
ment or flooding employed have come to be geomorphic engineering of temperate environ-
thought of as ‘hard engineering’ methods, subse- ments are increasingly important in view of the
quent approaches have been developed to be effects of global change, and the issues may be
more concerned with maintaining the maximum summarised as:
integrity and balance of the total land–water
ecosystem as it relates to landforms, surface 1. Some unchanged landscapes should be main-
materials and processes, an approach which is tained for future generations to appreciate —
called geomorphic engineering (Coates, 1976). including Sites of Special Scientific Interest
(SSSIs), heritage sites, wilderness areas and
Restoration and key approaches for museums.
a geomorphological foundation 2. Hard engineering methods should be avoided
for engineering except where absolutely essential.
High population densities in many parts of the 3. Any engineering solution for the problems of
temperate zone mean that the consequences of a specific area should be designed in relation
human impact have to be minimised and protec- to adjacent areas — an approach described
tion has to be given against natural hazards, as total catchment management or a holistic
which may include earthquakes, floods, ava- approach.
lanches, soil erosion, landslides, subsidence and 4. Management strategies should include an
dam disasters (Alexander, 1999). Since 1990 awareness of stability/instability, sensitivity,
there has been increasing pressure for the man- and risk and uncertainty.
agement of temperate environments to ensure not 5. Knowledge of past environmental systems
only that they utilise the approach of geomorphic should inform decision–making about future
engineering but also involve hard engineering management; acknowledging that the environ-
methods only where absolutely necessary. This ment has a memory and that it is essential that
has therefore meant that ‘soft engineering’ design is undertaken with an awareness of past
approaches have been sought and in some cases environmental change.
restoration, the complete structural and func- 6. Restoration in a general sense should be
tional return to a pre-disturbance state (Gregory, attempted. Wherever possible landscapes
2000), has been undertaken. Some rivers that and environments should be restored to more
were channelized, including straightening, by ‘natural’ conditions. To achieve restoration, a
hard engineering methods have now been variety of overlapping approaches can be envi-
restored to a meandering condition (Brookes and sioned which have been described as recovery,
Shields, 1996). It has been suggested (Brown, re-establishment, enhancement, rehabilitation,
1998) that multiple-channel systems charac- reinstatement, creation and naturalisation as
terised north-west European floodplains prior to well as restoration.
deforestation and river channelization, so that it 7. In determining the management or restoration
can be argued that multiple-channel systems policy for a particular location, community
which involve regular flooding of parts of the views should be taken into consideration.
GFE-15.qxd 6/24/05 4:04 PM Page 417
15.4 Summary Butzer, K. W. (1976) Environment and Archaeology.

Methuen, London.
Chandler, R. J. (1977) The application of soil mechanics
Temperate areas include ample evidence of devel-
methods to the study of slopes. In Hails, J. R. (ed.)
opment under former climates and often have been Applied Geomorphology. Elsevier, Amsterdam,
thought to be the norm against which other land- 157–180.
scapes should be considered. There are several Coates, D. R. (1976) Geomorphic engineering. In Coates,
distinctive types of temperate area. Deposits over D. R. (ed.) Geomorphology and Engineering.
Dowden, Hutchinson and Ross, Stroudsburg, 3–21.
the landscape may be comparatively thin, although
Coates, D. R. (1982) Environmental Geology. Wiley,
many are often inherited from earlier climatic Chichester.
conditions, and may locally be thick. It is often Dalrymple, J. B., Conacher, A. J. and Blong, R. J.
difficult to deduce quickly the character, extent, (1968) A hypothetical nine-unit landsurface model.
and nature of superficial deposits. Sediment trans- Z. Geomorphologie 12, 60–76.
Darby, H. C. (1956) The clearing of the woodland
fers are fairly slow and storage of sediment on
in Europe. In Thomas, W. L. (ed.) Man’s Role in
slopes and in flood plains is common, except Changing the Face of the Earth. University of
where human activity, especially by deforestation, Chicago Press, Chicago, 183–216.
has accelerated change. Stream runoff is depend- Derbyshire, E., Gregory, K. J. and Hails, J. R. (1979)
ent upon the extension and contraction of drainage Geomorphological Processes. Dawson, Folkestone.
Downs, P. W. and Gregory, K. J. (1995) Approaches to
networks; most runoff is generated close to the
river channel sensitivity. Professional Geographer
drainage lines, and most fluvial sediment sources 47, 168–175.
are also found close to streams and rivers. Human Dury, G. H. (1970) The Face of the Earth. Penguin,
activity is very extensive and, whereas hard engi- Harmondsworth.
neering solutions to problems are still necessary Early, K. R. and Skempton, A. H. (1972) Investigations
of the landslide at Walton’s Wood, Staffordshire.
in some situations, in others softer methods
Q. J. Eng. Geol. 5, 19–41.
sometimes involving restoration of more natural Graf, W. L. (1977) The rate law in fluvial geomorphol-
conditions are increasingly attempted. ogy. Am. J. Sci. 277, 178–191.
Gregory, K. J. (ed.) (1983) Background to Palaeo-
hydrology. Wiley, Chichester.
References Gregory, K. J. (1996) Introduction. In Branson, J.,
Brown, A. G. and Gregory, K. J. (eds) Global Con-
Alexander, D. (1999) Natural hazards. In Alexander, D. E. tinental Changes: the Context of Palaeohydrology.
and Fairbridge, R. W. (eds) Encyclopedia of Geological Society Special Publication 115, 1–8.
Environmental Science. Kluwer Academic Publishers, Gregory, K. J. (2000) The Changing Nature of Physical
Dordrecht, 421–425. Geography. Arnold, London.
Brookes, A. and Shields, F. D. (eds) (1996) River Channel Gregory, K. J. and Walling, D. E. (1973) Drainage
Restoration: Guiding Principles for Sustainable Basin Form and Process. Edward Arnold, London.
Projects. Wiley, Chichester. Hupp, C. R. (1999) Relations among riparian vegeta-
Brown, A. G. (1987) Long-term sediment storage in the tion, channel incision processes and forms, and large
Severn and Wye catchments. In Gregory, K. J., woody debris. In Darby, S. E. and Simon, A. (eds)
Lewin, J. and Thornes, J. B. (eds) Palaeohydrology Incised River Channels. Wiley, Chichester, 219–245.
in Practice. Wiley, Chichester, 307–332. Kozarski, S. (1983) River channel adjustment to climatic
Brown, A. G. (1991) Hydrogeomorphological changes change in west central Poland. In Gregory, K. J. (ed.)
in the Severn Basin during the last 15,000 years: Background to Palaeohydrology. Wiley, Chichester.
orders of change in a maritime catchment. In Lewin, J. (1987) Stable and unstable environments —
Starkel, L., Gregory, K. J. and Thornes, J. B. (eds) the example of the temperate zone. In Clark, M. J.,
Temperate Palaeohydrology. Wiley, Chichester, Gregory, K. J. and Gurnell, A. M. (eds) Horizons in
147–170. Physical Geography. Macmillan, Basingstoke,
Brown, A. G. (1998) The maintenance of diversity in 200–212.
multiple channel floodplains. In Bailey, R. G., Liu Changming (2000) A remarkable event of human
Jose, P. V. and Sherwood, B. R. (eds) United Kingdom impacts on the ecosystems: The Yellow River drained
Floodplains. Westbury, Otley, 83–92. dry. Paper presented to the 29th International Geo-
Büdel, J. (1982) Climatic Geomorphology (tr. L. Fischer graphical Congress, Seoul, Korea, 17th August 2000.
and D. Busche). Princeton University Press, Montgomery, D. R. (1997) What’s best on the banks?
Princeton. Nature 388, 328–329.
GFE-15.qxd 6/24/05 4:04 PM Page 418
Starkel, L. (1983) The reflection of hydrological change Trimble, S. W. (1983) A sediment budget for Coon Creek
in the fluvial environment of the temperate zone basin in the driftless area, Wisconsin 1853–1977.
during the last 15,000 years. In Gregory, K. J. (ed.) Am. J. Sci. 283, 454–474.
Background to Palaeohydrology. Wiley, Chichester, Trimble, S. W. (1999) Decreased rates of alluvial sedi-
213–235. ment storage in the Coon Creek basin, Wisconsin,
Stokes, S. and Walling, D. E. (2003) Radiogenic and 1975–1993. Science 285, 123–124.
isotopic methods for the direct dating of fluvial Walling, D. E. (1981) Yellow River which never runs
sediments. In Kondolf, G. M. and Piegay, H. (eds) clear. Geographical Magazine 53, 568–575.
Tools in Fluvial Geomorphology. Wiley, Chichester,
233–267.
Thornes, J. B. (1979) Processes and interrelationships,
rates and changes. In Embleton, C. and Thornes, J. B. Further reading
(eds) Process in Geomorphology. Arnold, London,
378–387. Butzer, K. W. (1976) Geomorphology from the Earth.
Thornes, J. B. (1983) Evolutionary geomorphology. Harper Row, New York, London.
Geography 68, 225–235. Gregory, K. J. and Benito, G. (2003) Paleohydrology.
Tricart, J. (1957) Application du concept de zonalité à Understanding Global Change. Wiley, Chichester,
la géomorphologie. Tijdschrift van het koninklijk 61–164.
Nederlandsch Aardrijkskunddig Genootschap, Lewin, J. (1987) Stable and unstable environments –
Amsterdam 422, 34. the example of the temperate zone. In Clark, M. J.,
Tricart, J. and Cailleux, A. (1972) Introduction to Gregory, K. J. and Gurnell, A. M. (eds) Horizons in
Climatic Geomorphology (tr. C. J. Kiewiet de Jonge). Physical Geography. MacMillan, Basingstroke,
Longman, London. 200–212.
GFE-16.qxd 6/27/05 2:27 PM Page 419
16. Hot Drylands

Mark Lee and Peter Fookes
16.1 Introduction Rainfall can also be subject to extreme spatial

variability. Yair and Lavee (1985) note that local
Hot drylands cover around one third of the earth’s variations in rainfall intensity in the Sinai can be
surface (Figure 16.1) and can be distinguished considerable, with some areas within the same
through the aridity index P/ETP (where P is the storm receiving twenty times more rainfall than
annual precipitation and ETP is the mean annual areas less than a few kilometres away, during a
potential evapotranspiration, based on the particular rainfall event.
Penman formula). Aridity increases as the P/ETP
ratio declines, defining four zones of aridity
(UNESCO, 1979): 16.2 Desert landscapes
1. sub-humid zone P/ETP ⫽ 0.50–0.75 The popular notion is that barren sand dune-
2. semi-arid zone P/ETP ⫽ 0.20–0.50 covered landscapes dominate hot drylands. The
3. arid zone P/ETP ⫽ 0.03–0.20 reality is somewhat different and much more com-
4. hyper-arid zone P/ETP ⫽ ⬍ 0.03. plex. The magnitude, frequency, duration and tim-
ing of rainfall and the scale and significance of
Most drylands are centred on the tropics where wind activity vary between arid zones. The range
the stable, descending air of the Sub Tropical High of bedrocks and geological histories and the avail-
Pressure Zone maintains aridity throughout the ability of mobile sediments all add to the variety of
year. Among the other important factors in control- desert environments. In general, however, there are
ling the distribution of arid zones are the presence two main types of desert landscape (Figure 16.2):
of large land masses which disrupt the zonal pattern
of the global pressure systems (e.g. the Sahara), 1. shield and platform deserts, including
mountain barriers that generate rain-shadow areas inselberg–pediment landscapes and canyon–
in their lee (e.g. the extra-tropical deserts of North scarp–pediment landscapes
and South America) and the influence of cold 2. basin and range deserts.
currents on the eastern margins of the oceans.
Arid areas experience the highest average It is important to bear in mind that much of the
deviation in annual rainfall (Morales, 1977). This current desert landscape detail is probably the
deviation tends to increase as the mean annual product of the major climatic changes experienced
rainfall decreases. For example, a 35 year rainfall throughout the Quaternary (i.e. the last 2–3
series for Elat in the Sinai desert (mean annual million years), during which there were many
rainfall of 25.3 mm), includes 9 years with a quar- phases of alternating semi-arid or humid and dry
ter of the mean, 5 years with double the mean and (arid) conditions. Gerson (1982) and Bowman
1 year with treble the mean (Schick, 1987). et al. (1986) have suggested that recent (i.e. the
‘Cloudbursts’ (associated with convection cells) last 100 000 years) landform development has
are a characteristic feature of the climate, asso- involved responses to varying climatic conditions:
ciated with almost instantaneous peaks in rainfall Semi-arid conditions: these conditions are
followed by a prolonged tail of low-intensity rain. believed to have promoted the development of
419
GFE-16.qxd 6/27/05 2:27 PM Page 420
Figure 16.1 World distribution of deserts (from UNESCO, 1977).
debris flows off hills and scarps, creating Fontes et al. (1985) and Gasse et al. (1987) iden-
relatively smooth talus aprons at the slope foot tified relict lake bed deposits on the northern
(Gerson, 1982) (i.e. a phase of talus apron cre- margins of the Grand Erg Occidental, Algeria.
ation). During this period, colluvial (e.g. debris These lakes are believed to have developed in
flow textured) and alluvial sediments are deposited enclosed depressions (possibly wadi channels
within wide, braided floodplains. blocked by dune movement) and were present
Arid and extremely arid conditions: on hill- between 9300–3000 BP, when groundwater levels
slopes and scarps, this phase would have been were probably close to the surface (nowadays they
characterised by slope wash and gullying, with the are around ⫺50 m).
dissection of the talus aprons producing ‘flatirons’ However, an important characteristic of the
(i.e. a phase of talus apron destruction). As debris desert environment is the marked contrast
flow activity declines, so wadi channels transport between the long-term morphological stability of
less sediment of smaller size and tend to be many landscape features (e.g. the scarps and the
‘misfits’ within the relict floodplains. Later, when pediments) and the dynamic geomorphological
the talus aprons have been largely removed, com- activity of the alluvial plains, sand seas and allu-
bination of flash floods with limited sediment load vial fans.
promotes incision across the floodplains. Most desert soils are granular, with the grad-
Lake formation occurred in enclosed basins ing related to the distance from the uplands.
and low-lying areas, during pluvial (wet) or humid Fine sediments are transported by water or wind
phases, and was accompanied by the depos- and stored in topographic lows or on base-level
ition of muds, usually carbonate. For example, plains. Here, the presence of high and saline
GFE-16.qxd 6/27/05 2:27 PM Page 421
HOT DRYLANDS 421
with typical surface features and with different

engineering behaviour (Figure 16.3; Fookes and
Knill, 1969).
A variety of issues can be encountered in dry-
lands, including:
1. surface erosion and instability
2. difficult excavation
3. behaviour of desert soils
4. availability of aggregates
5. water and sediment movement problems
6. wind blown sand
7. aggressive salty ground.
These issues are considered briefly after an out-
line of the main terrain zones.
16.3 Zone I: the uplands
Mechanical and chemical weathering are active

processes in uplands. However, these areas are
often characterised by bare rock and boulder-
strewn slopes as the rate of weathering is generally
less than the rate of removal by mass movement or
surface erosion. Many uplands comprise a rela-
tively simple landscape of slowly retreating scarps
with detached mesas and buttes, developed in near
Figure 16.2 Characteristic desert landscapes: horizontal sedimentary rocks or tabular volcanic
(A) basin and range topography with alluvial fans rocks. Elsewhere, as in the Zagros Mountains of
and playas; (B) inselbergs and pediment landscape and Iran, there may be complex assemblages of cuestas,
(C) canyon and scarp country (from Goudie and
hogs-backs and strike valleys developed in folded
Wilkinson, 1978).
or faulted rocks. Despite these variations it is possi-
ble to identify a number of broad terrain sub-zones
groundwater tables will create salt weathering (Figures 16.4 and 16.5).
problems. Wind-blown dust (loess) has accumu-
lated across many desert landscapes. For exam- Zone Ia: backslopes
ple, in the Middle East the material is believed to Backslopes are where the caprock, often a duri-
have derived from dust-laden storms from North crust (Chapter 2) is exposed as a low-relief
Africa between 20 000–80 000 years BP. This loess stripped plain (‘slickrock’ slope) that is typically
material is potentially metastable (i.e. will col- boulder or rock strewn, with little or no soil cover.
lapse on wetting, when under a higher load than it Away from the scarp face the backslope may
previously supported). grade into a pediment fronting another scarp
For engineering geomorphological purposes, slope (see Zone IIa below).
the key hot dryland issues can be related to a
number of distinctive terrain units that are char- Zone Ib: scarp slope and free face
acteristic of many desert environments (albeit These are often ‘composite scarps’ with a resist-
in varying combinations). Four main zones can be ant caprock underlain by more erodible strata.
recognised from uplands to base level plains, each This zone is prone to rockfalls and so-called
GFE-16.qxd 6/27/05 2:27 PM Page 422
Zone I Uplands
Zone Ia Backslopes
Z Ib S l
Zone II Pla
Zone I
us slope
irons
Zone IIIb Plain
Zone IV Base Level Plains
Figure 16.3 A simplified desert terrain model (based on Fookes and Knill, 1969; Fookes, 1976, 1978).
Rim
Figure 16.4 Components of a scarp–backslope terrain (from Howard and Selby, 1994).
block-by-block undermining and collapse. Major scarp retreat rates in the order of 0.1–10 m/1000
deep-seated landslides may be present, but these years may be anticipated, depending on the set-
are often an inheritance from previous wetter ting. Yair and Gerson (1974) estimated that the
climates (e.g. Figure 16.6). Where scarp faces average rate of scarp retreat in reef limestones in
comprise impermeable strata over massive per- the Sinai was probably around 1.2–2 mm/year
meable rocks they are often subject to basal over the past 250 000 years. Higher average rates
undermining by seepage erosion. This leads to of up to 6.7 mm/year have been suggested for
cave formation and ultimately collapse. Estimated shales in the Grand Canyon (Lucchitta, 1975).
GFE-16.qxd 6/27/05 2:27 PM Page 423
HOT DRYLANDS 423
the Nahal Yael catchment, Sinai occurred in

5 days and by only 7 runoff events that constituted
20% of the mean annual rainfall.
Runoff coefficients on the bare rock and debris
mantled (scree) slopes can be very high, generally
in the range 15–40% (Yair and Lavee, 1976).
Relatively frequent rainfall events will generate
runoff and overland flow on almost every slope.
The lag from cloudburst to peak flow can be as lit-
tle as 1 hour. However, up to 60–70% of runoff
Figure 16.5 Scarp development model, highlighting
different terrain elements. (A) Generalised form of generated in small catchments may not reach a
compound scarp composed of caprock face and sub- main channel because of infiltration losses (Yair
strate ramp. (B) Erosion of caprock brow and substrate and Lavee, 1985).
ramp, with downward cliff extension into substrate. (C) The probabilities of large, destructive flood
Failure of substrate face and collapse of unsupported
events are relatively high and the frequency of
caprock, producing talus embankment on substrate
ramp. (D) Erosion of caprock brow, talus embankment, historical events suggest that such events can be
and substrate ramp; erosion of talus embankment pro- expected to occur within the main mountain
duces talus flatirons; downward cliff extension into catchments once every 25–50 years.
substrate approaches threshold for next collapse (from Widespread availability of loose, coarse sedi-
Oberlander, 1997).
ment on the mountain slopes and the absence of
vegetation can lead to very large sediment yields
in large flood events.
Zone Ic: rampart Temporary dams across main stream channels
A rampart is an inclined rock surface at the scarp (wadis) can be formed as a result of debris flows
base, often mantled with rockfall debris. These or flash floods from tributary wadis. The subse-
slopes are believed to be formed at the character- quent breach of these dams in storms can give rise
istic friction angle of the debris, under conditions to near instantaneous rises in wadi flow, generat-
where debris removal just balances production ing large bores or ‘walls of water’.
(i.e. ‘Richter’ slopes; Figure 16.7). An example of the large, destructive floods
that occur in the mountains is provided by the
Zone Id: talus slopes and flatirons Wadi Mikeimin (a tributary of Wadi Watir, imme-
Many upland slopes are flanked by talus aprons or diately south of Nuweiba) flood of January 1971
screes, produced by scarp face collapse. These in Sinai (Schick and Lekach, 1987). Detailed
slopes are prone to gullying and debris flow activ- reconstruction of the event indicated that the flow
ity, producing so-called flatirons (i.e. talus rem- peaked at 91.9 m3/s, with a total flow in excess of
nants with marked upslope margins; Figure 16.5). 100 000 m3. The flow deposited 6200 m3 of coarse
Considerable hydrological research has been material as a fan at the junction with Wadi Watir.
undertaken in and around the uplands. Among the The fan dammed the main wadi for 21 months
characteristic features of this environment are during which time a lake, up to 400 m long,
convective cell cloudbursts; very high runoff formed. The first subsequent flood breached it in
coefficients; destructive flood events; large sedi- November 1972, resulting in large flood bores.
ment yields, and temporary dam formation.
Convective cell cloudbursts can have rainfall
intensities of up to 90 mm/hour and considerable 16.4 Zone II: footslopes and fans
spatial variability. Much erosion of channel banks
and beds, and deposition is achieved in rare storm The terrain surrounding the uplands may com-
events. For example, Schick (1977) estimated that prise two contrasting terrain units: gently sloping
99% of erosive work during a 10 year period at rock pediments and alluvial fans.
GFE-16.qxd 6/27/05 2:27 PM Page 424
Figure 16.6 Examples of deep-seated landsliding in the Sahara (from Grunert and Busche, 1980).
Figure 16.7 Development of talus and Richter slopes (from Selby, 1982).
Zone IIa: pediments lithology between the pediment and the adjacent
Pediments are gently sloping surfaces developed uplands. Pediments are believed to be formed by a
in bedrock, often cutting across a range of litholo- combination of lateral planation of mountain front
gies. On occasions there may be no change in streams and sheetwash (inter-rill) and rill erosion
GFE-16.qxd 6/27/05 2:27 PM Page 425
HOT DRYLANDS 425
of weathered materials. It has also been suggested Zone IIb: alluvial fans
that they are relict forms, created by the progressive Alluvial fans are cones of poorly sorted coarse
stripping of deeply weathered old tropical soils. sediment (boulders, cobbles, gravels and some
The surface may have irregular micro- sands) laid down where a channel emerges from
topography, with low knolls, hills (inselbergs) or an upland area onto a plain. When a confined
bedrock ridges that stand above the general pedi- stream leaves the uplands it spreads out, with the
ment level and a network of bedrock-lined and allu- increase in width accompanied by a reduction in
vial channels (wadis). Stream dissection tends to be flow depth and velocity, resulting in sedimenta-
greatest close to the mountain front. Un-dissected tion. In addition, as the stream crosses permeable
areas often have shallow, ill-defined drainage chan- fan sediments the discharge will decrease,
nels with low indistinct interfluves. The regolith is increasing the sediment concentration and fur-
dominated by weathered residual soils, often man- ther enhancing deposition. Not all channels con-
tled by a veneer of superficial deposits, including struct alluvial fans. Large perennial rivers are
blown sands, loess and alluvium (including infilled capable of maintaining channel banks and flow
bedrock channels). Duricrusts may be present. competency.
Erosion and debris flow activity from the The main morphological features of a fan are
upland areas and adjacent steep flanking slopes (Figure 16.8):
supplies sediment (predominantly sand and grav-
1. the drainage basin, typically with steep, debris-
els) to the surrounding plains. Where the sedi-
covered, unvegetated or partly vegetated slopes
ment has been transported further, extensive
2. a feeder channel, the main stream that delivers
gravel spreads (piedmont plains) can developed.
water and sediment to the fan—some fans may
Sediment can be moved across pediments and
have multiple feeder channels
deposited in Zone III because the velocity of
3. a fan apex, the point at which the feeder chan-
water flows and, hence, stream power may remain
nel emerges from the uplands
in the same order of magnitude on both the
4. an incised channel, the downslope extension of
uplands and plains. For example, with reference
the feeder channel across the fan
to the Manning equation (Chapter 11):
5. an intersection point, where the incised chan-
nel merges into the fan slope
r 2/3 S1/ 2
v⫽ 6. an active deposition lobe, below the intersec-
n tion point, flows leave the channel at this point
and spread out across the fan forming a zone
where r is the hydraulic radius, S is the slope gra-
of active deposition.
dient (m/m) and n is the Manning roughness coef-
ficient, if the gradient on the uplands is 0.8 and Fans often comprise a complex sequence of coa-
0.1 on the pediment, and n ⫽ 0.06 on the uplands lescing or segmented sections, often of different
and 0.02 on the pediment, the velocity of flow dates (Figure 16.9; Table 16.1).
would remain almost constant (Cooke et al., The dominant sedimentary processes are sum-
1993). This means that this Zone can experience marised in Figure 16.10 and can involve both
significant sediment transport and erosion as primary construction or aggradation and secondary
flood flow velocities remain high across the rela- fan erosion or degradation, especially by gullying or
tively smooth sloping surfaces. wind erosion. An important feature of fan develop-
The key engineering issues in this zone include ment is the constantly changing pattern of channels.
potentially difficult excavation of bedrock at or This is caused by the progressive infilling and over-
near the ground surface, flood hazard and chan- flows of channels, often as a result of a blockage by
nel instability. However, foundation conditions boulders or mudflows. The primary processes are:
away from drainage channels are likely to be
good, though dependent upon the local bedrock 1. debris flows, transporting sediment supplied
type. The groundwater table is likely to be deep. by a variety of landslide processes including
GFE-16.qxd 6/27/05 2:27 PM Page 426
Figure 16.8 Principal features of alluvial fans (redrawn from Blair and McPherson, 1994).
GFE-16.qxd 6/27/05 2:27 PM Page 427
HOT DRYLANDS 427
not believed to be related to fan construction. Fan

deposits do not generally show planar or trough
cross-bedding typically associated with braided
channel deposits.
Figure 16.11 indicates the relationship between
average fan slope, the dominant processes and
sediment size. Where one process is dominant
(e.g. debris flows) or there is a uniform available
sediment size, there may be a constant fan angle. In
contrast fans with a progressive decrease in sedi-
ment size may be characterised by concave profiles.
Two main fan types can be recognised (Blair
and McPherson, 1994):
Type 1: developed by cohesive clast-rich and
clast-poor debris flows. Fan morphology include
debris flow lobes and levees, and a constant 5–15°
slope. Distal sand-skirts are absent. Because of the
low frequency of debris-flow events the fan surface
is often marked by the effects of secondary
processes, especially gullies, rills and boulder lags
created by the removal of fines by wind action. The
fan sediments comprise stacked sequences of
debris-flow deposits separated by gravel lags pro-
duced by secondary erosion. This type of fan is
associated with catchments underlain by bedrocks
that weather to produce boulders, cobbles, silt and
clay, but little sand.
Type 2: developed by flash floods, especially
sheetfloods. A prominent sand-skirt forms the
Figure 16.9 The development of a segmented fan distal fan margin. The fan surface is smooth with
complex (from Denny, 1967). average slope angles of 2–8⬚ with a progressive
downslope decline in angle. Secondary processes
are effective because of the smaller fan sediment
rockfalls, rock avalanches and slides (Chapter
sizes. This type of sand fan is often associated
8; Costa, 1988); debris flows may range from
with catchments underlain by fractured or jointed
predominantly boulders or cobbles (clast-rich)
granitic plutons, gneiss or friable sandstones.
or muds (clast-poor mudflows)
Fan development may proceed from a rela-
2. sheetfloods (unconfined flows), extensive but
tively steep (10–25⬚) ramp abutting the mountain
short-duration events produced by very intense
front (Stage 1), primarily through landslide
rainstorms
processes. The emergence of debris flows and
3. incised channel flows, generally depositing only
sheetfloods as the dominant processes results in
coarse sediments because of the high flow veloc-
the creation of a gentler fan surface (Stage 2)
ities created by the confinement by near-vertical
(Figure 16.11). Over time fan enlargement may
channel walls. However, some finer sediments
involve the development of segmented fans in
may accumulate along the channel during the
which secondary fans are built on or beyond the
waning flood stage or during low flow events.
original segments (Figure 16.9) or telescope fans
Although many fans have a dense network of where successive fans occur within the bound-
braided distributaries, these are secondary features aries of the initial fan (Figure 16.12).
GFE-16.qxd 6/27/05 2:27 PM Page 428
Table 16.1 Characteristics of alluvial fans of different ages (from Christenson and Purcell, 1985).
Characteristic Young Intermediate Old
Drainage pattern Distributary; Tributary; dendritic Tributary; dendritic

anastomosing or parallel
or braided
Depth of incision ⬍1m Variable (1–10 m) ⬎ 10 m
Fan surface Bar-and-channel Variable, generally Ridge and valley, most
morphology smooth and flat of surface slopes
Preservation of Presently active Incised, but well Basically destroyed,
fan surface preserved wide locally preserved on
flat divides narrow divides
Desert pavement None to weakly None to strongly None (surface destroyed)
developed developed to strongly developed
(surface preserved)
Desert varnish None to weakly None to strongly None (surface
developed developed destroyed) to
strongly developed
(surface preserved)
Calcic horizon None to weakly Weakly to strongly None, carbonate rubble
developed; carbonates developed on surface (surface
disseminated destroyed) to strongly
throughout profile developed hardpan
(surface preserved)
The key engineering issues in this zone include generate the main geohazards and, hence, dominate
debris flows, flood hazard and channel instability. the engineering issues in this Zone.
Foundation conditions on old fan surfaces away
from drainage channels are likely to be good, Zone IIIa: alluvial plains
although rapid horizontal and vertical changes in The main fluvial processes and forms are related
engineering properties may occur. In general, fan to sediment transport and deposition within allu-
materials are a good source of ‘borrow materials’ vial channels. Many desert plains support a com-
for rockfills, gabions and aggregates, although plex dendritic drainage network, with numerous
their chemistry needs to be checked for use in braided stream channels (wadis) and extensive
concretes. floodplains; for the most part, these channels are
relict forms, created during periods of higher
rainfall, although they may be sustained by con-
16.5 Zone III: plains temporary processes. For example, much of the
gross form of the Sahara was probably fashioned
This is a zone of net deposition of sediments sup- during the early to mid-Tertiary, over 66–25 M
plied by the erosion of the uplands and footslopes, years ago, prior to the establishment of full desert
comprising alluvial (Zone IIIa) and wind-blown conditions (e.g. Williams and Faure, 1980).
sediments (Zone IIIc). Elsewhere stone-covered Braided streams tend to be the dominant chan-
rocky surfaces can dominate (Zone IIIb). These nel form, reflecting the interaction between slope
gently sloping plains may comprise considerable gradient, sediment availability, erodible banks
thicknesses of water-borne materials. Two types and highly variable flows. In floods, the flow
commonly occur: sandy stony and silty stony desert occupies the entire channel system between low
(Fookes, 1978). Fluvial and aeolian processes terraces.
GFE-16.qxd 6/27/05 2:27 PM Page 429
HOT DRYLANDS 429
Figure 16.10 The dominant processes operating on alluvial fans: (A) fans dominated by water flows (B) fans
dominated by debris flows (from Blair and McPherson, 1994).
GFE-16.qxd 6/27/05 2:27 PM Page 430
Dryland streams with permanent low flows flows, during which bed materials are completely
may develop compound forms. The low forms mobilised. After the flood has past a new domin-
occupy a single meandering channel whilst high ant channel will develop.
flows spread across a wider braided channel. Dryland flows are ephemeral (i.e. the channel
When the meandering channel is near bankfull is dry between runoff producing storms). Large
discharge it can transport large volumes of sedi- floods do occur. For example, the El Arish flood
ment and erode its banks. At higher flows the sys- of 1975 which resulted from a 48 hour storm
tem becomes braided and flows more directly across much of Sinai, with total rainfall of up to
downslope along the sub-channels. The whole 73 mm at Santa Katharina Monastery (Gilead,
compound system may be occupied by very high 1975). A discharge of 1650 m3/s was recorded
30 km south of El Arish. The flood wave needed
39 hours to cross 250 km from mid-Sinai where it
was first identified. A 1 m high wall of water was
observed. The floodwaters destroyed a railway
bridge and deposited a delta 500 by 300 m on the
Mediterranean shoreline. This is believed to have
been a 1 in 50–100 year event (Schick, 1987).
Major floods have also been reported in the
western Sahara (e.g. the Gur–Saoura–Messaoud
catchment floods of 1890 and 1915 (Mabbutt,
Figure 16.11 The association between fan form, 1977; Figure 16.13).
processes and slope angle, drawn with a 2 ⫻ vertical Sheetfloods occur across the ground surface
exaggeration (from Blair and McPherson, 1994). and when the channel system capacity is exceeded.
Figure 16.12 The evolution of telescopic fans in Death Valley, California (from Dorn, 1988).
GFE-16.qxd 6/27/05 2:27 PM Page 431
HOT DRYLANDS 431
Table 16.2 Changes in channel width along rivers in

the western USA (from Graf, 1988).
River Change Time
Canadian River, 0.8–3.2 km Flood in 1906

Oklahoma
Rio Salado, 15–168 m 1882–1918
New Mexico
Red River, No change 1874–1937
Texas–Oklahoma
Red River, 1.2–0.8 km 1937–1953
Texas–Oklahoma
Cimarron River, 15–366 m 1874–1942
Kansas
Cimarron River, 366–168 m 1942–1954
Kansas
Platte River, 1.16 km–111 m 1860–1979
Nebraska
S.Platte River, 790–60 m 1897–1959
Colorado
N.Platte River, 1.2 km–60 m 1890–1977
Wyoming
Gila River, Arizona 45–90 m 1875–1903
90–610 m 1903–1917
610–61 m 1917–1964
Salt River, Arizona No change 1868–1980
Fremont River, Utah 30–400 m Flood of 1896
become re-established over time as braided sub-

Figure 16.13 The limits of flooding in the Gur– channels infill with sediment, bars and islands
Saoura–Messaoud catchment, Algeria (from Mabbutt, expand and become attached to the banks. A new
1977).
floodplain will develop between the banks of the
braided system, across which the meandering
These floods comprise a sheet of unconfined channel will migrate.
turbulent water moving across the alluvial plain; A notable feature of dryland flows is the ‘trans-
large events are often highly erosive and laden mission losses’ due to infiltration through the chan-
with debris. nel bed. Thus, discharge may decrease or dry up
Large desert floods tend to be formative events, downstream (the reverse is true in temperate rivers)
as opposed to temperate rivers where bankfull and larger flows will extend further than smaller
discharge is believed to be the most effective floods (Figure 16.13).
discharge (Chapter 10), triggering significant
channel changes, including the switch from Zone IIIb: plains and plateau surfaces
meandering to braiding with the resulting increase Extensive, apparently monotonous flat plains are a
in channel width (Table 16.2). Occasionally plan- feature of many deserts. Often the plateaux sur-
form changes are accompanied by incision. For faces are bounded by major escarpments, believed
example, during a flood in 1896 the Fremont to be the product of a combination of differ-
River, Utah entrenched 7 m and a 1.5 km wide ential erosion and structural controls (Zone Ib).
braided channel replaced a 30 m wide meandering A distinctive feature of many desert plains is a sur-
channel (Graf, 1983). Meandering channels may face pavement of closely packed stones (reg or
GFE-16.qxd 6/27/05 2:27 PM Page 432
sarir in the Sahara, gobi in Asia and gibber plain former lake bed deposits (‘mud lions’) to features
in Australia), often in close association with boul- extending several kilometres in length, carved in
der-strewn rock outcrops with a scatter of stones harder rocks such as the Nubian Sandstone or
(hamada). The pavements are often a lag deposit granites (Breed et al., 1997).
left after wind deflation. In some regions, they are Hydrocompaction can be a significant problem
now believed to be the result of accretion of wind- on some desert alluvial soils. Ground fissuring
blown silt and dust. The gradual surface accumu- and subsidence of up to 3 m can result from com-
lation of fine material results in the upward growth paction as clay bonds supporting the voids within
of the soil profile beneath the stone layer which the soil are weakened by wetting (e.g. during
settles evenly on the surface. In this way the uni- floods or after the onset of irrigation).
form input of wind-blown material over a broad
area can lead to the development of monotonous Zone IIIc: sand dunes
level surfaces with relatively consistent thick- Many desert plains contain areas of wind-blown
nesses of silty or fine sandy soils. sand accumulation, either as thin sand sheets,
Streamlined wind-sculptured bedrock hills dune fields (a collection of dunes in an area less
(yardangs) often occur on desert plains and than 30 000 km2) or sand seas (ergs, areas of
plateaux, along with oddly shaped and fluted dunes of varying forms and sizes extending over
hillocks (demoiselles or hoodoos). Yardangs tend 30 000 km2; Figure 16.14). In general, dunes
to be associated with unidirectional wind regimes accumulate where concentrated sand flows con-
and most features are aligned parallel to these verge or where the wind energy declines, as in
winds. They range from small hillocks cut in topographic lows away from upland areas.
Figure 16.14 The Great Western Sand Sea, Algeria: dune assemblages (from Cooke et al., 1993).
GFE-16.qxd 6/27/05 2:27 PM Page 433
HOT DRYLANDS 433
Most dune material is quartz, although some (draa); mesodunes: individual simple dunes
dune areas are dominated by calcium carbonate superimposed on the megadunes; ephemeral
(e.g. the south Wahiba Sands, Oman), gypsum (e.g. dunes: reaching around 1 m in height and devel-
White Sands, New Mexico), sodium chloride oped on constantly changing slopes—they may be
(e.g. southern Tunisia) or clay. Dune sands com-
monly have median grain diameters between
0.2–0.4 mm and range between 0.1–0.7 mm. Dunes
are usually well graded, with a single dune having
particles of similar size and roundedness. Dunes
have poor load-bearing capacities (loose to medium
dense) and can prove difficult to compact as fill.
Dunes exhibit a variety of forms, including
barchans, crescentic ridges, linear, star, reversing
and parabolic (Figure 16.15; Lancaster, 1994;
Thomas, 1997). Transverse and barchan dunes tend
to be migratory forms (see Table 16.3 and follow-
ing discussion on due mobility), whereas linear
dunes are extending or sand-passing forms; star
dunes are accumulating forms. ‘Zibar plains’ are
rolling sand plains without distinct dune forms.
Dune types appear to be related to the volume
of sand and the wind direction variability, as
expressed by the ratio between the total sand drift
potential (DP) and the resultant drift potential RDP
(i.e. the dominant drift direction). Barchans occur
where there is little sand-moving wind variability
and limited sand supply (Figure 16.16). Linear
dunes are associated with restricted sand supply
but variable wind direction; star dunes occur where
sand is abundant with an unstable wind direction.
In sand seas there is a hierarchy of superim- Figure 16.15 Morphological classification of dunes
posed forms: megadunes: complex dune systems (from McKee, 1979).
Table 16.3 Dune activity rates (from Thomas, 1992).
Dune Type Location Dune height (m) Net migration

rate (m/year)
Barchan Pampa de la Joya, Peru 1–7 32–9

Pampa de la Joya, Peru 2.1–3.9 28–16
Algodones, California 6 20
Salton Sand Sea, California 3.1–8.2 27–14
Abu Moharic, Egypt 6–14 8–5
El-Arish, Sinai 1.9–4 13.1–6.2
Mauritania 3–17 18–63
Jafurah Sand Sea, Saudi Arabia ⬍ 30 15
Transverse Namib Sand Sea 2.5 5
Erg Oriental 35 0.3
Erg Oriental 240 0.16
GFE-16.qxd 6/27/05 2:27 PM Page 434
Figure 16.16 The occurrence of major dune types in

relation to sand volume and wind direction variability
(from Wasson and Hyde, 1983).
destroyed and rebuilt as the wind changes, and

wind ripples. Each form is a response to an elem-
ent of the wind regime, megadunes being related
to long-term wind patterns (over 1000s of years),
ripples to hourly or daily events.
Most sand seas and dune fields have sharply
defined margins. Often this is due to topographic
barriers such as scarps or rivers. Elsewhere the
sharpness is explained by ‘sand-shepherding’,
with sand moving quicker over stony plains than
sandy dune surfaces. Sand entering a dune field
suddenly decelerates and may be deposited,
whereas sand leaving the dunes accelerates.
Other important dune forms include: anchored
dunes, which are fixed by topographic obstacles
(e.g. scarp slopes, hills) or plants (Figure 16.17).
‘Climbing dunes’ bank up against a slope as a
sandy ramp. ‘Echo dunes’ form in response to
the wind vortices set up in front of and around
steep obstacles. In areas of high hills there can
be a 1 km wide dune-free corridor separating the
echo dunes from the high ground. ‘Nebkha
mounds’ are formed as individual desert plants Figure 16.17 Types of anchored dunes (from Howard,
trap mobile sand—they have been reported as 1985).
reaching in excess of 10 m in height. Stabilised
dunes are immobilised by cementation or vege- 16.6 Zone IV: base level plains
tation after their formation. The cementation
may be by clays, creating ‘lunettes’ (often found This zone is dominated by the effects of ground-
on the downwind margins of ephemeral lakes); water and salt accumulation, often with soils
carbonates, forming aeolinite (calcarenite) or dominated by wind-blown or water-lain silts and
gypsum. sands.
GFE-16.qxd 6/27/05 2:27 PM Page 435
HOT DRYLANDS 435
In areas of high saline groundwaters the fine- salt attack on roads, structures and buildings. Each
grained soils have restricted load bearing and site needs to be investigated and tested separately.
other engineering performance characteristics.
Dewatering of excavations may be needed (e.g. by Zone IVa: enclosed basins
using a well point system). Filter protection against The lowest areas within enclosed basins are charac-
migration of fines (e.g. geotextiles) may be neces- terised by almost flat vegetation-free ephemeral
sary in underground and surface drainage systems. lakes (playas, salar or pans; Figure 16.18). Similar
However, over time these can get clogged with salts. features may occur in topographic lows in both ped-
Calcrete duricrusts (carbonate cemented sands iment and alluvial plains (salinas) or on coastal
or clays) may be present in layered sediments, plains (sabkhas). Playas are net accumulation zones
especially in sabkhas. The local salt regime may be of both fine-grained clastic and non-clastic sedi-
very complex, varying with the seasons. Salts will ments with inputs received from aeolian and fluvial
contaminate fine and coarse aggregates and lead to processes and through groundwater exchanges
Figure 16.18a Idealised cross sections through sabkha, playa, salt playa and salinas for engineering purposes
(from Fookes, 1976).
GFE-16.qxd 6/27/05 2:27 PM Page 436
Figure 16.18b Simple playa basin model (from Hardie et al., 1978).
(capillary rise or from through-flow). Groundwaters 4. Phreatophyte discharge playas: where brack-
are often saline because of evaporite salts dissolved ish groundwaters are close to the surface,
from the local or regional bedrock or from sea water phreatophytes and other plants may flourish
intrusion. The predominance of silts and clays and form large mounds. Ring fissures and
means that playa surfaces tend to be impermeable, localised surface subsidence may develop as a
encouraging the accumulation of surface runoff. result of the groundwater being lowered
They are also prone to desiccation phenomena, beneath the plants.
including raised-rim polygons, ‘boils’ of deliques- 5. Spring discharge playas: mounds can develop
cent salts, drying cracks and salt crusts. around saline springs, rising up to 30 m high
Five principal types of playa can be recognised: (theoretically they can grow up to the level of
the piezometric surface). Where artesian water
1. Surface water discharging playas: the playa emerges onto a playa surface large spring pools
surface is above the capillary limit and tends to can develop, up to 4 m deep and 5 m wide, with
be hard, dry and composed of fine-grained near-vertical sides. These features probably
sediments. form through a combination of surface solution
2. Capillary movement playas: the surface tends and subsidence.
to be saline, soft, permeable and loosely com-
pacted, often with micro-relief of up to 15 cm. Sabkhas are saline coastal flats, noted for
Where there is capillary discharge, ‘self-rising’ their high concentrations of carbonates and evap-
ground and salt-thrust polygons can develop. orites. They often form part of a complex assem-
3. Direct groundwater playas: these features are blage of coastal landforms, including beach ridges,
characterised by thin salt crusts of halite or dunes, barriers and lagoons (Figure 16.19).
gypsum of thicker salt pavements. The surfaces Flooding may occur at high tides and during large
can be wet, soft and sticky. Solution pits and runoff events. Various types of salts accumulate as
sinkholes may be present. a result of capillary rise from the saline intrusion.
GFE-16.qxd 6/27/05 2:27 PM Page 437
HOT DRYLANDS 437
Figure 16.19 Generalised section through an Arabian sabkha (from Akili and Torrence, 1981).
Figure 16.20 The salt cycle in deserts (from Cooke et al., 1993).
The concentration of salts in playas, salinas plateaux, and further extended by mass movement,
and sabkhas can be viewed as part of a ‘salt- deflation and surface water erosion (Albritton
cycle’ (Figure 16.20). Salts are dissolved from et al., 1990). Wind erosion may have been signifi-
rocks in the uplands and transported by water to cant during the Quaternary arid phases, with
depositional basins where they precipitate out. 20 000 km3 of removed debris providing the source
Wind erosion results in their removal by deflation of sand for the downwind sand seas.
and transport either back to the upper catchment
or to an adjacent basin.
16.7 Surface erosion and instability
Zone IVb: depressions
Broad, low-lying depressions, bounded by steep A single layer of angular or sub-rounded gravels
scarps or terrace sequences, occur in many deserts immediately above stone-free material covers
(e.g. the Qattara, Siwa and Jaghbub depressions in many desert pediments and plains. Although
the eastern Sahara). Although these features have stone pavements provide excellent surface protec-
been viewed as the product of extensive and pro- tion against wind and water erosion they are easily
longed wind deflation to below the groundwater destroyed by vehicle traffic or during road and
table, they probably reflect a variety of factors pipeline construction. The resulting accelerated
working in combination. The Qattara depression surface erosion leads to increased dust and sedi-
(the lowest point is 134 m below sea level) in mentation problems and creates notable environ-
Egypt is believed to have been excavated as a mental impacts. Once disrupted, pavement
stream valley and subsequently modified by recovery may be very slow. Little is known of the
solution weathering of the adjacent limestone rate of pavement development in different desert
GFE-16.qxd 6/27/05 2:27 PM Page 438
environments, although evidence from stone Weak gypsum crusts are found in warm
pavements on sequences of alluvial terraces in deserts, where mean annual rainfall is less than
Israel suggests that a near-continuous pavement 250 mm (Watson, 1985). The gypsum accumula-
(i.e. stones touch each other) may take over tions may reach 5 m thick, ranging from bedded
14 000 years to form (Amit and Gerson, 1986). crusts, bands of large crystals (desert roses—
Elsewhere, off-road vehicle tracks in the Mohave these can reach up to 5 m high) or surface crusts
Desert, USA have recovered within decades of alabastine gypsum.
(Webb and Wilshire, 1983). An important distinction needs to be made
between duricrusts developed within the soil
profile (i.e. they form thin indurated layers or
16.8 Excavation pans which ‘float’ within the soil profile) and
those formed by the enrichment of the bedrock
The presence of bedrock at or close to the surface (i.e. ‘caprocks’). Floating duricrusts can
ground surface on upland ramparts and pedi- generally be ripped out by conventional excava-
ments may come as a surprise if, for example, tion equipment. Well-developed caprock duri-
talus or alluvium was anticipated. However, it is crusts can be strong and massive and, hence,
the occurrence of duricrusts that can create a present significant problems for excavation. They
challenge to planning the method and rate of often require heavy ripping with hydraulic break-
excavations (the need to bring in heavier equip- ing or even blasting, especially if they are
ment than originally expected, or to use blasting, unjointed. As excavation can be a major cost
can prove costly and cause delays). Duricrusts driver in pipeline projects it is prudent to establish
(Chapter 2) are generally the product of strong the nature and extent of duricrusts at an early
upward leaching in response to an excess of stage in project planning (e.g. Fookes et al.,
evapotranspiration over rainfall (e.g. Dixon, 2001).
1994; Watson and Nash, 1997). Desert duri-
crusts include silcrete, calcrete and weak gypsum
crusts. 16.9 Behaviour of desert soils
Silcrete is formed through the accumulation
of silica, creating a brittle, intensely indurated In general the soil grading, in situ density and
material comprising quartz clasts cemented by Casagrande classification are good guides to the
crystallised or amorphous silica. Well-developed potential engineering performance of desert soils.
silcrete horizons are between 0.5 and 3 m thick, However there are important exceptions to this in
though multiple crusts may exceed 5–7 m. Silcretes that soils may often appear to be of high strength
are believed to be ancient features and are rela- due to very low moisture contents (high suctions
tively common in Australia, southern Africa, parts in fine-grained soils) or cementing due to clay or
of the Sahara and the Arabian Gulf (e.g. Milnes salt bonds between larger particles. Whereas in a
and Thiry, 1992). temperate climate coarse-grained materials having
Calcrete is formed through the accumulation less than 12% fines will generally be free draining,
of calcium carbonate. Calcretes may range from in deserts some sands with up to 30% fines of
calcified soils, calcium carbonate nodules or con- low plasticity may behave as free-draining soils
cretions, honeycomb calcretes (coalesced nodules) (Fookes and Gahir, 1995).
to mature hardpans (Netterberg, 1980). Hardpans Unless truly cemented, such soils can either
are typically 0.3–0.5 m thick, although superimpos- collapse on wetting and loading if they are ini-
ition of sequences of calcretes can create thick, tially loose (e.g. aeolian deposits), or swell sub-
complex profiles. Calcrete can form relatively stantially on wetting if comprising high-plasticity
quickly (over the 100–1000 year timescale) and clay. There is a need in classifying arid soils to
is a common feature of deserts with extensive identify those that are truly cemented and those
carbonate-rich bedrocks. whose strength can decrease or volume change
GFE-16.qxd 6/27/05 2:27 PM Page 439
HOT DRYLANDS 439
dramatically on wetting (e.g. Fookes and Parry, 16.10 Aggregates

1994).
A simple slaking test whereby samples are left The principal aggregate sources in drylands
in water for a minimum of an hour is useful in dis- (Fookes and Higginbottom, 1980) are shown in
tinguishing between real and apparent cementa- Figure 16.21 and Table 16.4 and include alluvial
tion. Swelling potential of clays can be estimated sands and gravels, coastal beach sands, dune
from the difference in behaviour of samples in sands, and crushed rock aggregate.
water and carbon tetrachloride (Sridharan, 1998). Alluvial sands and gravels are sourced from
Once the effects of cementation have been taken fans and alluvial channels in Zones II and III.
into account, normal engineering classifications Leaching by floodwaters ensures that the chloride
are a good guide—the potential collapse of sandy and sulphate levels are relatively low. However,
and silty soil is indicated by low relative density the deposits may be cemented by carbonate or
and uniform grading, while the swelling potential gypsum and, hence, require suitable processing. It
of clay is indicated by high plasticity and low is usually essential to screen and wash fan gravels
in situ moisture content. carefully before use.
Some desert soils may be prone to collapse or Coastal beach sands are often carbonate grains
swelling affecting foundations, roads and pipe- and may have a narrower and finer grading than that
lines. These problems can be the result of changes preferred for concretes. Shells may need to be
in moisture content associated with, for example, screened out to avoid adverse effects on concrete
broken drains, irrigation of gardens, or failure to structure or density. Often chlorides and sulphates
deal correctly with surface water, air conditioner may be unacceptably high, especially where borrow
or roof drainage. In other situations the problems pits are worked close to the groundwater table and
may simply be the response to the presence of a where salts can accumulate through evaporation.
new building changing the balance between cap- Dune sands—wind-blown sands of Zone
illary rise and surface evaporation. IIIc—have narrow, fine grading (c. 50–600 ␮m)
Where potential problems of collapse or with silt contents of around 10% and, hence, are
swelling are identified, special precautions may not ideal sources, but can be used. The rounded
have to be taken with building foundations. These nature of the grains gives them poor binding qual-
may range from attempts at prevention of the ities. Experience has shown that it is possible to
problem (e.g. double-sleeving of water service identify dunes or parts of dunes with better grad-
connections to buildings) to the use of deep foun- ing or to improve the resource by blending with
dations, the design of which must take account of other material.
the expected soil movements. Crushed rock aggregate can be obtained from
The shortage of water in deserts is often a prob- the uplands (Zone I) sources or from duricrusts in
lem in applying normal compaction methods. Zones IIa and III. In some areas, such as the Gulf
However it has been found that many soils have a States, crushed calcrete has been an important
second maximum in their water content/dry source of aggregate (Fookes and Higginbottom,
density curves at near zero water content, giving 1980). However, dolomite-rich rocks may be
a dry density close to that achieved at the normal liable to alkali/carbonate reaction. Bedrock insel-
optimum moisture content. Dry compaction is bergs in Zone IIa may prove excellent sources of
often, therefore, a viable option. However, it must suitable material.
be recognised that such compacted materials
have relatively high air voids and may be prone
to collapse settlements on wetting. Special con- 16.11 Water and sediment problems
struction techniques may therefore be necessary
to prevent wetting of the material in, for example, Major flood events and the associated erosion and
road embankments during periods of heavy deposition are rare and, hence, are often unex-
rainfall. pected. Predictions of discharge and sediment
GFE-16.qxd 6/27/05 2:27 PM Page 440
Major landforms as aggregate resources in drylands
Mountains
Including peaks, ridges, plateau surfaces, steep (excluding precipitous) slopes,* deep valleys and canyons, wadis,
river terraces* and alluvial fans,* bounding scarp slopes.* Forms vary with rock type and the evolutionary history of
the area.
Pediments and Alluvial Fans
Rock pediment,† fan* and bajada,† with occasionally inselbergs* or salt domes0 forming locally high ground.
Plains
Occur downslope of pediments or alluvial fans without a distinct boundary and may include a whole variety of fea-
tures including: alluvial† and colluvial plains,† wadi channels and flood plains, dune fields,0 salt domes,0
inselbergs,* and extensive stone pavement surfaces.0
Playa Basins
Enclosed depressions receiving surface runoff from internal catchments or within escarpment zones.† They
frequently contain lakes (either temporary or permanent), lake beaches, evaporite deposits0 and may be strongly
influenced by aeolian, fluvial and salt processes in their baseland zones.
Coastal Zones
These include beach ridges† (formed at periods of higher sea-level or during exceptional storms), Sabkhas,0 mud
flats,0 beach0 and foreshore,0 estuaries0 and deltas.0
*Normally a major source of aggregate, conditional on suitable mineralogy
†
May be a reasonable source, depending on specific characteristics
0
Normally should not be used for aggregates.
Figure 16.21 Dryland landforms and aggregate sources (redrawn from Cooke et al., 1982).
yield can be severely hindered by the absence of drainage networks are relict forms, only partly
meteorological or hydrological records. In addi- active even in the largest floods. The limited spa-
tion, the assessment of runoff based on catchment tial extent of high intensity storms may mean that
area (Chapter 11) are less reliable in drylands only part of a catchment is actively generating
than temperate areas because most desert runoff at any one time.
GFE-16.qxd 6/27/05 2:27 PM Page 441
HOT DRYLANDS 441
Table 16.4 Landforms and aggregate potential (adapted from Fookes and Higginbottom, 1980; Cooke et al., 1982)
Feature Aggregate Type Nature of Material Engineering Properties/Fills Potential Volume
Bedrock Crushed rock Angular, clean and Depends on rock type and Very extensive
mountains suitable for all rough texture processing. Good fills
types of aggregate
Duricrust Road base and Often contains salts. May be self-cementing with Often only small
sub-base Needs crushing time. Quite good rock fill deposits of good
and processing quality
Upper alluvial Concrete and when When crushed and Good compaction as fill Often very
fan deposits crushed for road screened angular – extensive but good
base and clean. Otherwise quality material
sometimes dirty may only be in
and often rounded small deposits
Middle alluvial May make road Often high fines Often good compaction as Small to extensive
fan deposits base content fill. Good bearing
Lower alluvial Generally not High fines content. Poor bearing and as fill Small to extensive
fan deposits useful Needs processing
Other piedmont Variable, locally Dirty, rounded and Good bearing capacity Very extensive
plain alluvium good concrete well graded. May (dense sediments) and as but good quality
aggregate contain salts fill. Locally poor due to material in small
clay and silt layers deposits
Old river Concrete Variable Difficult to locate in field Often deposits
deposits patchy and thin
Dunes Generally not good Usually too fine and Fills of poor compaction Locally extensive
rounded
Interdunes Fine aggregate Coarse to fine, angular. Fills of poor compaction Very localised
Needs processing
Salt playas and Not suitable Very salty and Poor. Special random fill Locally extensive
sabkhas aggressive
Coastal dunes Generally not Generally too fine Special fills often poor Sometimes
suitable and rounded compaction extensive
Storm beach Fine aggregate may Sufficiently coarse Random fill Sometimes
be sharp and/or for concrete sand extensive
salty Clean after processing
Foreshore Generally not Fine, rounded sand. Salt contaminated, but might Locally extensive
suitable make random fill
In the absence of hydrological records, it is observations of channel-form (i.e. cross-section

necessary to use simple field observations to pro- area, bed roughness and estimated water depths).
vide an approximation of the flooding and scour The method involves the following steps:
hazard. Estimates of historical flood depths and
velocities can be made from the theoretical rela- 1. Field mapping of the channel cross-section: a
tionships between channel form and bed grain field sketch or surveyed profile (e.g. Figure
size. Richards et al. (1987) have suggested a 16.22) is required at representative channel
method for estimating wadi discharge through the cross-sections, defining separate morphological
reconstruction of credible flood events from field units within each cross-section (e.g. the active
GFE-16.qxd 6/27/05 2:27 PM Page 442
Figure 16.22 Sample cross section through Wadi Dhamad, Saudi Arabia illustrating the field data needed for
assessment of flows (redrawn from Richards et al., 1987).
channel, terraces, point bars etc.). For each unit, is then calculated from the continuity equation:
estimates need to be made of slope gradient,
Q ⫽ wRv
unit width, bed sediment size (D50 ⫽ median
grain size, and D84 ⫽ the 84th percentile grain where w is the channel width.
size) and the depth of recent and possible max- 3. Estimation of the overall discharge at a par-
imum flows through the channel section. ticular flood stage: this involves summing the
2. Estimation of the discharge passing through a discharges for each unit at a particular flood
unit at a particular flood depth (i.e. stage), from depth to estimate an overall discharge in a
the mean depth and channel width and estimates particular event.
of the mean velocity. This involves calculating 4. Development of a flood stage-discharge rela-
velocity of flow through a unit at a particular tionship: this can be achieved by assessing the
flood depth using the Darcy-Weisbach equation: discharge at a range of credible event sizes (i.e.
events that have left a morphological imprint
v ⫽ 8g R s /f
on the channel).
where g is the acceleration due to gravity; R is
the hydraulic radius (m); s is the slope gradient An indication of the maximum flood depth
and f is the friction factor. The friction factor experienced in a channel can be estimated from
(f ) can be calculated using the empirical rela- the largest mobile bed material. Baker and
tionship of Limerinos (1969): Ritter’s (1975) method involves an empirical rela-
1 冑 f = 1.16 + 2 log ( R D84 )
tionship between the largest grain size considered
mobile (Dmax, mm) and the bed shear stress (␶0):
As a check on the results, the friction factor ␶0 ⫽ 0.00044 Dmax1.85
can also be calculated from:
Given the channel slope (S ) the bed shear stress
n ⫽ 0.0151 D501/6 can be related to the channel hydraulic radius R by:
where n is Manning’s roughness coefficient, and ␶0 ⫽ ␥RS
f ⫽ 8 g n2/R1/3
where ␥ is the specific gravity of water. Alter-
where R is the hydraulic radius of the channel natively the critical velocity (Vc ) needed to move
(i.e. cross section area/wetted perimeter). The the largest bed material can be estimated from:
velocity of flow through a unit at a particular
Vc ⫽ 0.18Dmax0.49
flood depth can be calculated using either or
both estimates of the friction factor. Discharge Careful morphological mapping can also
(Q) through each unit at a particular flood depth be used to develop flood hazard maps (e.g.
GFE-16.qxd 6/27/05 2:27 PM Page 443
HOT DRYLANDS 443
Figure 16.23 Flood hazard across an alluvial fan (from Kesseli and Beaty, 1959).
Figures 16.23 and 16.24). Useful strategies for the road over controlled lengths (dependent on the
managing flood risk include avoidance, flood likely maximum flows). Although the road may be
warning systems (e.g. Figure 16.25) or adopting closed during the actual flood event, damage to the
route alignments that minimise channel-crossing road is localised and even prevented if the road is
problems (e.g. Figure 16.26). A technique some- armoured over the low lengths.
times used to protect minor roads crossing areas
prone to flash flooding is to have most of the road
on low embankment, but have sections at intervals 16.12 Dune mobility
at or slightly below normal surface level which act
as inverted bridges (referred to by engineers as Mobile dunes may blow over roads or accumulate
‘Irish bridges’) and allow flood water to pass over against buildings and structures. The migration
GFE-16.qxd 6/27/05 2:27 PM Page 444
Figure 16.24 Flood hazard map of Suez City, Egypt (from Cooke et al., 1982).
rate is inversely proportional to the dune wave- Alternatively, the change in dune location on
length and height (Simons et al., 1965): sequential aerial photographs or satellite imagery
can provide an indication of the actual migration
Migration Rate ⫽
rates during that period.
Sand Transport Rate
(Cross Section/Wavelength ⫻ Height) ⫻ Height ⫻ Bulk Density
Control measures include surface stabilisation
and the use of fences. Surface stabilisation of
Thus, the bigger the dune, the less mobile it is source areas can be achieved with gravel, gypsum,
(Figure 16.27; Table 16.3). However, above a saline water or a chemical or oil spray. Oil has been
certain size dune mobility appears to be fairly used successfully to stabilise large areas at low cost
uniform. (Kerr and Nigra, 1952). Permanent stabilisation
GFE-16.qxd 6/27/05 2:27 PM Page 445
HOT DRYLANDS 445
Figure 16.25 Flood warning system on the Gulf of Aqaba shoreline, Egypt (from Schick and Sharon, 1974).
can often only be achieved through the establish- A frequently used method of aeolian sand control
ment of a vegetation cover. However, in many dry- is the construction of panels or fences to divert or
lands this is not a straightforward task and requires stop sand transport reaching a vulnerable site (e.g.
careful consideration of suitable plant types and a village). It is common practice to use a system of
management. Control of movement can also be three parallel fences to provide effective protection
achieved through transport pathways with fences. to a site (Figure 16.28). Sand accumulates mainly
GFE-16.qxd 6/27/05 2:27 PM Page 446
Figure 16.26 Proposed alignments of the Super-Indus highway, Pakistan. Note that the alternative route minimises
channel crossings by crossing fans towards the apex (from Cooke et al., 1982).
and first around fence 1; as the effectiveness of the plains (Zone IV) and coastal margins creates
first fence declines, sand accumulates increasingly a range of engineering hazards related to salt
at fence 2 and so on (Cooke et al., 1982). Once the weathering (Figure 16.29). These problems largely
fence is covered it will no longer act as a sand trap, occur as a result of post-construction accumula-
as new sand arriving from upwind will simply tion of salts due to capillary rise of groundwater.
sweep across its smooth surface and be carried The impact of chlorides and sulphates on concrete
downwind (i.e. no reduction in transport capacity). and roads is perhaps the most significant problem
Increasing the height will restore the trap effi- (e.g. Fookes and Collis, 1975a, 1975b, 1976;
ciency. Fookes and French, 1977; French and Poole, 1976;
Fookes et al., 1985).
Figure 16.30 provides a simple model of
16.13 Aggressive salty ground the soil-groundwater conditions. The upper sur-
face of the capillary fringe (limit of capillary
The presence of saline groundwater close to the fringe) can vary from being at depth (below foun-
ground surface in topographic lows, base level dation level) to potentially above ground level
GFE-16.qxd 6/27/05 2:27 PM Page 447
HOT DRYLANDS 447
Figure 16.27 The relationship between dune size and mobility (redrawn from Cooke et al., 1993).
Figure 16.28 The use of porous fences as sand traps (from Kerr and Nigra, 1952).
GFE-16.qxd 6/27/05 2:27 PM Page 448
Figure 16.29 Salt problems in drylands (from Fookes and Collis, 1976).
GFE-16.qxd 6/27/05 2:27 PM Page 449
HOT DRYLANDS 449
Figure 16.30 A simple soil-groundwater model for drylands (from Jones, 1980; after Fookes and French, 1977).
route selection and the evaluation of road main-

tenance priorities. An alternative approach is to
convert the simple soil-groundwater model of
Figure 16.30 into hazard classes:
1. Zone I: no hazard from groundwater as the

limit of the capillary fringe is below foundation
depth. Protection may be needed in some cases.
2. Zone II: the limit of the capillary fringe is
below ground surface but sufficiently close to
Figure 16.31 The interrelationships between soil mois- it to affect foundations. Concrete and struc-
ture zones in drylands (from Fookes and French, 1977).
tures will need protection.
3. Zone III: the limit of the capillary fringe is
(resulting in salt crusts on the surface). Buildings potentially above ground level so that both
or structures can allow capillary rise to extend to foundations and susperstructures are at risk.
its full height, leading to damp-stained walls and Use full protective measures if construction is
the deposition of salts in the fabric. The capillary necessary.
fringe limit is the location where the capillary 4. Zone IV: the water-table is within 0.5 m of the
fringe intersects the rising ground surface and ground surface for most of the year so that
marks the boundary between areas where only foundations are emplaced in water and there is
foundations are at risk and areas where both foun- potential for capillary rise to well above
dations and the superstructure may be attacked. ground level. Avoid these areas.
Figure 16.31 shows how the aggressive
ground hazard can be related to observable soil Figure 16.32 presents a hazard zone map of Suez
moisture zones. This approach can be useful in City, Egypt, developed using this method.
GFE-16.qxd 6/27/05 2:27 PM Page 450
Figure 16.32 Aggressive ground hazard map, Suez City Egypt (from Cooke et al., 1985).
16.14 Summary resources. Indeed, deserts present a range of

unique problems to engineers, including surface
Although desert landscapes show considerable instability following disruption of stone pave-
variability, the simple classification of the terrain ments, difficult excavation of strong and massive
into uplands, footslopes and fans, plains and base- duricrusts, unfamiliar (to western engineers)
level plains provides a framework for identifying behaviour of the soils, restricted availability of
and evaluating engineering hazards and assessing aggregates, ephemeral flood activity, dune mobil-
GFE-16.qxd 6/27/05 2:27 PM Page 451
HOT DRYLANDS 451
ity and aggressive ground. In addition, active Denny, C. S. (1967) Fans and pediments. American
slope instability on upland slopes, hydro-com- Journal of Science 265, 81–105.
Dixon, J. C. (1994) Duricrusts. In Abrahams, A. D.
paction and dust generation are issues that need to
and Parsons, A. J. (eds) Geomorphology of Desert
be considered in many desert landscapes. Environments. Chapman and Hall, London,
82–105.
Dorn, R. I. (1988) A rock varnish interpretation of allu-
References vial fan development in Death Valley, California.
National Geographic Research 4, 56–73.
Akili, W. and Torrence, J. K. (1981) The development Fontes, J. Ch., Gasse, F., Callot, Y., Plaziat, J. -C.,
and geotechnical problems of sabkha, with prelimin- Carbonell, P., Dupeuble, P. A. and Kaczmarska, I.
ary experiments on the static penetration resist- (1985) Freshwater to marine-like environments
ance of cemented sands. Quarterly Journal of from Holocene lakes in northern Sahara. Nature
Engineering Geology 14, 59–73. 317, 608–610.
Albritton, C. C., Brooks, J. E., Issawi, B. and Swedan, A. Fookes, P. G. (1976) Road geotechnics in hot deserts.
(1990) Origin of the Qattara Depression, Egypt. Highway Engineer 23, 11–29.
Bulletin of the Geological Society of America 102, Fookes, P. G. (1978) Middle East—inherent ground
952–960. problems. Quarterly Journal of Engineering Geology
Amit, R. and Gerson, R. (1986) The evolution of 11, 33–49.
Holocene reg (gravely) soils in deserts—an example Fookes, P. G. and Collis, L. (1975a) Problems in the
from the Dead Sea region. Catena 13, 59–79. Middle East. Concrete 9, 12–17.
Baker, V. R. and Ritter, D. F. (1975) Competence of rivers Fookes, P. G. and Collis, L. (1975b) Aggregates and the
to transport coarse bedload material. Geological Middle East. Concrete 9, 14–19.
Society of America Bulletin 86, 975–978. Fookes, P. G. and Collis, L. (1976) Cracking and the
Blair, T. C. and McPherson, J. G. (1994) Alluvial fan Middle East. Concrete 10, 14–19.
processes and forms. In Abrahams, A. D. and Fookes, P. G. and French, W. J. (1977) Soluble salt
Parsons, A. J. (eds) Geomorphology of Desert damage to surfaced roads in the Middle East.
Environments. Chapman and Hall, London, 354–402. Journal of the Institution of Highway Engineers
Bowman, D., Karnieli, A., Issar, A. and Bruins, H. J. 24, 10–20.
(1986) Residual colluvio-aeolian aprons in the Negev Fookes, P. G. and Higginbottom, I. E. (1980) Some
Highlands (Israel) as a paleoclimatic indicator. problems of construction aggregates in desert areas,
Palaeography, Palaeoclimatology, Palaeoecology 56, with particular reference to the Arabian peninsula 1:
89–101. Occurrence and special characteristics. Proceedings
Breed, C. S., McCauley, J. F., Whitney, M. I., of the Institution of Civil Engineers, Part 1 68,
Tchakarian, V. P. and Laity, J. E. (1997) Wind ero- 39–67.
sion in drylands. In Thomas, D. S. G. (ed.) Arid Fookes, P. G. and Knill, J. L. (1969) The application of
Zone Geomorphology. Wiley, Chichester, 437–464. engineering geology to the regional development of
Christenson, G. E. and Purcell, C. (1985) Correlation Northern and Central Iran. Engineering Geology 3,
and age of Quaternary alluvial-fan sequences, Basin 81–120.
and Range Province, southwestern United States. Fookes, P. G. and Parry, R. H. G. (eds) (1994)
Geological Society of America Special Paper 203, Engineering Characteristics of Arid Soils. Balkema,
115–122. Rotterdam.
Cooke, R. U., Brunsden, D., Doornkamp, J. C. and Fookes, P. G. and Gahir, J. S. (1995) Engineering per-
Jones, D. K. C. (1982) Urban Geomorphology in formance of some coarse-grained arid soils in the
Drylands. Oxford University Press. Libyan Fezzan. Quarterly Journal of Engineering
Cooke, R. U., Brunsden, D., Doornkamp, J. C. and Jones, Geology 28, 105–130.
D. K. C. (1985) Geomorphological Dimensions of Fookes, P. G., French, W. J. and Rice, S. M. (1985) The
Land Development in Deserts—with special refer- influence of ground and groundwater geochemistry
ence to Saudi Arabia. Nottingham Monographs in on construction in the Middle East. Quarterly
Applied Geography No. 4, University of Nottingham. Journal of Engineering Geology 18, 101–128.
Cooke, R. U., Warren, A. and Goudie, A. S. (1993) Fookes, P. G., Lee, E. M. and Sweeney, M. (2001)
Desert Geomorphology. UCL Press. Pipeline route selection and characterisation, Algeria.
Costa, J. E. (1988) Rheologic, geomorphic and sedimen- In Griffiths, J. S. (ed.) Land Surface Evaluation for
talogic differentiation of water floods, hyperconcen- Engineering Practice. Geol. Soc. Engineering Group
trated flows and debris flows. In Baker, V. R., Kochel, Special Publication 18, 115–121.
R. C. and Patton, P. C. (eds) Flood Geomorphology. French, W. J. and Poole, A. B. (1976) Alkali aggressive
Wiley, New York, 113–122. aggregates and the Middle East. Concrete 10, 18–20.
GFE-16.qxd 6/27/05 2:27 PM Page 452
Gasse, F., Fontes, J. C., Plaziat, J. C., Carbonel, P., Lucchitta, I. (1975) Application of ERTS images and
Kaczmarska, I., De Deckker, P., Soulie-Marsche, I., image processing to regional geologic problems and
Callot, Y. and Dupeuble, P. A. (1987) Biological geologic mapping in Northern Arizona: Part IVB, The
remains, geochemistry and stable isotopes for the Shivwits Plateau. National Aeronautical and Space
reconstruction of environmental and hydrological Administration Technical Report 32-1597, 41–72.
changes in the Holocene lakes from North Sahara. Mabbutt, J. A. (1977) Desert Landforms. MIT Press,
Paleaogeography, Paleaoclimatology, Paleaoecology Cambridge, Massachusetts.
60, 1–46. Mckee, E. D. (ed.) (1979) A Study of Global Sand Seas.
Gerson, R. (1982) Talus relicts in deserts. Israel US Geological Survey Professional Paper 1052.
Journal of Earth Sciences 31, 123–132. Milnes, A. R. and Thiry, M. (1992) Silcretes. In
Gilead, D. (1975) A Preliminary Hydrological Appraisal Martini, I. P. and Chesworth, W. (eds) Weathering,
of the Wadi El-Arish Flood 1975. Mimeogr. Rep. Soils and Paleosols. Developments in Earth Surface
Israel Hydrological Service, Jerusalem. Processes 2. Elsevier, Amsterdam, 349–377.
Goudie, A. S. and Wilkinson, J. (1978) The Warm Morales, C. (1977) Rainfall variability—a natural
Desert Environment. Cambridge University Press. phenomenon. Ambio 6, 30–33.
Graf, W. L. (1983) Flood-related change in an arid region Netterberg, F. (1980) Geology of southern African cal-
river. Earth Surface Processes and Landforms 8, cretes. I: Terminology, description, macro-features
125–139. and classifications. Transactions of the Geological
Graf, W. L. (1988) Fluvial Processes in Dryland Society of South Africa 83, 255–283.
Rivers. Springer-Verlag, Berlin. Oberlander, T. M. (1997) Slope and pediment systems.
Grunert, J. and Busche, D. (1980) Large-scale fossil In Thomas, D. S. G. (ed.) Arid Zone Geomorphology.
landslides at the Msak Mallat and Hamadat Manghini Wiley, Chichester, 135–163.
escarpment. In Salem, M. J. and Busrewil, T. (eds) Richards, K. S., Brunsden, D., Jones, D. K. C. and
Geology of Libya. Academic Press, London. McCaig, M. (1987) Applied fluvial geomorph-
Hardie, I. A., Smoot, J. P. and Eugster, H. P. (1978) Saline ology: river engineering project appraisal in its
lakes and their deposits: a sedimentalogical approach. geomorphological context. In Richards, K. S. (ed.)
In Matter, A. and Tucker, M. (eds) Modern and River Channels: Environment and Process.
Ancient Lake Sediments. International Association of Blackwell, Oxford, 348–382.
Sedimentologists Special Publication 2, 7–41. Selby, M. J. (1982) Hillslope Materials and Processes.
Howard, A. D. (1985) Interaction of sand transport with Oxford University Press.
topography and local winds in the northern Schick, A. P. (1977) A tentative sediment budget for an
Peruvian coastal desert. In Barndorff-Nielson, O. extremely arid watershed in the southern Negev. In
E., Moller, J. -T., Rasmussen, K. and Willetts, B. B. Doehring, D. O. (ed.) Geomorphology in Arid
(eds) International Workshop on the Physics of Regions. Allen & Unwin, London, 139–163.
Blown Sand. University of Aarhus Memoir 8, Schick, A. P. (1987) Hydrologic aspects of floods
511–544. in extreme arid environments. In Baker, V. R.,
Howard, A. D. and Selby, M. J. (1994) Rock slopes. In Kochel, R. C. and Patton, P. C. (eds) Flood Geo-
Abrahams, A. D. and Parsons, A. J. (eds) Geomorph- morphology. Wiley, New York, 189–203.
ology of Desert Environments. Chapman and Hall, Schick, A. P. and Lekach, J. (1987). A high magnitude
London, 123–172. flood in the Sinai desert. In Mayer, L. and Nash, D.
Jones, D. K. C. (1980) British applied geomorphology: (eds) Catastrophic Flooding. Allen & Unwin,
an appraisal. Zeitschrift für Geomorphologie Suppl. London, 381–410.
36, 48–73. Schick, A. P. and Sharon, D. (1974) Geomorphology
Kerr, R. C. and Nigra, J. O. (1952) Eolian sand control. and Climatology of Arid Watersheds. Mimeo
American Association of Petroleum Geologists, Report Dept. Geography, Hebrew University,
Bulletin 36, 1541–1573. Jerusalem.
Kesseli, J. E. and Beaty, C. B. (1959) Desert Flood Simons, D. B., Richardson, E. V. and Nordin Jr, C. F.
Conditions in the White Mountains of California (1965) Bedload equation for ripples and dunes.
and Nevada. US Army Quartermaster Research and US Geological Survey Professional Paper 462H,
Engineering Center, Technical Report, EP-108. H1–H9.
Lancaster, N. (1994) Dunes. In Abrahams, A. D. and Sridharan, A. (1998) Volume change behaviour of
Parsons, A. J. (eds) Geomorphology of Desert expansive soils. In Eiji Yanagisawa, Nobuchika
Environments. Chapman and Hall, London, 474–505. Moroto and Toshiyuki Mitachi (eds) Problematic
Limerinos, J. T. (1969) Relation of the Manning coeffi- Soils. Balkema, Rotterdam.
cient to measured bed roughness in stable natural Thomas, D. S. G. (1992) Desert dune activity: concepts
channels. US Geological Survey Professional Paper and significance. Journal of Arid Environments 22,
650-D, 215–221. 31–38.
GFE-16.qxd 6/27/05 2:27 PM Page 453
HOT DRYLANDS 453
Thomas, D. S. G. (1997) Sand seas and aeolian Webb, R. H. and Wilshire, H. G. (eds) (1983)
bedforms. In Thomas, D. S. G. (ed.) Arid Zone Environmental Effects of Off-road Vehicles. Springer,
Geomorphology. Wiley, Chichester, 373–412. New York.
UNESCO (1979) World Distribution of Arid Regions. Williams, M. A. J. and Faure, H. (1980) The Sahara and
Laboratoire de Cartographie Thématique du the Nile. Balkema, Rotterdam.
CERCG, CNRS, Paris. Yair, A. and Gerson, R. (1974) Mode and rate of
Wasson, R. J. and Hyde, R. (1983) A test of granulo- escarpment retreat in an extremely arid environment
metric control of desert dune geometry. Earth (Sharm el Sheikh, southern Sinai Peninsula).
Surface Processes and Landforms 8, 301–312. Zeitscrift für Geomorphologie 21, 106–121.
Watson, A. (1985) Structure, chemistry and origins Yair, A. and Lavee, H. (1976) Runoff generative
of gypsum crusts in southern Tunisia and the process and runoff yield from arid talus mantled
central Namib Desert. Sedimentology 32, 855– slopes. Earth Surface Processes 1, 235–247.
875. Yair, A. and Lavee, H. (1985) Runoff generation in
Watson, A. and Nash, D. J. (1997) Desert crusts and arid and semi-arid zones. In Anderson, M. G. and
varnishes. In Thomas, D. S. G. (ed.) Arid Zone Burt, T. P. (eds) Hydrological Forecasting. Wiley,
Geomorphology. Wiley, Chichester, 69–108. Chichester, 183–200.
GFE-17.qxd 6/25/05 1:59 PM Page 454
17. Savanna
Michael F. Thomas
17.1 Introduction to savanna Australia, Eucalyptus savannas cover a transi-

environments tional zone between the ‘wet tropics’ of the north
and north-east and the interior steppe and desert.
The nature and status of ‘savanna’
Savanna environments are generally regarded as Geomorphology in the savanna zone
typifying the seasonal tropics, beyond the range The largest areas of savanna coincide with the
of the tropical forests, but excluding the extremi- extensive plateau surfaces of the southern,
ties of desert and steppe. The present-day vegeta- Gondwana continents, usually attributed to plana-
tion consists of a variety of open deciduous tion in the late Mesozoic or early Cenozoic eras.
woodlands, woodland/grassland mosaics and These are largely cratonic areas underlain by
areas of open grassland. In Brazil this vegetation Archaean rocks and by undeformed platform sed-
is known as ‘cerrado’ and drier thorn woodlands iments, though important areas of volcanic rocks
as ‘caatinga’. The transition to tropical rainforest occur, notably on the Deccan of India and in the
occurs where rainfalls approach c. 1500 mm/year east African rift zone. The ancient land surfaces
and the dry season is less than 4 months in dura- are dominated by leached, nutrient-poor soils
tion. Where dry season conditions become more (oxisols and ultisols), and are commonly under-
extreme and rainfall declines below 800 mm/year lain by deep saprolites (see Fookes, 1997), which
the deciduous broadleaved species give way to may be capped by duricrusts. They are also char-
Acacia thorn woodland and semi-arid steppe, acterised by groups of ‘inselbergs’, prominent
which includes the ‘Sahel’ of northern Africa, and rocky hills, often of granite and usually domed in
the ‘Sertao’ of north-east Brazil. Internally the so- profile (Figure 17.1). The combination of open
called savanna zone can be highly complex, the
physiognomy of the vegetation reflecting stress
factors such as seasonal soil moisture deficits,
natural and man-made fires, grazing and other
land-use pressures. Because much of this varia-
tion is between the content of trees, shrubs and
grasses in the ground cover, and this does not
reflect in a simple manner the rainfall amount,
the status of the savannas is uncertain. Savanna
woodland vegetation, dominated by Brachystegia
spp., covers a broad crescent-shaped zone of
Africa, stretching from Senegal in the north-west
to Moçambique in the south-east. In South America
the Llanos is an island of savanna (La Grand
Sabana) dominating parts of Venezuela and adja- Figure 17.1 Monolithic granite domes at
Dombashawa, Zimbabwe, typical of the ‘older granite’
cent countries, while the cerrado and caatinga plutons within the cratonic landscapes of former
stretch from north-east Brazil to Chile, covering Gondwanaland, associated with boulder strewn lower
important areas of Argentina and Paraguay. In slopes derived from former weathering.
454
GFE-17.qxd 6/25/05 1:59 PM Page 455
SAVANNA 455
vegetation, extensive plains and steep, isolated redistribution of these saprolites. The renewal of
hills led many geomorphologists to ascribe these weathering and saprolite formation in the later
features specifically to a climatic control that Neogene and throughout the Quaternary has been
reflected the strong seasonal contrasts in moisture demonstrated more recently from work in
supply (Tricart, 1965; Büdel, 1982). In turn this Queensland (Feng and Vasconcelos, 2001; Li and
led to the argument that a suite of landforms char- Vasconcelos, 2002; Vasconcelos and Conroy,
acteristic of savanna could be defined within a 2003). These finding are likely to apply to other
scheme of climatic geomorphology. There are, continents, indicating that weathering systems
however, serious objections to this approach. The have fluctuated in their effectiveness, responding
idea that there is a unique combination of features to regional climatic and hydrological controls
that characterise savanna areas is misleading; over long time periods.
many landforms and deposits extend into both During the Quaternary (the last 2 Ma) major
drier and wetter climatic zones. This is a result of fluctuations of rainfall accompanied the tempera-
changing global climates and plate positions ture variations corresponding to the ice ages
since the break-up of Gondwanaland in the late of higher latitudes. During the Last Glacial Maxi-
Mesozoic. The nature and extent of the savanna, mum (LGM), at 22–21 BP, much of the tropics
therefore, often reflects the geology and history of experienced a dry phase lasting several millennia,
the landscape, more than the operation of recent when rainfall may have been reduced by
processes. Nonetheless, when the details of local 30%–60% compared to present-day levels. The
landscapes are considered, the imprint of Neogene semi-deciduous forests were almost certainly con-
dissection, Quaternary climate changes and cur- verted to open savanna during the late Quaternary
rent processes all become increasingly important dry periods, such changes possibly affecting a
for a full understanding of the regolith and its 200–400 km depth of the rainforest. On the drier
chemical and physical properties. margins of the savanna these climates, although not
always in phase, were sufficiently arid in places to
Climate history and the distribution of convert the landscape to desert. The widespread
savanna residual and alluvial sands became subject to aeo-
The climates of areas that now support deciduous lian processes and now form linear or other dune
woodland and savanna have been both wetter and systems, which are extensive in the drier savanna
drier in the geological past. Humid tropical con- regions of Africa and Australia.
ditions were more widespread during the late
Mesozoic and Palaeogene. The growth of the
Antarctic ice sheet during the Miocene appears to 17.2 The nature of the regolith in
have been associated with a drying out of climate savanna areas
in many parts of the southern hemisphere tropics
and sub-tropics. According to studies in South This climatic history has created many complex-
America, the weathering systems were effectively ities in the altered substrates of savanna land-
‘switched off’ by this event (Alpers and Brimhall, scapes. In the first place, the zone experienced
1988; 1989; Vasconcelos et al., 1992; 1994). The prolonged weathering, often to great depths, not
Miocene aridity may also help explain why sedi- only beneath older planation surfaces, but also
mentary formations such as the Kalahari Sands in the growing mountain chains of the Andes
and Continent Terminal in Africa, and possibly (Alpers and Brimhall, 1988; 1989). Across the
the Barreras Formation in Brazil, are found today surviving plateaux a weathered mantle, in places
in wetter savanna and forest areas. The prolonged more than 100 m deep, remains in place. But
humidity of the Mesozoic and Palaeogene had left the geochemical and physical attributes of this
a legacy of widespread, very deep saprolites mantle vary greatly, not only according to the
across the continents. Declining rainfall during parent rocks, but also as a result of a complex
the Miocene may have led to the erosion and hydrological history.
GFE-17.qxd 6/25/05 1:59 PM Page 456
Climate and weathering products the wetter areas, alkali-poor 1 : 1 clays (kaolins)
Despite the complexities of climate change, it has progressively dominate the saprolite.
been shown that within the rainfall belts between The chemical composition of saprolite is of
750–1500 mm/year, kaolinite clays predominate in great significance to its behaviour as an engineering
well-drained surface soils (Pedro, 1968). However, material. The presence of smectite in significant
smectites become increasingly abundant at rain- quantities leads to marked increases in swelling in
falls below 1000 mm/year, especially in receiving the presence of moisture. Where Ca2⫹ is the domi-
sites for surface and groundwater flows. A 60 km nant cation, the figure ranges from 45–145%. In
west to east transect in southern India reported by Na⫹ rich clays it rises to ⬎ 1400%, and such sodic
Bourgeon (1991) and discussed by Pedro (1997) clays are highly dispersible. Kaolinite, as a 1 : 1
demonstrates a transition from highly evolved, lattice clay, absorbs less water and may expand by
kaolinised saprolites (French: alterites) in the wet less than 10%. Both smectite and kaolinite com-
zone (⬎ 2500 mm/year), towards friable sandy monly bond with free iron oxides, which are pro-
materials (French: arènes; German: grus) in the duced in the weathering of biotites and other dark
dry zone (⬍ 750 mm/year) (see Figure 17.2). The minerals. The Fe2O3 acts as a weak cement that
transition from ‘massive’ weathering to ‘limited’ binds kaolinite and other clay minerals, leading to
weathering (Bourgeon, 1991) depends on the their appearance in the silt fraction of the saprolite,
availability of drainage water. According to this and effectively reducing the apparent clay content
study, plagioclase will remain intact at depth in the of the material. The delicate aggregates formed in
driest zone, whereas biotite and amphibole will this way exhibit open, ‘cardhouse’ microfabrics.
alter to low-charge 2 : 1 phyllosilicates (smectite These produce collapsible soils, which react to
clays), which retain Calcium. As weathering inten- excess moisture or loading by rearrangement of the
sifies in the friable zone, so alteration of plagio- platy clay particles with a reduction of void space
clase, and the formation of high-charge 2 : 1 and loss of volume. Where the Fe2O3 is amorphous
phyllosilicates (sericite, illite), leads to the export rather than crystalline it can be remobilised under
of calcium and the retention of Potassium. In saturated (reducing) conditions. Iron behaviour is
complex, however, and Fe2O3 will crystallise, com-
monly as goethite or hemetite, to form segregations
(mottles) and pisolitic nodules, usually within a few
metres of the ground surface. Where the iron
content is high enough (20–40% ⫹) individual
mottles or pisoliths may aggregate to form laterite
(or ferricrete), which is resistant to erosion.
In drier areas and beneath more recent or dis-
sected land surfaces, sandy regolith low in clay
content, similar to the ‘grus’ found in colder
climates predominates. In contrast to clay-rich
saprolites that have 20–50% clay content, these
less weathered materials typically contain ⬍ 15%
clay. According to engineering weathering classifi-
cations, the grus generally corresponds with Zone
III, grading up into more clay-rich materials in
Zones IV and V (Geological Society, 1990; Migoń
and Thomas, 2002).
Figure 17.2 Geochemical and mineralogical charac- Duricrusts

teristics of the weathering zone along a transect in The regolith in savanna areas is often charac-
southern India (redrawn from Bourgeon, 1991). terised as having a widespread cover of surface or
GFE-17.qxd 6/25/05 1:59 PM Page 457
SAVANNA 457
Table 17.1 The occurrence of common duricrusts in disputes about the definition of ‘laterite’ (Taylor
relation to humidity of climate. and Eggleton, 2001). The behaviour of iron duri-
crusts, however, depends largely on the crys-
Humid
tallinity and continuity of the cementing oxides,
↓ Alucrete or ‘bauxite’ (Al2O3 as gibbsite and
boehmite) and the amount of clay. But in some detrital ferri-
↓ Ferricrete or ‘laterite’ (Fe2O3 as goethite cretes the clastic host sediment is an important
and hemetite) component. Duricrusts that overlie clay-rich
↓* Silcrete or ‘porcellanite’ (SiO2, saprolite are often subject to marginal collapse as
silica) the clay is washed out by underground seepage.
↓* Calcrete or ‘caliche’ (CaCO3, Pseudo-karst hollows are also found in duri-
with Mg as, dolocrete) crusted terrain, and appear to result from volume
↓ Gypcrete (CaSO4, gypsum) loss in the underlying saprolite, due in part to the
Arid
formation of underground drainage channels.
* Calcrete and silcrete do not follow a clear order of Both calcrete and silcrete deposits are found
formation in relation to present-day climate. widely in the drier savanna, where precipitation
falls below c. 800 mm/year. Calcium liberated dur-
near-surface, resistant duricrust. Duricrusts are ing the weathering of plagioclase, or derived from
exposed and hardened horizons first formed in carbonate rocks initially travels as the dissolved
saprolites or sediments, by the precipitation of bicarbonate in groundwater, but rapid degassing
hydrated oxides of Aluminium, Iron and Silicon, or occurs in shallow, aerated rivers, leading to pre-
of Calcium as the carbonate or sulphate (Geological cipitation of tufa deposits. CaCO3 precipitation
Society, 1990; Fookes, 1997). As a broad generali- also takes place as pods and domes of calcrete
sation, the formation of these materials follows a (also known as caliche or kunkar) in floodplain
humidity gradient, as indicated in Table 17.1. environments, probably due to water-table fluctua-
This sequence has been the subject of debate tions. The precipitation of secondary silica to form
and the true situation is complex. However, most silcrete can arise from different processes: by the
alucrete (or bauxite) deposits are found in rain- loss of Al2O3 from kaolinised saprolite, in which
forest environments, while it is possible that many case it forms within the profile, or by release of
ferricretes formed first as soft, mottled clay (often silica from rocks and saprolites and its subsequent
termed laterite) horizons under forests and have precipitation from silica-rich groundwater. This
become indurated, with the formation of crys- may possibly take place under conditions of
talline goethite and hemetite skeletons, as a result impeded drainage, where seasonal desiccation
of dehydration. This has been due both to climate prevents the export of the silica. But long-distance
change and to the lowering of the water-table transport of SiO2 has also been proposed to
during the fluvial dissection of landscapes. Laterite explain the occurrence of silcrete around fluviatile
in this sense was formed as part of a weathering and past lacustrine basins, as in central Australia
profile, but ferricrete also develops in transported (Stephens, 1971; Taylor and Eggleton, 2001).
sediments, particularly in savanna regions, wher- The removal of iron, silicon and even alu-
ever seasonal water-table fluctuations occur near minium from saprolites that already display
the surface. The distal ends of piedmont slopes, advanced weathering increases the void ratio, and
which are often alluvial fans or amorphous collu- collapsible fabrics beneath well-drained sites can
vium, are commonly observed to contain shallow result. Elsewhere, the precipitation of ions in
ferricrete. Alluvial deposits, especially coarse transport and their accumulation as smectite
sands and gravels are also readily cemented clays, creates environments where swelling clays
by iron in floodplains and river terraces. Such can present problems for both agriculture and
detrital ferricrete can later appear as a resistant road foundations. On and beneath some ancient
caprock as a result of relief inversion. The many surfaces, karstic and pseudo-karst (silica) features
varieties of Fe-duricrust have led to continuing have developed. These can lead to cavernous
GFE-17.qxd 6/25/05 1:59 PM Page 458
macro-voids and to surface features ranging from sediment downslope to form valley fills. These
micro-forms such as rillenkaren (surface channels processes contributed to the formation of surface,
or corrugations) through sinkholes to macro-tower or near-surface, ‘lag’ gravels.
karst. The rocks affected range from ultra-basic
igneous and metamorphic rocks to silica-cemented Regolith materials and landforms
sandstones and conglomerates, in addition to the The savanna zone is, therefore, associated with
true karst of carbonate rocks. the widespread occurrence of duricrusts that form
tabular residual hills and bench-lands in valleys
Climate change and regolith (Figure 17.3). In many cases these hills represent
transformation an inversion of relief, and the duricrusts com-
The mineralogical changes brought about by the monly contain transported gravels. In fact, many
climatic transitions induced by plate-tectonic ferricretes appear unconnected with a ‘laterite
movement and global climate changes have been profile’ and demonstrate iron cementation of a
documented by Tardy and Roquin (1998). transported regolith (Ollier and Galloway, 1990).
Weathering profiles developed during the Mesozoic Eroded and redistributed regolith materials accu-
have since experienced both geochemical destabil- mulate to form residual sands and gravels on
isation and physical dismantling as a consequence uplands (Figure 17.4) and valley-side slopes, while
of altered hydrology and increasing dissection of fine sands and clays accumulate in valley floors.
the continents during the last 100 Ma. Profiles This can occur on the scale of small, shallow val-
have been truncated by surface erosion (Butt, leys known as dambos in Africa (Figure 17.5;
1987) and the products of dismantling redistrib- Clark, 1974; Thomas and Goudie, 1985) and also
uted across the landscape by wind action and on the much broader scale of large alluvial basins,
sometimes by water. The Kalahari Sands probably as in the middle Nile valley (Williams et al.,
resulted from these processes, and ‘sandplains’ are 1998). The seasonal drying of soils and regoliths
encountered widely across the other continents to considerable depths beneath elevated sites is
(Fairbridge and Finkl, 1984). Those profiles that largely responsible for the formation of Fe-Mn
survived intact have experienced an altered hydrol- pisoliths, while the strong evaporation from
ogy, usually involving a fall of water-tables, spo- floodplains and valley floors leads to the incorpo-
radically through this 100 million years (Thomas, ration of divalent Ca2⫹ ions to form vertisols
1989; 1994; Twidale, 1991). This has been a nec- containing smectite clays.
essary consequence of fluvial dissection of the Strong catenary relationships or toposequences
continental margins, as well as a result of drier cli- of soils and slopes were first described in soils
mates. Bauxitisation may have occurred in wetter from the seasonal tropics, reflecting the strong
areas, by de-silicification of kaolinitic weathering
mantles while, in less humid areas (Valeton et al.,
1991), induration of Iron-rich zones led to the for-
mation of Fe-duricrusts. In a similar way silcretes
now form cappings to low hills formed in sapro-
lite, particularly in the interior of Australia.
Quaternary climate changes affected regoliths
mainly by the impacts of periodic aridity and
accelerated surface erosion, a result of reduced
vegetation cover. Some weathering profiles
became truncated, exposing horizons formed at
depth to the desiccating effects of surface envi-
ronments. Duricrusts became strongly indurated;
pisolitic gravels developed from iron-rich soil Figure 17.3 Duricrust (ferricrete) ‘breakaway’ in
mottles, and surface erosion removed fine northern Nigeria.
GFE-17.qxd 6/25/05 1:59 PM Page 459
SAVANNA 459
leads to the break-up of the ferricrete and the

redistribution of derived gravels across the lower
slopes of valleys and around isolated hills (mesas).
Subsequently, secondary or ‘footslope’ ferricretes
may form within these deposits (Dowling, 1966;
Figures 17.6 and 17.7).
It is, however, mistaken to think that the
savanna zone is blanketed by ferricrete. In most
areas these deposits occur as discrete caps or have
formed around the margins of larger hills. On the
other hand surface gravels are common on convex
summits and these materials contain duricrust
fragments, unrolled vein quartz and, commonly,
rolled pebbles of fluvial origin (Figure 17.4).
These gravels appear to mark the complex history
and lowering of the land surface over long time
periods. Related stonelines of gravel also appear
within soil profiles, usually within 1–2 m of the
surface, and these have led to considerable debate
concerning their origins (Figures 17.8; 17.9;
Alexandre and Symoens, 1989; Thomas, 1994).
Such stonelines can create hydraulic discontinu-
ities in the profile, encourage sub-surface water
flows and also precipitation of Fe3⫹ to form ferri-
crete. While thin stringers of stones may arise from
simple downslope processes, many stonelines are
Figure 17.4 Residual gravels forming a surface ‘lag’ more complex, commonly exceeding 50 cm in
deposit in northern Sierra Leone. The gravels contain thickness beneath surfaces of gentle slope. They
fragments of ferricrete, vein quartz and quartzite
pebbles. are often viewed as lag deposits, dating from
episodes of aridity and powerful sheetwash, which
have been followed by accumulation of finer sedi-
seasonal contrasts in slope and groundwater ment and soil, often by bioturbation (especially by
hydrology (Milne, 1935). Such contrasts are termites) and soil development. Others may be
heightened by the dissection of ancient saprolites, degraded river terrace gravels.
leading to the formation of residual hills with the The rate of topsoil formation by termites
exposure of different weathering zones across the and other soil fauna has been quoted as
slope. The exposure of duricrusts results from 0.05–0.5 mm/year (see Thomas, 1994), which
the loss of topsoil and dissection, which creates might imply that a topsoil of 2 m depth could
low cliffs or ‘breakaways’ (Figure 17.3). Water have formed during the Holocene (10 ka, at
penetrating the duricrust via numerous cavities and 0.2 mm/year). Large termitaria (e.g. Macrotermes
cracks will find subterranean pathways towards natalensis) can reach 10 m in height and 30 m
nearby valleys. This can create a form of lateritic in diameter, with densities of 2–5 per hectare
pseudo-karst and water will issue from tunnels (Figure 17.10; Aloni, 1975). The material brought
below the duricrust capping, often leading to its to the surface is fine grained (⬍ 2 mm, c. 20%
collapse. Closed depressions can form below sum- clay) and is washed over the intervening areas by
mit areas, where removal of material from beneath rainsplash erosion, but very large mounds may
the capping has led to sagging of the crust. The have an age of at least 700 years (Watson, 1974).
undermining and erosion of the breakaways The action of termites has a major impact on
GFE-17.qxd 6/25/05 1:59 PM Page 460
Figure 17.5 Dambo valley at Kankamo, Zambia, showing the flat floor and sedimentary fill.
sub-surface conditions, contributing to the parti- environments. But they are seldom described as,
tioning of the soil profile (Figure 17.11), with the or grouped with, savanna. For this reason, a range
penetration of chambers and channels continuing of forms and deposits that characterise the ‘old
to depths of many metres. The contribution of lands’ of the seasonal tropics will be emphasised
small Microtermes species to the total rate of soil here (see Table 17.2; Figure 17.12).
accumulation is important.
Regional weathering and landform
evolution
17.3 Landform assemblages and In essence Table 17.2 lists residual and trans-
surface materials ported materials derived from the weathering of
bedrock, and regarded as ‘regolith’ (Ollier and
Regolith features and landforms Pain, 1997; Taylor and Eggleton, 2001). Bedrock
Although savanna is associated with the broad forms are ignored. These materials co-exist
plateaux and lowlands of the ancient fragments of within regional landscapes as quazi-regular
Gondwanaland, there are of course grasslands patterns. But before these are discussed, the pat-
and grassland–tree mosaics present in montane tern of weathering of the mantle itself deserves
GFE-17.qxd 6/25/05 1:59 PM Page 461
SAVANNA 461
Figure 17.6 Particle-size distributions and plasticity data for high-level and footslope laterites from the Fika-Nafada
area, Nigeria (after Newill and Dowling, 1968).
Figure 17.8 Colluvium overlying stoneline gravel

above a granite saprolite, adjacent to a granite outcrop
on the Jos Plateau, Nigeria.
Figure 17.7 (A) Formation of footslope laterite circumstances) below the ground surface
(ferricrete) and relation to high-level laterites and (B) (Figures 17.13; 17.14). Many studies have indi-
distribution of ferricretes (laterites) in the present land- cated regional and local variations in the depth of
scape (after Newill and Dowling, 1968).
the weathered mantle. On a regional scale,
greater depths and differentiation correspond
comment. The base of the weathered zone is with the survival of ancient land surfaces of rela-
sometimes abrupt and referred to as the weather- tively low relief. Many parts of the central
ing front or basal weathering surface, and it African plateau (or ‘High Veld’), and significant
may lie more than 30 m (⬎100 m in special sections of other cratonic areas, such as east
GFE-17.qxd 6/25/05 1:59 PM Page 462
Figure 17.9 Thick stoneline gravels below a planate Figure 17.10 Large termitarium (of Macrotermes
land surface in eastern Zambia. natalensis) within a deciduous forest in Shaba
Province, Republic of the Congo.
Figure 17.11 Composite soil–saprolite profile showing typical combination of termite layer, colluvium, stoneline,
plinthite (incipient laterite), saprolite and transitional (saprock) zone, with hypothetical grading curves. The 5 m
depth may be much greater than shown, up to 15 m or more.
central Brazil and the Yilgarn Block in western most contain weathering products older than
Australia, fall within this category. Radioisotope 10 Ma (Vasconcelos et al., 1992; 1994).
and other studies converge to indicate that many During the ensuing period, partial dismantling
of these regoliths date from the Mesozoic and of these ancient regoliths has occurred, and led to
GFE-17.qxd 6/25/05 1:59 PM Page 463
SAVANNA 463
Table 17.2 Common associations of landforms and materials in the seasonal tropics. This list excludes bedrock
surfaces, and does not specifically list Landforms characteristic of erosion (modified from Thomas, 1994, after
Thomas, 1974 and Butt, 1987).
A. Residual materials
I. Interfluve and plateau residuum
1. Residual soil on rocks, duricrusts and saprolites
2. Duricrust of weathering profile (tabular summits and plateaux) including ferricrete, silcrete, and calcrete
cappings (alucrete rare in savanna)
3. Duricrust derived rubble (e.g. from in situ breakdown of fericrete)
4. Blockfields of rock cores (residuals from weathering)
5. Eluvial ‘white sands’, resulting from collapse of weathering profile
II. Hillslope residuum
6. Duricrust cliffs — cliffs around 1. — see also B I
7. Upper zones of weathering profile
8. Lower zones of weathering profile (saprolite and corestones on steep and lower slopes)
9. Boulder-controlled slopes; with tors as residual hills (convex summits and steep slopes ⬎ 20⬚)
B. Transported materials
I. Interfluve and plateau deposits
1. Duricrusts containing transported clasts (tabular summits) including ferricrete, silcrete cappings (often
impregnating sediments)
2. Gravel layers (exposed or as stone-lines) comprised of duricrust fragments, rolled and unrolled quartz
3. Sandplains: aeolian sand sheets, alluvial and colluvial sands (possibly from eluvial sands)
II. Hillslope deposits
4. Colluvium: often in hillslope hollows or former gullies
5. Landslide debris, including: boulder accumulations from rock falls or slides; earth flow debris from slumps
and translational slides, and debris flow and mudflow sediments, often channelised
III. Lower hillslope and piedmont zone deposits
6. Coarse talus and debris fan deposits, including: clast-supported talus and alluvial fans, and matrix supported
debris flow deposits
7. Colluvium, including: mudflow and sheetflood sediments; laminated sands and silts, and undifferentiated
colluvium including hillfoot ‘pedi-sediment’
8. Cemented deposits, including: ferricrete benches in colluvial and alluvial deposits; silcreted sands and
gravels, and calcrete layers and benches
IV. Valley floor deposits
9. Alluvial deposits, including: floodplain facies from meandering streams; deep channel fills; flood sediments;
shallow braided valley sediments; clay fills in channelless dambo valleys, and clay pans
10. Lacustrine deposits, often associated with alluvial plains
11. Aeolian deposits, including: linear dunes marking former extensions of desert conditions; sandplains,
see also 3, and lunette dunes associated with deflation pans
12. Cemented deposits, as in 8.
the range of surface deposits listed in Table 17.2. rocks to chemical weathering processes through
Degrees of regolith stripping vary and the land- time. A sequence of scale-dependent changes is
scapes produced have been described as varieties indicated in schematic terms in Table 17.3.
of etchplain (see Thomas, 1994). Although The results of this evolution are reflected in
brought about by fluvial dissection, outcrop pat- the detail of surface deposits. Many saprolite pro-
terns become accentuated during stripping, files deepen away from stream channels under
because of the differential response of various duricrust-capped hills and plateaux. Renewed
GFE-17.qxd 6/25/05 1:59 PM Page 464
Figure 17.12 Composite block diagram to show assemblage of typical savanna landforms and deposits.
floodplain sediments. In drier areas, past desicca-

tion has led to the formation of deflation pans and
lunette dunes, even to the formation of linear
dunes across large areas. Many such deposits are
functions of Quaternary climatic fluctuations and
they may not be in equilibrium with the balance
of present-day processes (Figures 17.14; 17.15).
Implications for foundations and

earthworks
In general, conditions for shallow foundations in
Figure 17.13 Core boulders and a sharply defined
savanna areas may be reasonably good; allowable
weathering front in granite, Jos Plateau, Nigeria.
bearing capacities tend to decrease with increas-
ing clay content. Problems that may be encoun-
tered in foundation engineering include excessive
weathering has led to the formation of troughs settlement of collapsible soils, or heave of expan-
and basins of weathering along shatter zones sive soils. There is now a considerable literature
or has been influenced by orthogonal joint sets related to these soils, and useful introductions are
and variations in mineralogy or rock fabric provided by ASCE (1982), Blight (1997) and
(Figure 17.14). The juxtaposition of outcrops and Fookes (1997). Undermining by erosion of duri-
deep regoliths is a function of all these factors. crust layers provides the possibility of local subsi-
The rapid variations in sub-surface conditions dence (methods of combating this are discussed
also make a detailed ground investigation diffi- in Chapters 12 and 24). Piling for heavy struc-
cult using conventional boreholes or geophysics. tures can be problematic, due to the variability of
The best approach is usually to undertake visual the ground both horizontally and vertically, with
inspections and block sampling in trial trenches the potential for zones of fairly strong material to
or shafts of sufficient diameter to allow man entry be underlain by more weathered and weaker
(with appropriate safety measures). ground. Expansive or collapsible soils also pro-
Sediment stores are found as plateau sands and vide problems for road construction, but duri-
gravels; hillslope debris, which may be cemented, crusts and pisolith gravels can provide good
and gully fills; piedmont fans and colluvium, and quality materials for road bases and surfacing.
GFE-17.qxd 6/25/05 1:59 PM Page 465
SAVANNA 465
Figure 17.14 Deep weathering patterns in fine-grained biotite granite near Jos, Nigeria. Sections A–B, C–D show
colluvial mantle (stipple) over saprolite (diagonal rule).
GFE-17.qxd 6/25/05 1:59 PM Page 466
Table 17.3 Schematic representation of the evolution of deeply weathered plateaux through time.
Geologic evolution Landform development
Regional cratonic surfaces (plateau Warped and faulted during late Mesozoic–Cenozoic
surfaces––Mesozoic or pre- Mesozoic in origin) plate tectonic movement
↓
Mid-late Cenozoic dissection (along continental 100 Ma of increasing relief; local planation within
margins; within interior basins) major catchments
↓
Geochemical differention in relief patterns during Duricrusted hills and ranges; elevated inselbergs;
dissection (relief sensitive to outcrop geology) weathered plains, and alluvial valleys
↓
Neogene weathering and erosion (renewed weathering Penetration of weathering beneath residual hills;
within relief compartments) weathering of new lavas and sediments;
new ferricretes
↓
Quaternary redistribution of regolith (formation of Formation of transported regolith stores on hillslopes,
present mosaic of surface deposits) benchlands and valley plains
17.4 Surface processes and the heavy rainfall is increased where soils are
geohazards in the savanna developed on sandy colluvium, and severe soil
erosion can result on slopes as low as 2o. Recent
Runoff and surface erosion work on the soil erosion history of central
Tropical savanna areas share rainfall characteris- Tanzania has shown (Eriksson et al., 1999) that
tics with more humid zones but experience these ‘modern’ soil erosion in this region is no more
processes for a shorter period, usually less than than 900 years old, but a major episode of colluvi-
6 months in the year. Rainfall frequently occurs as ation took place between 14.5 ka and 11.4 ka, a
intense convectional downpours, possibly deliver- period of rapidly changing climate in the late
ing more than 100 mm in a single storm lasting Pleistocene. However two other studies, from
2–3 hours, with intensities exceeding 100 mm/hour eastern Zambia and from central Tanzania have
over periods of 15–30 minutes. Prolonged heavy indicated a more continuous history of punctuated
rainfall lasting many days, characteristic of mon- colluviation spanning most of the last glacial cycle
soon climates worldwide, occurs less often in drier (Sørensen et al., 2001; Thomas and Murray, 2001).
areas. Because seasonal moisture stress limits the
density of the vegetation it is often argued that Slope failures
these characteristics lead to high runoff ratios and On the other hand slope failures are commonly
sediment yields. But there are problems with this restricted to shallow debris slides and flows that
argument because, under natural conditions, long occur on steeper hillslopes during heavier storms.
grass provides a protective ground cover, in the The restricted amounts of groundwater present
wetter savanna at least. If the grass cover is across much of the savanna zone appears to limit
removed by frequent fires, which are usually fol- the potential for deep-seated landsliding, which
lowed by grazing or cropping, then the impact of would otherwise be favoured by the occurrence of
the intense rainfall can be very severe. Sediment deep, clay-rich saprolites. Nonetheless, the impact
will be released from most well-drained, exposed of large intense storms can be devastating.
soils after 15–20 minutes of rainfall exceeding In February 1970, more than 100 mm of rain fell
only 25 mm/hour (Hudson, 1995). The impact of on the Mgeta area of Tanzania in less than 3 hours,
GFE-17.qxd 6/25/05 1:59 PM Page 467
SAVANNA 467
Figure 17.15 Erosion and deposition features common in savanna areas: (A) landslides and ‘lavaka’ type gullies
supply colluvium to lower slopes, here comprised of river terrace alluvium, sensitive to active gullying; (B) amor-
phous colluvium and more definite fan deposits occur widely in piedmont zones in savanna areas.
GFE-17.qxd 6/25/05 1:59 PM Page 468
leading to more than 1000, mainly small, debris

slides, which transformed into mudflows downs-
lope. Most slides were 5–20 m wide and their
average depths were only 1–1.5 m. More than 90%
of the slides occurred beneath open grassland or
cultivated fields (Temple and Rapp, 1972).
Some larger landslides found in the savanna
may be relict features from periods of more humid
climate during the Pleistocene. Well-dated land-
slide deposits from the tropics are rare. A few scat-
tered dates suggest that landslides may have been
favoured by the early Holocene pluvial conditions
(c. 10–8 ka). But, recent work in Zambia, where
rainfall is around 1100 mm/year, has revealed OSL
(Optically Stimulated Luminescence) dates
of ⬎ 200 ka for sediments associated with an
episode of widespread rotational sliding around
Chipata in the Eastern Province (Thomas, 1999;
Thomas and Murray, 2001). On the other hand,
large landslides may be triggered in monsoon and
cyclonic areas, even when total rainfalls are insuf-
ficient for rainforest. They are probably rare
events in the drier savanna, unless caused by
exceptional ground conditions.
Gully erosion Figure 17.16 Section through gully-fill sediments

Gullies commonly develop in deep colluvium, as near St Michael’s Mission, Zimbabwe. Deep vertical
observed above. At certain sites, as at St Michael’s cracks become enlarged into ‘pipes’ due to the presence
Mission in Zimbabwe, colluvial sediments are of dispersible Sodium-rich clays.
exposed in modern gullies and several episodes of
deposition and erosion have clearly taken place at
this site, old gully floors being marked by bands of
coarse sediment (Figure 17.16; Stocking, 1981).
The current phase of gully development appears to
have been triggered by cattle tracks. But there is
also an interaction between the sediments, soil
development and erosion. The colluvium contains
clays with a high exchangeable sodium percentage
(ESP), and is subject to deep cracking during the
dry season. As in many soils of the savanna zone
there is also a clay-rich or argillic horizon of low
permeability, found within 50 cm of the surface,
leading to sub-surface water flows and piping
(Figures 17.17; 17.18).
Once gully incision has taken place, the Figure 17.17 Partly collapsed ‘tunnel’ gully devel-
oping above an impermeable clay horizon in alluvial
hydraulic gradient is greatly increased at the head- sediments, Luangwa Valley, Zambia.
cut, and groundwater penetration via surrounding
cracks is rapid. If sodic clays are present, they are
GFE-17.qxd 6/25/05 1:59 PM Page 469
SAVANNA 469
Figure 17.18 Shallow gullying resulting from the

extension of the process described in Figure 17.18,
Luangwa Valley, Zambia.
highly dispersible and enter into suspension,

rapidly enlarging the soil macropores into pipes
(Figure 17.19). Flow through the pipes progres-
sively enlarges the cavities until roof collapse
becomes common, thus extending the gully sys-
tem. This destabilisation of valley fills can lead to
a rapid propagation of lateral gullies, fed partly by
groundwater flows. The extent of gullying will be
limited by the distribution of the colluvium, but
other factors operate. Away from the valley floors
Figure 17.19 Deep vertical fissure-like pipes devel-
the smectite clays may be replaced by more stable oped in the colluvium at St Michael’s Mission,
kaolinite within a more sandy sediment. In the Zimbabwe.
absence of deep cracks and pipes, groundwater
movement becomes more diffuse and less aggres-
sive, and more water is likely to enter the gully
by surface flow. As the gully system extends,
contributary slope catchments become reduced in
area and runoff lengths leading to the gully-heads
become shorter. These factors can reduce effective
water flows in individual gullies, which may then
evolve by wall collapse and stabilisation of debris
by vegetation growth (Figure 17.20).
The trigger mechanisms leading to gully inci-
sion may also be complex. Some gullies develop
from river banks, where the hydraulic gradient
is controlled by the seasonal rise and fall of the Figure 17.20 Inactive gully on the Jos Plateau,
river level (possibly ⬎ 10 m). Normally, adjacent Nigeria, now evolving by wall collapse and vegetation.
floodplain sediments will be stabilised by riparian
forest vegetation, but this can be breached by natu- in and out of water, are instrumental in causing
ral bank erosion and by animal tracks. Large game serious bank erosion. Trampling by these animals
animals such as buffalo, which travel in large also leads to the collapse of soil pipes, accelerating
herds, and hippopotamus, which move frequently the formation of open gullies. Intensive grazing of
GFE-17.qxd 6/25/05 1:59 PM Page 470
Figure 17.21 Development of deep gullies in

Quaternary colluvium on the Jos Plateau, Nigeria. The
gully is extending partly as a result of runoff via the
road culvert.
floodplain silts by domestic animals can have dev-

astating effects, as can be seen in resettlement areas
downstream of the Kariba dam in Zambia.
Road culverts and ditches designed for the dis-
persal of storm runoff from sealed road surfaces
are also liable to cause or exacerbate gully devel-
opment, and the tendency to lead road runoff into
depressions and stream heads, without adequate
protection is indefensible. Examples abound,
including sealed roads built in the 1970s on the
Jos Plateau of Nigeria (Figure 17.21). Spillways
from river impoundments, using low earth-fill
dams also have the potential to cause downstream
erosion during storms.
Although the role of colluvium in the extension
of gullying has been emphasised, the in situ sapro-
lite is also subject to slope failure and gully forma- Figure 17.22 Model for gully development on con-
tion. The lavaka gullies of Madagascar with their cave slopes, based on conditions near Bananal, São
distinctive ‘hourglass’ shape were thought by Wells Paulo, Brazil. I: Gullies connected to the drainage net-
et al. (1991) to start from bare patches on hill- work, extending by seepage erosion along discontinu-
slopes underlain by ferrallitic saprolites. In this case ities in colluvial sediments. II: Gullies not connected to
the drainage network, but forming on steepest slopes by
the development of the gullies by overland flow on runoff erosion. III: The combination of types I and II:
overgrazed slopes is the likely mechanism (Morgan, connecting upper and lower slopes. SU ⫽ summit;
1986). Similar, and even more aggressive gullying, SH ⫽ shoulder; BS ⫽ backslopes; FS ⫽ footslope; 1–4
can be seen in north-east Brazil, where convex superposes colluvial sheets (after De Oliveira, 1990).
relief compartments entirely comprised of granite
saprolite have become carved into deep canyons. connected the potential for serious land degrada-
Experience in a humid tropical area around tion is obvious. This model is equally relevant to
Bananal, eastern Brazil led De Oliveira (1990) the savanna areas of the tropics, but the occur-
to propose a model for the two major types of rence of deeply weathered relief compartments
gully formation: by seepage erosion, and by will become limited in the drier areas to hills
overland flow (Figure 17.22). Where these become protected by duricrust.
GFE-17.qxd 6/25/05 1:59 PM Page 471
SAVANNA 471
17.5 Summary Alpers, C. N. and Brimhall, G. H. (1989) Paleohydro-

logic evolution and geochemical dynamics of cumu-
lative supergene metal enrichment at La Escondida,
Savanna regions receive annual rainfalls varying
Atacama Desert, northern Chile. Economic Geology
from below 800 mm/year to as much as 1600 mm/ 84, 229–255.
year and comprise mixtures of deciduous woodland ASCE (1982) Engineering and Construction in
and grass. Their association with the plains and Tropical and Residual Soils. ASCE, New York.
plateaux of the ancient continents of Gondwanaland Blight, G. E. (ed.) (1997) Mechanics of Residual Soils.
Balkema, Rotterdam.
implies a long and complex history of development,
Bourgeon, G. (1991) See Pedro (1997) for a published
involving climate changes on different timescales. version of this thesis.
Extensive mantles of weathered rock and saprolite Büdel, J. (1982) Climatic Geomorphology (trans. L.
are legacies of past eras of humid climate, and more Fischer and D. Busche). Princeton University Press,
recent history has involved the erosion and redistri- 443pp.
Butt, C. R. M. (1987) A basis for geochemical explo-
bution of the regolith to form a variety of sediment
ration models for tropical terrains. Chemical Geology
stores in the landscape. Limited leaching in the sea- 60, 5–16.
sonal climates has retained much of the silicon, Clark, D. A. (1974) The Kankoma Clay Deposit.
iron, calcium, and magnesium liberated from rock Economic Report 49, Geological Survey, Zambia.
minerals during weathering, either as duricrusts or De Oliveira, M. A. T. (1990) Slope geometry and gully
erosion development: Bananal, São Paulo, Brazil.
as 2 :1 phyllosilicates (mainly smectite clays). The
Zeitschrift für Geomorphology N.F. 34, 423–434.
properties of these materials, when subject to the Dowling, J. W. F. (1966) The mode of occurrence of
prevailing regime of heavy tropical rainfall, and laterites in northern Nigeria and their appearance in
their occurrence within Quaternary sediments that aerial photography. Engineering Geology 1, 221–223.
accumulated under a different climatic environ- Eriksson, M. G., Olley, J. O. and Payton, R. W. (1999)
Late pleistocene colluvial deposits in central
ment, have created conditions for serious and rapid
Tanzania: erosional response to climate change?
extension of gully erosion. GFF 121, 198–201.
Savanna environments can be viewed as transi- Fairbridge, R. W. and Finkl, Jr, C. W. (1984) Tropical
tional between the humid tropics and the drylands stone lines and podzolised sand plains as palaeocli-
of the deserts and their borders. Active processes matic indicators for weathered cratons. Quaternary
Science Reviews 3, 41–72.
work on inherited materials that can be particu-
Feng, Y. X. and Vasconcelos, P. M. (2001) Quaternary
larly sensitive to destabilisation, and foundation continental weathering geochronology by laser
conditions reflect the complexity of geologic and heating 40Ar/39Ar analysis of supergene cryptome-
environmental changes over long time periods. lane. Geology 29, 635–638.
The deceptive simplicity of many of the great Fookes, P. G. (ed.) (1997) Tropical Residual Soils. The
Geological Society, London, 184pp.
savanna plains when viewed from afar, can,
Geological Society (1990) Engineering Group Working
therefore, be seriously misleading. Party Report: Tropical residual soils. Quarterly
Hudson, N. W. (1995) Soil Conservation (3rd edn).
References Batsford, London, 391pp.
Li, J. W. and Vasconcelos, P. (2002) Cenozoic conti-
Alexandre, J. and Symoens, J. J. (eds) (1989) Stone- nental weathering and its implications for palaeocli-
Lines. Journée d’Étude, Bruxelles, 24 mars 1987. mate: Evidence from 40Ar/39Ar geochronology of
Académie Royale des Sciences D’Outre-Mer, Geo- supergene K-Mn oxides in Mt Tabor, central
Eco-Trop. 11, 239pp. Queensland, Australia. Earth and Planetary Science
Aloni, J. (1975) Le sol et l’évolution morphologiques Letters 200, 223–239.
des ermitières géantes du Haut-Shaba (Rép. Zaire). Marsh, G. P. (1864) Man and Nature. Scribner, New York.
Pedologie 25, 25–39. Migoń, P. and Thomas, M. F. (2002) Grus weathering
Alpers, C. N. and Brimhall, G. H. (1988) Middle mantles––problems of interpretation. Catena 49,
Miocene climatic change in the Atacama Desert, 5–24.
Northern Chile: evidence from supergene minerali- Milne, G. M. (1935) Some suggested units of classifi-
sation at La Escondida. Geological Society of cation and mapping particularly for East African
America Bulletin 100, 1640–1656. soils. Soil Research 4, 183–198.
GFE-17.qxd 6/25/05 1:59 PM Page 472
Morgan, R. P. C. (1986) Soil Erosion (2nd edn). Tricart, J. (1965) The Landforms of the Humid Tropics,
Longman, New York, 298pp. Forests and Savannas (English trans. C. J.
Newill, D. and Dowling, J. W. F. (1968) Laterites in KiewietdeJonge, 1972). Longman, London, 306pp.
west Malaysia and northern Nigeria. In Engineering Twidale, C. R. (1991) A model of landscape evolution
Properties of Laterite Soils, Proceedings ICS MFE involving increased and increasing relief amplitude.
VII, 133–150. Zeitschrift für Geomorphologie, N.F. 35, 85–109.
Ollier, C. D. and Galloway, R. W. (1990) The laterite pro- Valeton, I., Beissner, H. and Carvalho, A. (1991) The
file, ferricrete and unconformity. Catena 17, 97–109. Tertiary Bauxite Belt on Tectonic Uplift Areas in the
Ollier, C. D. and Pain, C. F. (1997) Regolith, Soils and Serra da Mantiqueira, South-East Brazil. Contribu-
Landforms. Wiley, Chichester, 316pp. tions to Sedimentology 17, Schweizerbart, Stuttgart,
Pedro, G. (1968) Distributions des principaux types 101pp.
d’altération chimique à la surface du globe. Revue Vasconcelos, P. M. and Conroy, M. (2003) Geo-
de Géographie Physique et Géologie Dynamique X, chronology of weathering and landscape evolution,
457–470. Dugald River valley NW Queensland, Australia.
Pedro, G. (1997) Clay minerals in weathered rock Geochimica aet Cosmochimica Acta 67, 2913–2930.
materials and soils. In Paquet, H. and Clauer, N. Vasconcelos, P. M., Renne, P. R., Brimhall, G. H. and
(eds) Soils and Sediments. Springer, Berlin, 1–20. Becker, T. A. (1992) Age and duration of weathering
Sørensen, R., Murray, A. S., Kaaya, A. K. and Kilasara, by 40Ar/40Ar and 40Ar/39Ar analysis of potassium
M. (2001) Stratigraphy and formation of late manganese oxides. Science 258, 451–455.
Pleistocene colluvial apron in Morogoro district, Vasconcelos, P. M., Brimhall, G. H., Becker, T. A. and
central Tanzania. Palaeoecology of Africa 27, Renne, P. R. (1994) 40Ar/39Ar analysis of supergene
95–116. jarosite and alunite: implications to the weathering
Stephens, C. G. (1971) Laterite and silcrete in Australia: history of the western USA and west Africa.
a study of the genetic relationship of laterite and Geochimica et Cosmochimica Acta 58, 401–420.
silcrete and their companion materials, and their col- Watson, J. P. (1974) Termites in relation to soil forma-
lective significance in the weathered mantle, soils, tion, groundwater, and geochemical prospecting.
relief and drainage of the Australian continent. Soils and Fertilizers 37, 111–114.
Geoderma 5, 5–52. Wells, N. A., Andriamihaja, B. and Rakotovololona, H.
Stocking, M. A. (1981) Causes and prediction of the F. S. (1991) Patterns of development of lavaka,
advance of gullier. Proceedings of the South-East Madagascar’s unusual gullies. Earth Surface
Asian Regional Symposium on Problems of Soil Processes and Landforms 16, 189–206.
Erosion and Sedimentation 1981, Bangkok, 37–47. Williams, D., Dunkerley, D., De Dekker, P., Kershaw, P.
Tardy, Y. and Roquin, C. (1998) Dérive des Continents and Chappell, J. (1998) Quaternary Environments
Paléoclimats et Alterations Tropicales. editions (2nd edn). Arnold, London, 329pp.
BRGM, Orléans, France, 473pp.
Taylor, G. and Eggleton, R. A. (2001) Regolith Geology
and Geomorphology. Wiley, Chichester, 375pp.
Further reading
Temple, P. H. and Rapp, A. (1972) Landslides in the
Mgeta area, western Uluguru Mountains, Tanzania. Chorley, R. J., Schumm, S. A. and Sugden, D. E. (1984)
Geografiska Annaler 54A, 157–193. Geomorphology. Methuen, London, 606pp.
Thomas, M. F. (1974) Tropical Geomorphology. Douglas, I. and Spencer, T. (eds) (1985) Environmental
Macmillan, Basingstoke, pp 332. Change and Tropical Geomorphology. George
Thomas, M. F. (1989) The role of etch processes in Allen & Unwin, London, 378pp.
landform development II. Etching and the forma- Eggleton, R. A. (ed.) (2001) The Regolith Glossary.
tion of relief. Zeitschrift für Geomorphologie N.F. CRCLEME, CSIRO Mining, Canberra, ACT, 144pp.
33, 257–274. Goudie, A. S. and Pye, K. (eds) (1983) Chemical
Thomas, M. F. (1994) Geomorphology in the Tropics. Sediments and Geomorphology. Academic Press,
Wiley, Chichester, 460pp. London, 439pp.
Thomas, M. F. (1999) Evidence for high energy land- Summerfield, M. A. (1993) Global Geomorphology.
forming events on the central African plateau: Eastern Longman, Harlow, UK, 537pp.
Province, Zambia. Zeitschrift für Geomorphologie
N.F. 43, 273–297. Acknowledgements
Thomas, M. F. and Goudie, A. S. (eds) (1985) Dambos:
small channelless valleys in the tropics. Zeitschrift The author wishes to thank George Milligan for
für Geomorphologie, Supplementband 52, 222pp.
Thomas, M. F. and Murray, A. (2001) On the age and additional material, and William Jamieson for
significance of Quaternary colluvium in eastern cartographic assistance and for compiling the
Zambia. Palaeoecology of Africa 27, 117–133. figures for this chapter.
GFE-18.qxd 6/25/05 2:00 PM Page 473
18. Hot Wetlands

Ian Douglas
18.1 The tropical rainforest equatorial parts of the Pacific ‘rim of fire’, the
environment volcanic landscapes of Java and central America,
the old plateaux of the northern part of the
The high temperatures and humidity of hot Brazilian shield and the flooded forests and
wetlands support a dense, diverse forest cover freshwater swamp forests of the lower parts of the
where biological activity is continuous. The Amazon and Congo Basins and of eastern
apparent uniformity of the humid tropical rain Sumatra and southern Borneo. The tectonically
forest is deceptive. There are wide variations in active areas tend to have younger rocks that
forest type and in the engineering soils, rocks weather and break-up more rapidly than rocks of
and landforms beneath the forest (Figure 18.1). more stable areas. Given the same general rock
This diversity of humid tropical landforms must type, topography, runoff and temperature, rates of
be appreciated, as engineering experience in one chemical weathering of younger igneous rocks,
part of the hot wetlands is not necessarily trans- such as those in island arcs, are roughly twice
ferable to another. rates in old cratonic settings (Stallard, 2000). In
The hot wetlands embrace landscapes as tectonically active areas, earth movements fre-
diverse as the tectonically active mountains of quently trigger landslides and thus remove the
Figure 18.1 Block diagram of humid tropical terrain to show associations between vegetation, weathering,
sedimentary deposits and geomorphology.
473
GFE-18.qxd 6/25/05 2:00 PM Page 474
decomposing rock, supplying large quantities of the water-table is at or near the surface for most of
sediments to rivers. the year, the breakdown of clays characteristic of
The rate of chemical weathering in these hot well-drained sites does not occur, and smectite,
wetlands, and thus the preparation of rock swelling and cracking clays may accumulate. In
material for detachment and transport by erod- tectonically active areas mass movements are
ing agents, is largely determined by the combi- often frequent enough to prevent the accumula-
nation of tectonics, lithology, climate, relief and tion of more than a metre or so of soil and weath-
vegetation. The resulting weathering environ- ered material. In stable areas, deep weathering
ment can be viewed in terms of transport-limited profiles of 40 m or more occur.
and weathering-limited erosion regimes (Johnsson,
2000; see Chapter 11). Material loosened by The nature of the rainforest
weathering is carried downslope by transport The problems of engineering soils and geomor-
processes such as landslides and soil creep. If phology cannot be separated from the vegetation.
these slope erosion processes work faster than The rain forest plays a critical role in the hydro-
the weathering processes, then erosion is limited logical cycle, in pedogenesis and in landform
by the rate at which the rock is weathered. stability. Intense rainfall is a feature of the equa-
However, if the rate of weathering exceeds the torial tropics. Mean annual rainfall erosivity may
rate at which detachment and transport processes be 25 times greater than in humid temperate lati-
operate, erosion is limited by the efficiency of tudes (for example in the northeast of peninsular
the transport process and a deeper weathering Malaysia mean annual erosivity is over 25 kJ m⫺2,
profile develops (Johnsson, 2000). Vegetation while in western Britain it is just over 1.3 kJ m⫺2
cover is often the regulator of the efficiency of (Morgan, 1986)).
transport processes, preventing weathered parti- As elsewhere, rainfall varies with elevation, the
cles from being detached and trapping particles highest mean annual rainfall total recorded in the
carried downslope. When tree-fall or landslides hot wetlands of northeast Queensland, Australia,
expose bare weathered material, detachment and being 7664 mm (maximum annual total 11 521 mm
downslope transport are rapid. The occurrence in 1979) at the summit station on Mt. Bellenden
of erosion in areas of natural forest is often Ker, whereas at the eastern side of the mountain,
related to irregular disturbance of small parts of Babinda the annual mean is 4092 mm y⫺1.
the forest (see also Chapter 19). In the areas prone to tropical cyclones, high
Weathering in well-drained sites thus tends to rainfall intensities can occur for several hours
be highly efficient, reducing clay minerals to (Table 18.1). Such extreme falls have great
oxides of iron and aluminium. However, where geomorphological significance. Shorter return
Table 18.1 Examples of recorded hot wetland rainfall intensities for different storm durations.
Time period Rainfall (mm) Intensity mm h⫺1 Location Date
1 minute 33 1980 Barot, Guadeloupe 26 Nov 1970

3 minutes 63 1260 Porto Bello, Panama 29 Nov 1911
15 minutes 198 792 Plumb Point, Jamaica 12 May 1916
9 hours 1087 121 Belouve, La Reunion 28 Feb 1964
12 hours 1340 112 Belouve, La Reunion 28 Feb 1964
18.5 hours 1689 94 Belouve, La Reunion 28 Feb 1964
24 hours 1870 78 Cilaos, La Reunion 15–16 Mar 1952
GFE-18.qxd 6/25/05 2:00 PM Page 475
HOT WETLANDS 475
Table 18.2 2-year return period rainfall intensities for Natural forest disturbance and
selected tropical cities in mm/h. sediment yields
Place Continent 5 min 15 min 30 min
Overall, the forest operates as a pulse-driven
system (Table 18.3). Rainstorms can initiate geo-
Niamey Africa 160 110 79 morphological changes, such as small mass move-
Kinshasa Africa 126 93 ments, sudden stream rises and pulses of sediment
Kuala Lumpur Asia 250 148 110 moving downstream. The bigger the rain event, the
Babinda Australia 129 107 84 larger is the work done. However other pulses are
Paranaibo South America 124 88 partly driven by the biota. Not only are there ani-
Manaus South America 186 112 76
mals that create tracks and wallows where erosion
Belém South America 167 106 72
occurs, but tree falls knock down other plants and
create gaps in the canopy that allow light to reach
the forest floor. In the streams, falling trees create
period rainfall intensities (Table 18.2) indicate that debris dams that temporarily store up both plant
30 minute rainfalls of 80 to 100 mm h⫺1 are wide- material and sediment until either biotic decay or
spread in hot wetlands. In the Danum Valley area lifting of the wooden matter in a flood, releases a
of Sabah, Malaysian Borneo, 5-minute rainfall pulse of sediment to be swept downstream.
intensities rarely exceed 100 mm h⫺1 (Bidin, 2001). Disturbances of this type reach their greatest
In the period May 1995 to April 1998, 5-minute frequency on slopes affected by earth quake-
intensities of over 100 mm h⫺1 only occurred on 26 induced mass movements, in forests affected by
occasions and on only two of these days was an volcanic activity and on hillsides exposed to the
intensity of over 50 mm h⫺1 sustained for over full force of tropical cyclones making their landfall
25 minutes. These figures are within the general after a long path across the ocean. In such environ-
2-year return period ranges shown in Table 18.2. ments, the frequency and magnitude of disturbance
Depending on the type of forest, some 11 to 25 of the forest and soil may be so great that high nat-
per cent of the annual rainfall may be intercepted ural rates of erosion occur (see also Chapter 11).
by the foliage and returned to the atmosphere. Considering combinations of tectonic and
Transpiration may account for 40 to 65 per cent of lithology, it is possible to develop a schema of
rainfall. Although the high canopy breaks the fall how sediment yields will vary under natural con-
of raindrops, the coalescing of drops on leaves may ditions in the humid tropics that helps to explain
create large drops that can reach their terminal the broad inter-regional differences in sediment
velocity before hitting the ground. Localised yields in the humid tropics (Table 18.4). The high-
splash erosion of patches of bare soil in the forest est sediment yields in the humid tropics occur in
illustrates the erosivity of raindrops below the tectonically active areas, such as the mountains of
canopy. Overall, the forest retains water in the bio- New Guinea and Taiwan, experiencing sediment
mass, soils and weathering profile. Surface runoff yields of the order of 10 000 t km2 y⫺1 (Pickup
is rare, save during the most intense and prolonged et al., 1981; Shimen Reservoir Authority, 1975).
storms. Although small streams rise and fall rap- The lowest sediment yields are on old land
idly during short duration thunderstorms, the vege- surfaces or sedimentary basins of low relief.
tation reduces the height of storm runoff peaks. Areas like the Congo Basin in Africa and eastern
Most tropical forest soils are acid, low in South America (Orinoco and Negro basins) have
organic matter and available plant nutrients (i.e. sediment yields of the order of 100 t km2 y⫺1
oxisols and ultisols). Most of the nutrients that (Milliman and Meade, 1983).
support the forest are in the decomposing leaves
and other plant material on the soil surface. Once Quaternary environmental change in
this organic material is removed, new plant cover the humid tropics
is difficult to re-establish, but cover is essential if Despite claims for their ancient heritage, tropical
erosion is to be avoided. rainforests have changed greatly in extent and
GFE-18.qxd 6/25/05 2:00 PM Page 476
Table 18.3 Principal types of hot wetland (evergreen) rainforests.
Forest type Forest characteristics (after Prance, 2002) Engineering geomorphological significance
Lowland rain Multi-layered, many trees exceeding 30 m Predominant forest type in areas subject
forest in height, closed canopy with sparse to agricultural, urban and industrial
undergrowth, large lianas frequent, development; removal leads to accelerated
epiphytes relatively scarce erosion, slope instability and increased runoff
unless protective measures are undertaken.
Montane rain Few trees exceeding 30 m, palms and tree Often on steep slopes with considerable surface
forest ferns common in undergrowth, ground organic matter, due to slower decomposition
layer rich in herbs and mosses, epiphytes in cooler conditions, but with shallow soils
common; often subdivided into lower and shallow root systems. In exposed areas
and upper montane forest subject to windthrow or cyclone damage.
Variable foundation conditions.
Cloud forest High altitude closed forest with many gaps, Often associated with upland peat soils that
trees rarely over 20 m, often gnarled, take hundreds of years to develop and with
numerous lianas, tree ferns, herbs, shrubs, shallow depths to unweathered bedrock.
epiphytes and mosses. Highly susceptible to instability when
earthworks undertaken. Greatly sensitive to
expansion of temperate crop cultivation and
hill resorts and golf courses. Rapid erosion
follows cloud forest removal.
Alluvial forest Forest growing in seasonally inundated Widespread in Amazonia: in white water
areas along river banks, multi-layered, river areas known as várzea forest, in black
closed forest with numerous gaps, and clear water areas as igapó; seasonal
buttresses and stilt roots, palms, flooding disrupts mobility and restricts use
herbaceous undergrowth, epiphytes of heavy machinery; complex soil shrink-
and lianas common swell problems or surface loading issues
may arise; major impacts on aquatic life.
Swamp forest Forest on permanently wet areas, buttresses, Often traversed by blackwater streams with pH
stilt roots and pneumatophores common, as low as 3.0 which may be corrosive to
palms, ferns and herbs abundant concrete and metals; these forests often
overlie former distributary channel systems
whose coarse sands and gravels may pose
problems of stability and water retention
when canals and embankments are constructed.
Peat forest Forest over deep peat deposits in nutrient On drainage these lowland peats are likely to
poor soils. Rarely above 20 m in height, shrink and develop acid sulphate soil
ground cover mostly of ferns conditions and cracking clay, shrink-swell
phenomena. Low pH conditions may result
in corrosion of structures.
Mangrove forest Single layered forests up to 30 m in height Associated with marine clays and silts, but
growing in intertidal zones in salt water often conceal lenses of sand derived from
throughout tropics and subtropics, evergreen, former river channels. Greatly modified for
stilt roots and pneumatophores common, aquaculture, port development and tourist
little ground vegetation, few epiphytes resorts. Major installations usually require
deep piling, unless substantial layers of stiff
clay occur. Potential for disruption of natural
sediment fluxes and natural coastal protection
causing harbour siltation or coastal erosion.
GFE-18.qxd
Table 18.4 Schema of regional variations in sediment yields with tectonic situation in the humid tropics.
Tectonic setting Major channel Dominant sediment Estimated erosion General depth of Example
6/25/05
characteristics source rate (range of annual soil on slope
sediment yield)/(t km⫺2 a⫺1)
Active plate margin rift Braided channels, abundant Mass movement often up to 10 000 thin Markham River,
or half graben edge gravel triggered by earthquakes Papua New Guinea
2:00 PM
Active volcanic areas Braided channels Volcanic debris unstable up to 10 000 skeletal Toto Amarillo, Costa
of recent lava flows ash deposits Rica
Tectonically active Deep gorge sections where Valley wall failure, 7000–10 000 thin Upper Fly River,
mountain areas river usually occupies whole landsliding associated Papua New Guinea
valley floor in flood; some with seismic activity
Page 477
braiding, but lateral movement
of river restricted
Late Tertiary tectonic Wide channel with gravel; Bank erosion in main channel ca. 1000 ca. 1 m Segama River
activity and weak bare, frequent undercut and tributaries, erosion Sabah, Borneo
mudrocks banks and grassed flood by saturated overland flow
deposits in streamhead hollows
Passive margin with In uplands, boulder strewn Continuation of boulders, 50–100 up to 30 m Gombak, Malaysia
relief of 2000 m in channels, with mature trees surface wash on slopes,
equatorial climate right up to water’s edge; bank erosion
abundant quartz sand
between boulders
Passive margin with Upland boulders strewn or Disintegration of boulders, 50–200 up to 30 m Babinda and Behana
relief of 2000 m in rock-cut channels, with surface wash on slopes, Creeks, North
tropical cyclone zone areas of exposed rock or bank erosion Queensland
boulders at low flows;
channel capacities much
greater than previous case
Ancient craton with Upland channels guided by Little sediment supply except up to 50 up to 30 m Mahaweli Ganya
relief of 2000 m ancient structural lineaments, through river action on System, Sri Lanka
resistant angular blocks rock of channel wall
and some derived gravel;
HOT WETLANDS 477

monsoonal climates produce
large broad lowland valleys
Ancient craton with Wide sandy channels, little Wash from etchplain surface up to 30 up to 30 m Zaire, Africa
erosion surface incision; occasional rock bars
which suffer little erosion
Sedimentary basin Wide anastomising channels, Little locally derived up to 30 up to 30 m Amazon basin
on ancient craton often with legacies of past sediment, except from
phases of fluvial erosion bank erosion
GFE-18.qxd 6/25/05 2:00 PM Page 478
Figure 18.2 Typical boreholes on weathered granite in hot, wet lowland environments.
altitudinal range during the Quaternary as sea

levels and climatic conditions changed (Stanley
et al., 2002). Much of the area now covered by
tropical rain forest became more arid during
cooler periods with lower sea levels, with a mod-
erate extension towards the equator of the
savanna, semi-arid and arid areas (Haberle, 1997).
In many places soil formations and river channels
characteristic of these more seasonal, less wet,
climates are found. However, new palynological
evidence after 1990 has suggested that the
hypothesis of the Amazon rainforest being
reduced to several discrete ‘refuges’ may be no
longer tenable. (Colinvaux et al., 2000, 2001)
suggest that the lowland Amazon forests persisted
through glacial cycles, with some reassortment of
species as temperature fluctuated from the Late
Glacial Minimum to the present.
Many formations, from the Older Alluvium of
Malaysia and Singapore (Gupta and Pitts, 1992)
and the widespread ferricretes (the iron-oxide-
rich duricrust often termed laterite, Chapter 2)
under some rainforests in SE Asia and Africa are
more likely to have been formed under more
seasonally wet climates than those prevailing at
present. The laterite in the granite residual soil
borehole (Figure 18.2b) is probably a relic from Figure 18.3 Typical borehole in a humid tropical
such seasonally dry quaternary conditions. The coastal plain environment with evidence of former river
channels (sand and gravel beds).
Older Alluvium has depositional features typical
of large braided stream channels that probably
flowed across the exposed floor of the Sunda Sea
to a sea level well below the present. Not surpris- stream channels (Figure 18.3). In such former
ingly, therefore, boreholes through the marine channels, groundwater movement is relatively
clays of coastal plains often encounter beds of rapid, and clay tends to be washed, i.e. piped,
sand with some gravel that represent former between particles.
GFE-18.qxd 6/25/05 2:00 PM Page 479
HOT WETLANDS 479
Figure 18.4 Classification of tropical weathering grades (from Fookes, 1997a).
18.2 Weathering and soil undergone a decompositional change to a soil

characteristics type material (residual soil). Weathering is initiated
at the surface and from discontinuities in the rock
The deep weathering profiles of the humid tropics that allow the ingress of water to greater depths.
are probably the most well-known aspect of The weathering profiles in Figures 18.4 and 18.5
the engineering geomorphology of hot wetlands are schematic representations of a classification
(e.g. Lovegrove and Fookes, 1972; Fookes, 1997b). for the various stages of decomposition (Tables
However, while deep weathering is commonly 18.5 and 18.6). The result is a weakened ‘skin’ of
thought to be the norm and does prevail in much decomposed material which can be tens of metres
of the area covered by rainforest, it is variable, thick. In Hong Kong, for example, the rocks have
even sometimes absent. Parent material, rainfall, been deeply weathered, giving rise to depths of up
drainage and the age of the landsurface are to 60 m of silty-sand residual soil (Grades IV–VI)
the prime factors determining the character of with large corestones in the matrix or exposed on
weathering profiles. the surface.
The combination of high temperature and rain- The lower sections of many mountain slopes are
fall produces the conditions for intense tropical likely to be mantled by varying thicknesses of col-
weathering causing decomposition of the rock luvium (i.e. transported material ‘. . . heterogeneous,
forming minerals and a general decrease in rock generally structureless materials of soil/rock mate-
intact strength and interparticle bonding. This rial and sometimes organic matter, deposited on
weathered material has not been transported––it is and at the base of natural slopes by predominantly
in its original position (in situ) and has only mass-wasting’ Evans et al., 1997). Various forms of
GFE-18.qxd 6/25/05 2:01 PM Page 480
Figure 18.5 Examples of hillslope profiles illustrating common sequences of weathering and landforms (from
Fookes, 1997a).
colluvium are to be expected, relating to different and phyllite, for example, the depth of weathering
generic processes, including the products of soil may be as little as 0.1 m on the quartzite and as
creep, rock falls, and landslide activity. much as 15 m on the phyllite. Quartzite is resistant
Parent material character often produces abrupt to weathering but other common rock minerals,
changes in the depth of weathering. In a metamor- particularly feldspars, decompose rapidly in the
phosed sedimentary sequence containing quartzite high humidity, high temperatures and high organic
GFE-18.qxd 6/25/05 2:01 PM Page 481
HOT WETLANDS 481
Table 18.5 Scale of weathering grades of rock mass (from Fookes, 1997a).
Term Description Grade
Fresh No visible sign of rock material weathering; perhaps a slight discolouration on I

major discontinuity surfaces.
Slightly weathered Discolouration indicates weathering of rock material and discontinuity surfaces. II
All the rock material may be discoloured by weathering.
Moderately weathered Less than half of the rock material is decomposed or disintegrated to a soil. III
Fresh or discoloured rock is present either as a continuous framework or
as corestones.
Highly weathered More than half of the rock material is decomposed or disintegrated to a soil. IV
Fresh or discoloured rock is present either as a discontinuous framework or
as corestones.
Completely weathered All rock material is decomposed and or disintegrated to soil. The original mass V
structure is still largely intact.
Residual soil All rock material is converted to soil. The mass structure and material fabric VI
are destroyed. There is a large change in volume, but the soil has not been
significantly transported.
activity of rainforest areas. Where parent rocks movements and relatively high soil loss, prevent-
contain silicate minerals with a high proportion of ing a deep weathering profile from developing.
more soluble elements, such as calcium, magne- Topography and drainage influence weathering-
sium and potassium, then decomposition is most profile characteristics. On well-drained, upper
effective. However, even silica eventually becomes and middle sections of slopes on suitable rocks,
mobile, leaving behind oxides of iron and alu- kaolinite will be the dominant clay mineral, but
minium that eventually appear as nodules or con- downslope it will tend to be replaced by illite or
cretions in the weathering profile, and, finally, as montmorillonite (Table 18.7). Where drainage is
massive hardpan layers, or as duricrusts, termed impeded, montmorillonite becomes the dominant
ferricrete and bauxite respectively. The types of clay mineral. Thus quite different clay conditions
clay found in the weathering profile also reflect will be encountered in flat valley floors and on
the parent material. Although the trend in hot wet- adjacent spurs and hillslopes. Materials mobilised
lands is to the formation of kaolinitic clays (Table by weathering on the well-drained sectors of
18.7), such clays are by no means found every- slopes may accumulate in the valley floors to such
where in such regions. While soils developing on an extent that secondary ferricretes can be found
granites and pegmatites tend to be rich in kaolin- where the iron mobilised upslope is rede-
ite, those on slate and shales, perhaps because of posited in the fluctuating water-table zone of the
the character of the sediments from which these floodplain.
rocks were formed, are usually dominated by illite. The age of the landsurface is of major signifi-
Rainfall influences weathering through both cance for the depth of weathering. Many parts of
quantity and seasonality. In general, the propor- the hot wetlands lie in the zone of present day
tion of kaolinite in the clay increases with greater tectonic mountain building, the tectogene tropics
humidity and greater rainfall. In the wettest areas (Table 18.4). Other parts belong to the stable,
with high rainfall throughout the year, kaolinite is older cratonic landsurfaces of the cratogene trop-
broken down and gibbsite and haematite become ics. On the latter surfaces, weathering profiles
more prominent. Weathering profiles tend to be may be old and carry features inherited from past
deeper in the wetter areas, but as the highest climates. In the tectogene tropics, frequent tec-
rainfalls tend to occur on steep mountainsides, tonic activity may trigger earthquakes that induce
excessive water there may cause episodic mass so many mass movements that deep weathering
GFE-18.qxd
Table 18.6 Summary of Duchaufour tropical soil phases, location and climate (from Fookes, 1997b).
482
Factor/conditions: Mineralogy Climate needed to Typical locations of FAO/UNESCO equivalents
6/25/05
Soil phase reach the phase the phase (USA — Soil survey)
1. Fersiallitic Upper soils undergo decalcification Mean annual temperature (⬚C) Mediterranean, subtropical Cambisols, calcisols,
and weathering of primary minerals. 13–20 luvisols, alisols, andosols
Quartz, alkali feldspars and muscovite (alfisols, inceptisols)
2:01 PM
not affected. Free iron usually ⬎60% Annual rainfall (m)
of total iron. Main clay mineral formed 0.5–1.0
is 2:1 smectite; 1:1 kaolinite may
appear in older well drained surfaces.
Page 482
With recent volcanic ashes porous Dry season — Yes
andosol soils formed which are
eventually replaced by 1:1 halloysites.
2. Ferruginous More strongly weathered soils form Mean annual temperature (⬚C) Subtropical Luvisols, nitosols, alisols,
(ferrisols- but orthoclase and muscovite typically 20–25 acrisols, lixisols, plintha
transitional) remain unaltered. sols (alfisols,
ultisols, oxisols)
Kaolinite is the dominant clay mineral; Annual rainfall (m)
2:1 minerals are subordinate and 1.0–1.5
gibbsite usually absent.
On older land surfaces and more Dry season — Sometimes
permeable and base rich parent material,
ferrisols transitional to phase 3. Partial
alteration to gibbsite may occur.
3. Ferrallitic All primary minerals except quartz are Mean annual temperature (⬚C) Tropical
weathered by hydrolysis and much of ⬎ 25
the silica and bases removed by solution. Can occur in modern savanna
Remaining silica combines to form Annual rainfall (m) from previous wetter climate.
kaolinite but with excess aluminium ⬎ 1.5 Conversely, some currently
gibbsite is usually formed. Depending hot wet areas are still only in
on the balance between iron and Dry season — No the ferruginous phase (e.g. by
aluminium, iron oxide or aluminium climate change or by
oxide will predominate. Soils currently rejuvenation of slopes).
take 104 or more years to form.
GFE-18.qxd
Table 18.7 Clay-sized weathering products and humid tropical geomorphology.
Weathering product (in order of Character and composition Presence in hot wetlands Location and geomorphic significance
6/25/05
stage of weathering)
1. Gypsum, halite, etc. Gypsum, CaSO4 2H2O: evaporite Rarely found, being characteristic
from brine and salt solutions. of arid regions.
2. Calcite, dolomite, aragonite Calcite, CaCO3; dolomite, CaCO3. Widespread only in karst Primarily as a secondary deposit in
2:01 PM
MgCO3; aragonite, CaCO3 environments — readily weathered karst cavities or on tufa in rivers
(less stable form than calcite). from igneous rocks. draining karst areas.
3. Olivine––hornblende Olivine: magnesium silicate; Minerals readily weathered from Found on fresh rock exposures only.
hornblende: calcium silicate. igneous rocks with loss of magnesium, Rapidly decomposed.
Page 483
calcium and silica in solution.
4. Biotite, chlorite Ferromagnesian mica, KFe3 (AISi3 Minerals readily weathered from Found on fresh rock exposures only,
O10) Mg3 (OH, F)2; chlorite is a clay magmatic rocks due to solubility biotite granites likely to weather
mineral formed from weathering of of potassium and magnesium. readily.
hornblende, augite and hypersthene
5. Albite Sodium silicate, Na (AlSi3 O8); Readily weathered in decomposition
primary mineral in acid igneous of feldspars in igneous rocks.
and metamorphic rocks.
6. Quartz Silica, SiO2: quartz is a major rock Less readily weathered than preceding Primary mineral which is more resistant
constituent and quartzite veins minerals. than other common minerals in
often occur in metamorphosed weathering profile. May comprise a
rock. high proportion of the sand fraction in
many tropical deep weathering
profiles. Often the major component
of sand size bed material in tropical
rivers. Quartzite forms upstanding
bare ridges or hills in hot wetlands.
7. Illite, muscovite, sericite Clay minerals: secondary micas. Produced by breakdown of primary
Muscovite, K2O. 3Al2O3. 6SiO2. micas in igneous rocks and tend to
HOT WETLANDS 483

2H2O; Structure is platy with be weathered themselves by loss of
potassium ions sandwiched potassium.
between alumino-silicate layers.
8. Hydrous mica intermediates, Similar layer lattice structured clay Usually found as an intermediate
vermiculites minerals, often with similar stage in the breakdown of primary
composition to illite and muscovite micas in igneous rocks but can
but with less potassium and more be found as minerals of neoformation
hydroxyl (H3O⫹). in swamp environments.
GFE-18.qxd 6/25/05 2:01 PM Page 484
profiles do not develop or persist on steep slopes. year. This equates approximately to an average
Irregular regolith depth and instability of the density of one landslide per year for every
weathering mantle is likely to be the norm in square kilometre of the developed area.
these tectogene areas. In Papua New Guinea, for Almost all of the landslides are triggered by
example, most weathering profiles are shallow, heavy rainfall. Rainfall events that trigger
immature and kaolinitic (Löffler, 1977). However, landsliding at medium densities (1–10/km2)
rates of chemical denudation may be higher in the occur every 2 years at the local scale (i.e. any
tectogene tropics than in the cratogene tropics, given site), and 5-times/year for the region (i.e.
because of their association with high rates Hong Kong as a whole), as indicated in Table
of physical erosion (Riebe et al., 2001). Deep 18.8.
weathering profiles may thus indicate landscape Most of the landslides occur on man-made
stability, age and duration of weathering much slopes (i.e. cut slopes and retaining walls) and
more than rate of weathering. are relatively small (⬍50 m3), although some
On the cratogene land surfaces, legacies from can be 5000 m3 or more. Until recently attention
drier, more seasonal Quaternary climates can be has focused on the stability of man-made slopes.
found. While some basalts, shales and similar However, landslides on natural terrain also pres-
rocks have iron accumulations in the form of ent a significant hazard, as highlighted by:
gravels in the upper part of the present-day weath-
ering profile (Figure 18.6), such accumulations – the 1.2 km long debris flow on the eastern
may well be broken fragments from a ferricrete slopes of Tsing Shan, above Tuen Mun
formed under an earlier, seasonally dry climate. New Town, that occurred in 1990 and
Duricrusts develop in some hot wetland environ- reached a planned development site (King,
ments at the present time, such as bauxites on well- 1996);
drained basic rocks in Sarawak, or ferruginous – numerous debris flows on the northern
concretions found in the weathering profiles on slopes of Lantau Island, near Tung Chung
basalts and andesites (Figure 18.6). It is wrong in 1992/1993 (Franks, 1996), which would
to think that a particular type of duricrust is an indi- have presented a threat to the North Lantau
cator of a specific palaeoclimate. Topography, Expressway had it been opened at the time.
drainage, parent material and time all play a role in
weathering and duricrust formation. Concerns for the safety of new development
The combination of tropical climate, deeply sites led the Geotechnical Engineering Office
weathered rocks and steep slopes generate a num- to undertake a programme of region-wide
ber of active surface processes that are of import- landslide hazard mapping – the Natural Terrain
ance to engineering in hot wetlands (e.g. Fookes Landslide Study (Evans and King, 1998; Evans
1997a; Figure 18.7). These are: et al., 1997; King, 1997, 1999). As a first step,
an inventory of landslide features and areas of
1. Widespread landslide activity. In Hong Kong, intense gullying was compiled from the inter-
for example, the combination of widespread pretation of high-level aerial photographs
landslide activity, dense development of a hilly (1 : 20 000 to 1 : 40 000 scales) taken in 1945,
terrain, highly variable seasonal rainfall 1964 and annually from 1972–1994 (excluding
regime and a large number of potentially sub- 1977). Most parts of Hong Kong appeared on
standard man-made slopes (mostly formed between 20–23 sets of photographs.
before the 1970s without proper geotechnical A total of 26 870 natural terrain landslides
input and control) have created acute slope were identified within an area of 640 km2. Of
safety problems. Landslides have caused over this total, 8,804 landslides were described as
470 deaths since 1940. having occurred within the last 50 years (i.e.
On average, some two to three hundred ‘recent’), with evidence for 17 976 slides over
slope failures are reported in Hong Kong every 50 years old. Most failures were debris slides
GFE-18.qxd 6/25/05 2:01 PM Page 485
HOT WETLANDS 485
Figure 18.6 Examples of weathering profiles and topographic relationships in humid tropical terrain.
or debris avalanches within weathered rock (e.g. gullying) and landslide activity. Note that
(i.e. residual soil) or overlying colluvium. large run-out flow slides can develop where
Mapped landslide frequencies on different relatively saturated metastable soils fail under
geological units range from 0.3/km2 per year undrained conditions (e.g. during excavation of
to 2.8/km2 per year, with an overall average of cuttings).
0.8/km2 per year for all landslides. The aver-
age density of recent slides is 0.275/km2 per
year. A clear association between landslide 18.3 Soil characteristics on
density, geological unit and slope angle was different rock types
identified and provided the basis for a
regional hazard assessment. Granitic rocks contain quartz, feldspars and micas
2. Areas of severe soil erosion hazard, especially which weather to produce a soil with a bimodal
after vegetation has been removed prior to con- grain size distribution, a clay fraction from the
struction. Soil erosion can lead to major land decomposition of the feldspars and micas, and a
degradation and off-site sedimentation prob- coarse sand fraction of relatively unweathered
lems (e.g. decline in reservoir water quality, quartz grains. Such weathering profiles are likely
increased flood risk because of siltation within to contain boulders, or corestones, of unwea-
river channels etc.) and have significant visual thered rock one or more metres in diameter. Such
impact (Chapter 11). corestones are often exposed as outcrops on
3. The potential for metastable collapse of residual hillsides (Figure 18.6) whence they may eventu-
soils, leading to a combination of erosion ally roll on to the valley floor. Such corestones
GFE-18.qxd
486
6/25/05
2:01 PM
Page 486
Figure 18.7 Wetland weathering and landscape model (from Fookes, 1997b).
GFE-18.qxd 6/25/05 2:01 PM Page 487
HOT WETLANDS 487
Table 18.8 Rainfall thresholds for landslide activity in Hong Kong (Evans, 1997; Evans et al., 1997)
Threshold I––start of Threshold II––start of Threshold III––start of

landsliding landsliding at medium densities landsliding at high densities
(1–10/km2) (⬎10/km2)
Rainfall Return period Rainfall Return period Rainfall Return period
Local 60–70 mm 2.5 times/year 180–220 mm Every 2 years 380–450 mm Every 20 years
Regional 25 times per 5 times per Every 2 years
year year
can cause problems in excavations and in the sta- that accumulates at the base of dolines or cockpits
bilisation of road cuts. Boreholes can hit core- in the karst surface (see Chapter 24). Clays are
stones, giving the impression that solid rock has washed through surface openings and accumulate
been reached. in the cavern systems inside karst hills (see
Similar corestones are likely to be found in below). In high rainfall areas a steep, sharp
sandstone and conglomerate weathering profiles, edge pinnacle or needle karst develops on pure
but in mudstones, shales and schists, the tran- limestones, producing an almost impenetrable
sition between weathered and unweathered mate- knife-edge terrain (see also Chapter 24).
rial is gradual. Weathered shale materials retain
the lineaments and hues of the parent rock but are
friable and loose to touch. Where such stratified 18.4 Rivers and alluvial landforms
rocks contain vein quartz or quartz grains in a
coarse sand horizon, quartz fragments can form Most water falling on the rainforest floor infil-
a characteristic stone-line in the soil. Some shales trates, then is either transpired by the plants or
have ferruginous gravel in the upper part of the finds its way by subsurface routes to the water-
weathering profile. table or to stream channels (see Chapter 11).
Basalt weathers into deep red-brown, uniform Streams emerge as seepages or small springs in
soil, but under some circumstances rounded stream-head hollows, which may be choked with
basalt boulders may form. Although usually boulders, depending on the local terrain and parent
smaller than the granitic corestones, such boul- material. Under normal conditions, a concentrated
ders can be inconvenient when excavating. Being stream will begin some distance down the channel
rich in basic minerals, basalt regoliths tend from the stream head, but after prolonged rain,
to develop bauxitic concretions in the wettest, with a return period of once a year or less fre-
well-drained areas, iron-rich concretions in quently, concentrated surface runoff may occur
slightly less wet and more seasonal conditions along every declivity where slopes converge above
and montmorillonitic clays generally. Some the hollow, producing a cascade of water into the
basaltic soils have marked shrink-swell character- stream head hollow. Along the banks of such small
istics. Depth of weathering and regolith develop- headwater streams, several natural pipe outlets
ment on such volcanic rocks vary with the age of may be active after a major storm, indicating that
the lava flows, as well as with other environmental quick subsurface return flow operates.
characteristics. The heads of streams tend to be the most sensi-
Unlike most other rocks in tropics, save for tive part of the landscape, being likely to erode rap-
quartz and quartzite, limestone appears in bare idly if surface runoff above it is accelerated by
vertical cliffs protruding from rainforests. The human activity. Gullies frequently develop above
solution process on calcareous rocks is so effective stream head hollows following removal or severe
that the only soil material is the insoluble residue disturbance of the forest, becoming the prime
GFE-18.qxd 6/25/05 2:01 PM Page 488
pathway into the stream for the eroded material as lift the jammed logs, setting them and all the sed-
land is developed in these hot, humid regions. iment behind them in motion, sometimes provok-
Disturbance can also lead to the enlargement of nating bank scour and often leading to the floating
ural pipes and their collapse to form gullies, espe- timber building up against bridges or other struc-
cially in the colluvial fill of stream-head hollows. tures further downstream. Careless tree removal
Often, in extremely wet conditions, the walls and grading operations close to rivers can lead to
of stream-head hollows are so saturated that frag- extremely serious downstream consequences.
ments of weathered and colluvial material fall to Usually the concentrations of dissolved solids in
the floor of the hollow and are carried down- tropical rainforest streams are low. There is plenty of
stream. Along small channels, rivers erode their rain to dilute ground and soil water. However, the
banks, undercutting trees and scouring around waters tend to corrode both natural rocks and indus-
logs that have fallen in the river. Some of the logs trial materials in rivers. Appropriate concrete is
trap other branches and plant fragments, develop- needed for bridge abutments and similar structures
ing debris dams which may eventually form a if corrosion is not to become a serious problem in a
small step in the channel long profile with sand few years. Boulders of crystalline rocks develop
and gravel accumulating upstream of the block- irregular surfaces as the more soluble minerals are
age. Logging and land clearance operations can broken down and the resistant minerals protrude
increase the number of these temporary sediment from the remaining boulder (Figure 18.8). In many
stores. A high flow following a major storm can small streams emerging from hilly terrain, boulders
Figure 18.8 Boulders in a tributary of the Sungai Langat, Selangor, Malaysia. The stream bed here is largely
quartz sand, the suspended sediment load mainly kaolinitc clay, reflecting the bimodal grain size of the weathered
granite regolith in the area.
GFE-18.qxd 6/25/05 2:01 PM Page 489
HOT WETLANDS 489
do not persist far below the last rock outcrop and the means of establishing the positions of former
sand in the stream bed gradually decreases in grain channels and evidence of past flood heights (see
size. Larger rivers, however, can carry boulders, Chapter 10).
sand and gravel considerable distances, as demon- Removal of vegetation and shifting of earth alter
strated by the in-stream gravel abstraction close to rainfall: runoff: sediment yield relationships.
many lowland towns in Asian rainforest regions. Erosion plot experiments suggest that, depending
Although movement of bed material may contribute on parent material and slope, soil loss under forest
a relatively small part of the total sediment transport may be of the order of 5–100 t km⫺2 y⫺1, but
by equatorial rivers, it is a powerful agent that needs around 2400–4000 t km⫺2 y⫺1 from bare soil. This
to be considered whenever structures are built along 100-fold increase in the erosion rate may exagger-
or across rivers. ate what happens over larger areas, but nonetheless
Large flood events such as those of a 1 in 50 indicates the potential severity of the sediment
recurrence interval (or less frequent) can lead to problems created by ground clearance. Road build-
massive avulsion, river terrace dissection and ing and the use of unsealed roads with inadequate
channel change. Such events can disrupt commu- drainage in rural and forest areas are major sources
nications, especially bridges and their approaches. of sediment, particularly when steep road-cuts and
Often the available hydrological data have gaps embankments along roads fail (Figure 18.9).
caused by recorder damage during floods, while Stabilisation of cuttings and embankments by turf
many remote areas only have rainfall information or spraying grass seed and the provision of ade-
for villages on valley floors. Detailed hydro- quate cross drains and safe disposal of roadside
geomorphological surveys are often the only runoff should be part of the construction brief of all
Figure 18.9 Stabilising a fill slope in the Kinabalu Highlands.

GFE-18.qxd 6/25/05 2:01 PM Page 490
roads, even those for temporary use in forestry and surfaces are raised, but also of the probability of
similar operations. shift in channel position.
High sediment yields often last for a relatively In the deltaic areas, the problems become even
short period during road building and clearance. more complex as changed sediment and solute
The disturbance usually shifts, in activities like loads coming from the upper catchment interact
forestry and plantation development or the build- with the brackish water from the sea, as modified
ing of new residential areas to another part of the by canal building, flood defences and irrigation
catchment. Around a large city, the outward works (Chapter 20). Flooding can last several
movement of bare ground, cleared for new con- months in major river deltas such as the Mekong.
struction, can be readily traced by remote sensing River channel change occurs through erosion and
(Figure 18.10). Eventually a whole catchment accretion. Shifts of river banks in the Mekong
could be come less prone to erosion as the ground delta of up to 1000 m laterally have occurred over
is covered by forest regrowth, or new forms of reaches of up to 6 km in length in the period
vegetation, or by paved areas. However, extreme 1966–1999 (Viet et al., 2000). Sand and gravel
events can always trigger the collapse of some islands in the delta rivers change position, altering
apparently stabilised slopes. Failure of aban- conditions for navigation. These naturally unsta-
doned forest roads, partly through the decay of ble areas show constant adjustment to present day
hollow log culverts and the blocking of subsur- conditions, but also overlie complex sedimentary
face drainage by cut and fill, following rainstorms sequences of former channels and backswamps
of 1 in 5 year, or less frequent, recurrence interval, sloping down to former shorelines, associated
can set in train almost as much sediment move- with lower sea levels during the Quaternary. In the
ment into channels as the original road building dry season salt water penetration becomes a prob-
and logging operations (Douglas et al., 1999). lem and changes in land use and reservoir capac-
These changes in land cover and the irregular ity upstream may alter low flows in major rivers
pulsed nature of major sediment inputs to rivers such that salinity begins to affect new areas fur-
complicate the estimation of sediment yields to ther inland.
reservoirs and peak flows in rivers. In places, low- When rivers reach the coast, their dry season
land channel siltation affects irrigation channels flow may be insufficient to prevent the virtual
in ricefields such that there is no longer sufficient closure of the mouths by beach ridges developed
head for the water to reach all the planted area. by longshore drift. Many humid tropical shore-
Water intakes for urban supplies have become lines have a series of parallel sand beach ridges
silted. Engineering design and maintenance has to with intervening peat filled backswamps. The
cope with the consequences of changing land latter play important flood storage roles but are
cover and land use. often reclaimed as ports, industries and coastal
settlements expand.
Rivers in alluvial lowlands
In the lowlands, rivers often meander dramatic-
ally, or develop anastomosing channels, through 18.6 Volcanic terrain
freshwater swamps and brackish water mangrove
systems (Chapter 10). The várzea and igapó The varied character of the volcanoes dominating
forests of the Amazon probably represent the many humid tropical landscapes produces a variety
most complex such environment with changes of of landforms. Active volcanoes create new land-
several metres in water levels between wet and forms, burying older landscapes, often blocking
dry seasons. Traditional farmers have many ways watercourses, and creating temporary lakes which
of adapting to the seasonal flooding, but modern can overflow and cause major floods. Many volca-
settlements and transportation systems often noes have irregular periods of activity between
require careful estimation not only of the foun- which people occupy the fertile soils that
dation problems and ground settling when land develop on their flanks. The density of agricultural
GFE-18.qxd 6/25/05 2:01 PM Page 491
HOT WETLANDS 491
Figure 18.10 Map to show the outward shift of bare ground (mainly on construction sites) since the 1960s in the
Kelang Valley of Malaysia.
GFE-18.qxd 6/25/05 2:01 PM Page 492
settlements often places people at risk from karst pinnacles, and the sheer size of many under-
future lava flows or nuées ardentes, as on the ground caverns (Figure 18.11). High rainfall
flanks of Mt. Merapi in Java, Indonesia (see areas may have large underground river systems,
Chapter 23). The burning of the vegetation that often discharging through major cave openings or
follows showers of nuées ardentes exposes the springs. The density of enclosed depressions in
soil to erosion. pyramid and doline and cockpit karst may be as
Deep gulleys on the volcano flanks provide high as 12 per km2, making the centripetally
pathways for the lahars that develop when water is focusing polygonal karst drainage net probably
caught up with molten lava. The lahars move rap- the most efficient natural drainage system
idly down valley with the power to obliterate (Williams, 1978). Solution rates often exceed
everything in their path. Large blocks several 300 t km⫺2 y⫺1 (Laverty, 1980), and are associated
metres in diameter are moved by these flows. with decay of the limestone surface, particularly
Concrete dams to slow lahar movement have been at the base of dolines. Thus partial collapse of
built in many channels on the blanks of Mt. cavern roofs often change landforms leaving piles
Merapi, but Engineering works further downval- of boulders choking many parts of tower karst
ley need to take account of the lahar risk. caves. Residual soils frequently accumulate on the
Bandung, Java provides a good example of a floor of enclosed depressions, supporting a denser,
humid tropical city with complex volcano-related taller forest vegetation than the adjacent limestone
geomorphological problems. The city municipal- slopes.
ity and district cover some 2200 km2, with nearly Around the karst hills, karst plains have often
6 million people living at a density of over 17 500 developed, probably by lateral corrosion by rivers.
per km2, overlying important aquifers of volcanic Sometimes outcrops of limestone occur at the sur-
deposits or fluvial sediments (Dam, 1994). The face of the ground. In other places, the limestone
city essentially is built on a fan of volcanic mater- surface has been buried by alluvium and is only
ial from Gunong Tangkubanparahu to the north exposed at the base of excavations such as tin
that becomes inter-fingered with alluvial material mines or gravel pits. This buried karst is a major
and is eventually overlain with lacustrine deposits engineering problem in many parts of the tropics,
that formed in the Pleistocene Bandung Lake. particularly where pinnacle karst surfaces have
This sequence of volcanic and alluvial deposits been buried or where large cavern systems lie
gives rise to several good aquifers yielding high underground. Boreholes sunk into the buried
quality water. However, that water supply is being karst may penetrate a succession of voids in the
threatened by pollutants from the expanding urban limestone (Figures 18.1 and 18.12). Some of these
area, siltation from erosion and the runoff of voids have a clay infill, others contain so little
excess agricultural chemicals from surrounding material that borehole loggers record that the drill
farmland. bit ‘fell under its own weight’. These irregular
For further discussion of the engineering pro- cavities may only affect part of a site and in the
perties of volcanic soils and their role in earth- past have been missed where exploratory bore-
works and foundations, see Chapter 23. holes have been widely spaced. Details of the sit-
ing of several major buildings in cities like Kuala
Lumpur have had to be changed because of prob-
18.7 Karst terrain lems associated with buried karst. Grouting tech-
niques to fill such cavities are not always
Limestone landscapes in the humid tropics take a successful, as material may be lost into deeper
variety of forms (Table 18.9), all reflecting com- cave systems. As the sub-surface water circulation
binations of lithology, structure and climate (see is not well understood, it is possible that solution
Chapter 24). The most striking general features of and cavern formation are continuing and that sub-
karst in these hot, wet environments are the great surface cavern collapse may lead to subsidence of
vertical development of the residual hills and surface structures.
GFE-18.qxd
6/25/05
Table 18.9 Karst landscapes of the hot wetlands.
2:01 PM
Type Description Problems Typical localities
1. Polygonal General term for any system of closed Complex drainage likely to be cut by New Guinea.
depression tropical karst where divides excavations. Fill material may be
between depressions form a cellular washed into underground cavern
Page 493
network in plan. Subdivided into five systems.
major categories.
a. Pitted undulating Surfaces incised by close-set depressions Rapid runoff in gullies after rain.
surfaces with gully-like channels leading to
roughly central stream sinks.
b. Pyramid and doline Bowl-shaped depressions and Steep-hillsides, irregular Lake Kutubu area, New Guinea.
concave pyramidal hills. depressions, chaotic relief.
c. Cockpit (cone or kegal) Conical or hemispheroidal residual hills Water-logging of depressions in wet Gunung Sewu, Java, Indonesia;
with closed depressions between them. season. Minimal surface drainage. Puerto Rico; Jamaica; Kikori,
Red earths or rendzinas in depression Underground water connections New Guinea.
floors. Forest covering craggy slopes. usually unknown.
d. Tower Steep-sided, vertical or overhanging Massive caves in hills, complex Porema, New Guinea
residual hills, surrounding deep enclosed subsurface water circulation, collapse (Löffler, 1977).
depressions or rising from karst margin debris and secondary formations often
plains, often covered with alluvium, filling caves.
cliff-foot caves or notches.
e. Arête and pinnacle Spectacular, knife-edge ridge country Extremely difficult country. Limestone Gunung Mulu, Sarawak;
consisting of nearly vertical limestone edges so sharp they can cut boots! Mt Kaijende, New Guinea.
ridges with sharp crests surrounding Vegetation usually obscures true height
HOT WETLANDS 493

enclosed depressions. of pinnacles.
GFE-18.qxd
494
6/25/05
Table 18.9 (continued).
2:01 PM
Type Description Problems Typical localities
2. Karst margin plains Flat to gently undulating plains Massive caves in residual hills, pinnacles Kinta Valley and Kuala Lumpur
developed along river channels and beneath adjacent alluvial plain. Summits areas of Malaysia; Sulawesi;
Page 494
coastal margins, usually covered with of hills often form cockpit karst. Hills Cuba; Vietnam; southern
alluvium, often with isolated upstanding frequently exploited for limestone China; Sarawak.
residual tower karst hills. Buried karst aggregate. Foundation problems on
plains often have cavern systems in the plains acute due to difficulty of
underlying limestone. locating buried karst pinnacles and
partially filled caverns. Virtually nothing
known of deep karstic water circulation
below buried plains.
3. Corridor and crevice Flat-topped plateaux with narrow, often Irregular spacing of corridors. Kikori area, New Guinea
discontinuous corridors cut in the Some may be obscured by vegetation. (Löffler, 1977).
limestone preferentially along joints and
faults and other lines of weakness. Crevice
karst is usually at more closely dissected
scale, akin to the grikes of northern
England.
4. Doline Enclosed depressions pitting a plateau Dolines often filled with alluvium. Müller Plateau,
surface. Relief akin to that of limestone Extremely variable in depth. New Guinea (Löffler, 1977).
areas of Europe, with occasional Connections to subsurface drainage
enclosed depressions and dry valleys. complex.
GFE-18.qxd 6/25/05 2:01 PM Page 495
HOT WETLANDS 495
A good example of the problems in these buried a complex series of buried pinnacles and cavities
karst plains is provided by the building of a 4 km that pose tremendous challenges to engineers.
tunnel for the Light Rail Transit (LRT) System in Overlying the limestone in places is the Kenny Hill
Kuala Lumpur. The metamorphosed limestone has Formation consisting of interbedded shales and
sandstone of upper Silurian-Devonian age. Alluvial
deposits mainly consisting of loose silty sand and
gravels were deposited over these formations dur-
ing the Quaternary. The building of a tunnel
through these diverse materials saw the collapse of
thin sandy alluvium over the weathered Kenny Hill
Formation with excessive settlement and sinkhole
development near Masjid Jamek Station. When a
large limestone pinnacle 33 m below ground level
was encountered, high rates of water seepage
occurred through re-cemented weathered lime-
stone material of high permeability than the adja-
Figure 18.11 Solutionally rounded ceilings distinguish
cent hard rock. Surface investigations failed to
the large passages of the Niah Caves in Sarawak’s Subis show many of the karst features encountered or
Hills. Photograph courtesy of Tony Waltham Geophotos. indicate the presence of variable strengths of rock.
Figure 18.12 Typical borehole in buried karst terrain overlain with alluvium in the humid tropics.
GFE-18.qxd 6/25/05 2:01 PM Page 496
Awareness of the problems of buried karst and its Hong Kong has had to cope with similar prob-
complex relationship to the overlying formations lems. In the 1950s and 1960s the Hong Kong
was highly important in this case (see Chapter 24 Government built resettlement estates for immi-
for discussion of problems of site investigation and grants. Unfortunately, some of the earthworks
foundation engineering in Karst terrain). were not designed and constructed well enough
to cope with severe wet season rainstorms. The
result was frequent failures of artificial slopes,
18.8 Slope instability and landsliding culminating in 1972 in two major disasters on
the same day. On 18 June 1972 in Sau Mau Ping
Any hill, or steepland area, is difficult to develop Estate in Kowloon, a 40 m high road embank-
without having a risk of causing environmental ment collapsed, killing 71 people. This was fol-
damage. Throughout the humid tropics granite is lowed a few hours later by the collapse of the
often deeply weathered and can be liable to mass hillside above a steep temporary excavation on
movement, through soil slips or landslides, if Po Shan Road in the Mid-Levels area of Hong
vegetation is removed and if excavations are made Kong Island which triggered a landslide that
into steep slopes. The depth of unconsolidated demolished a 12-storey residential building and
material on such slopes may be well over 30 killed 67 people. Four years later, another severe
metres. While the bulk of this profile may be rainstorm hit Hong Kong and brought down
undisturbed weathered rock, some of it is likely to three fill slopes in Sau Mau Ping Estate
be colluvial and taluvial material, weathered rock which again were constructed without proper
and soil material that has been washed down from compaction. The resulting landslides killed 18
higher upslope. The presence of these materials people.
above weathered rock can create a zone of lateral After these tragic events the Hong Kong
water flow at the discontinuity. This lateral move- Government developed a policy on landslide risk
ment may facilitate soil slips and landslides. The management entitled ‘Slope Safety for All’
colluvial material usually found only on the lower aimed at meeting the highest standards of slope
parts of slopes is a problem when construction safety. The Slope Safety System is managed by
involves cut and fill techniques. the Geotechnical Engineering Office (GEO) of
In the areas subject to tropical cyclones the Civil Engineering Department (CED), with
(typhoons), landslides triggered by extreme rainfalls the overall target of achieving reduction in risk
are always a threat to urban areas. Throughout the to the whole community (e.g. GEO, 1984).
Philippines, landslides often occur, whether caused The System is based on seven key result areas
by typhoons or earthquakes. Roads, railways, and (Table 18.10).
other key transport systems can be rendered unus-
able, hampering rescue and relief operations and
isolating disaster areas from the rest of the com- 18.9 Extreme events and major
munity. In January 1999 at least 200 000 people geomorphological problems in
were made homeless by landslides in and around hot wetlands
the towns of Viga, Bagamanoc, Payo, Caramoran
and San Miguel on the island of Catanduanes. In The humid tropics are often classified as an
August 1999, the Manila area experienced the extreme or risky environment. This view empha-
heaviest rainfalls for 25 years. A torrent of mud sises the climatic hazards associated with heavy
engulfed the Cherry Hills housing estate in rainfalls, tropical cyclones, coastal storm surges
Antipolo, 36 people being lost. Elsewhere in the and El Niño Southern Oscillation (ENSO, see
area 46 people were killed in other floods and Chapter 2) related floods and droughts. In terms
landslides. The most vulnerable areas were the of engineering geomorphology it is important to
shanty towns which have grown up along rivers recognise what these climatic hazards mean for
and on steep hillsides (Gittings, 1999). earth surface processes. Wet and drying of the
GFE-18.qxd 6/25/05 2:01 PM Page 497
HOT WETLANDS 497
Table 18.10 Strategies and action to cope with landslides in Hong Kong.
Seven point strategy Other measures
Standards and Improve slope safety standards, technology, Publish geotechnical standards
regulations and administrative and regulatory
frameworks
Implementing Ensure safety standards of new slopes Catalogue some 54 000 sizeable man-made
standards slopes and retaining walls in Hong Kong
Improving slope Rectify sub-standard Government slopes Identify the maintenance responsibility
stability of all the catalogued slopes
Maintenance and Maintain all Government man-made slopes Extensive network of automatically recording
warning raingauges: real time rainfall data for
emergency warnings
Public Ensure that owners take responsibility 24-hour year-round emergency service
responsibility for slope safety
Awareness and Promote public awareness and response in Investigation of serious landslides
help slope safety through public education,
publicity, information services and public
warnings
Good design Enhance the appearance and aesthetics of Auditing the design and supervision of
engineered slopes construction of all new slopes
Prevention Landslip Preventive Measures (LPM)
Programme
Squatter Inspect squatter villages on steep hillsides:
settlements guidance to residents; clearance if necessary
land surface takes place rapidly. Most of the time River in Sabah (Douglas et al., 1999) show that
the climate above the tropical rainforest canopy the sediment load carried by a 1 in 10 year flow
is like that of a desert. In the fierce sunlight, the on a single day may exceed that carried in a whole
humidity is low and temperatures are high. Once year (Figure 18.13). Attention to records of past
the forest is removed, the surface of the ground is floods, extreme rainfalls, and mass movements is
subjected to this daytime aridity. of great importance in developing plans for new
When the rain starts, conditions change infrastructure or urban developments.
rapidly, ground water tables in stream head and Foundation conditions are complicated not
below valley floors respond rapidly to even a few only by the energy of the present-day environ-
millimetres of rain. Infiltrating water moves rap- ment, but also by the legacies of Pleistocene
idly to streams which begin to rise quickly; a river climatic and sea-level changes. Understanding of
only 4 m wide can rise 1 metre in 15 minutes. Quaternary events in the humid tropics has
Heavy rains lead rivers to flow into riverine vege- advanced greatly since 1980, but there are still
tation. Channels often do not have a well-defined uncertainities about the extent of climate and
bankfull stage of precise recurrence interval. vegetation changes. Buried former river channel
Flood height prediction from morphological and karst features are however a practical issue
features is not always easy. for anyone concerned with water supplies, major
Much of the sediment load is carried in a short infrastructure developments, city sewerage schemes
period of time. Close observations on the Segama or construction of high rise buildings.
GFE-18.qxd 6/25/05 2:01 PM Page 498
Figure 18.13 Cumulative plot of sediment yield over time for the Segama River at the Danum Valley Field Centre,
Sabah, Malaysian Borneo emphasizing the days on which extreme events occurred.
18.10 Summary
with inidividual rock types, such as the buried
Most hot wetlands, with their green covering of karst plains of limestone areas. The legacies of
tropical rain forest, appear similar to each other. the past are still less well understood in the
However, the jungle conceals a wide variety of humid tropics than in most other places and
forest types, engineering soils, rocks and land- planned new developments require careful and
forms. Engineering experience in one part of cautious investigation.
the hot wetlands is not necessarily transferable The humid tropics are an environment to be
to another. High mountain areas were subject to treated with care and caution. In particular, the
shifts in climatic and vegetational belts during downslope, downstream changes that result from
the Quaternary. Lowland areas were greatly changes in runoff and sediment yield must
affected by Quaternary sea-level changes. Soil always be borne in mind. Good foundation con-
formations, relict river channels and deposits of ditions and effective stabilization of one site may
sand and gravel typical of more seasonally, less mean the creation of off-site problems for some-
wet environments are frequently encountered. one else.
Large river flows, major mass movements, and
rockfalls tend to happen more frequently than
elsewhere on similar terrain, just because of the References
amount, intensity and magnitude of rain events.
Many humid tropical landsurfaces are old, Asdak, Chay., Jarvis, Paul G., Gardingen, Paul V. and
with deep weathering profiles. Others are Fraser, A. (1998) Rainfall interception loss in
unlogged and logged forest areas of Central
young, steep and unstable, with relatively thin Kalimantan, Indonesia. Journal of Hydrology 206,
regoliths. Particular problems are associated 237–244.
GFE-18.qxd 6/25/05 2:01 PM Page 499
HOT WETLANDS 499
Bidin, K. (2001) Spatio-temporal variability in rain- Haberle, S. (1997) Upper Quaternary vegetation
fall and wet-canopy evaporation within a small and climate history of the Amazon Basin: correlat-
catchment recovering from selective tropical ing marine and terrestrial pollen records. In Flood R.
forestry. Unpublished PhD. thesis, University of D., Piper, D. J. W., Klaus, R. and Peterson, L. C.
Lancaster. (1997) (eds) Proceedings of the Ocean Drilling
Colinvaux, P. A., Irion, G., Rasanen, M. E. and Programme, Scientific Results 155, 381–396.
Bush, M. B. and De Mello, J. (2001) A paradigm Johnsson, M. (2000) Chemical weathering and soils. In
to be discarded: geological and paleoecological Ernst, W. G. (ed.) Earth Systems: Processes and
data falsify the Haffer & Prance refuge hypothe- Issues. Cambridge University Press, Cambridge,
sis of Amazonian speciation. Amazoniana 16, 119–132.
609–646. King, J. P. (1996) The Tsing Shan Debris flow. GEO
Colinvaux, P. A., De Oliveira, P. E. and Bush, M. B. Special Report SPR 6/96.
(2000) Amazon and Neotropical plant communities King, J. P. (1997) Natural Terrain Landslide Study: The
on glacial time scales: the failure of the aridity and Natural Terrain Landslide Inventroy. GEO Report
refuge hypotheses. Quaternary Science Reviews 19, No. 74.
141–169. King, J. P. (1999) Natural Terrain Landslide Study: The
Dam, M. A. C. (1994) The Late Quaternary evolution Natural Terrain Landslide Inventory. GEO Technical
of the Bandung Basin, West java, Indonesia. Ph.D. Note TN 1/97.
thesis, Vrije Universiteit, Amsterdam. Laverty, M. (1980) Water chemistry in the Gunung
Douglas, I., Kawi Bidin, G., Balamurugan, N. A., Mulu National Park including problems of inter-
Chappell, R. P. D., Walsh, T., Greer and Sinun, W. pretation and use. Geographical Journal 146,
(1999) The role of extreme events in the impacts of 232–245.
selective tropical forestry on erosion during har- Löffler, E. (1977) Geomorphology of Papua New
vesting and recovery phases at Danum Valley, Guinea. ANU Press, Canberra.
Sabah. Philosophical Transactions of the Royal Lovegrove, G. W. and Fookes, P. G. (1972) The
Society of London B, 354, 1749–1761. planning and implementation of a site investiga-
Evans, N. C. and King, J. P. (1997) The Natural Terrain tion for a highway in topical conditions in Fiji.
Landslide Study: Debris Avalanche Susceptibility. Quarterly Journal of Engineering Geology 5,
GEO Technical Note TN 1/98. 43–68.
Evans, N. C., Hung, S. W. and King, J. P. (1997) The Milliman, J. D. and Meade, R. H. (1983) World-wide
Natural Terrain Landslide Study: Phase III. GEO delivery of river sediment to the oceans. Journal of
Special Project Report SPR 5/97. Geology 91, 1–21.
Fookes, P. G. (ed.) (1997a) Tropical Residual Soils. Morgan, R. P. C. (1986) Soil Erosion & Conservation.
Geological Society Engineering Group Working Harlow: Longman.
Party Revised Report. Geological Society Pickup, G., Higgins, R. J. and Warner, R. F. (1981)
Publishing, London. Erosion and sediment yield in the Fly River
Fookes, P. G. (1997b) First Glossop Lecture: Geology drainage basin, Papua New Guinea. International
for Engineers: the geological model, prediction and Association Hydrological Sciences Publication 132,
performance. Quarterly Journal of Engineering 436–456.
Geology 30, 290–424. Prance, G. T. (2002) Tropical Forests. In Mooney, H. A.
Franks, C. A. M. (1996) Study of Rainfall Induced and Canadell, J. G. (eds) The Earth System: biological
Landslides on Natural Slopes in the Vicinity of Tung and ecological dimensions of global environmen-
Chung New Town, Lantau Island. GEO Special tal change. (Vol. 2 of Encyclopaedia of Global
Project Report SPR 4/96. Environmental Change (ed.) Munn R. E.) Chichester,
GEO (1984) Geotechnical Manual for Slopes. Geo- Wiley, 582–586.
technical Control Office, Hong Kong Government. Riebe, C. S., Kirchner, J. W., Granger, D. E. and
Gittings, J. (1999) Floods engulf East Asia. The Finkel, R. C. (2001) Strong tectonic and weak cli-
Guardian, August 5th 1999, 13. matic control of long term chemical weathering
Gue See Sew and Muhinder Singh (2000) Design & rates. Geology 298, 511–514.
Construction of A LRT Tunnel in Kuala Lumpur, Shimen Reservoir Authority (1975) Shimen Reservoir
Seminar on Tunnelling, IEM, Kuala Lumpur, 28 Catchment Management Work Report. Taipei: The
February, 2000. Authority (in Chinese).
Gupta, A. and Pitts, J. (eds) (1992) The Singapore story: Stallard, R. F. (2000) Erosion. In Hancock, P. L.
Physical adjustments in a changing. Singapore: and Skinner, B. J. (eds) The Oxford Companion to
Singapore University Press. the Earth. Oxford University Press, Oxford, 314–318.
GFE-18.qxd 6/25/05 2:01 PM Page 500
Stanley, S., Chopra, P. and De Deckker, P. (2002) Sea Further reading

level changes in SE Asia over the last 150 000 years.
Implications for geographical settings and oceanic Douglas, I. and Spencer, T. (1985) Environmental
currents. In Chen-Tun, A. C. (ed.) Marine Envir- Change and Tropical Geomorphology. Allen &
onment: The Past, Present and Future. Sueichan Unwin, London.
Press, Taipei, 76–78. Reading, A. J., Thompson, R. D. and Millington, A. C.
Viet, Pham Bach, Nguyen, Lam Dao and Duan, Ho (1995) Humid Tropical Environments. Blackwell,
Dinh (2002) Using remotely sensed data to detect Oxford.
changes of riverbank in Mekong River, Vietnam. Styles, K. A. and Hansen, A. (1989) Geotechnical Area
www.gisdevelopment.net/application/natural_ Studies Programme: Territory of Hong Kong. GASP
hazards/floods/nhcy0009.htm. Report XII, Geotechnical Control Office, Hong
Williams, P. W. (1978) Interpretations of Australasian Kong.
karst. In Davies, J. L. and Williams, M. A. J. Landform Thomas, M. F. (1994) Geomorphology in the Tropics: A
evolution in Australasia. ANU Press, Canberra, Study of Weathering and Denudation in Low
259–286. Latitudes. Wiley, Chichester.
GFE-19.qxd 6/27/05 2:29 PM Page 501
19. Mountain Environments

John Charman and Mark Lee
19.1 Introduction still occurring. The main active mountain belts are
the Alpine–Himalayan belt, running from Borneo
It has been estimated that around 36% of the land through northern India into Iran, Turkey and
surface of the earth is covered by mountains, high- through southern Europe, and the Circum–Pacific
lands and hills (Fairbridge, 1968). Around 10% of belt, encompassing the Andes of south America
the world’s population live in mountain regions and the Rocky Mountains of the USA, and
and more than 40% are believed to be dependent encompassing the island arcs of the west Pacific.
on mountain resources (Messerli, 1983). They are These mountain belts are located close to tectonic
extremely dynamic and sensitive environments, plate boundaries and continuing seismic events
reflecting the combination of steep slopes, high are a firm indication that mountain building is
altitudes and relative relief, together with the still active. Their relative immaturity is reflected
presence of numerous relict landforms often in high relative relief, steep slopes, folded and
inherited from previous phases of glaciation. This fractured rocks and high rates of weathering and
landscape sensitivity is easily disrupted by defor- erosion.
estation, land use change and construction pro- Mountain belts formed in earlier orogenies,
jects, often leading to dramatic impacts on the principally the Caledonian and the Variscan
scale and intensity of landslide activity and soil
erosion.
Table 19.1 Classification of mountain types (from
A great variety of classification systems have
Fairbridge, 1968).
been developed to try and capture the distinctive-
ness of mountain environments (e.g. Table 19.1). Type Sub-type
Perhaps the most pragmatic are the simple div-
isions based on relative relief: 1. Structural, tectonic A: volcanic mountains
or constructional B: fold or nappe mountains
forms C: block mountains
1. hill: ⬍ 700 m D: dome mountains
2. highland and upland: 700–1000 m E: erosional uplift or outlier
3. mountain: ⬎ 1000 m. mountains
F: structural outlier
Alternatively, absolute height can be used: or klippe mountains
G: polycyclic tectonic
1. low mountain: ⬍ 1000 m mountains (Alpinotype)
2. intermediate mountain: 1000–1500 m H: epigene mountains
3. high mountain: ⬎ 1500 m. 2. Denudational, A: differential erosion or
subsequent, relict mountains
The locations of the world’s major mountain destructional or B: exhumed mountains
belts are shown in Figure 19.1. Many have been sequential forms C: igneous (plutonic) and
formed in the relatively recent geological past, metamorphic complexes
D: polycyclic denudational
originating in the Tertiary Alpine orogeny but in
forms
many areas, for example the Himalaya, uplift is
501
GFE-19.qxd 6/27/05 2:29 PM Page 502
MOUNTAINS
OF THE WORLD
Mountains
Widely–spaced
mountains
Submarine
connections of
fold systems
Submarine
rift systems
Figure 19.1 Distribution and types of mountain (from Gerrard, 1990).
(Hercynian), exist elsewhere but a longer history degree of resistance to weathering and erosion.
of weathering and erosion have worn them down Near the core of the mountain belt both volcanic
and the extreme relative relief, characteristic of (fine-grained) and deep-seated (coarse-grained)
the Alpine belts, is no longer present. igneous rocks tend to form. Intense metamorph-
ism also occurs, yielding slates, schists and
gneisses. The result is a suite of minerals that
19.2 Rocks, climate and formed in conditions far removed from present
weathering day atmospheric conditions and they are highly
susceptible to weathering (Chapter 2).
In young fold mountains the landscape is charac- In the outer regions sedimentary rocks, some-
terised by the interaction of several factors, times lightly metamorphosed, are more dominant.
including tectonic activity, geology, topography These rocks are often buried beneath igneous and
and climate. The most important geotechnical metamorphic rocks transported from the core
problems relate to the nature of the near surface zone by gravity slide and nappe structures.
rocks and soils, slope instability and drainage. In Mountain slopes are often controlled by dif-
tectonically active zones rapid rates of uplift lead ferential erosion of the various rock types.
to over-steepened slopes (Chapter 9). In active However, lithological variability can lead to a
mountain belts major shear zones and faults are complex pattern of weathering and soil forma-
initiated and continue to develop. Accompanying tion, with marked changes in landslide and
metamorphism produces rocks with preferred surface erosion potential. An indication of the
structural trends, for example, slaty cleavage and relative susceptibility of some of the more
schistocity. These structural planes exhibit lower common minerals is given in Table 19.2 (Chapters
shear strength than the intact rock. 3 and 4).
The influence of geology is reflected both in Table 19.3 gives an indication of the relative
rock type and rock structure. Mountain belts con- weathering resistance of the main rock types in
tain a variety of rock types which vary in their relation to their intact rock properties.
GFE-19.qxd 6/27/05 2:29 PM Page 503
MOUNTAIN ENVIRONMENTS 503
Table 19.2 Susceptibility to chemical weathering of limestones and gypsum-rich rocks such as evapor-
common rock minerals. ites tend to dissolve easily.
Fine-grained Weathering Minerals
in minerals susceptibility in igneous rocks Metamorphic rocks
sedimentary rocks Metamorphic rocks possess a texture that depends
on the character of the original rock and the par-
Primary minerals Most Primary minerals ticular conditions of temperature and pressure
Gypsum ↑ Olivine under which it has been modified. For example,
Calcite Ca-Plagioclase rocks that have been modified under high tempera-
Olivine, feldspar
tures and pressures during mountain building
Amphiboles Na-Plagioclase
episodes are often coarsely crystalline, such as
Biotite feldspar
Alkali feldspar Biotite gneisses.
Alkali feldspar
Rock fabric
Secondary minerals
Quartz The fabric of a rock is the spatial arrangement of
Illite ↑ the textural features. The rock structure is the
Hydrated mica result of processes that have impacted on the rock
Montmorillonite during and after deposition. Igneous rocks may
Hydrated contain flow bands, sedimentary deposits may
aluminium contain alternating beds of differing grain size
oxide and metamorphic rocks may contain a preferen-
Hydrated iron Least tial mineral orientation as a result of the dominant
oxide stress pattern during formation. Major faults and
joints result from post-depositional processes and
are a major factor in controlling the mass stability
of the rock mass. The major geological structural
Igneous rocks trends affect the major valley profiles, the mass
The texture of igneous rocks depend on the rate at stability mechanisms active on the slope and the
which the magma cools. Granites and gabbros are depth to which weathering will penetrate.
coarsely crystalline because they are emplaced
below the earth’s surface and cool relatively slowly. Climate
Basalts are finely crystalline because they are Climate varies considerably in mountainous
ejected onto the earth’s surface and cool quickly. regions because of the effects of altitude and slope
The coarser grained varieties, such as gabbros, orientation. The active mountain belts straddle the
weather more quickly than the finer grained var- main climatic zones and weathering is at its most
ieties, such as basalts, because they possess a higher intense in the tropical and equatorial regions.
porosity. Altitude is also an important climatic influence
and over relatively short distances a traverse across
Sedimentary rocks an area of high relative relief can experience sev-
Sedimentary rocks have a texture that depend on eral climatic zones. Landscape forming processes
the mode and distance of sediment transport and depend on the interaction of temperature and rain-
the conditions under which they were deposited fall and on the relative interaction of weathering
and subsequently buried. Such rocks may be (in situ decomposition) and erosion (sediment
loosely compacted and voided, densely compacted transport and mass wasting).
with a range of grain sizes or cemented with a sec- Temperature, both seasonal and daily, together
ondary constituent. Sedimentary mudrocks such with rainfall influences the rate and type of weather-
as clays and shales also contain clay minerals but ing. Mechanical weathering may cause breakage of
weather less quickly. Carbonate-rich rocks such as rock into more closely fractured components while
GFE-19.qxd 6/27/05 2:29 PM Page 504
Table 19.3 Resistance to weathering related to rock properties (modified from Cooke and Doornkamp, 1990).
Rock Physical weathering (disintegration) Chemical weathering (decomposition)
properties
Resistant Non-resistant Resistant Non-resistant
Mineral High feldspar content High quartz content Uniform mineral Mixes/variable mineral
composition Calcium plagioclase Sodium plagioclase composition composition
Low quartz content Heterogeneous High silica content High CaCO3 content
Ca CO3 composition (quartz, stable Low quartz content
Homogeneous feldspars) High calcic plagioclase
composition Low metal ion content High olivine
(Fe–Mg) Unstable primary
Low biotite igneous minerals
High aluminium ion
content
Texture Fine-grained Coarse-grained Fine-grained dense Coarse-grained igneous
Uniform texture Variable texture rock Variable texture
Crystalline or tightly Schistose Uniform texture (porphyritic)
packed clastics Coarse-grained Crystalline Schistose
Gneissic silicates Clastics
Fine-grained silicates Gneissic
Porosity Low porosity High porosity Large pore size Small pore size
Free-draining Poorly draining Low permeability High permeability
Low internal surface High internal surface Free-draining Poorly draining
area area Low internal surface High internal surface
Large pore diameter Small pore diameter area area
permitting free hindering free
drainage after drainage after
saturation saturation
Bulk Low absorption High absorption Low absorption High absorption
properties High strength, Low strength High compressive, Low strength
elasticity Partially weathered tensile strength Partially weathered
Fresh rock rock Fresh rock rock
Hard Soft Hard Soft
Structure Minimal foliation Foliated Strongly cemented Poorly cemented
Clastics Fractured, cracked Dense grain packing Calcareous cement
Massive formations Mixed soluble, Siliceous cement Thick-bedded
Thin-bedded insoluble mineral Massive Fractured, cracked
sediments component Mixed soluble,
Thin-bedded insoluble
sediments mineral component
Representative Fine-grained granites Coarse-grained Acidic igneous Basic igneous varieties
rocks Some limestones granites varieties Limestones
Diabases, gabbros Dolomites, marbles Crystalline rocks Marbles, dolomites
Coarse-grained Many basalts Rhyolites, granites Poorly cemented
granites Soft sedimentary Quartzite sandstones
Rhyolites rocks Granitic gneisses Slates
Quartzites Schists Metamorphic rocks Carbonates
Strongly cemented Schists
sandstones
Slates
Granitic gneisses
GFE-19.qxd 6/27/05 2:29 PM Page 505
chemical weathering causes decomposition of the rates. In this dynamic environment any rural
rock and the disaggregation of minerals into a soil management programme or new engineering
comprising a collection of discrete particles. project, such as a road or a hill irrigation canal
Given the role of weathering in producing a benefit from a careful evaluation of landslide
residual mantle of potentially erodible disaggre- and erosion hazard, allowing them to be planned
gated particles, rainfall is probably the most accordingly.
important climatic factor governing whether this The cyclic nature of mountain development
mantle is subject to soil erosion or landsliding. provides the basis for developing a model in
While annual rainfall totals have some influence which the system can be divided into land units
the greater role is provided by seasonal rainfall (zones) and further sub-divided into land facets.
patterns, particularly when the rainy season is Figure 19.2 is based on a mountain system classi-
populated by short intense storms which can pro- fication developed in Nepal (Fookes et al., 1985).
duce catastrophic slope erosion. The onset of The land units and sub-units (facets) are des-
intense periods of rainfall provides the medium to cribed in Table 19.4.
transport the weathered materials. In temperate The cycles of high tectonic activity lead to the
and colder climates the rate of weathering is conforming of narrow incised valleys. The steep
siderably slower so that significant thicknesses of slopes of these valleys, immediately bordering the
weathered materials do not form. In these regions main rivers, are often very unstable, depending on
transported soils are more prevalent. the underlying geological structure, and can be
With such a diversity of topography, geology areas of high landslide hazard. These are desig-
and climate a variety of issues can be encountered nated as land unit 4, characterised by slopes
in mountain regions, including: steeper than 35⬚, often steeper than 45⬚, and
actively degrading to shallower slope angles.
1. access and routing In periods of lower activity and relatively slow
2. large landslides uplift, continuing landslide activity eventually
3. debris flows and torrents produces shallower and more stable slopes. These
4. snow melt and flash floods less active areas are subject to a longer period
5. landslide dams of chemical weathering and because erosion is
6. glacial lake outburst floods less intense a mantle of weathered residual soil
7. snow avalanches develops. These are designated as land unit 3,
8. ice avalanches characterised by slopes shallower than 35⬚ and
9. suitable sources for construction materials. chemically weathered to produce friable and eas-
ily erodible soils.
These issues are considered briefly after an out- During these periods the river may begin to
line of the main terrain zones. widen the valley floor and deposit alluvium. The
alluvial areas are designated as land unit 5, char-
acterised by flat tracts of granular material, the
19.3 Mountain terrain model higher, older terraces having steep frontal slopes,
and the tops of the terraces being subjected to
In active young fold mountain belts, cycles of chemical weathering. The next phase of high
relatively rapid uplift initiate a period of intense activity initiates another cycle in which the
erosion as rivers cut down to lower base levels river cuts down through the alluvium, which is
and produce steep-sided valleys. Intervening left as a depositional terrace above newer steep
more dormant periods allow weathering agen- slopes of land unit 4 dropping down to the new
cies to dominate and cause rock decomposition, river level.
and the reduction in shear strength causes land- When considering an engineering project,
slide activity in the valley sides. Meanwhile, particularly linear projects which traverse sev-
periods of intense rainfall initiate high erosion eral land units, the use of the mountain zone
GFE-19.qxd 6/27/05 2:29 PM Page 506
Figure 19.2 The mountain terrain model (from Fookes et al., 1985).
classification to differentiate individual land ‘straight-line’ approach to pipeline alignment is

units and facets is invaluable. rarely feasible in rugged terrain. The most efficient
Terrain maps based on this system, with knowl- routing approach to minimising risk in such areas
edge of the distribution of land elements and involves adopting ‘ridge and spine’ or ‘ridge and
typical engineering approaches in each, provides spur’ alignments, wherever possible. These are the
engineers with the information to prioritise the topographic positions where landslide and ero-
location of individual sites or route alignments. sion problems are considerably lower than on the
adjacent hillslopes or valley side-slopes. The most
suitable ridgelines are those which are generally
19.4 Access and routing aligned with the desired alignment direction.
Where it is not possible to use spurs or ridgelines,
Steep, dissected terrain presents extreme difficul- it is desirable to climb or descend steep slopes
ties in providing safe, maintained access to a site normal to the contours, minimising the exposure
and associated borrow or spoil disposal sites to landslide hazard.
throughout the year. On road and pipeline sites it The principles of road alignment through
is often impossible to provide temporary access, mountainous terrain have been set out in Fookes
parallel to the alignment, without endangering the et al. (1985), based on their experience on the
integrity of the route. Construction materials and Dharan–Dhankutra Road, Nepal. The ideal corri-
other essential supplies must be transported along dor is along the mature slopes, stale terraces and
the right-of-way, potentially causing disruption to plateaux of land unit (zone) 3. Where it is neces-
operations. sary to traverse all the zones, the key principle is
Pipeline routes through mountainous terrain make distance in land unit (zone) 3 and make
should avoid sidelong ground and the crossing height in land unit (zone) 4, avoiding extended
of deeply incised gullys and stream valleys. The runs of the line across land unit (zone) 4 slopes
GFE-19.qxd 6/27/05 2:29 PM Page 507
Table 19.4 A mountain system classification: description of terrain units (see Figure 19.2).
Land Unit Land Facet
No. Description No. Description
1 High altitude glacial and periglacial areas

subject to glacial erosion, mechanical
weathering, rock and snow instability
and solifluction movements with thin rocky
soil, boulder fields, glaciers, bare rock
slopes, talus development and debris fans.
2 Free rock face and associated steep debris
slopes subject to chemical and mechanical
weathering, mass movement, talus creep,
freeze-thaw, and debris fan accumulation.
3 Degraded middle slopes and ancient valley 3A Ancient eroded terraces covered with a weathered
floors forming shallow eroded surfaces subject residual soil mantle generally up to 3 m thick.
to chemical weathering, soil creep, sheetflow, Slope angle generally ⬍ 35⬚ and stable. Often
rill and gully development and stream incision. farmer terraced. Highly susceptible to water
erosion.
3B Degraded colluvium comprising landslide debris
of gravel, cobbles and boulders in a matrix of
silt and clay. Slope angle ⬍ 35⬚. Relatively stable.
Often farmer terraced. Variable permeability.
4 Steep active lower slopes with chemical and 4A Bare rock slopes. Steep slope angles ⬎ 60⬚.
mechanical weathering, large-scale mass Stability dependent on orientation of
movement, gullying, undercutting at base discontinuities, such as joints and
and accumulation of debris fans and flows bedding planes.
of marginal stability. 4B Rock slopes with mantle of residual soil usually
⬍ 2 m thick. Steep slope angles ⬎ 45⬚.
Prone to extensive shallow debris slides.
Deeper instability as for 4A.
4C Active colluvium. Thick landslide debris
often at base of slope and subject to active
river erosion. Slope angle ⬎ 35⬚. Highly
unstable, particularly during wet season.
4D Degraded colluvium. Thick landslide debris.
Slope angle ⬍ 35⬚. Marginally stable and
susceptible to gradual downslope creep
during wet season.
5 Valley floors associated with fast flowing, 5A Top of old alluvial terraces above present river
sediment laden rivers, and populated by level. Generally flat to shallow, ⬍ 10⬚. Coarse
sequences of river terraces. granular and permeable soils. May be covered
by a less permeable residual soil mantle.
5B Front scarp face of old alluvial terraces.
Steep slope angle ⬎ 65⬚, but subject to
sudden collapse when cementation breaks
down under weathering or when subject
to toe erosion.
GFE-19.qxd 6/27/05 2:29 PM Page 508
(Figure 19.3). In addition, experience suggests the masses, intense physical weathering, harsh cli-
following principles: mates and, in neotectonic regions, seismic activity.
The full variety of landslide forms described in
1. Avoid structures and rock cuts, both of which Chapter 8 can be found in mountain regions, how-
are costly, unless the alternative is clearly more ever it is worth highlighting a number of land-
expensive. Rock cuts will be unavoidable on slide forms that are perhaps unique to this type of
slopes over 45⬚. environment.
2. Retaining structures will be needed on slopes
greater than 30⬚. Rock spreading and sagging
3. Earthworks problems will arise when the Rock spreading and sagging involves the large-
alignment generates a surplus or deficit of scale (volumes can exceed 1 M m3), extremely slow
material. On such slopes the location of bor- deformation of mountain ridges and scarps. This
row and spoil disposal sites should be given as often results in the development of double ridges,
much attention as the alignment itself. trenches (grabens) and uphill facing scarplets.
Hutchinson (1988) suggested that the features
represent the early stages of major landsliding. A
19.5 Large landslides number of different mechanisms can be recognised:
lateral spreading in homogeneous rocks and brittle
Mountains are extreme, high-energy environments rocks overlying ductile materials, and rock flows.
and probably the most landslide prone landscape Lateral spreading in homogeneous rocks
in the world. This is because of the combination of (Figure 19.4a) can involve ridge core subsidence
steep slopes, high relative relief, active river inci- and downward movement along low-angled shear
sion, widespread structural weaknesses in the rock surfaces, causing rock spreading at the ridge
Figure 19.3 Route selection through mountainous terrain (from Fookes, 1997).
GFE-19.qxd 6/27/05 2:29 PM Page 509
double ridge garben-like

(a) d i
squeezing out
Figure 19.4 Lateral spreading (A) in homogeneous rocks and (B) in a brittle formation over a ductile layer (from
Pasuto and Soldati, 1996).
crest. Examples have been described in the jointed or stratified hard rock masses (Bisci et al.,
Carpathians, USA, New Zealand and Venezuela 1996). Typical features include trench-like
(Pasuto and Soldati, 1996). grabens, double ridges and counterscarps, with
Lateral spreading in brittle rocks overlying bulges and other compressional features on the
ductile materials (Figure 19.4b) leads to the devel- lower slopes (Figures 19.6 and 19.7). It is
opment of trenches, gulls, grabens and pseudo- believed that the rock mass deforms through vis-
karst depressions in the competent rocks and cous flow because the confining pressures are
bulges in the softer materials. Movements may high and the deviator stresses too low to produce
extend for several kilometres back from a plateau shearing. On the upper and lower slopes where
edge (the zone of decompression). Figure 19.5 the confining pressures are low, the rock mass
illustrates the development of this style of move- moves along shear surfaces and deforms in a
ment which involves progressive visco-plastic dilatant mass.
deformation of the ductile material along a system As rock spreading and flow movements tend to
of partial shear surfaces or a band of softened be extremely slow, they do not generally present a
material. Over time these partial shear surfaces significant problem. However, as development
may combine to generate distinct block slides encroaches onto mountain regions there is a need
(Záruba and Mencl, 1982). to ensure that dams or hydro-electric power (HEP)
Rock flows (sackung) are creeping flow-type, stations, for example, are not located at the toe of
deep-seated deformations affecting densely these features. In some instances the creeping
GFE-19.qxd 6/27/05 2:29 PM Page 510
Figure 19.5 Schematic sequence of rock spreading in brittle and ductile rocks (from Cancelli and Pellegrini, 1987).
deformations may be indicative of the potential for long-term remedial measures are unlikely to be
the development of a catastrophic landslide that effective.
could cause widespread damage and loss of life.
Field investigations are needed to assess the risk Catastrophic landslides
associated with these sites. Monitoring of the Many high-impact landslide events in mountain-
displacements through the use of GPS, geodetic or ous regions involve rapid displacement and long
topographic surveys or crack monitoring can be an run-out of the debris (Chapter 8, Table 8.1; see
important tool in managing the risks. However, Voight, 1978, for an indication of the enormous
because of their scale and the depth of movement, scale of some historical rock avalanches). The
GFE-19.qxd 6/27/05 2:29 PM Page 511
Figure 19.6 Block model of a rock flow (sackung) (from Bisci et al., 1996).
1970 Huascaràn disaster, for example, destroyed village of Morignone (Govi, 1999). The front of
the town of Yungay, Peru, killing between the rock mass continued up the opposite side of
15 000–20 000 people. An offshore earthquake the valley, where the runup was 300 m above the
(M ⫽ 7.7) triggered a massive rock and snow ava- valley floor. Part of the avalanche fell into a small
lanche from the overhanging face of the mountain lake and caused a 95 m high wave of muddy water
peak. The resulting turbulent flow of mud and and debris to travel upstream, flattening the vil-
boulders (estimated at 50–100 ⫻ 106 m3) passed lages of Poz, San Antonio and Tirindre. Twenty-
down the Rio Shacsha and Santa valleys as a 30 m seven people were killed when the mud wave hit
high wave travelling at an average speed of Aqmlone 2.1 km upstream. The main avalanche
270–360 km/hour in the upper 9 km of its path. continued downstream, travelling 1.4 km and
The landslide run-out buried the towns of Yungay forming a landslide dam across the valley.
and Ranrahirca beneath 10 m of debris (Plafker An earthquake in 1987 triggered landslides in
and Ericksen, 1978). the north-eastern Ecuador Andes which resulted
The Bairaman landslide, Papua New Guinea, in the destruction or local severance of nearly
in May 1985 in the Nakanai Mountains of New 70 km of the trans-Ecuadorian oil pipeline and the
Britain was 200 m thick and covered 1 km2, with only highway from Quito to the eastern rain
an estimated volume of 180 ⫻ 106 m3. The slide forests and oil fields (Nieto and Schuster, 1999).
probably began as a down-dip failure of saturated Economic losses were estimated at $1B, with
weak biospartite. The mass then broke down into between 1000–2000 deaths caused by the land-
a debris avalanche flow that travelled distances of slides and associated flooding.
2 km upstream and 1.5 km downstream, damming In the 1991 rock avalanche off the top of
the Bairaman River (King et al., 1999). Mount Cook in the Southern Alps of New
The 1987 Mount Zandila rock avalanche in the Zealand (McSaveney et al., 1999), an estimated
Valtellina of the central Italian Alps destroyed the 14 ⫻ 106 m3 of rock cascaded downslope, with an
GFE-19.qxd 6/27/05 2:29 PM Page 512
Figure 19.7 The main types of rock flow (from Hutchinson, 1988). Notes 1 ⫽ normal, down movement scarps;
2 ⫽ upslope, down movement scarps; 3 ⫽ upslope, up movement scarps; G ⫽ graben.
initial fall of 1500 m and an average speed of 1997 rapid Harihara debris slide-flow in Izumi
300 km/hour. Part of the avalanche overtopped an City killed twenty-one people and destroyed nine-
adjacent ridge, the remainder was deflected down teen homes and eleven non-residential buildings.
the Tasman Glacier. The runup at the far side of The slide was triggered by heavy rainfall during
the glacier was 70 m. the rainy season and involved 1.3 ⫻ 105 m3 of
Large run-out landslides are a regular hazard debris (Sassa et al., 1998).
in Japan. For example, the 1995 Nikawa land- The threat posed by such events has led to con-
slide, in the Kobe area destroyed eleven houses siderable research being undertaken to explain the
and buried thirty-four people. The landslide was speed of movement. A failure mechanism was
triggered by the Hyogoken–Nanbu earthquake — proposed by Sassa (1996) by which ‘sliding-
7.2 Richter magnitude (Sassa et al., 1999). The surface liquefaction’ occurs when a shear surface
GFE-19.qxd 6/27/05 2:29 PM Page 513
develops in sandy soils and the grains are crushed flow disaster in the mountainous region of Serra
or comminuted in the shear zone. The resulting do Mar, south-eastern Brazil (Ogura and Filho,
volume reduction causes excess pore pressure 1999), heavy rainfall triggered a rockslide on the
generation which continues until the effective slopes above Petropolis City, which slid from
stress becomes small enough that no further grain approximately 500 m above street level. The slide
crushing occurs. This mechanism is common in moved downslope as a debris flow, mobilising
earthquake-triggered debris slides and flow. The further colluvial material. The flow destroyed or
mechanism can also take place in rain-induced damaged housing at the slope foot and killed six
landslides if the amount of grain crushing in the people.
shear zone is high, as in andesitic debris. In 1988 an earthquake (magnitude 5.5) south-
Large runout, rapid landslides generally pres- west of Dushanbe, Tajikistan triggered a series of
ent a very low probability, high consequence landslides in loess country (Chapter 25). The
threat to mountain societies. Landslide hazard slides turned into a massive mudflow (20 ⫻ 106 m3)
assessment and avoidance of critical settings are which travelled around 2 km across an almost
likely to prove the most effective means of man- flat plain (Ishihara, 1999). It buried more than
aging the risk, along with monitoring of vulner- 100 houses in Gissal Village under 5 m of debris;
able slopes linked to early warning and evacuation 270 villagers were killed or reported missing.
procedures (Chapter 8). In Japan, Sassa (1996) In 1993 heavy rainfall in Kagoshima Prefecture,
has suggested that those slopes that are prone to Japan caused widespread debris flows in the
sliding surface liquefaction are potential sites Ryugamizu District (Sasahara and Tsunaki, 1999).
of large runout landslides. The key slope condi- Soil from shallow hillside failures transformed
tions are the presence of a saturated layer into channelled debris flows that seriously dam-
within the slope and soils susceptible to sliding aged a hospital, houses, roads and railways. The
surface liquefaction, such as weathered granitic flows killed 20 people; 2500 were cut off from
soils, volcanic deposits and sandy river terrace their homes.
deposits. In 1994 an earthquake (magnitude 6.4) in
the Paez Valley, south-west Colombia triggered
landslides and debris flows over an area of around
19.6 Debris flows and torrents 250 km2 (Martinez et al., 1999). Some debris
flows travelled some 120 km downstream, destroy-
Debris flow (debris torrents) comprise a slurry of ing villages, six bridges and 100 km of road before
fine (sand, silt and clay) and coarse material (cob- being dumped into a reservoir. The maximum
bles and boulders) mixed with varying quantities debris flow reached 10–40 m above the pre-earth-
of water — natural debris flows may have viscosi- quake river level, with a velocity of 8–12 m/s.
ties of over 1000 poise and densities of Table 19.5 summarises the main protection
2–2.4 g/cm3, whereas water has a viscosity of 0.01 measures that can be use to mitigate debris flows
poise and a density of around 1.0 g/cm3 (Costa, hazards (Hungr et al., 1987); all are reliant upon a
1984). The mixture moves downslope, often along thorough investigation of the problems (Chapter 8).
pre-existing drainage paths, in surges induced by So-called ‘passive’ measures either involve avoid-
gravity and channel bank collapses. They are a ing the risk areas or minimising the impact of
common and highly destructive feature in moun- debris flows through the use of early warning
tain environments, where earthquakes, heavy systems and evoking evacuation procedures.
rainfall or snowmelt can mobilise surface debris ‘Active’ measures can be directed towards the dif-
and the thin soil cover and incorporate it into a ferent parts of a debris flow catchment: the source
flow (Costa and Wieczorek, 1987). Observed areas, transport zones and deposition zones.
velocities are in the range 0.5–20 m/s. Source areas: check dams can be installed along
There is much damage associated with debris debris flow channel sections to prevent scour and
flow activity for example in the 1988 Morin debris remobilisation of the bed and bank materials.
GFE-19.qxd 6/27/05 2:29 PM Page 514
Channel linings (e.g. grouted riprap) can also be areas (Figure 19.11). Closed debris barriers and
used to stabilise the channel bed (Figure 19.8). storage basins (Figure 19.12) can be constructed
Transport zone: as the blockage of a debris to store design flows while passing water dis-
flow channel can lead to avulsion (uncontrolled charges; spillways can act as debris overflows in
channel branching) or flooding, measures can be an emergency.
used to improve the ability of the channel to pass
the debris surges downstream. Options include
debris chutes (Figure 19.9), channel diversions or 19.7 Snow melt and flash floods
bridges designed to allow the free passage of
flows (e.g. Figure 19.10). Sacrificial bridges can Mountain streams and rivers are strongly influ-
be used that can pass minor surges but will fail in enced by the seasonal impact of snowmelt and
large flows: the bridge can be anchored at one glacier melt. When combined with steep, bare
abutment with a cable to ensure that it does not rocky and boulder strewn catchments, these
join the flow, adding to its destructive power. processes tend to give rise to flash floods and
Deposition zone: debris sheds or galleries compressed annual hydrographs. The discharge
can be used to protect transportation routes characteristics of the Upper Hunza River in the
where they cross debris flow paths. Open basins Karakorum Mountains illustrate this point (Figure
bounded by dykes can be used to divert and store 19.13). Between November and February, the
debris, preventing the accumulation in high-risk river is fed by groundwater flows. As snowmelt
Table 19.5 Debris flow protection measures (from Hungr et al., 1987).
Measure Purpose
Passive measures
Hazard mapping and zoning Restrict use of endangered areas
Warning systems: advance, Facilitate evacuation at times of danger
during event or post-event
Active measures: source area
Reforestation/controlled harvest Reduce debris production due to logging or natural
loss of forest cover
Forest road construction Eliminate unstable cuts and fills that could be
debris sources
Stabilisation of debris sources Stabilise channel bed and side slopes in source reaches
(channel linings or check dams)
Active measures: transport zone
Training by chutes, channels, deflecting Ensure passage of debris surges, without blockage
walls or dykes or overflowing
Channel diversion Change path of flow away from endangered area
Bridges designed for passage Protect traffic and prevent blockage due to bridge obstruction
‘Sacrificial’ bridges Prevent channel blockage due to bridge obstruction
Bypass tunnels beneath creek bed Protect transport route without modifying stream channel
Active measures: deposition zone
Open debris basins, dykes and walls Control extent of natural deposition area by shaping
and dyking
Closed retention basins and barriers — full Create a controlled deposition area fronted by a barrier
or partial volume and spillway
Bridges or other structures designed for burial Prevent damage during burial by a flow
Debris sheds, galleries or cut and cover tunnels Place transport route beneath deposition area
GFE-19.qxd 6/27/05 2:29 PM Page 515
Figure 19.8 Channel stabilisation measures, Alberta

Creek, British Columbia (from Hungr et al., 1987).
Figure 19.10 Bridge on Coquihalla Highway, British

Columbia, designed to allow debris flow passage (from
Hungr et al., 1987).
British Columbia, the maximum discharge occurs

around 19:00–20:00 hours, with a minimum around
09:00–10:00 hours (Slaymaker, 1974). There can
also be pronounced year-to-year fluctuations in
discharge as a result of annual variations in snow
cover and glacier mass balance.
Rainfall-triggered flash floods are common,
especially during the latter part of the snowmelt
season. Flash floods are characterised by the rapid
rise and fall of the floodwaters, with peak flows
often occurring within hours of the onset of heavy
rain. The combination of steep terrain, very high
runoff, small catchments and episodic heavy rain-
fall creates the unique flash flood problems of
upland areas. The rapid rise of floodwaters is a
reflection of the speed at which water is delivered
to the stream channel because of the very high
Figure 19.9 Debris chute on Alberta Creek, British runoff rates and the fact that water arrives at the
Columbia (from Hungr et al., 1987). channel at nearly the same time throughout the
small catchment (i.e. a short time of concentra-
tion: Chapter 11). The absence of significant
begins in the lower parts of the basin in March, so floodplain areas adjacent to many mountain
the discharge increases. Between 40–70% of the streams tends to intensify the flood conditions as
total annual discharge occurs in July and August, floodwaters are not ‘stored’ in relatively slow-
where the flows can be up to forty times those flowing backwaters.
earlier in the year. The Highlands of Scotland are noted for
Marked diurnal fluctuations in discharge are frequent flash flooding (Table 19.6); in one event,
common in mountain streams, reflecting both in 1829, the River Findhorn rose 15 m above its
day–night variations in snowmelt and interruptions normal level during the course of a storm (Nairne,
in melting conditions during periods of cloud cover 1895). In many cases the impact of flash floods
or during snowfalls (Figure 19.14). In Miller Creek, may extend beyond the inundation of communities
GFE-19.qxd 6/27/05 2:29 PM Page 516
Figure 19.11 Layout of upstream barriers, debris basin and impact wall around a property site in San Mateo
California (from Baldwin II et al., 1987).
adjacent to the affected watercourses. Such events The area at risk within a vulnerable catchment
can also be very effective in mobilising sediment is likely to be a narrow strip alongside stream
in upland areas and initiating both hillslope and channels or within dry valleys. Such areas can be
channel erosion, as in the 1956 Cairngorm floods, readily defined from geomorphological evidence.
and deposition of spreads of gravel and cobbles It is also important to recognise the potential
across the river floodplains. An example of the hazards within valleys below dams and reservoirs.
large, destructive floods that occur in arid moun- An alternative, more sophisticated and costly
tains is provided by the Wadi Mikeimin (a tribu- approach to managing the flash flood risk is
tary of Wadi Watir) flood of January 1971 in the through the implementation of a flood warning
Aqaba Mountains (Schick and Lekach, 1987; see system, as in the Aqaba Mountains (Chapter 16;
Chapter 16). Figure 16.25). In the USA, the National Weather
GFE-19.qxd 6/27/05 2:29 PM Page 517
Figure 19.12 Debris flow barrier on Coquihalla Highway, British Columbia, designed to store debris flow material
(from Hungr et al., 1987).
Service issues a flash flood watch bulletin that therefore, a need for careful design and regular
signifies that rains will cause flooding in a partic- maintenance of bridge piers, trash screens and
ular area, whereas flash flood warnings are issued sediment traps upstream of culverted and other
when flooding is imminent. However, perhaps the restricted stream sections.
greatest problem associated with flash floods is
their unpredictability.
Bridges and culverted sections of streams can 19.8 Landslide dams
collect the debris swept downstream by the flood-
waters, forming temporary dams. The subsequent Landslides that occur in mountainous terrain
build up of water and dam breach can cause can block narrow, steep-sided valleys and form
destructive surges of water, greatly enhancing the landslide dams (Schuster, 1986). Large rock ava-
destructive potential of a flood event. There is, lanches and rock slides are particularly likely to
GFE-19.qxd 6/27/05 2:29 PM Page 518
Figure 19.13 Discharge characteristics of the Upper Hunza River, Karakorum Mountains (from Ferguson, 1984).
form high dams because they tend to be high- generally occurs within a year of dam formation
velocity events that can cause complete stream or (Figure 19.16). The effects of the resultant flood-
river blockage. Tables 19.7 and 19.8 provide an ing can be devastating, as indicated by the
indication of the scale of some recent landslide following examples.
dams and a simple classification, respectively. In the winter of 1840–41 a spur of the Nanga
Landslide dams give rise to two important Parbat Massif failed during an earthquake, com-
flood hazards: upstream or backwater flooding, pletely blocking the River Indus and causing the
and downstream flooding. impoundment of a 60–65 km long lake. The dam,
Upstream or backwater flooding occurs as a which had been up to 200 m high, breached in
result of the relatively slow impounding of water early June 1841. The lake emptied in 24 hours
behind the dam. In 1983, the small town of causing what has been called ‘the Great Indus
Thistle, Utah was inundated by a 200 m long, flood’ during which hundred of villages and
50–60 m deep lake that formed behind a landslide towns were swept away. A Sikh army encamped
dam on the Spanish Fork River (Figure 19.15; close to the river about 420 km downstream was
Kaliser and Fleming, 1986). A lake that had overwhelmed by a flood of mud and water esti-
formed behind a rock avalanche dam in 1513 led mated to be 25 m high; about 500 soldiers were
to the droning of the Swiss hamlets of Malvaglia killed (Mason, 1929).
and Semione (Eisbacher and Clague, 1984). Failure of the Deixi landslide dam on the Min
Downstream flooding occurs in response to the River, China in 1933 resulted in a wall of water
failure of the landslide dam. The most frequent that was 60 m high 3 km downstream (Figure
failure modes are overtopping because of the lack 19.17). The floodwaters had an average velocity
of a natural spillway, or breaching due to erosion. of 30 km/hour and reached the town of Maowen
Failure of the poorly consolidated landslide debris (58 km downstream) in 2 hours. The total length
GFE-19.qxd 6/27/05 2:29 PM Page 519
Figure 19.14 Daily variations in discharge of glacial fed rivers (redrawn from Ferguson, 1984; Slaymaker, 1974).
of the flood was 253 km and at least 2423 people preventing overtopping, either through the con-
were killed (Li et al., 1986). struction of a spillway or draining the lake waters
Between thirteen and forty people were killed through pipes or tunnels. Although many dams
in the Canadian Cordillera in 1891, when a land- have been successfully controlled, in some cases
slide dam burst engulfing an Indian village near overtopping has occurred before stabilisation
the North Pacific Cannery on Inverness Passage measures were fully implemented.
(Evans, 1986).
As indicated in Figure 19.18, the peak dis-
charge associated with the failure of landslide 19.9 Glacial lake outburst floods
dams is lower than the equivalent man-made
structures. This is because full breach tends Proglacial lakes impounded behind moraines or
to develop slower in landslide debris and, hence, an ice dam can present a major flood risk to
the discharge is spread over a longer time downstream communities. Failure of the
period. moraine or ice dam can lead to a catastrophic
There is an urgent need to undertake a outburst flood (aluvión, débâcles or glacier
rapid risk assessment once a landslide dam has lake outburst floods) with the discharge reaching
formed, because of their temporary nature and the up to several thousand cubic metres per second
potentially devastating consequences of down- (Reynolds, 1998), often accompanied by large
stream flooding. Engineering options generally volumes of mobile sediment. Perhaps the great-
involve controlling the dam discharge and thus est impacts have occurred in the Cordillera
GFE-19.qxd 6/27/05 2:29 PM Page 520
Table 19.6 Examples of flash flood events in the Scottish Highlands (from Nairne, 1895 and later sources).
Date Location Comment
3–4 August 1829 Moray district Around 90 mm of rain; the R. Findhorn rose 15 m above its normal
level causing immense damage; bridges swept away, crops and
farms destroyed or ruined by deposited gravels. Numerous families
left destitute and damage estimated at £20 000 (1829 prices).
Severe floods on the Nairn and Spey; ‘great landslips’ occurred,
farms swept away. In Spey valley damages estimated
at over £37 000, plus countless livestock and several lives lost.
27 August 1829 Inverness district Considerable flood damage in Inverness; crops flattened, numerous
bridges lost, mills and homes damaged. Estimated as
several thousand pounds damage.
24–26 January 1849 Inverness district The ‘Inverness Flood’: most disastrous flood in the NW Highlands.
Bridges lost at Aberchalder and Forst Augustus, Caledonian Canal
breached. In Inverness the stone bridge was lost and a third of the
town flooded by the combined waters of the Ness and the canal.
Immense damage but no loss of life.
30 January – Inverness district Farms damaged, crops lost, bridges swept away throughout district.
1 February 1868 Inverness flooded with extensive damage to property.
29 January 1892 Strathglass, Great, extensive flood following unpredicted snow falls for ten days.
Strathspey Damage extensive especially in Strathglass, Bonar-Bridge and
Strathspey, but no fatalities. Railways washed away on Skye.
25 May 1953 Lochaber, Appin Road bridges destroyed, disruption to road and rail traffic;
and Benderloch extensive damage to forestry property through floods and landslips.
In Argyllshire damage to roads estimated at £130 000.
30 July 1956 Cairngorm and Flooded houses and bridges damaged throughout region, especially
Moray around Forres. Main railway line from Inverness washed away.
Livestock swept away. Extensive erosion and deposition of
gravels on agricultural land. 72 hr maximum rainfall of 250 mm.
17–18 August 1970 Moray 72 hr maximum rainfall of 150 mm. Extensive damage to roads and
bridges, agricultural land flooded and covered by gravels.
Blanca of Peru, where there are some 722 gla- freight train; Jackson, 1979), Washington State
ciers covering an area of 723 km2 (Lliboutry et and Alaska in the USA (O’Conner and Costa,
al., 1977). An outburst flood in 1725 killed 1993; Stone, 1963) and the Alps (e.g. Tufnell,
between 1500 and 2000 people in the village of 1984), where the Vernagt glacier has caused seri-
Ancash. In 1941 an outburst from Laguna de ous débâcles in 1600, 1678, 1680, 1773, 1845,
Palcacocha destroyed part of the city of Huaraz 1847 and 1848.
and killed 6000. In 1945 the ancient town of Failure can result from a surge of the glacier or
Chavin de Huantar was destroyed by an outburst an ice avalanche into the lake, causing waves that
from the Quebrada Huachescsa. Similar floods either overtop or erode the dam. Moraine dams
have been widely reported in the Himalayas can be affected by piping of the heterogeneous
(Table 19.9; Reynolds, 1998), the Canadian materials, weakening the feature until it fails (e.g.
Cordillera (Clague and Evans, 1994; an outburst during an earthquake). If the moraine has a core
in 1978 damaged the Canadian Pacific Railroad of stagnant ice, its melting can result in the grad-
and the Trans-Canadian Highway and derailed a ual lowering of the crest height until the lake can
GFE-19.qxd
Table 19.7 A selection of historic landslide dams (from Schuster, 1986 and Sassa, 1999).
Landslide Year Dammed river Landslide Dam Dam Lake length Lake volume Dam
6/27/05
volume height width (km) (m3) failure
(m3) (m) (m)
Slumgullion Earth 1200–1300 Lake Fork, 50–100 ⫻ 106 40 1700 3 No
2:29 PM
Flow, USA Gunnison River
Usay landslide, Tadzhikistan 1911 Murgab River 2–2.5 ⫻ 106 300–550 1000 53
Partial
Lower Gros Ventre 1925 Gros Ventre River 38 ⫻ 106 70 2400 6.5 80 ⫻ 106 Yes
Page 521
landslide, USA
Deixi landslide, China 1933 Min River 150 ⫻ 106 255 1300 17 400 ⫻ 106 Yes
Tsao-Ling rockslide, Taiwan 1941–42 Chin-Shui-Chi River 250 ⫻ 106 217 2000 157 ⫻ 106 Yes
Cerro Codor Sencca 1945 Mantaro River 5.5 ⫻ 106 100 580 21 300 ⫻ 106 Yes
rockslide, Peru
Madison Canyon rockslide, 1959 Madison River 21 ⫻ 106 60–70 1600 10 No
USA
Tanggudong debris slide, 1967 Yalong River 68 ⫻ 106 175 3000 53 680 ⫻ 106 Yes
China
Mayunmarca rock slide, 1974 Mantaro River 1.6 ⫻ 109 170 3800 31 670 ⫻ 106 Yes
Peru
Gupis debris flow, Pakistan 1980 Ghizar River 30 300 5 No
Polallie Creek debris 1980 East Fork Hood River 70–100 ⫻ 103 11 230 105 ⫻ 103 Yes
flow, USA
Thistle earth slide, USA 1983 Spanish Fork River 22 ⫻ 106 60 600 5 78 ⫻ 106 No
Pisque River landslide, 1990 Pisque River 3.6 ⫻ 106 56 60 2.6 3.6 ⫻ 106
MOUNTAIN ENVIRONMENTS
Partial
Ecuador
Tunawaea landslide, 1991 Tunawaea Stream 4 ⫻ 106 70 80 0.9 9 ⫻ 106 Yes
New Zealand
Rio Torro landslide, 1992 Rio Torro 3 ⫻106 100 75 1.2 0.5 ⫻ 106 Yes
Costa Rica
La Josefina rockslide, 1993 Paute River 20–44 ⫻ 106 100⫹ 500 10 177 ⫻ 106
Partial
Ecuador
521
GFE-19.qxd 6/27/05 2:29 PM Page 522
Table 19.8 Classification of landslide dams (after Swanson et al., 1986).
Speed of Landslide type Comment

movement
Slow landslides A1 Basal shear surface emerges at Slides does not impinge on channel, but delivery of
ground surface upslope of channel colluvium may occur by secondary failures.
A2 Basal shear surface emerges in A common setting for landslide dams.
channel Potential for damming can be represented by:
Annual Speed of landslide toe
⫽
constriction ratio Channel width
Ratios in excess of 100 appear to be required
for development of lakes.
A3 Basal shear zone emerges on far Landslide movement can result in an upstream back-
side of channel water effect from upward heave of the streambed.
Limited storage capacity behind these dams.
Rapid landslides B1 Landslides small in relation to Typically cause only minor damming, although
width of valley floor channel can be diverted.
B2 Landslide deposits spanning the Typically produces a single impoundment which may
valley floor flood broad areas upstream.
B3 Landslide deposits covering valley Deposits may dam both main channel and tributaries,
floor over valley length much greater tributaries, forming multiple lakes. Volumes of
than the width impounded water may be very large where dams
are high.
Figure 19.15 Key events in the rise and fall of the Thistle landslide dam lake, 1983 (redrawn from Kaliser and Fleming,
1986).
GFE-19.qxd 6/27/05 2:29 PM Page 523
Figure 19.16 Survival time of landslide dams (redrawn from Schuster and Costa, 1986).
Figure 19.17 Downstream heights of the frontal wave of the 1933 Min River flood caused by the breaching of the
Deixi landslide dam, China (redrawn from Li et al., 1986).
overtop the dam. Climate change and glacier fluc- the moraine dam to lower the lake levels; the con-
tuations are key controls on the occurrence of struction of siphons, as used to drain the Tsho
outburst floods and it is worth stressing that Rolpha lake, Nepal in 1994; tunnelling through
Reynolds (1998) notes that the number and size bedrock beneath or beside a moraine dam into the
of high altitude lakes in both the Andes and lake to drain it (Figure 19.19); and the construc-
Himalayas has increased in recent decades tion or restitution of the natural dam by installing
because of climate change. a culvert or spillway across the moraine. The need
According to Reynolds (1992; 1998) and for ongoing monitoring of the whole glacier/
Reynolds et al. (1998), mitigation of the flood lake system is essential for successful hazard
hazard can involve: the excavation of open cuts in mitigation.
GFE-19.qxd 6/27/05 2:29 PM Page 524
Figure 19.18 Flood peak discharge versus dam factor (height ⫻ volume ⫻ 106) for constructed, landslide and
glacial dams (from Schuster and Costa, 1986).
Table 19.9 Recent glacier lake outburst floods in Nepal (from Reynolds, 1998).
Year Location River system Principal damage
1964 Gelhaipco Lake Arun/Pumqu End moraine collapsed due to ice avalanche into the lake;
road damaged and 12 trucks lost
1964 Arun/Pumqu Damage to forest, bridges and trucks
1964 Longda Glacier Trisuli River No data
1964 Zhangzangbo Glacier Sun Kosi/Poiqu Moraine collapsed due to seepage
1968 Aycio Lake Arun/Pumqu Damage to roads and bridges
1977 Nare Glacier Lake Dudh Kosi Ice-cored moraine collapsed. Damage to mini-HEP
station, road, bridges, fields etc.
1980 Phuchan Glacier Lake Sapta Kosi Damage to forest and river bed etc.
1981 Zhangzangbo Glacier Sun Kosi/Poiqu Moraine collapse due to glacier front calving. Damage
to Arniko Highway, bridges, Sun Koshi HEP station,
fields, killed livestock and caused casualties
1982 Jinco Lake Arun/Pumqu Moraine collapsed due to glacier tongue sliding into lake.
Damage to eight villages, livestock killed, fields roads
and bridges damaged
1985 Dig Tsho Dudh Kosi Moraine collapse after rock avalanche. Destroyed Namche
HEP plant, damaged roads, bridges, fields, houses and
caused casualties
19.10 Snow avalanches 1. Powder avalanches, in which an aerosol of fine

snow is carried in the air as a snow cloud. At
The rapid movement of detached masses of snow the front of the avalanche the snow may reach
presents a serious hazard in many mountain 20–70 m/s and victims can die by inhaling the
regions, often causing loss of life to both local snow particles.
inhabitants and tourists. Three main types of 2. Dry-flowing avalanches, involving dry snow
avalanche motion are recognised: travelling over steep irregular terrain with
GFE-19.qxd 6/27/05 2:29 PM Page 525
Figure 19.19 The tunnel design for draining the glacial lake at Hualcán, Cordillera Blanca, Peru (from Reynolds
et al., 1998).
particles ranging from powder grains to snow Table 19.10 Impact pressures and potential damage
blocks. These events tend to follow well defined from snow avalanches (from Perla and Martinelli, 1976).
channels such as gullys. Typical speeds range
Impact pressure (t/m2) Potential damage
from 15–60 m/s up to 120 m/s while travelling
through the air. 0.1 Break windows
3. Wet flowing avalanches, composed of wet 0.5 Push in doors
snow or a mass of sludge. These features tend 3.0 Destroy wood-frame houses
to flow in stream channels. Wet avalanches 10.0 Uproot mature trees
tend to be relatively slow moving (5–30 m/s) 100.0 Move reinforced
concrete structures
and can have a high density (300–400 kg/m3,
compared with 50–150 kg/m3 for dry flows).
As they support large volumes of debris and
rises further and meltwater appears within the
boulders they can cause considerable erosion
snow mass, so the risk of failure increases.
along the avalanche track.
Avalanches tend to be triggered by heavy snow-
Avalanche impacts can exert high external fall, rain, thaw or an increase in dynamic loading
loadings on structures in the path (typically in the such as skiers moving across the slope. Failure
range 5–50 t/m2, but up to 100 t/m2) and, hence, tends to start at a fracture point where there is high
can cause considerable damage (Table 19.10). In tensile stress, as at a break of slope or an overhang,
addition to the direct impact, avalanches may also and involves either: loose snow avalanche, which
exert both upwards and downwards forces which develops in cohesionless snow where the inter-
have lifted vehicles and buildings. granular bonding is very weak giving rise to
Although the nature and conditions vary behaviour similar to dry sand, or slab avalanche,
between avalanches (see Figures 19.20 and 19.21 where a cohesive snow layer breaks away from a
for avalanche classifications), all are associated weaker underlying layer (the ‘lubricating layer’).
with the weight of the snow exceeding the fric- The initial slab may be up to 10 000 m2 and involve
tional resistance of the underlying surface. Snow a 10 m thick slab. When a slab breaks loose it may
layers can sustain large density changes; a layer mobilise additional snow, with the avalanche
deposited with an original density of 100 kg/m3 increasing in volume by as much as 100 times.
may densify to 400 kg/m3 during the course of a Slope angle is a key factor in avalanche sus-
winter largely due to the weight of overlying ceptibility. Most events occur on slopes between
snow. Shear strength, however, decreases, as the 30–45⬚; slopes less than 20⬚ tend to be too
temperature rises towards 0 ⬚C. As temperature slack for failure to occur, whereas slopes above
GFE-19.qxd 6/27/05 2:29 PM Page 526
Figure 19.20 Classification of avalanches (redrawn from Embleton and Thornes, 1979).
Figure 19.21 Avalanche classification (from Verstappen, 1983).

GFE-19.qxd 6/27/05 2:29 PM Page 527
50⬚ rarely accumulate sufficient snow to present a that steer the avalanche by less than 15–20⬚
hazard. However, dry snow can avalanche on 4. direct protection structures such as snow sheds
slopes as low as 25⬚ and wet snow on slopes as low and galleries designed to allow the flow over
as 10⬚ (Mellor, 1978). As avalanches tend to recur railway line or roads (Figure 19.23) — in
at the same sites, the potential threat can often be France buildings in avalanche-prone commu-
readily defined by the mapping of past avalanche nities are oriented to allow the snow to pass
tracks and runout zones (e.g. from breaks of slope, between them (De Crecy, 1980).
eroded channels, damaged vegetation or different
Avalanches can also be managed by artificial
aged trees in forests).
release, through the use of explosives or weapons.
A variety of methods of avalanche mitigation are
This allows the unstable snowpack to be released
available, including the construction of defence
in a series of small manageable failures rather
structures (Figure 19.22; La Chapelle, 1977a;
than a single major event. Also the release can
1977b):
be timed to occur when the ski runs and highways
1. snow fences above the starting zone to trap and are closed. For example, Parks Canada and the
hold up snow that otherwise would have fallen Canadian Armed Forces trigger avalanches in the
onto the avalanche slope Rogers Pass where the Canadian Pacific railway
2. supporting structures such as snow rakes to pro- and the Trans-Canadian highway pass through the
vide external support for the snowpack — these Selkirk Mountains (Smith, 1992).
can stop small avalanches before they gain
destructive momentum 19.11 Ice avalanches
3. deflectors (e.g. wedges pointing upslope) or
retarding devices (e.g. earth mounds) built Unstable ice on steep mountain slopes can lead to
within the avalanche track or runout zone — the initiation of ice avalanches through the calving
the most successful deflectors have been those and free fall of ice from a hanging glacier or the
Figure 19.22 Avalanche mitigation measures (from Smith, 1992).

GFE-19.qxd 6/27/05 2:29 PM Page 528
vertical elevation of 4000 m. The scale of the dev-

astation was immense: 4000 people were killed, 6
villages were destroyed and 3 partly destroyed;
600 ha of agricultural land was lost as well as
many thousands of livestock; 10 flour mills and 4
bridges were destroyed (Morales, 1966).
In most instances the greatest threat is to
climbers and trekkers; an avalanche off the Le
Tour glacier, France killed six walkers in 1949,
and in 1981 part of a glacier in the Swiss Alps
swept down the mountain side killing six climbers
(Tufnell, 1984).
Little can be done to control the hazard other
than avoidance of avalanche-prone areas of the
Figure 19.23 Various approaches to avalanche pro- mountains or the monitoring of glacial flow
tection (from Verstappen, 1983). patterns and crevasse development to forecast
potential events.
detachment and sliding of tabular masses of ice.
Often such events fail to extend beyond the imme- 19.12 Construction materials
diate glacier margins and, hence, have little impact
on society. However, when the avalanche travels The low and moderate grade metamorphic rocks
further the results can be devastating. The 1962 that dominate many mountain areas present a series
Huascaràn avalanche in the Santa Valley, Peru of problems for construction (Fookes and Marsh,
began as a displacement of up to 3 million m3 of 1981a; 1981b). Typically these rocks show consid-
ice. As the ice moved downhill it was transformed erable variations in strength and durability with
into a mudflow of some 13 million m3 which orientation. Natural rock fragments and crushed
attained a maximum velocity of over 100 km/hour aggregates often have flaky and elongated, shapes
(McDowell, 1962; Morales, 1966). The avalanche reflecting the metamorphic fabric (Table 19.11).
travelled 16 km from its source and descended a The poor shape characteristics severely limits the
Table 19.11 Excavated rock characteristics from the Lower Himalayas (from Fookes and Marsh, 1981b).
Rock Type Shape characteristics of slightly to highly Soundness 10% fines:

weathered rock of slightly slightly to
to highly highly
weathered weathered
rock rock
Axial Ratio Numerical Numerical Loss, % Average

flakiness elongation passing value (kN)
a/b a/c b/c (Na2SO4)
Middle Siwalik sandstone 1.4 2.3 1.6 33 18 1.7–29.9 150

Sangure Series phyllite 1.7 4.3 2.5 69 40 0.2–22.7 87
Sangure Series metasandstones 1.6 3.0 1.9 52 23 2.0–6.9 123
Sangure Series white quartzite 1.5 2.4 1.6 35 13 1.5–3.1 188
Sangure Series purple quartzite 1.6 2.6 1.6 32 28 ⬍ 1.0 353
Lower Himalayan Unit plyllonite 1.8 5.9 3.3 95 43 0.5–14.2 70
and schist
Lower Himalayan Unit gneisses 1.6 2.7 1.7 39 21 0–5.2 79
GFE-19.qxd 6/27/05 2:29 PM Page 529
quantities of well shaped stone for masonry, gabion area of adhesion of bitumen to aggregate in surface
and rockfill; reduces the effectiveness of normal dressing, increasing the binder requirement, and
compaction procedures to produce dense, stable substantially increases the water requirement in
layers of granular material; reduces the effective concrete mixes for a given degree of workability.
Figure 19.24 Idealised block model of the Leoti Valley to Sangure Ridge, Lower Himalayas — see Table 19.12
for the accompanying legend (from Fookes and Marsh, 1981a).
Table 19.12 Mountain zones between the Leoti Valley and Sangure Ridge, Lower Himalayas — see Figure 19.24
(from Fookes and Marsh, 1981a).
Mountain zone 5. Valley floor 4. Active lower slopes 3. Degraded middle slopes 2. Free rock faces and
and ancient valley floors associated debris slopes
Topography Low angle Steep undercut slopes Undulating stepped slopes at Steep and very steep
slope ⬍15⬚ at 25–45⬚ moderate to low angles ⬍35o slopes ⬎25o
Foundation Sands and Coarse taluvium and Coarse and finer colluvium Coarse talus, taluvium
soils gravels, silty colluvium soils and residual soils with and some colluvium
sand paddy some taluvium and over residual soils and
unweathered sandy gravel unweathered rock
terraces and fans
Drainage Good Often poor with springs High rate of runoff and High rate of runoff and
and seepage generally poor subsoil good subsoil drainage
drainage
Available sub- Sands and Sands and gravel from Naturally well graded
base material gravel ancient terraces and fans quartzitic soils
Available Crushed gravels Crushed rock Crushed and uncrushed Crushed rock
base material and boulders, gravels occasionally
uncrushed crushed rock
gravels
GFE-19.qxd 6/27/05 2:29 PM Page 530
Table 19.13 Classification of common engineering soil types in the Lower Himalayas (from Fookes and Marsh,
1981a).
Soil type Description Weathering state of Typical slope

coarse fragments angles (degrees)
Residual Material remaining in situ during the weathering of rock Highly to completely
weathered
Colluvium An accumulation of usually well graded soil from the Moderately through ⬍35
weathering of rock which has been transported downslope to completely
mainly by gravity and slope wash processes, destroying weathered
the original rock structure; fine to coarsely graded
Taluvium Soil formed by the chemical and mechanical weathering Moderately to 45–25
of scree; often medium to coarsely graded or gap-graded completely weathered
Scree or An accumulation of angular rock fragments formed Slightly to moderately 38–33
talus by weathering and rockfall processes on exposed rock weathered
faces; often medium to coarsely graded
Gully infill An accumulation of sub-angular debris in stable gullies Slightly to completely ⬍25
and natural depressions which has been transported a weathered
short distance by water; fine to coarsely graded
Alluvium An accumulation of water transported weathering Slightly to completely ⬍15
products in either primary or reworked river deposits; weathered
fine to coarsely graded
Poor particle shape can hinder the achievement for pipelines. Grading is related to the source rocks
of a dense, fully compacted, fully interlocking base and the position within the catchment. Boulders
layer in road construction (Fookes and Marsh, are common within all the transported soils and
1981b). The flaky and elongated particles tend to can present problems during excavation. The
rotate to lie parallel to the surface during rolling and equipment used to exploit these sources should be
greatly reduce the effectiveness of binding tech- capable of handling very large boulders. Screening
niques in filling voids between the crushed stone. In uncrushed river gravels from coarsely graded river
a compacted, well graded, wet mix or crusher run deposits will result in a relatively low overall yield
base, flakes lying on the surface cannot be ade- of processed aggregates; two-stage crushing with
quately bound into the compacted mass and are eas- suitable equipment is more efficient.
ily worked out by subsequent construction traffic.
The potential for quarry development is often 19.13 Summary
limited because of the very steep natural slope
angles which would lead to unworkable quarry Mountains are one of the most dynamic environ-
faces. Local practice often involves exploiting river ments in the world and, as a result, can present a
boulder deposits for construction stone and range of major challenges to engineering projects.
crushed rock. However, working these deposits can Many of the hazards faced in these areas are
be wearing on mechanical plant. Seasonal floods complex, with earthquakes, snowmelt or heavy
will also restrict the safe period of working. rainfall being the trigger for potentially massive
Widespread slope instability and erosion tends events involving both landslides, debris flows and
to produce extensive fan and terrace deposits flooding that could cause widespread damage. The
within mountains and around their margins (Figure simple classification of the terrain into a series of
19.24; Tables 19.12 and 19.13). These generally distinct mountain zones provides a framework for
offer the best sources of mechanically stable fills, identifying and evaluating engineering hazards
natural sub-base and bases for roads and backfill and assessing resources. Figure 19.25 illustrates
GFE-19.qxd
6/27/05
2:29 PM
Page 531
MOUNTAIN ENVIRONMENTS
Figure 19.25 Classification of mountain hazards (redrawn from Messerli and Ives, 1984).
531
GFE-19.qxd 6/27/05 2:29 PM Page 532
how many hazards correspond with different Ferguson, R. I. (1984) Sediment load of the Hunza River.
zones and emphasises the interaction between In Miller, K. J. (ed.) International Karakorum Project
Vol. II. Cambridge University Press, 581–598.
natural processes and human activity.
Fookes, P. G. (1997) Geology for engineers: the geo-
logical model, prediction and performance. Quarterly
Fookes, P. G. and Marsh, A. H. (1981a) Some charac-
References teristics of construction materials in the low to mod-
erate metamorphic grade rocks of the Lower
Baldwin II, J. E., Donley, H. F. and Howard, T. R. Himalayas of East Nepal. 1: Occurrence and
(1987) On debris flow/avalanche mitigation and geological features. Proceedings of the Institution of
control, San Francisco area, California. In Costa, J. E. Civil Engineers, Part 1 70, 123–138.
and Wieczorek, G. F. (eds) Debris Flows/ Fookes, P. G. and Marsh, A. H. (1981b) Some charac-
avalanches: Processes, Recognition and Mitigation. teristics of construction materials in the low to mod-
Geological Society of America, Reviews in erate metamorphic grade rocks of the Lower
Engineering Geology VII. Boulder, Colorado, Himalayas of East Nepal. 2: Engineering character-
223–236. istics. Proceedings of the Institution of Civil
Bisci, C., Dramis, F. and Sorriso-Valvo, M. (1996) Engineers, Part 1 70, 139–162.
Rock flow (sackung). In Dikau, R., Brunsden, D., Fookes, P. G., Sweeney, M., Manby, C. N. D. and
Schrott, L. and Ibsen, M.-L. Landslide Recognition: Martin, R. P. (1985) Geological and geotechnical
Identification, Movement and Causes. Wiley, engineering aspects of low-cost roads in mountain-
Chichester, 150–160. ous terrain. Engineering Geology 21, 1–152,
Cancelli, A. and Pellegrini, M. (1987) Deep-seated Elsevier, Amsterdam.
gravitational deformations in the Northern Gerrard, A. J. (1990) Mountain Environments: an
Apennines, Italy. Proceedings 5th ICFL, Australia Examination of the Physical Geography of Moun-
and New Zealand, 1–8. tains. Belhaven Press, London.
Clague, G. J. J. and Evans, S. G. (1994) Formation and Govi, M. (1999) The 1987 landslide on Mount Zandila
failure of natural dams in the Canadian Cordillera. in the Valtellina, Northern Italy. In Sassa, K. (ed.)
Geological Survey of Canada Bulletin 464. Landslides of the World. Kyoto University Press,
Cooke, R. U. and Doornkamp, J. C. (1990) Geo- 47–50.
morphology in Environmental Management. Oxford Hungr, O., Morgan, G. C., VanDine, D. F. and
University Press. Lister, D. R. (1987) Debris flow defences in British
Costa, J. E. (1984) Physical geomorphology of debris Columbia. In Costa, J. E. and Wieczorek, G. F. (eds)
flows. In Costa, J. E. and Fleisher, P. L. (eds) Debris Flows/avalanches: Processes, Recognition
Development and Applications of Geomorphology. and Mitigation. Geological Society of America,
Springer–Verlag, Berlin 268–317. Reviews in Engineering Geology VII. Boulder,
Costa, J. E. and Wieczorek, G. F. (eds) (1987) Debris Colorado, 201–222.
flows/avalanches: processes, recognition and miti- Hutchinson, J. N. (1988) General report: morphological
gation. Geological Society of America, Reviews in and geotechnical parameters of landslides in relation
Engineering Geology VII. Boulder, Colorado. to geology and hydrogeology. In Bonnard, C. (ed.)
De Crecy, L. (1980) Avalanche zoning in France — Landslides. Balkema, Rotherdam, Vol. 1, 3–35.
regulation and technical bases. Journal of Ishihara, K. (1999) Liquefaction-induced landslide and
Glaciology 26, 325–330. debris flow in Tajikistan. In Sassa, K. (ed.) Landslides
Eisbacher, G. H. and Clague, J. J. (1984) Destructive of the World. Kyoto University Press, 224–226.
Mass Movements in High Mountains: Hazard and Jackson, L. E. (1979) A catastrophic glacial outburst
Management. Geological Survey of Canada Paper flood (jokulhlaup) mechanism for debris flow gen-
84–16, 230pp. eration in the Spiral Tunnels, Kicking Horse River
Embleton, C. and Thornes, J. B. (eds) (1979) Process in Basin, British Columbia. Canadian Geotechnical
Geomorphology. Arnold, London. Journal 16, 806–813.
Evans, S. G. (1986) Landslide damming in the Kaliser, B. and Fleming, R. W. (1986) The 1983 land-
Cordillera of Western Canada. In Schuster, R. L. slide dam at Thistle, Utah. In Schuster, R. L. (ed.)
(ed.) Landslide Dams: Processes, Risk and Landslide Dams: Processes, Risk and Mitigation.
Mitigation. Geotechnical Special Publication No. 3, Geotechnical Special Publication No. 3, American
American Society of Civil Engineers, 111–130. Society of Civil Engineers, 59–83.
Fairbridge, R.W. (1968) Mountain and lilly terrain, King, J. P., Loveday, I. and Schuster, R. L. (1999) The
mountain systems: mountain types. In Fairbridge, Bairaman river landslide and natural dam, Papua
R.W. (ed.) Encyclopedia of Geomorphology. New Guinea. In Sassa, K. (ed.) Landslides of the
Reinhold, New York, 745–761. World. Kyoto University Press, 51–53.
GFE-19.qxd 6/27/05 2:29 PM Page 533
La Chapelle, E. R. (1977a) Snow avalanches: a review O’Conner, J. E. and Costa, J. E. (1993) Geologic and
of current research and applications. Journal of hydrologic hazards in glaciated basins in North
Glaciology 19, 313–324. America resulting from 19th and 20th century
La Chapelle, E. R. (1977b) Alternative methods for the global warming. Natural Hazards 8, 121–140.
artificial release of snow avalanches. Journal of Ogura, A. T. and Filho, O. A. (1999) The Morin debris-
Glaciology 19, 389–397. flow disaster at Petropolis City, Rio de Janeiro State,
Li Tianchi, Schuster, R. L. and Wu Jishan (1986) Brazil. In Sassa, K. (ed.) Landslides of the World.
Landslide dams in south-central China. In Kyoto University Press, 289–292.
Schuster, R. L. (ed.) Landslide Dams: Processes, Pasuto, A. and Soldati, M. (1996) Rock spreading.
Risk and Mitigation. Geotechnical Special Publica- In Dikau, R., Brunsden, D., Schrott, L. and
tion No. 3, American Society of Civil Engineers, Ibsen, M.-L. Landslide Recognition: Identification,
146–162. Movement and Causes. Wiley, Chichester, 122–136.
Lliboutry, L., Arnao, B., Morales, A., Pautre, A. and Perla, R. I. and Martinelli, M. Jr (1976) Avalanche
Scneider, B. (1977) Glaciological problems set in Handbook. Agriculture Handbook 489 USDA
the control of dangerous lakes in the Cordillera (Forest Service), Washington DC.
Blanca, Peru. Part 1 Historical failures of moraine Plafker, G. and Ericksen, G. E. (1978) Nevados
dams, their causes and prevention. Journal of Huascaràn avalanches, Peru. In Voight, B. (ed.)
Glaciology 18 (79), 239–254. Rockslides and Avalanches, 1 Natural Phenomena.
McDowell, B. (1962) Avalanche! National Geographic Elsevier, Amsterdam, 277–314.
121, 855–880. Reynolds, J. M. (1992) The identification and mitiga-
McSaveney, M. J., Chinn, T. J. and Hancox, G. T. (1999) tion of glacier-related hazards: examples from the
Mount Cook rock avalanche of 14 December 1991, Cordillera Blanca, Peru. In McCall, G. J. H.
New Zealand. In Sassa, K. (ed.) Landslides of the Laming, D. J. C. and Scott, S. C. (eds) Geohazards.
World. Kyoto University Press, 56–59. Chapman and Hall, London, 143–157.
Martinez, J. M., Avila, G., Agudelo, A., Schuster, R. L., Reynolds, J. M. (1998) High altitude glacial lake haz-
Casadevall, T. J. and Scott, K. M. (1999) Landslides ard assessment and mitigation: a Himalayan per-
and debris flows triggered by the 6 June 1994 Paez spective. In Maund, J. G. and Eddleston, M. (eds)
earthquake, southwestern Colombia. In Sassa, K. Geohazards in Engineering Geology. Geological
(ed.) Landslides of the World. Kyoto University Society, London Engineering Geology Special
Press, 227–229. Publication, 15, 25–34.
Mason, K. (1929) Indus floods and Shyock glaciers: the Reynolds, J. M., Dolecki, A. and Portocarrero, C. (1998)
Himalayan Journal. Records of the Himalayan The construction of a drainage tunnel as part of gla-
Club, Calcutta, Vol. 1, 10–29. cial lake hazard mitigation at Hualcan, Cordillera
Mellor, M. (1978) Dynamics of snow avalanches. In Blanca, Peru. In Maund, J. G. and Eddleston, M. (eds)
Voight, B. (ed.) Rockslides and Avalanches. Elsevier, Geohazards in Engineering Geology. Geological
Amsterdam, 753–792. Society, London Engineering Geology Special
Messerli, B. (1983) Stability and instability of Publication, 15, 41–48.
mountain ecosystems: introduction to the work- Sasahara, K. and Tsunaki, R. (1999) Landslide disas-
shop. Mountain Research and Development 3, ters triggered by heavy rainfall in Kagoshima
81–94. Prefecture, July to September 1993. In Sassa, K.
Messerli, B. and Ives, J. D. (1984) Gongga Shan (ed.) Landslides of the World. Kyoto University
(7556 m) and Yulongxue Shan (5596 m). Geoeco- Press, 253–256.
logical observations in the Hengduan Mountains of Sassa, K. (1996) Prediction of earthquake induced
Southwestern China. In Lauer, W. (ed.) Natural landslides. In Senneset, K. (ed.) Landslides. Balkema,
Environments and Man in Tropical Mountain Ecosys- Rotterdam, 115–132.
tems. Springer-Verlag, Berlin, 55–77. Sassa, K. (1999) Geotechnical classification of land-
Morales, B. (1966) The Huascaràn avalanche in the slides. In Sassa, K. (ed.) Landslides of the World.
Santa valley, Peru. In International Symposium on Kyoto University Press, 139–145.
Scientific Aspects of Snow and Ice Avalanches. Sassa, K., Fukuoka, H. and Wang, F. W. (1998)
International Association of Hydrological Sciences Mechanisms of rapid long run-out motion in the
Publication 69, 304–315. May 1997 Sumikawa reactivated landslide in Akita
Nairne, D. (1895) Memorable Floods in the Highlands Province and the July 1997 Harihara landslide-
during the Nineteenth Century. Northern Counties debris flow, Kagoshima Prefecture, Japan. Journal
Printing and Publishing. of Japan Landslide Society, 35, 29–37.
Nieto, A. S. and Schuster, R. L. (1999) Mass wasting Sassa, K., Fukuoka, H. and Sakamoto, T. (1999) The
and flooding induced by the 5 March 1987 Ecuador rapid and disastrous Nikawa landslide. In Sassa, K.
earthquakes. In Sassa, K. (ed.) Landslides of the (ed.) Landslides of the World. Kyoto University
World. Kyoto University Press, 220–223. Press, 27–31.
GFE-19.qxd 6/27/05 2:29 PM Page 534
Schick, A. P. and Lekach, J. (1987) A high magnitude Stone, K. H. (1963) Alaskan ice-dammed lakes. Annals
flood in the Sinai desert. In Mayer, L. and Nash, D. of the Association of American Geographers 53,
(eds) Catastrophic Flooding. Allen & Unwin, 332–349.
381–410. Swanson, F. J., Oyagi, N. and Tominaga, M. (1986)
Schuster, R. L. (ed.) (1986) Landslide Dams: Processes, Landslide dams in Japan. In Schuster, R. L. (ed.)
Risk and Mitigation. Geotechnical Special Publica- Landslide Dams: Processes, Risk and Mitigation.
tion No. 3, American Society of Civil Engineers. Geotechnical Special Publication No. 3, American
Schuster, R. L. and Costa, J. E. (1986) A perspective on Society of Civil Engineers, 131–145.
landslide dams. In Schuster, R. L. (ed.) Landslide Tufnell, L. (1984) Glacier Hazards. Longman, London.
Dams: Processes, Risk and Mitigation. Geotechnical Verstappen, H. Th. (1983) Applied Geomorphology.
Special Publication No. 3, American Society of Civil Elsevier, Amsterdam.
Engineers, 1–20. Voight, B. (ed.) (1978) Rockslides and Avalanches, 1
Slaymaker, O. (1974) Alpine hydrology. In Ives, J. D. and Natural Phenomena. Elsevier, Amsterdam.
Barry, R. G. (eds) Arctic and Alpine Environments. Záruba, Q. and Mencl, V. (1982) Slope movements
Methuen, London, 134–155. caused by squeezing out of soft rocks. In Záruba, Q.
Smith, K. (1992) Environmental Hazards: Assessing and Mencl, V. (eds) Landslides and their Control.
Risk and Reducing Disaster. Routledge, London. Elsevier, Amsterdam, 110–120.
GFE-20.qxd 6/25/05 2:03 PM Page 535
20. Estuaries and Deltas

Warren E. Grabau, H. Jesse Walker and Mark Lee
20.1 Estuaries and deltas: northern Botswana, which has formed on the
introduction margins of an enormous dry depression.
The matter is still further confused by the fact
The very beginnings of human civilizations were that some definitions incorporate tides as an essen-
nurtured by deltas and estuaries, and that relation- tial component of both estuaries and deltas, thus
ship has continued unbroken through the ages up focusing on the ceaseless combat between fresh-
to and including the present time. Many of the water streams and saltwater tides. However, that
world’s major cities are sheltered by them. Their view seems too restrictive because the ancient Nile
waterways carry the cargoes of the world, and in delta, which gave its name to the landform, formed
some cases whole nations depend upon their for- at the margin of the nearly tideless Mediterranean.
tunate combinations of fertile soil and abundant Accordingly, the following discussion will be
water for their food supply. Given such a concen- confined to those features, both sedimentary and
tration of interests, it is hardly surprising that hydrologic, occurring where freshwater rivers meet
mankind, through combinations of engineering the saltwater sea.
works and management, has tried throughout the
ages to bend the natural environments better to
suit the needs of their human populations. 20.2 Estuaries
Estuaries and deltas are surprisingly difficult
to define, in part because they so often occur as The definitions of estuary are many and varied,
components of a single compound feature which but nearly all reflect one element of commonality,
consists of an association of a sub-aerial and namely the concept that an estuary is a semi-
sub-aqueous sedimentary structure (the delta) enclosed body of water which has at least one free
with a semi-enclosed sub-marine structure filled connection with the sea, as well as a supply of
with salt water (the estuary). However, it is fresh water, thus implying the presence of features
entirely possible for an estuary to occur without consequent to the mixing of saline and fresh
an accompanying delta, and a delta may occur water. Some extend the definition to include not
without an associated estuary. Thus, while the only the arm of the sea but all portions of the
two features are commonly closely associated, related freshwater streams that are affected by
and perhaps inextricably linked conceptually tides. However, the implication that tides are
within the human mind, such an association is essential would eliminate consideration of the
not a requirement. many river-meets-the-sea situations around the
Estuaries and deltas are usually defined as fea- nearly tideless Mediterranean. To avoid such an
tures that form where freshwater rivers meet the obvious imbalance, the zero-tide examples are
salt sea, ignoring the obvious fact that features simi- included as end-members of a continuous series
lar in almost every way, save for the presence of ranging from tideless to very high tides indeed.
salt water, form on the margins of freshwater lakes. Other definitions reflect in some degree the vary-
Indeed, features analogous in many ways to deltas ing points of view of the various scientific discip-
can form even without the presence of a water lines of those who study them, and thus focus on
body, as witness the eerie Okavango Swamp in the complexity of the interactions of hydrological,
535
GFE-20.qxd 6/25/05 2:03 PM Page 536
geological, chemical and biological processes that The modern sea level was reached about 6000
operate within them. years ago, and since that time the drowned valleys
have been modified in four principal ways, thus
Basic morphology producing four clearly recognisable types (Figure
In most cases sedimentation pushes the shoreline 20.1):
seaward from the head of the estuary, creating a
positive fill estuary. There are a few exceptions, in 1. unmodified drowned river valleys
which the estuaries receive the bulk of their sedi- 2. barred (drowned river valleys which have been
ment from offshore sources, thus creating nega- partially walled off from the sea by bars built
tive fill estuaries. There are even a few neutral across their mouths)
estuaries in which the delta front remains in equi- 3. fjords and fiards (glacier-scoured valleys
librium because the rate of sedimentation is invaded by the sea after the melting of the ice)
matched by the rate of flushing (transport of the 4. tectonic (tectonic structures, not necessarily
sub-aqueous portion of the delta to the open sea). originally associated with a river, that has been
The rate at which the sub-aerial delta margin is flooded by the sea).
pushed towards the sea varies greatly from estu-
ary to estuary. The most significant factors are the Two additional forms are sometimes recognised:
rate at which sediment is delivered to the delta by
the stream, the volume of the estuary which must 5. riverine estuaries
be filled and the flushing rate. 6. compound estuaries.
Many of the glacial-era valleys, such as those
of the Mississippi and the Nile, have already been Unmodified drowned river valleys (Figure 20.1)
filled and overfilled, so that the modern delta are especially common on temperate zone low-
extends well beyond the mouth of the original relief coastlines and are accordingly often referred
estuary. If sea level remains essentially unchanged, to as coastal plain estuaries, even though they are
the ends of all estuaries are foreordained: all will also often found on coastlines bordering low
eventually fill completely to the mouth, and most plateaux. In cross-section they tend to be shaped
will overfill, the exceptions being those where like an open V, like the original river valley, and
waves and currents are able to sweep sediments normally widen and deepen as they approach the
away as rapidly as the river deposits them. open sea. Because the invading sea usually covered
Estuaries vary enormously in size, shape, and the river floodplain as well as the river channel, the
environmental setting, with much of the variabil- deepest part of the estuary is commonly the sinu-
ity dependent on the mode of formation. In prin- ous channel of the pre-flooding river. However,
ciple, all estuaries are drowned valleys. Most such is not always the case, because sedimentation
existing estuaries, but by no means all, are arte- after the drowning sometimes obscures the old
facts of the depressed sea level during the last channel. The planimetric shapes of this type of
episode of the Pleistocene glaciation and the estuary are often very complex because tributary
resurgence of the sea level during postglacial valleys were also often flooded. Depths at the
times. During the long glacial period, many mouth vary widely, in some instances reaching in
streams carved deep valleys or gorges in the mar- excess of 150 metres. Much shallower ones are
gins of the continents and formed deltas at their very common, especially in those estuaries where
mouths at the margins of a sea that was perhaps as sedimentation has been rapid.
much as 160 metres lower than the modern level. Barred estuaries (Figure 20.1), which are often
The rising sea that accompanied the melting of referred to as bar-built estuaries, or depositional
the glaciers flooded those ancient deltas as well barrier estuaries, can form on any gently shelving
as the glacial-era valleys and gorges upstream, coast that is subjected to even modest wave attack.
producing the countless estuaries that notch the In such cases, wave action mobilises sediment
coasts of all of the continents. (usually sand) from the sea floor and constructs a
GFE-20.qxd 6/25/05 2:03 PM Page 537
ESTUARIES AND DELTAS 537
o
88 Alabama River
Te
.
o bile R
nsas River
M
MOBILE
N
o
rt
h
LONDON Se
M obile
a 30.5
o o
30.5
Thames Estuary
E n g l a n dDOVER
Ba
y
o
51
o
0 1o
1: Unmodified drowned river valley estuary Gulf of Mexico

o
88
166
o o
2: Barred estuary
167
Doubtful
Sea
N
Sound 123o Sacramento
River
San Pablo
Bay
Zealand
Breaksea
Ta s m a n
Sound 38 o 38 o
Sa
n Joa
Pacif
West Cape San Francisco

qu
Bay
ic
New
in
Co
Oce
R.
o o
46 46
yo
te
an
o o
166 167
Cr
123o
.
3: Fjord estuaries
4: Tectonic estuary
o
90 d
Lake oun
Pontchartrain ippi S
iss N
o iss o
30 M 30
NEW Cho
ORLEANS Albemarle
wa
M
nd
Sound
n R.
si
is
u
ss
So
36 36
ip n
pi to
R
iv B re Ro
anoke R.
er
Tidal
TarSub-estuaries Inlets
R.
Ne u
Mex i c o se
R.
Pamlico
of
Gulf 90
o Sound
5: Riverine estuary 6: Compound barred estuary

Figure 20.1 Types of estuaries.
GFE-20.qxd 6/25/05 2:03 PM Page 538
barrier beach island across the mouth of the so they typically have substantial streams dis-
stream valley, so that the estuary communicates charging into them at or near their head. Fiards,
with the sea only through one or more tidal inlets however, are glacially scoured depressions
which penetrate the barrier. However, a variant which are not related to pre-glacial stream val-
form can also develop in drowned valleys whose leys. Most of them do not have steep sides and
mouth is exposed to along-shore currents strong they are normally much shorter, broader and
enough to mobilise and transport sediments, creat- shallower than fjords.
ing a bay-mouth bar. Such estuaries also commu- Tectonic estuaries (Figure 20.1) are created
nicate with the open sea through tidal inlets. Not when diastrophism drops a segment of the Earth’s
uncommonly, the barrier beach islands form so far crust below sea level in a position such that the
offshore that several river valleys are blocked off sea can flood the depression. Down-dropped fault
by a single extensive barrier, so that in effect the blocks (grabens) and fold valleys are the most
mouth of the individual estuaries open, not into the common causes, although volcanic activity and
open sea, but into a lagoon or sound, the outer general crustal subsidence are also responsible for
limit of which is the enclosing barrier beach some estuaries. Because tectonic forces do not
island. This often has profound effects on the follow the rules of normal weathering and erosion
dynamics of the component estuaries, because processes, there are no preferred shapes or fea-
such marine influences as salinity, wave action, tures, with the possible exception being that the
tidal variations and along-shore currents may be shorelines of those formed in grabens and struc-
strongly moderated in the lagoon behind the bar- tural troughs are often relatively simple and regu-
rier islands. Indeed, it is not unknown for the tidal lar. The same cannot be said for those formed by
inlets to seal over completely from time to time volcanic action or general subsidence.
and for extended periods, during which intervals Although the classic estuary is formed when
the river estuaries are discharging into a brackish, the sea enters a pre-existing depression, some
tideless lake instead of the open sea. Such estuar- investigators recognise another way in which
ies, having formed on gently shelving coasts, are estuaries may form. As rivers create deltas which
usually very shallow (often less than 10 metres), extend into the open sea (e.g. through the over-
one consequence being the absence of high- filling of the original flooded valley), they flow
energy waves at their mouths. Nearly the entire on their own sedimentary deposits seaward of the
Atlantic coast of North America south of the mouth of the original valley. Such river channels
mouth of the Hudson River is characterised by are called riverine (Figure 20.1), or deltaic dis-
barred estuaries, as is much of the coast of the tributary estuaries. The classic examples are, of
Gulf of Mexico. course, the Nile and the Mississippi, although
Fjords (Figure 20.1) typically have U-shaped there are many others.
cross-sections resulting from glacial scour prior Most estuaries are characterised by a single
to glacial melting and subsequent flooding by large freshwater stream entering an arm of the
the sea. They are typically steep-sided, narrow in sea, but there are many examples of situations in
relation to length, and straight or slightly sinu- which several streams of approximately equal size
ous. They are often very deep, up to as much as enter a lagoon (or sound) which has been created
800 metres, but depth is usually not continuous. by the development of an offshore barrier bar. In
The bottoms of many, and perhaps most, are such cases, the ‘mouth’ of the estuary is the tidal
characterised by rock sills (transverse rock bars) inlet that pierces the barrier bar. However, each of
which often rise nearly to the surface, especially the streams entering the lagoon is itself an inde-
at or near their mouths. The bottoms of pendent sub-estuary which may behave quite
many fjords consist of linear basins strung end differently than its neighbours because of local
to end and separated from each other by rock conditions. Albemarle Sound, in North Carolina,
sills which may nearly reach the surface. Fjords is of this kind. It is also possible for a single
are the glacially scoured courses of rivers, complex river system to be have been flooded in
GFE-20.qxd 6/25/05 2:03 PM Page 539
such a way that several major tributaries coalesce A = Transverse components (Cross-section)
just before passing through a gorge or narrow val- Mud flats
Salt marshes
ley into the open sea. An example of this variety
High Tide
is Chesapeake Bay, in which the bay mouth com-
Low Tide
prises the mouth of the primary estuary, and each
major tributary, such as the James, York, Potomac,
Subtidal area
and so on is a sub-estuary. Such systems may be Intertidal area
called compound estuaries (not illustrated). Supratidal area
B = Longitudinal components (Planimetric view)

Estuarine classifications Transition zones
(Shifts with tide, season, river flow)
Given the extraordinary complexity of the estuar-
Marine
ine system, it is perhaps to be expected that clas- or Fluvial
effects
Lower Middle or
sifications of estuaries are also highly varied. Upper
However, most classifications reflect, to some
Limit of tidal
Estuarine Basin
Mouth
degree, the fact that an estuary is a semi-confined Mixing zone
Fluvial zone
(Mostly fresh water)
body of water in which sea water is mixed Marine dominance zone
(Mostly brackish water)
(Mostly sea water)
with salt water. The potential variations in size
and shape are almost limitless in number (as Figure 20.2 Estuarine components (modified from
described in Basic morphology, above). Some are Fairbridge, 1980).
affected by enormous tidal ranges, while others
have little or none. Some are conditioned by great
ranges of meteorological conditions, while others in part because the mixing processes are so
rest in regions having climates that vary little complex as to defy ready description.
from day to day or season to season. Some are in Every individual estuary displays marked tem-
the tropics, and some are in the Arctic. Some have poral changes in the position and extent of the
abundant biotas, while others are impoverished. transitions between the longitudinal categories.
All have been used at one time or another to Further, depending on the criteria being applied
classify estuaries. and the methods of measurement or detection
Despite the complexities, there is more or less being used, it is often difficult for two independent
general agreement that useful characterisations investigators to agree on the location of the transi-
can be achieved by describing transverse and tion zones at any arbitrary time. It is also worthy
longitudinal components (Figure 20.2). In this of note that not all longitudinal components are
scheme the transverse components describe the necessarily present at all times. For example, a
configuration of the estuarine basin, not the water, river flood period may force both the middle and
while the longitudinal components describe, lower components seaward past the mouth of the
somewhat indirectly, the salinity of the water. estuary. Conversely, a desert estuary with an inter-
Thus, the generally recognised transverse compon- mittent stream may permit the entire estuary to be
ents are supratidal (the part of the basin above tide ocean dominated for an extended period. Further,
level), intertidal (the part of the basin alternately an estuary at the margin of a tideless sea, such as
flooded and drained during the tidal cycle) the Mediterranean, will have no intertidal zone.
and subtidal (continuously below water level).
Similarly, the commonly recognised longitudinal Estuarine dynamics
zones are upper (freshwater dominated, or river The common feature of all estuaries on marine
dominated), middle (the primary zone of mixing littorals is the interaction of sea and river. It has
of saline and fresh waters) and lower (saltwater been said that sea and river are in an eternal
dominated, or ocean dominated). Profound dis- conflict that neither can wholly win. Of all the
agreements arise when attempts are made to factors that are germane to the struggle, the most
define the longitudinal components quantitatively, important are relative water densities, tidal action,
GFE-20.qxd 6/25/05 2:03 PM Page 540
river flow cycles, the size and three-dimensional Limit of tidal effects
SEA RIVER
shape of the estuarine basin, and wave climate,
because those factors collectively control both
the mixing of fresh and salt water as well as
stuary
the erosive and depositional processes. All these Bed of e
factors — and more — are linked in exquisitely Microtidal conditions
complex interactions within the estuary. SEA RIVER
While there is no universal system for classi-
fying the characteristics of the tidal cycle, three
stuary
categories are often used to describe amplitude: Bed of e
microtidal (0–2 m), mesotidal (2–4 m), and Mesotidal conditions
macrotidal (⬎ 4 m), and the temporal characteris- SEA RIVER
tics are classified as: diurnal (one high and one
low water per day), semidiurnal (two high and two
stuary
low waters per day with small diurnal inequality), Bed of e
and mixed (two high and two low waters per day Macrotidal conditions
with one high water very significantly higher than Water salinity
Currents
the other). (length of arrow indicates
Fresh Brackish Salt relative velocity)
The beds of streams (especially large ones)
approaching an estuary are very commonly below
Figure 20.3 Fresh water/salt water mixing modes in
sea level for significant distances upstream. As a
estuaries (modifed from Biggs, 1982).
result of the differences in specific gravity (the
ratio of the specific gravity of fresh water to sea
water is approximately 1.0/1.03), a salt wedge brackish-water layer also moves seaward. To
often forms because the fresh water tends to over- replenish the water lost in the mixing process,
ride the salt water, with the result that within the seawater flows up the wedge, so that the result is
inflowing stream, the water is layered, or strati- an upstream bottom current of salt water, and a
fied, with fresh water on top flowing seaward and downstream primary current of fresh water.
the salt water at the bottom, producing a vertically Clearly any sediment being transported down-
stratified estuary (Figure 20.3). This effect per- stream by the freshwater flow, either in saltation
sists even when the fresh water is carrying signif- or suspended as bedload, will strike the point at
icant amounts of materials (chiefly silt and clay) which the currents reverse direction, come out of
in suspension. suspension and begin the construction of a bar
In principle such a salt wedge can extend as far transverse to the bed of the stream. Eventually the
upstream as the point at which the bed of the river rise in elevation of the stream bed generates a
rises to an elevation above sea level. The effect is ramp which directs the freshwater flow upward,
most pronounced in estuaries subject to relatively which creates significant amounts of turbulence,
modest tidal oscillations (i.e. microtidal situ- which in turn enhances the mixing of the fresh
ations). For example, the salt wedge in the and salt waters. The consequence of this seems to
Mississippi River may extend 150 km upstream. be that the head of the saltwater wedge is driven
In practice such an extreme rarely occurs because downstream at the same time that the zone of
the rush of fresh water tends to push the end of the mixing is greatly enlarged.
saltwater wedge towards the sea. As tidal oscillations increase, salt wedge
At the contact between the fresh and salt water formation is reduced because the increasingly
(Figure 20.3), turbulence develops because of rapid tidal currents increase turbulence, which
the differences in velocity, and as a result a layer enhances vertical mixing. Estuaries exposed to
of brackish water develops between the salt- mesotidal ranges are often characterised as
water wedge and the freshwater outflow, and the partially mixed estuaries (Figure 20.3) because
GFE-20.qxd 6/25/05 2:03 PM Page 541
there tends to be a relatively smooth vertical For example, estuaries with wide, unobstructed,
salinity gradient, saltiest at the base (but not as deep mouths, but gradually shoaling and con-
salty as sea water) and brackish to nearly fresh at stricting channels, like the Severn Estuary, often
the top. In these estuaries, the transition zone create a strongly asymmetrical tidal surge. This is
between the upper and middle estuary is both because the gradually narrowing and shoaling
indistinct and highly mobile, moving upstream at channel allows the crest of the tidal wave to travel
low tide and downstream at high. Estuaries with faster than the more slowly moving water in the
macrotidal ranges may have no perceptible salt trough of the tidal wave. This produces a
wedge formation, because the very rapid tidal markedly asymmetric tidal wave, with a steep ris-
currents thoroughly mix the entire water column, ing limb of short duration, while the ebb duration
thus producing a vertically homogeneous estuary increases. In extreme cases this combination of
(Figure 20.3). The only detectable salinity gradi- conditions may produce a bore, a breaking wave
ent in this type of estuary is longitudinal, and the marking the front of the tidal surge as it moves up
transition zones between fresh and salt water are the channel. In such cases the rising tide is a
so indistinct as to be undefinable. period of high current velocities and extreme
turbulence, while the long ebb-tide period is
Adjustments between processes accompanied by relatively modest turbulence and
and forms current velocities. Such estuaries are almost
‘Neither the channel morphology nor the tidal invariably characterised by a cross-section such
properties . . . can be explained solely in terms of that the average water depth is greatest at high tide.
the other, though the two are mutually interde- When such flood dominant estuaries are asso-
pendent.’ (Wright et al., 1973). ciated with a net input of sediment from the sea,
as they often are, the estuary becomes a sediment
Because estuaries are links between a river and sink with rapid deposition rates on the intertidal
the sea, the reversing flows of fresh and salt water areas. Thus, in such estuaries, the elevations of
interact in tidal estuaries to produce unique depos- intertidal areas gradually increase.
itional environments. The three-dimensional At the other extreme are those estuaries in
configuration of the estuary is of critical impor- which the cross-section is such that the mean
tance as it may profoundly modify the shape of water depth is greater at low tide than at high tide.
the tidal surge as it enters the relatively shallow This apparent contradiction results because the
and confined estuary basin. In the open ocean the high tide cross-section is dominated by shallow
tide may be thought of as two waves, one sub- water over intertidal flats, whilst the low tide
lunar and the other anti-lunar, which travel around cross-section is confined to the deeper subtidal
the earth in approximately 24.8 hours. Thus, at the channel. In this condition, the asymmetry of the
equator, the wavelength (the distance from crest tidal wave is reversed, with the passage of the tidal
to crest) is approximately 20 000 km, and a tidal wave crest being slower than the trough (Dronkers,
limb (the distance from crest to trough) is about 1986). In such ebb dominant estuaries, the inter-
10 000 km, because the waves are essentially sym- tidal areas are often actively eroded by the high
metrical. current velocities of the ebbing tide, and the net
However, as the tidal wave approaches the movement of sediment tends to be out to sea.
shore of an island or continent, the front of the It is at least conceptually possible for a single
wave (the rising limb) commonly steepens as estuary to alternate between flood dominant and
the water becomes more shallow, and accordingly ebb dominant over long periods of time. However,
at the shoreline the tide often requires less time to many estuaries are characterised by a long-term
go from trough to crest than from crest to trough: balance between tidal processes and form. That is,
that is, the tidal wave has become asymmetrical. deposition within the estuary during the period of
Complex configurations of the shoreline and rising tide is more or less precisely balanced by
ocean floor may magnify this effect enormously. erosion and sediment removal during the period
GFE-20.qxd 6/25/05 2:03 PM Page 542
of ebb tide. In other words, these are neutral estu- area alone is not a sufficient measure, because
aries, in which sedimentation rate is equal to (all other conditions being equal) a broad and
flushing rate. Pethick (1996) hypothesised that the shallow cross-sectional shape will not transmit as
system works somewhat as follows: much water as a narrower but deeper channel.
Estuary depth and tidal wave length: in the The wetted perimeter, defined as the contact
open ocean the tidal wave propagates in response length (in a cross-section) between the water and
to extraterrestrial (especially lunar) gravitational the channel bed, is a useful shape parameter
attractions. However, as it approaches the shore, because maximum hydraulic efficiency occurs
other factors such as decreasing water depths and when it reaches a minimum for a given cross-
shoreline configuration, decouple the tidal wave sectional area.
from its lunar influence. After the wave enters the The shape and size adjustment of the mouth
mouth of an estuary, its velocity slows in propor- occurs because the current velocity through
tion to the shallowing of the estuary from mouth the opening will be a function of the tidal prism
to head. This reduction in velocity is accompanied (the total amount of water which flows through the
by deposition which, in turn, reduces the depth of opening) and the cross-sectional size and shape of
water. If, and when, the deposition rate balances the opening. Note that the tidal prism is also a
the flushing rate, the estuary can remain in a rela- function of the magnitude of the tide. If the
tively static status for a long period of time. The opening is too small, the current velocity may be
precise mechanisms of this process are not well high enough to mobilise sediments in the opening,
understood, but it appears that deposition on the thus changing the shape or enlarging the opening
intertidal areas eventually reaches elevations such until the maximum current velocities are no longer
that they are flooded only at spring tides (maxi- capable of mobilising the sediments. Of course the
mum high tide, which occurs once per lunar contrary is also true: if the opening is too large,
month), thus effectively reducing the area of the then the tidal currents flowing through it will
intertidal zone, and at the same time modifying not only fail to mobilise the sediments already
the cross-sectional shape such that rising tide and there, they will also deposit sediments (chiefly the
falling tide current regimes are approximately coarser fractions) which are already entrained
equal. within the tidal waters until the available current
Estuary mouth and tidal prism: in every estu- will pass through the opening at velocities capable
ary a relationship exists between the size and of carrying the pre-existing sedimentary load but
shape of the cross-section of the estuary mouth, incapable of mobilising new material from the bed.
and the volume of the tidal prism (defined as the However, it is important to note that any appar-
total volume of water injected into, and drained ent stability is dynamic. During the period of
from, the estuary during each tidal cycle). It rising tide the currents in most estuaries will be
should be noted that the tidal prism is never a con- flowing inland, and the size and shape of the open-
stant, because its value changes (in some ing will adjust to the maximum current velocities
instances very significantly) over the course of achieved during some (usually brief) part of the
each approximately 28-day lunar cycle, with a rising limb of the tide. Following that period, sedi-
minimum value at neap tide and a maximum mentation will again occur as the current through
value at spring tide. the mouth slows to a complete stop during high
If the estuary mouth is formed of unconsoli- tide, at which time the highest sedimentation rate
dated materials, any long-continued period of will occur within the mouth. When the tide begins
environmental stability will produce a cross- to ebb, gradually accelerating current velocities
sectional shape and area for the estuary’s mouth through the mouth will again begin to mobilise
that is adjusted to the tidal prism, although it will materials within the mouth and again reach a max-
exhibit small-scale variations produced by the imum at some point on the falling limb of the tidal
difference between neap-tide and spring-tide cycle. Thus, if the estuary is exposed to semi-
prisms. It should be noted that the cross-sectional diurnal tides, the cross-sectional area and shape of
GFE-20.qxd 6/25/05 2:03 PM Page 543
the mouth will actually go through four periods of Other types of estuary, notably those with
expansion and contraction during each 24-hour markedly constricted openings, such as the tidal
period. Added to the normal daily cycle is the inlets of most barred estuaries, obey very different
28-day lunar cycle of extreme high and extreme rules. Such restricted openings often produce sec-
low tides, which of course exacerbates the effect ondary bottom structures called channel-mouth
because the tidal prisms are respectively larger bars or tidal deltas resulting from the drop in cur-
and smaller than normal. This effect may be very rent velocity as the tidal flow passes through the
limited in many estuaries with very wide mouths, constricted opening and emerges into either the
but it is present all the same. open sea (during the ebbing limb of the tidal
Nevertheless, certain kinds of estuaries, cycle) or the interior of the estuarine basin (during
notably the unmodified drowned river valley type, the rising limb of the tidal cycle). Thus, while it
have apparently evolved relatively stable relation- seems counter-intuitive, channel-mouth bars com-
ships between the cross-sectional area of the monly form both inside ( flood tide deltas) and
mouth and the volume of the tidal prism. For outside (ebb tide deltas) the mouths of estuaries
example, thirteen estuaries along the east coast of with restricted openings. The mouth of St Johns
England demonstrate such a relationship even River estuary in north-east Florida is an example
though the estuaries are widely different in size of this condition, although in this case the interior
and configuration (Pethick, 1996) (Figure 20.4). bar is less well developed than the exterior bar
However, even in estuaries characterised by such because the current of St Johns River is strong
broadly similar environmental settings, variations enough to inhibit its full development.
occur. For example, two of the estuaries exhibit As a general principle, any substantial modifi-
mouth cross-sections larger than necessary to cation of the three-dimensional geometry of the
accommodate the tidal prism, suggesting that estuary may result in a cascade of secondary and
rapid accumulation of sediment within the estuary, tertiary effects. Geometric changes may be caused
coupled perhaps with extensive human reclama- by increased sedimentation, land reclamation of
tion of the intertidal flats over the last 500 years, intertidal areas, enhancement or degradation of
has significantly reduced the tidal prism, and tidal deltas, or dredging and channelling.
that this reduction has not yet been accompanied Sedimentation within an estuary results in a
by compensatory reduction in the cross-sectional gradual reduction of the tidal prism, which creates
area of the mouth. a complex chain of responses. As previously
noted, the reduced water flow through the mouth
(which may be thought of as an inlet jet) may
result in changes in cross-sectional shape and
reduction of cross-sectional area of the estuary
mouth, as well as modifications to the tidal deltas.
For example, if the estuarine system provides an
abundant supply of sediment that can be trans-
ported by the reduced velocity of the inlet jet,
the ebb tide delta may actually increase in size
because the reduced inlet jet velocity permits an
increase in sedimentation as it spreads and dissi-
pates in the sea. Alternatively, a well-developed
flood tide delta implies the existence of a high-
velocity inlet jet and a copious supply of sediment
derived from marine deposits outside the estuary.
Figure 20.4 Best-fit logarithmic relationship between If the process continues, the flood tide delta may
tidal prism and cross-sectional area of mouth for British grow so large that it significantly reduces the tidal
east coast estuaries (modified from Pethick, 1996). prism, which will in turn reduce the velocity of the
GFE-20.qxd 6/25/05 2:03 PM Page 544
inlet jet. The mouths of many estuaries are was followed by accelerated coastal erosion,
restricted by the growth of spits which are created culminating in the loss of much of Dunwich
usually on the seaward side of the inlet by combin- village (Gardner, 1754; Bacon and Bacon, 1988).
ations of strong along-shore currents, wave action
and wind. Because the restricted opening is inca- Estuarine sedimentation
pable of transmitting the necessary volume of The nature of the mixing processes has important
water to either fill or empty the estuary within a consequences for the sedimentology of the
tidal cycle, the water level on the upstream side estuary, especially if the suspended load in the
rises and creates a hydraulic ‘head’ which converts freshwater stream contains large amounts of clay.
the inlet to a short stretch of stream with a very The carrying capacity (whether in suspension,
high gradient and an accordingly high current bedload or saltation) of both salt and fresh water is
velocity, which in turn affects the morphology of dependent on current velocity. Because the current
both the ebb-tide and flood-tide deltas, as well as velocity of a freshwater stream entering an estuary
the cross-sectional size and shape of the inlet. is reduced as the estuary broadens toward the sea,
Land reclamation within an estuary can result the sedimentary deposits within the estuary tend to
in dramatic changes throughout the estuarine sys- exhibit a marked longitudinal grain-size gradient,
tem, including within the estuary, at the mouth, from large upstream to small toward the sea. This
and even along the adjacent coastline. The inter- tendency is especially marked in estuaries with lit-
actions are complex and not fully understood. tle tidal action. Increasingly large tidal oscillations
However, it is conjectured that the large changes progressively destroy the grain-size gradient, in
in the tidal prism often associated with reclama- part because tidal currents may be rapid enough to
tion of the intertidal areas may result in greatly mobilise and carry even large particle sizes.
reduced tidal current velocities through the Moreover, the depositional dynamics are also
mouth, as well as within the estuary itself, thus altered by the electrochemical action of salt
reducing the capacities of those currents to water, which produces flocculation (aggregation)
mobilise and transport sediments. The reduced of clay particles. Because the flocs have densities
sediment quantities delivered to the ebb-tide delta significantly higher than water, they come out of
thus results in such impoverishment that wave suspension as soon as the current velocities fall
action and along-shore currents are capable of below a critical value, and in consequence they
removing and rearranging the deltaic materials settle out in the seaward parts of the estuary as a
faster than they are replenished. This process may low-density clay depositional layer. All estuaries
result in the partial closure of the inlet by features experience this phenomenon, the intensity being a
such as spits, bars, and dune systems. The reduc- function of the proportion of clay minerals in sus-
tion in sediment transport out of the mouth of the pension in the freshwater stream and the degree to
estuary implicit in a reduced current velocity may which the water column is mixed. Because very
result in accelerated coastal erosion, because little of the fresh water is mixed with salt water in
there may no longer be a supply of sediments vertically stratified estuaries (usually those with
large enough to replenish the losses along the little tidal action), such estuaries may send plumes
nearby seashore due to wave and along-shore cur- of fine-grained sediment well beyond the mouth
rent action (Figure 20.5). For example, it has been of the estuary, while at the same time accumulat-
postulated that this sequence was a major factor ing nearly all of the coarse-grained sediments
in the events in the vicinity of the former port of within the estuary itself. Progressively more of
Dunwich, Suffolk (Pethick, personal communica- the clay-sized materials are deposited within the
tion). The concept is that reclamation in the Blyth estuary as tidal ranges increase, because the
and Dunwich estuary on the coast of England increasingly well-mixed fresh and salt water
north-east of the mouth of the Thames, triggered results in higher flocculation rates. Thus, a verti-
the southwards growth of a shingle spit across the cally homogeneous estuary can be a highly effi-
mouth and the collapse of the ebb tide delta. This cient sediment trap for all grain sizes. However, it
GFE-20.qxd 6/25/05 2:03 PM Page 545
Reclamation
Post-reclamation saltmarshes
Post-reclamation
sedimentation in
estuarine basin
Flood-tide
delta
Shoreline erosion
Post-reclamation
spits; partial
Sediment transport closure of mouth
into estuary of estuary
Ebb-tide delta
Present shoreline
(original position)
Sediment transport
(present position) into estuary
Original shoreline
Reduction of sediment
Land Reduction of Shrinkage of
replenishment to
reclamation current velocities ebb-tide delta
adjacent coastline
through estuary
Reduction of mouth Partial closure
Coast-line
tidal prism of estuary mouth
erosion
Sediment build-up Enlargement Along-shore

in estuarine basin of flood-tide sediment
delta transport
Figure 20.5 Concept for explaining the evolution of meso/macrotidal estuarine environments along the east coast
of Britain (modified from Carter, 1992).
must be emphasised that this tendency may be the rising and falling stages and slowest during
interrupted or even reversed if tidal currents are the periods of high and low water, the largest
very rapid, because the high-velocity currents grain sizes, as well as the largest amounts of sus-
flush the low-density clay-rich sediments accu- pended sediments, tend to be moved during the
mulated because of flocculation out to sea. rising and falling limbs of the cycle. Because the
Estuaries are found along most coastlines of the current velocities of flood tides are usually
world and are therefore subjected to the full spec- greater than those of ebb tides, tides sometimes
trum of tidal cycles and ranges. In general, the cumulatively move sediment upstream, a process
greater the tidal range, and to a somewhat lesser which produces an inverse-fill estuary.
extent the greater the tendency toward semidiurnal Tidal effects are greatly complicated by the
tides, the greater the amount of energy available size and three-dimensional shape of the estuary,
for mixing. Thus, estuaries exposed to semidiurnal because complex shapes tend to produce greater
macrotides tend to be the best mixed. turbulence and therefore enhanced mixing. Thus,
Tidal currents are often strong enough to act as estuaries with extensive intratidal zones tend to be
agents of erosion as well as transport and depos- better mixed than those in which intratidal zones
ition, and because the currents are fastest during are small or absent. In funnel-shaped estuaries
GFE-20.qxd 6/25/05 2:03 PM Page 546
(unmodified drowned river valleys) the tidal range McDowell and O’Conner (1977), the following
is frequently accentuated from mouth to head. In sources have been identified (Figure 20.6):
some estuaries the development of a bore greatly
enhances mixing. Tidal enhancement of mixing is 1. stream erosion (sediment carried by streams
normally minimal in microtidal estuaries, but such entering the estuary SR)
estuaries usually fill rapidly with coarse-grained 2. littoral drift (sediment carried by along-shore
sediment because tidal currents are inadequate to currents SL)
transport such materials to the mouth. As tidal 3. bank (derived from the banks of streams
ranges rise through mesotidal to macrotidal, ever within the estuary SB)
larger areas of intertidal zones are affected, and 4. shore erosion (derived from erosion of shores
tidal currents increasingly dominate circulation, of the estuary SS)
so that current pulses (generated by rising and 5. erosion of near-shore ocean bed (SO)
falling periods of the tide) transport pulses of 6. aeolian sources (wind erosion of coastal dunes
sediment through the estuarine channels and out and intertidal shoals SW)
to sea. Because the high-tide period is a time of 7. biological sources (decomposition and excre-
near-zero current velocities, fine-grained sedi- tions of aquatic biota SA)
ment is deposited in the intertidal zone, especially 8. human activities (derived from domestic and
on mud flats and marshes. Such tidal marshes and industrial effluents and solid wastes SP)
flats are very extensive in estuaries such as the 9. dredged spoil (derived from re-mobilisation of
barred estuaries of the Atlantic coast of the United dredged materials SD).
States. Tidal effects tend to be minimal in fjords,
which are usually very deep and steep-sided, The relative importance of these sediments varies
because the intertidal zones are very restricted or with the size and nature of the estuarine drainage
even absent. basin (the combined drainage basins of the inflow-
Barred estuaries constitute a special case with ing streams, plus that of the estuary itself), the
respect to tidal influence, because in those nature of the associated marine environments and
instances in which the lagoon behind the barrier the degree and nature of human development of
bar is large, the small tidal inlets may so restrict the estuarine drainage basin, as well as with the
the entrance of the tidal surge that the tide is estuarine dynamics discussed above. Indeed, espe-
scarcely felt. In effect, regardless of the amplitude cially in large estuaries, the sources of sediments
of the tides in the open sea beyond the barrier bar, may vary significantly within various parts of a
the estuary may become microtidal. single estuary. For example, in Chesapeake Bay
The flow cycles of the rivers which feed fresh (Figure 20.7), a very large compound estuary,
water into the estuaries are at least as variable and there is a large contrast in the relative contribu-
far more irregular and unpredictable than the tions of inorganic and organic materials between
tides. Streams may actually cease to flow in estu- the upper and middle parts of the estuarine system.
aries along arid and arctic coasts, in which case Typically the depositional facies gradually
the estuary temporarily becomes nothing more change from marine (usually sand, and occasionally
than an arm of the sea. On the other hand, the with marine-derived gravels) through estuarine
flows of some rivers are so powerful, especially (combinations of detrital and organically derived
during flood periods, that the mixing zone is materials) to fluvial (usually fine-grained clas-
pushed well out to sea beyond the mouth of the tics) from mouth to head (Figure 20.8). If tidal
estuary. A notable example is the Amazon. All flats exist, they are normally covered with silts
possible intermediate conditions may occur, and clay and often mixed with substantial
depending on the local conditions within any amounts of organic material. However, there is
given estuary. normally much overlapping and inter-fingering
Sediment entrained in estuarine waters come of sediment types because the upper, middle,
from a wide variety of sources. As modified from and lower estuarine zones shift up or downstream
GFE-20.qxd 6/25/05 2:03 PM Page 547
Seaward limit of estuary Limit of tidal effects Sources % of total Fate of materials
ESTUARY SR
INORGANIC
SA Rivers 61
Upper Bay
SP Shore erosion 13
SS suspended
SP ORGANIC
material
SO SD Rivers 22
SA SB
Shells, tests, etc. 2
Production deposition
2
SL within
transport Upper Bay
4% of total 3.7mm/yr
SW
Internal sources External sources

INORGANIC
SB ⫽Bank erosion, incl. tidal channels SO⫽Erosion of sea bed Circulation 5
Middle Bay
SS⫽ Shore erosion SW⫽Aeolian sources Shore erosion 52
SA⫽ Biological sources SA⫽Biological sources ORGANIC suspended
SD⫽Dredged spoil SL⫽Littoral drift material
SP⫽ Human activities SR⫽ Stream sediment Production 22
SP⫽Human activities Shells, tests, etc. 18
Circulation 3 deposition
Figure 20.6 Sources of sediment in an estuary within
transport to Middle Bay
(modified from McDowell and O’Connor, 1977). Lower Bay 1.1mm/yr
32% of total
with every change of tide or river flow. Because

Figure 20.7 Differences in sources of suspended
the number of combinations and permutations is sediments between upper and Middle Chesapeake Bay.
nearly infinite, the depositional characteristics of (modified from McLusky, 1981).
each estuary are unique and often surprising. For
example (Figure 20.9), the Tay estuary on the east
coast of Scotland is a long, narrow, macrotidal Mudflat surfaces are steepened by storm
estuary with extensive intertidal areas, complex surges, probably because sediment near the open
tidal channels and a mouth constricted in both water edges of the surface is mobilised by the
width and depth. Contrary to expectations, the extreme turbulence generated by the breaking
maximum suspended sediment load and mini- storm waves, and the sediment thus entrained
mum salinity occurs within the estuarine basin is deposited farther inland because water turbu-
(i.e. the deep basin within the constricted mouth) lence decreases as the wave moves inland and
at low tide. As the tide ebbs, fine sediment is loses energy. The original nearly horizontal sur-
entrained from the mudflats and tidal channels face may be restored during the long intervals
and concentrated in the estuarine basin, from between storms by erosion of the high parts and
which it cannot readily escape because of the con- re-deposition of the materials on the low parts.
stricted mouth and relatively high sill at the The mechanisms are not well understood, but
entrance. During flood tide, the suspended load is among those postulated is wave planation by
diluted by inflowing sea water and dispersed over wind-generated waves in the shallow water cover-
the mudflats, where some of it is deposited during ing the mudflat during periods of high tide.
the slack-water high-tide period. Thus, both erosive and depositional processes are
Mudflats (horizontal or near-horizontal inter- fundamental to their long-term stability.
tidal surfaces devoid of vegetation and composed Saltmarshes are much less subject to gradient
of fine-grained sediments) and saltmarshes (simi- changes in their surfaces, presumably because the
lar to mudflats, but covered with salt-tolerant surfaces are ‘armoured’ by the root-masses of
reeds and grasses) are formed by complex alter- the vegetation, but their surfaces nevertheless
nations between deposition and erosion of fine rise and fall in response to local conditions.
sediments (Pethick, 1992a; 1996). While many Investigations, primarily in small estuaries along
exceptions occur, there is a normal sequence from the east coast of England, have demonstrated that
open water to mudflat to saltmarsh within most a major storm event will significantly reduce the
estuarine intertidal zones. average elevation of a saltmarsh (Pethick, 1992b),
GFE-20.qxd 6/25/05 2:03 PM Page 548
SCALE
0 5 10 15 km
N
77˚ Origins of bottom sediments

37˚
Residual (wave- and current-eroded
bank and bottom materials)
Fluvial-estuarine sediments
Estuarine sediments
Marine sediments
Biogenic materials (shell beds, etc.)
No data
Figure 20.8 Distribution of depositional facies in James estuary (modified from Nichols, 1972).
but that it will recover and in some cases exceed over the neighbouring saltmarshes. In brief, high-
the original surface over a period of several years energy events cause saltmarsh erosion, and the
if the recovery is not interrupted by a second storm entrained sediments are transported to and
event. The reduction resulting from the storm is deposited upon the fronting mudflats, resulting in
apparently the product of the extreme water turbu- mudflat accretion (Figure 20.10), with the process
lence at the water–soil contact during the period reversed during intervals of calm.
of storm-wave activity. Despite the protection As with all sedimentation processes, mudflat
afforded by the vegetation, sediment particles are sedimentation is controlled primarily by two sets
mobilised and removed during the progress of the of factors: the concentration and grain sizes of the
storm. The surface of the saltmarsh begins to rise suspended sediment, and the velocity of the water
again as soon as the storm has passed, probably in contact with the mudflat surface. Grain size is
because the water after the storm is slightly deeper important because, all other things being equal,
than before. If entrained sediment is evenly dis- large particle sizes settle more rapidly than small
tributed through a water column, then the deeper sizes. Current velocity is important because, all
the water, the greater the amount of sediment other things being equal, the faster the current, the
available for deposition. greater the quantity of material that can be kept in
Curiously, while the surfaces of the salt- suspension and the larger the size of the particle
marshes decline during a storm event, the sur- that can be mobilised. However, it is not the mass
faces of mudflats actually rise, probably because velocity of the water (a measure of the total
the materials entrained by the storm waves are amount of water passing through a given cross-
re-deposited on the adjacent mudflats because the section per unit of time) that is important, but
water is deeper, and therefore less turbulent, than rather the velocities of the individual packets of
GFE-20.qxd 6/25/05 2:03 PM Page 549
Salinity Suspended load

(in % ) (in mg/l)
˚
14
5 – 40
40 – 100
100 – 150
Flood tide (high water) Flood tide (high water)
10
Water surface
24
14
8 Neap tide (low water) 17
Depth in metres
12
6 10
8 3 24
21
4
2
9
27 13 10 18 24
6
0700 0900 1100 1300 1500 1700 1900
Time
Figure 20.9 Variations in salinity and suspended load in the Tay estuary, Scotland, during one tidal cycle (after
Sholkovitz, 1979 as modified from McLusky, 1981).
water surrounding the individual sediment grains. turbulence. For the most part all of these factors
Thus laminar flow can mobilise particles of sizes reinforce each other in such a way that there is a
appropriate to the flow velocity, but transport is marked tendency for turbulence to increase with
confined to saltation (particles tumbling along the mass velocity.
channel floor), because the mobilised particles As a rapidly moving tidal current slows (and
settle out of the laminar flow as soon as they are becomes less turbulent) as it flows across the pro-
no longer in contact with the bottom. However, if gressively shallower water of a mudflat, first the
water having the same mass velocity is also tur- larger grains, followed by progressively smaller
bulent, the carrying capacity, as well as the parti- ones will settle out. Sometimes the rates are
cle mobilisation capability, will be a function of phenomenal, especially during storm events.
the mean velocities of the innumerable packets of Centimetres of sedimentation have been recorded
water in the turbulence vortices. A particle of in the course of a single tide cycle. However, the
appropriate size injected into such a flow may be average rate of sedimentation for most mudflats is
suspended indefinitely because it will be carried unlikely to be greater than about a centimetre per
along a path that is closely analogous to the ran- year. This value may reflect a balance between
dom Brownian motion of a particle suspended in erosion and deposition, both of which occur at
a gas. timescales reflecting both the local tide cycle and
True laminar flow is rare, because irregular- the local storm events. For example, most of the
ities in bottom configuration, the presence of sediment deposited on the mudflats during high
vegetation, inequalities due to density variations, tide may be remobilised on the succeeding ebb
wind at the surface, and many other factors create tide and carried toward the sea until the low-tide
GFE-20.qxd 6/25/05 2:03 PM Page 550
5Apr90(?)
25Jun89
8 Apr89
13Dec88
3
9Oct89
24Jan89
21Feb89
Elevation of
28Mar88
Surface height in centimetres
2 saltmarsh
9Jan90(?)
12Feb67
surface
1
⫺1
Elevation of
⫺2
mudflat surface
⫺3
0 100 200 300 400 500 600 700 800 900 1000 1200
Time in days
Figure 20.10 Changes in mudflat and saltmarsh surfaces over a 2.9-year period. Extreme storm event 20 December
1988 (modified from Pethick, 1992a).
slack water occurs, at which time much of it will and exposure to waves originating in the open sea.
be deposited. The currents of the next incoming Thus saltmarshes in mesotidal and macrotidal
tide may then remobilise much of it and carry it estuaries with relatively broad mouths, such as
back to the mudflats, with the process repeating those along the eastern shore of England, develop
endlessly. In stable or accreting estuaries, some of primarily on the landward edges of mudflats, as
the material is lost by long-term deposition on the described above. However, the very extensive
mudflats and saltmarshes, and some lost by being saltmarshes which commonly develop within
carried out to sea, with the loss being made up by microtidal estuaries with narrow mouths, espe-
new sediment added to the estuarine system by cially such barred estuaries as those along the
the inflowing freshwater stream, or from the sea south-eastern coast of the United States, develop
carried in through the mouth by strong high-tide simply by the gradual accumulation of sediments
currents. on previously shallow portions of the estuarine
As the mudflats grow in height and extent, they basin. Lacking both strong tidal currents and
eventually assume the form of a wide inter-tidal energetic wave action, sedimentation simply
‘beach’ which is capable of dissipating the wave raises the general floor until the process slows and
energy which is imposed on it. The reduction in stops as the sedimentary surface rises to a level
wave energy caused by these wide intertidal flats such that it is exposed to the atmosphere for a sig-
allows rapid sedimentation at their upper, shore- nificant length of each tidal cycle. Colonisation
ward extremity where only infrequent high energy by salt-tolerant vegetation usually occurs well
storm waves act to remobilise the deposited mater- before that stage is reached, so the saltmarshes
ial. As their elevation increases, their tidal inunda- originate and evolve without the prior develop-
tion period is lowered until they may be covered by ment of mudflats.
tidal water for only relatively short periods during Once a saltmarsh is established it acts as an
the 12.4-hour tidal cycle. Under these conditions, efficient sediment trap, because the vegetation
vegetation may develop on these upper mudflats, stems reduce the velocities of any currents flow-
thus forming embryo saltmarshes. ing across them. This process can be so efficient
The mechanisms of formation and evolution of that the surface can be raised to approximately the
saltmarshes is strongly dependent on tidal range same level as that of spring tides, at which point
GFE-20.qxd 6/25/05 2:03 PM Page 551
they are often colonised by salt-tolerant shrubs to produce ebb asymmetry (i.e. faster ebb-tide
and trees. This evolution of saltmarsh to tree than flood-tide currents), then there is a high prob-
island occurs chiefly in microtidal environments ability that active erosion of the intertidal zone
lacking frequent high-energy wave events. will occur, resulting in a net loss of sediment from
Conversely, especially in mesotidal and the estuarine basin, a slowly enlarging tidal prism,
macrotidal estuaries with wide mouths, the marsh and marked changes in the size and configuration
can act as a reservoir of sediment, as previously of the estuary mouth. A channel confined by
described. The extreme turbulences and high jetties may have similar effects. Without the
current velocities produced by storm waves are enhanced hydraulic efficiency of the new channel,
capable of mobilising sediment despite the sediment would be more likely to be contained
inhibiting effect of the vegetation. During such within the estuarine system, thus reducing the
events the sediments derived from the saltmarsh tidal prism, with all the secondary effects that can
surfaces are distributed over the adjacent mudflow therefrom.
flats as well as the estuary basin as a whole, thus On the other hand, if the deepened or confined
effecting an overall widening and flattening of channel produces flood asymmetry, there is a rea-
the intertidal zone. sonable probability that sediment will be brought
into the estuarine basin by the enhanced flood-
Estuaries and engineering tide currents, resulting in a long-term reduction
Estuaries are so complex and dynamic that it is of the tidal prism, followed by increased sedi-
impossible to generalise on engineering require- mentation within the estuarine basin. A classic
ments within them. Current velocities and direc- example is provided by a study of the Thames
tions change in unpredictable patterns because of estuary (Inglis and Allen, 1957), which suggested
chance combinations of river flows, tidal cycles that the removal of large volumes of dredged sed-
and meteorological conditions. Near random vari- iment to the outer estuary merely resulted in its
ations in water salinity create extraordinarily immediate transfer back into the estuarine basin.
harsh environments for many structural materials. This consequence was avoided by transporting
Foundation conditions may change both vertically dredged spoil entirely out of the estuary, with the
and laterally over astonishingly short distances as result that dredging (required to maintain an ade-
a result of the complex dynamics of the system. quate ship channel) was reduced by a factor of
Even more troubling is the fact that even rela- twelve.
tively small changes in the configuration of the Even such seemingly innocuous things as boat
estuary, such as those which might be created by wakes, if repeated over a long span of time,
the emplacement of piers or retaining walls, or can affect sediment deposition patterns, current
the dredging of new channels, may result in dra- dynamics and salinity variations. Upstream (even
matic large-scale changes in estuarine dynamics, outwith the estuary) engineering projects such as
with consequences that may not be entirely pre- river diversion, reclamation, dredging, navigation
dictable. Channel creation such as dredging and channel maintenance and dam construction may
jetty construction, can result in important modifi- result in profound changes in the behaviour of
cations to the natural regime. The hydraulic effect estuaries, because they induce changes in river
of deepening or otherwise constraining the flow flows and sediment loads. To summarise, an engi-
of a subtidal channel is dependent on the relative neering project within an estuary may require the
size of the channel to the tidal prism (if any), as services of a very broad spectrum of engineering
well as to the overall geometry of the estuarine specialisations.
basin. For example, if a deep longitudinal channel Estuaries also have a profound significance to
is dredged in the bed of a wide and shallow estu- conservation and wildlife management. It is clear
ary, one effect will be to enhance the velocities that any inhibition or enhancement of sediment
of both the flood tide and ebb tide currents movement within the intertidal zone will affect
within the newly created channel. If the result is the stability of both the mudflat and saltmarsh
GFE-20.qxd 6/25/05 2:03 PM Page 552
forms and, consequently, the habitats which they the region of the delta may have a significant
support. In this context, fixed flood defences may impact, the most obvious example being the cli-
prevent the free adjustment of the saltmarsh and matic and geologic conditions over the drainage
mudflat balance and, hence, may lead to the basin of the stream, because those factors
degradation of these features. control both the grain-size distribution, mineral-
The complex of subtidal, intertidal, and terres- ogy and the quantities of the sedimentary load
trial habitats characteristic of estuaries often carried by the stream. Because the number of
support unique indigenous communities of plants, combinations and permutations of variables is so
invertebrates, birds and animals, including fish, very large (Figure 20.11), deltas display wide
amphibians and reptiles. Mudflats and saltmarshes variations in all aspects of their composition and
are rich habitats for invertebrates and plants morphology.
which supply food and shelter for other animals Deltas occur in two very broad contexts.
(Davidson et al., 1991). Many estuaries are also of Estuarine, or bay-head deltas are those which
major national and international importance as form where a stream enters a drowned valley of
sites for rookeries, as well as for resting places for some type (see above). Continental shelf deltas
migrant and wintering waterfowl. In recognition are those which are constructing a sedimentary
of their importance to wildlife, many estuaries structure out onto the continental shelf beyond the
have been designated or otherwise identified mouths of bays, fjords, or other coastal re-entrants.
under a variety of national and international meas- While the dynamics of delta formation and
ures, including: Ramsar Sites, which are wetlands growth are very similar, the lateral restrictions
of international importance designated under imposed by a drowned valley, of whatever form,
the 1971 Ramsar Convention; Special Protection result in a very significant simplification of the
Areas (SPAs) designated under the EC Birds processes. Consequently, the following discussion
Directive (79/409/EEC Conservation of Wild will focus primarily on the phenomena associated
Birds); Special Areas of Conservation (SACs) with classic deltas, which are those that extend
designated under the EC Habitats Directive beyond the normal continental shoreline onto the
(92/43/EEC Conservation of Natural Habitats and continental shelf.
of Wild Fauna and Flora).
Delta components
The components of a delta system are easy to con-
20.3 Deltas ceptualise, but extraordinarily difficult to identify in
the field because the various parts grade smoothly
Deltas are accumulations of sediment, primarily into each other, without any sharp lines of demar-
of fluvial origin, which are created where a cation. The delta includes the sub-aerial deltaic
stream discharges into a receiving basin such as plain and a sub-marine sub-aqueous delta (Figure
an ocean, inland sea or lake. While virtually 20.12). The deltaic plain commonly consists of
identical expressions form at the margins of both three components: a) the upper deltaic plain, which
fresh and saltwater bodies, the following discus- is normally an older part of the deltaic structure,
sion will be restricted to those at the margins of inland from and topographically higher than the
salt water. Both the materials and the mor- remainder of the delta, as well as above the limit of
phology of deltaic systems are functions of the tidal influences; b) the lower deltaic plain, which
complex interactions of the hydraulic regime lies within the realm of river–marine interaction and
and sedimentary load of the river including: the normally extends seaward of the limit of tidal
waves, tides and coastal currents of the sea; the effects and c) the marginal deltaic plain, which is a
temperature, wind, and precipitation regimes of somewhat indeterminate zone that normally lies on
the local climate; the morphology and tectonic the flanks of the lower deltaic plain.
stability of the earth’s crust over the area occu- The upper deltaic plain may include both
pied by the delta. Indeed, conditions well beyond active and abandoned channels and is comprised
GFE-20.qxd 6/25/05 2:03 PM Page 553
River
River-dominated deltas
1. Mississippi River
2. Po River 1
3. Danube River
4. Ebro River Tide-dominated deltas
2 12. Fly River
Wave-dominated deltas 3 13. Ganges –
5. Senegal River Brahamaputra River
6. San Francisco River 14. Mekong River
7. Rhone River River-dominated
8. Nile River 4 15. Copper River
9. Burdekin River
10. Orinoco River 9
11. Niger River
7
10
Wave-dominated Tide-dominated
11 14
6 13
5 8 15 12
Wave Tide
1: Qualitative ternary diagram illustrating how deltas may be
categorized in terms of predominant processes operating at the
delta front (modified from Galloway, 1975 and Elliott, 1978b).
Climatic regime:
River regime:
a. Storm surges
1. Flow magnitude
b. Precipitation rates
2. Flow periodicity
c. Precipitation periodicities
3. Sediment load
d. Temperature regimes
Tidal regime:
Wave regime: 1. Amplitudes
1. Wave direction 2. Periodicities
2. Wave amplitudes 3. Current velocities
3. Wave priodicities
Tectonics:
Rates and magnitudes of
Marine current regime: subsidences and uplifts
1. Directions
2. Velocities Delta morphology
3. Periodicities and sedimentation
2: Qualitative illustration of the interactions among major factors affecting

delta morphology and sedimentation (modifed from Elliott, 1978a).
Figure 20.11 Illustrations of the complexities of interactions among enviromental factors affecting the morph-
ology and sedimentation of deltas on marine shorelines.
of sediments and landforms resulting from the environments are normally very complex as well
migrations or abandonments of both main and as gradational, and include braided channels,
distributary channels, overbank flooding and point-bars in meander belts, lakes, backswamps
periodic crevasses (breaks) in the natural levees and marshes.
that characteristically line both banks of both The lower deltaic plain is normally even more
main channel and distributaries. These processes complex than the upper plain, in part because it is
distribute river sediments into adjacent inter- directly affected by all of the vagaries of the asso-
distributary basins (topographic basins of ciated marine environment, as well as those of the
very low relief between the natural levees of river. Channels are usually numerous and form
active or abandoned channels). The depositional complex bifurcating and anastomosing patterns,
GFE-20.qxd 6/25/05 2:03 PM Page 554
it consists of two quite different depositional

environments. The delta front is characteristically
Floodplain a relatively steep slope dropping into deep water
Abandoned channel
Active channel
from the seaward margin of the sub-aerial delta.
Limit of tidal However, the ‘steep’ slope may exhibit an incli-
Upper deltaic inundation
plain nation of 0.5⬚, or even less, especially in front of
Marginal Marginal
Limits of
deltaic deltaic very large deltas such as that of the Mississippi
plain basin
sub-aerial
delta
River. In general, the sediments tend to be
Lower
coarse-grained near the mouths of the distribu-
deltaic
plain
taries, becoming finer with greater distance from
the shore. Beyond the relatively steep delta front
Sub-aqueous delta Active delta
(‘‘bird's-foot’’) slope, the bottom slopes flatten still further. This
D o mi n a n t a l o
pro-delta area, sometimes called bottomset beds,
n g - s h o r e c u rr e n t
Continental shelf Continental shelf receives only the very finest grain sizes, includ-
ing fine silt and flocculated clays.
Figure 20.12 Components of a delta (modified from
Finally, there is a highly specialised compon-
Coleman and Prior, 1982).
ent of the sub-aqueous delta known as the channel
mouth bars, which form as highly localised arcu-
with each channel paralleled by natural levees of ate sea-bed structures just offshore of the mouths
varying degrees of development. As a direct con- of all riverine channels reaching the sea, includ-
sequence of the natural levees, there are normally ing both the main channel as well as the distribu-
extensive inter-distributary basins, some of which taries. In general they consist of ridges of the
are flooded by the sea to form inter-distributary most coarse-grained sediments transported by the
bays, while others form marshes, swamps and streams.
lakes. Within these basins and bays there are com-
plex, intergrading and overlapping sedimentary Delta dynamics
deposits called ‘overbank splays’ and ‘crevasse Deltas will build seaward from stream mouths
splays’. only if two conditions are met. First, the supply
The marginal deltaic plains are features which of sediment delivered by the stream at its
form chiefly under the influence of the sea, and mouth must exceed the amount that is mobilised
thus consist primarily of extensive saltmarshes and removed by the waves and currents of the
cut by complex patterns of tidal channels, and offshore marine environment. Second, the sup-
interrupted by active or abandoned wave-created ply of sediment to the sub-aerial portion of the
beach ridges. In some places the beach ridges delta must more than compensate for any subsi-
form offshore of the marsh edge, creating semi- dence caused by crustal down-warping or com-
landlocked salt bays and lagoons between the paction of previously deposited deltaic sediments.
open sea and the saltmarshes. Occasionally the The process of progradation (compensating for
rapid growth of a delta lobe will almost com- subsidence by the progressive superimposition of
pletely cut off a large area adjacent to the lower younger sedimentary layers on older deposits) can
deltaic plain, thus creating extensive brackish- be extraordinarily complex. For example, cores
water lakes in areas that would otherwise develop from the modern Mississippi delta often reveal
the more characteristic saltmarsh. An example of complex sequences of strata, each characteristic of
this development is Lake Pontchartrain, adjacent a unique depositional environment, and often
to the Mississippi River on the Gulf Coast of the exhibiting shifts from a shoreline environment to a
United States. delta front, back to a shoreline and ending with a
The sub-aqueous delta is that portion of the modern marsh veneer (Figure 20.13). The continu-
delta that lies below sea level, but nevertheless ally shifting geometry of channels, natural levees,
actively receives fluvial sediments. In principle, inter-distributary bays, crevasse splays, and all of
GFE-20.qxd 6/25/05 2:03 PM Page 555
the other depositional and dynamic features so the original channel is abandoned. A new delta
characteristic of the deltaic environment results in lobe is then initiated at the point where the new
considerable three-dimensional variability in the main channel reaches the sea. The Mississippi
sequences (Figure 20.14). In fact, it may be said provides a text-book example of the process.
with some confidence that no two cores taken Over the past 7000 years the river has switched
50 metres apart will display precisely the same seven times, each delta remaining active only
sequence of deposition. about 1000 to 1500 years. The last switch,
The proclivity for rapid spatial variability which began the formation of the modern Balize
within a delta is even further exacerbated by delta, began only about 600 to 800 years ago
a process known as delta switching (Figure (Figure 20.15).
20.15). It is essentially a consequence of the With the abandonment of each delta lobe,
process of progradation. As the delta front regional subsidence, which in the Mississippi
moves seaward, the efficiency of the discharge delta is caused by a combination of crustal warp-
channels is reduced because their thalweg ing and sediment compaction, results in marine
(i.e the planimetric expression of the thread of transgressions. Thus, a coastal plain formed by
maximum current velocity) length is increased delta switching will exhibit an orderly repetition
and their gradient is decreased. This reduces of depositional events characterised by inter-
the current velocity in the channels, which fingering and overlapping deltaic sequences
increases the flood heights, which results in separated by shallow-water marine facies.
a more frequent inundation of the delta surface, Channel-mouth bars are formed where the
which results in higher sedimentation rates. confined current of the stream, whether main
Eventually the river is flowing down the crest of channel or distributary, slows as it spreads into
a low ridge which projects into the sea. The end the sea beyond the mouth. The rapid drop in cur-
result is that at some arbitrary point upstream, rent velocity results in rapid deposition of the
the gradient of a distributary which departs at an coarsest components of the transported sediments
angle from the main channel is steeper than the closest to the river mouth, with progressively
gradient of the main channel because the dis- finer sediments being deposited with increasing
tance to the sea is shorter, at which time the distance from the mouth. Theory predicts a
river switches to the distributary channel, and smoothly arcuate ridge concave to the mouth of
LITHOLOGY ENVIRONMENT STRATIGRAPHIC UNIT

MSL bay and marsh organic clay
⫺20 channel and LEGEND
channel mouth sands
⫺40 bar plant roots
⫺60 delta front laminated
silts and clays clay
⫺80 prodelta
silt and sand
Depth in metres
⫺100 clay unit I

⫺120 burrows
continental
⫺140 shelf shell horizon algae
clay unit II
⫺160
shell
⫺180 clay unit III
peat and wood
⫺200
algal reef zone fragments
⫺220
deltaic and strand plain sands
⫺240 shallow marine
Figure 20.13 Vertical sediment sequence in modern Mississippi River delta (modified from Coleman and
Prior, 1980).
GFE-20.qxd 6/25/05 2:03 PM Page 556
distributary distributary
inter-distributary
lake channel freshwater channel inter-distributary
swamp bay
subaqueous
limit of
crevasse splay
tidal marsh
inter-distributary bay
lay
marsh deposits
sp
natural jetties e (mostly organic
a ss
c re v clays and peat)
beach deposits
(mostly sands)
crevasse splay deposits

(mostly silts and sands)
pre-delta marine deposits (mostly clays) distributary mouth deposits (mostly sand)
shells
continental shelf deposits (mostly clays and fine silts)
biogenic deposits (mosly clays with shells, tests, other biogenic material)
prodelta deposits (mostly clays)
delta front deposits (mostly clays, silts, and fine sands)
inter-distributary bay deposits (mostly clays)
Figure 20.14 Isometric block diagram illustrating rapid vertical and lateral facies changes in deltaic deposits
(modified from Coleman and Prior, 1980).
the stream, but the actual form is greatly modified open bodies of water, bounded on the seaward
by such factors as wave climate, tidal range, and side by open sea and on the landward side by
along-shore currents. As the margin of the sub- saltmarsh penetrated by bifurcating and anasto-
aerial delta moves seaward, the river discharge mosing tidal channels and by reaches of natural
channels are extended, and in the process a chan- levee. While the water is usually saline, or nearly
nel is progressively scoured through the channel- so, brackish periods may occur, especially when
mouth bar, while at the same time new deposition the river is in flood, or when a crevasse in a
on the bar extends further toward the sea. The natural levee permits large amounts of river
forms remain, but the entire structure gradually water to enter the bay. Such breaks in the natural
moves seaward. levees initiate bay-filling episodes, because the
sediment-laden river water creates a fan-shaped
sub-delta known as a crevasse splay, which itself
20.4 Depositional environments develops a radial system of bifurcating channels.
On a large scale, entire delta systems may
Inter-distributary bays enlarge by sequences of bay-filling episodes
Inter-distributary bays, which form between the (Figure 20.16).
natural levees paralleling the main channel or the Typically, each crevasse splay evolves from
distributaries of the river, are normally shallow, an initial break in the natural levee during a flood
GFE-20.qxd 6/25/05 2:03 PM Page 557
o o o
91 90 89
Baton Rouge
Mississippi River hartrain
ontc
(present course) keP
La New Orleans
o
2
o 92
30 30
3
1
5
o
6 o
29 29
7
Mississippi River deltaic lobes
(in order of activity)
Gu N
lf o
1
2
Sale Cypremort
Cocodrie
5 Lafourche
6 Plaquemine
of Mexic
3 Teche 7 Balize (present) SCALE IN KILOMETERS
0 25 50 75 100
4 St. Bernard 90o
Figure 20.15 Delta switching in the Mississippi River delta (modified from Kolb and Van Lopik, 1966).
episode. If conditions are propitious, the cre- commonly a highly organic peaty clay or more
vasse enlarges during successive flood episodes, rarely, a thick peat, which presumably formed
and sediment deposition in the resultant splay while the area was a saltmarsh, prior to the forma-
proceeds rapidly until the distributary channels in tion of the crevasse splay. The organic layer is then
the splay offer no hydraulic advantage over the overlain by alternating silts and silty clays which
main channel. Thereafter water volumes and sedi- are extensively burrowed by marine organisms.
ment transport decrease, until at last the system These may be interpreted as pro-delta sediments
becomes inactive. The evolution of such crevasse deposited on the floor of the inter-distributary bay
splays may be nearly as complex as that of a pri- in advance of the sub-delta front. These beds are in
mary delta (Figure 20.17). If regional subsidence turn followed by layers of coarser silt and fine
is significant, the entire sub-delta may eventually sand which are only slightly disturbed by burrow-
be lowered below sea level, thus returning the ing marine organisms. As the sub-delta front
area to an approximation of the original configu- advances into the bay, sands and silts form graded
ration and condition. This may set the stage for beds and small-scale climbing ripples may be pre-
another episode of sub-delta formation in the served. The capping unit usually consists of either
same inter-distributary bay, as illustrated in clay which has been extensively burrowed and
Figure 20.16. reworked, or highly organic peaty clays. If the for-
Bay-filling episodes are usually readily iden- mer, the implication is that regional subsidence
tifiable in the stratigraphic record, because dropped the sequence below sea level, where it
they result in a coarsening upward sequence of remained without further sedimentation for a con-
deposits (Figure 20.18). The lowest member is siderable time; if the latter, the implication is that
GFE-20.qxd 6/25/05 2:03 PM Page 558
o Crevasse splays (in order of activity) o

89 30' 89
Conjectural complexes (identified by drill records)
A Dry Cypress Bayou
B Grand Liard
Historical complexes
C West Bay (1838)
D Cubite Gap (1862)
E Baptiste Collette (1874)
F Garden Island Bay (1891)
B
E
o
29 15'
D
NOTE: This map is

a synthesis of the F
shoreline from 1836
to the present. No such
shoreline ever actually
existed. The various
C
historical crevasse splays have Garden Island
been superimposed on a modern map. Bay
East Bay o
West Bay N 29
SCALE IN KILOMETERS
o 0 5 10 15 o
89 30' 89
Figure 20.16 Crevasse splays in the modern Mississippi River delta (modified from Coleman and Prior, 1980).
it remained at or near the surface, and thus because the sedimentary structure is partially
provided the base for a tidal saltmarsh. water supported, inter-distributary bay deposits
Engineering works in inter-distributary bay are universally in the process of compaction,
sediments and environments must be undertaken which results in continuous surface subsidence.
only with full awareness that such deposits con- The only variable is the rate of subsidence, which
sists of materials that have extremely high poros- may vary over a broad range, depending on local
ities and permeabilities, and that their structures circumstances.
are partially water supported. The combination
makes them extremely prone to shrinkage on Abandoned distributaries
draining or drying. This is especially true of Distributary channels are the conduits for sedi-
the peats and highly organic clays. In addition, ment transport within deltaic systems. They rarely
GFE-20.qxd 6/25/05 2:03 PM Page 559
o
89 25' M 89o15' o
89 25' 89o15'
is
sis
o o o
um
29 15' 29 15' 29 15'
sip
The J
pi
Riv
er
o
29 10' West Bay o
29 10'
N N
Head of Passes
SCALE, km SCALE, km
0 10 0 10
o
29 05'
1 La Tourette (1845) 2 Howell (1875) o
89 15'
o o
89 25' 89 15'
o o
29 15' o
29 15'
29 15'
o o
29 10' 29 10'
N N
SCALE, km SCALE, km
0 5 10 o
29 05' 0 5 o
29 05'
o o
3 USC&GS (1922) 89 15' 4 Aerial photo (1958) 89 15'
Figure 20.17 Development of the West Bay crevasse splay (bay-filling episode), 1845 to 1958 (modified from
Coleman and Prior, 1982).
exhibit point-bar or meander-belt deposits distributary may be so reduced by such obstruc-

because they do not usually migrate laterally. The tions as tree growth that the river blocks off the
reasons remain obscure. The reasons for the aban- head of the channel by the construction of a bar,
donment of such channels is equally uncertain, in a process analogous to the creation of ox-bow
and it is likely that local circumstances dictate lakes.
individual instances. The following reasons have Deprived of its river source, the water flow
been suggested for certain Mississippi delta dis- becomes very slow and may cease entirely during
tributaries: log jams blocking the entrances may certain seasons. Once abandoned, the channel
become so impervious that flow is prevented; the begins to fill with mostly locally derived sediments,
thalweg gradient of the distributary may become although significant amounts are usually derived
less than that of the main channel; changes in the from overbank flood flows in the main channel.
course of the main channel may isolate the head Because all sediment sources, including the over-
of the distributary; the channel efficiency of the bank flood flows, consist of fine-grained materials,
GFE-20.qxd 6/25/05 2:03 PM Page 560
Proportion of sand
Generalised to clay in sediments
stratigraphic (silt and other
column materials not included)
Peat and highly organic clays: present salt marsh.
Interbedded fine sands, silts, and clays, extensively

re-worked by marine organisms: basis for tidal salt
marsh.
Ripple-marks
sand
Silty fine sands, slightly disturbed by burrows of
marine organisms.
Silty clays with a component of fine sand, extensively

borrowed by marine organisms: pro-delta sediments.
Shells, tests, etc., of

marine organisms clay
Peat, or highly organic clays: surface of marsh in
Burrows of inter-distributary bay
marine organisms Silty clays with many burrows and much organic
material: bottom of inter-distributary bay.
Figure 20.18 Characteristic sediment sequence in a crevasse splay (bay-filling episode). Sample from West Bay
episode, Mississippi River delta (modified from Coleman and Prior, 1982).
the channels eventually fill with fine silts and clays. the edge of the sub-aerial delta advances seaward,
Without significant currents, the infilling process is the channel is defined by natural jetties which
hastened by water-tolerant trees and shrubs which are extensions of the natural levees which normally
colonise the shallow edges and act as highly effi- border the river channel as it crosses the sub-aerial
cient sediment traps, gradually constricting the delta. However, the function of the natural levee
channel to scarcely more than a thread. Significant changes subtly as the river builds out into open
amounts of wood and other vegetal debris are some- water. As long as the land on either side is above
times incorporated, and in some instances the final sea level, the natural levees serve to guide and
depositional layer is peat or highly organic clay. constrain the current only during high water
Thus, the abandoned distributary channels are stages, but as soon as the sedimentary surface
represented by channel plugs consisting almost recedes below sea level, the natural levee, now a
entirely of fine-grained silts and clays, often natural jetty, serves to constrain and direct the cur-
slightly coarser near the base, incorporating vege- rent during all water stages. The natural jetties are
tal material and often capped with peat or highly continued as sub-aqueous forms as the river cur-
organic clay (Figure 20.19). rent escapes from confinement and spreads out,
and therefore reduces in velocity. This mechanism
Channel-mouth bars is responsible for the ‘toes’ of the ‘bird’s-foot
Channel-mouth bars are an integral part of the deltas’ that are so prominent at the mouth of the
stratigraphic record of the sub-aqueous delta. As Mississippi and other deltas. The sands which
GFE-20.qxd 6/25/05 2:03 PM Page 561
Limits of active distributary

channel before abandonment
Inter-distributary bay
Residual channel
(often absent)
Peat or organic
clay capping
Depth in metres
16 Sandy silts, silts, silty clays, and clays,

24 of inter-distributary bay deposits and
0 1 2 3 4 5 6 delta front deposits
SCALE IN KILOMETRES
Fine sands and silty sands
deposited while distributary Silty clays and clays filling
was active channel of abandoned distributary
Figure 20.19 Isometric block diagram of deposition in channels of abandoned distributaries in the lower deltaic
plain of the Mississippi River (modified from Coleman and Prior, 1982).
comprise the dominant materials of the natural continually sweep the upper portions of the channel
jetties are deposited over, and generally become mouth bars. Bar deposits are usually capped by the
indistinguishable from the sub-aqueous channel- silts and silty clays of inter-distributary bay sedi-
mouth bars. Some of the sand comprising the bar ments, which normally include sand grains, shells
itself is usually mobilised and transported in all and shell fragments and organic debris.
directions away from the river mouth by wave Engineering problems associated with channel-
action, along-shore currents and tidal currents, the mouth bars are in most instances only those
extent and direction being a function of the local related to the properties of unconsolidated sands
conditions. As the channel mouth advances sea- and silty sands. However, local diapirism around
ward, the sandy bar and the natural jetty materials the channel mouths of the Mississippi River can
are emplaced over the silts and clays of the delta produce extensive and relatively rapid ground
slope (Figure 20.20). The result is a feather-edged motions, including both uplift and horizontal dis-
‘tongue’ of sandy sediments that gradually pro- placements which can rapidly destroy buildings,
grades, bounded by feather edges as it merges wharfs and pipelines. The cores of the diapirs are
with the more fine-grained inter-distributary bay the infamous mudlumps almost unique to the
sediments on either flank. Mississippi River.
The bar sediments typically show a coarsening
upward sequence, with pro-delta silty clays at Sub-aqueous slopes
the base. Sequentially above these are strata of Rapid sedimentation on the sub-aqueous delta
increasingly silty cross-laminated sand and silt slopes, including both the delta front and the pro-
layers with burrows and shell fragments, followed delta, results in soft silts and clays with low
still higher in the sequence by fine sand structures strengths. For reasons not fully understood, the
such as ripples caused by the oscillatory currents sediments are often rich in interstitial methane gas
generated by waves, and cross-laminae and cut- and can have extremely high pore fluid pressures
and-fill features resulting from the unidirectional (Figure 20.21). Under these conditions, even slopes
tidal and along-shore currents which more or less as low as 0.5⬚ are subject to a number of different
GFE-20.qxd 6/25/05 2:03 PM Page 562
Natural jetties Inter-distributary bay outer limit of distributary

bar deposits
Distributary channels
0 Silty sands and sands of

Depth in metres
distributary bar deposits

12
Silts and silty clays of inter-distributary
24
bay and delta slope deposits
36
Silty clays and clays of prodelta deposits
48
0 1 2 3 4 5 6
Distance in kilometres
Figure 20.20 Isometric block diagram illustrating characterisitic sediment sequence in distributary mouth bar
deposits (modified from Coleman and Prior, 1982).
kinds of slope instabilities which collectively move design, installation and maintenance of platforms
large quantities of sediment downslope into deeper and pipelines. The instabilities characteristic of
water (Figure 20.22). Sub-aqueous landslide sub-aqueous deltaic sediments pose particular
varieties include: shallow rotational sliding, which difficulties because they are especially prone to
occurs most commonly near the seaward edges of both vertical subsidence and lateral displacement.
channel mouth bars; collapse depressions resulting Many existing structures have been damaged, and
from the subsidence of inter-distributary bay some have been lost due to some combination of
sediments; elongated mudflows or mudslides these factors. Modern geological and geotech-
which may carry very large quantities of sediment nical capabilities, particularly the mathematical
downslope for very long distances, resulting in modelling of tide and wave interactions with
mudflow lobes or debris lobes that prograde across the sea bottom, can be used to provide useful
the continental shelf, or even into the offshore deep engineering solutions.
sea basin. The causes of the initiation of sediment
movement are not fully understood, but they are
known to include the complex interaction of sev- 20.5 Summary
eral processes (Figure 20.23), including excessive
pore pressures, which are enhanced by the genera- Estuaries and deltas comprise only a small per-
tion of methane gas within the sediments them- centage of the total area of the coastal zone of
selves. Surface waves, especially if amplified by the world. Yet, because of their attractiveness to
storm surges during great storms, create high- humans, they are among the most intensively
frequency cyclic loadings which produce instabil- utilised landscape on the planet. Part of the reason
ities. Typically delta-front sediments have such they are so desirable relates to the biological,
high water contents (often in excess of 50%) that hydrological, climatological, sedimentological
they are nearly always near the plastic limit, and and morphological dynamics that results from
frequently exceed the liquid limit. their position astride the boundary of sea and
The development of oil and gas resources in land. Many of the ways in which the estuaries and
sub-aqueous deltas has encountered problems of deltas have been used, or are proposed for use, are
GFE-20.qxd 6/25/05 2:03 PM Page 563
gas content
(% of pore volume
at 1 standard
composition shear strength, kN % water content atmosphere)
0 2 4 6 8 10 0 50 100 150 0 10 20 30
0
maximum miniature vane
residual
plastic limit
25 water
content
liquid limit
50
DEPTH IN METRES
75
100
125
LEGEND
clay with >50% water content fine sand
clay with <50% water content shells
Figure 20.21 Soil characteristics on the delta front, Mississippi River delta (modified from Bea and Audibert, 1980).
mutually incompatible, and methods of reconcil- Biggs, R. B. (1982) Estuaries. In Schwartz, M. L. (ed.)
ing such diversity in the extraordinarily complex The Encyclopedia of Beaches and Coastal
environments of the land–sea interface are likely Environments. Hutchinson Ross, Stroudsburg, Pa.,
393–397.
to provide unique challenges to the engineer for Carter, R. W. G. (1992) Coastal conservation. In
the foreseeable future. Barrett, M. G. (ed.) Coastal Zone Planning and
Management. Thomas Telford, 21–48.
Coleman, J. M. and Prior, D. B. (1980) Deltaic Sand
References Bodies. Am. Assn. Petrol. Geol. Continuing
Education Series, No.15, Tulsa, Oklahoma.
Bacon, J. and Bacon, S. (1988) Dunwich Suffolk. Coleman, J. M. and Prior, D. B. (1982) Deltaic envir-
Segment Publications, Colchester. onments of deposition. In Scholle, P. A. and
Bea, R. G. and Audibert, J. M. E. (1980) Offshore plat- Spearing, D. (eds) Sandstone Depositional
forms and pipelines in the Mississippi river delta. Environments. Am. Assn. Petrol. Geol., Memoir
Proc. Am. Soc. Civil Engrs. 106 (GT 8), 853–869. 31, 139–178.
GFE-20.qxd 6/25/05 2:03 PM Page 564
bar-front slumps mudflow scars

overlapping
(peripheral slumps) elongate mudflow nose and
collapse retrogressive slide (toe lobe) coalescing
depression mudflow gully mudflow toe lobes
clay diapirs
("mud lumps")
channel-mouth
bar sands
river
mouth
m.s.l.
Depth in metres
100
remolded sediments
pro-delta clays
shelf-edge
200 normal fault
Pleistocene silts and clays system
300 continental slope
mud diapir
delta slope prodelta
CONTINENTAL
SUB-AQUEOUS DELTA
SLOPE
Figure 20.22 Block diagram illustrating soil failure modes on the sub-aqueous delta of Mississippi River (modifed
from Coleman et al., 1998).
deltaic in situ geochemical surface wave Coleman, J. M., Roberts, H. and Stone, G. (1998)
sedimentation changes activity Mississippi River delta: An overview. Journal of
Coastal Research 14 (3), 698–716.
sedimentary methane gas cyclic Davidson, N. C., d’A. Laffoley, D., Doody, J. P., Way,
loading generation loading L. S., Gordon, J., Key, R., Drake, C. M., Pienkowski,
M. W., Mitchell, R. and Duff, K. L. (1991) Nature
excess pore gas bubble cyclic pore
water pressure water pressure Conservation and Estuaries in Great Britain. Nature
phase
generation Conservancy Council, Peterborough.
Dronkers, J. (1986) Tidal asymmetry and estuarine
decrease morphology. Netherlands Journal of Sea Research
in cohesion gas mobility
(metastability) 20, 117–131.
Elliott, T. (1978a) Deltas. In Reading, H. G. (ed.),
pore water effective Sedimentary Environments and Facies. Oxford
mobility stress reduction
University Press, 97–142.
strength
Elliott, T. (1978b) Clastic shorelines. In Reading, H. G.
stress
reduction increase (ed.), Sedimentary Environments and Facies.
Oxford University Press, 143–77.
initial sediment failure Fairbridge, R. W. (1980) The estuary: its definition and
slumping plowsliding geodynamic cycle. In Clausson, E. and Cato, I. (eds)
Chemistry and Biogeochemistry of Estuaries. Wiley,
Figure 20.23 Summary of factors contributing to fail- New York, 1–31.
ure of marine sediments on low slopes on the Mississippi Galloway, W. E. (1975) Process framework for describ-
River delta (modifed from Prior and Coleman, 1982). ing one morphologic and stratigraphic evolution of
the deltaic depositional system. In Broussard, M. L.
(ed.), Deltas, Models for Exploration. Houston
Geol. Soc., Houston, Texas, 87–98.
Gardner, T. (1754) Historical Notes on Dunwich.
Blythburgh and Southwold.
GFE-20.qxd 6/25/05 2:03 PM Page 565
Inglis, C. and Allen, F. (1957) The regimen of the Further reading

Thames estuary as affected by currents, salinities
and river flow, Proceedings of the Institution of Bowden, K. F. (1967) Circulation and diffusion. In
Civil Engineers 7, 827–868. Lauff, G. H. (ed.) Estuaries. American Association
Kolb, C. R. and Van Lopik, J. R. (1966) Depositional for Advancement of Science, Washington DC, 15–36.
environments of the Mississippi river deltaic plain, Coleman, J. M. (1988) Dynamic changes and processes
southeastern Louisiana. In Shirley, M. L. and in the Mississippi River delta. Geological Society of
Ragsdale, J. A. (eds) Deltas. Houston Geol. Soc., America Bulletin 100, 999–1015.
Houston, Texas, I 7–62. Dronkers, J. and van Leussen, W. (eds) Physical
McDowell, D. M. and O’Connor, B. A. (1977) Processes in Estuaries. Springer-Verlag, Berlin.
Hydraulic Behaviour of Estuaries. Wiley, New York. Dyer, K. R. (1986) Coastal and Estuarine Sediment
McLusky, D. S. (1981) The Estuarine Ecosystem. Dynamics. Wiley, Chichester.
Blackie, Glasgow and London; Wiley, New York. French, P. W. (1997) Coastal and Estuarine Management.
Nichols, M. M. (1972) Sediments of James River estu- Routledge, London.
ary. In Nelson, B. W. (ed.) Environmental Frame- Hobbie, J. E. (ed.) (2000) Estuarine Science: a Synthetic
work of Coastal Plain Estuaries. Geol. Soc. of Approach to Research and Practice. Island Press,
America, Inc., Boulder, Colorado, 169–212. Washington DC.
Pethick, J. (1992a) Salt marsh geomorphology. In Ippen, A. T. (1966) Estuary and Coastline
Allen, J. R .L. and Pye, K. (eds) Saltmarshes: Hydrodynamics. McGraw-Hill, New York.
Morphodynamics, Conservation and Engineering Lauff, G. H. (1967) Estuaries. American Association
Significance. Cambridge University Press, 41–62. for the Advancement of Science,Washington DC.
Pethick, J. (1992b) Natural change. In Barrett, M. G. Leeder, M. R. (1982) Sedimentology, Process and
(ed.) Coastal Zone Planning and Management. Product. George Allen & Unwin, London.
49–63. Thomas Telford, London. National Research Council (1983) Fundamental
Pethick, J. (1996) The geomorphology of mudflats. In Research on Estuaries. National Academy Press,
Nordstrorn, K. F. and Roman, C. T. (eds) Estuarine Washington DC.
Shores: Evolution, Environments and Human Nelson, B. W. (ed.) (1972) Environmental Framework
Alterations. Wiley. of Coastal Plain Estuaries. Geol. Soc. of America,
Prior, D. B. and Coleman, J. M. (1982) Active slides Boulder, Colorado.
and flows in underconsolidated marine sediments Perillo, G. M. E. (ed.) (1995) Geomorphology and
on the slopes of the Mississippi delta. In Saxov, S. Sedimentology of Estuaries. Elsevier, Amsterdam.
and Nieuwenhuis, P. K. (eds) Marine Slides and Shirley, M. L. and Ragsdale, J. A. (eds) (1966) Deltas.
other Mass Movements. Plenum, New York, 21–49. Houston Geol. Soc. Houston, Texas.
Sholkovitz, E. R. (1979) Chemical and physical Stanley, D. J. (1997) Mediterranean deltas: subsidence
processes controlling the chemical composition of as a major control of relative sea-level rise. Bulletin
suspended material in the River Tay estuary. de l’Institut Océanographique, Monaco, Numero
Estuarine and Coastal Marine Science 8, 523–545. special 18, 35–62.
Wright, L. D., Coleman, J. M. and Thom, B. G. (1973) Stone, G. W. and Donley, J. C. (eds) (1998) The World
Processes of channel development in a high- Deltas Conference. Journal of Coastal Research
tide range environment: Cambridge Gulf-Ord 3 (14), 695–915.
River Delta, W. Australia. Journal of Geology 81, van Westen, C. J. and Scheele, R. J. (1996) Planning
15–41. Estuaries. Plenum Press, New York.
GFE-21.qxd 6/27/05 4:37 PM Page 566
21. Coastal Environments

Julian Orford
21.1 Introduction spatially contiguous geomorphological compon-

ents) more than it does the terrestrial element that
This chapter is about how engineers can use geo- theoretically defines the landward zone of the CZ.
morphology in support of their activities in the The shoreline is the pivotal position with respect
coastal zone. There is a range of texts (e.g. Komar, to hazardous contact between the human and
1998; Horikawa, 1988) for engineers working on physical worlds, hence maximal engineering
coastal problems that identify empirical and theo- intervention at this position.
retical approaches to wave and tidal processes and The importance of the CZ in the future should
associated sediment transport. Considerable under- not be underestimated, as the physical challenges
standing of the geomorphological basis of coasts to societies adjusting to CZ changes accelerated
can also be found (e.g. Davies, 1972; Carter, 1988; by climate change, will occupy a central position
Trenhaile, 1997). There are texts that attempt to in future environmental agendas. It has been esti-
straddle both approaches (van Rijn, 1998), however mated that 37% of the world’s population lives
there is a scarcity of material that specifically iden- within 100 km of the coastline (Cohen et al.,
tifies the geomorphological approaches and diffi- 1997) and is growing at a rate greater than the
culties that engineers need to identify when dealing overall population rate. This fact alone serves to
with coastal problems in terms of causes as well as emphasise the strategic need to be able to deal
symptoms. This chapter outlines a framework in with the range and magnitude of the impacts of
which engineers might consider the nature and future human occupancy of the coastline.
solution of coastal problems with which they deal.
There is little space available to cover the variation
and complexity of the world’s coastlines, so this 21.2 Perspectives affecting our
chapter will emphasise the features of mid-latitude understanding of the
coasts where engineering intervention is at its most coastal zone
complex and advanced.
The seaward boundary of the coastal zone Human interventions in the coastal zone
(CZ) is given, for convenience, as the depth of the The perils posed by living next to a hazardous and
effective wave base related to extreme storms, unstable coastal zone have not prevented a tradi-
while the landward boundary is often well land- tion of positive intervention into the coastal zone
ward of coastal deposition, in that terrestrial by most coastal societies. The success of inter-
sources of sediment and water masses need to be vention has been constrained by the technical effi-
included in the CZ system. The importance and ciency of the moment. Lack of ability did not
role of energy and mass exchanges from terres- prevent past societies from taking risks in estab-
trial geomorphological systems into the CZ, as lishing a presence in hazardous positions with
well as exchanges between it and offshore envir- respect to the CZ. But prior to the Industrial
onments should be recognised as part of the Revolution, there was a more fatalistic acceptance
background to understanding the dynamics of the that nature could not be dominated, such that soci-
CZ. Engineering geomorphology of the CZ con- ety had to live (and die) with the impact of
cerns the actual shoreline (i.e. the beach and its extremes of coastal hazards. This attitude still has
576
GFE-21.qxd 6/27/05 4:37 PM Page 567
COASTAL ENVIRONMENTS 567
echoes in some contemporary societies that are of natural resources can be assured. This means
striving to achieve an economic transformation of that societies will have to work with the coast
their CZ. and not block or deflect coastal processes and
The nature of human intervention in the CZ responses. This idealistic approach is unlikely to
reflects the cultural imperatives of each society. happen overnight, if at all, given the history of
In western-orientated societies, technological human investment in the CZ. The pace of natural
advances in engineering design and material coastal change is, however, speeding up as glob-
behaviour over the last two centuries support the ally accelerated climate change is translated into
view that CZ vicissitudes could be contained. For rising sea levels and transgressive shifts of coastal
most of the twentieth century a persistent culture environments onshore. This means that there is a
of living by the seashore was not balanced by a rising risk associated with the shoreline as a point
recognition of the limits of technology, nor by an of hazardous interaction between natural and
understanding of how coastal systems were cultural activities.
changed by human intervention. However, experi-
ence over the last fifty years, whereby engineering Scenarios for future intervention in
intervention has generally proven to be unsus- the coastal zone
tainable per se, over-expensive for the benefits There are four scenarios that societies can explore
derived, and incapable of withstanding the in response to these shoreline pressures:
extremes of coastal hazards, has led to the recog-
nition that intervention carries a sustainability cost 1. Hold the line by trying to prevent all potential
to the coastal environment. Whilst it is likely that changes.
debates over the realities and costs of engineered 2. Do nothing and let society take a progressive
coastal living will continue into the future (e.g. the loss of function and value due to coastal
conflicts on the value of sea walls between Kraus, change.
1988 and Pilkey and Wright, 1988), most coastal 3. Accommodate the pressures by spatially mix-
engineers accept that intervention into the CZ is at ing holding the line and doing nothing in pro-
a price. Once started, engineering needs constant portion to the risk and utility of maintaining
support as it rarely creates a mutually balancing specific coastal functions.
cultural–physical system. As Doyle et al. (1985) 4. Undertake managed retreat by shoreline
commented in their ‘Laws of the Coast’, once you realignment.
start intervening in the coast you can never stop.
Whether this cost is acceptable, depends on the The last approach is to ‘give up’ marginal coastal
perspective and relative valuation of natural and areas by allowing them to form the accommoda-
human environments. tion space for coastal environments associated
with the present (and future?) shoreline, and by so
To protect or allow change? doing, allow an increasing revision to a natural
The scale of perceived damage to coastal systems and self-sustaining coastal system. Although this
by human intervention has generated tension. is an ideal state and in theory allows for the reten-
Some engineers believe that their mission is tion or expansion of natural habitats that have
foremost to protect the landward investment while come under severe pressure in the twentieth cen-
others would argue that such protection has been tury, there has to be a realistic sense of the likeli-
at the cost of the seaward environments. Rather hood of society allowing sufficient release of
than continue this debate, it is now time for both coastal space to keep pace with the landward shift
sides to reflect on the best ways to achieve a sus- of coastal environments under the predicted
tainable accommodation of human coastal usage. eustatic sea-level changes of the twenty-first cen-
It has become increasingly recognised that human tury (Houghton, 1997). This means that engineer-
coastal activity has to be conducted within self- ing will be found in the future, but that its mode
regulating coastal systems so that the sustainability is diversifying from outright resistance to subtle
GFE-21.qxd 6/27/05 4:37 PM Page 568
Increasing acceptability
Preferred policy
Hold the line
Management realignment
No active intervention
Limited intervention /
advance the line
0 10 20 30 40 50
Years
Figure 21.1 The most acceptable coastal management policy pathway for the next half century in the UK. Note
the switch from holding the line to managed realignment.
reinforcement of ongoing processes. Figure 21.1 created when society intrudes into highly mobile
shows how the acceptance of these modes of and dynamic natural coastal systems. In this
response varies over time, and that the selected sense, S1 reflects the need to mitigate, deflect or
priority method is likely to change due to chang- stop a natural tendency of the system that is per-
ing society and coastal circumstances. ceived to be causing human problems; S2 reflects
on controlling the consequences of intervening
Coastal engineering as a function of into an existing self-regulating natural coastal
geomorphology system that is likely to be thrown out of balance
The reasons for engineering in the CZ constrain by intervention. To work with these system per-
the type of geomorphological information that is spectives requires an understanding of the con-
required. In outline, engineers design coastal trols and especially the timescales by which the
structures to the needs of two main generic processes and responses are structured.
situations (S):
Time and space scales in the coastal zone
1. S1: to prevent loss of both coastal land and Geomorphic elements of the coastal zone respond
built infrastructure by erosion and coastal differentially to process timescales. Individual
flooding. sand grains (mm scale) move in response to water
2. S2: to support new developments in the CZ by current activity that is driven by periodicities of
reclaiming land from areas that are active geo- seconds to hours (waves and tides). Aggregate
morphological units of coastal environments. sand grain morphology (⬎10 m) responds more
slowly and may require seasonal, annual or
These two requirements can be reduced to annual plus event time frames before change is
geomorphological terms to identify the problems noticeable. In this way, time and space scales of
GFE-21.qxd 6/27/05 4:37 PM Page 569
Time
scale GEOMORPHOLOGICAL /
Macro-scale
GEOLOGICAL
millennium
inner continental
shelf
century ENGINEERING
lower shoreface
Meso-scale
transgressive dune
decade
tidal basin
EVENT
year inlet-rivermouth
upper shoreface
season frontal dune

Micro-scale
day surfzone bars

beachface
hour INSTANTANEOUS
ripple
second
0.1 1 10 100
Length scale (km)
Figure 21.2 Time and space scales controlling coastal response (after Cowell and Thom, 1994).
morphology are positively associated (Cowell and emphasis on the effects of accelerated climate-
Thom, 1994; Figure 21.2) such that geomorphic induced change has alerted coastal geomorpholo-
analysis needs to be set in the correct time frame. gists to potential timescales of coastal change that
Engineering responses are often critically short of had been previously ignored. In particular, the
observation time, yet estimates of long-term consideration of meso-scale processes (decades to
change need to be central to the study. Engineers centuries) (De Vriend et al., 1993) characteristic
should be aware of the potential for time-dependent of large spatial scales (km), have proven fruitful
change and variability of natural systems. in providing a framework for the analysis of
A principal element of most coastal problems coastal changes (e.g. beach profile volumes along
is insufficient longshore sediment supply (cf. situ- the central Dutch coast: Wijnberger and Terwindt,
ation 1 above), yet information on the natural 1995).
annual variation of the supply is rarely available, Coasts are dynamic, with extreme events of
while investigations into the longer-term (decade, one to two days’ duration affecting morphology
century or even millennium) sediment budget that may remain in evidence for decades to cen-
context are usually absent. It is a fallacy to con- turies. Many events of longer-term processes, e.g.
sider that all coastal situations are in a steady state sea-level change and sediment supply, define the
or dynamic equilibrium (Chapter 1) such that boundary conditions in which problems of a
instantaneous change rates are representative of smaller time span, for example storms and their
long-term change. Such situations are rare given products, are considered critical by coastal engin-
that a stationary process set has only a short-term eers. Coastal problems exercising engineers rarely
context (hours at best). The forcing of coastal involve just short-term or spatially isolated
processes is dependent on atmospheric and oceano- processes. An understanding of geomorpholog-
graphic conditions that are highly vulnerable ical constraints on coastal evolution is an essential
to non-stationary changes over time. The recent prerequisite for successful engineering in the CZ.
GFE-21.qxd 6/27/05 4:37 PM Page 570
Sensitivity and response in presents a problem for associated human activ-

the coastal zone ities. Low coastal resilience is usually related to
As a further refinement on process–response falling beach face sediment volumes due to fail-
timescales, engineers need to be aware of the sen- ing supply. Environments with low resilience may
sitivity issues of processes related to change show a major shift in the structure of the coast to
(Thornes and Brunsden, 1977). The magnitude of a new domain in which the coastal sediments are
the forcing process has an associated reaction self-organised into new stable configurations
time, which is the timescale before the morphol- (Forbes et al., 1995; Orford et al., 1996). Such
ogy being forced starts to change, and a relaxation reorganisations on low-lying coasts are often
time, over which the morphology still responds to associated with transgressive movements of geo-
the forcing event’s occurrence even after the for- morphic units. This can lead to severe impact
cing has ceased. problems when any potential accommodation
Extreme storms are likely candidates for for- spaces are already occupied by human activity —
cing coastal changes that have long meso-scale hence the issue of ‘coastal squeeze’, whereby
relaxation times (Orford et al., 1999a). Coastal coastal defences hold a stable line in the face of a
changes in relaxation phases are rarely linear in rising sea level and as a consequence the natural
response over time. This can lead to false perspec- on–offshore sequences of geomorphic environ-
tives on the rate and route of coastal change, if ments are differentially squeezed spatially.
such changes are measured over short timescales An added problem in understanding coastal
(⬍ 5 years). The sensitivity of geomorphic units sensitivity is the deflection of natural process–
can be characterised by the reaction times. response timescales by human intervention.
Saltmarshes and sand dunes are likely to respond Reaction and relaxation times to dis-equilibria
almost instantaneously to direct wave attack car- created by man’s interventions can range from the
ried by storm surges and would be defined as instantaneous to well beyond the design life of the
highly sensitive environments, while the same structure installed. In general, open coast changes
storm may have no impact on a hard rock cliff as a function of changes in the landward elements
(Pethick and Crooks, 2000). (e.g. estuary reclamation and wetland draining)
A further dimension of sensitivity can be seen take longer to appear (⬎ 100 years) than do the
in the concept of coastal vulnerability, which is changes incurred by intervention on the beach
the balance between coastal susceptibility and face per se. For example, a cessation of beach
coastal resilience. Whereas sensitivity reflects a sediment supply by sea-wall protection of eroding
temporal dimension to change, vulnerability car- cliffs may show an impact on beach volume
ries a value that any sensitive change is sudden within a year.
and detrimental in a human context.
1. Vulnerability is a function of coastal suscepti- 21.3 Process and response controls

bility and resilience. in the coastal zone
2. Susceptibility is an indication of the forcing
conditions to be experienced. Relative sea-level change
3. Resilience is the ability of coastal environ- Relative sea-level change (RSLC) concerns long-
ments to absorb forcing change and return to term variation (usually annual) in the vertical pos-
the existing state once the forcing relaxation ition of mean water level relative to the land. Such
phase is over. change refers to the apparent rise and fall of mean
sea level due to geodetic, sea volume (eustatic)
A vulnerable coastal environment is one that and land surface (isostatic/tectonic variation)
shows a low and diminishing resilience in the face changes. As these water and land movements are
of consistent forcing, often associated with independent, the final resolution of sea-level pos-
persistent coastal change, and as a consequence ition and any change per se, can only be relative,
GFE-21.qxd 6/27/05 4:37 PM Page 571
hence Relative SLC. Determination of annual be the cause of shoreline behaviour out of step
RSLC rate over decades provides a measure of the with the general decadal trend. The superimposi-
long-term shift in a statistical position of sea level tion of differing coastal responses at different
that is the effective datum for all periodic activity timescales is a problem in understanding coastal
at the shoreline. The movement of this position behaviour, especially when relaxation phases may
defines whether the coastal zone responds trans- overlap.
gressively (movement upwards) or regressively The importance of RSLC for engineers in
(downwards) over the long term. As this posture tectonically mobile coastal zones is limited,
will influence the long-term morphological con- though seismic disturbance of water elevation in
text of the CZ, it is essential to know RSLC before the form of tsunami is a sizable threat (Dawson,
attempting engineering intervention. As RSLC is a 1996). In tectonically stable areas, the rate of
statistical concept, its value is open to movement RSLC over a structure’s design life is important.
due to a variety of water level forcing influences at Southern England is currently experiencing a
radically different time periods: seconds (waves); RSLC of ⫹1.5 mm/year, which increases the risk
hours to years (tides); seasonal to inter-annual of estuary flooding and structures overtopping
(oceanic circulation, e.g. up-welling and seasonal during surges. Barkham et al. (1992) estimated
oceanic density changes due to temperature varia- the change in return period of the present 100-year
tion). There is a general consensus that accelerat- extreme water elevation (storm plus surge) for
ing atmospheric–oceanic heat exchange accounts ports around the UK, given projected rises in
for 50% of eustatic change in the late twentieth RSL by AD 2040 of c. 50 cm. The return period
century, the rest relating to increased release of ice falls to 1 in 45 (Glasgow) as one of the least,
melt (Raper et al., 2000). RSLC is statistically de- while the potential reductions in return period of
trended to remove the periodic effects of tidal 1 in 3.5 for Milford Haven and less than 1 in 10
activity, but inter-annual variation related to for other southern British ports indicates the
atmospheric and oceanographic conditions can problems that lie ahead for assets in areas at risk
persist (Figure 21.3), which in the short term may from flooding.
0.20
MEAN SEA LEVEL: NEWLYN, UK, 1916–1992
y = 0.0016x ⫺ 3.1229
0.15 ⭸ annual MSL rate ⫽ 1.6mm/yr
R2 ⫽ 0.9819
Elevation (m OD)
0.10
0.05
Annual MSL
0.00
De-trended Nodal MSL
Predicted Nodal MSL
⫺0.05
1910 1920 1930 1940 1950 1960 1970 1980 1990 2000
Year
Figure 21.3 Inter-annual to decadal scales of variation in residuals of annual mean sea level from the empirical
secular trend in sea level based on the nodal tidal cycle (18.6 years). Clustering of residuals relates to other oceano-
graphic forcing factors. Annual MSL data (1916–92) are from the Newlyn (UK) tide gauge.
GFE-21.qxd 6/27/05 4:37 PM Page 572
A rising sea level is a passive factor per se; its domain changes (Orford et al., 1996) that could
importance is that it transmits onshore energetic spell dramatic instability for the CZ of the twenty-
geomorphological processes at positions shore- first century. The potential for increased sediment
wards of a minimum lateral position due to inun- in the intertidal zone with accelerated RSLC pre-
dation alone. Bruun (1962; 1983) has attempted dicted for the next century, is dependent on the
to relate this horizontal shoreline displacement to degree to which new terrestrial sources are
RSLC. He empirically related coastal retreat to already protected from this future erosion.
sea-level rise, weighted by the depth offshore at
which sediment transport under prevailing local Tidal activity
conditions starts (Figure 21.4). Bruun’s Rule is a Tidal action concerns the effects of diurnal and
mantra for planners along the heavily developed semi-diurnal periodic vertical changes in still
eastern USA barrier islands where future progres- water level due to the gravitational effects of the
sive shoreline erosion as a result of RSLC has Moon and Sun on the Earth’s water bodies
been predicted (Leatherman, 2000). Despite the (Pethick, 1984). Tidal range translated into hori-
relationship between RSLC and shoreline erosion zontal displacement of the breaking wave on
being equivocal (SCOR, 1991), the debate on coastal slopes is important in designing struc-
Bruun underlines the role of RSLC as a major tures. Tidal range controls the rate of wave
coastal control element. Since the last glacially energy applied per unit area of beach face. In
related sea-level minimum, RSLC has caused a semi-diurnal conditions the high and low neap
substantial volume of unconsolidated sediment to tidal positions are occupied for the longest times
be reworked in the CZ. (Pugh and Faull, 1983). Tides are not entirely
Generally during the middle and late Holocene passive, but on the beach face their value lies in
(last 6000 years) there has been a deceleration in the elevation of waves and surges, and as such,
RSL rise rates, such that reworking has led to joint tide and wave plus surge probabilities are
reduced sediment volumes on contemporary important in the design of structures. Severe
beaches, which allows greater potential for storms superimposed on a rare high-tide level
increased geomorphic sensitivity in the CZ (e.g. can have more significance than frequent lower-
Carter and Wilson, 1993). energy events at neap-tide positions (e.g. the
In areas with near-stationary RSL, the rework- 1953 East Coast Flood in England).
ing and reduction of non-renewable sediment vol- Tidally induced movement of water masses
ume has also led to increased longshore erosion generates currents capable of sediment transport
of existing coastal sediment (cannibalisation). that are enhanced in shallow seas. Such currents
Many of the current coastal engineering problems show a prevailing bi-directional structure related
result from prevailing beach sediment scarcity to flood and ebb directions. The net velocity dif-
related to low RSL change over the last two ference in these currents is mirrored in an asym-
millennia. metrical morphology of offshore and estuary sand
The IPCC (Inter-governmental panel on bodies (Chapter 22). The position of shoals may
climate change) predicts (Houghton, 1997) that offer shoreline coastal protection from wave
future eustatic rise rates may be up to 8 mm/year. attack, as well as being a source and sink of
During the middle Holocene a similar variation in littoral mobile sediment. The influence of tidal
RSL rise may have led to the discontinuous currents on coastal morphology appears maximal
rolling onshore of beaches and barriers (Figure for beaches with a tidal range between 2–4 m
21.5: Jennings et al., 1998). There is a suggestion (Hayes, 1979), especially on offshore barrier
that some barriers can be overstepped and beaches rather than fixed/fringing beaches (see
‘drowned’ during periods of fast rise rates (Forbes below). On beaches with high tidal ranges
et al., 1991). Shoreline deposition may not be able (⬎ 5 m) there is a likelihood of tidal currents
to react coherently to the speed of future sea-level interacting with wave-generated currents to pro-
change, which opens up the issue of coastal vide greater transport potential around the low
GFE-21.qxd 6/27/05 4:37 PM Page 573
A.
Equilibrium profile
Initial Water surface
Closure
Depth
hc
B.
Increase of water S Final Water
level and profile surface
elevations
Initial Water
surface
Required Volume R
C.
Rh
Recession of
S
profile
Available Volume
D. Eroded
Net results
S
LS
R⫽
H
H
Closure
R Deposited Depth
Figure 21.4 Schematic development of shoreline recession as a function of rising sea level (after Bruun, 1962).
Equilibrium beach profile (A) moves upwards (B) and onshore (C) with a rise in sea level. Eroded shoreline sedi-
ment is required to build up the beach face to maintain an equilibrium beach profile. Recession is a function of
length of profile times sea-level rise increment divided by height of profile (D).
GFE-21.qxd 6/27/05 4:37 PM Page 574
⌬RSL rise thresholds

> 8.5 mm/yr < 2 mm/yr
Sea Level
Sed
ime
nt
Sediment supply
RSL
HIGH SLOW HIGH

morphological morphological morphological
retreat potential retreat retreat potential
DOMAIN 1 DOMAIN 2 DOMAIN 3
open barrier open/closed barrier open barrier
10 8 6 4 2 0
Time (1000 years cal. BP)
Figure 21.5 Domains of relative instability and stability of a gravel barrier as a function of RSLR over the last
10 k years (after Jennings et al., 1998). Barrier stability depends on longshore sediment supply, which is best with
high RSLR. However high RSLR discourages barrier spatial stability due to rapid barrier rollover. Barriers appear
to have been most stable under medium RSLR, although intermittent breaches in the barrier are likely.
tide position. Flood–ebb flow asymmetry depends The impact of empoldering was a reduction of the
on the relative balance of an estuary’s tidal prism tidal volume in the estuary (prism) and associated
retention capacity over the tidal period, and dic- disturbance of the flood–ebb fluid flow asymme-
tates the relative size of any sub-aqueous deltas at try at the estuary mouth or tidal exit, as reclama-
the estuary mouth (Chapter 20). Ebb-tide deltas tion especially reduced the intertidal friction
influence the consistency of longshore sediment generating areas of the estuary. The cross-section
transport in the near-shore zone. The lack of wave of the inlet mouth is proportional to the prism
activity and the dominance of fine sediment depos- (O’Brien, 1969). As the prism decreases so does
ition in the upper reaches of tidal estuaries the inlet width, and by association the size of the
encourage saltmarsh growth. ebb-delta diminishes, as does its protection of
History has tended to view tidal marshes as open-coast flanking (Nordstrom et al., 1986).
being of little value, allowing them to became Major post-seventeenth-century marsh reclama-
waste sites for growing estuary communities. In tion around the Wash and along the north Norfolk
the post-1850s, many marsh sites became poten- barrier coast has been considered responsible for
tial sites for reclamation, culminating in the ongoing coastal changes (Orford et al., 2000).
Dutch agricultural expansion as a by-product of Post-nineteenth-century marsh reclamation in
the impounding of the Zuider Zee for coastal pro- north-west Lancashire has been thought to
tection purposes. The east coast of the USA has account for substantial open-coast erosion (cf.
seen considerable reclamation in the post-1950s Pethick, 2001), originally thought to be a function
of back-barrier marshes for urban expansion of the engineering protection of Victorian holiday
(Lins, 1980). facilities. The problem with this reclamation
Reclamation has had a long history with trend is the sub-century timescale by which the
respect to agricultural uses of estuaries by the open-coast erosion problems became apparent
rapid enclosure of shallow areas (polderisation). (Orford, 1988). The wave of nineteenth-century
GFE-21.qxd 6/27/05 4:37 PM Page 575
reclamation optimism never had to deal with its which refract more and hence have smaller
twentieth-century impact. breaker approach angles. The majority of the
annual longshore sediment transport in southern
Wave energy and storm activity California is estimated to occur in c. 10% of the
Breaking wave height (Hb) is the driving force of time indicating the importance of storms in the
coastal zone morphological variation as wave coastal climate (Seymour and Castel, 1985).
energy is proportional to wave height squared. Onshore storms also generate a rise in water
Longshore gradients of Hb can be related to vari- level above the tidal prediction (i.e. storm surge).
ation in beach ridge height and beach sediment This rise is both a response to the low barometric
size. On gravel beaches, Hb correlates positively pressure accompanying storm centres and the
with mean sediment size. Engineers should, in geostrophic effect of wind activity over the sea
particular, be aware of two related aspects of surface. Elevations of 0.5 m are common with
wave activity: the range of high-magnitude, low- westerly-dominated depressions in the UK, with
frequency storms that occur, and the temporal supra-elevations of several metres being recorded
and spatial distribution of wave steepness (wave in severe storms. Surges are amplified when
height over wave length). depressions drive into enclosed basins and can
Storms dominate CZ geomorphological devel- be major flood hazards for low-lying coastal
opment. Figure 21.6 shows which coasts are margins, as witnessed by the destruction asso-
affected by winter storms. Most offshore sedi- ciated with the 1953 storm surge around the
ment losses from the CZ are due to return flows southern North Sea basin and the catastrophic
from breaking storm waves. Littoral sediment succession of cyclone surges along the Orissa and
drift rates are proportional to Hb2 ⫻ cosine of the Bangladesh coastlines of the Bay of Bengal
breaker crest angle with the shoreline, and are (Murty et al., 1986). The risk, scale and economic
several orders of magnitude greater in storms than impact of such flooding can account for major
under fair-weather swell, as the wave types of the protective measures like the Thames Barrier
latter have longer associated periods (⬎12 s) scheme to prevent extreme surges flooding central
60˚ ? 0˚ 60˚ 120˚ 180˚ 120˚
60˚ JANUARY 60˚
40˚ 40˚
20˚
0˚ 0˚
20˚ 20˚
40˚ 40˚
60˚ ? ? ? ? 60˚
I c e L im i t
60˚ 0˚ 60˚ 120˚ 180˚ 120˚
% Frequency
0–5 6–10 11–20 21–30 >30 Tropical storm tracks
Figure 21.6 Storm wave zones: frequency of ocean gales (Force 8 winds) in January. This is the basis of storm
wave environments (⬎ 20% gales). Note the tropical storm tracks that generate short-term severe storm wave activ-
ity in otherwise swell wave dominated mid- and low-latitude coasts (after Davies, 1972).
GFE-21.qxd 6/27/05 4:37 PM Page 576
London. Such massive engineering structures are SWL

rarely viable options for economically developing 0.100
Bar
countries of the low latitudes where greater mag- Profile
nitude surges are common as tropical storms (hur- Ho
ricanes, cyclones and typhoons) impinge on the 0.010 SWL
Lo
CZ (Figure 21.6). Step ␤
Profile
Storm-induced surges are often the principal
marine agent in back-beach and back-barrier 0.001
10⫺4 10⫺3 10⫺2 10⫺1
dynamics and deposition. Beach crest heights on
⫺0.67
barriers (ignoring any windblown element) are [tan ␤]⫺0.27 [d/Lo ]
related to storm wave plus surge elevations, though

Figure 21.7 Beach profile response to wave steep-
the highest combinations may overwash the crest ness (Ho/Lo). Step (constructional) profiles appear with
and produce remnant crest heights. Therefore low wave steepness and bar (destructional) profiles
design of elevated coastal structures to resist flood- appear with high wave steepness (after Sunamura,
ing requires a confidence band of water elevations 1975).
above existing beach crest heights. Prohibition on
building within a set-back zone that recognises the
Latitude
back beach spatial limits to severe surges may need 0˚ 20˚ 40˚ 60˚ 80˚
to become mandatory in some settings, so as to free 100
Shell
the accommodation space for geomorphic activity Coral
associated with extreme events. Maximum surge 80
Rock & gravel
penetration may be spatially progressive along a ?
shallow sea coastline, when a storm moves in con-
% Frequency
60 Mud
junction with the peak of a moving tidal wave
(Orford et al., 1999b). Surges may show a rhythmic
preferential longshore penetration (Dolan and 40
?
Hayden, 1983), though over time such positions Sand
appear to be self-balancing. Artificial breaches on 20
the beach crest due to vehicle, pedestrian and serv- ?
ices’ access can be critical to accelerating erosion
0
positions during storms.
Wave steepness has been used as an indicator Figure 21.8 Relative textural composition of contin-
of the likelihood of a beach eroding. Steep waves ental shelf sediment (after Hayes, 1967).
associated with storms tend to erode sediment
from the beach face and transport it into tempor-
ary sinks in the sub-tidal zone, often in longshore 21.4 Sediments in the coastal zone
bar form. Waves of low steepness attributed to
fair-weather swell tend to be constructional and Modern coastal zone sediment distributions are
rebuild the beach, in particular creating step-like related to products both of Pleistocene glacial
berms that act as reservoirs of finer sediment that events (Chapter 6) and of current terrestrial
may be deflated into back-beach dunes. The criti- weathering zones (Chapter 2). Figure 21.8
cal value of wave steepness that discriminates (Hayes, 1967) shows proportions of material
between step and bar profiles also depends on types on the world’s continental shelves. The
sediment size, wave length and beach slope substantial element of rock and gravel in the mid
(Figure 21.7; Sunamura, 1975). The value of such to high latitudes is the result of Pleistocene
predictive relationships for long-term planning is glacigenic activity. Bioclastics (coral) are
however being questioned and they should only restricted to equatorial warm waters, though car-
be used as guides to beach behaviour. bonate material (shell) is equally present at all
GFE-21.qxd 6/27/05 4:37 PM Page 577
latitudes. Low-latitude mud results from fluvially variation of the eroding terrestrial cover, as well
transported weathering products of equatorial as wave and current energy assemblages. These
chemical weathering regimes. assemblages help to sort and transport initial het-
Sediments in the CZ range from boulders to erogeneous influxes of sediment into more textu-
mud. The coarsest material exists in the most rally selective environments. McLaren and
energetic zones unless it is a lag deposit in situa- Bowles (1985) have shown how spatial residual
tions with sufficient energy only to remove sediment deposition can be seen as parts of sedi-
smaller sediment sizes. Silts and clays (i.e. mud) ment pathways that link coastal geomorphic
are rarely found at the shoreline of wave- units, although as Carter (1988) shows, similar
dominated coasts. Mud dominates in peri-coastal coastal sediment pathways may emerge from
(marsh and estuary) settings where wave energy is different energy gradients, such that uni-dimen-
expended seawards and tidal current activity is sional sediment selection is not a consistent
maximised, as well as offshore beneath the effec- mechanism.
tive wave-base. The continental shelves’ sediment
zones are observed to a large extent at the adja-
cent coastline (Figure 21.8). Figure 21.9 shows 21.5 Coastal organisation
the importance of pebbles on a world scale as a
beach-forming material and reflects the direct One of the difficulties with assessing human
impact of upper to mid latitude glacial deposits. intervention in the coastal zone is the lack of
This should not underestimate sand-sized mate- recognition given to the linkage of coastal geo-
rial as the dominant sediment by volume of beach morphic units and the cumulative effect that inter-
material at almost all latitudes. vention can have spreading across spatially related
Zonal patterns of sediment transport pathways units. Determination of linkage, therefore, needs
are disturbed at the local scale by the textural to be a central theme of coastal engineering. In
60˚ 0˚ 120˚ 180˚ 120˚
60˚
40˚ 40˚
20˚
0˚ 0˚
20˚ 20˚
40˚ 40˚
60˚ 60˚
60˚ 0˚ 60˚ 120˚ 180˚ 120˚
Important Occasionally important Unimportant Ice limit
Figure 21.9 Relative importance of gravel (pebbles) to world shoreline sediment (after Davies, 1972).
GFE-21.qxd 6/27/05 4:37 PM Page 578
this sense there is a difference between beach and wave–sediment cells (and also ‘littoral’ cells or
coast scales. The former is concerned with short- coastal process units).
term geomorphic analysis dominated by within- Most coastal cells rely on coastal erosion of
unit activity (i.e. how does the specific cliff or terrestrial sediment by waves as the contributing
beach or dune system change?). Coastal scale is source. Where sand is the dominant sediment type
more concerned with inter-unit linkages that then the cell sinks may be either onshore sand
respond in the meso-scale. dunes or offshore shoals. A beach represents a
There are three basic methodologies that focus transport corridor or sediment conveyor. They are
on coastal linkage: not necessarily the most efficient conveyor as
maximum transport rates may occur only for a
1. facies analysis low proportion of the time. The persistence of a
2. coastal cell analysis beach is either dependent on consistent up-drift
3. coastal behaviour systems. sources of sediment, or on less efficient longshore
transport conditions that leave sediment behind.
Facies analysis is a traditional sedimentological A beach’s formation and deposition can therefore
method by which vertical sequences of sediment- be viewed as temporary, representing a hiatus or
ary deposits are interpreted into horizontal suc- waiting stage in the long-term transport of sedi-
cessions of environments. The facies model is an ment alongshore.
attempt to reconstruct the energy gradients and Adjacent cells can be linked longshore by a
sediment flows (process and response) connect- constant drift direction, especially where geologic-
ing this spatial sequence (Reading, 1996; Relson, ally variable (crenellated or closed) coastlines
1979 for barriers). There is nothing unique to the occur (see below). Directional changes of drift in
coastal zone in facies modelling per se but as its adjacent cells will lead to either progradation or
outcome is to interpret vertical successions as erosion boundaries (Figure 21.10). The volume of
reflecting contemporary spatial sequences, its sediment moving along the transport corridor is
usefulness is beyond the requirement to establish reflected in beach morphology: single beach
linkages. Coastal cell analysis is a methodology ridges with a concave plan view form, open to the
specific to the CZ and concentrates on the coast, suggest limited longshore drift volumes,
wave–sediment linkages of the shoreline and as while multiple ridge systems suggest high trans-
such is now a standard perspective from which to port volumes. Pulses of sediment moving wave-
understand coastal organisation and shoreline like along the shore can indicate irregular temporal
behaviour. episodes of cliff erosion by major storms. These
features should not be confused with the regular
Coastal cells shoreline perturbations generated by reflected
Coastal cells are defined by: sediment inputs to wave energy (see below).
the shoreline (source); sediment movement along The recognition of cell element boundaries is
the shoreline (transport corridor or pathway); crucial in addressing S1-type problems (see
sediment loss (sink) from the shoreline. The con- above). Cells are delimited by positions where the
centration on beach sediment in cell analysis value of longshore directed wave power (Pl) is
reflects the importance of beach width and height zero. This occurs where there is no breaking wave
as a control on terrestrial coastal erosion. Cell or the angle of breaker approach to the shoreline
analysis, therefore, is a major tool for coastal zone is zero, regardless of the height of the breaking
management with respect to set-back and other wave (Hb). Longshore variations in both Hb and
controls on human activity, as cell delimitation approach angle are caused by wave refraction off-
identifies areas of potential coastal erosion and shore (Komar, 1998). Wave refraction is the
deposition. Given the importance of sediment to process by which deepwater waves differentially
coastal cells, as well as breaking wave activity, adjust to a varying bathymetric topography as the
coastal cells are sometimes referred to as wave moves onshore. The effect of offshore rises
GFE-21.qxd 6/27/05 4:37 PM Page 579
e/e
e/e
c e/a c
a /a
c Pl
Pl c
Pl a/a
Pl
Net beach erosion a/a Low beaches
Net beach deposition c Equilibrium

Pote
al
P o t e n ti n deposnitti al e/e High beaches
o s i o ion
er
e/a Longshore sediment protrusion
Increasing Pl decreasing Pl
Figure 21.10 Relative structure of coastal cells as a function of longshore variation in wave power: see text for
development (after May and Tanner, 1973).
and hollows under different points of the wave defined by all wave directions, leads to a residual
crest causes a variable response in wave crest cell structure (usually for a year). Figure 21.11
deformation along the crest length. Acceleration shows this analysis for part of Northern Ireland’s
and retardation of different elements of the crest Irish Sea coast (Bowden and Orford, 1984). Note
length cause it to deform when viewed in plan, that multiple drift reversals can be seen; such
thus varying the angle of breaker approach to the refraction-generated reversals are common on
shoreline. Wave crest distortion leads to bunching irregular or crenellated coasts like this one. Bray
and stretching of the crest length, which translates et al. (1995) offer further evidence of cell delimi-
into a variable longshore Hb and Pl. tation exercises at differing scales. Empirical for-
Cell delimitation requires increasing or mulae for the calculation of Pl and its connections
decreasing Pl values in the down-drift direction to longshore sediment drift rates abound. That of
(May and Tanner, 1973). Boundary a is the up-drift Komar (1998) has been widely used, though
cell limit where erosion starts, while e is the Horikawa (1988) cites alternative equations that
down-drift cell limit where no further longshore underline the very imprecise predictive nature of
transport occurs (Figure 21.10). Between a and e drift calculations. It is best to use non-calibrated
is c where Pl values are insufficient to maintain coastal cell analysis only as a long-term indica-
the sediment supply already being transported, tion of what might be expected with respect to
thus c marks the start of net deposition. The trans- cell structure.
port corridor stretches between the source and
sink zones. Two contiguous cells with a common
a/a boundary results in an erosion node with low Cells and coastal geomorphology
or non-existent beaches, due to drift moving away Usually the development of cells can be seen in
from the boundary in both shore parallel direc- the geomorphological structure and disposition of
tions. A common e/e boundary results in a depo- the cliff, beach, barriers and dunes. Figure 21.12
sitional node associated with high-volume beaches. shows the structure and relationship between ele-
An e/a boundary marks a sub-cell boundary of a ments of a sediment cell on a sand and gravel
major drift cell with a consistent drift direction. fringing beach system that is common to UK
A budgetary amalgamation of Pl results, for cells shorelines. It depends on cliff sediment sources
GFE-21.qxd 6/27/05 4:37 PM Page 580
Millisle
N
Knockinelder Ballywalter
Cloghy
Ballyhalbert
Portavogie
0 km 3
C. Sediment cell structure

e/e
e/e
e/e
e/e
a/e
a/e
a/e
e/a
e/a
a/a
a/a
a/a
a/a
a/a
c
c
c
c
MR
B. First difference model

6
deposition erosion
1 2 3 4 5 6 7 8
⌬Pl (KJ/ms)
A. Weighted average model

6
north
Pl (KJ/ms)
0
south
Figure 21.11 Coastal cell modelling along part of the Northern Ireland’s Irish Sea coast (after Bowden and Orford,
1984). A budgetary analysis (based on 400 m coastal units) of annual longshore wave power experienced at the coast
(A) is reduced to adjacent unit differences in wave power (B), which allow residual coastal cells (source and sink
elements) and boundaries to be identified (C).
and reflects the separation of gravel and sand on area and sediment size. This process has com-
the beach and the development of dunes as the pounded an ongoing reducing beach sediment
down-drift sink. This typology is conditional on availability due to the reduction in the rate of rel-
sediment volume and in most UK cases the source ative sea-level rise over the last two millennia.
areas are insufficient to maintain the cells as Figure 21.13 shows an example of cell devel-
coastal defences incrementally seal the source opment from the high-energy west coast of
areas off, such that beaches are moving into Ireland, where it is possible to recognise the cell
deficit sediment volume and thereby reduce in elements from the geomorphology of the beach.
GFE-21.qxd 6/27/05 4:37 PM Page 581
nt
Sedime
sink
or
ort corrid
Transp Low angle sandy
Cusp morphology beach in sink area
nt indicates reflective showing dissipative
Sedimee status of beach face status; intertidal
sourc ridges or subtidal
Lagoon trapped bar systems likely;
Fringing gravel beach ridge and multiple by ridge
onshore sediment
levels of cusp morphology indicate tidal development
Cliff movement in the
range; height of crest relates to surge
recession form of dunes
level and breaking wave height
due to Sediment size of ridge
subaerial Stabilised cliffs controls seepage and
processes due to toe-slope prevents breach point
and over- beach protection development
steepening
by toe Developing
trimming gravel beach
ridge
Movement of
fine sediment
to offshore sink
Toe-slope
marine
trimming Resultant longshore transport
of fines in the surf zone and
of gravel in the swash zone
Obliquely impinging waves leading to

general movement longshore in the swash zone
Figure 21.12 Generalised coastal cell typical of the British Isles.
Exposure to Atlantic storm waves has established Depletion and disturbance of this rubble mound is
a net longshore transport corridor (A) from the common under severe Atlantic storms, and such
source cliffs south of the village of Lahinch (B), material has to be renewed to maintain the
to the sediment sinks of the dune field north of defence. The depletion in beach sediment is now
Lahinch (C) and to the nearshore shoals in front progressing along in front of the dunes and its
of the dunes (C). Sediment source depletion has associated golf course where further coastal
resulted in a variable volume of beach sediment defences (late 1990s) have now been constructed
along the shoreline. The higher beach face in to prevent storm erosion of the dune given the
front of the dunes (D) as opposed to the lower decreasing sediment on the beach. The high cost
beach face in front of the village (E) reflects the of coastal protection for the village, given the
net deposition of the sink and the net erosion of static and massive structure involved, generated
the source. This difference in beach height is much debate over its construction.
picked out by the extent of draining water on the
lower beach face (E). To avoid storm flooding as Coastal cells and relative sea-level
well as erosion of the village foundations, a con- change
tinuous sea wall with all the attendant difficulties One of the unresolved issues about coastal cells is
of arranging access to the beach (steps and slip- their stability and relationship to rapid sea-level
ways) has been constructed. The rockfill (G) at change. Forbes et al. (1995) have indicated that a
the base of the wall is designed to prevent wave rising sea level will by accelerated erosion gener-
reflection off the wall, which can cause beach ate more beach face sediment than a stationary
scouring and undermining of coastal defences. sea level. Some form of equilibrium will evolve
GFE-21.qxd 6/27/05 4:37 PM Page 582
Figure 21.13 Coastal cell elements observed at Lahinch, western Ireland; positions A to G are identified in the text.
between the number of source areas and the beach beach sediment were materials supplied by coastal
volume that is being dispersed alongshore, which cliff and slope erosion of Pleistocene glacigenic
can maintain protection against further erosion. deposits plus unlithified Tertiary deposits in the
Fast relative sea-level rise (RSLR) is likely to be south and south-east. Much of the fine sediment
associated with smaller longshore cells due to from cliff erosion would have been transported off-
multiple erosion sources producing higher vol- shore while long-term sediment attrition would
umes of sediment than can be dispersed by the have depleted the beach-retained sediment. Little
beach conveyor. As the rate of RSLR decelerates, fresh sediment would have moved ashore in the
beach volumes are likely to reduce and greater storm wave dominated conditions. The retained
longshore cell elongation occurs. This is because beach volume was quickly dispersed longshore,
as beach volume falls there is a reorganisation of but it was rarely sufficient to maintain other than
the beach face away from the drift-aligned high- swash-aligned beaches and in some cases insuffi-
volume beaches to swash-aligned low-volume cient to do even that. The English south coast
beaches characterised by a zero-net longshore (specifically west Sussex) reflects a chequered his-
transport rate (Orford et al., 2001). Swash- tory of depleted cells where e/a positions should
aligned beaches tend to attach themselves to any have set a template for swash-alignment beaches,
differentially resistant coastal feature; therefore but for human intervention whereby beach drift-
their persistence depends on the sensitivity of the alignment has been maintained over the last
feature to which they are attached. century, originally by groyning, now by beach
During periods of low RSLR in the late re-profiling and nourishment (Gifford Associated
Holocene around the English coast, the sediment Consultants, 1997). This is a somewhat inefficient
supplied to the shoreline during the main Holocene approach, as a controlled movement to swash-
transgression would have been reworked by long- alignment (cf. 21.2 Perspectives affecting our
shore transport. In general, apart from limited understanding of the coastal zone) would reduce
beach progradation associated locally with fluvial the requirement for contemporary sediment main-
inputs, most beach progradation was at the cost of tenance, which will only accelerate in the light of
adjacent longshore erosion of cliff-supplied sedi- future predicted RSLR.
ment or of existing beaches (Lee, 2001). Hard rock The west Sussex coast exemplifies a lack of
cliffs provided little sediment volume for beach understanding of cell analysis in coastal manage-
replenishment. The principal terrestrial sources of ment. Although the concept was taken up as the
GFE-21.qxd 6/27/05 4:37 PM Page 583
basis for shoreline management planning in the local scale. A further problem of cell defini-
England and Wales (MAFF 1995: Figure 21.14), tion relates to the movement of substantial coastal
the delimitation of cells was based on a regional sediment offshore of the beach and beyond
analysis of major headland controls (macro- the nominated down-drift cell boundary. This is
geology) and major estuary sinks (Motyka and evident along eastern England where major vol-
Brampton, 1993). This delimitation can be criti- umes of mud derive from the Holderness coast-
cised as being too coarse to offer specification at line, move south past several cells into the
the local scale, given that only eleven cells cover southern North Sea and potentially end up in the
c. 1800 km of highly variable coastline. This Wadden Sea of northern Holland and Germany. It
coarseness dissuaded a furtherance of the proac- could be argued that this negates the value of
tive use of cells as an instrument of positive coastal cells, though it does not disallow the use
shoreline planning, though the human dislocation of shoreline cells wherever the issue of coastal
in realigning the Sussex coast would have created erosion and fluctuating beach sediment volumes
a political barrier to its implementation at even has to be considered.
St. Abb's Head

N
Sub-cell boundary
Cell boundary
Flamborough
Solway Firth Head
Great 11
Orme 2
The
10 Wash
Bardsey
3
Sound
9
The
St. David's Thames
Head
The Severn 8
7
5
Land's End Selsey Bill
Portland Bill 0 km 100
6
Figure 21.14 Macro-scale coastal cells identified by Motyka and Brampton (1993) and used as the basis for
Shoreline Management Plans by MAFF during the 1990s.
GFE-21.qxd 6/27/05 4:37 PM Page 584
Coastal behaviour systems above. The specification of behaviour at the meso-

Almost by definition coast cell analysis concen- scale allows attempts to parameterise processes
trates on beach and shoreline activities, which for predictive estimation at a similar scale
should not be surprising given beaches’ crucial (Terwindt and Kroon, 1993). Such parameters are
position as an indicator of likely coastal erosion often statistical statements of forcing and though
problems. However the beaches’ behavioural they have no physical manifestation, they do have
character may well be related to activity across the a characterising power. An example is the use of a
whole range of the coastal zone as well as to surge-forcing coefficient (Wheeler et al., 1999) to
regional-scaled activities that transcend single reflect annual storminess.
cells. Coastal behaviour systems (CBS) is a form The value of CBS is enhanced when the coastal
of analysis which attempts to broaden the investi- system is variable in form and type. Figure 21.15
gator’s perspective beyond the day-by-day activity exemplifies schematically the coastal system
and to use meso-scale plus timeframes in order to of the north Norfolk coast (UK) consisting of
set a geomorphological context to the problems. eroding and accreting sand-dominated barriers,
There is also a requirement to consider behaviour dune fields, tidal inlets, ebb-tide deltas and back-
as a linkage between what might be considered barrier marshes in a meso-tidal environment,
an ensemble of geomorphic units. Although the which all interact in terms of sediment pathways.
activity of a beach in terms of particle movement Cell analysis will work at the shoreline scale, but
is a legitimate study, it is the aggregate behaviour in this case is clearly limited to supporting a struc-
of a beach that is the required scale for CBS. tural analysis of the barrier beaches per se and
Establishing the controls on aggregate behaviour ignores the relationships between RSLR, histori-
clearly lift the analysis away from the instanta- cal sediment supply variations, barriers, tidal
neous scale and into contexts identified under inlets and associated sedimentation, tidal prisms
‘Time and space scales in the coastal zone’ and and back-barrier marshes: areas that are essential
‘Sensitivity and response in the coastal zone’, in understanding Norfolk barrier extension and
Salt
Marsh
Salt Marsh
Tidal creeks
Salt Marsh
Wave crests
Estuary/inlet flank
dune ridges Wave orthogonals
Central barrier linear
Aeolian transport vectors
dune and overwash fans
Sand–inter-tidal Littoral transport vectors
Figure 21.15 Schematic structure of the north Norfolk barrier coast to show strong interlinked relationship between
open-coast and tidal-dominated passes and back-barrier areas, which control sediment pathway and coastal evolution.
Coastal cell analysis understates the intimate connection of open coast and estuary (after Orford et al., 2000).
GFE-21.qxd 6/27/05 4:37 PM Page 585
dune emplacement (Orford et al., 2000). The CBS at high tide act as influential extensions of the
scenario should provide the holistic context in shoaling slope. Shallow slopes and wide offshore
which coastal cell development and contribution zones dampen wave energy, so that relatively high
can be understood for a range of time and space wave-lain beach ridges on the upper beach are
scales. unlikely except with extreme events (Aagaard
et al., 1997). Low shoaling slopes are often asso-
ciated with sediment sinks, and subsequent defla-
21.6 Effects of morphology on tion of the intertidal zone leads to back-beach
coastal zone structure dunes.
Differentiation between the slope of the inter-
Terrestrial morphology tidal beach and the adjacent offshore shoaling
Further constraints on coastal zone morphology slope is needed as they can reflect different
are imposed by the terrestrial morphology that the circumstances. The steepness of the intertidal
CZ borders. A coastline which is geologically beach face needs to be assessed against the width
complex, with folding and faulting paralleling of the area immediately landward of the high-tide
beach dip (discordant), and which exposes a wide position. Low-angle beach faces with no back-
range of lithologies of varying durability to wave beach volume can be indicative of reduced beach
attack, is likely to be a crenellated one with volume, while a low beach face and a wide back-
numerous fixed or fringing beaches with geologi- beach area (where foredunes may be developing)
cal structures differentiating coastal cells. Bray is indicative of a radically different condition of
et al. (1995) have identified how the discordant positive sediment supply to the beach. It is usual
coastal lithologies along the Devon and Dorset to find with sand beaches, holding both waves and
south coast (UK) have contributed to uneven cell tidal range constant, that steeper beaches are asso-
development as a function of the episodic retreat ciated with reducing sediment supply. As the tidal
of coastal cliffs. Accordant (open) coastlines, range expands, the likelihood of a single slope
especially those cut in unconsolidated glacigenic across the beach face reduces such that an upper
deposits, can show elongated cells unless facies steeper beach slope relative to a lower low-angle
variation (Chapter 7) offers initial perturbations slope appears.
that generate small disturbances to the macro-cell The nearshore slope strongly controls the rates
structure. Orford et al. (1996) stress how drumlin and modes of breaking wave deformation. Beaches
density along the Atlantic Nova Scotia coast sets can be placed on a continuum between total
the focus of coastal structure as the drumlins pro- reflected wave energy (steep beaches dominated
vide the point sources for barrier feeding, while by gravel) and total dissipated wave energy (low-
the inter-drumlin areas provide the accommoda- angle sand beach). Incident wave energy can gen-
tion space in which the barriers move under erate a range of secondary water movements that
RSLR. Irregularities in the offshore bathymetry have a range of morphological responses in the
can cause wave focusing leading to coastal cell type and spatial arrangement of inter and subtidal
drift reversals. sand bodies. Figure 21.16 shows this morpholog-
ical variation as a function of the surf-similarity
Wave shoaling slope index (Wright and Short, 1984) that characterises
The offshore shoaling slope of beaches affects the beach as a statement of wave structure relative
the translation of sea-level rise, tidal and surge to slope. Pertinent to engineers from this type of
amplification and rates of wave attenuation. The surf-zone analysis is the likelihood of longshore,
beach slope per se can be viewed as the final part spatially periodic, high and low beach morphol-
of the shoaling slope, whose importance in final ogy. Beach cusps on reflective gravel beaches and
wave transformation grows as the tidal range rip-current spacing on sandy dissipative beaches
increases. Macro-tidal (⬎ 4 m) sand beaches can control topographic cross-beach lows which may
have intertidal widths of hundreds of metres and act as avenues for breaker access that overtop
GFE-21.qxd 6/27/05 4:37 PM Page 586
Outer breaker zone Dissipative domain
5
Low frequency
Trough swash bores Spilling breakers
Dissipative 0
Inner breaker zone Flat concave
Trough
⫺5 beach
Beach
no longshore variability 0 Metres 300
20
Reflective dom. Dissipative domain
Bar Straight Cresentic Berm and cusps

Intermediate 3 Surging Plunging
Trough 0
longshore
⫺3 Steep
bar-trough reflective Bar
Cusps may be present inner beach face
0 Metres 250
Mod ref D Mod dis D
Normal waves Oblique waves

breaker amplitude (2⌸ / T)2
3
A B
A´ A
0
Migrating bar
Intermediate ⫺3
rhythmic bar Crescentic bar Rip Low ref D Mod dis D

g tan2 ␤
and beach Mega 3

cusp Scarp B´ B
Mega 0
horn Rip
cusp
A´ B´
embayment ⫺3
0 Metres 200
Var dis D
Normal waves Oblique waves 3
A B MHW
Surf similarity ⑀⫽
A´ A
0 MLW
Transverse ⫺3 Mod dis D
intermediate Rip
3 Dis or ref D
bar and rip Welded High MHW
B´ B
0 MHW
A´ B´ bar tide
cusps
⫺3 0 Metres 200
Normal waves Oblique waves 3 Ref D

MHW
0 A´ A
A B MLW
Intermediate Runnel ⫺3 Dis D

ridge-runnel or Ridge Terrace
3 Ref D
low tide terrace MHW
High tide Skewed B´ B
Runnel 0 MLW
cusps mini rip
A´ B´ Dis D
⫺3 0 Metres 150
1
Reflective domain
Wide berm
3 Cusps
Steep beach face
Surging
Reflective Berm crest 0 MLW
Cusps Berm
Runnel Steep
beach face
⫺5
0 Metres 200
Figure 21.16 Integration of beach face morphological variation through the concept of surf similarity (after
Wright and Short, 1984).
GFE-21.qxd 6/27/05 4:37 PM Page 587
beach crests. Otherwise temporal changes in sediment deficit leads to an erosive pseudo-
beach-face morphology affect rates of longshore transgressive morphology (usually in the form
sediment transport and longshore Hb gradients. of a single beach ridge dominated by crest
These morphologies are variable as a function of overtopping as a function of extreme events,
wave climate, however less variation is found hence ‘pseudo-transgressive’). The rate of
with beaches that are at the continuum end posi- transgression is controlled by RSL rise while
tions (reflective and ultra-dissipative: Wright and progradation can be conditional on RSL fall
Short, 1984). Wave climate generally refers to the (forced regression) or on sediment deposition
distribution of wave heights and periods that can pushing the shoreline seawards (normal
be expected over a year at any one site. While regression).
regional wave climate refers to offshore condi- 4. Terrestrial basement morphology upon which
tions, more specific wave climates apply at differ- the coastal environments are formed: in par-
ent positions along a coastline when wave ticular the degree of coastal crenellation or
transformation through shoaling and refraction irregularity indicates the likelihood of open
would have taken place. It is this distribution that (wave dominated) and protected (tide domi-
will specify the beach morphological range, con- nated) depositional environments. However,
trolled by available sediment size and offshore note that protection is often associated with
slope. At a meso-scale level it is important to the ability of processes to use the available
recognise that offshore, intertidal and back-beach sediment to form wave-lain structures that
slopes are rarely in step due to the variable sedi- offer seaward protection to areas where wave
ment volume working through a coastal system. energy is reduced or negligible and tidal
These slopes would only become balanced when currents enhanced.
sediment input was either zero, or when sediment 5. Tidal range, which has the effect of enhancing
input balanced output across the coast, and dislocation of longshore beach continuity as it
process conditions were stationary. The improba- increases.
bility of these conditions means that slopes are
usually out of step such that any one slope is not The combination of these variables leads to several
a safe guide for characterising the whole coastal basic beach forms that, once established, control
system. the spatial disposition of other environments
dependent on beach morphology for protection
from wave energy.
21.7 Controls on coastal morphology
The morphology of coastal environments is too 21.8 Key clastic coastal depositional
varied for discrimination on other than a multi- morphology
variate basis. Coastal morpho-sedimentary envi-
ronments reflect five interacting factors: Mid-latitude coastal morphology dominated by
clastic sediments varies from fringing beaches
1. Mean sea level acts as the datum for all activity, (Figure 21.12) to beaches found on barriers stand-
while its relative change rate provides the tempo ing seaward of the terrestrial shoreline (Figure
control on change and sediment availability. 21.17). In general, fringing or fixed beaches will
2. Wave climate controls the retained sediment result from:
size as a function of wave height. Climate also
identifies the extreme events that control the 1. higher median wave energy (maximised in
basic depositional framework. storm wave environments: Davies, 1972)
3. Sediment source, supply and size range: 2. higher tidal range (meso–macro)
excess sediment leads to progradation (usually 3. low sediment availability
in the form of multiple beach ridges), while 4. steeper offshore slope.
GFE-21.qxd 6/27/05 4:37 PM Page 588
Figure 21.17 A natural coastal barrier at Long Island, New York, USA (photograph courtesy of J. Allen, US Parks).
Coastal barriers are more likely with the of accumulated wealth into coastal real estate
reverse of the above conditions. The origin of (Figure 21.18) and thus create the potential for
coastal barriers is still debated (Leatherman, 1982) major protection issues in the future.
though a single origin is now considered unlikely.
Although their presence is limited to about 15% of Fringing or fixed beaches
the world’s coastline, they are the physical founda- Fringing or fixed beaches are a common feature
tion to some of the world’s most valuable real of meso- to macro-tidal (⬎ 3 m), storm-wave
estate: property worth ⬎ $5 ⫻ 109 can be found (Hb ⬎ 1.5 m) dominated coasts (e.g. north-west
along the barrier coast of the eastern USA alone. Europe). This mid-latitude zone tends to have
Even barriers in economically developing coun- experienced a decelerating rate of relative sea-level
tries like Brazil can attract a disproportional share rise since the mid-Holocene. These beaches,
GFE-21.qxd 6/27/05 4:37 PM Page 589
Figure 21.18 Urbanised barrier west of Rio de Janeiro. There is a future issue of beach face protection to be faced
as well as an existing pollution problem in the lagoon separating the two phases of urbanisation on the seaward and
landward barrier ridges.
particularly those that are gravel-dominated (shin- linked to the rear of once-larger bay-wide barrier
gle), are often remnants of a rolling-onshore beaches controlled by a hinge-point terrestrial mor-
process associated with a weak transgressive phology (e.g. Start Bay, Devon). Most of these lat-
shoreline. The growth and diminishment of fixed ter beach and barrier forms were transgressive and
or fringing beaches is strongly controlled by long- have now overridden the once back-barrier wetland
shore supply. Around the UK coastline much of deposits, generating an unstable unconsolidated
that supply comes from cliffs formed from sub-beach stratigraphy. Gravel barriers driven
Quaternary glacigenic deposits and unlithified onshore quickly lose their coherence when pushed
Tertiary sands and clay that have been weakened by against headlands of a crenellated coast or when
Quaternary periglacial activity. The episodic fail- RSLR slows and longshore sediment supply
ure of such cliffs as a function of toe-slope erosion diminishes (Jennings et al., 1998). Headlands act
has been noted as a formative process of coastal as refraction hinge-points, upon which the remnant
scenery (Brunsden and Jones, 1980). Lee (2001) barrier pivots, to be left stranded in the bay head as
offers an illuminating perspective of how cliff a fixed beach.
supply fluctuations of the last two millennia have On most contemporary fringing beaches
contributed to the strength or weakness of contem- around the UK, reducing longshore sediment
porary fringing beaches along the coast of south supply and longshore transport specify beach
central England. Sediment reductions in the face of budgets. Few beaches show sediment progradation
reducing RSLR identify a period of beach consoli- as most are in deficit mode, so that the best depo-
dation, characterised by a single beach ridge that sitional coast and beach morphology is usually
has to act as the basic coastal protective bulwark to exhibited at the down-drift cell boundary (e.g.
flooding and to wave energy along the coast. There sand dunes, Figure 21.12). However, many dune
is little coastal zone related deposition landward of systems around the British Isles show a lack of
these fixed units, though limited wetland develop- modern foredune development reflecting a con-
ment may form at the rear of such beaches where temporary reduction in beach volume. Fringing
seaward terrestrial drainage is impeded. Extensive beaches show marked spatial separation of fore-
wetland areas are limited to major depositional shore sand and rear beach gravel elements, with
sinks in crenellated and estuary coasts (e.g. The the highest wave-lain ridges found on gravel-
Wash, outer Thames estuary), or are remnants dominated beaches. The beach crest height is
GFE-21.qxd 6/27/05 4:37 PM Page 590
governed by extreme events (i.e. related to the ver- high-energy beaches. Sedimentation depths of
tical limit of swash run-up in storms). Run-up can modern fringing beaches are small, with the sweep
overtop the beach crest by which sediment can be zone (difference between the lowest and highest
incrementally added to the ridge top, while over- beach profile at the same position over time) far
washing of crests in severe storms can remove less than the absolute beach crest height.
crest sediment to a back-beach position (hence Stratification of beach ridges is common, with
beach ridge rollover). The rate of rollover retreat wave-lain sand-based ridges usually incremented
for gravel-dominated ridges has been related to by blown sand. Such aeolian development has been
RSLR rate: c. 1 m barrier retreat per 1 mm of termed dune decoration. Fringing gravel beaches
RSLR per year (Orford et al., 1995). It is not a contain predominant seaward dipping sedimentary
continuous process as the interplay of overtopping units with limited landward dipping ones at the rear
and overwashing is dependent on the size and of the crest related to the rollover process. Gravel-
spacing of storms. Overtopping can also occur dominated remnants of bay-mouth barriers now
under swell wave conditions, as the high reflective stranded can show major landward dipping, planar
status of gravel beaches can generate secondary and lenticular stratification (Figure 21.19) with
water motions that can amplify swell run-up suffi- only superficial surface seaward dipping units.
ciently to overtop. Most fringing beaches are still Back-beach stratification formed by storm wash-
quasi-transgressive due to overtopping, despite over (⬍ 5 cm thick), show a coarsening upwards of
longshore-dominated transport on the lower beach. grain size, and have been used as groundwater or
Beach face constructional berm height and storm water seepage corridors.
beach slope angle are a function of Hb, wave period Beach ridge breaching by fluvial sources
and grain size (Figure 21.12). Mature gravel ridges depends on the sediment-controlled percolation
down-drift of the source area can show alongshore rates (Figures 21.20 and 21.21). A transition from
sediment-size grading related to longshore Hb gra- seepage to open-channel flow (breach) for gravel
dients. This is best observed on beaches without can occur when percolation rises in excess of
sediment recharge, and points to the difficulty of 1–2.5 m3/s per metre beach width, and for sand
nourishing gravel beaches without the correct beaches around 0.3–0.8 m3/s per metre beach
specification of depleted sediment size elements. width. Fluvially derived breaches usually occur at
On less energetic gravel beaches, cross-beach cell divides (a/a point) or at low beach positions
particle-form grading occurs, related to size. due to low Hb values. The absence of distinctive
Beach-face sand size depends on wave energy and breaches in gravel ridges means that overtopping
sediment renewal. A typical median size for by storms leads to a constant onshore movement
high-energy beaches ranges from 0.15 mm to of sediment, as there is no feasible sediment
0.18 mm, with the best sediment sorting found on return to the near-shore, as happens with tidal
South North
10 Crest Washover fans
Atlantic
m I.O.D.
beach Lake swash bars

5 Lady's island
lagoon
Beach gravel bar Washover flat
0 stratification stratification
0 50 100 150 200 250

Metres
Figure 21.19 Internal structure of a gravel-dominated barrier in southern Ireland (‘m I.O.D.’ refers to metres above
the Irish Ordinance Datum, the local levelling datum in Ireland, approximately mean sea level).
GFE-21.qxd 6/27/05 4:37 PM Page 591
A. Seepage B. Seepage lagoon
Barrier
Barrier
lagoon
Seepage
Seepage
Sea rills
Sea rills
C. Channel outlet D. Channel inlet
Estuarine (tidal) Swash bars

lagoon (flood)
Sea
Sea
Flow Flow
lines lines
Swash bars Swash bars
Outlet channel (ebb)
Figure 21.20 The development of fluvially generated breaches in coastal barriers as a function of seepage (A); back-
barrier freshwater lagoon development (B); overtopping in storms cutting channels enlarged or maintained by fluvial
flow (C), and substantial maintenance by tidal action forming a brackish lagoon (D) (after Carter et al., 1984).
passes on sandy barriers. Seaward dip of fringing of major beach sediment depletion. Dunes and
beaches stops excessive landward seepage under beaches may also show rhythmic variations in
storms, while a lack of distinctive seepage points volumes related to seasonal wave climate varia-
on the landward side of beach ridges reduces the tion. If there is long-term sediment loss, then
possibility of incipient breach points and land- resorting to primary engineering structures (i.e.
ward flooding. walls and groynes) will at best only reduce rates
of loss in the short term, while at worst accelerate
Engineering intervention on beaches erosion at other points. The uses of structures that
Engineering intervention should be assessed in de facto prevent further sediment passage to the
the light of the sediment balance in the coastal beach (walls, bulkheads, revetments), or interrupt
cell. Successful attempts at controlling cliff ero- longshore sediment transport (groynes), have
sion may well cause sediment supply problems been commonplace in the engineer’s armoury.
for down-drift beaches, though cliffs of a high However useful these have been in reducing land-
clay content are unlikely to be the source for ward erosion rates, they have serious detrimental
down-drift sandy beaches. The requests for beach consequences for the beach sediment system, as
engineering defences are usually prompted these interventions in longshore sediment supply
because of a lack of beach volume that can act as only cause further depletion down-drift. The role
either a natural adjusting buffer to breaking wave of such structures should now be more a question
energy, or an elevation to prevent storm overwash. of regional perspective than local need.
Such sediment losses need to be carefully The priority of engineering solutions has in the
assessed, as seasonal differences in beach-face past two decades given way to the use of a ‘greener’
volume should not be taken as the only indication solution, that of beach renourishment. It is usually
GFE-21.qxd 6/27/05 4:37 PM Page 592
6
Sand Gravel
Maximum potential head (m)
5 Seepage
4
25
01
5
0.0
0.0
0.0
02
Channel
01
0.0
0.0
0.0
4 Transitional channel/seepage
K⫽
3
Grave
l
2
Sand
1
0
0.001 0.01 0.1 1 10
Discharge cumecs/1000 m length/100 m width barrier
Figure 21.21 Seepage flows required to form channels in sand and gravel barriers (after Carter et al., 1984).
(‘Discharge cumecs’ is a measure of discharge in cubic metres per second.)
nourishment at point sites rather than nourishment distinctive channels can cut back into the beach
for the whole beach conveyor per se, given that the crest and become early potential avenues for over-
volumes required for the latter are prohibitive. One- washing and back-barrier flooding. Engineers
off exercises at beach rebuilding with sand have should aim at preserving the natural pattern of
been restricted to large sites (km) to ensure beach form and material in the transport corridor of
economies of scale (e.g. Bournemouth, southern fringing beaches — though this is virtually impos-
England), as obtaining sand coarser than the origi- sible to do. The disturbance of a beach structure by
nal (for beach retention) is environmentally diffi- cross-beach trenching will, despite backfilling, also
cult. England also exemplifies the undertaking of lead to localised seepage disturbances.
extensive gravel nourishment, notably the west
Sussex coast. The problem of attempting to match Barriers
sediment in terms of size and shape is reduced The geomorphic attributes of fringing beaches
given the anisotropic dominant beach lithology — can be applied equally to beaches on barriers.
flint — that supports a low level of shape selectiv- However, the engineer has to widen the reference
ity in beach sedimentation. This issue gains almost frame when working with barriers by considering
national importance at Dungeness where a nuclear the barrier’s regional setting and its position rela-
power station needs protection from longshore tive to the terrestrial shoreline, back-barrier water
gravel depletion, with down-drift sediment being bodies (lagoons or bays) and water connections
transported by vehicle to an up-drift re-injection between the open sea and lagoons. Barrier types
site. Cynics debate whether the recharged sediment can be represented as a continuum from free-
or vehicles reach the down-drift sink sooner! When standing offshore to fixed position onshore, but
sediment recharging and profile remodelling is are best differentiated on the basis of tidal range
undertaken to raise beach crest levels against storm and sediment size. Although there are numerous
flooding, it is inevitable that upper beach stratifica- gravel-dominated barriers in the mid to high lati-
tion will be lost. Raising beach crests in this way tudes (Orford et al., 2001), they pose substantially
means that upper-beach backwash cannot dissipate fewer engineering problems than sand barriers of
via the subsurface, such that sediment starts to the mid-latitudes
move down-beach under even constructive swell The best freestanding and continuous sand bar-
conditions, and by a pseudo-headward erosion riers are in micro-tidal conditions, with medium to
GFE-21.qxd 6/27/05 4:37 PM Page 593
low wave energy (Hb⬍ 0.5 m) punctuated by and transgressive islands and reflect swash bar
peaks of high wave energy in tropical storms and accretion of longshore drifted sand moulded by
hurricanes, sand dominated, and on a low-angle wave refraction around the ebb-tide deltas.
shoaling slope (e.g. Gulf of Mexico). Barriers are Although island positions tend to be stable, the
absent from the waters of north-west Europe, erosion of centre beach material for barrier
except where substantial fine sediment and low- growth at terminal positions means that the
angle shoaling slopes coincide (e.g. the southern islands appear to oscillate in place. Oblique wave
North Sea: FitzGerald et al., 1984). Hayes (1979) approaches (e.g. east coast swell from the south-
and Leatherman (1982) indicate basic controls on east along the US barriers) cause island elonga-
barrier type. Leatherman (1982) divides barriers tion and down-drift thickening, leading to islands’
on the basis of their regressive or transgressive ‘drumstick’-shaped plan view. Storm-generated
status (in terms of sediment supply) and micro- or overwash is not as geomorphologically important
meso-tidal range (Figure 21.22). As the tidal on meso-tidal as on micro-tidal barriers, as barri-
range increases towards meso-tidal, then tidal ers are wider and sediment can be moved to back-
inlets that have a strong spatial stability segment barrier positions via inlet activity.
the barrier’s length. Barriers in micro-tidal envi-
ronments tend to be wave dominated, long, nar- Engineering intervention on barriers
row, low lying and continuous. Such inlets as do The US east coast barriers have had a major his-
occur are generated initially as washover sites tory of engineering intervention during the
during severe storms (usually hurricanes) and twentieth century. Intervention on barriers has
then enlarged into outlets by out-flowing storm followed the pressure of human recreational
waters trapped in the lagoon. The inlets are often occupancy. There has been a strong tradition of
unstable, as they can migrate rapidly in the face of building right at the shoreline edge to take
high-volume longshore drift, or seal up quickly advantage of the view, but also to maximise the
given there is no tidal exchange to keep them effect of onshore coastal breezes during other-
open. Sediment washed over the low barrier by a wise stifling and humid summers experienced on
storm surge may return to the beach face via tem- the eastern seaboard. The density of barrier
porary outlets, flowing out with the storm water recreational use is a result of the proximity of the
return. Fine sediment can also return to the beach main urban centres of the north-east with the
face by aeolian action. Lack of lagoon tidal action northern barriers of New Jersey (Nordstrom et
means little marsh growth. al., 1986) and New York having a long history of
Meso-tidal (2–4 m) environments show more traditional coastal defences: walls, revetments
stable barrier islands fronting substantial wet- and groynes. This is a micro-tidal area with long
lands caused by tidal exchanges of water and sed- transgressive barriers that experience severe
iment into the lagoon from the open sea via pressure for defences, as they are both low and
relatively fixed inlets (Figure 21.22). Inlets are narrow. The amount of natural dune activity —
more common, with barriers readily defined as that might elevate the barriers — is limited, and
islands. Most inlets are linked with flood and ebb- attempts to maintain any dune line as a natural
tide deltas as a consequence of cross-barrier tidal defence have had a varied history of success
flows. Ebb deltas are important for the open coast using mixtures of dune fences, vegetation seed-
as their morphology influences nearshore wave ing and fixing, as well as the near-universal dune
refraction, and hence beach development at the access control. The concept of shoreline-parallel
island–inlet boundary. Switches in ebb-delta dep- boardwalks for recreational promenading has
osition may change the wave refraction focus been extended to major structures overarching
positions and move erosion or deposition along- the dunes to prevent dune trampling, but allow
shore; it is not uncommon for these zones to jump beach access (Figure 21.23).
spatially longshore. Multiple beach ridges at The main engineering issue is about maintain-
island longshore ends are common in regressive ing beach width in the face of barrier rollover and
GFE-21.qxd 6/27/05 4:37 PM Page 594
A. Transgressive barrier: microtidal (⬍ 2 m) C. Regressive barrier: microtidal (⬍ 2 m)
Long, narrow, low-

lying barriers. Wave Positive sediment
dominated with storm balance allows
surges overtopping parallel beach
to cause maximum ridges to develop
Few inlets relative to
barrier damage
barrier length. Most
dominated by flood-tide
deltas associated with
storms
Limited marsh Multiple ridges
due to low controls washover
Overwash fans breach tidal range and access
entire barrier due to single Sand provides dune limited water
ridge and narrow width accretion to wave lain exchange
lagoon beach ridges
Transgressive beach Multiple arcuate

overrides any marsh Limited back barrier ridges with
sedimentation. Overwash Engineers need to sediment for marsh dominant longshore
fans coalesce to form an seal fans/breaches to build upon. drift
apron. with sand fences
and plants to
encourage sand
deposition Sealed inlet due to
longshore drift.
Inlets do not stay
Barrier beach open in face of low
complex
tidal exchange
B. Transgressive barrier: mesotidal (2–4 m) D. Regressive barrier: mesotidal (2–4 m)
As tidal range As tidal range

Generally stable
increases then increases then
inlets unless
continuity of barriers continuity of barriers
Multiple ebb-dominant influenced by
in the longshore in the longshore
tidal passes positive balance of
direction is lost by direction is lost by
longshore drift.
the presence of the presence of
numerous tidal inlets numerous tidal inlets
– which define barrier – which define barrier
islands islands Multiple prograding
beach ridge systems.
Dominent uni-
Island centre is most
Multiple beach ridge directional drift can
stable position. Island
development around lead to skewed
ends align with storm
tidal pass flanks due barrier island –
and inlet control
to localized refraction or drum-stick islands.
around ebb delta
Tidal inlet shoal
Tidal channel Refraction around
ebb-tide delta leads
Longshore Erosion centre of
Extensive wetland to swash bar welding
drift barrier island - lowest
development based direction at island terminus
point is centre for
on major tidal overwash fans –
exchange of water though fans are rare Stable islands with
and sediment. in other than rare overwash
transgressive barriers sedimentation
Shore-
face
Figure 21.22 Structure of micro-tidal and meso-tidal barriers along the US east coast (after Leatherman, 1982).
predominant longshore transport. The ‘northern elevation and prevent overwashing are actively
states’ beaches are relatively free from hurricanes employed. Re-profiling (bulldozing) has been a
but ‘Northeasters’ (intense, fast moving storms) common method of raising beach ridge elevation
are more than capable of generating overwash. as a defence against storm surge. Beach sediment
Not only is overwashing the only viable means for renourishment has also been widely employed as
coherent barrier retreat, but it is also seen as the a means of expanding the beach width to act as a
major disturbance factor to human barrier activ- wave buffer, though its efficiency in the long
ity, so that attempts to build up the beach ridge term has been widely debated in recent years.
GFE-21.qxd 6/27/05 4:37 PM Page 595
underlining the need for beach nourishment;

(B) the initial 16 km beach replenishment
costing $60 ⫻ 106 in 1980 and (C) the state of
the beach over a decade later. The sheer size of this
scheme and the lack of any hurricane activity on
Miami Beach probably accounts for its persistence,
although it has been prone to some longshore reor-
ganisation. However, other US experiences show
that few beach renourishment projects last for
longer than 10 years with nearly 50% of projects
failing within 5 years. This high failure rate con-
firms that continual renewal of beach sediment is
Figure 21.23 Possible misplaced concern for barrier required and identifies how ineffective nourish-
dune protection shown by over-elaborate boardwalks ment can be if the offshore extension of re-profiling
when the broader issue of urbanisation and its effect on is not fully undertaken. Even then a sequence of
dune evolution has been ignored. Fire Island, New York
stormy events may reduce nourished volumes to a
State, USA.
position of non-seasonal recovery.
The use of beach nourishment has received con-
Figure 21.24 shows the changing state of the siderable support from conservationists, relative to
biggest USA renourishment project at Miami the alternatives of fixed engineering and is still
Beach, Florida: (A) shows the human investment seen by many as the only effective sustainable con-
at risk in 1972 due to the loss of the natural beach, trol on coastal erosion despite its cost ($160 ⫻ 106
A C
Figure 21.24 The use of beach nourishment as a means of coastal defence at Miami Beach Florida. Miami before
nourishment (A) relied on sea walls and groynes. The threat to major real estate investment generated the then (late
1970s) biggest US renourishment scheme designed to act as a barrier to hurricanes (B). The scheme in 1992 (C) shows
a substantial element of the beach in place, however Miami Beach has not been tested by a hurricane as yet. Photograph
(A) Courtesy of RWG Carter; photogaph (B) Courtesy of S Royle.
GFE-21.qxd 6/27/05 4:37 PM Page 596
for US national renourishment schemes). However, unsuccessful given the lack of sediment volume
it is important that this approach should not be seen currently in the transport corridor and the virtual
as a one-off exercise and that remedial renourish- absence of a drying area for sediment deflation. The
ing will be needed, despite the fact that the US threat of ‘Northeasters’ is not only an ocean-side
federal authorities are reducing their support of this problem, the associated surges in the tidal lagoon
cost. The presence of nourishment is often notice- can flood and erode bay-side developments, many
able by the down-drift structural containment of which are built on dredge spoil. The hurricane
barrier that marks an administrative or property that intermittently reaches this far north further
boundary, rather than a natural sediment cell compounds this bay surge problem. Evacuation of
boundary (Figure 21.25). the island’s population in the face of a hurricane
Figure 21.26 shows Long Beach, mid New landfall is a real possibility for many US barriers.
Jersey, an example of a transgressive micro-tidal The limited access to and from barriers (one bridge
barrier island that has been extensively urbanised for Long Beach’s 31 km shoreline) presents a bot-
for recreational purposes. Property development tleneck for all evacuation plans. The lack of hurri-
has been a continuing aspect of the island’s history cane-proof buildings on the barriers offers little
since the late nineteenth century, and represents hope for those people caught on them.
the main incentive for Federal-funded coastal A major threat to barrier stability and human
defences to counter shoreline recession (⬎ 2 m/year use arises with hurricane landfall on barriers. The
prior to groynes) as a function of RSLR of threat of hurricane landfalls rises in central and
⫹2–3 mm/year. Rock groynes have greatly southern east coast states and the risk of severe
reduced longshore drift rates, but the terminal scour storm-induced surges (⬎ 2–3 m) strongly influ-
problem necessitates their presence along the entire ences engineering responses. The emphasis is on
island’s shoreline. The narrow beach indicates hurricane proofing properties rather than hurri-
attempts to control shoreline recession. Attempts to cane proofing barriers, as the latter is near impos-
recreate a dune ridge by erecting sand fences proved sible in view of the accumulated experience over
Figure 21.25 Beach renourishment has become a major element of US coastal protection, but issues of federal sup-
port relative to private and public beach access inevitably leads to piecemeal beach nourishment that leads to a dis-
tortion of beach widths given longshore redistribution of sediment. Public and private access along St Petersburg’s
(Florida) shoreline can account for inconsistencies in beach width post nourishment. Source of photograph unknown.
GFE-21.qxd 6/27/05 4:37 PM Page 597
Figure 21.26 The scale of urbanisation on barriers is exemplified on Long Island Beach, New Jersey, USA.
Controlling the loss of beach by multiple phases of groyning can be seen, though the success is negligible given that
the loss is due to human attempts to prevent beach movement onshore with RSLR. The loss of beach width makes
the attempts to form dunes by dune fencing (at the high water positions) a pointless exercise, indulged in as a means
of morale building more than beach building.
the twentieth-century, while the massive urban to be built, but that more strategic zoning and
development of vulnerable barriers underlines the better building regulations are established and a
need for the former. Further emphasis is placed on better awareness of normal coastal behaviour is
forecasting hurricane landfall positions in order to encouraged.
give time-dependent warnings for evacuation. The The level of urban use of gravel-dominated
difficulty of obtaining precise landfall positions barriers is virtually negligible in comparison to
other than within 12–24 hours of the event and the sand barriers. The main problems posed by gravel
limited evacuation routes means that warnings are barriers are related to a reducing sediment supply
not spatially specific and require substantial popu- and the retention of spatial position under rising
lation movement that might not be justified in light RSL. Gravel-based barriers in the UK that have
of the actual landfall. People resist calls for evacu- acted in the past as flood barriers are now being
ation and so instead of horizontal evacuation, some rolled landwards under rising sea levels. The asso-
areas legislate for vertical evacuation by proofing ciated problems are not always with the barriers
buildings to resist the force of surge flows and ele- per se, but rather with the wetlands (both fresh
vating dwelling positions above predicted surge and brackish) that they protect, which form bird
levels. Coastal zone management on the eastern habitats and sites of special interest with respect to
US seaboard is among the most proactive in the bio-diversity. As the barriers migrate, the wetlands
world with emphasis now on planning preventative are caught in the ensuing coastal squeeze, and
measures and educating the population to be more often their legal designation means conservation-
responsive and flexible to the hazards of the CZ ists have to argue for replacement areas (‘no net-
rather than expecting engineers to defend them loss’ policy). The difficulty in finding effective
totally. This is not to say that more defences are replacement areas for the loss of old habitats
GFE-21.qxd 6/27/05 4:37 PM Page 598
inevitably puts pressure back on resisting barrier engineers need to understand the impact and
retreat — yet resisting retreat only builds up the change to coastal zone systems caused by prior
likelihood of catastrophic change when the barrier attempts at engineered solutions. Figure 21.27
is next breached. This type of complex dilemma is shows the variety of ways in which a single
becoming more characteristic of the next genera- basic problem of building a coastal promenade in
tion of coastal problems. Barrier mobility is the Victorian times (S1) was the trigger for subsequent
key to invigorating back-barrier habitats, but if phases of engineering to counter S2 problems thus
humans still maintain the landward limit of initiated. Inherited sequential engineering struc-
contemporary wetlands then international legal tures and their effects are a constant problem for
obligations relating to coastal habitats will be coastal engineers. The loss of fresh sediment to the
invoked, as a consequence of which engineering coastal cell and the interruption of longshore sedi-
modifications of barriers will be forced upon us. ment are the two principal generic results of most
intervention. If we wish to return to more sustain-
Inherited engineering structures able shores then the next generation of engineers
The range of intervention that has been undertaken will have to be more effective in advocating the
over the last century means that contemporary removal of much of their predecessors’ work.
Natural sediment cell
Sediment Sediment transport corridor Sediment sinks and

source and transient sinks decreasing sediment volume
Gravel beach
ridge
Dunes
Lake
Dunes
Foreshore sand
Sea wall protection of built Mining losses of gravel/sand

environment – induces Sand losses due to
scouring of corridor sediment recreational degradation
Gablons to protect Beach ridge
depleted flank height reduced
of built environment due to washover
Terninal scour
accelerated erosion
Limit of funding authority
New
Offset groynes Wave focusing =
deposition
Sea wall Incremental retreat Offshore losses to counter accelerated erosion
shuts off as maximum erosion with storms progressive
sediment point pushed along- terminal
source shore by littoral drift Depleted beaches lead to erosion Dredging
steeper slopes allowing hole
Offshore sediment losses
higher waves to break
due to creation and
closer to the coast
stabilization of an inlet
Protected sediment sources Deflected downdrift erosion New net- Sediment transport
Depleted beach leading to net sediment losses sediment corridor and
sediment source and transport corridor source transient sink
Sediment cell distortion due to ‘ad hoc’ coastal protection measures
Figure 21.27 Engineers need to consider how incremental coastal protection measures have contributed to a restruc-
turing of coastal cells. The complexity of distorted cells makes the impact of new protection measures difficult to gauge.
There is an inevitability of progressive sterilisation of the natural system after a century of engineering intervention.
Is it time to consider what may happen if defences were rationalised by realignment?
GFE-21.qxd 6/27/05 4:37 PM Page 599
21.9 Conclusion barriers. Controls on the development of either

type reflect the terrestrial plan-view morphology,
The coastal zone presents major dilemmas for offshore shoaling slope, relative sea-level change,
engineers as intervention generally diminishes the tidal range and sediment supply. Most engineering
sustainability of the natural environment. The use structures have associated effects down-drift of the
of the CZ as a backdrop to major extensions of structure due to controls imposed on longshore
human activity is accelerating, while the require- sediment supply.
ments made of the CZ are increasingly outside of It is impossible to ‘fossilise’ coastal environ-
the coast’s sustainability. In the past, this imbal- ments by engineering design. Good design should
ance was masked by engineering support that treat the coastal zone as a highly mobile and
tried to maintain human occupancy, albeit unsuc- dynamic environment. The absence of man-made
cessfully in terms of supporting naturally func- disturbance to longshore processes is a key to
tioning coastal systems. This lack of success has good coastal engineering. Engineering practice
become readily observed over time as engineering has radically changed over the last two decades.
interventions increasingly failed to resolve symp- The emphasis has shifted from engineering struc-
toms of coastal problems and even generated new tures for direct protection, to using less obtrusive
ones. It is now requisite that engineers understand ‘greener’ methodologies by which beaches can be
the underlying causes of coastal situations that are maintained as protective buffers.
perceived as problems through human usage. This The future will undoubtedly see more proactive
understanding can only be viable within a coastal shoreline realignment with coastal geomorphol-
geomorphological context that emphasises time ogy being the basis for behaviour-orientated pre-
and space controls on the interaction of energy dictions to help plan such realignments. However,
and sediments. Without this understanding most as society has increasingly entrenched its position
coastal engineering intervention will be, to quote along the world’s shoreline over the last two hun-
Kaufman and Pilkey (1979), ‘for naught’. dred years, there will be no viable alternative to
hard engineering structures as the first and only
line of defence in many places, unless society
21.10 Summary comes to terms with moving away from the coast.
Successful engineering intervention in the coastal

zone requires a clear understanding of the time
References
and space scales of causative processes as well as
the nature of geomorphological controls on Aagaard, T., Nielsen, J., Nielsen, N. and Greenwood, B.
coasts. Beaches are controlled by relative sea- (1997) Suspended sediment transport and morpholog-
level change, sediment supply, tidal and wave ical evolution on an intertidal beach. Proc. Coastal
Dynamics ’97. Amer. Soc. Civ. Engr., 824–833.
activity, and basement configuration. They are Barkham, J. P., MacGuire, F. D. and Jones, S. J. (1992)
part of a coastal sediment cell system that com- Sea-Level Rise and the UK. Rept. for Friends of the
bines a source area, a transport corridor (i.e. the Earth, London, 82pp.
beach) and a depositional sink. Source areas and Bowden, R. and Orford, J. D. (1984) Residual sediment
sinks may be either offshore or onshore. Erosion cells on the morphologically irregular coastline of
the Ards Peninsula, Northern Ireland. Proc. Roy.
at one coastal location is invariably related to dep- Irish Acad. 84B, 13–27.
osition elsewhere, so that few coastal engineering Bray, M. J., Carter, D. J. and Hooke, J. M. (1995)
problems can be studied in spatial isolation. Few Littoral cell definition and budgets for central
mid-latitude beaches are currently growing: most southern England. Journal of Coastal Research 11,
show long-term sediment deficits. 381–400.
Brunsden, D. and Jones, D. K. G. (1980) Relative time
Clastic beaches that are a source of engineering scales and formative events in coastal landslip sys-
interest in the mid-latitudes are divided into two tems. Zeitschrift für Geomorphologie 34 Supplement
types: fringing or fixed beaches and free-standing Band 1, 1–15.
GFE-21.qxd 6/27/05 4:37 PM Page 600
Bruun, P. (1962) Sea level rise as a cause of shore ero- Gifford Associated Consultants (1997) Final report.
sion. J. Waterways Harbours Div. 88, 117–130. South Downs Shoreline Management Plan.
Bruun, P. (1983) Review of conditions for uses of the Hayes, M. O. (1967) Relationship between coastal cli-
Bruun rule of erosion. Coast. Engr. 7, 77–89. mate and bottom shelf sediment type on the inner
Carter, R. W. G. (1988) Coastal Environments. Academic continental shelf. Mar. Geol. 5, 111–132.
Press, London. Hayes, M. O. (1979) Barrier island morphology as a func-
Carter, R. W. G., Johnston, W. and Orford, J. D. (1984) tion of tidal and wave regime. In Leatherman, S. P.
Formation and significance of stream outlets on (ed.) Barrier Islands. Academic Press, New York,
mixed sand and gravel barrier coasts of southeast 1–27.
Ireland. Zeitschrift für Geomorphologie 28, Horikawa, K. (ed.) (1988) Nearshore Dynamics and
427–442. Coastal Processes. Univ. Tokyo Press, Japan.
Carter, R. W. G. and Wilson, P. (1993) Aeolian Houghton, J. T. (1997) Global Warming (2nd edn).
processes and deposits in north-west Ireland. In Cambridge University Press.
Pye, K. (ed.) The dynamics and environmental con- Jennings, S. C., Orford, J. D., Canti, M., Devoy, R. J.
text of aeolian sedimentary systems. Geol. Soc. N. and Straker, V. (1998) The role of relative sea-
London, Spec. Publ. 72, 173–190. level rise and changing sediment supply on
Cohen, J. E., Small, C., Mellinger, A., Gallup, J. and Holocene gravel barrier development; the example
Sachs, J. (1997) Estimates of coastal populations. of Porlock, Somerset, UK. The Holocene 8,
Science 278, 5341, 1211–1212. 165–181.
Cowell, P. and Thom, B. (1994) Morphodynamics of Kaufman, W. and Pilkey, O. H. Jr. (1979) The Beaches
coastal evolution. In Carter, R. W. G. and Woodroffe, are Moving. Doubleday, New York.
C. D. Coastal Evolution: Late Quaternary Shoreline Komar, P. D. (1998) Beach Processes and
Morphodynamics. Cambridge University Press, Sedimentation (2nd edn). Prentice-Hall. Englewood
33–87. Cliffs, New Jersey.
Davies, J. L. (1972) Geographical Variation in Coastal Kraus, N. C. (1988) The effects of seawalls on the
Development. Oliver and Boyd, Edinburgh. beach: an extended literature review. J. Coastal
Dawson, A. G. (1996) The geological significance of Research, Special Issue No. 4, 1–28.
tsunamis. Zeitschrift für Geomorpologie, Supple- Leatherman, S. P. (1982) Barrier Island Handbook
ment band 102, 199–210. (2nd edn). National Park Service, Boston.
De Vriend, H. J., Latteux, B., Capobianco, M., Leatherman, S. P. (2000) Social and economic cost of
Chesher, T., Stive, M. J. F. and de Swart, H. E. sea level rise. In Douglas, B. C., Kerney, M. S. and
(1993) Long-term modelling of coastal morphol- Leatherman, S. P. (eds) Sea Level Rise: History and
ogy. Coastal Engineering 21, 225–269. Consequences. Inter. Geophysics Series 75,
Dolan, R. and Hayden, B. (1983) Pattern and prediction 181–223.
of shoreline change. In Komar, P. D. (ed.) Coastal Lee, E. M. (2001) Estuaries and coasts: morphological
Erosion Handbook. CRC Press, Boca Raton, adjustments and process domains. In Higgitt, D. L.
Florida, 123–149. and Lee, E. M. (eds) Geomorphological Processes
Doyle, L. J., Sharma, D. C., Hine, A. C., Pilkey, O. H. and Landscape Change: Britain in the Last 1000
Jr., Neal, W. J., Pilkey, O. H. Sr., Martin, D. and Years. Blackwell, London, 147–189.
Belknap, D. F. (1985) The coast, land use and the Lins, H. F. (1980) Patterns and trends of land use and
law. In Doyle, L. J., Sharma, D. C., Hine, A. C., land cover on Atlantic and Gulf coast barrier
Pilkey, O. H. Jr., Neal, W. J., Pilkey, O. H. Sr., islands. US Geol. Survey, Prof. Paper, 1156, 1–12.
Martin, D. and Belknap, D. F. Living With the West MAFF (1995) Shoreline Management Plans: a Guide
Florida Shore. Duke University Press, Ch.7, for Coastal Defence Authorities. UK Ministry of
171–192. Agriculture, Fisheries and Food, London, PB 2197,
FitzGerald, D. M., Penland, S. and Nummedal, D. 24pp.
(1984) Changes in tidal inlet geometry due to back- May, J. P. and Tanner, W. T. (1973) Littoral power gra-
barrier filling: East Friesian islands, West Germany. dient and shoreline changes. In Coates, D. R. (ed.)
Shore and Beach 52, 3–8. Coastal Geomorphology. State University of New
Forbes, D. L., Taylor, R. B., Orford, J. D., Carter, R. W. G. York, Binghampton, New York, 43–60.
and Shaw, J. (1991) Gravel barrier migration and over- McLaren, P. and Bowles, D. (1985) The effects of sed-
stepping. Mar. Geol. 97, 305–313. iment transport on grain-size distributions. Journal
Forbes, D. L., Orford, J. D., Carter, R. W. G., Taylor, R. B., of Sedimentary Petrology 55, 457–470.
Shaw, J. and Jennings, S. C. (1995) Morphodynamic Motyka, J. M. and Brampton, A. H. (1993) Coast
evolution, self-organisation and instability of coarse- Management: Mapping of Littoral Cells. HR
clastic barriers on paraglacial coasts. Marine Geology Wallingford Rept. SR 328, Hydraulics Research,
126, 63–85. Wallingford, 102pp.
GFE-21.qxd 6/27/05 4:37 PM Page 601
Murty, T. S., Flather, R. A. and Henry, R. F. (1986) Pilkey, O. H. and Wright, H. L. (1988) Seawalls versus
The storm surge problem in the Bay of Bengal. beaches. J. Coastal Research, Special Issue No. 4,
Progr. Oceanography 16, 195–233. 41–66.
Nordstrom, K. F., Gares, P. A., Psuty, N. P., Pilkey, O. H. Pugh, D. T. and Faull, H. E. (1983) Tides, surges and
Jr., Neal, W. J. and Pilkey, O. H. Sr. (1986) Living with mean sea-level trends. In Shoreline Protection.
the New Jersey shore. Duke Univ. Press, Durham, Proceedings of the Institution of Civil Engineers,
North Carolina. Thomas Telford, London, 59–70.
O’Brien, M. P. (1969) Equilibrium flow areas of inlets Raper, S. C. B., Cazenave, A., Dawson, A. G., Frezzotti,
on sandy coasts. J. Waterways, Harbours Coasts. M., Long, A. J., Reeh, N., Tooley, M., deWold, J. and
Eng Div. ASCE 95, 43–52. Woodworth, P. (ed.) (2000) Global changes in the
Orford, J. D. (1988) Alternative interpretation of volume and the mass of the ocean. In Smith, D.,
man-induced shoreline changes in Rosslare Bay, Raper, S. C. B., Zerbini, S. and Sanchez-Arcilla, A.
southeast Ireland. Trans. Inst. Brit. Geogr. 13, (eds) Sea Level Change and Coastal Processes:
65–78. Implications for Europe. Luxembourg, Office for
Orford, J. D., Carter, R. W. G., McKenna, J. and Official Publications of the European Communities,
Jennings, S. C. (1995) The relationship between the EUR19337, 9–80.
rate of mesoscale sea-level rise and the retreat rate Reading. H. G. (1996) Sedimentary environments:
of swash-aligned gravel-dominated coastal barriers. processes, facies and stratigraphy (3rd edn).
Marine Geology 124, 177–186. Blackwell Science Publ., Oxford.
Orford, J. D., Carter, R. W. G. and Jennings, S. C. Relson, G. E. (1979) Barrier island systems. In Walker,
(1996) Control domains and morphological phases R. G. (ed.), Facies Models. Geoscience Canada
in gravel-dominated coastal barriers. Journal of Reprint Series 1, 57–74.
Coastal Research 12, 589–605. SCOR (Scientific Committee on Ocean Research)
Orford, J. D., Cooper, J. A. G. and McKenna, J. (1999a) (1991) The response of beaches to sea level
Mesoscale temporal changes to foredunes on Inch changes: a review of predictive models. Working
Spit, south-west Ireland. Zeitschrift für Geomorpho- Group 89, chair: P. D. Komar, J. Coastal Res. 7,
logie N.F. 43, 439–461. 895–921.
Orford, J. D., Cooper, J. A. G., Jackson, D., Malvarez, G. Seymour, R. J. and Castel, D. (1985) Episodicity in
and White, D. (1999b) Extreme storms and thresh- longshore sediment transport. Amer. Soc. Civ. Engr.,
olds on foredune stripping at Inch Spit, south-west J. of Waterway, Port, Coastal and Ocean Engr. 111,
Ireland. Coastal Sediments ’99. Amer. Soc. Civ. 542–551.
Engr. vol. 3, 1852–1866. Sunamura, T. (1975) Static relationship among beach
Orford, J. D., Wilson, P., Wintle, A. G., Knight, J. and slope, sand size and wave properties. Geogr. Rev.
Braley, S. (2000) Holocene coastal dune initiation Japan 48, 485–489.
in Northumberland and Norfolk, eastern UK: cli- Terwindt, J. H. T. and Kroon, A. (1993) Theoretical con-
mate and sea-level changes as possible forcing cepts of parameterization of coastal behaviour. In
agents for dune initiation. In Shennan, I. and List, J. H. (ed.) Large Scale Coastal Behaviour ’93.
Andrews, J. (eds) Holocene Land–Ocean Interac- US Geol. Surv., Open File Reprt. 93–381, 193–196.
tion and Environmental Change around the North Thornes, J. T and Brunsden, D. (1977) Geomorphology
Sea. Geological Society, London, Special and Time. Methuen, London, 208pp.
Publications, 166, 197–217. Trenhaile, A. (1997) Coastal Dynamics and Landforms.
Orford, J. D., Jennings, S. C. and Forbes, D. L. (2001) Clarendon Press, Oxford.
Origin, development, reworking and breakdown of van Rijn, L. C. (1998) Principles of Coastal Morpho-
gravel-dominated coastal barriers in Atlantic logy. Aqua Publications, Amsterdam.
Canada: future scenarios for the British coast. In Wheeler, A. J., Orford, J. D. and Dardis, O. (1999)
Packham, J. R., Randall, R. E., Barnes, R. S. K. and Saltmarsh deposition and its relationship to coastal
Neal, A. (eds) The Ecology and Geomorphology of forcing over the last century on the north-west
Coastal Shingles. Westburg Academic and coast of Ireland. Geologie en Mijnbouw 77,
Scientific Publishing, UK, 23–55. 295–310.
Pethick, J. (1984) Introduction to Coastal Geomorpho- Wijnberger, K. M. and Terwindt, J. H. J. (1995)
logy. Arnold, London. Extracting decadal morphological behaviour from
Pethick, J. (2001) Coastal management and sea level high-resolution, long-term bathymetric surveys
rise. Catena 42, 307–322. along the Holland coast using eigenfunction analy-
Pethick, J. and Crooks, S. (2000) Development of a sis. Mar. Geol. 126, 301–330.
coastal vulnerability index: a geomorphological Wright, L. D. and Short, A. D. (1984) Morphological
perspective. Environmental Conservation 27, variability of surf zones and beaches: a synthesis.
359–367. Mar. Geol. 56, 1–18.
GFE-21.qxd 6/27/05 4:37 PM Page 602
Further reading Kaufman, W. and Pilkey, O. (1979) The Beaches are

Moving. Doubleday, New York.
The following are some standard texts that ident- A stimulating, but partisan account of coastal engi-
ify the range of approaches to coastal environments neering disasters along the USA shoreline. A more
up-to-date statement can be found in Dean, C.
and their management. Clearly some sources are (1999) Against the Tide: The Battle of America’s
dated, but scanning the historical development of Beaches. Columbia University Press, New York,
procedures can tell us about pitfalls, rather than 278pp.
experiencing them directly! Komar, P. D. (1998) Beach Processes and Sedimentation.
(2nd edn). Prentice-Hall. Englewood Cliffs, New Jersey.
Carter, R. W. G. (1988) Coastal Environments. Academic A major authority on approaches to wave mechan-
Press, London. ics and sediment transport. It is less developed
A substantial statement about coastal processes, with respect to longer-term approaches to beach
responses and management. Inevitably somewhat development.
dated but still a sound geomorphological synthesis Komar, P. D. (ed.) (1983) Coastal Erosion Handbook.
of coastal problems. CRC Press, Boca Raton, Florida.
Davies, J. L. (1972) Geographical Variation in Coastal A selection of mainly US-orientated papers con-
Development. Oliver and Boyd, Edinburgh. cerned with the then modern approaches to beach
Still one of the most readable geomorphological and coastal dynamics that still act as the mainstay of
accounts of worldwide variation in coastline devel- today’s approaches.
opments. Silvester, R. (1972) Coastal Engineering. Elsevier,
DEFRA (ex MAFF) http://www.defra.gov.uk; English Amsterdam, 2 vols.
Nature http://www.English-nature.org.uk, and the An engineer’s approach to the understanding of
UK Environment Agency http://www.environment- wave dynamics. A good statement of a theoretical
agency.gov.uk all provide full web sites where issues approach to coastal engineering that has not always
of coastal protection and management in the UK are been successful in the light of empirical evidence.
explored with differing perspectives. For compari- Trenhaile, A. (1997) Coastal Dynamics and Landforms.
son of a US perspective try http://www.ocrm.nos. Clarendon Press, Oxford.
noaa.gov/czm. A good systematic review of coastal environments,
Horikawa, K. (ed.) (1988) Nearshore Dynamics and especially in terms of geomorphology.
Coastal Processes. University of Tokyo Press, Japan. US Army (1977) Shore Protection Manual (3rd edn).
Modern numerical review of processes in the Coastal Eng. Res. Centre, Fort Belvoir, Virginia,
nearshore and on the beach. Pertinent overview of 3 vols.
Japanese contributions to CZ engineering. The standard reference for US coastal engineering
Institution of Civil Engineers (1983) Shoreline design, though now updated and issued as software
Protection. Thomas Telford, London. that can be accessed by means of web site
A useful UK view of shoreline protection of fixed http://www.veritechinc.com.
or fringing beaches, probably somewhat dated now van Rijn, L. C. (1998) Principles of Coastal Morphology.
but a reflection of how coastal defence policy is Aqua Publications, Amsterdam.
changing. A comprehensive approach to coastal morphology
based on numeric and dynamic principles.
GFE-22.qxd 6/27/05 2:32 PM Page 603
22. Continental Shelves

Colin Jago
22.1 Introduction below a thin veneer of modern sediments. The

geotechnical properties of the substrate may
The continental shelves do not constitute a quan- then be dramatically different from the properties
titatively impressive proportion of the oceans. In of the surface. It is essential that the often com-
geological and morphological senses, the shelves plex three-dimensional array of shelf deposits be
are not even part of the ocean basins, being thoroughly explored in any site investigation. The
essentially the drowned margins of the conti- engineer should consider the geological, as well
nents. But in economic and social terms, the as the oceanographic, controls of continental shelf
shelves are the most important parts of the sedimentation.
oceans. They are the primary source of fish and
shellfish, and a major source of hydrocarbons,
hard minerals, sand and gravel. They are sites of 22.2 Morphology of continental
increasing navigation, engineering activity, waste margins
disposal, and recreation.
A major concern of the coastal and offshore The continental margin consists of three seg-
engineer is the nature of the shelf floor — its ments: shelf (of average gradient 7 minutes),
morphology and its sediments — and its slope (average 4⬚) and rise (less than 1.5⬚). The
response to both natural processes (waves and shelf break between shelf and slope lies at an
currents) and human interference (emplacement average water depth of 132 m (Figure 22.1).
of structures). To this end, it is crucial to under- Together these three segments amount to 21%
stand not only the distribution and properties of of the sea surface, but they hold 73% of total
shelf sediments, but also the dynamics of sedi- marine sediments. It has been estimated that
ment entrainment, transport and deposition. The there are 150 ⫻ 106 km3 of sedimentary material
shelf floor is a complex physico-chemico-biolog- on the continental margins; this compares with
ical environment: the shelf hydraulic regime 125 ⫻ 106 km3 on the continents. Given the
drives the sediment system; the sediment flux mineral wealth associated with sediments, the
carries nutrients and contaminants, affects the potential economic value of the margins is clear,
composition of the benthic community and deter- although as yet almost all engineering activity
mines the suitability of the sea floor for building has been confined to the shelves. As a result of
and dumping. However, the response of the shelf tectonic controls (Chapter 4), there is a wide
floor to the hydraulic regime is partly determined physiographic diversity among margins, but a
by chemical and biological processes within the small number of basic categories can be identified
surficial sediments. (Inman and Nordstrom, 1971).
Inasmuch as the shelves are the recently
drowned margins of continents, they carry sedi- Tectonically active margins
ments that were deposited in a range of coastal Tectonically active margins occur at destructive
environments — deltas, estuaries, barriers and plate boundaries and subduction zones, and at
lagoons. These sediments, modified by modern conservative plate boundaries and transform
processes, lie either on the shelf floor or just faults, and are thus subject to volcanism and
603
GFE-22.qxd 6/27/05 2:32 PM Page 604
Tectonically passive margins

Tectonically passive margins are mainly the trailing
edges of continents. Their continental slopes mark
the original points of rifting at constructive plate
boundaries. Most of these margins are now distant
from spreading centres and are seismically passive.
They border coastal plains of subdued relief, and
are characterised by wide shelves, continental
slopes of low gradient and extensive continental
rise wedges of sediment (Figures 22.1 and 22.2).
Although the major mountain belts, and hence
sources of sediment, form on the leading edges of
the continents, some major rivers and sediment
suppliers (e.g. the Amazon) disgorge on the trail-
ing edges, because of the asymmetry of the water-
sheds. As the trailing edges move away from
spreading centres, cooling and isostatic sinking
occur, although subsidence is of the order of
100 mm/1000 years, much less than on active
margins. As a consequence of sedimentation on a
stable sea floor coupled with slow subsidence,
such shelves consist of a seaward-thickening
wedge of sedimentary material. Examples of pas-
sive margins are the east coast of the Americas,
North West Europe (‘Amero’ subtype) and much
of Africa (‘Afro’ subtype). Juvenile forms, where
Figure 22.1 Morphotectonic classification of conti- spreading was initiated in geologically recent
nental margins. After Hayes, 1967a. times, are the Red Sea and Gulf of California
(‘Neo’ subtype). Passive margins are described as
Atlantic-type margins.
seismicity (Chapter 23). They are characterised
by mountainous hinterlands, narrow shelves, Tectonically buttressed margins
steep fault-defined slopes and often deep-sea Tectonically buttressed margins (Figure 22.1) are
trenches (Figures 22.1 and 22.2). They fre- associated with destructive plate boundaries
quently constitute the leading edges of con- behind island arcs. They are less clearly cat-
tinents, and are thus adjacent to mountain belts egorised than the other types, but frequently con-
(e.g. Andes, Rockies). Large numbers of high- sist of wide and shallow shelves (e.g. Yellow Sea,
gradient streams provide sediments (often East China Sea, South China Sea, although the
sandy) to the narrow shelves. A combination of shelf is narrow at Taiwan). These Asian marginal
rapid sedimentation, steep gradients and seismic seas are supplied by two of the major sediment
activity produces an unstable substrate prone to suppliers to the oceans (Huanghe and Changjiang
failure. Tectonic forces give rise to rapid vertical rivers) and sedimentation rates are high. Coupled
movements of the margin; subsidence rates with the seismic activity that results from plate
of 300 cm/1000 years have been measured. subduction, such shelves may be high-risk areas
Examples of active margins are the west coast of from the engineering point of view, with sedi-
the Americas, much of the Mediterranean, South ments liable to liquefaction. Good examples are
East Australia and New Zealand. These are often most of the margins of the Western Pacific and
described as Pacific-type margins. the Gulf of Mexico.
GFE-22.qxd 6/27/05 2:32 PM Page 605
CONTINENTAL SHELVES 605
Figure 22.2 Worldwide distribution of margin types. From Inman and Nordstrom (1971).
Many shelves are cut by deep valleys (submar- a larger than basinal scale (tectonics, climate,
ine canyons) which can transfer large amounts of sea-level change), some of which are restricted to
sediment from the shelf to the rise. Shelf sedi- a particular region (tide- or storm-driven hydro-
ments, moving with a shore-parallel component, dynamics).
tumble into the canyons; after a period of tempor- Those controls which operate on a scale that
ary storage, the sediments are dislodged (by is larger than a localised area of a continental
earthquakes or storm waves) and travel at high shelf are often designated as external controls.
speeds in turbidity currents which carry enor- These controls are tectonics, climate and sea-
mous quantities of sediment to the rise. Large level change.
submarine fans develop at the canyon mouths.
Canyons are especially important in shelf sedi- Tectonics
mentation on Pacific-type margins where the Global tectonics are the fundamental control of
canyon heads can be very close to the shore, shelf sedimentation. Terrigenous sediments are
thereby intercepting longshore drift. The canyon ultimately derived from erosion of continental
slopes are steep, and the sediments deposited on areas, primarily areas of elevated topography, and
them are gravitationally unstable and of low shear mountain building is driven by movement of
strength. Failure of the sediments is most likely lithospheric plates. As a result, the continental
on active margins, and on margins where the rate shelves which receive most terrigenous sediments
of sedimentation is high. are those which are adjacent to young mountain
belts — the faster the rate of uplift, the faster the
rate of supply of sediments to the shoreline and
22.3 Major controls of shelf shelf. So the shelves of South East Asia, fed by
sedimentation rivers from the Himalayas and from mountainous
island arcs, receive large quantities of modern
The nature of shelf sediments and sedimentation sediments while the shelves of North West Europe
is exceedingly varied as a result of the interaction and the Atlantic shelf of North America are sedi-
of a number of processes, some of which occur on ment starved.
GFE-22.qxd 6/27/05 2:32 PM Page 606
Climate sand is increased under climates that have an

Climate exerts a profound influence on the wea- intense dry season rather than a uniformly distrib-
thering of the continents which constitute the uted rainfall. So the sediment supply is high along
fundamental source of supply of sediments coasts with monsoon climates (e.g. South Vietnam,
(Chapter 2). It further determines the mode of ero- India). A combination of climate and mountainous
sion by wind, water or ice of the weathered mate- hinterlands puts the Ganges-Brahmaputra and
rial. Climate accordingly controls the rate and type Mekong among the major sediment carriers.
of terrigenous sediment supply to the shelves and Climate also affects the rate of production and
produces broad Latitude zones of shelf sediment distribution of biogenic, chiefly carbonate, sedi-
types. Polar climates discourage chemical weath- ments on the shelves. Carbonate sedimentation
ering and so production of clay minerals is mini- is encouraged by a low influx of terrigenous
mal; production of rock fragments by mechanical material and high water temperatures (the latter
weathering due to ice action is paramount. Land reduces carbonate solubility — organisms can
ice and floating ice are effective transporters of a then more readily extract carbonate to build their
range of sediment grain sizes, including gravels shells, and the shells of dead organisms dissolve
too coarse for transport by most water currents. more slowly). The temperature control means that
High-latitude shelves are therefore characterised there is an increase in carbonate sedimentation in
by maximum concentrations of gravels and min- low latitudes; furthermore, since the westward-
imum concentrations of muds (Figure 22.3). flowing waters of the equatorial currents are
Tropical rainy climates, however, accelerate the progressively heated before turning polewards at
rate of chemical weathering and the production of the western margins of the oceans, the highest
clay minerals. Low-latitude shelves are therefore temperatures and most abundant carbonates are
areas of maximum mud concentrations. found at these western margins in equatorial
A high proportion of sand and the maximum latitudes (Figure 22.4).
yield of sediment are produced where the mean
annual precipitation is about 300 mm/year. On both
high- and low-latitude shelves, seasonal climatic
variations will give rise to corresponding fluctu-
ations in the rates of sediment supply. The yield of
Figure 22.4 Idealised distribution of shelf sediment

types. White arrows––warm water; dotted arrows––
Figure 22.3 Climate control of shelf sediment distri- upwelling water; black arrows––cold water. From
bution. After Hayes (1967a). Reineck and Singh (1975).
GFE-22.qxd 6/27/05 2:32 PM Page 607
On trailing-edge shelves in low latitudes, car- geographical disparities in this flux: the shelves of
bonates are deposited as the initial shoulder of South East Asia receive large quantities of terrigen-
the rifting coast (e.g. Red Sea). Off-lying banks, ous sediment while the shelves of North West
born early in the rifting process, drift close to, Europe and the Atlantic shelf of North America are
but detached from, the margin; continuous depos- starved of sediment. Although comparable with
ition of carbonates on the banks perpetuates respect to their dominant tidal regimes, the Yellow
shallow-water conditions despite the subsidence and North Seas are remarkably different in sedi-
that results from spreading (e.g. the Bahamas). mentary regimes because of gross differences in
Peninsulas and adjacent shelves distant from sediment supply (⬎ 2000 ⫻ 106 and ⬎ 10 ⫻ 106
terrigenous sources have also maintained a long tonnes/year, respectively). The well-supplied
history of carbonate deposition (e.g. Florida shelves, rather than the starved shelves, present
shelf, Yucatan shelf). Carbonates also occur on most engineering problems since they are floored
some mid-latitude shelves where the terrigenous with rapidly deposited, under-consolidated muds
supply is low (e.g. off southern Australia, west- prone to failure when dynamically loaded by
ern Ireland, western Scotland). The nature of waves, earthquakes or man-made structures.
carbonate sediments is often largely determined Tectonics and sea-level change determine the size
by the supplying organisms rather than by the and morphology of continental shelves. Shelves
hydraulic regime — coarse-grained carbonates are open (e.g. Atlantic and Pacific shelves of
can occur on low energy shelves. North America) or semi-enclosed (e.g. North Sea,
Yellow Sea). Of the major sediment suppliers, only
Sea-level change the Ganges, Brahmaputra, Indus and Amazon
Eustatic sea-level changes occur in response to supply open ocean shelves.
climatic variations (e.g. glaciations) and to tec-
tonic processes (e.g. sea-floor spreading and the
formation of mid-ocean ridges). Local changes 22.4 Shelf hydrodynamic regime
can occur as a result of isostatic adjustments (e.g.
because of accumulation and subsequent melting Shelf sediments are subjected to various water
of ice sheets) and tectonic uplift or downwarping. currents acting alone or in concert (Figure 22.5).
Such changes obviously control the position of Of these, meteorological currents and tidal cur-
the shoreline; most of the continental shelves were rents are the most important on most shelves, but
exposed to sub-aerial processes at various times others, especially oceanic currents, can affect
during the Pleistocene (Chapter 6) when a complex some shelves.
mosaic of glacial, fluvial and lacustrine sediments
was deposited on them. Sea-level changes, espe- Oceanic currents
cially when rapid, alter river-mouth morphology Differential solar illumination gives rise to a heat
and equilibration which in turn control the supply flow from the equator to the poles, accomplished
of terrigenous sediment to the shelf. Thus modern by shallow wind-driven currents and deep thermo-
river mouths are predominantly estuarine, because haline currents. The major geostrophic currents
of the Flandrian transgression, and are effective hug the continental margins, flowing mostly paral-
sediment traps (Chapter 20). Much of the shelf lel to the shelf break. These currents undoubtedly
is therefore deprived of a supply of terrigenous deposit sediments on the continental rise — thinly
material. Sea level determines the water depth and bedded fine-grained contourites — but they can
hence available hydraulic energy at the sea bed, also intrude onto the shelf as a result of eddy shed-
and many shelf sediments are now at depths too ding and large-scale meandering. The resulting
great to be frequently affected by modern shelf flows can transport suspended sediment and
processes. sometimes cause scour (e.g. the Gulf Stream over
The foregoing factors combine to determine the the Blake Plateau on the US Atlantic shelf. The
flux of sediment from land to shelf. There are gross Canaries Current generates megaripples on the
GFE-22.qxd 6/27/05 2:32 PM Page 608
Figure 22.5 Summary of the main physical processess of the shelf hydraulic regime. Redrawn from Swift
et al., 1971.
outer Saharan shelf, and the Agulhas Current diameter equal to the wave height at the surface
produces spectacular bed forms (sand ribbons, but this diameter decreases with depth. At depths
sand waves) on the South East African shelf. The greater than about half the length of the wave,
Agulhas Current flows at up to 2.5 m/s at the sur- the movement of water particles ceases. At shal-
face, but oceanic currents usually flow at lower lower depths, the particle motion extends to the
velocities. The slower flows may not in themselves sea floor, which then modifies the orbits from
be competent to transport much sediment, but they circular to elliptical and absorbs energy through
may be significant when superimposed on other friction. The orbits become increasingly ellip-
shelf currents. tical with depth until they are linear oscillations
at the sea floor of 10 cm/s or more. These oscil-
Meteorological currents lations constitute wave surge currents which may
Meteorological forces generate a range of cur- have sufficient energy to entrain sediment. But
rents that become the paramount energy input on wave surge has little direct effect on sediment
some shelves. Wind drift currents arise from shear transport since the movement is only back and
stresses at the air–sea boundary. Such currents forth.
deepen with time, as turbulent mixing induces a However, sediments set into motion by wave
transfer of momentum. The subsurface currents surge may then be moved by residual components
are not, however, parallel to the wind direction, of the velocity field which by themselves would
since the Coriolis effect and the Ekman spiral be below the sediment movement threshold.
cause a deviation of the surface and subsurface Moreover, percolation of water through the
flows respectively. Surface wind drift current porous sediment induced by the pressure field of
velocities are some 2–5% of the surface wind passing storm wave crests and troughs induces
velocities. Bottom currents so produced are much cyclic loading, liquefaction, and loss of sediment
less, though flows up to 80 cm/s have been strength. This may lead to the collapse of engin-
recorded at depths of 50–80 m. eering structures placed on the sea floor.
Currents are also produced by waves whose In shallow water, the surge currents exhibit
motions are accompanied by an orbital motion time–velocity asymmetry; the shoreward oscillation
of water particles. The circular orbits have a is stronger and shorter than the seaward, and the
GFE-22.qxd 6/27/05 2:32 PM Page 609
water particle orbits no longer close. This residual is On all shelves, wave energy is attenuated as the
the wave drift current which produces a net shore- waves approach land and water depth decreases;
ward movement of sediment on the shore face the wider and shallower the shelf, the greater the
under some conditions. attenuation. Wave attenuation is therefore greatest
The capability of waves to entrain sediment on on Atlantic-type margins, least on Pacific-type.
the shelf floor depends on their energy (Figure Thus the energy loss across the wide shelf of the
22.6). The largest waves are generated in mid to South East United States is up to 85%, compared
high latitudes by winds associated with mid- to a loss across the narrow shelf of South East
latitude cyclones. These waves travel to the north- Australia of only 4%.
east till they break in storm-wave and west-coast Storm surge, due to reduced barometric pres-
swell environments. They also move along some sure or high wind stress, produces a long-period
east-coast shelves where they are augmented oscillation of water level that may not only flood
by local summer sea-breeze waves. The average coastal regions but may also induce sediment
annual wave energy input is therefore high on transport on the shelf. Currents of 1 m/s, both
storm-wave shelves and west-coast swell shelves onshore and offshore, can be generated by hurri-
where wave surge currents augmenting other cur- canes and the resulting transport of sediments,
rents can stir sea-floor sediments down to water especially offshore, may be intense. Storm surges
depths of 200 m or more, rather less on east-coast are especially spectacular on the trailing edges of
swell shelves. In the subtropics, the trade winds continents which have low-lying constructional
and seasonal monsoons create moderate energy coasts.
waves, while in the tropics wave energy further
diminishes (though punctuated by occasional Tidal currents
hurricanes). Polar shelves are protected from Tides, the result of gravitational attractions
waves by ice; shelves in semi-enclosed seas are between Moon and Earth, and Sun and Earth,
protected by the surrounding land masses and occur as a wave of long period propagates on to
limited fetch. the shelf from the adjacent ocean. The tidal wave
Figure 22.6 Wave energy input to the continental shelves.

GFE-22.qxd 6/27/05 2:32 PM Page 610
represents an equilibrium attained by the sea sur- 12.58 hours, close to the lunar semidiurnal period
face between the centripetal and gravitational of 12.42 hours.
forces of the Earth–Moon system and, to a lesser The principal tidal components are reversing
extent, the Earth–Sun system. This equilibrium so that, in the first approximation, there is no net
gives rise to spring tides, when the gravitational flow of water — but there are invariably residu-
forces of the two systems are additive, and to neap als caused by basin topography. In coastal
tides when they are in opposition: this is a fort- waters, a time–velocity asymmetry develops
nightly cycle. Further variations arise from the because of the shoaling effect of the sea bed and
eccentricity of the lunar orbit (the highest tides constriction of the tidal wave by the enclosing
occurring when the perigee falls at either new or coastal regions. Fast flows of 1–2 m/s can then
full Moon), and from variations in the Earth–Sun be generated. The resulting net transport of
distance in summer and winter. water may be accompanied by a net transport of
The tide is much modified by the distribution sediment. On the open shelf and in wide bays,
of the continents and the shape of the ocean the Coriolis effect brings about a constant
basins. If a sea is semi-enclosed with only a small change of direction of the tidal currents which
connection to the open ocean, propagation of the are accordingly rotary. Close to shore, the tide
tidal wave is hampered; the sea is microtidal, like becomes strongly elliptical to rectilinear.
the Mediterranean (Figure 22.7). If the length of Macrotidal shelves are high-energy shelves in
the basin is such that its natural period is close to terms of sediment transport since sediment
that of the principal tide-producing force, then entrainment can occur twice daily.
resonance occurs and a large tidal range and
strong tidal currents result, as in the Southern Density currents
Bight of the North Sea. Such macrotidal shelves Density currents occur in response to differential
are to be found in the Gulf of Korea, the Gulf of densities of sea water of varying salinity, tempera-
California, the Arabian Gulf, the Gulf of Maine ture or suspended sediment load. The density strati-
and elsewhere. The highest tides of all are in the fication characteristic of estuarine circulation
Bay of Fundy (Canada) whose resonance period is (Chapter 20) can continue across the shelf, with
Figure 22.7 Tidal ranges: microtidal (< 2 m), mesotidal (2–4 m) and macrotidal (> 4 m). After Davies (1964).
GFE-22.qxd 6/27/05 2:32 PM Page 611
surface water flowing offshore and bottom water Non-cohesive siliciclastic sediments
onshore. The bottom drift is slow and is not gener- The dynamics of non-cohesive sands and gravels
ally able to move bedload, but it will affect sedi- have been extensively researched over past
ment already entrained by other stronger currents. decades. The transport of non-cohesive sediments
Stratification is destroyed by the turbulence from is controlled by inertia forces dependent on grain
high-velocity tidal currents. size. Both the critical (or threshold) conditions for
initial transport and the transport rate can be
parameterised in terms of the bed shear stress
(usually expressed as ␶0 in Pascals), the shear
22.5 Properties of shelf sediments
velocity (expressed as U ∗ in m/s or cm/s, where
U ∗⫽ ␶ / ␳ ), or the flow speed expressed as U1 in
Shelf sediments may be genetically classified as:
m/s or cm/s, the flow speed 1 m above the bed).
The shear velocity is proportional to the flow
1. detrital (i.e. supplied by rivers, coastal erosion, speed. Reasonably accurate estimates (to within a
aeolian activity or glacial activity) factor or two) can be made of the critical condi-
2. biogenic (i.e. whole and fragmented shells; tion for transport using the grain diameter as the
faecal pellets) sole descriptor of the bed sediment for both
3. authigenic (i.e. formed, like phosphorite, unidirectional currents (Miller et al., 1977) and
within a host sediment — typical in areas of oscillatory currents (Komar and Miller, 1975).
upwelling where cold, nutrient-rich water Simple equations of the form:
from depth rises to the near surface at mid-
latitudes on the eastern sides of the oceans:
q ⫽ k [ (U1 ⫺U1c )/U1c ]
n
Figure 22.4) (22.1)
4. volcanic (predominantly at active margins)
5. residual (i.e. from in-situ weathering of rock have been derived (Bagnold, 1963; Hardisty, 1983;
outcrops). Jago and Mahamod, 1999) for sand transport by
unidirectional currents where q is sand transport
While geologists have traditionally classified rate, c indicates critical values for transport, and k
sediments in terms of texture (mainly grain size) and n are dependent on grain diameter and bed
and mineralogy, a more useful classification for shear stress. It is generally agreed that n ⫽ 3 for
oceanographers and engineers, who are concerned bedload transport, increasing to 4 or 5 for total
with the properties and behaviour of sediments, is load transport (which includes suspension). Such
based on sediment cohesion. Non-cohesive sedi- equations are not very accurate but will predict
ments are those in which there is no effective transport to within an order of magnitude. The
cohesion between particles. Sands and gravels are major problem is the non-linear dependence of
non-cohesive; they are composed of silicates (pri- q on U ∗ or U1 so small errors in the latter produce
marily quartz), carbonates (calcite and aragonite) large errors in q.
and rock fragments. Individual grains are, by Further difficulties arise when steady
definition, coarser than 63 ␮m diameter. Cohesive currents are enhanced by waves. Heathershaw
sediments are those in which there is significant (1981) showed that the sand transport rate
cohesion between adjacent particles: muds are increases many fold when even quite small
cohesive. Cohesion is dependent on electrochem- waves are superimposed on steady currents.
ical forces, van der Waals forces and stickiness Waves, with some help from the current, entrain
imparted by organic compounds created by marine the sediment and the current transports it. The
organisms (mostly algae and bacteria). Muds, with combined bed stress is not simply additive
particles finer than 63 ␮m diameter, are composed but involves non-linear interactions. There are
typically of clay minerals (complex alumino- several formulations which consider sediment
silicates), organic matter and water. transport by combined waves and currents
GFE-22.qxd 6/27/05 2:32 PM Page 612
including the following given by Soulsby cemented, varying from slightly cemented to
(1997). strong calcarenite rock. As a result, settling vel-
ocity, critical shear stress and transport rates are not
2.4
 0.018 
0.5
 accurately parameterised by grain diameter alone.
qt ⫽ AsU   U 2 ⫹ U rms2 ⫺U c  Therefore, equations 22.1 and 22.2 are likely to be
  CD  
in error when applied to carbonates.
⫻ (1⫺1.6 tan b ) (22.2) There are also important differences in the geo-
technical properties of siliciclastic and carbonate
where Asb ⫽ [0.005h(d/h)1.2]/[(s⫺1)gd]1.2, Ass ⫽ sediments. Up until the late 1980s, information
[0.012dD*]⫺0.6 and As = Asb + Ass . In equation on the geotechnical properties of carbonates was
22.2, qt is total sand transport rate, U is depth- scarce. However exploitation of the continental
averaged current velocity, Urms is root-mean-square shelf off North West Australia and in the Bass
wave orbital velocity, CD is a drag coefficient due Strait required installation of offshore production
to the current alone, Uc is the critical current vel- platforms in deep carbonate deposits. Major
ocity required for transport, b is the slope of the problems were encountered with the installation of
seabed, h is water depth, d is the median grain piles for the foundations of the North Rankin A
diameter and D* depends on the relative density of platform, with piles penetrating many metres under
the sediment and the kinematic viscosity of the their own weight and generating negligible cap-
water. The term Asb gives bedload transport and acity in either end bearing or skin friction. These
term Ass gives suspended load transport. failures led to important programmes of research
The sand transport rate predicted using an and a dedicated conference on engineering with
equation of the form of equation 22.2 is orders of carbonate sediments (Jewell and Andrews, 1988).
magnitude greater than the rate predicted using a The proceedings of this conference include useful
simple power relationship of the form of equation reviews of the engineering geology of carbonates
22.1. This suggests that sand transport on shelves by Fookes (1988) and their geotechnical properties
is dominated by extreme events and on many by Semple (1988). A second conference followed
shelves most sand transport does occur during in 1999 (Al-Shafei, 1999). Ongoing research has
storms. However, on shelves with large tides, sought to define the mechanics of carbonate
extreme events are less significant: moderate sediments and rocks in relation to critical-state
waves superimposed on average tidal currents frameworks, see for example Coop (1990), Coop
occur frequently, while storm waves superim- and Atkinson (1993), Lagioia and Nova (1995) and
posed on peak spring tidal currents occur rarely, Cuccovilli and Coop (1997).
and both combinations may make comparable Uncemented carbonate sands generally display
contributions to long-term sand transport. much higher critical-state friction angles than
silica sands, around 40⬚ compared with 33⬚ to 35⬚,
Carbonate sediments due to the interlocking of the shelly particles.
Although nominally non-cohesive, carbonate sands However the shear strength decreases with
and gravels differ from their siliciclastic counter- increase in stress more rapidly than for silica
parts in both hydrodynamic and geotechnical sands due to particle crushing. The in-situ void
properties. This is because carbonate grains, which ratio may be very high in comparison with silica
are essentially biological in origin, have irregular sands due to the very angular and plate-like shape
shape and variable effective density, both proper- of the particles and to a lesser extent the presence
ties being determined by the nature of the organism of intra-particle voids. Initial compressibility is
that produced the grain. Carbonate grains are also very high — this is the reason for very low
whole shells, fragments of shells, tests, skeletons, confining stresses (and hence skin friction) being
etc. and these may contain chambers and holes. developed on piles after driving.
Furthermore, due to diagenesis in the upper centi- For slightly cemented materials the effects of
metres of the seabed, carbonate grains are often cementing are difficult to separate from those of
GFE-22.qxd 6/27/05 2:32 PM Page 613
particle breakage. In general, three phases of 0.15 m/s), but deposition from very high concen-
behaviour are observed: an initial elastic phase trations may occur under much faster currents
where the material behaves essentially as a rock with t0 up to 1 Pa. A thin region of quasi-laminar
and the stiffness is related to the degree of cemen- (non-turbulent) flow develops at the seabed if bed
tation rather than the porosity; a yield or ‘destruc- roughness is small and particles that fall into this
turation’ stage when rapid compression occurs at region will be deposited (McCave, 1971). Mud
constant stress and the behaviour changes from deposition generally occurs in low energy parts of
rock-like to soil-like; the final stage when the shelf below wave base but deposition can
the behaviour is the same as for an initially occur under higher energy conditions if there is
uncemented material. sufficient material in suspension.
Once deposited on the bed, the structure of
aggregates changes over time due to compaction
Cohesive sediments and the erosion stress likewise changes over time.
The foregoing equations cannot be applied to The early hypothesis of Krone (1963) that critical
cohesive sediment transport. While non-cohesive erosion stress is controlled by the strength of
sediment transport is dominated by inertia forces, aggregated matter on the surface of the seabed
cohesive sediment transport is controlled by cohe- remains the most plausible physical model. While
sion. Cohesion depends on many factors including a freshly deposited mud with high water content
water content (i.e. compaction), mineralogy, tem- may be readily eroded, a compacted mud will
perature, salinity and organic mucus content. resist erosion; compacted mud may require a
Grain size per se is not important (see also substantially greater erosion stress than sand. A
Chapters 3 and 7). This means that it is difficult to freshly deposited mud may erode at rates greater
parameterise and predict cohesive sediment prop- than 1 mm/hour but a compacted mud, under the
erties and transport. A fundamental property of same excess shear stress, may erode at less than
cohesive sediments is their propensity to form 1 mm/day. Critical erosion conditions and trans-
aggregates — weakly bound flocs of clays, port rate have been shown to be related empir-
organic matter and water — which form in sus- ically to yield strength (Migniot, 1968) and density
pension under the influence of turbulence, differ- (Delo, 1988) but universally-applicable algorithms
ential settling and organic binding. Aggregation have not been developed (Vanoni, 1975; Thorn
varies with the concentration of suspended matter and Parsons, 1980; Mehta et al., 1982; Villaret and
and aggregates are easily broken by turbulent Paulic, 1986, and Mehta, 1988).
shear. Consequently, the size and behaviour of
aggregated matter change rapidly on short time Other geotechnical problems
and length scales and it is difficult to model and Large-scale offshore development, initially on the
predict settling and resuspension fluxes. The diffi- continental shelf but progressively moving into
culty is compounded by biophysical interactions deeper water, has been a driving force behind
particularly during plankton blooms. Blooms give research on many areas of soil mechanics and
rise to phytodetritus which, by combining with geotechnical engineering of sediments. Examples
cohesive matter in the water column, settles rap- have been the behaviour of sediments under
idly to the seabed as a layer of benthic fluff (Jago cyclic loading due to wind and waves, the
et al., 1993). Large aggregates exceeding 600 ␮m undrained behaviour of sand under very large
are common in shelf seas. These have low density platform foundations experiencing short-term
but fast settling velocity, typically faster than fine loading and design and installation methods for
non-cohesive sand grains (settling velocities of very large piles.
5 mm/s are not uncommon: Jago et al., 1993). Following loss and damage involving oil
Deposition of cohesive particles from low platforms in the Gulf of Mexico due to the under-
concentrations occurs under low shear stress water mud slides triggered by Hurricane Camille
conditions, typically when t0 ⬍ 0.1 Pa (u1 ⬍ c. (Sterling and Strohbeck, 1973), attention was also
GFE-22.qxd 6/27/05 2:33 PM Page 614
turned to the potential of wave loading to cause stabilise the host sediment. Others, or the same,
massive slides of very soft sediments on even organisms effect grain agglomeration and floccu-
very shallow slopes following initial work by lation, burrow through and track the surface of the
Henkel (1970). As offshore developments extend sediment and produce faecal pellets; as a result
across the shelf break into deeper water, the prob- the water content of the sediment is increased,
lems of underconsolidated sediments on the rela- shear strength decreased and scour by marine cur-
tively steep gradients of the continental slopes rents accelerated (see Rhoads and Boyer, 1982,
will become paramount. The continental margins for a review of animal–sediment interactions).
are characterised by hydrodynamic and sedi- The magnitude of such biological processes
mentological regimes that differ considerably from should not be underestimated: a typical shelf sedi-
those of the continental shelves: these present geot- ment contains 150–500 g/m2 of biomass in its
echnical problems and engineering scenarios are upper few centimetres and, for example, mussels
beyond the scope of this chapter. in the Wadden Sea generate some 150 000 metric
tonnes of sediment (i.e. faecal pellets) per year.
22.6 Organism–sediment interaction Biological alteration of the sediment’s geotech-
nical properties may become critical on shelves
This is a factor that is too often overlooked by with inherent sediment instability due to earth-
oceanographers and engineers, perhaps because it quakes, storms or steep sea-floor slopes.
is notoriously difficult to quantify. However, the
marine biomass is at a maximum on the shelf 22.7 Patterns of shelf sedimentation
where infauna and epifauna continually modify the
upper 10–15 cm of the substrate (Figures 22.8 and The patterns of sedimentation on modern shelves
22.9). Organism activities alter the strength and have been established in varying measure by
stability of the sediments, while the strength the sea-level oscillations of the recent past, by
and stability of the sediments control the diversity the quantity and quality of sediment supply and
and density of the organisms. by the hydrodynamic regime. Shelves were period-
Benthic organisms may secrete organic films, ically exposed to sub-aerial processes, then crossed
produce dense fields of vertical tubes and form by transgressing shorelines and finally swept by
calcareous crusts, all of which activities may shelf currents and waves. Shelves adjacent to the
Figure 22.8 Organism–sediment interaction. From Webb et al. (1976).

GFE-22.qxd 6/27/05 2:33 PM Page 615
Figure 22.9 Bioturbation (sediment reworking by organisms) in intertidal sands. Primary sedimentary structures
(cross-bedding and plane bedding) progressively destroyed by burrowing tellinids and worms. Note that surface bed
forms and structures may not be affected. Scale in cm. From Jago and Hardistry (1984).
mouths of major sediment-carrying rivers receive is achieved, and the nature of the resulting patterns
large supplies of sediments, while shelves decou- of sedimentation, depend on the rate and type of
pled from any terrestrial source of sediment are sediment supply (primarily whether it is cohesive
effectively sediment starved. There are large tem- or non-cohesive) and the nature of the hydraulic
poral and spatial variabilities of controlling factors, regime (primarily whether it is tide dominated or
so shelf morphologies and sediment distributions storm dominated). A genetic classification of shelf
are complex and variable. sedimentary regimes has been developed, based on
Most shelf sediments are not in equilibrium with these controls, which differentiates muddy and
present-day conditions because they were emplaced sandy shelves, while sandy shelves are differenti-
during periods of low sea level. Many shelves ated into tide- and storm-dominated types.
have sediment distributions that owe more to past
processes than to modern processes. Sediments that
are now in deep water may have been emplaced in 22.8 The shoreface and coastal
glacial, fluvio-glacial, fluvial or coastal environ- zone bypassing
ments. Thus coarse sands occur on outer shelves.
Such sediments are designated as relict (i.e. rem- Under most conditions, the inner shelf hydraulic
nant from a different earlier environment and now regime pins sands against the shoreline and
in disequilibrium with the shelf hydraulic regime). maintains the nearshore sand prism. During major
On high energy shelves, much of this relict material storms, the role of wind-driven currents becomes
has been reworked to form palimpsest sediments important. More sand is entrained because of
(i.e. possessing vestiges of a former environment increased wave energy. Winds drive surface water
but with the imprint of the present environment). shorewards, creating downwelling and offshore
Where the shelf is subjected to a high-energy bottom flow; massive amounts of sand may then
hydraulic regime, the relict sub-aerial and nearshore be transported seawards. This loss of sand may be
morphologies and sediments become progressively greater than can be recovered by the prevailing
modified by modern waves and currents. How this onshore wave drift currents. The seawards escape
GFE-22.qxd 6/27/05 2:33 PM Page 616
V.C. SAND ⴝ 1–2 mm, C.SAND ⴝ 0.5–1.0 mm, M.SAND ⴝ 0.5–0.25 mm,
F.SAND ⴝ 0.25–0.125 mm, V.F. SAND ⴝ 0.063–0.125 mm, C.SILT ⴝ 0.031–0.063 mm
Figure 22.10 Seaward-fining textural gradient due to waves, in Bristol Bay, southern Bering Sea. f ⫽ ⫺log2 D
(where D is grain diameter in mm). After Sharma et al. (1972).
seawards and is deposited as mud, usually beyond

the shoreface where wave surge currents stir the
bottom only rarely. This route becomes dominant
where a river is supplying large amounts of mater-
ial to the coast. Sediment passes through the
littoral energy fence via jets associated with flood
stages of the river flow and ebbing tides. River-
mouth bypassing (Swift, 1976a, b) of modern
sediments produces a mud blanket that steadily
grows towards the shelf break, gradually covering
older sediments; few shelves have yet reached this
equilibrium situation because significant river-
mouth bypassing occurs in only a few areas.
Figure 22.11 Seaward-coarsening textural gradient in The inner shelf is affected at most times by the
tide-dominated Carmarthen Bay, Bristol Channel. f ⫽ passage of fairweather waves and hence by turbu-
⫺log2 D (where D is grain diameter in mm). lence. On most shelves this turbulence prevents
the permanent deposition of fine-grained sedi-
ments, so that the innermost shelf is clothed by
of sand becomes permanent, especially during a sand and muds are deposited in deeper water. If,
marine transgression when sea level is rising. The however, the supply of fine-grained sediment is
shoreface then serves as a zone of sediment very large — as off the Amazon — then muds
bypassing as it erodes and retreats. Such shoreface may prevail even in shallow water. Usually,
bypassing (Swift, 1976a, b) leaves a thin sand though, sands are deposited on the intertidal fore-
sheet in the wake of the transgression. shore and the subtidal shoreface. The shoreface
The fine-grained sediments escape from the consists of an upper part which is subject to a
coasts more easily. Suspended material disperses wave drift current and a lower part which is
GFE-22.qxd 6/27/05 2:33 PM Page 617
Figure 22.12 Sediment-free shoreface, North Yorkshire shelf, North Sea. From Jago (1981).
agitated by wave surge. Depths of 10–15 m mark may be swept free of sand altogether so that a
the division between these two parts, though this wide wave-cut platform extends seawards from a
depends on the rigour of the wave climate. Sand rocky shoreline (Figure 22.12).
in shallow water is pushed shorewards by wave
drift currents and seawards by rip currents, so that
much of the shoreface is the dynamic surface of a 22.9 Muddy shelves
nearshore sand prism (Chapter 21), sand moving
shorewards during fairweather conditions and Muds dominate the shelf sediment regime in
seawards during storms. regions where major rivers discharge large quan-
On many shorefaces, there is a seaward-fining tities of cohesive sediments. The sedimentary
textural gradient in response to seaward dimin- regime of these shelves is autochthonous. The
ishing sea bed shear stresses and to rip current pattern of mud deposition depends on the rate of
fallout (Figure 22.10). However, where tidal cur- supply of sediment. Where this is relatively low,
rents augment wave surge currents, there may be mud deposition occurs on the lower shoreface (typ-
a seaward coarsening of sands that corresponds ically at water depths of 5–20 m) while sands are
to increasing tidal flows in deeper water (Jago, deposited on the upper shoreface. The mud belt is
1981 and Figure 22.11). On inner shelves where replaced by relict or palimpsest sands on the outer
there is a paucity of sand, much of the shoreface shelf. Examples are the North Sea shelf of the
GFE-22.qxd 6/27/05 2:33 PM Page 618
and shelf mud blanket (McCave, 1972). The outer

shelf relict sand blanket escapes mud deposition
because the rate of supply of mud is small relative
to the resuspension capacity of tides, internal tides
and internal waves (Huthnance, 1981).
Where the supply of fine sediment from rivers
is high, the mud blanket progrades towards the
outer shelf. This occurs on parts of the north-west
Gulf of Mexico (Curray, 1960; van Andel, 1960)
and the Oregon shelf (Kulm et al., 1975) of the
USA. On the shelf of the East China Sea, where
there is a large supply of fine sediment from the
Changjiang and Hwanghe, a lens of mud is rapidly
accumulating on the inner shelf, which thins sea-
wards towards the mid shelf where it is replaced
by relict/palimpsest sands. Because of the semi-
enclosed morphology of this shelf, more muds
supplied by the Hwanghe encroach on the outer
shelf (Butenko et al., 1985). The resulting distri-
bution of sediments is essentially an autochthon-
ous regime (modern muds) with autochthonous
windows (relict sands) (Figure 22.13). The inner
shelf mud deposit is accumulating at rates up to
Figure 22.13 Allochthonous sedimentation showing 5cm/year and is dispersed by a southward-flowing
autochthonous sand ‘windows’, East China Sea. From coastal current during the stormy winter months
Nittrouer et al. (1984).
(DeMaster et al., 1985). These deposits have high
water contents and porosities, with average values
Netherlands and Belgium and the inner German of 111% and 72%, respectively; shear strength
Bight (Dorjes et al., 1970), the Adriatic shelf south is as low as 1.5 kPa (Keller and Yincan, 1985). The
of the Po, parts of the north-west Gulf of Mexico maximum mud accumulation rate on the outer
(Curray, 1960) and the Washington shelf of the shelf is 0.3 cm/year (DeMaster et al., 1985).
USA (Nittrouer and Sternberg, 1981) the shelf The most dramatic examples of muddy shelves
north of the Zaire (Congo) river (Giresse and are those adjacent to major fluvial sediment
Kouyoumontzakis, 1973) and the Golfe du Lion sources such as the Amazon shelf (Nittrouer and
adjacent to the Rhone (Jago and Barusseau, 1981). DeMaster, 1986), Ganges–Brahmaputra system
Generally, only the inner shelf sands and shelf (Kuehl et al., 1990) and the Yellow Sea shelf
muds are zones of active sedimentation. Wave adjacent to the Hwanghe river (Alexander et al.,
energy is the most critical factor: in parts of the 1991). For obvious reasons, it is difficult to quan-
German Bight of the North Sea, mud is accumu- tify the sediment loads of large rivers but it is
lating at a rate of 15.5 cm/100 years even though estimated that it is up to 1300 ⫻ 106 tonnes/year
near-bottom tidal currents reach 0.3 m/s, but wave (Meade et al., 1985) for the Amazon, so two orders
effectiveness is low. The seaward extension of the of magnitude greater than the sediment supply to
mud belt depends on the concentration of sus- the North Sea. In the case of the Amazon, more
pended sediment (which determines the deposition than 80% of the drainage basin is low-lying rain-
rate) relative to wave- or current-generated bed forest and accounts for most of the water discharge,
shear stress (which determines the resuspension but more than 80% of the sediment is derived from
rate); the greater the wave energy, the further sea- the Andes; up to 95% of the sediment is silt and
ward the boundary between inner shelf sand prism clay-size material carried in suspension.
GFE-22.qxd 6/27/05 2:33 PM Page 619
In these systems, muds are deposited on the coast. Muds dominate the shelf out to a water
shelf faster than they are resuspended even in depth of 60 m and are then replaced in deeper
shallow water stirred by waves. Aggregation and water by palimpsest sands, relict sands reworked
settling of suspended sediment occurs as the rivers into a transgressive sand sheet as sea level rose
reach the sea so suspended sediment concen- (Nittrouer et al., 1983). The delta is defined by
trations decline seawards. For example, on the topset beds between shore and 40 m isobath, fore-
Amazon shelf surface suspended sediment concen- set beds between 40 m and 60 m isobaths, and thin
trations decrease from 400 mg/l near the coast to bottomset beds in deeper water. The beds are
10 mg/l within 200 km of the river mouth (Curtis defined acoustically by sandy interbeds which
and Legeckis, 1986) though higher concentrations pinch out towards the north-west along the trans-
near the seabed extend further offshore. The port path.
seabed, made up of very fluid muds, effectively Rapid settling of aggregates and frequent
attenuates wave energy and net accumulation resuspension by tidal currents and waves generate
occurs as a mud blanket that extends from the a very turbid layer at the seabed above the mud
shore towards the outer shelf. The muds form large blanket. The bottom turbid layer occurs with
sub-aqueous deltas defined by gently dipping increasing thickness shoreward of the 30 m iso-
topset beds, more steeply inclined foreset beds and bath until it fills the entire water column shore-
thin bottomset beds (Figures 22.14 and 22.15) ward of the 10 m isobath. The bottom turbid layer
which can be identified through high-resolution grades exponentially into a fluid mud on the
seismic studies. Sedimentation rates vary from seabed. Sediment remains on the seabed due to a
small or negligible on the topset and bottomset combination of drag reduction by high suspended
beds to a maximum on the foresets (e.g. 10 cm/year sediment concentrations, yield stress development
and 8 cm/year for the Amazon and Ganges– during periods of slow flow speeds, and shear
Brahmaputra systems, respectively). Since most thickening of the fluid mud (Faas, 1986). Intensely
deposition occurs on the foresets, the deltas remain reworked mud forms a surficial layer up to 2 m
sub-aqueous and prograde seawards over older thick at the 15 m isobath (Kuehl et al., 1986).
transgressive sand sheets. Limited sands occur seaward of the river mouth
Although dominated by fine sediments, these where interlaminated mud and sand is present
shelves are not low energy environments. Thus on (Kuehl et al., 1982; Nittrouer et al., 1983).
the Amazon shelf, the best documented of these Much of the north-westward advection of fine
systems, tides are energetic and current speeds up sediment appears to occur in a belt shoreward
to 2 m/s occur within 2 m of the seabed near the of the 10 m isobath (Nittrouer et al., 1986). This
river mouth (Nittrouer et al., 1986). The large dis- gives rise to a 30 km wide and tens of metres thick
charge of freshwater produces an estuarine-like mud deposit on the coastline and inner shelf of the
circulation on the inner shelf, with offshore flows Guianas. The Guiana shoreline has been prograd-
in surface waters and onshore flows in bottom ing seawards at rates up to 250 m/year in associ-
waters. An intruding oceanic current, the South ation with growth of oblique, shoreface mud banks
Equatorial Current, gives rise to the North which are typically 5 m high, 10–20 km wide, and
Brazilian Coastal Current, which sweeps the shelf 50–60 km long (Rine and Ginsburg, 1985).
in a north-westward direction with surface speeds Sediment is eroded from the upstream side and
in excess of 0.75 m/s. Combination of this current deposited on the downstream side of the mud-
and tidal currents produces net advection towards banks resulting in a migration rate of 1.5 km/year.
the north-west with superimposed onshore– The downstream muds are exceedingly fluid with
offshore tidal excursions (Nittrouer et al., 1986). bulk densities of 1.03–1.22 g/cm3 and concentra-
A combination of the buoyancy-forced circulations exceeding 104 mg/litre (Wells and Coleman,
tion, particle aggregation and settling ensures that 1981). These fluid muds attenuate surface gravity
most sediment does not escape from the shelf and wave energy so that, although some resupension
a high concentration belt is maintained along the occurs on the shoreface, the shoreline is protected
GFE-22.qxd 6/27/05 2:33 PM Page 620
Physically Stratified sand
Interbedded mud and sand
Proximal-shelf sandy silt
Faintly laminated mud
Mottled mud
Organic-rich laminae
Figure 22.14 Surficial sediment distribution on the sub-aqueous delta of the Amazon shelf (from Kuehl et al.,
1986).
and is able to prograde. By contrast, the upstream as well as shore-parallel components. High sus-
side of the mudbanks, and inter-mudbank areas, pended sediment concentrations, up to 10 kg/m3
contain overconsolidated muds so that wave during normal conditions, move alongshore due
energy attenuation by the seabed decreases and to tidal currents up to 1 m/s and plunge down the
erosion of the adjacent shoreline occurs. delta front beneath the ambient waters as gravity-
While along-shelf advection of fine sediment driven hyperpycnal underflows (Wright et al.,
is important on the Amazon shelf due to the influ- 1990); storm resuspension of delta front sedi-
ence of shelf edge processes on this open shelf, on ments can increase the suspended sediment con-
a more enclosed shelf, such as the Yellow Sea, centration of underflows by an order of
the dispersal system has important shore-normal magnitude. The paths of the underflows are
GFE-22.qxd 6/27/05 2:33 PM Page 621
Figure 22.15 Surficial sediment distribution in the Yellow Sea (from Alexander et al., 1991).
curved. They decay rapidly due to entrainment of resuspension which give rise to turbid waters and
ambient mass so that rapid deposition occurs on a substrate with a high water content. The inner
the delta front. The geometries of the deltas are shelf muds of the Amazon shelf contain a sparse
controlled by the nature of the dispersal system. benthic community due to the unstable nature of
In the pericontinental Amazon and Ganges– the seabed and poor nutrient supply from the
Brahmaputra systems, since sediment dispersal overlying turbid waters (Aller and Aller, 1986).
follows the dominant shelf circulation along the Allied to intense physical reworking, this results
shelf perpendicular to the progradation direction, in a deposit in which primary sedimentary struc-
the bottomset deposits extend less than 50 km tures prevail. At water depths greater than 30 m,
across the shelf. By contrast, in the epicontinental the influence of bioturbation increases. Along
Yellow Sea, the dispersal system is parallel to the shelf, in the net sediment transport direction,
the progradation direction and both axially sands become less common so the deposits
and laterally extensive bottomset deposits have change from interbedded sands and muds to
developed (Alexander et al., 1991). laminated muds (Kuehl et al., 1986). The sandier
Characteristic features of these muddy sediments near the river mouth have low water
shelves include rapid sedimentation and frequent contents, high shear strength and normal to over
GFE-22.qxd 6/27/05 2:33 PM Page 622
consolidation as opposed to the high water con- tidal currents coincide with large storm waves,
tents, low shear strength and normal to under the resulting sediment transport is intense. With
consolidation of the muds (Faas, 1986). The such a high rate of energy expenditure, morpho-
foreset regions, with seabed gradients up to 1⬚, logical and textural patterns inherited from the
are potential sites for sediment failure although retreating nearshore zone have been effectively
there is no evidence of slumping or mass erased.
movement (Adams et al., 1986). Tide-driven, Although tidal currents are reversing, a sedi-
asymmetric megaripples occur on a sandy shoal ment transport asymmetry develops from several
at the mouth of the Amazon; the outer shelf sands causes: flood and ebb currents that are unequal in
have bedforms but these appear to be generally speed and duration; flood and ebb currents that
moribund. follow mutually evasive transport paths; lag
Most allochthonous muddy shelves, although effects resulting from rotary currents; the
constructional in nature, tend to be smooth and enhancement of the flood or the ebb currents by
featureless in areas of mud accumulation due to other currents. Well-defined transport paths are
deposition of sediments too fine grained for the the result, with sand streams which diverge from
formation of large-scale bedforms. Such shelves ‘bed load partings’ and flow down a hydraulic
are in many respects sedimentologically and geo- gradient until either the shelf break or a ‘bed load
morphologically less complex than autochtho- convergence’ and sand accumulations are reached
nous shelves, but in engineering terms they are (Figure 22.16).
likely to be more problematical. This is partly due Each stream is defined by a characteristic pro-
to the high rates of sedimentation. Allochthonous gression of bottom morphologies and sediment
shelves on tectonically active margins are prone textures (Figure 22.17). This begins in high
to earthquakes; the combination of rapid sedi- velocity zones (maximum surface currents
mentation and seismic activity produces a geot- exceeding 1.5 m/s) where the sea floor is scoured
echnically difficult substrate. The liquefaction down to bedrock locally and thinly veneered with
potential of the sediments further increases on lags of shells and gravel. Further down the trans-
shelves attacked by storm waves and swept by fast port path, at maximum surface velocities of
tidal currents. 1.3–1.5 m/s, sand ribbons become the dominant
bedform. These are longitudinal bedforms, up to
15 km long and 200 m wide, though usually less
22.10 Tide-dominated sandy than 1 m high. They consist of sands in transit
shelves over a pavement of shell and gravel and they can
be identified using side-scan sonar by their
The most intensely studied shelf in its response to textural contrast with the underlying gravel. They
modern hydraulic processes is the tide-swept shelf originate in secondary flows generated by vari-
around the British Isles (Stride, 1982). Parts of this able turbulence over a substrate of variable rough-
shelf support intense activity by the oil and off- ness (McLean, 1981). Sand ribbon morphology
shore industries, as well as some of the world’s varies somewhat depending on the current veloc-
busiest shipping lanes. Active sediment transport ity and the sand availability. Beyond the sand
on much of the shelf produces scour around ribbons are sand waves where maximum surface
pipelines, platforms and previously buried wrecks, velocities are 0.5–1.0 m/s. Sand waves are
while deposition on shifting shoals presents a transverse bed forms with straight crests and well-
constant hazard to shipping. defined avalanche faces, wavelength 30–500 m,
This shelf is exposed to frequent storms and and height ⬍ 1–20 m. There is disagreement
surges but much more work is done on the sea about whether all sand waves, regardless of their
floor by semidiurnal tidal currents which regu- size, are formed in the same way and some
larly exceed 0.5 m/s over vast areas and locally authorities classify megaripples (which have
2.0 m/s (surface currents). In areas where such wavelengths up to 2 m) as separate bedforms
GFE-22.qxd 6/27/05 2:33 PM Page 623
has outrun the bedload stream. The sands may be

as thick as 10 m, but where they do not continue
into mud, they break up into irregular patches less
than 2 m thick on a gravel substrate. Many trans-
port paths are terminated by mud zones formed
by deposition of the suspensive load in areas
where suspended sediment concentrations are
high and bottom stirring by waves infrequent. The
most extensive mud areas are at water depths
greater than 30 m. Wave effectiveness, rather than
tidal current intensity, is a critical control of mud
deposition (McCave, 1971). Mud is absent from
large areas of the northern North Sea because of
increased wave activity.
The transport paths are predominantly parallel
to the tidal streams and to the coastline, up to
400 km long though usually incomplete. Areas of
bed-load convergence are sites of deposition
marked by fields of sand waves. Areas of bed-
load parting are usually erosive, although some
Figure 22.16 Sand transport paths on the UK con- deposition may occur where there is a local influx
tinental shelf. From Stride (1982). of sand from the shoreline. The transport paths
imply that the shelf is moving toward a state of
equilibrium with its tidal regime.
Bedforms larger than sand ribbons and sand
waves occur in the Southern Bight of the North
Sea (e.g. the Norfolk Banks and Flemish Banks,
Figure 22.18) and in the Celtic Sea. These are
sand ridges or sand banks. In the North Sea they
are up to 40 m high, 6 km wide and 60 km long,
with their crests spaced at 4–12 km. Water depth
between the ridges is 30–50 m but over their
crests only 3–13 m; they clearly present a naviga-
Figure 22.17 Schematic model of sand transport path
down a hydraulic gradient which is matched by the
tional hazard. They have their long axes orien-
textural gradient. In Swift (1976b). tated almost parallel to the direction of the
strongest tidal currents, which suggests that they
are at least in part related to present-day hydraulic
(Belderson et al., 1982) while others view them as conditions. It is not clear however whether they
essentially the same as the larger sand waves were initiated by modern processes or whether
(Allen, 1980). Although sand waves are usually they are modified relict features responding to
considered as actively migrating features modern processes.
responding to the present-day hydraulic regime, The discovery of a slight obliquity (of usually
data on their migration rates are lacking. Next 7⬚–15⬚) of sandbank axes to peak tidal current
down the velocity gradient are sand patches; these direction led to a model for sand-ridge genesis by
are sheets of fine sand and muddy fine sand lack- a rectilinear tidal flow (Huthnance, 1982). This
ing bedforms other than ripples and supporting a gives a residual flow due to the effects of the
varied infauna. They seem to be essentially the earth’s rotation and bottom friction. Given an
product of suspensive transport of sediment that initial bed waviness of small amplitude with
GFE-22.qxd 6/27/05 2:33 PM Page 624
contours making a small angle with the peak cur- where an intruding oceanic current, the Agulhas
rent direction, a series of bed forms should grow Current, has sufficient energy to generate large
in amplitude. Once formed, helical circulations sand waves and sand ribbons in palimpsest car-
may develop in conformity with the ridge axes, bonate sands moving over a gravel pavement.
and these may assist in building up the banks fur- Field et al. (1981) identified large sand waves
ther (Figure 22.19). The establishment of mutu- formed by a steady northerly coastal current on
ally evasive flood and ebb channel systems could the northern Bering Sea shelf.
then aid the growth and proliferation of sand
ridges (Caston, 1972; Figure 22.20).
It was at one time thought that the lateral 22.11 Wave and storm-dominated
sequences of bedforms seen in the North Sea sandy shelves
were diagnostic of a tide-driven dispersal system.
But Flemming (1980) identified very similar Wave dominance per se is generally confined to
sequences on the shelf of South East Africa the shoreface during fair-weather conditions but
Figure 22.18 Bathymetry, erosional features and constructional features of the north-west European shelf.
GFE-22.qxd 6/27/05 2:33 PM Page 625
Figure 22.19 Schematic representation of helical flow

structure believed to be associated with sand ridges.
Figure 22.20 Model for growth and development of
linear tidal sand ridges. (A) Development of mutually
evasive ebb and flood channels. (B, C) Inequality of
can affect the whole shelf during storms. Wave secondary cross-shoal currents leads to destruction of
straight crest line. (D) Double curve develops into incipi-
action on the sandy inner shelf of the Golfe du ent ebb and flood channels. (E) Lengthening of channels
Lion in the western Mediterranean generates a so that centre ridge becomes parallel to adjacent ridges.
strongly defined seaward-fining textural gradient (F) Continuation of cycle. From Caston (1972).
between the beach and the start of the shelf mud
blanket (Jago and Barusseau, 1981). Such a tex-
tural gradient also occurs on the sandy, open shelf
of the Bering Sea (Figure 22.10) but here it Swift, 1976a, b) but comparable conditions occur
extends into deeper water due to the more intense elsewhere — for example the Gulf of Mexico
wave action; note the absence of muddy sedi- (Hayes, 1967b) and the Bering Sea (Nelson,
ments on this high latitude shelf (Sharma et al., 1982). Under fair-weather conditions, the prevail-
1972). On sandy shorefaces, waves generate a ing movement of shoreface sand is probably
shore-normal sequence of bedforms which reflect onshore, but during storms a downwelling jet can
the decreasing bed shear stresses seawards of the move large quantities of sand offshore. Storm
breaker zone. On high energy shorefaces of open wind stress leads to storm set-up on the coast
shelves, the bedforms include asymmetric lunate which in turn gives rise to a compensatory return
dunes and upper flow regime plane beds (Clifton flow at the seabed. This geostrophic flow has a
et al., 1971). On low energy shorefaces, the bed- large shore-parallel component as well as a shore-
forms are small asymmetric ripples. Modified normal component. The shoreface is charac-
shoreface bedforms are exposed on intertidal terised by sand ridges which are oblique to the
foreshores in macrotidal regions (Jago and shoreline (Duane et al., 1972). These form under
Hardisty, 1984). the opposing forces of the wind-driven coastal
The best-documented sandy shelf dominated current (moving sand seawards) and wave drift
by storms is the Middle Atlantic Bight of the currents (moving sand shorewards). The process
North American Atlantic shelf (summarised by of sediment convergence is analogous to the
GFE-22.qxd 6/27/05 2:33 PM Page 626
process which forms sand ridges in tidal seas. basins (e.g. Long Island Sound, the Gulf of
When the storm set-up abates, the offshore flow is Maine).
augmented by a further pulse as the piled-up Shelves of lower relief consist of flat plateau
water relaxes. Under these conditions, fast divided by shelf valleys cut during low stands of
nearshore currents in excess of 1 m/s are gener- the sea. Thus glacier tongues excavated former
ated (Morton, 1981) which can rapidly move river valleys to produce the lochs and fjords of
large quantities of sand (up to 103 tonnes/m per Scotland and Norway, and their seaward exten-
day) over distances of several kilometres. The sions in the Minches, the Sea of Hebrides and the
sand is deposited as a sand sheet across the shelf. Norwegian Trough (Figure 22.18). Glacial melt-
This mechanism is an extreme form of shoreface water, flowing into an island lake located over the
bypassing which allows sand to escape the Fladen Ground of the North Sea, cut long deep
nearshore wave drift. The sand sheet is charac- channels known as tunnel valleys which remain as
terised by small dunes (⬍ 0.3 m high) covered by local basins (e.g. the Buchan Deep and Devil’s
small ripples (Swift et al., 1979). Formed during Hole: Figure 22.18).
winter storms, these tend to be flattened during In addition, there are many examples of
summer. drowned river valleys, surface and subsurface
channels, such as in the English Channel and
the Middle Atlantic Bight of North America,
22.12 Preservation of relict where the valleys terminate in low-stand deltas
features whose fronts are seaward-bulging shelf-edge
scarps (Figure 22.21). Buried river channels
The preservation of relict features is most likely that were formed and filled during the last low
where the influx of detrital sediment is small and stand of sea level occur below the Holocene
the speed of the transgression is rapid. In such sediment sequence on the shelf of the East
cases, coastal sediments enter the shelf system China Sea (Butenko et al., 1985). These consti-
by shoreface bypassing; since these are of local tute a major geological hazard for seafloor
origin, the sedimentary regime is described as engineering such as emplacement of drilling
autochthonous. The preservation potential of platforms.
relict features is low if the detrital sediment Depositional features are also preserved on
influx from rivers is high (so that river-mouth glaciated shelves; in the northern North Sea glacial
bypassing is enhanced) and the speed of trans- moraines were deposited by successive ice sheets
gression is slow; this is an allochthonous regime. and these survive as prominent banks (e.g. the
The proportion of relict sediment decreases, Dogger Bank and the Jutland Bank). Smaller-scale
and of palimpsest sediment increases, if the shelf features seem in general to have been destroyed
is transgressed slowly or exposed to a high-energy by erosive retreat of the shore face, so that most
hydraulic regime. shelves are somewhat featureless when compared
Sub-aerial morphologies have survived on to adjacent coastal plains, unless subsequently
autochthonous shelves. On high-latitude shelves, covered by complex sand ridges that are largely
relief may exceed 200 m, a consequence of fluvial the result of marine processes initiated after the
and glacial erosion of bedrock. A good example is transgression.
the North American Atlantic shelf where the Fall Shelves covered with biogenic carbonates are
Line (the demarcation between piedmont and exceptional in that relict features are more readily
coastal plain provinces) intersects the shoreline at preserved owing to the early lithification of car-
New Jersey and runs across the shelf. Sub-aerial bonate sediments. Thus, on the Yucatan Shelf of
cuesta have been either cut off (e.g. Long Island) the Gulf of Mexico, reefs have developed on the
or drowned (e.g. Georges Banks) by the trans- cemented remnants of low-stand coastal ridges or
gression. The ridges are separated by drowned dunes.
GFE-22.qxd 6/27/05 2:34 PM Page 627
Figure 22.21 (A) Main morphological features of the Middle Atlantic Bight, USA. (B) Schematic profile
across the shelf showing transgressive sheet sand overlying lagooned and pre-Holocene deposits. After Swift
(1976b).
22.13 Preservation of nearshore commemorated by scarps 10 m or more in height

marine patterns (Figure 22.21).
Superimposed on the sand sheet of the USA
As the Flandrian coasts transgressed the shelves, Atlantic shelf are various shoals formed in zones
relict morphologies were modified and coastal of nearshore sand storage associated with river
sand bodies deposited. The most characteristic mouths and headlands. These form shoal-retreat
feature of shelves in an autochthonous regime is massifs (Figure 22.21). The first type, called
a discontinuous sand sheet up to 10 m thick that inlet-associated bars, consist of seaward convex
was laid down during the erosive retreat of the bars at the mouths of major estuaries and tidal
shoreface by a seaward transfer of sand during inlets. They extend seawards for several kilome-
storms. On shelves bordering low, constructional tres. Formed as ebb tide deltas, they were dis-
coasts (such as the USA Atlantic shelf), this sand sected by mutually evasive flood and ebb
sheet overlies back barrier sands and lagoonal channels flaring outwards from the estuary
muds. Pauses in the general rise of sea level mouths, passing seawards into storm-dominated
enabled shorelines to temporarily stabilise, so sand ridges. Coastal retreat during transgression
that shoreface profiles built upwards by means lessened tidal dominance and increased rework-
of upper shoreface and barrier aggradation, ing by storms, so that shoal-retreat massifs remain
rather than retreating landwards. With the reworked with superimposed linear sand ridges
resumption of the transgression, the coasts moulded by unidirectional storm-generated cur-
resumed their migration through shoreface ero- rents (Figure 22.22). The second type are cape-
sion and storm washover; the short hiatuses were associated bars: these are adjacent to prominent
GFE-22.qxd 6/27/05 2:34 PM Page 628
Figure 22.22 Schematic model for evolution of

inlet-associated bars into offshore linear sand ridges
during a marine transgression From Swift et al.
(1972). Figure 22.23 Cape-associated sand bars forming at
sites of littoral drift convergence (sand transport rates
in yd/y ⫻ 10⫺3). From Swift (1976b).
headlands and possess a distinctive hammer-head

shape (Figure 22.23). Common on the shelf south In the Middle Atlantic Bight, shoreface sand
of the Middle Atlantic Bight of North America, ridges were detached from the coast by sea-level
they are the response of the shoreface regime to a rise and then maintained by helical currents
wave climate that includes a high directional vari- (Figure 22.19). It is unlikely that the abundant
ance. Waves approach from both the north-east mid-shelf ridges are simply inherited shoreface
and south-east and cuspate forelands arise by features. Once formed, the ridges are maintained
littoral drift convergence; this creates shoals by the shelf hydrodynamic regime which may even
which then maintain patterns of wave refraction create secondary ridges (Figure 22.24) during
that further drive littoral drift convergence. intense winter storms. Evidence of sand transport
Probably initiated at low-stand deltas, tidal inlets on the ridges is provided by current ripples, sand
or estuary mouths, the cape-associated bars were ribbons and occasionally sand waves.
self-maintaining features throughout the trans-
gression, with subsequent modification by storm-
generated currents which developed superimposed 22.14 Summary
sand ridges and troughs. The shoal-retreat massif
therefore marks the retreat path of the drift con- The pattern of shelf sedimentation and the prop-
vergence. erties of the shelf substrate vary in time and
GFE-22.qxd 6/27/05 2:34 PM Page 629
References
Adams, C. E., Wells, J. T. and Coleman J. M. (1986)
Transverse bedforms on the Amazon shelf.
Continental Shelf Res. 6, 175–187.
Alexander, C. R., DeMaster, D. J. and Nittrouer, C. A.
(1991) Sediment accumulation in a modern
epicontinental-shelf setting: the Yellow Sea. Mar.
Geol. 98, 51–72.
Allen, J. R. L. (1980) Sand waves: a model of origin
and internal structure. Sediment. Geol. 26, 281–328.
Aller, J. Y. and Aller, R. C. (1986) General characteris-
tics of benthic faunas on the Amazon inner conti-
nental shelf with comparison to the shelf of the
Changjiang River, East China Sea. Continental
Shelf Res. 6, 291–310.
Al-Shafei (ed.) (1999) Engineering for Calcareous
Figure 22.24 Evolution of storm-generated sand Sediments. Balkema, Rotterdam.
ridges. From Stubblefield and Swift (1976). Andel, T. H. van (1960) Sources and dispersion of
Holocene sediments, northern Gulf of Mexico. In
Shepard, F. P., Phleger, F. B. and van Andel, T. H.
space. Young mountains are in places providing (eds) Recent Sediments, Northwest Gulf of Meixco.
Amer. Ass.Petroleum Geol., Tulsa, 44–55.
large quantities of terrigenous sediments to the
Bagnold, R. A. (1963) Mechanics of marine sedi-
coast and shelf via rivers carrying large quantities mentation. In Hill, M. N. (ed.) The Sea, 3. Wiley
of suspended fine sediment. However, as a result Interscience, New York, pp. 507–582.
of sea-level changes, many river mouths are now Belderson, R. H., Johnson, M. A. and Kenyon, N. H.
estuarine, trapping sediments from their rivers (1982) Bedforms. In Stride, A. H. (ed.) Offshore
Tidal Sands. Chapman and Hall, London, 27–57.
and poaching sediments from the shelves. Many
Butenko, J., Milliman, J. D. and Yincan, Ye. (1985)
shelves are therefore sediment starved. Most Geomorphology, shallow structure, and geological
shelves have inherited relict morphologies and a hazards in the East China Sea. Continental Shelf
mosaic of sediments deposited in a range of Res. 4, 121–141.
coastal environments. Allochthonous shelves, Caston, V. N. D. (1972) Linear sand banks in the southern
North Sea. Sedimentology 18, 63–78.
dominated by fine cohesive sediments supplied
Clifton, H. E., Hunter, R. E. and Phillips, R. L. (1971)
by rivers, exhibit very different properties and Depositional structures and processes in the non-
dynamics compared to autochthonous shelves, barred high energy nearshore. J. Sedim. Petrol. 41,
characterised by relict and palimpsest non-cohe- 651–670.
sive sands. Muddy shelves, on which sedi- Coop, M. R. (1990) The mechanics of uncemented
carbonate sands. Geotechnique 40, No. 4, 607–626.
mentation rates are high, develop sub-aqueous
Coop, M. R. and Atkinson, J. H. (1993) The mechanics of
deltas which are prograding across the shelf; the cemented carbonate sands. Geotechnique 43, No. 1,
sediments have high water content and low shear 53–67.
strength. Rapid sedimentation on shelves on seis- Cuccovilli, T. and Coop, M. (1997) Yielding and
mically active margins gives rise to a substrate pre-failure deformation of structured sands.
Geotechnique 47, No. 3, 491–508.
with a high liquefaction potential. Reasonably
Curray, J. R. (1960) Sediments and history of Holocene
accurate estimates of the transport of sands can be transgression, continental shelf, northwest Gulf of
made but it is not yet possible to quantify the Mexico. In Shepard, F. P., Phleger, F. B. and can
dynamics of cohesive sediment transport. The Andel, T. H. (eds) Recent Sediments, Northwest
surface veneer of sediments on some shelves may Gulf of Meixco. Amer. Ass. Petroleum Geol., Tulsa,
221–266.
be atypical of the subsurface sediments and it is
Curtis, T. B. and Legeckis, R. V. (1986) Physical obser-
therefore vital that the three-dimensional geome- vations in the plume region of the Amazon river
try of the shelf deposits be thoroughly investi- during peak discharge — 1. Surface variability.
gated prior to any engineering activity. Continental Shelf Res. 6, 31–51.
GFE-22.qxd 6/27/05 2:34 PM Page 630
Davies, J. L. (1964) A morphogenetic approach to world Huthnance, J. M. (1982) On one mechanism forming
shorelines. Z. Geomorph. 8, (Sp. No.) 127–142. linear sand banks. Estuarine Coast. Mar. Sci. 14,
Delo, A. (1988) Estuarine muds manual. HR Wallingford, 79–99.
England, Report No.SR164. Inman, D. L. and Nordstrom, C. E. (1971) On the tectonic
DeMaster, D. J., McKee, B. A., Nittrouer, C. A., and morphologic classification of coasts. J. Geol. 79,
Jiangchu, Q. and Guodong, C. (1985) Rates of sedi- 1–21.
ment accumulation and particle reworking based on Jago, C. F. (1981) Sediment response to waves and cur-
radiochemical measurements from continental shelf rents, North Yorkshire Shelf, North Sea. In Nio, S. D.,
deposits in the East China Sea. Continental Shelf Schuttenhelm R.T.E. and van Weering, T.C.E. (eds),
Res. 4, 143–158. Holocene Marine Sedimentation in the North Sea
Dorjes, J., Gladow, S., Reineck, H.-E. and Singh, I. B. Basin. I.A.S. Spec. Publ. 5, 283–301.
(1970) Sedimentologie und Makrobenthos der Jago, C. F., Bale, A. J., Green, M. O., Howarth, M. J.,
Nordergrunde und der Aussenjade (Nordsee). Jones, S. E., McCave, I. N., Millward, G. E.,
Senckenberg. Mar. 2, 31–59. Morris, A. W., Rowden, A. A. and Williams, J. J.
Duane, D. B., Field, M. E., Meisburger, E. P., (1993) Resuspension processes and seston dynamics.
Swift, D. J. P. and Williams, S. J. (1972) Linear Phil. Trans. Royal Soc. 343 (1669), 475–491.
shoals on the Atlantic inner continental shelf, Jago, C. F. and Barusseau, J.-P. (1981) Sediment
Florida to Long Island. In Swift, D. J. P., Duane, D. B. entrainment on a wave graded shelf, Roussillon,
and Pilkey, O. H. (eds) Shelf Sediment Transport: France. Mar. Geol. 42, 279–299.
Process and Pattern. Dowden, Hutchinson and Jago, C. F. and Hardisty, J. (1984) Sedimentology
Ross, Stroudsburg, 447–499. and morphodynamics of a macrotidal beach,
Faas, R. W. (1986) Mass physical and geotechnical Pendine Sands, SW Wales. Mar. Geol. 60, 123–
properties of surficial sediments and dense nearbed 154.
sediment suspensions on the Amazon continental Jago, C. F. and Mahamod, T. (1999) A total load algo-
shelf. Continental Shelf Res. 6, 189–208. rithm for sand transport by fast steady currents.
Field, M. E., Nelson, C. H., Caccione, D. A. and Estuar. Coast. Shelf Sci. 48, 93–99.
Drake, D. E. (1981) Sand waves on an epicontinen- Jewell, R. J. and Andrews, D. C. (eds) (1998) Proc. Int.
tal shelf: northern Bering Sea. Mar. Geol. 42, Conf. on Calcareous Sediments Perth, Australia,
233–258. March 1988.
Flemming, B. W. (1980) Sand transport and bedform Keller, G. H. and Yincan, Y. (1985) Geotechnical
patterns on the continental shelf between Durban properties of surface and near-surface deposits in
and Port Elizabeth (southeastern African continental the East China Sea. Continental Shelf Res. 4,
margin). Sediment. Geol. 26, 179–205. 159–174.
Fookes, P. G. (1988) The geology of carbonate soils and Komar, P. D. and Miller, M. C. (1975) On the compari-
rocks and their engineering characteristics and son of the threshold of sediment motion under waves
description. Proc. Int. Conf. on Calcareous Sediments and unidirectional currents. J. Sediment. Petrol. 45,
Perth, Australia, March 1988. 362–367.
Giresse, P. and Kouyoumontzakis, G. (1973) Krone, R. B. (1963) A study of rheologic proper-
Cartographie sédimentologique des plateaux con- ties of estuarial sediment. Sanit. engrg. Res. Lab.
tinentals sud du Gabon, du Congo, du Cabinda et du Rept. University of California, Berkeley, 63–8,
Zaire. Cah. ORSTOM, Ser. Geol. 5, 235–257. 91 pp.
Hardisty, J. (1983) An assessment and calibration Kuehl, S. A., Hariu, T. M. and Moore, W. S. (1990)
of formulations for Bagnold’s bedload equation. Shelf sedimentation off the Ganges–Brahmaputra
J. Sediment.Petrol. 53, 1007–1010. river system: evidence for bypassing to the Bengal
Hayes, M. O. (1967a) Relationship between coastal Fan. Geology 17, 1132–1135.
climate and bottom sediment type on the inner con- Kuehl, S. A., Nittrouer, C. A. and DeMaster, D. J.
tinental shelf. Mar. Geol. 5, 111–132. (1982) Modern sediment accumulation and strata
Hayes, M. O. (1967b) Hurricanes as geological agents, formation on the Amazon continental shelf. Mar.
south Texas coast. Bull. Am. Ass. Petrol. Geol. 51, Geol. 49, 279–300.
937–942. Kuehl, S. A., Nittrouer, C. A. and DeMaster, D. J.
Heathershaw, A. D. (1981) Comparisons of measured (1986) Distribution of sedimentary structures in the
and predicted sediment transport rates in tidal cur- Amazon subaqeous delta. Continental Shelf Res. 6,
rents. Mar. Geol. 42, 75–104. 311–336.
Henkel, D. J. (1970) The role of waves in causing sub- Kulm, L. D., Roush, R. C., Harlett, J. C., Neudeck, R. H.,
marine landslides. Geotechnique 20, No. 1, 75–80. Chambers, D. M. and Runge, E. (1975) Oregon
Huthnance, J. M. (1981) Waves and currents near the continental shelf sedimentation: inter-relationships
continental shelf edge. Prog. Oceanog. 10, of facies distribution and sedimentary processes.
193–226. J. Geol. 83, 145–175.
GFE-22.qxd 6/27/05 2:34 PM Page 631
Lagioia, R. and Nova, R. (1995) An experimental and Nittrouer, C. A., DeMaster, D. J., McKee, B. A.,
theoretical study of the behaviour of a calcarenite in Guodong, C. and Jiangchu, Q. (1984) Formation of
triaxial compression. Geotechnique 45, No. 4, sedimentary strata in the East China Sea. In Luo
633–648. Yuru (ed.) Proc. Int. Symp. on Sedimentation on the
McCave, I. N. (1971) Wave-effectiveness at the sea-bed Continental Shelf with Special Reference to the East
and its relationship to bed-forms and deposition of China Sea. Beijing, pp. 696–704.
mud. J. Sedim. Petrol. 41, 89–96. Nittrouer, C. A., Sharara, M. T. and DeMaster, D. J.
McCave, I. N. (1972) Transport and escape of fine (1986) Concentration and flux of suspended sedi-
grained sediment from shelf areas. In Swift, D. J. P., ment on the Amazon continental shelf. Continental
Duane, D. B. and Pilkey, O. H. (eds), Shelf Sediment Shelf Res. 6, 151–174.
Transport: Process and Pattern. Dowden, Hutchinson Reineck, H. E. and Singh, I. B. (1975) Depositional
and Ross, Stroudsburg, 225–248. Sedimentary Environments. Springer-Verlag, Berlin.
McLean, S. R. (1981) The role of non-uniform roughness Rhoads, D. C. and Boyer, L. F. (1982) The effects of
in the formation of sand ribbons. Mar. Geol. 42, marine benthos on physical properties of sediments:
49–74. A successional perspective. In McCall, P. L. and
Meade, R. H., Dunne, T., Richey, J. E., de M.Santos, U. Tevez, M. J. S (eds) Animal–Sediment Relationships.
and Salati, E. (1985) Storage and remobilisation of Plenum, New York, 3–52.
suspended sediment in the lower Amazon river of Rine, J. M. and Ginsburg, R. N. (1985) Deposi-
Brazil. Science 228, 488–490. tional facies of a mud shoreface in Suriname,
Mehta, A. J. (1988) Laboratory studies on cohesive South America — a mud analogue to sandy
sediment deposition and erosion. In Dronkers, J. shallow marine deposits. J. Sediment. Petrol. 55,
and van Leussen, W. (eds) Physical Processes in 633–652.
Estuaries. Springer, Berlin, Heidelberg, New York, Semple, R. (1988) State of the art report on engineer-
pp. 427–445. ing properties of carbonate soils. Proc. Int. Conf. on
Mehta, A. J., Parchure, T. M., Dixit, J. G. and Calcareous Sediments Perth, Australia, March
Ariathurai, R. (1982) Resuspension potential of 1988.
deposited cohesive sediment beds. In Kennedy, V. Sharma, G. D., Naidu, A. S. and Hood, D. W. (1972)
(ed.) Estuarine Comparisons. Academic Press, Bristol Bay: a model contemporary graded shelf.
New York, pp. 591–609. Bull. Am. Ass. Petrol. Geol. 56, 2000–2012.
Migniot, C. (1968) Étude des propriétés physiques de Soulsby, R. L. (1997) Dynamics of marine sands:
différents sediments très fins de leur comportement A manual for practical applications. HR Wallingford
sous des actions hydrodynamiques. Houille blanche Report No. 14, 85 pp.
7, 591–620. Sterling, G. H. and Strohbeck, E. E. (1973) The failure of
Miller, M. C., McCave, I. N. and Komar, P. D. (1977) the South Pass ‘B’ Platform in Hurricane Camille.
Threshold of sediment motion under unidirectional Proc. 5th Offshore Technology Conf. Houston 2,
currents. Sedimentology 24, 507–527. 719–730.
Morton, R. A. (1981) Formation of storm deposits by Stride, A. H. (1982) Offshore Tidal Sands. Chapman
wind-forced currents in the Gulf of Mexico and the and Hall, London.
North Sea. In Nio, S. D., Schüttenhelm, R. T. E., and Stubblefield, W. L. and Swift, D. J. P. (1976) Ridge
van Weering, T. C. E. (eds) Holocene Marine development as revealed by sub-bottom profiles on
Sedimentation in the North Sea Basin. I.A.S. Spec. the central New Jersey shelf. Mar. Geol. 20,
Publ. 5, 385–396. 315–334.
Nelson, C. H. (1982) Modern shallow water graded Swift, D. J. P. (1976a) Coastal sedimentation. In
sand layers from storm surges, Bering shelf: a mimic Stanley, D. J. and Swift, D.J.P. (eds), Marine
of Bouma sequences and turbidite systems. J. Sedim. Sediment Transport and Environmental Management.
Petrol. 52, 537–545. Wiley, New York, 255–310.
Nittrouer, C. A. and DeMaster, D. J. (1986). Sedi- Swift, D. J. P. (1976b) Continental shelf sedimentation.
mentary processes on the Amazon continental shelf: In Stanley, D. J. and Swift, D. J. P. (eds), Marine
past, present and future research. Continental Shelf Sediment Transport and Environmental Management.
Res. 6, 5–30. Wiley, New York, 311–352.
Nittrouer, C. A. and Sternberg, R. W. (1981) The for- Swift, D. J. P., Stanley, D. J. and Curray, J. R. (1971) Relict
mation of sedimentary strata in an allochthonous sediments on continental shelves: a reconsideration.
shelf environment: the Washington continental shelf. J. Geol. 79, 322–346.
Mar. Geol. 42, 210–232. Swift, D. J. P., Freeland, G. L. and Young, R. A. (1979)
Nittrouer, C. A., Sharara, M. T. and DeMaster, D. J. Time and space distribution of megaripples and
(1983) Variations of sediment texture on the associated bedforms, Middle Atlantic Bight, North
Amazon continental shelf. J. Sediment. Petrol. 53, American Atlantic shelf. Sedimentology 26,
179–191. 389–406.
GFE-22.qxd 6/27/05 2:34 PM Page 632
Swift, D. J. P., Kofoed, J. W., Saulsbury, F. P. and Sears, P. various aspects of shelf sedimentation in the East
(1972) Holocene evolution of the shelf surface, China Sea which receives large quantities of fine sed-
central and southern Atlantic shelf of North America. iments from the Changjiang and Hwanghe rivers.
In Swift, D. J. P., Duane, D. B. and Pilkey O. H. (eds) Nio, S. D., Schuttenhelm, R. T. E. and van Weering
Shelf Sediment Transport: Process and Pattern. (eds) (1981) Holocene Marine Sedimentation in the
Dowden, Hutchinson and Ross, Stroudsburg, North Sea Basin. Inter. Ass. Sedimento-logists, Sp.
499–574. Pub. 5. A collection of research papers which covers
Thorn, M. F. C. and Parsons, J. G. (1980) Erosion and important aspects of North Sea sediments.
cohesive sediments in estuaries: an engineering guide. Nittrouer, C. A. (ed.) (1981) Sedimentary Dynamics of
Proc. 3rd Int. Symp. Dredging Technol. Bordeaux, Continental Shelves (Developments in Sedimentology
349–358. 32). Elsevier, Amsterdam. Mostly research papers
Vanoni, V. A. (1975) Sedimentation engineering. ASCE that describe developments in instrumentation, tech-
Task Comm. New York, 745 pp. niques, and scientific rationale.
Villaret, C. and Paulic, M. (1986) Experiments on the Nittrouer, C. A. and Demaster, D. (eds) (1986)
erosion of deposited and placed cohesive sediments Sedimentary Processes on the Amazon Continental
in an annular flume and a rocking flume. Rep. Shelf. Continental Shelf Research 6 (1/2). A collection
UFL/COEL–86/007 Coast. Oceanogr. Eng. Dep. of research papers providing detailed information on
University of Florida, Gainesville. Amazon shelf sediments, properties and dynamics.
Webb, J. E., Dorjes, D. J., Gray, J. S., Hessler, R. R., Reading, H. G. (ed.) (1996) Sedimentary Environ-
van Andel, T. H., Rhoads, D. C., Werner, F., Wolff, T. ments and Facies (3rd edition). Blackwell, Oxford.
and Zijlstra, J. J. (I976) Organism sediment relation- An excellent review of modern (and ancient)
ships. In McCave, I. N. (ed.) The Benthic Boundary sedimentation in both continental and marine
Layer. Plenum, New York, 273–295. environments.
Wells, J. T. and Coleman, J. M. (1981) Physical Stanley, D. J. and Swift, D. J. P. (eds) (1976) Marine
processes and fine grained sediment dynamics, coast Sediment Transport and Environmental Management.
of Surinam, South America. J. Sediment. Petrol. 51, Wiley, New York. A collection of papers on the phys-
1053–1068. ical oceanography and sedimentology of continental
Wright, L. D., Wiseman, W. J., Yang, Z.-S., Bornhold, margins. Includes papers on problems facing the
B. D., Keller, G. H., Prior, D. B. and Suhayda, J. N. offshore engineer.
(1990) Processes of marine dispersal and deposition Stride, A. H. (ed.) (1982) Offshore Tidal Sands.
of suspended silts off the modern mouth of the Chapman and Hall, New York. An excellent synthesis
Huanghe (Yellow River). Continental Shelf Res. 10, of our knowledge of the tidal shelves of north-west
1–40. Europe.
Swift, D. J. P., Duane, D. B. and Pilkey, O. H. (eds) (1972)
Shelf Sediment Transport: Process and Pattern.
Further reading Dowden, Hutchinson and Ross, Stroudsburg.
Collection of research and review papers on sediment
Dalrymple, R. W. (1992) Tidal depositional systems. In dynamics and sedimentation on continental shelves.
(eds) Walker, R. G. and James, N. P. Facies Models: Swift, D. J. P., Oertel, G. F., Tillman, R. W. and Thorne,
Response to Sea Level Change. Geological J. A. (1991) (eds) Shelf Sand and Sandstone Bodies:
Association of Canada, Waterloo, Ontario, 195–218. Geometry, Facies and Sequence Stratigraphy.
New information on sediments in tidal environments. Special Publication of International Association of
Flemming, B. W. and Bartholoma, A. (1995) Tidal Sedimentology 14, 532pp. Blackwell Scientific
Signatures in Modern and Ancient Sediments. Publications, Oxford. A good synthesis of informa-
Special Publication of International Association of tion on sandy sediments in shelf seas.
Sedimentology, 24, 358pp. Blackwell Scientific Walker, R. G. and Plint, A. G. (1992) Wave- and storm-
Publications, Oxford. Diagnostic features of sedi- dominated shallow marine systems. In (eds) Walker,
ment deposits in tidal environments. R. G. and James, N. P. Facies Models: Response to
Milliman, J. D. and Qingming, J. (1985) Sediment Dyna- Sea Level Change. Geological Association of
mics of the Changjiang Estuary and the Adjacent Canada, Waterloo, Ontario, 219–238. A facies
East China Sea. Continental Shelf Research 4 (1/2), approach to shallow marine sediments forced by
251pp. A collection of research papers documenting waves and storms.
GFE-23.qxd 6/27/05 2:12 PM Page 633
23. Volcanic Landscapes

Tony Waltham
23.1 Plate tectonics and volcanoes Viscosity of the parent magmas accounts for the
main contrasts between volcanoes and owes its
The popular concept of a volcano as a tall conical properties to origins in different plate tectonic
mountain, such as Fujiyama, Vesuvius or environments. Plate boundaries are the loci of most
Popacatapetl, belies the variety of terrain styles active, dormant and recently active volcanoes
that are created by different types of volcanic (Figure 23.2), notably over hot spots. Away from
activity. High symmetrical cones are the the active plate margins, and a few intra-plate sites,
conspicuous product of viscous magmas and there are no true volcanic landscapes. Old volcanic
explosive activity (Figure 23.1), but equally rocks lose their shapes and create high ground,
important are the effusive volcanoes that produce merely on account of their resistance to erosion.
fluid and basaltic lavas to build the gentle profiles Volcanic materials fall into four main groups:
of shield volcanoes, typified by the island moun-
tains of Hawaii. Add to these the great sheets of 1. Lava: rock crystallised by direct cooling of a
flood basalts, notably India’s Deccan Traps, then liquid (also known as lava once it emerges
the ignimbrite eruptions of caldera collapses, above ground and therefore ceases to be
epitomised by Yellowstone, and finally the many known as magma), including bubbly scoria and
small scoria cones, to appreciate the full range of frothy pumice.
volcanic landscapes. Each type has its own impli- 2. Tephra: all forms of exploded and partially or
cations for the engineering geomorphologist. totally cooled volcanic debris, including airfall
Figure 23.1 The active andestic volcano of Koryaksky forms a classic cone to 3456 m high on Russia’s Kamchatka
peninsula. In the foreground, a dome of dark andesitic lava rises almost to the rim of the summit crater of Avacha.
633
GFE-23.qxd 6/27/05 2:12 PM Page 634
Figure 23.2 The major volcanic belts of the world in relation to the active plate boundaries. Named volcanoes are
referred to in the text.
ash, cinder cones and tuffs of various grain Table 23.1 A classification of the eight most import-
sizes, and individual volcanic bombs. ant types of igneous rocks (named in the shaded box)
3. Pyroclastic flows: the hot gas and debris mix- related to their magma properties, chemistries, mor-
tures that flow down a volcano’s flanks, and the phologies, eruption styles and appearances. Granite and
basalt are the most abundant rocks as the dominant mor-
deposits left by these.
phologies are determined by the contrasting viscosities
4. Fluvial deposits: mudflows and lahars rich in
of the acidic and basic magmas.
volcanic debris, and jökulhlaup deposits from
subglacial volcanic meltwater floods. Classification Acid Intermediate Basic
The proportions of these materials on any one Lavas rhyolite andesite basalt
volcano depend largely on its explosivity, and Dykes porphyry dolerite
therefore largely on the chemistry of its parent Batholiths granite diorite gabbro
magma (Table 23.1). Magma continent mountain ocean
source chain floor
Plate divergence and basaltic volcanoes Silica content 70% 60% 50%
The crustal construction that takes place on a diver- Viscosity very viscous viscous fluid
gent plate boundary, such as the Mid-Atlantic Dominant batholiths pyroclastics lavas
Ridge (Figure 23.2), is almost entirely the emplace- form
ment of new igneous material derived from the Eruption style violent explosive effu-
uprising mantle convection flow that lies beneath. sive
Mantle fractionation produces basaltic magma that Iron content 3% 8% 10%
Lava colour pale grey black
is low in silica and therefore very fluid. Most of it
is emplaced as dyke swarms that infill tension
fissures and form most of the new lower crust as the volcanoes. Gases escape from the fluid magma
sheeted dolerites and coalesced gabbros of ophio- before pressures can build up and there is, there-
lite sequences. Magma that emerges above its dyke fore, little scope for explosive eruption activity.
feeders becomes the basalt lava of mild and effusive Most of these volcanoes are on mid-ocean floor
GFE-23.qxd 6/27/05 2:12 PM Page 635
VOLCANIC LANDSCAPES 635
Figure 23.3 Mount Etna, Sicily, with an ash plume blown to the east by the wind. The upper slopes of the
volcano are bare rock and tephra, but the lower, slightly more gently graded slopes are cultivated between
many villages and towns. In the foreground, Giardini-Naxos stands on limestone just clear of the basalt.
sites and their morphology has little direct impact and airfall tephra. Together with giant explosions,
on the human world. Where lava builds up above lateral blasts and flank collapses, these eruptions
sea level it creates shield volcanoes over the create an extremely dangerous environment.
divergent plate boundaries; the prime example is Convergent plate boundaries form very narrow
Iceland, which lies above an over-productive hot zones across the Earth’s crust, and these explosive
spot within the divergent boundary. volcanoes are all within them, notably around
Basalt lavas may also be produced where sub- the Pacific Ocean’s ‘Ring of Fire’. The sites of all
ducted ocean floor melts and rises through the the famous and recent seriously destructive
crust on a convergent plate boundary. Mt Etna in eruptions — Krakatoa, Mt St Helens, Vesuvius,
the Mediterranean (Figure 23.3) and Mt Adams in Katmai, Pinatubo, Mt Pele, Santorini and many
the Cascades are two examples of these basaltic more — are landmarks on the convergent plate
volcanoes amid lines of andesitic cones. boundaries (Figure 23.2).
The older convergent boundaries are also the
Plate convergence and explosive
sites of the major granites, which have the most
volcanoes
extensive outcrops of the intrusive igneous rocks
Within the orogenic belt of any convergent plate
exposed by erosion (Chapter 4).
boundary a major process is the downward move-
ment of crustal material. Some is truly destroyed Intra-plate plumes and flood basalts
when it is subducted into the mantle but much is One of the world’s most famous volcanoes,
melted and rises back to the surface. Because Kilauea on Hawaii, is conspicuously remote from
this material originates wholly or partly from con- any plate boundary. This is one of a group of vol-
tinental crust, its composition is normally close to canoes formed over hot spots within the mantle,
that of andesite or dacite. The high silica content where rising thermal plumes have burned holes
ensures a magma viscosity that is high enough to through the overlying plates. Movement of the
minimise gas escape. Close to the surface gas plates has then created lines of volcanoes with the
pressures can rise enormously, and the eruption activity continuing at the trailing end.
activity is typically violent. Lavas are produced, These hot spots are now recognised as repre-
but are normally outweighed by pyroclastic flows senting the trailing tails of massive blobs of hot
GFE-23.qxd 6/27/05 2:12 PM Page 636
magma that have risen slowly through the mantle.

When the main blobs of magma first melted
through the crust, they produced enormous erup-
tions of extremely fluid lava — the flood basalts
that now form massive lava plateaux. None of these
is active at present, perhaps fortunately, as their
impacts spread to world climates and the entire
Earth environment. Plumes within plate boundary
structures may account for centres of greater activ-
ity, notably Iceland within the Mid-Atlantic Ridge.
23.2 Basaltic volcanic landscapes

Eruption styles
Individual eruptions from basaltic volcanoes vary
from purely effusive Hawaiian and Icelandic styles
to regularly and mildly explosive Strombolian
activity (Figure 23.4), both named after their type
localities. More violently explosive events are not
normally associated with fluid basaltic magmas,
though submarine sites can create Surtseyan erup-
tions where steam-driven explosions are caused by
the meeting of water and magma (Table 23.2).
Figure 23.5 Basaltic lava emerges from a small para-

sitic vent on Sicily’s Mount Etna. The paler material
crossing the top of the picture is bright red and moving
at about 0.1 m/s. The dark material in the foreground
forms a splendid pahoehoe flow about 2 m wide, whose
surface was rolled into classic ropy structure. It has
cooled and solidified, as it flowed from the vent
towards the camera the previous day.
Eruptions may be from the summit vents of

volcanoes, but often occur where the magma
finds an exit at lower altitude. Flank eruptions
may be from single parasitic vents or from long
fissure vents. A variation in the Hawaiian erup-
tion style is the lava fountain, where molten lava
is driven by hydrostatic pressure into fountains
over 100 m high, but these rarely last more than a
Figure 23.4 A modest night-time explosion hurls few minutes or hours.
glowing lumps of lava 100 m into the air, and some
fallen blocks remain incandescent as they roll down the
flanking tephra slopes. This is the summit crater of Lava flows
Stromboli, off the north coast of Sicily, the type-site of Basaltic lava that emerges from either a sum-
Strombolian activity. mit crater or a flank vent (Figure 23.5) flows
GFE-23.qxd 6/27/05 2:12 PM Page 637
Table 23.2 Classification of volcanic eruptions (based largely on Francis, 1993). The Volcanic Explositivity Index
(VEI) is a relative scale based on the volume of material ejected into the air (thereby excluding lava volumes), the
scale of the explosions and the height of the eruptive cloud.
Type Style VEI Product Magma
Icelandic Outpouring of lava from 0 Thick flood lava flows basalt

fissure vents building large plateaus
Hawaiian Outpouring of lava from 0 Thin lava flows building basalt
central vent shield volcanoes
Strombolian Periodic mild ejections 1–2 Spatter, cinders and bombs basalt
of pasty lava building small steep
cones
Surtseyan Larger explosions over 2 Tephra and glass building basalt
subaqueous vent flat cone up to water level
Vulcanian Repeated explosions 2–3 Airfall ash, some pyroclastic basalt/andesite
blasting debris flows and lava, building
high in the air composite volcanoes
Vesuvian Strong Vulcanian events 4 More airfall ash and andesite
(term not widely used) pyroclastic flows, building
composite volcanoes
Plinian Continuous powerful jets 4–5 Pyroclastic flows and airfall andesite
blasting out debris ash, building composite
volcanoes
Pelean Magma dome squeezed 5 Magma dome or spine, andesite/rhyolite
out of vent with nuees ardentes and
pyroclastic flows
Krakatoan Multiple giant blasts and 6 Airfall ash and pyroclastic andesite/rhyolite
caldera collapse flows, blast crater or
collapse caldera
Super-volcano Cataclysmic eruptions 7–8 Thick and extensive rhyolite
and giant caldera ignimbrite sheets, collapse
collapse of large calderas
downslope until it cools enough to preclude fur- lava remains hot, continues to flow and may
ther movement as it solidifies. Flow rates may be entrench itself by thermal erosion of the tube
tens or thousands of cubic metres per hour, but floor (Figure 23.6). The gradient on a shield vol-
lava cannot flow far when it is exposed to atmos- cano can allow a tube to drain out at the end of an
pheric cooling. All long flows are tube-fed, in eruption, so that it remains as a cave tunnel down
that new molten lava is supplied from the vent to the core of its host flow. Access to these tubes is
the growth zone through tubes insulated and generally via skylights that never fully roofed
maintained within the older cooled lava. In some over, or via collapse holes that failed subse-
flows, lava streams that are typically 5–20m wide quently.
form levees of abandoned, chilled rock, and then A lava lake can develop within a vent crater,
roof over; normally as crusts grow by accretion, until it is drained either by overflow or back into
levees coalesce and rafted blocks jam against the vent, or cools into the top of a vent plug.
each other. Many flows are inflated by the injec- Equally, lava can pond within a valley or depres-
tion of molten lava beneath an already solid crust sion, and the cooling area of flat lava may col-
(Walker, 1991), when extensive or braided tube lapse or be uplifted as the balance of flow changes
systems develop. Inside its insulated tube, the beneath it.
GFE-23.qxd 6/27/05 2:12 PM Page 638
Figure 23.6 A skylight in an active lava flow from the

Kupianaha vent of Hawaii’s Kilauea volcano. The sky-
light is about 3 m across, and through it the yellow-hot
lava can be seen flowing in a tube over 4 m high and
wide. The surface of the flow is smooth and ropy
pahoehoe, much of which formed as mini-flows downs-
lope of breakouts from the tube when lava filled it
under pressure.
Tephra falls
Lava (and included vent wallrocks) blasted into
the air, and cooled before landing back on the
ground, is collectively known as either tephra
or pyroclastic material; only the fine-grained
component is labelled volcanic ash. Figure 23.7 A shower of basaltic tephra on the town
Basaltic volcanoes commonly have a minor of Zafferana Etnea, on the downwind flank of Sicily’s
component of pyroclastic production. A cascade Mount Etna. This was produced a few hours earlier by
a single fountaining eruption, which lasted just 15 min-
of lava from an active fountain maintains enough utes, from Etna’s summit vent, 12 km away.
heat to weld together when it lands as a pile of
spatter. The same material thrown higher by the
fountain and blown further by the wind may land Eruption hazards
as a fan of loose ash just downwind of the vent. A basaltic eruption presents almost no threat to
Lava ejected in cooler Strombolian explosions human life. A lava flow never advances faster than
may partially weld on landing, but most drapes walking pace, and tephra falls are rarely more
the volcano as a debris slope of loose tephra, iso- than a minor inconvenience (Figure 23.7). Gas is
lated bombs and rolled blocks; this builds up to only an occasional hazard. These eruptions are
form cones as steep as, but smaller than, andesitic generally safe enough to be tourist attractions,
volcanoes. Many eruption events that initially with positive economic impact; Costa Rica’s
produce lava flows progressively lose energy and Arenal has become almost ringed by hotels since
heat, so that a cone of scoria is created in a dying going into steady Strombolian mode (even though
phase of mild explosions as gas escapes from the Arenal’s lava is a basaltic andesite with the poten-
solidifying lava. tial for more violent eruptions that could render
Distal airfall ash is minimal from basaltic the tourist development most inappropriate).
eruptions, which normally lack the explosive Lavas can however totally destroy landscapes
power to blast debris high enough for redistribu- and built structures. Vent production is unstoppable,
tion by the wind. Similarly, gas-rich pyroclastic but there is scope for controlling or diverting
flows are not a component of basaltic eruptions. lava flows that threaten houses and villages. Lava
GFE-23.qxd 6/27/05 2:12 PM Page 639
Figure 23.8 Diversion of a lava flow, in 1983, around the Rifugio Sapienza on the southern slope of Sicily’s Mount
Etna. The dark material is the basaltic lava that flowed across the site a few weeks before the picture was taken. The
bank of lighter material on the right is old lava and tephra bulldozed into place to protect the main Sapienza building,
whose top storey is visible beyond the bank.
flows gravitate to valley floors, but the gentle both on Etna in 1983, had debatable impact on the
topography of shield volcano flanks commonly lava’s progress.
allows scope for diverting lava away from its nat- A special eruption hazard is generated by sub-
ural line. Some basalt lavas are so fluid that they glacial volcanoes where huge volumes of meltwater
can flow through the windows and doorways of a are created over erupting lava. Eventually water
house, but more viscous lavas can push down pressure floats the ice, so that the water escapes
structural walls (Bolt et al., 1975). under the glacier to emerge as a sudden and
Some basaltic lava flows have been success- potentially catastrophic flood perhaps far from
fully managed by engineered diversion methods. the causative volcano. These jökulhlaup are com-
Bulldozed banks of earth and lava debris, built mon in Iceland (where the name means ‘glacier
obliquely across slopes have proved very effective flood’). Comparable melting of summit icecaps
on Hawaii and Etna (Figure 23.8); they cannot on high volcanic cones create lahars, usually with
dam lava flows, but can divert them towards less less water and more debris.
developed land. Permanent barriers have been Predictions of volcanic eruptions rely largely on
proposed above Hilo (below Hawaii’s Mauna comparison with recorded data from previous
Loa), but successes to date have been with smaller events and are therefore reliable for Kilauea, the
banks built immediately in front of wayward world’s most monitored volcano on Hawaii. There
flows. Diversion can also be achieved by cooling and elsewhere, inflation of the volcano and
one flank of a flow, and a massive water-spraying increased numbers of micro-earthquakes, both
operation contributed to keeping the lava out of caused by rising magma inside, are the best indica-
the Icelandic town of Heimay during the 1973 tors of impending eruption. But they can only allow
flank eruption of Helgafell (Williams and Moore, broad predictions of increasing activity with mini-
1973). Attempts at blasting holes in flow levees, mal practical use. Even when an eruption is in
and blocking tubes with dropped concrete blocks, progress, day-to-day predictions remain impossible.
GFE-23.qxd 6/27/05 2:12 PM Page 640
Figure 23.9 Typical structure in pahoehoe lava, with small open tubes in the cores of lava lobes that are now shells of
strong and minimally fractured basalt, with more broken lava between the lobes. Based on a road-cut exposure in Hawaii.
emerges from it. Much of the stretched skin is

rolled into ropy pahoehoe by plastic movement of
the hot lava beneath (Figure 23.5). Successive
pulses of lava build the pahoehoe into a welded
mass with a lenticular structure. Where it lies on
almost flat ground, inflow of molten lava continues
beneath the cooled crust, so that the whole flow
can swell; the interior lava then cools with a mas-
sive structure, which is typically more vesicular
towards its top. A lava pile of pahoehoe can be
uniform and strong, but it can contain open lava
tubes and variable fracture patterns (Figure 23.9).
More viscous lava may continue to flow on a
steeper slope or where it is driven by a high rate of
vent outpouring. It is continually ruptured by shear
within the stressed flow, so that it breaks into aa
(rubble of 100–500 mm size, also known as sco-
ria). The front of an aa flow appears as an advanc-
Figure 23.10 The advancing front of a small aa flow,
about 4 m thick, on Sicily’s Mount Etna. The pale patch in ing rubble pile (Figure 23.10), and the molten
the centre of the picture is red lava, not yet chilled, where interior of the lava advances over this rubble. The
it was exposed by blocks tumbling off the moving face. final cooled structure of the flow is a solid interior
between a top and base of loose or partly welded
aa (Figure 23.11). The resultant lava pile has alter-
Lava landforms nating layers of strong, massive, basalt rock and
There are three main types of small-scale basalt sur- weak, loose, aa scoria. Sequences of both aa and
face morphology — ‘pahoehoe’, ‘aa’ (both terms pahoehoe can be interrupted by weaker horizons
are adopted from the Hawaiian language, so proof weathered soil, which may be lateritic in warm
nounced ‘pa-hoy-hoy’ and ‘are-are’) and ‘pillow’. environments, or tephra or both.
Fluid lava develops a smooth pahoehoe surface Whether basalt cools as pahoehoe or aa depends
as it forms a thin, viscous crust. This may be on the original chemistry, viscosity and gas
stretched until it ruptures as a new pahoehoe toe content, but eruption rate is the key background
GFE-23.qxd 6/27/05 2:12 PM Page 641
Figure 23.11 The internal structure of aa lava flows. The upper flow has a sheared but relatively solid core between a
rubble top and a rubble base, and is advancing over an older flow with a core that has poor columnar jointing.
flows develop into aa, as they cool downflow

while movement is maintained down a hillside,
but aa rarely reverts to pahoehoe. The surfaces of
some large basalt flows can break up into blocks
larger than those of aa, thereby approaching the
morphology of blocky andesites.
Pillow lavas are a larger scale of pahoehoe toes
caused by the stretching of the lava crusts cooled
during underwater eruptions. Individual pillows are
less than a metre across, and a mass of pillow lavas
has a bulbous lenticular structure (Figure 23.12).
Shield volcanoes
The low slopes necessary to maintain flows in
fluid basalt lava ensure that basaltic volcanoes
typically develop with the low profile of a shield
(Figure 23.13); Mauna Loa is the type example
Figure 23.12 Pillow lavas exposed just north of San (Figure 23.14), with slopes nowhere steeper than
Francisco, USA. The pillows create a crude bedding, about 8⬚. Only a slight increase in lava viscosity,
which dips to the right due to subsequent folding. and the proportion of ejected tephra and spatter,
creates the steeper profile of a very flat cone. Etna
is a fine example of a composite volcano with its
factor. A flow rate of ⬎ 5–10 m3/s generally steeper summit standing on a flatter shield base.
forms aa, while lower effusion rates permit Large shields are generally dotted with cones and
flow without internal shear and thereby create craters on parasitic vents: Hallasan, on Korea’s
pahoehoe (Rowland and Walker, 1990). Within a Cheju Island has 360 parasitic vents.
single flow, ground slope and degree of cooling In a sub-glacial eruption, lava is chilled and
influence the lava style. Many pahoehoe lava broken into glassy fragments by steam explosions
GFE-23.qxd 6/27/05 2:12 PM Page 642
Figure 23.13 Block diagram showing various landforms associated with basaltic volcanoes.
Figure 23.14 The basaltic shield volcano of Mauna Loa, on Hawaii’s Big Island. The darker lava flows are
the more recent, and there is almost no tephra on it. The paler vegetated area in the foreground is the moss-covered
weathered slope of the older Mauna Kea shield, with many small cinder cones that were formed in the dying stages
of parasitic events.
GFE-23.qxd 6/27/05 2:13 PM Page 643
Figure 23.15 The tuya of Blafjall, a Pleistocene sub-glacial volcano just south of Myvatn, Iceland.
on contact with ice and meltwater. This accu- Proximal ejecta from individual vents, central
mulates within a glacial lake of its own making, or parasitic, create symmetrical cones of loose
and is eventually capped by lava when the pile rock, known as tephra, scoria or cinder cones
of ejecta rises above lake level. The lava cap (Figure 23.16). Typically, these have side slopes
creates the table mountain profile of the tuya, of about 33⬚, and most are produced in single
in place of a shield; tuyas have been exposed eruptions that last just a few days or weeks. More
by post-Pleistocene ice retreat in Iceland (Figure gassy eruptions leave larger craters within the
23.15). cones, in extreme cases creating open tuff rings
(Figure 23.13), of which Hawaii’s Diamond Head
Vent landforms is the classic example appreciated best from
Volcanic craters are either excavated as debris is above by passengers flying into Honolulu. Spatter
blasted out from the vent, or are created as ejecta cones are smaller features with steeper slopes in
accumulates around the vent. Summit craters are the welded spatter (Figure 23.16).
commonly blast features, maybe hundreds of Many vents are finally filled by solidified lava.
metres deep, but they may be left as lava drains This forms very strong plugs in the cores of the
back to a lower outlet; all of them suffer subse- extinct volcanoes which are frequently exhumed
quent wall collapse that ultimately leaves them as by differential erosion to create rocky crags or
debris-floored bowls. towers with steep sides (Figures 23.16 and 23.17).
Figure 23.16 Cross-section through contrasting landforms developed on minor vents of basaltic volcanoes. Each
feature can be of any height between a few metres and a few hundred metres. The composite cone is similar to a
small andesitic strato-volcano except that its slopes are less steep due to the fluidity of its lava. Based on coastal and
quarried exposures in Sicily, Costa Rica, Idaho and Arizona.
GFE-23.qxd 6/27/05 2:13 PM Page 644
Plateau basalts
Continental flood basalts represent enormous out-
pourings of lava. Single events may last for many
months and total thousands of cubic kilometres.
The source vents are fissures, commonly associ-
ated with plate rifting, and the lava is so fluid that
it spreads into almost horizontal sheets (Self
et al., 1996). The lava first spreads out as a thin
sheet, which develops a chilled crust and then
grows by sheet flow and injection beneath this
protective insulation. Individual flows may inflate
to over 80 m thick. The hot buried lava is
extremely fluid, and large flow sheets are almost
horizontal; their cores do not drain out, and they
contain no tubes. None has been active in
recorded history, and Tertiary examples have been
stripped by erosion to stand as huge plateaus,
notably the Deccan of western India and in the
Columbia River basin, USA (Figure 23.18).
During the Laki eruption of 1783 in Iceland,
Figure 23.17 Mount Popa, in Burma, a volcanic plug flows were partly tube-fed on sloping ground, but
exhumed by erosive removal of its original tephra cone, they spread out and inflated on level ground to
and now standing high above the Ayerwadi lowlands
with a group of temples on its summit.
create a small-scale example of flood basalt.
Figure 23.18 The thick sequence of flood basalts of the Columbia River basin form the cliffs of Cove Palisades
in Oregon, USA. The skyline cone of Mt Jefferson is not related, as it is one of the andesitic volcanoes of the
Cascades.
GFE-23.qxd 6/27/05 2:13 PM Page 645
Figure 23.20 Basalt columns, each about 600 mm

across, forming the Devil’s Postpile in California. The
flow is not extensive, but the columns are very well
developed, and their fallen neighbours form the debris
slope below the cliff.
Figure 23.19 Profile through a typical sheet of flood

length by concave or convex joints that are also
basalt, with a total thickness of 10–40 m.
formed by cooling contraction.
Engineering implications
The thick, horizontal flows of the plateau lavas Intact basalt is a very strong rock with a
contain some of the finest of the columnar joint- compressive strength normally ⬎ 200 MPa. It com-
ing that is so well known in basalt (Figure 23.19). monly has a fracture spacing of ⬍ 1 m, and its rock
No two flows are the same, but most have an inte- mass strength is typically less than that of more
rior entablature that is either massive or with massive granite. Basaltic lava can form strong
curved fractures and poor columns (Figure ground with a safe bearing capacity of 6–10 MPa,
23.20). The columns grow by extension of polyg- except where horizons of scoria or open tubes cre-
onal contraction fractures as a cooling front ate structural weakness. A sequence of aa flows,
migrates from the base of the flow, or from its top. each with a rubble top and a solid core, creates
Some flows are entirely columnar with two sets of ground that is notoriously difficult to excavate or
unequally sized columns meeting in the flow core. tunnel through (Figure 23.21). Excavation of a
Each column is terminated and broken across its water-collection tunnel into the flank of Mt Etna
GFE-23.qxd 6/27/05 2:13 PM Page 646
Figure 23.21 Very mixed ground exposed in a road cutting 5 m high on Hawaii. The rock is all basalt, but the solid
cores of two lava flows are separated by partly weathered aa rubble, and the thin lower flow has a small open tube
with only a thin roof of solid rock.
Figure 23.22 The extent of the open lava tubes in two basalt flows: the complex branching system of lava caves
at Cuernavaca, Mexico; and the single very long tube of Kazumura Cave in an old lava from Kilauea, Hawaii.
GFE-23.qxd 6/27/05 2:13 PM Page 647
Figure 23.23 A lava tube in Korea with a main road safely standing on a thin basalt roof adjacent to two small
natural collapses (from Waltham and Park, 2002).
was abandoned when blasting of the solid basalt branches (Figure 23.22). The special danger of lava
lava cores caused unmanageable overbreak and tubes is that many lie close to the surface, beneath
drill-hole collapse in the rubbly lava of the flow tops thin crusts with unknown fracture patterns; injec-
and bases. Outcrops of pahoehoe basalts can have tion tubes can have thick and very massive roofs,
well-defined platy structure, and may include areas while tubes formed by the crusting of lava streams
of heavily disturbed ground, typically up to 100 m commonly have thin roofs of more broken rock.
across, caused by crustal decline and deflation Basalt can create a strong rock arch: many roads
where underlying pools of lava drained out. and light structures rest on less than 2 m of sound
Lava tubes can be a foundation hazard on shield rock over voids 5 m or more wide (Figure 23.23).
volcanoes; they are not found in level plateau Different roof origins create very variable ground
basalts. Drained flows on very gently sloping conditions, and the basalt above any tube should be
shield flanks can contain multiple braided tubes; individually assessed before load is imposed over it.
the Cuernavaca lavas in Mexico have more than The only lava tube roof collapses yet recorded
10 km of braided and distributary tubes, each have been of very heavy machines crashing
2–10 m in diameter, in an 800 m long segment of a through during construction. Precautionary engi-
single flow 600 m wide (Figure 23.22). In contrast, neering may be required where a road unavoidably
a flow on the east slope of Kilauea contains a single crosses a tube. Reinforced concrete slabs have
continuous tube 41 km long with relatively few been placed over some lava caves on Cheju Island,
GFE-23.qxd 6/27/05 2:13 PM Page 648
Korea, and cave roofs have been broken in to cre- 23.3 Explosive volcanoes
ate a stable rock pile within tubes on Hawaii.
Heavy buildings demand a safer thickness/width Eruption styles
ratio, and the usual precaution is to avoid the lava While all the powerful explosions owe their origin
tubes in selecting the site. This demands careful to accumulated gas pressure in viscous magmas,
ground investigation, with direct underground variations in the eruptions occur as the style of
mapping, systematic drilling or microgravity sur- gas release varies. Vulcanian eruptions have brief
veys as appropriate in the light of the frequency of explosions that blast debris to heights of many
observed tube occurrences around a particular site. kilometres. Plinian events last longer with power-
Basaltic pyroclastic material behaves as a loose ful jets spewing out vast amounts of debris. As a
granular soil, while tephra cones are commonly variant, Pelean eruptions are those with nuées
quarried for hard granular aggregate. Inter-bedded ardentes, the most lethal of the descending clouds
lava and tephra sequences can create unstable of gas and debris (TABLE 23.3). The names are
slopes where masses of strong lava fail over weaker from the type localities of Vulcano and Pele, and
tephra; deglaciated valley sides in Iceland have from Pliny’s description of the a d 79 eruption of
some very large rotational slides that occurred Vesuvius. Even larger events are described as
soon after ice support melted away. Basalts tend to ultraplinian or as super-eruptions, causing total
weather to smectite-rich soils that provide excellent destruction on national or continental scales. The
arable land, but can create engineering instability last eruption of this size was that of Toba, on
through their high shrink–swell capability: the Sumatra, 73 500 years ago. There is no engineering
unstable ferrisiallitic soils of Australia are a notable defence against super-volcanoes, and the rarity of
example (known as black cotton soils after their events makes response programmes in civil logis-
colour and widespread crops). tics impractical for the foreseeable future.
Basalt lava sequences are very permeable and Because these eruptions are so explosive, little
surface water retention can be difficult on them. magma emerges as liquid lava. When it does, as
Work on the reservoirs of the Columbia River andesite, dacite, rhyolite or obsidian, it is so viscous
basin, USA, has shown that permeabilities can be that much of it remains as almost hemispherical
very high, variable and unpredictable within the domes or upheaved plugs. True lava flows can only
layers of rafted crustal slabs and rubbly lava move on steep slopes, and then for no more than a
between the flow cores; columnar jointing within few kilometres. Slopes of ⬎ 40⬚ are normal on
the cores is generally tight. angular blocky andesitic lava, and also survive on
Table 23.3 Relationships between volcanic landforms, magma types and eruption scales. VEI is
Volcanic Explosivity Index (see also Table 23.2); eruption types are in italics.
Magma VEI Quantity of magma produced

type small → → → → → → large or repetitive → → → → → → very large
fluid 0 lava flows flood basalts — Icelandic

basic
shield volcanoes — Hawaiian
↓ 1–2 scoria cones — Strombolian
↓
3 tephra cones composite volcanoes — Vulcanian
↓
↓ 4–5 — Plinian
↓
5 tephra maars plug domes — Pelean
↓
6 calderas — Krakatauan
acid
7–8 ignimbrite sheets
viscous
GFE-23.qxd 6/27/05 2:13 PM Page 649
some conical volcanoes armoured by lava flows. vent explosions that did not have the energy to rise
Modest andesitic flows have been observed on high into the atmosphere. Individual pyroclastic
some of the Pacific rim volcanoes, but no lavas of flows may be limited in size, but they can occur in
the other materials have been observed within a rapid series of explosions that can cover the vol-
recorded history. cano flanks. In AD 79, Pompeii (Figure 23.24) was
completely buried within about seven hours by a
Pyroclastic flows steady rain of pumiceous tephra from the Plinian
Explosive volcanoes eject mixtures of hot gas and eruption of Vesuvius, but most of the people died
incandescent lava fragments high into the atmos- in the hot gases of a series of short-lived surges
phere. As the blast energy is dissipated, the tephra that interrupted the steady fall of pumice.
of lava fragments is blown by the wind and cas- The morphology of pyroclastic flow deposits
cades back to the ground as an airborne deposit. varies considerably. A typical flow unit, 5–20 m
While a metre of coarse fragments may accumu- thick, has fine ashes at both base and top, with a
late close to the volcano, a thin layer of fine ash core that has particles graded by density (not
may reach 100 km from the vent. mass) so that lithic fragments near its base give
These eruption columns are intrinsically unsta- way upwards to pumice fragments (Figure
ble, with their high content of solids supported 23.25). Little sediment remains from the short-
only by the blast and heat energy. Many therefore lived, low-density surges whose hot gasses are so
collapse — into surges of hot, turbulent clouds lethal. Nuées ardentes are glowing clouds with a
that roll down the volcano flanks under their own higher proportion of solid debris than the more
weight. These are one type of pyroclastic flow. common surges. Pumice flows, and the ign-
Others are produced when expanding lava domes imbrites that they form, are the product of unre-
collapse and thereby release debris-charged gas strained lava frothing and even more dense
clouds, and yet others are the direct products of eruption clouds.
Figure 23.24 The ruins of Pompeii excavated from the pyroclastic debris of the AD 79 eruption of Vesuvius in
southern Italy. In the background, Monte Somma, on the extreme right horizon is a part of the rim of the crater
formed in that catastrophic event, and the higher summit cone of Vesuvius has grown almost entirely since then.
GFE-23.qxd 6/27/05 2:13 PM Page 650
avalanches from dome collapses create nuées

ardentes, and a spectrum of intermediate materials
can be produced within complex eruptions.
Lahars are mudflows and saturated debris
flows that descend the lower flanks of erupting
volcanoes. Their water content is derived from
crater lakes, from ice caps melted by volcanic
heat, from eruption-induced thunderstorms or
from flank lakes and rivers. Their solid fractions
derive from debris avalanches, pyroclastic flows
and airfall pyroclastics scoured from the flanks.
Once mobilised, lahars can travel many kilometres
down river valleys (Figures 23.26 and 23.27)
before their sediment settles, and they evolve into
very muddy flood pulses. They can become the
most destructive of an eruption’s elements when
they are channelled into populated valleys. Over
20 000 people died in Armero, Colombia, when
Nevado del Ruiz erupted 50 km away in 1985;
they died in lahars, without ever seeing either lava
or a pyroclastic event (Williams, 1990).
Eruption hazards
Large explosive eruptions are massively destruc-
tive. Once an eruption is in progress, the only safe
response is to be somewhere else. Consequently,
predictions of events become extremely valuable,
as they may allow timely evacuation of potential
danger zones. Predictions are based largely on
inflation of the volcano and increased seismic
activity. Both can be monitored, but the accuracy
of a prediction generally depends on having
recorded data for comparison and this is rarely
available for explosive volcanoes that typically
have eruption intervals of 100 years or more.
Figure 23.25 Diagrammatic profile through the ele-
ments of a pumiceous pyroclastic flow that emerged
Monitoring of vent gases and ground temperatures
during an explosive eruption. can support earlier predictions, but suffer even
more from the lack of comparative databases.
A volcano is described as ‘dormant’ through
Avalanches and lahars the times in between active periods of eruptions.
Viscous lava and proximal pyroclastics from explo- It is only considered ‘inactive’, ‘dead’ or ‘extinct’
sive eruptions build steep conical mountains whose when there are no historical records of activity.
slopes progressively become unstable. Collapse is a There are no clear boundaries between active,
natural event on conical volcanoes. Massive land- dormant and extinct, and use of the terms is dis-
slides may occur as part of an eruption, as famously couraged as they can mask the real situation.
at Mt St Helens in 1980 (Figure 23.26), and the Poorly documented volcanoes that were thought
debris avalanches from these constitute the most to be extinct have suddenly erupted, notably
voluminous of volcanic sediments. Hot debris Lamington (Papua New Guinea) in 1951 and
GFE-23.qxd 6/27/05 2:13 PM Page 651
Figure 23.26 Features of the 1980 eruption of Mt St Helens, USA. The landslide debris from the failure of the
northern flank fell into the valley of the Toutle North Fork. The debris flow dammed, and created lakes in, three trib-
utary valleys and then developed into a lahar which extended west of the map and evolved downstream into a muddy
flood pulse. Spirit Lake existed before the eruption, but was raised in level. The lateral blast overtook the landslide
and destroyed the forest over hills and valleys alike. Deposits from pyroclastic flows created the Pumice Plain and
also fed smaller lahars on the south flank, while the main airfall ash landed 250 km downwind to the east.
Figure 23.27 The debris of mud, tephra, soil and broken trees remaining from the lahar that swept down the Smith
Creek valley from America’s Mt St Helens during the 1980 eruption.
GFE-23.qxd 6/27/05 2:13 PM Page 652
Bezymianny (Kamchatka, Russia) in 1956. The August. The evacuation could have been judged as
island volcano of Taal, in the Philippines, erupted unnecessary, but was a wise precaution.
in 1911 and killed nearly all the 500 people who It is significant that the main destructive ele-
had resettled on the island during a century of ments from an explosive eruption are directed
dormancy, when memories and folklore of earlier largely down valleys. Pyroclastic flows can
massive eruptions were long forgotten. Such is descend a volcano at over 200 km/hour, but they
less likely to happen today due to the more intense rarely travel far from the edifice flanks (Figure
monitoring and documentation of all volcanoes. 23.28). Though many are confined to valleys, they
Evacuation of any area around a volcano is are so mobile that they can override interfluves and
often not practically possible until people can per- large base surges spread radially from the eruption
ceive a risk, or unless they have seen other recent, column over ridges and valleys alike. Lahars can-
destructive events. Fortunately, nearly all large and not escape from their valleys, but travel at
dangerous, explosive eruptions are preceded by 50 km/hour or more and can be extremely destruc-
days or weeks of milder activity, which provide this tive up to 100 km from their source volcano.
perception. Evacuation is likely to proceed when Hazard zoning around a volcano is therefore influ-
advice from an official source is backed by the enced by topography, with the absolute distances
sight of a tall ash column or the first pyroclastic determined by evidence from previous events and
flows high on the volcano slopes. Increased steam recognisable deposits. Open-framed barriers in
plumes and seismic activity on Ecuador’s volcano ravines work well in ameliorating lahars on
of Tungurahua prompted evacuation of 22 000 Sakurajima and other Japanese volcanoes.
people from the town of Banos in mid-September Symmetrically concentric hazard zones can
1999. Only in mid-October did the first major be rendered useless by a flank collapse and
explosive eruption occur. Pyroclastic flows and subsequent lateral blast. The 1980 event at
lahars were generated, but Banos was only Mt St Helens was the first to be observed and
impacted by airfall ash, and most inhabitants had interpreted (Lipman and Mullineaux, 1981), but
returned to their homes and fields by the following subsequent research has revealed that similar
Figure 23.28 A small pyroclastic flow hurtles down a hillside on Montserrat. It originated as a gas blast released
by a partial collapse of the chilled crust over the expanding lava dome.
GFE-23.qxd 6/27/05 2:13 PM Page 653
destructive events have taken place on dozens of from the eruptions of Krakatoa and El Chichon,
the world’s volcanoes (Siebert, 1984; Crandell, and realistic debate continues unresolved on the
1989). A rising magma dome, offset from the cen- impacts of much larger prehistoric events.
tral vent, caused the disturbance and subsequent
landslide failure of the flank, which allowed the
initial blast to escape sideways instead of upwards Volcanic landforms
(Figure 23.26). Greater attention to the shape of Accumulations of alternating lava and tephra
the inflation is now routine in volcano monitor- build the classic composite cones (Figure 23.1),
ing. St Helens’ landslide scar was modified into a also known as ‘strato-volcanoes’, whose simple,
new crater by subsequent explosive action, and symmetrical profiles belie a complexity of inter-
this now contains a new lava dome that has risen nal structure (Figure 23.30). Variations in cone
within the vent (Figure 23.29). This dome tem- and crater dimensions relate largely to the explo-
porarily blocks the vent, and the mechanism of its sivity of eruptions (Table 23.3).
eventual failure will determine the scale of the Viscous lavas, of andesite, dacite, rhyolite or
next eruption. obsidian only flow on steep slopes, where they
Airfall ash is rarely more than an inconven- break into chaotic masses of blocks that are typi-
ience, unless it is as thick as that which blanketed cally 1–5 m across; this is auto-brecciation on a
villages and rice paddies below Galunggung, in grand scale. Like giant aa, these flows can have
Java, in 1982. With respect to roof loading, the more solid cores but their surfaces create extra-
density of airfall tephra is typically about five ordinarily inhospitable terrain. They form lobes
times that of snow. High eruption clouds of finely down the volcano flanks, some elongated with high
dispersed ash can seriously impede aircraft jet levee margins, and others more bulbous with cres-
engines and volcanoes below air routes of the cent pressure ridges like super-giant ropy pahoe-
western Pacific are now carefully monitored for hoe. The inter-bedded tephra range from coarse
even modest upward blasts. Very large explosive lahar deposits to laminated airfall ashes. These fill
eruptions can inject so much ash into the upper valleys or blanket terrains and may be loose sedi-
atmosphere that they can influence global cli- ment or may gain strength where indurated or
mates. There were measurable short-term impacts partly welded.
Figure 23.29 The 1980 crater of Mt St Helens, USA, seen from its rim. Essentially a blast-modified landslide scar,
the break in its far wall is where the landslide and lateral blast headed out northwards. The valley beyond is floored
by pale pyroclastic flows on top of the landslide debris. This dams Spirit Lake, the dark patch below Mt Rainier,
another andesitic cone 80 km away. The lava dome has grown in the crater since 1980, and is now surrounded by
dust and debris from the crumbling crater walls, inter-bedded with winter snow layers.
GFE-23.qxd 6/27/05 2:13 PM Page 654
Figure 23.30 Block diagram showing various landforms associated with explosive volcanoes.
The craters of explosive volcanoes are rarely ground collapse, creating a deep caldera ringed by
deep pits, as they tend to be backfilled by slumped steep walls along marginal faults (Figure 23.31).
tephra to create only shallow bowls. Many vents are The only historical eruption of this style was of
plugged by lava, which may have risen in the late Mt Katmai, Alaska, in 1912 (Hildreth, 1983). More
stage of an eruption to create a short-lived spine (as than 10 km3 of pumice emerged from a parasitic
at Mt Pele until it collapsed) or the longer-lived vent (Figure 23.32) to fill the Valley of Ten
dome at Mt St Helens since 1980 (Figure 23.29). Thousand Smokes to a depth of over 200 m (Figure
23.33), while the summit of Katmai collapsed into
Calderas and ignimbrites a caldera 600 m deep and 2000 m across. It was a
In the most massive volcanic eruptions, vast tiny event in comparison with ancient eruptions
amounts of gas-rich magma emerge so fast from from super-volcanoes, which were two or three
multiple vents that magma chambers are drained orders of magnitude greater (Table 23.2). The
and the land above collapses into a caldera. The planet’s most recent super-eruption was at Toba, in
best analogy is with a frothing champagne bottle. Sumatra, 73 500 years ago (Chesner and Rose,
Liquid lava flows from the vents as rhyolite or 1991). It formed a caldera 100 km long, when
obsidian during the milder phases of the eruption. 2800 km3 of pyroclastic debris were ejected.
These are overshadowed by the climax events Sulphuric acid aerosols created from the sulphurous
when the extremely mobile froth emerges as a gas, together with the ash in the high atmosphere,
fluid mass of rhyolitic pumice fragments sup- appear to have caused the worldwide cooling of
ported by hot expanding gas. On settling, this 3–5 ⬚C that is recognised as Stage 4 in the oxygen
forms an ignimbrite (Sparks et al., 1973); some of isotope record from Greenland ice cores.
it may weld into a very strong, glassy rock, but Volcanic activity of this style is rare; the world-
most remains as a partly indurated, granular sand. wide frequency of super-eruptions is less than
Evacuation of the magma chamber then causes once every 100 000 years. This still represents a
GFE-23.qxd 6/27/05 2:13 PM Page 655
Figure 23.31 Geological map of the Uzon caldera, Kamchatka, Russia. Ignimbrites outside the caldera were
formed during its succession of collapses, starting 40 000 years ago. The rhyolite platforms, Belaya lava dome and
Dalneye tuff ring all lie inside, and post-date, the last caldera collapse. The Valley of Geysers lies close to the
caldera’s marginal faults, where groundwater circulation and geothermal heat are both at their maximum.
significant annual probability, but is very difficult and prediction of ground conditions very difficult
to relate to daily, individual or personal hazards. for the engineering geologist.
These eruptions can influence world climates and Under warm and wet weathering conditions,
they would severely impact human life and civili- basic volcanic rocks weather to produce swelling
sation on a continental scale. They also create smectite clays, creating moisture-sensitive, com-
entire new landscapes. The rhyolites survive as pressible and subsidence-prone soils within adja-
high plateaus, and the caldera walls remain as cir- cent basins (Chapter 18). Andesites can weather to
cular cliffs until steadily degraded by erosion. The unstable andisols rich in halloysite clay minerals.
weakly indurated ignimbrites are commonly gul- Acid volcanics tend to weather to kaolinite clays,
lied into badland topography, of which the dis- which are relatively stable and engineering-
sected plateaus and remnant pillars of Capadoccia friendly. However, many volcanic soil sequences
form a splendid example in eastern Turkey (Figure are notoriously variable, due to their accumula-
23.34). tion as layers of volcanic debris from events at
different times at varying distances, with periods
Engineering implications of weathering and erosion between and subse-
The explosive volcanics create notoriously vari- quent to the volcanic events. This complexity can
able ground conditions. Very strong rhyolitic cause significant problems in ground investiga-
lavas and welded ignimbrites alternate with loose tion and for all forms of construction.
tephra and coarse pyroclastic flow deposits con- Perhaps the most important feature of many
taining large lava blocks in a fine matrix. Steep residual soils derived from volcanic material is
initial dips in loose tephra can create slope insta- that they do not fit well into the pattern of behav-
bility. At the volcanic centres, vertical vent struc- iour established for alluvial soils (Wesley, 1998;
tures, lobate and lenticular lavas, infilling tephra Shimizu, 1998). This is mainly because of the
and hard intrusive dykes can create a three- porosity and crushability of the soil grains in the
dimensional complexity that makes interpretation silt, sand sizes and the presence of particular clay
GFE-23.qxd 6/27/05 2:13 PM Page 656
Figure 23.32 The lava dome inside the Novarupta vent, Alaska, from which emerged 10 km3 of pumice during the
1912 eruption, when the summit of Mount Katmai (on the horizon), collapsed into a small caldera.
Figure 23.33 The Valley of Ten Thousand Smokes, Alaska, filled to a depth of 200 m with pumiceous ignimbrite
in the 1912 eruption, and since entrenched by meltwater rivers. Novarupta, the 1912 vent, is off to the right;
Mt Griggs, in the background, is a dormant volcano.
GFE-23.qxd 6/27/05 2:13 PM Page 657
minerals such as allophane. In situ water content Low particle densities can make volcanic soils
and liquid limit tend to be much higher than usual, very susceptible to surface erosion, by water or
due to the predominance of hollow or porous par- wind, which can be prevented by rapidly estab-
ticles that hold water within the particles. Particle lishing vegetation or by placing a surface layer
crushing gives rise to a number of features: of soil-cement. Internal erosion can be equally
serious where water is able to flow through
1. investigation by cone testing shows very little embankments, though the hazard is reduced by
difference between dense and loose states, the adequate drainage measures. Liquefaction of
behaviour in both cases being similar to that loose volcanic soils during earthquakes is another
for loose quartz sand potential hazard.
2. the reduction in friction angle with increasing Some fine-grained volcanic soils in Indonesia
pressure is greater than for quartz sands, and it have been observed to have much higher strengths
is therefore more important to ensure that tri- than expected in the steep terraces of the rice
axial strength tests are carried out at an appro- fields. Their strength does not appear to be
priate confining stress dependent on plasticity in the manner of alluvial
3. during consolidation, the void ratio has a linear soils. Peak friction angles may be as high as
(not logarithmic) relationship with pressure, so 37–40⬚, while back analysis of slopes suggests
that compressibility is described by a constant that values of effective cohesion (c⬘) must be
value of the coefficient of volume change (mv) above 10–15 kPa.
4. during compaction, soils may soften so that no Many sequences of primary pyroclastics are
peak is obtained on the curve of dry density well bedded, with variable strengths and steep
against moisture content; the soil is easily initial dips (Figure 23.35), but some large areas
damaged by earth-moving machinery leading of ignimbrite can offer very simple ground con-
to loss of trafficability and reduced strength ditions. The thick and extensive Neapolitan Tuff
(Tonogaito et al., 1998), however the drying that underlies the western part of Naples, Italy, is
of the soil leads to improved performance, partly indurated, so that it is easily excavated yet
which is not reversible, while addition of has a bearing capacity of over 1 MPa, can be cut
quicklime is effective in improving stability. back to stable vertical faces and can also be cut
Figure 23.34 Gully erosion of weakly indurated ignimbrites has created the towers and pinnacles of Capaddocia,
Turkey. Houses and churches carved inside some of the pinnacles are exposed where they have partly collapsed.
GFE-23.qxd 6/27/05 2:13 PM Page 658
Figure 23.35 A sequence of coarse, poorly sorted deposits from four small pyroclastic flows are separated by thin-
ner beds of pale, fine-grained surge deposits. The exposure is a roadcut about 8 m high on the entirely volcanic
island of Lipari, off the north coast of Sicily.
into building blocks. The main geohazard of the one or more vents accumulates over time to form
Naples tuff is now the collapses over the hun- conical or gently domed edifices. Many mud vol-
dreds of ancient tunnels and mines which were canoes are associated with magmatic volcanoes,
left with thin, unstable roofs prone to piping but others are cold and are driven by gas that
failure. At the same time, some ignimbrites are emerges from diapiric intrusions of underconsol-
notable for their deep open fissuring; these idated mudrocks. The main gases are methane
include the fissured rhyolitic flow deposits that and carbon dioxide.
initiated the failure of the dam at Teton Falls, A typical mud volcano has a summit source
Idaho, in 1976. area that comprises one or more vents in a circu-
Hydrothermal alteration can occur anywhere, lar or elongated crater. Individual vents may be
but is particularly common under and around cones just a few metres high or bubbling mud
water-rich explosive volcanoes, including those pools just a few metres across. They may form
no longer active. Its main effect is the hydration of cones, shields, maars or basins comparable to
silicates, producing clay minerals that inevitably lava features but on a smaller scale. Ground rup-
reduce the strength of the rocks. The alteration is turing is a common feature of eruption activity
commonly along relatively narrow zones assoc- and fissures may extend for hundreds of metres.
iated with fissure lines that were pathways for the The body of a mud volcano may be a steep-sided
hot and corrosive escaping fluids. cone of viscous sediment that may reach 400 m
high. More fluid muds create low-profile
shields with base widths that commonly exceed
23.4 Mud volcanoes a kilometre. Mudflows commonly reach over
100 m from the source crater (Figure 23.36),
Mud volcanoes occur where mud (and some and more voluminous eruptions may mantle the
larger clasts), liquids and gasses erupt from the entire cone in a single event. The Otman Bozdag
ground surface, so that the sediment forms mud volcano in Azerbaijan has a mudflow
downslope mudflows. Sediment emerging from 2900 m long.
GFE-23.qxd 6/27/05 2:13 PM Page 659
Sediment ultimately accumulates as a series of

overlapping, low-angled (generally ⬍5⬚) flow
lobes, with deformational structures that match
those in basalt lavas. Recent lobes tend to be un-
vegetated and dark grey with distinct morpholo-
gies. Over time the lobes degrade, the surface
weathers to a lighter grey and plants colonise the
site. The extent of this zone of mudflow lobes is
controlled largely by the magnitude of the erup-
tions and the water content of the emitted material.
Though mud volcanoes are recorded in many
countries, the world’s greatest assemblage is in
Azerbaijan, where over 1800 recorded eruptions
have occurred within the last 190 years. The erup-
tions fall into four types:
1. Type I: eruptions of large volumes of mud

breccia with numerous rock fragments, accom-
panied by explosions of varying strength, the
emission of a powerful gas jet, with or without
combustion, and the formation of fissures.
Flame heights can exceed 500 m. In the 1902
eruption of Bozdag-Gezdeg, 6 men and 2000
sheep were burned to death.
2. Type II: explosions of gas, and the formation
of large fissures, without the emission of flow-
Figure 23.36 One of the Macalube mud volcanoes
ing mud. These approach the characteristics of near Agrigento, in Sicily, with flows of cold liquid mud
the more violent mud-laden eruptions of some radiating from a small vent pool.
of New Zealand’s geysers.
3. Type III: relatively small outflows of mud,
without intense gas emission. that are potentially catastrophic, only broadly pre-
4. Type IV: extrusion of sediment, with negligible dictable and barely controllable. While some lava
gas emission. These match the style of mud flows may be diverted, many eruptions must sim-
volcanoes on sedimentary basins in Sicily ply be avoided. Inactive volcanic landscapes are
(Figure 23.36) and Alaska, that are in a state of more widespread; they offer no dramatic hazard,
almost continuous mild eruption with modest but can provide difficult ground conditions where
rates of mud emission. areas and horizons of very strong rock mask loose
and weak materials.
23.5 Summary References

An engineer’s appreciation of a volcano starts with Bolt, B. A., Horn, W. L., Macdonald, G. A. and
recognising its type and the potential form of its Scott, R. F. (1975) Geological Hazards. Springer-
eruptions. The hazard contrast between the mild Verlag, New York.
Chesner, C. A. and Rose, W. I. (1991) Stratigraphy of
basaltic eruptions and the dangerous andesitic or the Toba Tuffs and evolution of the Toba caldera
rhyolitic explosions is significant (Table 23.4). complex, Sumatra, Indonesia. Bulletin of Volcan-
Active volcanic landscapes provide geohazards ology 53, 343–356.
GFE-23.qxd 6/27/05 2:13 PM Page 660
Table 23.4 A summary of the general features of volcanoes and volcanic landscapes. The two main
types are best described as basaltic (or effusive) and explosive (or mainly andesitic). There is some over-
lap with respect to features and landforms, and some volcanoes fit in between these two broad categories.
Magma type Basaltic Andesitic to rhyolitic
Activity style Effusive Explosive

Plate environment Divergent boundaries Convergent boundaries
Rock types Basalt Andesite
Basaltic andesite Dacite
Rhyolite
Obsidian
Eruption events Hawaiian: fountain or flow Vulcanian: brief explosions
Strombolian: mild explosions Vesuvian: powerful jets
Surtseyan: phreatic explosions Plinian: larger Vesuvian
Pelean: very large explosions
Krakatoan: giant explosion
Super-eruptions
Duration of eruption Strombolian: minutes Hours to days
Hawaiian: hours to years
Eruption interval Strombolian: minutes to hours 1– ⬎100 years
Hawaiian: weeks to years Super-eruptions: ⬎ 100 000 years
Lava flows Aa: rubbly surfaces Blocky surfaces
Pahoehoe: smooth surfaces
Flood basalt: large volumes
Volcano profiles Shield: low profile Conical strato-volcanoes
Lava plateau Caldera collapses
Tuya: table mountain
Explosive activity Airfall ash: minor Airfall ash: major
Tephra fountain Vertical eruption column
Spatter fountain Lateral blast
Pumice frothing
Pyroclastic deposits Airfall tephra: proximal Airfall tephra: proximal and distal
Pumice
Ignimbrite: loose or welded
Small landforms Tephra cone: loose debris Lava dome
Spatter cone: welded debris Lava spine (temporary)
Tuff ring: thrown loose debris
Slopes Lava flows: 3–8⬚ Lava flows: 40⬚
Flood basalts: stepped profile Tephra: 30–33⬚
Tephra: 33⬚ Welded ignimbrite: 90⬚
Ground conditions Strong lava cores Strong lavas
Rubbly lava tops Granular pyroclastics
Loose tephra Strong welded ignimbrites
Rock density (Mg/m3) 2.9 2.7 (pumice: 0.9)
Intact rock strength (MPa) (UCS) Lava: 200–300 Lava: 100–300
Tephra: 0–1
Welded ignimbrite: 5–200
Typical permeability Solid lava cores: ⬍ 0.01 Fractured lavas: 0.1–5
(m/day) Rubbly lava tops: 10–100 Tephra: 0.1–50
Ground hazards Inter-beds of loose lava rubble Hydrothermal alteration
Lava tubes
Eruption hazards Invasion by lava flows Major explosions
Proximal airfall tephra: modest Pyroclastic flows and surges
Nuées ardentes
Lateral blasts
Distal airfall ash: major
GFE-23.qxd 6/27/05 2:13 PM Page 661
Crandell, D. R. (1989) Gigantic debris avalanche of Waltham, A. C. and Park, H. D. (2002) Roads over lava
Pleistocene age from ancestral Mount Shasta volcano, tubes in Cheju Island, South Korea. Engineering
California, and debris avalanche hazard zonation. Geology 66, 53–64.
U. S. Geological Survey Bulletin 1861, 32 pp. Wesley, L. D. (1998) Some lessons from geotechnical
Francis, P. (1993) Volcanoes: A Planetary Perspective. engineering in volcanic soils. In Yanagisawa, E.,
Oxford University Press. Moroto, N. and Mitachi, T. (eds) Problematic Soils.
Hildreth, W. (1983) The compositionally zoned eruption Balkema, Rotterdam, 851–863.
of 1912 in the Valley of Ten Thousand Smokes, Williams, R. S. and Moore, J. G. (1973) Iceland chills
Katmai National Park, Alaska. Journal of Volcano- a lava flow. Geotimes 18 (8), 14–17.
logical and Geothermal Research 18, 1–56. Williams, S. N. (ed.) (1990) Nevado del Ruiz volcano,
Lipman, P. W. and Mullineaux, D. R. (eds) (1981) The Colombia. Journal of Volcanological and Geo-
1980 eruptions of Mt. St. Helens, Washington. U. S. thermal Research 42.
Geological Survey Professional Paper 1250, 843 pp.
Rowland, S. K. and Walker, G. P. L. (1990) Pahoehoe
and aa in Hawaii: volumetric flow rate controls the
lava structure. Bulletin of Volcanology 52, 615–628.
Self, S., Thordarson, Th., Keszthelyi, L., Walker, G. P. L., Further reading
Hon, K., Murphy, M. T., Long, P. and Finnemore, S.
(1996) A new model for the emplacement of Chester, D. K., Duncan, A. M., Guest, J. E. and Kilburn,
Columbia River basalts as large inflated pahoehoe C. R. J. (1985) Mount Etna: The Anatomy of a
lava flow fields. Geophysical Research Letters 23, Volcano. Chapman & Hall, London.
2689–2692. Fisher, R. V. and Schmincke, H. U. (1984) Pyroclastic
Shimizu, M. (1998) Geotechnical features of volcanic- Rocks. Springer-Verlag, New York.
ash soils in Japan. In Yanagisawa, E., Moroto, N. Green, J. and Short, N. M. (eds) (1971) Volcanic
and Mitachi, T. (eds) Problematic Soils. Balkema, Landforms and Surface Features. Springer-Verlag,
Rotterdam, 907–927. New York.
Siebert, L. (1984) Large volcanic debris avalanches: char- Kauahikaua, J., Cushman, K. V., Mattox, T. N.,
acteristics of source areas, deposits and associated Heliker, C. C., Hon, K. A., Mangan, M. T. and
eruptions. Journal of Volcanological and Geothermal Thornber, C. R. (1998) Observations on basaltic
Research 22, 163–197. lava streams in tubes from Kilauea volcano, island
Sparks, R. S. J., Self, S. and Walker, G. P. L. (1973) of Hawaii. Journal of Geophysical Research 103
Products of ignimbrite eruptions. Geology 1, (B11), 27 303–27 323.
115–118. Kilburn, C. R. J. and Guest, J. E. (1993) Aa lavas
Tonogaito, M., Mishima, N. and Kawai, Y. (1998) of Mount Etna, Sicily. In Kilburn, C. R. J. and
Design and construction of volcanic soil embank- Luongo, G. (eds) Active Lavas: Monitoring and
ments on expressways. In Yanagisawa, E., Moroto, Modelling. UCL Press, London.
N. and Mitachi, T. (eds) Problematic Soils. Balkema, Lipman, P. W., Self, S. and Hieken, G. (1984). Calderas
Rotterdam, 929–944. and associated igneous rocks. Journal of Geophysical
Walker, G. P. L. (1991) Structure and origin by injec- Research 89 (B10), 819–841.
tion under surface crust of tumuli, lava rises, lava- Pyle, D. M. (1989) The thickness, volume and grain
rise pits and lava-inflation clefts in Hawaii. Bulletin size of tephra fall deposits. Bulletin of Volcanology
of Volcanology 53, 546–558. 51, 1–15.
GFE-24.qxd 6/27/05 4:40 PM Page 662
24. Karst Terrains

Tony Waltham
24.1 Soluble rocks terrains Karst occurs worldwide, virtually wherever

limestone is at outcrop. It is therefore less com-
Karst mon in the Precambrian blocks, such as Northern
Limestone, and a few other rock types that are Canada and much of Africa, and is notably
almost completely soluble in natural waters, may abundant in south-east China, south-east Asia, the
be almost entirely removed in the weathering Balkans and south-east USA, but karst can create
process, leaving only very small insoluble residues. difficult ground for engineers on even the smallest
The effect of this on the ground surface is the outcrop of limestone.
development of a distinctive assemblage of land-
forms that create a range of characteristic mana- Soluble rocks
gement problems. These processes contribute By far the most widespread soluble rocks, that
minimal detrital sediment to the river systems and form most of the world’s karst terrain, are the car-
soils are therefore thin and very slow to accumu- bonates. They are soluble in natural waters that
late. The effect underground is that the ubiquitous contain carbon dioxide, with which they combine
fractures are enlarged by chemical dissolution to form the soluble bicarbonate. Soils are the
by slow-moving groundwater as the fracture walls main source of groundwater carbon dioxide, as
are removed in solution. Over time, the initial they contain up to 3% of the gas, compared with
cracks in soluble rocks are enlarged into wide only 0.03% in normal atmosphere. The main karst
fissures and then into open caves that are capable rock is, therefore, limestone. The largest caves
of carrying all natural drainage underground. This and the most rugged surface karst landforms
widening process is impossible in insoluble are formed in the older, stronger limestones.
rocks: surface tension prevents water moving fast Compact, well-lithified, strong limestone has
enough to erode mechanically in fractures that are negligible permeability in the intact rock, so that
initially narrow. all groundwater flow and dissolution effort is
A karst terrain (karst is the German form of a concentrated in the fissures, creating large caves
Slovene word meaning bare stony ground) is and efficient drainage paths. Marble may be simi-
defined as one with underground drainage and larly cavernous and karstic; there are large cave
also a distinctive landform that is the consequence systems just behind the marble quarries in Italy’s
of this (Figure 24.1). Caves, closed depressions famous Carrara district.
and sinkholes are therefore essential components Dissolution effort is dispersed within the more
of a karst terrain; dry valleys, isolated hills and porous, softer limestones, which therefore contain
bare rock outcrops are also characteristic but are numerous small cavities in place of fewer large
not present in all karsts. For the engineer, karst ones, though their bulk permeabilities are compar-
provides unique styles of difficult ground condi- able. Chalk is the softest limestone and generally
tions with open voids, potential collapse and has only small caves beneath a distinctive karst
subsiding soil cover, and also a special case of landscape with few rock outcrops on account of
hydrological condition where surface water is dif- its low strength and rapid degradation.
ficult to retain and underground water is difficult Dolomite, with magnesium replacing half the
to exploit. calcium to form the mineral dolomite in place of
662
GFE-24.qxd 6/27/05 4:40 PM Page 663
KARST TERRAINS 663
Figure 24.1 Bare limestone deeply fretted by dissolution so that steep runnels drain into deep fissures, forming
the pinnacle karst of a stone forest at Shilin, China. Comparable deep fretting of a buried rockhead provides the
difficult ground conditions typical of karst.
calcite, along with partially replaced dolomitic subsidence depressions over areas of rockhead
limestones, are also soluble, but less so than pure dissolution.
limestone. They can support karst landforms but Quartzite, composed almost entirely of silica,
they are generally more resistant, therefore form- is so insoluble that only tiny amounts of siliceous
ing crags and residual hills and containing only cement are dissolved over very long periods of
smaller caves. time. However, this can then permit piping failure
The main evaporite rocks, gypsum of the disaggregated quartz grains — which is the
(CaSO4.2H2O) and halite (NaCl, rock salt), are origin of the pseudo-karst of deep fissure caves
both much more soluble than limestones. Gypsum and giant sinkholes in the quartzite plateaux (the
may contain extensive caves within fine karstic tepuis) of Venezuela and just a few other sites.
terrains, and creates special engineering problems
by way of its rapid dissolution and its low
mechanical strength. Anhydrite is also soluble in 24.2 Surface landforms of karst
water but only occurs at depth because it converts
to gypsum at depths less than about 100 m. Halite Karren and rockhead
is so rapidly dissolved in normal weathering that Dissolution of strong and massive limestone is
it cannot survive at outcrop except in deserts. Its highly selective as fractures are etched out and
surface landforms are therefore rare, and caves enlarged into open fissures, while the intervening
are never large because halite is so ductile that rock is minimally eroded and retains its high
it will deform and close voids within erosion strength. The normal result is a rock surface
timescales. The main features of salt karst are fretted by open fissures and entrenched runnels,
GFE-24.qxd 6/27/05 4:40 PM Page 664
both of which are known as types of karren and draining into wide fissures, which can create
(Figure 24.2). Rillenkarren are channels only a terrain that is extremely difficult to traverse.
10–20 mm wide with rounded floors between Soil water is normally rich in biogenic carbon
sharp interfluve ridges; they are formed by sub- dioxide and continues the dissolution attack of the
aerial rainwater runoff which deepens the channel rockhead, except in rare situations beneath totally
floors by dissolution. Rundkarren are channels impermeable clay soils. Consequently, rockhead
(or runnels) 100–400 mm wide with rounded on limestone is typically highly irregular and may
floors and interfluves; they are formed beneath a contain wide buried fissures, creating difficult
cover of soil or vegetation that supplies corrosive ground conditions for heavy foundations that
water to the whole surface. Kluftkarren are gener- have to bear on rock. Karst in tropical regions
ally straight fissures which have been etched out may develop pinnacled rockheads with tens of
along tectonic fractures by sub-aerial or sub-soil metres of buried relief.
dissolution. A surface fretted by karren may also
be known as lapies (or lapiaz). Sinkholes and dolines
The sizes of the karren features on an outcrop Surface hollows or closed depressions are the
increase with age, rainfall supply and water ubiquitous landform of a karst terrain, as they can
aggressiveness (which increases with biogenic only survive (outside deserts) where underground
carbon dioxide in a plant-rich environment). Bare drainage prevents them filling with rainwater to
alpine terrains have only small rillenkarren excav- become lakes. They may be 1 m or many kilo-
ated by snowmelt water. Rainforest terrains have metres across and 1–300 m deep with sides that are
giant karren runnels cutting back into steep slopes gentle slopes or rocky cliffs. They may contain
Figure 24.2 A limestone pavement on the Astraka plateau of northern Greece, which was glaciated during the
Pleistocene. Deep kluftkarren fissures have been opened by postglacial dissolution, and are fretted with small
rillenkarren formed by modern snowmelt drainage.
GFE-24.qxd 6/27/05 4:40 PM Page 665
KARST TERRAINS 665
Figure 24.3 Profiles through the six types of sinkholes (or dolines) formed in karst. The dropout and suffosion
sinkholes may be known collectively as subsidence sinkholes.
sizeable sinking streams or rivers, or may absorb than 50 m across in the karst lowlands of central
all rainfall by percolation and fissure flow. To Florida.
geomorphologists, these closed hollows are Dissolution, collapse and caprock sinkholes
known as dolines, but most engineers refer to are all dependent on processes that mature over
them as sinkholes (regardless of whether or not geological timescales in limestone; although they
they contain sinking streams). form significant and abundant landforms, their
There are six types of sinkholes (Waltham et al., collapse events are so rare that they present only a
2005), classified on the processes by which they very minor hazard to engineering projects. Halite
develop (Figure 24.3); they have been given vari- may be dissolved so rapidly at a rockhead beneath
ous other names in the past (Table 24.1). Suffosion a permeable soil that dissolution sinkholes can
and dropout sinkholes both form entirely within form within engineering timescales; along with
the soil profile; infiltrating rainfall washes soil areal and linear dissolution at rockhead, this may
down into pre-existing rockhead fissures at rates create a significant form of ground subsidence in
that can be significant to engineered structures; salt regions (Chapter 12). Gypsum dissolution
they are known collectively as subsidence sink- is too slow to create new sinkholes by rockhead
holes. The slow slumping of a non-cohesive sandy lowering, but rapid corrosion by a flowing cave
soil produces suffosion sinkholes that may dam- stream can undermine gypsum fast enough to
age structures but are not life threatening. In a create collapse sinkholes along with the dropout
cohesive clay soil, cavitation initiates immediately failure of any soil cover (Figure 24.4). Buried
above a rockhead fissure and grows slowly beneath sinkholes are features of an irregular karst rock-
an arched soil roof. It propagates upward until the head that may be relevant to foundation condi-
surface collapses instantly and without warning — tions (see below).
such a dropout sinkhole can be a major engineer- Valleys are entrenched in karst terrains only
ing hazard in a soil-covered karst, particularly if it where and when their formative rivers cannot sink
is large. There are many dropout sinkholes more underground. This may occur before significant
GFE-24.qxd
666
6/27/05
4:40 PM
Table 24.1 Nomenclature of sinkhole types with past and present classifications compared.
Lowe & Waltham 2002 Ford & White Jennings Bogli Sweeting Culshaw & Beck & other
Page 666
Waltham & Fookes 2003 Williams Waltham Sinclair terms
Waltham et al., 2005 1989 1988 1985 1980 1972 1987 1986 in use
dissolution solution solution solution solution solution solution solution

collapse collapse collapse collapse collapse collapse
(fast) or
caprock – subjacent subsidence solution – collapse
collapse collapse (slow) subsidence
dropout cover cover

collapse subsidence alluvial alluvial subsidence collapse
subsidence
suffosion suffosion cover cover ravelled,
subsidence subsidence shakehole
buried – – – – – buried – filled,
palaeo-
GFE-24.qxd 6/27/05 4:40 PM Page 667
KARST TERRAINS 667
Figure 24.4 A ground failure that destroyed buildings in Ripon, Yorkshire (UK), was due to a series of sinkhole
processes. Rapid dissolution of a gypsum bed caused collapse of a cave within it, collapse of the overlying cover
rocks and then a dropout sinkhole in the soil cover, so that the ultimate ground failure was an instantaneous event.
Sinkholes had been recorded at the site before the houses were built, and part of the failure was due to collapse of
the fill within a buried sinkhole.
karstic permeability is developed, but under- Deep and spectacular gorges are common
ground capture of the rivers is inevitable with features in karst terrains. None of the world’s
maturation of the karst. Dry valleys are then left larger gorges originated as a collapsed cave. They
as abandoned fluvial features. By subsequent were entrenched by rivers whose incision was far
karstic modification, these may evolve into linear more rapid than was slope degradation, where
systems of dolines, as are commonly found in surface water is minimal because it sinks under-
mature tropical karsts (see below). Dry valleys ground. Many were cut, at least in part, when the
may also originate from periglacial interludes ground was sealed by permafrost and others were
when groundwater flow was temporarily pre- meltwater channels from Pleistocene glaciers.
vented by permafrost. Surface rivers do exist The largest gorges are cut by large allogenic rivers
where major flows of allogenic drainage are just that flow from outcrops of insoluble rock and
too large to sink into available caves, especially remain on the surface throughout the maturation
where low profiles create minimal hydraulic gra- of tropical karst.
dient through the rock. Ultimately many of these Dissolution planation by hillside undercutting
do find underground routes, but the timescale of is concentrated at valley floor level by surface
cave enlargement generally restricts very large water on an alluvial plain. An alluviated closed
cave rivers to the tropical karsts (see below). depression may, therefore, widen into a polje,
GFE-24.qxd 6/27/05 4:40 PM Page 668
with a flat floor and abrupt margins to steep build into extensive terraces, as along the south-
confining slopes and an outlet cave drain whose ern coast of Turkey. In cooler climates, tufa is
evolving morphology controls the level of the generally restricted to small stream dams and thin
polje floor. The many large poljes in the Dinaric crusts. Engineering problems can be encountered
karst, in the Balkans, are partly fault controlled, where a buried layer of strong travertine, or tufa-
with their levels constrained by underground cemented gravel, is mistaken for rockhead when it
drainage over impermeable basements in the conceals underlying sediment of low bearing
adjacent fault blocks. Poljes in the tropical karsts capacity or high permeability.
expand at local base level regardless of the Rapid deposition builds travertine dams, typ-
basement position. ically in long flights that hold back a staircase of
small lakes. These tend to create rather beautiful
Tufa and travertine tourist sites (Figure 24.5), but most cases owe
These two materials are essentially similar, both their origin to sources of geothermal water with
formed by calcite deposition along stream greater potential for carbonate dissolution and
courses. They are the sub-aerial equivalent of cave precipitation.
stalagmite, except that their precipitation is nor-
mally due to algae extracting the carbon dioxide Engineering implications
from carbonate-saturated water. The same algae Karren runnels are merely attractive landform
create the open porous texture typical of tufa, details in bare rock outcrops but they create dif-
while massive banded material is generally ficult ground conditions where structures have
described as travertine. Algal growth is stimulated to be founded on a very irregular rockhead.
in warmer environments, where travertine may Pinnacled rockhead is the extreme form, found
Figure 24.5 Travertine dams create a staircase of small lakes at Pammukale in western Turkey.
GFE-24.qxd 6/27/05 4:40 PM Page 669
KARST TERRAINS 669
largely in tropical karsts, where the depth to rock- settle to cause subsidence depressions when they
head beneath level ground may vary by tens of are disturbed by loading or drainage changes within
metres (Figure 24.6). A construction project in construction projects. Electrical contrast between
Kuala Lumpur, Malaya, had to place end-bearing the chalk and the pipe fills make it possible
piles through soft sediment and found rockhead at to detect these features reliably by the resistivity
depths of 5 m and 80 m at adjacent pile sites only techniques of geophysics.
few metres apart. Costs become high and unpre- Subsidence sinkholes are the most frequent
dictable in such karst terrain. Rafts or mattresses engineering hazard in karst terrains, especially the
may bear on the tops of pinnacles, except on those rapidly formed dropouts. The scale of potential
that are undercut to become large unstable boul- hazard is best assessed from local records of previ-
ders within the lower soil profile. ous ground failures. The sites of new failures are
Buried sinkholes are an added component of impossible to predict, except that zones of water
rockhead relief. The only indication of their scale input across the limestone boundaries are more
is from local exposures and drilling records, but at risk, especially where soil cover is within the
they may be up to 100 m wide and deep. Soft fills range 1–15 m thick (though ground collapses
can cause differential settlement in structures built have occurred where the cavernous limestone lies
across their margins. In the karst of South Africa, beneath more than 100 m of soil cover). Desk
buried sinkholes over 100 m wide and nearly as studies of recorded ground failures and a field geo-
deep have been found to be the sites of reactivated morphological survey may identify areas of greater
subsidence and new dropout sinkholes. The or lesser risk that may be used to plan site develop-
sediment-filled pipes that are common in chalk ment. Any amount of further ground investigation
are buried sinkholes that may be almost cylindri- cannot reliably indicate future sites of sinkhole
cal in shape and extend downwards into choked failure in undisturbed ground. Dynamic com-
cave passages. Their sediment fills commonly paction with drop weights can collapse existing
Figure 24.6 Rockhead profiles beneath soil cover in karst terrains, all drawn to the same scale. These range from
minimal dissolution fissuring to pinnacled conditions. The codes kI–kV refer to the engineering classification of
karst (see Table 24.4).
GFE-24.qxd 6/27/05 4:40 PM Page 670
soil cavities at shallow depths but does little to flow may create new voids within timescales of
preclude future void development. only a few years, and then cause renewed ground
The better engineering response to sinkhole failure adjacent to the original site.
risk is to design structures that will survive any
undermining by the size of sinkhole collapse that
is likely to occur within the area, as judged from 24.3 Underground features of karst
existing features and past records. Roads can be
effectively protected by underlayers of geogrid; Caves
these will span a new cavity for long enough to Underground drainage through a karstic lime-
recognise the hazard by a sag in the road before stone is largely through conduits that have been
total collapse occurs (Villard et al., 2000). enlarged by dissolution; those large enough to
Buildings may be protected by reinforced strip be entered by man are caves, which are intercon-
footings or rafts that will span a potential new nected with smaller un-enterable fissures. Caves
failure, and bridges can be designed so that they may form anywhere within a limestone mass,
will survive the loss of any single pier footing. wherever there is, or has been, through-drainage
Dropouts are caused by water flow, and of groundwater. They may form above, below or
therefore occur most frequently during rainstorms, at a conceptual base-level water table. Individual
or where drainage paths have been modified or passages may have freely draining flow beneath
water tables have declined due to over-abstraction. an air surface or may be filled by water under
Most sinkhole collapses are induced by engineer- hydrostatic pressure. Vadose caves, formed by
ing activity and are therefore avoidable (Newton, free flowing streams, are underground canyons
1987). Proper drainage control is critical in karst interrupted by waterfall shafts along continuously
terrains, as it is uncontrolled drainage that causes descending profiles. Phreatic caves, formed
the great majority of new sinkhole collapses beneath a local or regional water table, generally
and the reactivation of existing sinkholes. The have tubular cross-sections, and their long profiles
same applies to sinkholes in gypsum karst. The may switchback up and down. Both types may
Dserzhinsk region of Belarus averages more than include rift fissures opened on joints or faults,
four new sinkholes per km2 per year, over both larger chambers formed by coalesced passages
limestone and gypsum: this is a very high rate, modified by roof breakdown, dendritic tributary
but it does include many construction-induced systems and passage loops that are either braided
collapses. An entire machine factory was lost in a channels or maze networks. The dimensions of
sinkhole in 1992. When it was built, four adjacent caves can be spectacular: ten of the world’s caves
sinkholes were filled in and the factory floor had each have more than 100 km of interconnected
at least four small failures before the 1992 event, passages and 72 caves reach depths of 1000 m.
but all the warning signs had been ignored. Passage widths are typically 1–20 m, but the
Sinkholes in drift soils over limestone may be largest cave chamber (in the Mulu karst of
remediated by excavation to rockhead, then chok- Sarawak) is 700 m long and 300 m wide.
ing the open fissures with coarse rock fill. The Cave systems can also be of spectacular com-
covering fill should then be stabilised with geo- plexity. Surface lowering and valley entrenchment
textiles or by appropriate grading, so that over long periods of time mean that most limestone
drainage can still pass through to the cave below, masses have evolved through an earlier phase when
as a seal could divert water to form new sinkholes they were saturated beneath a water table and a
over nearby fissures (Sowers, 1996). Remediation subsequent phase when they were largely free-
of sinkhole-prone sites (as opposed to individual draining into adjacent valleys. Most cave systems
collapsed features) may also be achieved by com- are therefore multi-phase, with an early network of
paction grouting of the soils immediately above tubular phreatic caves modified and entrenched by
the rockhead. Repair and filling of sinkholes over later phases of vadose canyon caves. The older pas-
gypsum can be difficult; any diverted drainage sages are generally modified by roof breakdown
GFE-24.qxd 6/27/05 4:40 PM Page 671
KARST TERRAINS 671
debris, allogenic clastic sediments and also stalac- of the host limestone (Figure 24.7). Though this
tites and stalagmites formed by calcite preciptation can be recognised in all mapped caves, locations of
from saturated percolation waters that drip or seep unknown caves cannot be predicted, except in the
into the cave. Patterns, shapes and profiles of caves broadest of terms. Limestones have too many
are guided by structural and stratigraphic features structural elements to consider and fissures may
Figure 24.7 The pattern of passages in a cave in China reveals their guidance by features of the limestone geology.
GFE-24.qxd 6/27/05 4:40 PM Page 672
develop on any or all of them, and there are too phreatic tubes and open stream passages. Passage
many choices for subsequent cave development widths rarely exceed about 10 m before roof spans
from only some of the fissures. in the weak rock collapse and the cave is choked
The initiation, or inception stage of cave by breakdown and trapped sediment.
formation is dependent on water that is moving
extremely slowly through very narrow fractures or Engineering implications
bedding planes within tight limestone. The open- Caves constitute an engineering hazard in that they
ing of the first millimetre of a fissure width may are the ultimate form of zones of weak ground
take a million years, but exactly where it occurs within terrains of strong rock. Construction over a
influences the pattern of the cave systems that cave relies on the integrity of its rock roof under
subsequently develop. The slightest chemical con- imposed load. A broad guideline is that a cave roof
trasts in the rock or groundwater cause enhanced in karstic limestone is stable under engineering
dissolution related to carbon dioxide. In addition, loading where the thickness of rock roof is greater
dissolution by sulphuric acid, derived from con- than the cave width. This concept is conservative in
nate waters and pyrite breakdown, is probably many cases, and a cover/width ratio of 0.7 is gen-
significant in the initial stages of most caves’ erally more realistic (Waltham et al., 2005). Strong
origins. Some caves also owe their subsequent beds of intact limestone are stable in very thin
enlargement to sulphuric acid migrating into a spans, but the degree of fracturing and fissuring
karst limestone from adjacent sedimentary basins. must be assessed for each site and inspection of a
These include the well-known Carlsbad Caverns, cave roof may indicate variance from the guide
in New Mexico, whose huge banks of gypsum are ratio. The Dublin–Sligo railway in Ireland stood for
a by-product of the sulphate reactions. many years on limestone only 2.5 m thick over a
Once groundwater flow is established, cave cave more than 6 m wide. When the single
passage enlargement is mainly by dissolution track was replaced by a main road around 1960, a
in drainage waters containing carbon dioxide ground-slab of reinforced concrete effectively car-
derived from soil air and, to a lesser extent, ried the road over the cave in order to minimise the
directly from the atmosphere; there is an element perceived risk. Most caves lie at depths far greater
of mechanical abrasion and clastic removal in than their width, where the guideline indicates that
larger river caves. A cave passage 1 m in diameter they offer no danger to ground loading.
may be formed from an initial fissure within Near-surface caves are a ubiquitous hazard
about 5000 years. The rate increases with large (Figure 24.9), and locations of any without surface
drainage flows and also with waters that are more entries are almost impossible to predict. Cavernous
aggressive due to higher carbon dioxide contents karst warrants a clear philosophy of ground inves-
collected from basins with lush vegetation. Cave tigation, but this may have to lean heavily on local
passages are commonly 50–100 m high and wide knowledge and experience of karst conditions.
in limestone karst beneath tropical rainforests Without either of these, engineers may waste their
(Figure 24.8), where their expansion has also investigation budgets or may rely on inadequate
been aided by long periods of uninterrupted dis- assessments of the ground’s bearing capacity.
solution attack. In temperate environments, where Adequate investigation may demand extensive and
Pleistocene evolution of karstic landforms was expensive exploration by drilling to prove the
constrained by glacial and periglacial interludes, absence of caves. The number of exploratory drill-
caves are rarely more than 15 m wide. holes can only be defined in terms of the known
Extensive cave systems can also form in local site conditions (including its geomorpho-
gypsum. Rapid dissolution of the rock means that logical history), the sensitivity of the structure to be
networks of fractures are widened simultaneously built and the results achieved as the investigation
at shallow depth beneath the water table creating proceeds in stages. Belgium’s Remouchamps
very complex maze caves in areas of low relief. Viaduct provided a classic example of the unpre-
Inputs of allogenic water can create individual dictable nature of karst (Waltham et al., 1986): 31
GFE-24.qxd 6/27/05 4:40 PM Page 673
KARST TERRAINS 673
Figure 24.8 Gruta de Janaleo, in Brazil, is an indication of the size that cave passages can reach where a river of
chemically aggressive water has drained through a limestone karst from a tropical forest environment for over a
million years. Note the person standing on the sandbank for scale.
boreholes on the initial investigation found no a given point. Site-specific risk assessment can
caves; excavation of the pier footings found 2 then indicate whether a typical or a maximum
unknown caves; 308 boreholes in a second investi- cave width is used to determine drill hole probing
gation found no more caves. depths beneath foundation sites.
Local data is the only guide to local cave pas- A similar approach may have to be adopted
sage widths and also to the extent of caves with beneath the bases of end-bearing piles, by sub-
respect to the statistical chance of one lying under drilling to prove sound rock at whatever depth they
GFE-24.qxd 6/27/05 4:40 PM Page 674
Figure 24.9 A cave passage lies only a few metres below ground surface in the Yorkshire Dales (UK). Its roof
contains tectonic fractures and fissures widened by dissolution which weaken the limestone roof beam and make its
bearing capacity difficult to assess.
terminate (Foose and Humphreville, 1979). In this voids. Three-dimensional cross-hole seismic
respect, shallow caves are a component of any tomography (3dT) is a new technique that
pinnacled rockhead where the fissures and caves exploits improved computer analyses of massive
are components of a single three-dimensional void banks of data to image cavities by their seismic
network in an eroded limestone. Additionally, shadows. With data points in boreholes or a
inter-pinnacle soils and soft cave sediments tunnel it can produce spectacular results, but is
extracted in excavation may be unsuitable for use limited in application to surface investigations of
in a balancing fill operation. flat ground. Any geophysical survey is best used
Detection of caves by geophysical surveys only as a guide to ground conditions. Exploratory
would be a welcome engineering aid, but the sci- borehole programmes are more efficient and
ence is not easy and no single technique has instructive when targeted on geophysical ano-
proven totally reliable. Useful case studies and malies, and every anomaly should routinely be
experience with electrical, seismic, radar and verified by drilling (Waltham et al., 2005).
gravity surveys are reviewed in Beck et al. (1999) Caves are notoriously unpredictable. A single
and Beck and Herring (2001). Micro-gravity cave was found, purely by chance during routine
surveys provide the most consistently useful data, maintenance, just a few metres beneath the main
as negative anomalies always indicate the absence runway of Palermo airport, on Sicily. It was 25 m
of rock and therefore some potential hazard. They wide, and though there were no signs of break-
may be due to open cave passages, fissure net- down, the consequence of even partial failure was
works, filled sinkholes, sediment-choked caves so severe that it was filled with concrete (Jappelli
or water-filled caves. With sufficient collected and Liguori, 1979). The site is on a coastal plat-
data, an analysis of anomaly amplitude and wave- form of young limestone, where wide cavities are
length may deduce the size and depth of the notably prone to development by dissolution at
‘missing rock’, but drilling is normally required the interface between salt and fresh water at either
to ascertain the details of any detected karstic current or past sea levels.
GFE-24.qxd 6/27/05 4:40 PM Page 675
KARST TERRAINS 675
Caves in gypsum can be a greater hazard than concepts of diffuse groundwater flow in uniform
they are in limestone. The lower rock strength aquifers cannot be applied to karst terrains
requires greater thickness of rock roof, but there is drained by caves. Ox Bel Ha is one of a series of
little available data on loading over gypsum caves. very long cave systems that drain Mexico’s
Concrete filling of caves and fissures in gypsum Yucatan peninsula (Figure 24.10). The map of 70
may not be a safe option where diverted ground- km of its flooded tubes, mostly 2–20 m wide and
water flow or cave streams could excavate an entirely below the water table, is a fine demon-
alternative route through the gypsum within the stration of conduit flow in a karst aquifer. If the
lifetime of an engineered structure. Remediation scale is changed, it is representative of many other
of subsiding control structures on the River Neckar, karst aquifers which are not accessible for com-
in Germany, required filling voids in the gypsum parable mapping by divers.
with concrete and also placing an upstream grout Limestone and gypsum may transmit ground-
curtain to prevent continuing groundwater flow water by all types of rock permeability (Table
and subsequent renewed dissolution (Wittke and 24.2). Young poorly lithified limestones, includ-
Hermening, 1997). ing chalk, have high inter-granular porosity,
which is replaced in older limestones by fracture
Groundwater flow porosity. It is the fractures that are enlarged by
The hydrology of karst aquifers is primarily groundwater flow to create enhanced secondary
distinguished by extreme variability between permeability. Networks of fractures and fissures
intact rock and large open caves. Traditional may be so dense that they provide diffuse flow,
Figure 24.10 Outline map of the Ox Bel Ha cave with its system of flooded cave conduits that drain the Yucatan
karst in Mexico. The cenotes are collapse sinkholes that provide access to the cave passages. Passage widths and
cenote sizes are exaggerated to make them visible at this scale (after maps by the cave divers of Grupo de
Exploracion Ox Bel Ha).
GFE-24.qxd 6/27/05 4:40 PM Page 676
Table 24.2 Typical hydrogeological properties of the different groundwater elements that occur within a limestone
karst.
Cavity type Typical size Morphology Hydrogeology
Inter-granular 0.001–0.1mm Pores between grains Porosity up to 40%

Skeletal voids Slow laminar flow in small openings
in reef limestones Diffuse permeability
Fractures 0.01–10 mm Joints, faults and Network of planar voids
bedding planes Effectively diffuse flow
Porosity ⬍ 1% (fractures only)
Fissures 10–100 mm Fractures widened Greatly increased permeability
by dissolution Porosity may be up to 3%
(fissures only)
Flow is both diffuse and conduit
Rapid laminar or turbulent flow
Conduits/caves 100 mm–10 m Maze and network caves Underground rivers or random
Dendritic tube and flow through maze cave
canyon caves Porosity 1–4% (higher only in small zones)
Turbulent flow
except when viewed on the smallest of scales. The hydrology of gypsum karst is analogous to
Increasing drainage through fissure systems is that of limestone. Deeply buried salt beds are
normally accompanied by local variations in per- almost impermeable, and groundwater normally
meability as a result of flow concentrations and impacts on salt only at its boundaries with perme-
subsequent enlargement by dissolution along able cover rocks or sediments. Very high secondary
favourable routes. In a mature karst, the hydrol- permeability is then created by rapid dissolution,
ogy evolves towards efficient conduit flow in notably in breccia zones of collapsed insoluble
large open caves — many caves in the wet trop- inter-beds that evolve over a salt rockhead and are
ics carry base flows of ⬎ 1 m3/s. It is significant the prime sites of ground removal and surface
that cave drainage patterns are normally related subsidence (Chapter 12).
to geological structure and may have no relation-
ship to topographic divides and surface basins Engineering implications
(Figure 24.11). The large groundwater flows through discrete
Typical values for overall porosity and permea- cave conduits in karst are difficult either to con-
bility of limestones are indicated in Figure 24.12. strain or exploit. Chalk and other young porous
Values for hydraulic conductivity and flow rates limestones have significant diffuse permeability
cannot be assigned to cavernous limestones that make them excellent productive aquifers,
because they vary enormously with cave morph- though enhanced flows in fissures in chalk
ology. Completely vadose cave streamways may are well known. The major hand-dug wells of
transmit water at many kilometres per day. Flow Victorian London were sunk to just below the
rates are much lower where water is ponded in water table. Adits were then driven at a depth
flooded networks of phreatic caves, but flood where they could be kept dry by pumping. Each
pulses are transmitted instantly through a flooded was driven just as far as a chance encounter with
conduit. Most drainage paths through karst lime- a major fissure, where inflows increased to over
stone are a mixture of stream canyons, under- 0.1 m3/s, an order of magnitude greater than
ground lakes and phreatic loops, with hydrological seepage inflows from un-fissured chalk.
properties unique to each and only determined by A cavernous limestone has minimal diffuse
observation. flow, and wells and boreholes are only productive
GFE-24.qxd 6/27/05 4:40 PM Page 677
KARST TERRAINS 677
Figure 24.11 The main stream passage of the Guanyan cave system carries drainage through the fengcong karst
beside the Li River in Guangxi China. The cave is unrelated to the surface topography, as it passes beneath the polyg-
onal basins of internal drainage between the conical hills. Geological influences on the cones, basins and caves are
no longer recognisable in the mature karst.
Figure 24.12 Broad relationships between porosity, pore size and permeability in contrasting types of karst
aquifer. The permeability contours (in m/day) are theoretical values based on laminar flow in parallel straight tubes,
while the numbers indicate typical values found in each lithology.
GFE-24.qxd 6/27/05 4:40 PM Page 678
Figure 24.13 Elements of groundwater flow and storage in the limestone of Gunung Sewu, an area of fengcong
cone karst in Java.
if they penetrate a cave conduit. Vadose fissure tracing. These have to be designed so that they
storage is small in karstic limestone, and flows identify flow paths that may be active only at
into boreholes below the regional water table flood stage, when water (and pollutants) may
may be limited where fissures are poorly inter- cross over within a simpler pattern of flow paths
connected or are choked with sediment (Figure that are active at low stage. The town of Horse
24.13). Well yields are small, and systematic Cave, Kentucky, has long had problems with
exploitation of a cavernous aquifer is difficult contamination of the karst aquifer and the caves
unless it targets conduits and resources identi- beneath the town emanating from its own
fied by underground exploration and mapping sewage plant and industrial wastes. The pollu-
(Waltham et al., 1985). tion has been transmitted to eight springs along
Discharges from karstic springs can measure the Green River, but these have unpolluted
tens of cubic metres per second. Water may springs lying between them fed by separate dis-
cascade out of open cave mouths or pour from crete conduits within the limestone (Quinlan and
smaller fissures against an impermeable bound- Ewers, 1989).
ary. Alternatively, water may rise under pressure Abstraction of karst water requires careful mon-
from deep phreatic loops at sites known as ‘vau- itoring and control of potential pollution sources.
clusian risings’, named after the Fontaine de Water resources from caves under Kentucky’s
Vaucluse in southern France. Hydrographs of Sinkhole Plain are susceptible to fuel pollution
karst springs generally have high peaks when from roads across the source area; however, point
flood pulses are transmitted rapidly through the spillages drain into discrete conduits and therefore
efficient cave conduits. Percolation water pro- usually contaminate only one resource point,
vides an element of base flow, but storage is which is predictable now that the drainage paths
largely within the soil cover and shallow fissures have been traced (Quinlan and Ewers, 1989).
from where most of it drains rapidly into the cave Similar problems can be created by fuel draining
conduits. into sinkholes from roads across the chalk out-
Fissures and caves have almost no filtration crop, where there is adequate fissure flow to trans-
capability. Pollution is, therefore, transmitted mit pollutants many kilometres underground
easily and rapidly through a karst aquifer once it (Atkinson and Smith, 1974).
has escaped through or round any soil horizon. Reservoir impoundment on a karst lime-
Disposal of waste and sewage requires consider- stone may be appropriate where limestone can
able care on karst. Underground drainage paths provide strong foundations for a concrete dam.
are directly determined by programmes of dye Successful dams have been built, but extensive
GFE-24.qxd 6/27/05 4:40 PM Page 679
KARST TERRAINS 679
grout curtains are generally required and these ment. Many countries have dams on gypsum that
can be difficult to place where large caverns can have never retained water.
absorb immense volumes of concrete. Drilling at
the Hales Bar dam site in Tennessee found 8411
cavities in 2000 boreholes (Schmidt, 1943), some 24.4 Types of karst terrain
of which were then exposed in the cut-off
trenches (Figure 24.14). Repeated grouting pro- Any unit of soluble rock at or close to the ground
grammes failed to seal the leakage and the dam surface develops a range of karst features that
was subsequently demolished. Leakage is self- mature with time. Immature karst may have few
increasing where sediment is washed out of pre- caves, some degree of remnant surface drainage
viously choked fissures, and the same sediment and landforms dictated largely by resistance of
may prevent effective grouting. The best dam the rocks to surface denudation. A very old lime-
sites in karst are where the geological structure stone outcrop may be reduced to a karst plain with
and adjacent insoluble beds create natural hydro- thick soils and only some of its drainage under-
logical barriers to which shorter grout curtains ground in phreatic loops. Between these end
can be connected; alluvial fills are not adequate stages, a mature karst has large cave systems
seals as they are soon lost in sinkholes beneath within terrains of high, irregular and spectacular
impounded water. China has over 5000 reservoirs relief.
impounded on karst in the provinces of Guangxi, Most of the world’s karst terrains have been
Guizhou and Hunan alone, and about a third of maturing throughout Pleistocene times, and the
these suffer significant leakage (Yuan, 1991). main factor that determines their morphology is
The Chinese have also impounded groundwater the history of their climatic environments (Table
flow by building dams in caves; some have suc- 24.3). Hot and wet tropical (and equatorial) zones
ceeded, but others have failed, mainly due to are covered in lush vegetation and have developed
leakage through unrecognised sediment chokes the most mature karst landforms. Temperate
in branch cave passages (Waltham and Smart, regions have mature karsts, generally with less
1988). spectacular landforms. Glaciokarst occurs at
Reservoirs on gypsum are generally undesir- higher latitudes or altitudes, has a history inter-
able. Any fissures that carry leakage are widened rupted by glacial or periglacial conditions during
by dissolution within the lifetime of an impound- Pleistocene cold stages and still bears the scars of
Figure 24.14 Dissolution cavities exposed in the limestone beneath the Hales Bar Dam, Tennessee (after Schmidt,
1943).
GFE-24.qxd 6/27/05 4:40 PM Page 680
Table 24.3 Representative features of the five major types of karst terrain.
Glaciokarst Fluviokarst Doline karst Fengcong karst Fenglin karst
Positive rock scars interfluve polygonal cones towers

landforms ridges ridges
Hill distribution – – – clustered isolated
Relief inherited alpine up to 300 m less than 100 m up to 1000 m up to 300 m
Negative stream sinks dry valleys dolines dolines alluviated
landforms plains
Hill slopes stepped 20–30º 20–30º 30–45º 45–90º
Soil cover bare rock, thin thin patchy in fissures, thick
pavements on plains
Rockhead exposed fissured fissured pinnacled pinnacled
Cave locations anywhere within ridges under dolines anywhere within towers
Cave passage 1–5 m 1–5 m 1–10 m 1–30 m 1–30 m
width
Surface flood streams ephemeral minimal none on alluviated
drainage streams plains
Main origins ice scouring periglacial pure karst tropical tropical, with
streams uplift
Examples Yorkshire Peak District (UK) Sinkhole Plain Guizhou (China) Guangxi (China)
Dales (UK) Causses (France) (Kentucky) Sewu (Java) Halong Bay
Calcareous Chalk Downs Kras (Slovenia) Cockpits (Vietnam)
Alps (Austria) (UK) Yucatan (Mexico) (Jamaica)
ice cover. Dissolution processes virtually cease in an alpine karst. Bare rock outcrops are conspicu-
both desert and glacial environments, where the ous and include terraced scars left by ice plucking.
few karst features are generally inherited from Limestone pavements fretted with small karren
Tertiary climates that were warmer and wetter. are the result of glacial scouring followed by min-
These may include the very large but very shallow imal soil formation and slow karren evolution
sediment-floored depressions known as ‘dayas’ in (Figure 24.2). Typically small vadose cave pas-
the North African deserts; these appear to repre- sages carry underground drainage efficiently on
sent very old buried features but their depths steep gradients in the mountainous terrains, and
and floor structures are largely unknown. Even generally intersect old, dry cave passages aban-
older palaeokarst can survive beneath lithified doned as a result of multiple glacial rejuvenations.
sediments to create difficult ground. Such was Gorges were cut by proglacial meltwater rivers, at
encountered in Devon (UK) where road excava- least partly when the bedrock was frozen and the
tions were hampered by a deeply eroded and caves inactive, but many are dry now that more
locally pinnacled surface on Devonian limestone, modest rain-fed drainage has sunk underground.
found infilled and buried by Cretaceous sands
(Fookes and Hawkins, 1988). Temperate karst
The natural landscape on limestone in a temperate
Glaciokarst environment is doline karst, of which Kentucky’s
In many high mountain terrains of limestone, the Sinkhole Plain is a prime example. The entire sur-
larger landforms relate to glacial erosion during face is pitted by dolines (also known as sinkholes)
the Pleistocene cold stages, while the smaller and where the doline rims coalesce, a polygonal
landforms comprise a glaciokarst, also known as karst forms having almost no flat ground in a
GFE-24.qxd 6/27/05 4:40 PM Page 681
KARST TERRAINS 681
terrain of interfluve networks. All drainage is sides of the dry valleys; the typical result is the
underground into mature cave systems, except for downland of eastern England.
short lengths of temporary surface flow feeding to
sinks in the floor of the larger dolines. Most gyp- Tropical karst
sum karsts fall into this category. A variety of Massive production of carbon dioxide in the
doline karst is the cenote karst of the Yucatan, in hot-house environment of the wet tropics provides
Mexico — there is minimal surface erosion on the the acidic waters that create the most spectacular
dry plains but numerous cenotes are collapse karst landforms. The natural landscape of mature
sinkholes into long caves that carry drainage from tropical karst has an ‘egg-box’ topography of con-
inland mountains (Figure 24.10). ical or hemispherical hills separated and drained
Fluviokarst is dominated by dendritic systems by dolines (Figure 24.11). Known in the western
of dry valleys, as in the Derbyshire Pennines of world as cone karst (or cockpit karst), it is better
England (Figure 24.15). Some valleys are blind described as fengcong karst, after the Chinese
where they terminate at old stream sinks. Many landscape fengcong which has clusters of hills
valleys have sinkholes along their floors, fed by with little flat land between them (Figure 24.16).
short lengths of surface streams active only Fenglin karst has isolated hills with steep sides
in storm events. Extensive cave systems are undercut by planation at the level of the interven-
developed within the interfluve ridges and also ing alluviated plain across which the limestone
pass beneath the valleys. Most of the valleys towers are scattered; it is essentially equivalent
were cut as fluvial features when the caves were to the western concept of ‘tower karst’ (Figure
sealed by ice in the permafrost of Pleistocene 24.17). The Chinese terms should be used since the
periglacial stages, but others are relics from ero- huge karst areas of Guangxi and Guizhou have been
sion before the limestone developed secondary recognised as critical to the understanding of karst
permeability. Chalk karst is a type of fluviokarst, processes and landforms (Table 24.3).
where the low strength of the host chalk pre- There is an evolutonary progression from
cludes development of bare rock scars along the fluviokarst to doline karst to fengcong karst to
Figure 24.15 The dry valley of Cressbrook Dale in the Peak District fluviokarst of the southern Pennines. Streams
re-occupy stretches of the thalweg after heavy rainstorms and drain through the thin soils on the valley floor into
limestone fissures which are connected to deeper cave conduits.
GFE-24.qxd 6/27/05 4:40 PM Page 682
Figure 24.16 Fengcong karst of conical hills in Guizhou, China.
Figure 24.17 Fenglin karst with isolated limestone towers around Yangshuo in Guangxi, China.
fenglin karst to karst plain. Valleys are broken into Chinese as fenglin with cones. True fenglin, with
lines of dolines, and as these deepen the remnant tall narrow tower hills, only breaks into the evolu-
intervening cones become the dominant land- tionary path where tectonic uplift is matched by
forms; traces of the early valleys are recognisable planation and alluviation at the declining base level
through many fengcong karsts, though they are and is matched by undercutting at the margins of
unrelated to the subsequent underground drainage. the emerging towers. The long evolution of fenglin
As fengcong karst degrades, the hills diminish karst is matched by major surface lowering, and it
while flats and poljes expand at base level between only matures in limestone sequences thick enough
them until a karst plain evolves with few remnant to not be entirely removed in the process. This and
hills; the intervening stage may be described by the the necessary coincidence of various parameters
GFE-24.qxd 6/27/05 4:40 PM Page 683
KARST TERRAINS 683
accounts for the almost complete absence of classification can only define a broad suite of
fenglin karst outside China and south-east Asia. ground conditions. A full assessment of any site
Pinnacle karst and shilin are very similar terrain should have numerical values for critical parame-
types with sharply pointed limestone pinnacles ters appended to the karst class description. The
10–30 m high. These are effectively giant karren important quantifiable parameters include:
ridges created in massive limestones in wet trop-
ical regions. The pinnacles are micro-landforms 1. amplitude of rockhead relief
that may form the main surfaces over fenglin 2. typical and maximum cave passage widths
towers. Shilin is the Chinese variety, whose name 3. frequency of new sinkhole collapses.
translates as ‘stone forest’ (Figure 24.1).
Though these extreme forms of karst landscape Even with these parameters defined, the complexity
are largely confined to the tropical regions, geo- of karstic ground must be recognised. The
logical structure can locally influence landforms approach to terrain assessment by the construction
to mimic the tropical features. In the Nahanni of geological models (Fookes, 1997) is beneficial
karst of northern Canada, remnant towers amid in karst because it directs the engineer’s thoughts to
polje basins owe their slope profiles to control the various possible conditions that can occur on
by strong beds of horizontal limestone. Geology cavernous limestone. It is however notable that a
is also influential in that towers rarely develop in large site, such as a transport corridor, may gener-
dolomite, where conical hills dominate even in ate many ‘sub-models’ where variations are created
fenglin terrains. by the many factors involved in karstic evolution.
Perhaps the most successful approach to engineer-
Engineering implications ing on karst is an observational method, where
Ground conditions on karst can be extremely there is full scope for adaptation to ground condi-
difficult for engineers, and each site may warrant a tions that cannot fully be appreciated in advance.
prolonged ground investigation. Bearing capacity A safe bearing pressure for strong un-fissured
and potential instability of karstic ground are only limestone may be 4 MPa, but this is normally found
fully appreciated after careful consideration of rock only in footings that have been excavated to below
structure, drainage conditions and geomorphic all rockhead pinnacle and fissures, and must still
evolution. This does benefit from a thorough be proven by sub-drilling to ensure sound rock with
understanding of the local karst conditions, prefer- no caves. More typically, foundations on karst
ably by an engineering geologist with karst experi- require innovative designs adapted to the specific
ence. Engineering difficulties broadly increase in site in order to successfully transfer loads to stable
warmer and wetter climatic zones where rock dis- rock (Figure 24.19).
solution is more rapid and karst landforms are Karst on gypsum can have all the difficulties of
more mature. Relict landforms can occur, whereby a limestone karst. The absolute size of features is
a large cave in a glaciated terrain may be inherited generally smaller, but rapid dissolution of gypsum
from an ancient phase of warmer conditions, but may allow new cavities to form within engineering
such features are statistically rare. timescales. Salt karst has few caves and the engin-
An engineering classification of ground eering hazards are created by ongoing dissolution
conditions on limestone karst identifies five and ground subsidence (Chapter 12).
classes of terrain (Waltham and Fookes, 2003).
These cover the considerable range of karstic con-
ditions, with increasing engineering difficulties 24.5 Summary
created by more mature karstic erosion, cave and
landform evolution (Figure 24.18), and can be Karst terrains are distinguished by landforms that
delimited by broad groups of terrain features and are unique to soluble rocks and by engineering dif-
landform dimensions (Table 24.4). However, karst ficulties that are unique to those landforms. No
features do not all evolve in strict unison, and the other terrain has such a marked contrast in bearing
GFE-24.qxd
684
6/27/05
Table 24.4 The engineering classification of karst. The table provides outline descriptions of selected parameters, which are not mutually exclusive, and give
only broad indications of likely ground conditions that can show enormous variation in local detail. It should be viewed in conjunction with Figure 24.18 that
shows some of the typical morphological features. The comments on ground investigation and foundations are only broad guidelines to good practice in the
4:40 PM
various classes of karst. NSH is the rate of formation of new sinkholes per km2 per year.
Karst class Locations Sinkholes Rockhead Fissuring Caves Ground investigation Foundations
kI Juvenile Only likely in deserts Rare NSH ⬍ 0.001 Almost uniform; Minimal; low Rare and small; Conventional Conventional
Page 684
and periglacial minor fissures secondary some isolated
zones, or on permeability relict features
impure carbonates
kII Youthful The minimum in Small suffusion or Many small Widespread in Many small caves; Mainly conventional, Grout open fissures;
temperate regions dropout sinkholes; fissures and the few metres most ⬍ 3 m probe rock to 3 m, control drainage
open stream sinks depressions nearest surface across check rockhead
NSH 0.001–0.05 fissures
kIII Mature Common in temperate Many suffosion and Extensive fissuring; Extensive Many ⬍ 5 m Probe to rockhead, Rafts or ground
regions; the minimum dropout sinkholes; relief of ⬍ 5 m; secondary across at probe rock to 4 m, beams, consider
in the wet tropics large dissolution some loose blocks opening of multiple levels micro-gravity geogrids, driven
sinkholes in cover soil most fissures survey piles to rockhead;
NSH 0.05–1.0 control drainage
kIV Complex Localised in Many large Pinnacled; relief Extensive large Many ⬎ 5 m Probe to rockhead, Bored piles to
temperate regions; dissolution and of 5–20 m; dissolution across at Prove rock to 5 m rockhead, or cap
normal in tropical numerous loose pillars openings, on multiple levels with splayed grouting at
regions subsidence and away from probes, micro-gravity rockhead; control
sinkholes major fissures survey drainage and
NSH 0.5–2.0 abstraction
kV Extreme Only in wet tropics Very large sinkholes Tall pinnacles; Very complex Complex 3-D cave Make individual Bear in soils
of all types; relief of ⬎ 20 m; dissolution systems, with ground investigation with geogrid, load
remanent arches; loose pillars cavities galleries and for every pile site on proven pinnacles,
soil compaction undercut between chambers ⬎ 15 m or on deep bored
in large buried deep soil fissures across piles; control all
sinkholes drainage and
NSH ⬎⬎1 control abstraction
GFE-24.qxd
6/27/05
4:40 PM
Page 685
KARST TERRAINS 685
Figure 24.18 Profiles through the five classes within the engineering classification of karst ground conditions. The five classes provide guideline subdivi-
sions within a continuous spectrum of evolving karst terrains, and not all the landforms shown appear in all areas of each karst class.
GFE-24.qxd 6/27/05 4:40 PM Page 686
Figure 24.19 A variety of design options appropriate for foundations on karst (based largely on Sowers, 1996).
capacity — between strong rock and open caves. Fookes, P. G. (1997) Geology for engineers: the
Soil collapses over unseen ground voids add to the geological model, prediction and performance.
Quarterly Journal of Engineering Geology 30,
problems, and every site has to be investigated as
293–431.
if it was unique. The hydrology of karst is also Fookes, P. G. and Hawkins, A. B. (1988) Limestone
unlike any other, in that high groundwater flows weathering: its engineering significance and a pro-
are concentrated in conduits that make control and posed classification scheme. Quarterly Journal of
abstraction difficult. Engineering Geology 21, 7–13.
Foose, R. M. and Humphreville, J. A. (1979) Engineering
geological approaches to foundations in the karst ter-
rain of the Hershey Valley. Bulletin of the Association
References of Engineering Geologists 16, 355–381.
Ford, D. C. and Williams, P. F. (1989) Karst
Atkinson, T. C. and Smith, D. I. (1974) Rapid ground- Geomorphology and Hydrology. Unwin Hyman,
water flow in fissures in the chalk: an example from London, 601pp.
south Hampshire. Quarterly Journal of Engineering Jappelli, R. and Liguori, V. (1979) An unusually complex
Geology 7, 197–205. underground cavity. Proceedings of International
Beck, B. F. and Sinclair, W. C. (1986) Sinkholes in Symposium on Geotechnics of Structurally Complex
Florida: an Introduction. Florida Sinkhole Formations, Associazione Geotecnica Italiana, Rome
Research Institute Report 85-86-4, 16pp. 2, 79–90.
Beck, B. F. and Herring, J. G. (eds) (2001) Jennings, J. N. (1985) Karst Geomorphology.
Geotechnical and Environmental Applications of Blackwell, Oxford, 293pp.
Karst Geology and Hydrology. Balkema, Rotterdam, Lowe, D. and Waltham, T. (2002) A dictionary of karst
437pp. and caves (2nd edn), British Cave Research
Beck, B. F., Pettit, A. J. and Herring, J. G. (eds) (1999) Association Cave Studies 10, 1–40.
Hydrology and Engineering Geology of Sinkholes Newton, J. G. (1987) Development of sinkholes resulting
and Karst. Balkema, Rotterdam, 477pp. from man’s activities in the eastern United States.
Bögli, A. (1980) Karst Hydrology and Physical US Geological Survey Circular 968, 1–54.
Speleology. Springer-Verlag, Berlin, 284pp. Quinlan, J. F. and Ewers, R. O. (1989) Subsurface
Culshaw, M. G. and Waltham, A. C. (1987) Natural and drainage in the Mammoth Cave area. In White, W. B.
artificial cavities as ground engineering hazards. and White, E. L. (eds) Karst Hydrology: Concepts
Quarterly Journal of Engineering Geology 20, from the Mammoth Cave Area. Van Norstrand
139–150. Reinhold, New York, 65–103.
GFE-24.qxd 6/27/05 4:40 PM Page 687
KARST TERRAINS 687
Schmidt, L. A. (1943) Flowing water in underground the Remouchamps Viaduct, Belgium. Ground
channels, Hales Bar dam, Tennessee. Proceedings of Engineering 19 (8), 16–18.
the American Society of Civil Engineering 69 (9), Wittke, W. and Hermening, H. (1997) Grouting of cav-
1417–1446. ernous gypsum rock underneath the foundations of
Sowers, G. F. (1996) Building on Sinkholes. American the weir, locks and powerhouse at Hessigheim on
Society of Civil Engineers, New York, 202pp. the River Neckar. Proceedings of 19th Congres,
Sweeting, M. M. (1972) Karst Landforms. Macmillan, Florence, Commission Internationale des Grands
London, 362pp. Barrages, 613–626.
Villard, P., Gourc, J. P. and Giraud, H. (2000) A geosyn- White, W. B. (1988) Geomorphology and Hydrology of
thetic reinforcement solution to prevent the forma- Karst Terrains. Oxford University Press, 464pp.
tion of localized sinkholes. Canadian Geotechnical Yuan, D. (ed.) (1991) Karst of China. Geological
Journal 37, 987–999. Publishing House, Beijing, 232pp.
Waltham, A. C. and Fookes, P. G. (2003) Engineering
classification of karst ground conditions. Quarterly
Journal of Engineering Geology Hydrogeology 36,
101–118. Further reading
Waltham, T, Bell, F. and Culshaw, M. (2005) Sinkholes
and Subsidence. Springer-Verlag, Berlin and Gunn, J. (ed.) (2003) Encyclopedia of Caves and Karst
New York, 382pp. Science. Routledge, New York, 950pp.
Waltham, A. C. and Smart, P. L. (1988) Civil engineering Palmer, A. N. (1991) Origin and morphology of lime-
difficulties in the karst of China. Quarterly Journal of stone caves. Geological Society of America Bulletin
Engineering Geology 21, 2–6. 103, 1–21.
Waltham, A. C., Smart, P. L., Friederich, H. and Atkinson, Waltham, A. C. (1989) Ground Subsidence. Blackie,
T. C. (1985) Exploration of caves for rural water sup- Glasgow, 202pp.
plies in the Gunung Sewu karst, Java. Annales de la Wilson, W. L. and Beck, B. F. (1988) Evaluating sink-
Société Géologique de Belgique 108, 27–31. hole hazard in mantled karst terrane. American
Waltham, A. C., Vandenven, G. and Ek, C. M. (1986) Society of Civil Engineers Geotechnical Special
Site investigations on cavernous limestone for Publication 14, 1–24.
GFE-25.qxd 6/27/05 2:30 PM Page 688
25. Loess
Edward Derbyshire and Xingmin Meng
25.1 Introduction across oceans including the Atlantic and Pacific,

with deposition occurring some two or three weeks
Loess is a clastic sediment formed from wind-lain after initial entrainment (Pye, 1987). It is important
deposits predominantly of silt-size particles that to discriminate between regional and extra-regional
have subsequently been altered to varying degrees events (Pye, 1995), as the former occur at lower
by diagenetic processes, principally weathering atmospheric levels and transport coarser silts while
and pedogenesis (soil formation). the latter carry finer fractions at high levels within
The dominant mineral in most of the world’s the troposphere and are readily observed on earth-
loess is quartz (⬃50–70%), with a significant orbiting satellite images (Figure 25.1). In North
content of feldpars and mica, as well as minority China, for example, the driving force is the north-
carbonates and clay minerals. In some loess westerly hemispherical wind system controlled by
regions of the world, however, distinctive source the Mongolian and Siberian high pressure anticy-
materials yield quite different mineralogical popu- clonic systems and the Aleutian low pressure cell.
lations. This is notably the case adjacent to active The regional transport path (source proximal) in
volcanic belts, which inject rock fragments and this case is broadly north-west to south-east, as
glass shards into accumulating loess, as in demonstrated by the systematic fining of loess
Argentina and New Zealand for example. grain size across the Chinese Loess Plateau from
The quartz grains making up the loess ‘skeleton’ sandy loess in the north-west to ‘typical’ loess and
are mainly angular to sub-angular and blade-like in then, in the south-east, to clayey loess (Figure 25.2).
form. Such characteristics, taken together with the The long distance, high altitude transport path
characteristic airfall depositional process, are con- (source distal) is one in which central Asian dusts
ducive to a rather loosely packed, moderate to are carried by the upper Westerlies sometimes as far
poorly sorted deposit lacking any visible stratifica- as the western United States. A number of factors
tion, and with distinctive geotechnical behaviour. influence the location and rate of loess depos-
Surface detachment, entrainment and transport ition, including trapping of dust by vegetation or
of mineral dust by the wind are a function of surface water bodies, a reduction in wind speed and
several variables, notably critical wind speed, turbulence, formation of agglomerations caused by
particle exposure (for example when freshly increase of atmospheric moisture and washing out
deposited — seasonally, and when a vegetation of particles by precipitation (Pye, 1995), the latter
cover is lacking as in dryland regions) and particle process being well characterised by the ancient
size. Dust storms vary in scale between local vor- Chinese descriptor ‘loess rain’. There is strong sup-
tices (‘e.g. dust devils’, ‘willy-willies’, etc.) only a port for the view that long-term vertical accretion
few metres in diameter that raise dust for a few sec- of dust at rates sufficiently high to form loess is
onds to a few hours, and extensive atmospheric tur- likely when there is some surface vegetation cover
bulence associated with air-mass frontal systems (Tsoar and Pye, 1987; Pye and Tsoar, 1987).
that may transport mineral dust for great distances. The process relationship between wind-blown
Air-mass frontal systems associated with the dust transport and loess deposition was recognised
hemispherical wind regimes, notably the upper at least 2000 years ago in China (Liu, 1985),
Westerlies, frequently transport terrestrial dust although this link did not become widely accepted
688
GFE-25.qxd 6/27/05 2:30 PM Page 689
LOESS 689
Figure 25.1 SeaWiFS (sea-viewing Wide Field-of-view Sensor) image, showing a large dust storm over and the
leeward of the Chinese Loess Plateau (lower centre of photograph) on 16 April 1998. The outline of Chinese coast-
line, and the Bohai Gulf and East China Sea (largely covered by broken clouds) may be clearly seen in the lower
right quadrant. The storm is driven by a cold front, with its outburst of cold air from the ‘Mongolian High’, that can
be traced by the leading (eastern) edge of the dust cloud and its continuation as a zone of white cloud running north-
east (towards the upper right of the photograph). Driven by a vigorous depression (low pressure system), some of
the finer component dusts crossed the Pacific Ocean to reach North America on 25 April. (The SeaWiFS Project
(Code 970.2) and the Goddard Earth Sciences Data and Information Services Center/Distributed Active Archive
Center (Code 902) at the Goddard Space Flight Center, Greenbelt, MD 20771, USA, are gratefully acknowledged
for the production and distribution of these data, respectively, within NASA’s Earth Science Enterprise.)
in Europe and elsewhere until the publication of relief forms extending over a wide altitude range.
von Richthofen’s work on the Chinese loess (Pye, Both the thickness of this mantle and the loess
1995; Richthofen, 1877–85; 1882). Loess drapes, mean grain size decrease progressively with dis-
including the classic Chinese deposits, have been tance from the silt sources. Tracts of loess and
attributed to numerous other processes, including associated deposits such as wind-blown sands fre-
lacustrine and alluvial deposition. In addition, quently bear a clear geographical relationship to
there is a strong body of opinion in eastern Europe present or former prevailing wind tracks. Loess is
and middle Asia favouring attribution of the term unstratified and lacks the pebble stringers so com-
loess to aeolian silts only after they have been mon in water-lain deposits. The mineralogical
modified by in situ weathering (the process known composition of loess is often quite different from
as loessification: e.g. Lozek, 1965; cf. Smalley, that of the subjacent rocks or sediments, making
1971). However, a number of landscape and com- an in situ origin unlikely. Accretion of loess by
positional characteristics leave no doubt about the deposition of wind-blown dust can be observed in
wind-blown origin of loess, as Pye (1987; 1995), some parts of the world today (Figure 25.3),
among others, has shown. Typically, loess mantles detailed study of individual, hemispherical-scale
the landscape as a drape, and includes a variety of dust storm events being possible with the advent
GFE-25.qxd 6/27/05 2:30 PM Page 690
drapes are found in many mountain regions of the

world, especially in High Asia (e.g. the Karakoram
(Owen et al., 1992) and the Anyemaqen Mountains
of north-eastern Tibet) but also including loess
in sub-Andean montane basins as in north-west
Argentina (e.g. Sayago, 1995; Iriondo, 1997).
Finally, a loessic silt component has been detec-
ted in some surface soils outside such generally
recognised loess regions (e.g. Catt, 1978; 2001).
Numerous processes contribute to loess forma-
tion, but provenance (including particle gener-
ation as well as the geology of the source regions),
transportation by the wind, deposition and post-
depositional modification, especially by the
Figure 25.2 Gradational change in the particle size climatically driven processes of weathering and
population from north-west to south-east across the pedogenesis, are of particular importance. Opinion
Loess Plateau of China, expressive of the two major varies on the question of the relative effects of
factors of dominant regional winds and the humidity
gradient from arid in the north-west to sub-humid to these processes on some of the bulk properties
humid in the south-east (after Liu et al., 1964). of silt deposits. Wentworth (1933) favoured the
transport mechanism as the major determinant of
the grain size. More recently, attention has been
of high quality satellite imagery (Figure 25.1). directed particularly to wind transportation as a
Finally, the particle size distributions of typical sorting mechanism (e.g. Tsoar and Pye, 1987) and
loess and modern, source-adjacent aeolian dusts a case has been made for regarding particle
are identical (Pye, 1987). formation as at least as important a determinant
Siltstones are found throughout the geological of the grain size of silts as the transportation process
column, but conditions in the Quaternary (approxi- (Assalay et al., 1998). Fookes and Best (1969)
mately the last 2 Myr) appear to have been par- favoured processes operating during deposition as
ticularly conducive to silt generation and its the principal determinant of the engineering pro-
accumulation on the land surface, especially as perties of loess. In extreme cases of modification
alluvial silts and as loess. Although aeolian silts of a unit of aeolian silt by advanced pedogenesis,
have been found on all the world’s landmasses, the dominant control on bulk properties may be
large concentrations of loess are associated with largely attributable to post-depositional events
particular regions of the globe. Of the estimated (see below). It follows, therefore, that careful
10% of the earth mantled by loess, by far the account should be taken of the process history of
greatest volumes are to be found in the great accu- individual loess units during site assessment.
mulations making up the Eurasian (western and When loess accumulates to a thickness that
central Europe, central Asian and north China), effectively masks some or all of the morphology
Siberian, North American and South American of the underlying terrain, it assumes a series of ter-
loess regions (Figure 25.4). Less extensive, rain types that reflect its distinctive composition,
although regionally important deposits are also structure and behavioural properties, especially in
found in New Zealand and Alaska. Also less exten- response to water and seismic shock.
sive, but more variable in particle size composi-
tion, are the silt-rich deposits of fine sands
associated with the world’s sandy deserts, notably 25.2 Loess Terrain
around the Sahara (Nigeria, Tunisia, Libya,
Israel, Iran), frequently described as loess in the lit- The great loess deposits of the Earth mantle
erature. In addition, thin, discontinuous loess extensive areas of variegated terrain, but thicken
GFE-25.qxd 6/27/05 2:31 PM Page 691
LOESS 691
Figure 25.3 The year-round pattern of dust falls at four stations from WNW (top) to ESE (bottom) along the Hexi
Corridor (west of the Chinese Loess Plateau) for the four-year period 1988–1991 (after Derbyshire et al., 1998).
GFE-25.qxd 6/27/05 2:31 PM Page 692
significantly in broad basins and on gently surfaces, locally masking irregularities in the under-
sloping plains and plateaus. The thickest loess on lying bedrock. Thick loess (tens to hundreds of
Earth has accumulated in the continental interior metres thick) is found on plains and within basins
basins of Eurasia, as part of the great loess belt where it may completely mask the subjacent relief;
that stretches from the Atlantic coast of Western such plateau-like loess has extensive planar sur-
Europe to the Yellow Sea (Figure 25.4). Within this faces, interrupted to varying degrees by the action
10 000 km long tract, the Chinese loess reaches a of water, and other processes including tectonics.
thickness of c. 500 m towards the western margin The major morphological features found in thick
of the Loess Plateau, and the central Asian loess is loess terrain are thus plateaus, fluvially cut valleys
about 200 m thick in the Tajikistan–Uzbekistan and interfluvial ridges and, with more advanced
region. degradation, linear ridges and conical hills. Long
At the macroscopic scale, the geomorphology slopes in loess exposed by rapid river incision,
of loess terrain is relatively simple, consisting of a faulting and related mechanisms are vulnerable
limited number of elements. Unlike the distinctive to degradation by landslides, subsidence, mass
suite of depositional landforms associated with flowage and associated processes, especially in
aeolian sand, for example, the surface forms of environments with high-magnitude, low-frequency
loess landscapes rarely reflect dominant wind rainfall events and seismic activity. Complexity is
direction. Thin loess (a few metres in thickness) introduced in loess terrain by certain ground and
commonly drapes rocky footslopes and terrace surface water regimes, giving rise to distinctive
Figure 25.4 The distribution of thick loess in Eurasia, indicated by the darker shading (abstracted and simplified
from the map in Rózycki, 1991).
GFE-25.qxd 6/27/05 2:31 PM Page 693
LOESS 693
and sometimes complex morphology at the field cross-profile as a result of an irregular annual
scale. In general, complex loess morphology is a cycle of flood discharges in the summer wet sea-
function of the bulk properties of loess, and it is son followed by a long winter dry season in which
this relationship that is emphasised here. gully-side slopes relax and readjust (Figure 25.6),
The regional-scale morphology of loess is best in some regions with the help of additional pro-
illustrated by the thick loess terrains in central and cesses such as freeze–thaw and snowmelt. More
eastern Asia. The Chinese Loess Plateau is a clas- stable loess slopes in areas protected from active
sic region exemplifying most of the characteristics undercutting show varying degrees of departure
arising from the fundamental relationship between from the model slope set.
the bulk properties of loess and the morphology The dominant slope facet on many stable liang
of loess at all scales of enquiry. The macroscopic and mao forms in North China frequently lies in
morphology of this region is traditionally charac- the range 35–45. Aside from the flat-topped
terised using the descriptive native terms ‘yuan’, (yuan) surfaces, many slopes in this region thus
‘liang’ and ‘mao’. Yuan are flat plateaus in thick fall into the third category of Lohnes and Handy.
loess, best seen in southern Shaanxi and western Derbyshire et al. (2000), working on the very
Gansu Provinces (Figure 25.5a). Liang are ridges, thick loess in the western part of the Loess
common in northern Shaanxi and Shaanxi Plateau, showed that the slope population is
Provinces (Figure 25.5b). A mao is a rounded almost normal in distribution, with a modal
loess hill; mao hills often occur in liang country, category at 36–43, and a principal mode of 39.
and dominate the landscape only east of the There are conflicting views on the question of the
Yellow River (Hwang He) in Shanxi Province. optimal gradient for cut slopes in loess. Vertical
Although many view this set of forms as an ero- slopes, believed by some to minimise erosion by
sive sequence, neotectonics has certainly played a water, are subject to shear failure of their upper
role in its development. A fourth morphological parts in uniform, thick loess ( 10 m) as an
type, termed ‘cone and plain’, was recognised on unloading response not necessarily related to
the western edge of the Loess Plateau north of rainfall events (see ‘Mass movement types’
Lanzhou (Derbyshire, 1983a). Broadly similar below). In North China, slopes are commonly cut
landform suites occur in other regions of thick at grades of 1 : 1.2 to 1 : 0.25, depending on age,
loess, including central Asia and North America, slope height, local mean annual precipitation and
albeit often on a smaller scale. loessic facies (notably primary air-fall loess and
Lohnes and Handy (1968) proposed a general reworked material — loessic alluvium). The com-
model for slope form and type in loess terrain that monest grade used by engineers in the Loess
takes into account the bulk behaviour of loess and Plateau of China is 1: 0.75 (c. 53, which approxi-
the surface and sub-surface processes, notably the mates the basal shear plane angle of Lohnes and
hydrology. Fluvial incision by gully systems leads Handy, 1968).
to two common slope elements. First, there is a The relationship between formative wind sys-
steep upper facet with gradients of 70–85 that tems and loess landscapes is subtle and, in detail,
reflects sets of stress relief joints induced by quite complex. A distinctive zoning in the particle
tensional stresses near the ground surface. Retreat size of loess from source regions to the most dis-
of this upper cliff is largely by slab fall and slid- tal loess-covered terrain has been demonstrated in
ing. The variously disaggregated slabs of loess many loess regions of the world, including China
mantle the basal shear plane sloping at 51–59o, and Tajikistan–Uzbekistan. In China, the north-
depending on the values of and c. Accumu- west to south-east gradient in climate from the
lation of these slabs and blocks may obscure desert basins west and north-west of the Yellow
the lower shear slope at, or close to their angle River, and the dominant north-westerly winds in
of repose (38), explaining the large number of winter and spring, are reflected in a progressive
observed slopes in the range 34–38. Many active fining of the mean grain size of the surface
gullies in thick loess terrain maintain a rectilinear aeolian deposits from sandy silts near the sources to
GFE-25.qxd 6/27/05 2:31 PM Page 694
Figure 25.5a Thick loess mantling the landscape to form a plateau (yuan), and dissected to varying degrees to
produce deep valleys. Zhenyuan, Gansu Province (reproduced from Wang and Zhang, 1980).
Figure 25.5b Ridges (liang) in the thick loess near Ansai, Shaanxi Province. Note the alternation of groups of
thick palaeosols (dark) between loess units, the deep incision and the parallel valley-side gullies (reproduced from
Wang and Zhang, 1980).
GFE-25.qxd 6/27/05 2:31 PM Page 695
LOESS 695
Figure 25.6 Active erosion of loess by a tributary of the Yellow River (Hwang He), c. 50 km SW of Lanzhou city,
western Loess Plateau. The channel is characteristically flat-floored, with classic ‘upper cliff and lower debris slope’
morphology on the channel sides (centre and right). There has been notable re-entrainment of loessic silt by fluvial
action, as can be seen in the bedded silts in the bluff (‘loessic alluvium’: left of photograph).
wind-lain silts (‘typical’ loess) to progressively source areas are characterised by a mixed suite
more clay-rich loess towards the south-east (Liu, of landforms including sand plains and dunes,
1985). Similar systematic changes in sedimentary extensive (up to 50 km in diameter) silt-rich allu-
properties with distance from source have been vial fans and silt river terraces and plains. Mixed
demonstrated for loess units of several ages back gravels, sands and silts are commonly differentially
to c. 7.7 Myr (Ding et al., 1998). In addition to eroded by the wind, frequently leaving a protective
this effect of differential sorting during transport, and stable surface lag known by the Mongolian
enrichment with clay minerals is progressively word gobi (stony desert). The regional gradient in
important south-eastwards as rates of weathering mean particle size is also reflected in a progressive
and pedogenesis increase as climatic conditions change in landform types. On some piedmont
become warmer and wetter. fans within or close to silt source areas, locally
Considerable volumes of silt enter the geomor- reworked and alluvially re-deposited loessic silts
phological system from the seasonal melting are found inter-bedded with sand and gravel units.
cycle in both high, glaciated mountains and lower, In rare instances, relatively thin drapes of sandy
seasonally snow-covered hills. Using northern loess show linear forms parallel to dominant wind
China as a basis for a general model (Figure 25.7), directions, and some small barchan-like forms
GFE-25.qxd 6/27/05 2:31 PM Page 696
Figure 25.7 A generalised model of a mid-latitude loess landscape, based on northern China, showing glacial,
nival and seasonal alluvial sources, entrainment, the relationship to sand dune desert and plateau-like accumulation
of thick loess and its dissection.
have been reported from the mountains of central a trigger causing failure of loess slopes (Dijkstra
Asia (Keš, 1972). Some thin ( 1 m) sandy loess et al., 1993).
drapes mantle steep ( 20) rock slopes near Given the high rainfall intensities required to
source areas, as in the eastern Kunlun Shan in saturate and mobilise the surface of sparsely vege-
Tibet, attesting to its direct aeolian accretion tated thick loess (Muxart et al., 1994), slopes
(Figure 25.8). With increasing distance from the arising from mass slurrying of loess are relatively
sources, typical loess may accumulate to great rare, but rilling and gullying become progres-
thicknesses. However, loess thickness is affected sively more notable in the finer, clay-rich (lee-
significantly by the underlying geomorphology in ward) loess zone, as exemplified by the southern
the depositional zone. The thicker loess areas in and eastern Loess Plateau of North China. The
northern China, for example, have been correlated combination of joint enlargement by water, suffos-
with broad, usually structurally defined bedrock ion, development of sinkholes and major systems
basins (Liu, 1985; Zhu and Ding, 1994). It is in of large pipes and associated subsidence, together
such areas that the thick loess-palaeosol sequences, with gravitational failure by toppling, falling and
as seen in Asia, are cut by river systems, exposing sliding (as thin loess zones above clay-enriched
long slopes subject to several mass failure bedrock or palaeosols give rise to localised
processes, so completing the model landform liquefaction of the base of the loess), are the
sequence (Figure 25.7) Outside such favourable major land-forming processes in loess terrain.
regions, as in plateau and mountainous bedrock ter- Rivers that drain thick loess landscapes are
rain, the deposition–erosion balance may be more among the most turbid on earth.
variable, leading to thinner loess and greater fluvial The tendency to collapsibility in loess affects
reworking, but often adding earthquake shock as plain and plateau surfaces well away from
GFE-25.qxd 6/27/05 2:31 PM Page 697
LOESS 697
Figure 25.8 Accretion of loess of uniform thickness (25–40 cm) draping a rectilinear rock slope in the eastern
Kunlun Mountains, Tibet. Sub-parallel rills eroding the loess drape can be seen on the slopes above and to the left
of the Kirghiz herdsmen.
hillslopes. Movement along active faults certainly vary in important respects with age, location, and
plays a role in generating large and potentially site. Determination of the depositional and weather-
destructive fissures in flat loess terrain, although ing history of loess underlying a site may thus
severe seasonal rainfall has similar capability as provide a deeper appreciation of the range of bulk
shown by fissures 1 km long and up to 14 m properties and the likely behaviour of specific
wide in China’s Shaanxi province. Such fissure loess units.
systems, often enhanced by extraction of ground- The history and type of source rock (e.g. gran-
water for domestic and industrial purposes, are a ites, gneisses, vein-quartz) dictates the concentra-
continuing threat to the great historical city of tion of defects within quartz, the dominant mineral
Xi’an, where subsidence in the period 1962–83 in much of the world’s loess. These ‘Moss defects’
totalled 777 mm, with measured annual extreme may control the minimum size attainable during
values of up to 123 mm/yr (Sun, 1988). natural crushing and abrasion (the ‘comminution
limit’) of quartz (Moss, 1966; Moss and Green,
1975), thought to lie in the range 10–30 m. This
25.3 Engineering properties of loess may explain differences between loess derived
from the Laurentian and Baltic shields, and loess
The behavioural properties of loess are influenced adjacent to the Himalaya and Tibet. Other regional
by several factors, past and present climate being distinctions arise from different process histories,
particularly important. Often regarded as a homo- such as the dominance of glacial comminution as a
geneous material, the bulk properties of loess may source in the mid-western loess of North America.
GFE-25.qxd 6/27/05 2:31 PM Page 698
The adobe of the desert south-west of the United is 15–20 m (6 phi). In northern Africa, aeolian
States (Rogers and Smalley, 1996) is an example silts are much coarser than the Asian average,
of a non-glacial loess. The combination of a suite being dominated by the fine sand size fraction, as
of processes including tectonic crushing, frost and in Tunisia (e.g. Coudé-Gaussen, 1987) and Libya
salt weathering and glacial grinding, as in the High (Assallay et al., 1996), while they are finer than
Asia region including Tibet and the Himalayan the central Asian–Chinese average in northern
tract, is the source of the greatest piles of loessic Nigeria (McTainsh, 1987). Assalay et al. (1998)
silts on Earth. provide a detailed discussion of the possible role
of silt formation processes in contributing to such
Particle shape diversity in silt sizes from region to region.
The quartz grains making up the skeletal frame- The range of variation in the particle size
work of much of the world’s loess show a mix- envelopes of loess from central China, Tajikistan,
ture of forms including equant, tabular and blade Uzbekistan, Italy, France and North Africa is con-
shapes, with the latter two being dominant siderable (Derbyshire et al., 2000; Figure 25.9a;
(Smalley, 1966a; Krinsley and Smalley, 1973; Figure 25.9b). Fine aeolian deposits referred to as
Derbyshire, 1988). Rogers and Smalley (1993) loess from some parts of the climatically diverse
obtained probabilities of 1% cubic, 26% tetrago- region of Western Europe and North Africa may
nal/disc and 73% blade shapes, the axial ratios of contain up to 70% fine sand grade and only
the latter category being 8 : 5 : 2. Such grain mor- c. 20% silt (Figure 25.9c); elsewhere in Europe
phology is typical of loess in Asia and parts some deposits regarded as loess contain 50% clay
of eastern Europe, for example, the occasional grade ( 2 m) material. Inter-regional compari-
rounded or irregular forms probably being sons of grading envelopes for loess from China,
attributable to partial weathering of non-quartz
components and the varying extent of cementa-
tion or clay minerals (Derbyshire et al., 1988).
Most loess particles are transported in suspension
and thus lack the edge-rounding and ‘frosted’
surface texture so characteristic of aeolian sand
grains.
ND
SA
%
Particle size
%
CL
sp
The term ‘loess’ has been applied rather loosely to
AY
a broad range of silt-rich sediments. Many of

these are calcareous, with a broad particle size s
sp
sees ee s e
envelope in which clay-size content varies q
g es q
gg eeeeeeqe e
ee e g sse
between c. 8 and more than 30%, and sand may s e see eeg e
eeg
ee g e
q e gg e
exceed 60%. Published mean grain sizes of mater- g
ial described as loess range from at least 5.2 to

% SILT
8 phi, with sorting coefficients (Folk and Ward,
1957) between 3 and 1.3. Views on the mean or Gansu, Shaanxi and Qinghai
modal particle size of loess have been greatly g and e : Lanzhou loess (Gansu)
s: Luochuan loess (Shaanxi)
influenced by regional factors. For example, in sp : Louchuan palaeosol
regions dominated by ice sheet glaciation in the q: Qinghai silts
no. = 55
Quaternary (notably Western Europe and North
America) the modal grain size of loess appears to Figure 25.9a Ternary diagram showing typical closely
average about 30 m (5 phi). In central Asia and clustered envelope for the Chinese Loess Plateau (with
North China (where loess often occurs along sand some divergence in Qinghai, Tibet) (after Derbyshire
desert margins), in contrast, the modal grain size et al., 2000).
GFE-25.qxd 6/27/05 2:31 PM Page 699
LOESS 699
ND
ND
SA
SA
%
%
U
%
U
CL
CL
U
U
AY
AY
UU
M B B U
M LL M U
LM UU
U
OM B
OO r R UUU
O U
M O H
L OO B R
R
H V BK IARS
HK NII R Z
HDH JA
H NSFS F Z
KS R K SKZ
HH NN
NSK S JK Z
H A C
H TJ Z
A J
C J
T
C
% SILT % SILT
H Chirchick (Tashkent) W. Europe and Tunisia
O Orkutsai (Tashkent) B Barlassina (Italy)
L Keles (Tashkent) C Copreno (Italy)
T Karatau V Bella Vista (Italy)
M Karamaidan palaeosols R Rocourt (Belgium) (r = Rocourt palaeosol)
D Karadarya (Samarkand) H Harmignies (Belgium)
no. = 26 K Kesselt (Belgium)
N Nuth (Netherlands)
Figure 25.9b Ternary diagram showing consistent, S Saint Roman (Normandy)
but less closely clustered envelope for the loess of I Iville (Normandy)
F Saint Pierre les Elbeuf (Normandy)
Tajikistan–Uzbekistan (after Derbyshire et al., 2000). J Jersey (U.K.)
A Kent (U.K.)
U Primarette (SE France)
Western Europe and North Africa present some Z Saint Prim (SE France)
T Matmata Plateau (Tunisia)
interesting contrasts (Figure 25.10).
no. = 72
Most loess is poorly to very poorly sorted when
measured in the laboratory using conventional Figure 25.9c Ternary diagram showing the much
indices such as those of Folk and Ward (1957) more divergent particle size pattern found in Western
(cf. Derbyshire et al., 1993), an apparently anom- Europe and Tunisia (after Derbyshire et al., 2000).
alous outcome in view of the apparent effective-
ness of wind as a sorting agent. Once deposited, the Late Pleistocene loess (Figure 25.11a). Mean
many of these silt-sized aggregates are broken versus sorting co-plots reveal several loess groups
down and redistributed (dispersed or redeposited ranging from better-sorted and coarse sandy loess,
in higher concentrations) in response to overbur- through the ‘typical’ loess (the ‘D1’ population at
den pressures, shearing, leaching and diverse reac- 25–30 m of Jefferson and Smalley, 1995: cf.
tions during in situ weathering. Thus, standard Browzin, 1985), to fine-grained and very poorly
granulometric methods present the constituent sorted palaeosols (Figure 25.11b).
rather than the dynamic grain size of loess
(cf. Derbyshire et al., 1980, p.185). Soil fabric
Particle size distribution also varies with age, The granular skeleton of silt-sized particles
i.e. the depth-down profile. The mean grain size of usually show variable amounts of finer parti-
the North China loess lies in the range 10–60 m cles occurring as both loosely adhering clusters
(phi-value range 5.0–8.0; Derbyshire et al., 2000: and cements, as buttresses and bridges linking
cf. 25 m mode of Jefferson and Smalley, 1995). the skeletal grains, and as occasional silt-sized
The older loess formations are clearly finer than aggregations. Fabric refers to all directional
GFE-25.qxd 6/27/05 2:31 PM Page 700
A
4.0 W. Europe & Tunisian = 71
A
China n = 60
U U
B U U
U
3.5 A U UU U U
U
U U
A B B
B G
A V Y Y B
X
A Y
3.0 AY XY
K HA B G
Sorting coefficient
G
A R G G GG X G
H S K
G Y GG G
R AK Y X G
I I IV A X G
X G Y Y X
S NY R A GG GX G X Y
2.5 NC S G
K RK I G GG G G
HS Q Z GG GGG G G
G G
S K G GG X X
G
Z G Q
2.0 T G G G G
C S ZZ
J
J A Q K G
J Z G
G
1.5 J
J
A
J
1.0
3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5
Mean ¯
Typical Loess Chinese convention 'Typical' Loess

Sandy Loess Palaeosols
Figure 25.10 Plot of mean particle size versus sorting coefficient for samples of loess and palaeosols from China
and Western Europe. The consistency of the envelope for Gansu province (‘G’), China, contrasts with the rather dif-
fuse Western European envelope (site nomenclature as in Figure 25.9c). The grain size median zone for ‘typical’
Western European loess (left, horizontal axis) and western Chinese Loess Plateau loess (bold print) are also shown.
properties of a sediment (Derbyshire et al., 1976), grain coatings, and redistribution of clay-sized
although particle orientation and packing (Rogers particles often results in aggregation and clay coat-
et al., 1994a) have attracted most attention in ings on larger grains (cf. Grabowska-Olszewska,
geological and geotechnical studies of loess. 1988; Derbyshire and Mellors, 1988; Billard
Fabric is influenced by several factors including et al., 1993).
grain shape, particle size distribution and the The sequence dispersion — redistribution–
stress history of the material. cementation increases fabric compaction and
Following deposition of wind-blown mineral strengthens inter-particle bonds in some hori-
dust as a loosely packed agglomeration of angu- zons. In addition, the process of normal consoli-
lar, platy quartz grains and silt-sized aggregates dation is associated with inter-granular shearing.
of fine silt and clay-sized material, several surface Such disruption of aggregates in pore fillings
processes progressively affect changes in the may further strengthen loess structure. Increasing
deposit. Wetting and drying induce flocculation normal loads or porewater pressures lead to irre-
of constituent fines, as cationic concentration versible collapse of the characteristic edge-to-face
increases with any decline in porewater content and face-to-face fabrics considered syngenetic
(Grim, 1953). The fine silts and clay-sized grains in undisturbed loess (Derbyshire and Mellors,
making up aggregates are drawn to pore margins 1988; Gao, 1983; Derbyshire et al., 1988).
by porewater menisci, yielding fine-particle Eluviation of the finer particles tends to break
bridges, buttresses and adhering aggregates down the characteristic clay bridges, which do
(Derbyshire, 1984). Soluble salts become redis- not reform in the horizons below. Loess fabrics
tributed to form inter-particle cements and silt thus show both intra- and inter-site variation.
GFE-25.qxd 6/27/05 2:31 PM Page 701
LOESS 701
9 7 5 3φ described and illustrated using both optical thin

100 sections (micromorphology) and scanning elec-
tron microscopy (Derbyshire et al., 1995a).
90
The loess of Western Europe also varies with
80 age and geographical location. Loosely packed
fabrics in coarse to medium angular silts, with little
70 clay mineral content and limited, often localised
60
cementation (comparable to much Asian loess) can
be found at many sites in continental Europe
% finer
50 (Figure 25.12). However, fabric varies with cli-

matic diversity and past variations, shifting soil
40
water regimes from saturation to desiccation
30 (including freeze-drying associated with cryotur-
Lishi
bation: van Vliet-Lanoë et al., 1984), bioturbation,
20 leaching and re-deposition, snow meltwater infil-
Wucheng
tration, mineral weathering, natural loading and
10
Malan unloading, and reworking by running water and
0 mass movements. Particle packing, the nature
2 8 16 64 250µm and degree of cementation, and the content of clay-
Clay Silt Sand grade material provide a short but useful list of
comparative indices. An additional variable in
Figure 25.11a Particle size variation with age and
degree of soil development in the loess of the western loess fabrics is preferred fabric trends (anisotropic
Loess Plateau of China: particle size envelopes for Malan fabrics). These vary from visible lamination (as in
(youngest), Lishi and Wucheng (oldest) loess units (after alluvially re-deposited loessic silts: Derbyshire
Derbyshire and Mellors, 1988). 1983b) to strongly parallel particle fabrics, such as
those generated by cyclic freezing and thawing
Malan (Figure 25.12). The quartz grains in the loess of
Lishih and Wucheng Normandy, France and the nearby Channel Islands
Palaeosols
appear rounded because of an almost complete
Sorting coefficient
3
clay encrustation rich in Si, Al and Fe (Figure
25.13a). Limon à doublets, a loess facies that
2 Malan
Lishih
Wucheng
Ho series
2001 series
2561 series
Shaanxi
1
5 6 7 8
Mean diameter
Figure 25.11b Discrimination, using a co-plot of mean

versus sorting coefficient, between older loess with
palaeosols, and the younger loess (after Derbyshire and
Mellors, 1988).
The progression of fabric changes in the loess

and intercalated palaeosols from the drier west-
ern margins of the Chinese Loess Plateau to
the warmer and more humid east have been Figure 25.12 (Continued).
GFE-25.qxd 6/27/05 2:31 PM Page 702
Figure 25.12 (a) Scanning electron micrograph of Last Glacial loess from Göllersdorf, c. 30 km NW of Vienna,
Austria, showing a fabric skeleton made up of very angular quartz silt particles relatively poor in clay minerals. The
rather coarse grain size, angularity of grains and generally ‘clean’ appearance are a product of proximity to a source
characterised by a grain crushing geosystem (glacier and ground ice in the nearby Alps), as well as a freely draining
fabric. Sample provided by David Boardman. (b and c) Fabrics in loess produced by the development of ground ice.
Both micrographs are from samples located 1 m below the surface at Kesselt, Belgium. This calcareous loess dates
from the upper part of the Last Glacial. The distinctive fabric produced by segregated ground ice development is
shown. Sub-horizontal lens-like structural units are delineated by overlapping, slightly curved, sub-horizontal dis-
continuitities that mark the surfaces of former segregated ground ice lenses. (c) In this detailed micrograph (scale bar
20 m), the size sorting affected by ice segregation is clearly evident, with coarser silts at the top and bottom of the
micrograph. In between (lower centre), fine silts densely packed by ice-growth compression, and the characteristic
thin clay film draping the top of this compact fine silt unit are clearly visible. Sample supplied by Sanda Balescu.
occurs across Europe as far as the Russian Plain,

consists of thin, gently dipping alternations of clay-
rich brownish laminae and grey, clay-poor layers
between 1 mm and 1 cm in thickness (Figure
25.13b). The grey layers often have high void
ratios and lack clay bridges between silt grains
(Figure 25.14), while both clay bridges at silt grain
contacts and clay-coated silt grains are common in
the brown layers, and sometimes more concen-
trated at the base (Figure 25.14). This distinctive
loess is considered to be a periglacial facies.
The dominant feature of the macrofabric of
loess is a system of sub-vertical joint sets. These
arise from several processes including tectonic Figure 25.13a Loess from Green Island, Island of
Jersey (UK), off the coast of Normandy. Although the
shock and dilation (associated with erosive particle size is similar to other lowland loess in Western
unloading, natural and artificial), and are of Europe, the fabric is distinctive. The silts appear sub-
importance because of the effects on permeability rounded, but this arises from the presence of widespread
and bulk strength. At shallow depths, plant roots surface mantle on the silt grains rich in Si, Al and Fe (as
may directly affect vertical permeability and, in can clearly be seen in the upper right centre, where the
coating has been broken during sample preparation). The
giving rise to sub-vertical contraction cracks, they absence of any scattered or aggregated fine silt and clay
may constitute an indirect factor. The relative particles/aggregates indicates severe alteration of this
importance of these factors varies regionally in shoreline loess.
GFE-25.qxd 6/27/05 2:31 PM Page 703
LOESS 703
Figure 25.13b An exposure of the distinctively banded limons à doublets at Port Racine, Normandy, France.
Figure 25.14 Limon à doublets from Belcroute, southern Jersey, UK Channel Islands: (a) dense lamina of the
finer, more compact, clay-rich ‘brown doublet’ (lower part of photograph) compared to the more open fabric of the
‘grey doublet’ above it; (b) Sharp contact between grey doublet (top of photograph) and brown doublet. Some
cemented bridges occur in both types, but the greater void ratios in the grey layers are clearly evident.
response to climate, seismic conditions and the such as that found in parts of dryland Asia decline
erosion regime. In general, rates of vertical per- rapidly with depth. In North China, Kv : Kh ratios
meability (Kv) in loess are higher than horizontal decline with depth, there being little difference
permeability (Kh) rates; reported differences are below 100 m (Derbyshire et al., 2000). At the
high in relatively thin loess in which there is tree same time, marked variations in both Kv and Kh
growth, but differences in thick, massive loess, values and ratios arise from the presence of
GFE-25.qxd 6/27/05 2:31 PM Page 704
sub-horizontal discontinuities in the loess, notably occurring as concretions (‘loess dolls’), pore lin-
intercalated palaeosols. ings, encrustations and inter-granular cements,
are an important source of strength in loess
Mineralogy (Figure 25.15). Variations in calcium carbonate
In addition to quartz, a range of subsidiary min- content in the loess of the western Loess Plateau
erals is present in loess, notably feldspars, micas, of China are greatest in the palaeosol horizons.
calcium carbonate and gypsum. Various heavy Billard and Muxart (2000) have shown that dry-
minerals (density 2.7) also occur, but often ing of the palaeosols within loess successions
constitute only a few percent by weight of the leads to supersaturation of the soil solution and
total mass. Calcite is an important constituent of precipitation of carbonate components and
much loess and certain horizons within loessic cementation bonds. This strengthens the loess
palaeosols, occurring as both clastic and second- skeleton and the aggregates of finer particles in
ary types. Secondary carbonate precipitates, particular. Leaching of such cements, and any
associated clay minerals, is inhibited in semi-arid
conditions, so preserving the openwork fabric.
Carbonate contents also tend to rise with increas-
ing loess age (Derbyshire and Mellors, 1988;
Derbyshire et al., 1991). Such ‘strength stratigra-
phy’ is revealed in some cone penetrometer
profiles (Figure 25.16). Pedogenic gypsum occa-
sionally occurs as asymmetrical void fills (Kemp
et al., 2001) and sometimes as small concretions
(Derbyshire, 1983a) in the semi-arid parts of the
western Loess Plateau of China.
Distinctive mineralogy is found in loess
enriched with volcanic tephra, as in parts of South
America. Near Córdoba, close to the western
margin of the Argentine pampa, the loess has the
following mineralogical composition: quartz
27%, volcanic glass 20%, altered feldspar 20%,
plagioclase 14%, lithic fragments 10%, potassic
feldspar 7% and biotite 2% (Iriondo and Kröhling,
1995). Loess situated at greater distances from
volcanic centres, such as in the sub-Andean basins
of the north-west of the country for example,
contain only 3–6% of volcanic fragments and
glass shards (Ovejero, 1980).
Dissolution of salts and carbonate precipitates
during eluviation by percolating groundwater
considerably reduces soil strength, a factor affect-
Figure 25.15 These two micrographs (from Longpré- ing the known high sensitivity values of loess
les Corps Saint, Picardy, France) have a common centre (Derbyshire et al., 1993). In the dry state, carbon-
point. They provide a good example of a Western ates act to strengthen loess, but at higher moisture
European calcareous loess. It consists of a skeleton of contents they may be leached out, eventually con-
coarse angular to sub-angular silt grains with calcite tributing to collapse of the loess skeleton (Wang
cementation, a moderately open structure, and well-
developed, carbonate-strengthened inter-silt buttresses and Derbyshire, 1994).
and bridges made up of aggregates of very fine silt and Quartz is also frequently the principal mineral
clay grade particles. Sample supplied by Sanda Balescu. in the finest ( 2 m) loess fractions, occurring
GFE-25.qxd 6/27/05 2:31 PM Page 705
LOESS 705
Litho- Thick- Depth Cone Resistance (x1000KPa)

logy ness (m) Description
(m)
2 4 6 8 10 12 14 16
2
6.2 Re-worked loess
4
6
Malan loess: carbonate nodules
8 increase with depth
5.0
10
12 Malan loess: distinctive carbonate

3.2 nodule layers found between
14 13.0 and 13.3 m; and at 14 m
2.1 Malan loess: fewer carbonate

16 nodules
2.8 Malan loess with abundant

18 carbonate nodules
20
Carbonate nodule layers found
at 20 and 22.2 m.
22 Carbonate nodules increase
6.7
below 23.5 m.
24
Figure 25.16 Penetrometer profile to a depth of 24 m in the Late Pleistocene loess near Lanzhou city, western
Chinese Loess Plateau. The general increase of cone resistance with depth is clear, much of the variance arising from
the presence of carbonate-rich horizons in the intercalated palaeosol units.
in a range from lithic fragments to biogenic opal. Plateau, although the present and past north-
Kaolinite and illite are the most frequently occur- west to south-east climatic gradients in China
ring clay minerals, with chlorite, vermiculite, are broadly reflected in the clay mineralogy
smectite and several mixed layer clays, but there (Derbyshire, 1983a; 1983b). In Argentina, the
are some significant regional variations ari- dominant clay mineral appears to be illite in both
sing from the dual influence of the rock composi- the Pampean (Karlson et al., 1993) and the north-
tion in the dust source area and the effects of western loess regions, with both kaolinite and
post-depositional weathering. The pioneering smectite being minority minerals (Sayago, 1995).
work of Frye et al. (1962) showed that the amount Smectite/illite is prominent in the loess of Poland
of montmorillonite in the Peoria (late Last (50–80%, being highest in the older loess units),
Glacial) loess in the United States varies system- with illite (up to 40%) and minor kaolinite
atically depending on the different sources pro- (2–5%: Grabowska-Olszewska, 1988).
vided by different tributaries of the Mississippi Determination of the mineralogy of the clay size
River, rising to as much as 70% of all clay miner- fraction is an important precursor to geotechnical
als present in the loess. The range of variation is testing. For example, the tensile strength of loess
generally less than this on the Chinese Loess is significantly affected by the presence of clay
GFE-25.qxd 6/27/05 2:31 PM Page 706
particles in response to a complex of variables. 1.48–1.69 Mg/m3 in Kent, England (Derbyshire

These include soil moisture content and the clay and Mellors, 1988), 1.57–1.77 in south-west cen-
cohesion (including the degree to which clay min- tral Iran (Fookes and Knill, 1969), and 1.36–1.46
eral particles adhere both to each other and to the in Libya (Assallay et al., 1996). In New Zealand,
silt-sized primary mineral particles making up the bulk densities range more widely in response to
silt skeleton: Smalley, 1966b; Derbyshire, 1983b; the variable content of allophane, halloysite, glass
Tan, 1988; Derbyshire et al., 1994). and other volcanic materials in the loess, as well
as to different climatic environments (Stevens,
Natural moisture content 1988; Eden and Hammond, 2003).
Ambient natural moisture content in loess is
broadly controlled by the climatic regime, and Atterberg limits
specifically the annual water balance, but it is The range of liquid and plastic limits in loess
also affected to varying degrees by the detailed reflects the clay content, the highest values in most
geomorphology, and the loess sedimentary char- regions occurring in the palaeosols. In the western
acteristics including particle size, percentage Chinese Loess Plateau, the lowest mean liquid (LL)
clay content, degree of consolidation (age and and plastic limits (PL) (28.4% and 18.9% respec-
depth) and the surface and sub-surface hydrology. tively) are found in the youngest, and coarsest,
Because of the complexity of patterns of hydraulic loess. Values for LL and PL in the older and rather
conductivity in some loess, the relationship finer-grained units are c. 27% and 18% respec-
between natural moisture content and loess thick- tively, with values in the more clay-rich palaeosols
ness is not a simple linear function. Loess of dif- being c. 30% and 20%, respectively (Derbyshire
ferent ages and different degrees of consolidation et al., 2000). In more humid parts of the world,
may have widely different mean natural moisture including the central and south-eastern Chinese
contents. For example, Lin (1995) and Derbyshire Loess Plateau, values are higher (LL 29.9–31.1%;
et al. (2000) provide mean values of 7.4–20.1% PL 18.6–19.4%: Lin, 1995). Reported values of LL
for the Loess Plateau of China; Phien-wej et al. for the Vicksburg loess of the lower Mississippi
(1992) quotes 8–12% for the ‘fine sandy silt’ of valley (USA) fall in the range 28–43%, with a PL
Thailand; Assallay et al. (1996) cites 2–6% for range of 25–29% (Krinitzsky and Turnbull, 1967).
North African loess (Libya); Feda (1966) gives The range of LL and PL values for loess in the
14.8–17.0% for the Czech Republic; Lautridou Mississippi catchment is 23–58% and 17–29%. In
(1993) cites 17–25% for Normandy (France), and contrast, the loesses of Alaska and Washington
Grabowska-Olszewska (1988) quotes 3–26% for State are of very low plasticity (Sheeler, 1968). LL
the diverse loess deposits of Poland. and PL ranges in the loess of south-east England are
30–46% and 17–23%, respectively.
Bulk density To make a general case, the dominance of
The specific gravity of loess particles varies silt-sized primary minerals and the generally low
according to the mineralogy of the source materi- percentage of clay minerals in many loesses,
als, much lying in the range 2.5–2.8. Bulk density yields LL values of between 25 and 35% and low
values increase with age, and so with several other plasticity index (PI) values. PI values for loess
factors, most notably degree of consolidation. sampled in many parts of the world lie in the
Values for the loess of the western Loess Plateau range 10–25, encompassing the traditional
of China are 1.38 Mg/m3 (Upper Pleistocene), descriptive classification, based on particle size,
1.56 (Middle and the upper part of the Lower of sandy loess, ‘typical’ loess and clayey loess
Pleistocene) and 1.71 (Lower Pleistocene: (Figure 25.17; Gibbs and Holland, 1960).
Derbyshire et al., 2000). Porosity values show the
same trend (0.53, 0.47 and 0.43, respectively). Permeability and infiltration
Palaeosols have the highest bulk densities Given the tendency of loess to develop vertical
(c. 1.70 0.05). Similar trends occur elsewhere: joint systems, it is important to distinguish
GFE-25.qxd 6/27/05 2:31 PM Page 707
LOESS 707
Figure 25.17a Plasticity chart characterising the range of variation in the major loess deposits in the Missouri
River basin, USA. The largest cluster is found between the zone demarcated by PI values of 5 to 12%, and LL val-
ues between 25 and 35% (‘typical’ or silty loess). The two other groups, sandy loess and clayey loess are minority
populations. Based on eleven sites, redrawn after Gibbs and Holland (1960).
Figure 25.17b Plasticity chart for samples taken from a single site (Dawan, 20 km WSW of Lanzhou in the semi-
arid western part of the Chinese Loess Plateau), for comparison with the ‘silty loess’ category in Figure 25.17a.
GFE-25.qxd 6/27/05 2:31 PM Page 708
between intact permeability and bulk permeabil- rise to criteria taking account of several variables,
ity that includes bypass drainage by way of joints in the following form:
and fissures. Coefficients of permeability in
intact loess are 1.6–3 10
8 m/s and 8 10
9 c (hp
hp)/ho (ep
ep)/(1 eo)
m/s for vertical and horizontal cases, respectively.
where c is the coefficient of collapsibility; hp and
Rates of permeability involving bypass drainage,
ep are the sample height and voids ratio, respec-
however, may be orders of magnitude higher.
tively, under a loading pressure p; hp and ep are
Rates generally decrease with depth below the
the sample height and voids ratio under pressure
surface in loess with few or closed joint systems,
p in a saturated condition; ho and eo are the origi-
any modification of the latter increasing the vari-
nal sample height and voids ratio (CSCCC,
ance in values and sometimes locally reversing
the gradient of rate with depth. Well-developed
palaeosols, being more clay rich, are also a
complicating factor. In addition, various types of
structural anisotropy may lead to the development
of preferred routes for sub-surface water move-
ment. At the field scale, for example, the presence
within thick loess of more compact or more plas-
tic horizontal layers, including palaeosols, may
influence the geometry of water-enlarged joint
sets (piping: see ‘Loess hydrology and pseudo-
karst’ below), a process that is thought to
contribute to destabilisation of some loess slopes.
The tendency for piping to develop appears to be
true of the microfabric as well as the field scale in
Chinese loess, as shown by artificial eluviation Figure 25.18a Undisturbed Upper Pleistocene loess,
showing typical intact fabric (right) adjacent to a large
experiments (Figures 25.18a and 25.18b: Muxart natural drainage channel (left).
et al., 1995).
Collapsibility
When loess under a load reaches critical moisture
contents close to saturation, it displays collapse
behaviour sometimes referred to as hydrocom-
paction or hydroconsolidation. Collapsibility at
the field scale is a visible characteristic of loess,
and one with important implications for engineer-
ing works. Collapsibility has been quantified
using a number of criteria. Denisov (quoted in
Feda, 1966) used void ratio above a critical value
(ec) and at the liquid limit (el), while Feda (1966)
also used critical and liquid limit void ratios but
added void ratio at the plastic limit (ep), in the
following relationship:
Figure 25.18b Locally compacted inter-silt fabric and
ec 0.85 el 0.15 ep large, ovoid drainage pipes (both vertical and transverse)
developed during 240 hours of controlled laboratory
Widespread engineering problems involving col- leaching of Lanzhou loess. For details, see Muxart et al.
lapse in thick loess terrain in China have given (1995).
GFE-25.qxd 6/27/05 2:31 PM Page 709
LOESS 709
(Collapsibility Standard for Chinese Civil Constr- Shear strength

uction) 1978, cited in Derbyshire et al., 1995b). In general, shear strength rises with a decline
It has been pointed out by Lutenegger (1981) in moisture content and a rise in bulk density.
that loess, consisting predominantly of silt, Moisture content exerts a fundamental influence
would not be regarded as a sensitive soil in the upon the shear strength of loess. In laboratory tests
classic Terzaghi sense, because the ratio of undis- on Chinese loess, the internal friction angle rose as
turbed to remoulded strength at a constant mois- moisture content increased from zero to 5%, after
ture content is approximately 3, placing loess in which internal friction angle declined with rise
the medium sensitivity class only. Using the in the moisture content (Dijkstra et al., 1994). The
inference of Feda (1966) that sensitivity can be range of internal friction angles in the essentially
expressed by the ratio between undisturbed and uncemented loess of North China is 25–40. The
saturated strength (in unconfined compression), primary cohesion is dominated by particle size,
many loess soils would be classified as ‘quick’. mineral composition and material density, and
Classic quick-clays, however, are saturated, rises with percentage clay content. Consolidation
whereas some loess displays collapse behaviour cohesion occurs when chemical cementation is
when wetted to some critical but unsaturated present, common cementing agents in loess being
level and beneath a load, as shown in Feda’s calcium carbonate, magnesium sulphate, gypsum
oedometer tests. Handy (1973) noted that it is the and sodium chloride. These minerals are taken
loess soils deficient in clay minerals that tend to up and re-precipitated by soil moisture solutions
collapse in Iowa (USA). He went on to argue (e.g. Muxart et al., 1995), so that consolidation
that, because of the increase in liquid limit and cohesion is frequently more marked in the older
decline in saturation moisture content with dis- loess units.
tance from the loess source regions, the extent of Rogers et al. (1994b) discussed the relation-
collapsible loess in Iowa may be defined at a ship between cementation and clay mineral bond-
point when the saturation moisture content ing in loess. The presence of significant amounts
equals or exceeds the liquid limit. of clay mineral allows deformation of loess fabric
The collapsibility of loess varies geographi- without any accompanying decrease in strength
cally in response to climate and related factors, because the clay bonds are long range in type.
and also with depth below the surface (i.e. age). However, the percentage of clay mineral is low in
Lin and Liang (1982) and Qian et al. (1985) show most loess, so that it is dominated by short-range
how collapsibility in the loess of northern China bonds, has low plastic limits and is of low plastic-
varies with climatic region and mean thickness of ity (Ip range 6–8%). At the same time, varying
collapsible loess (Figure 25.19). Tests on the loess degrees of cementation bonding are commonly
at Lanzhou show that, below a void ratio of c. 0.95, present so that loess may achieve greater yield
coefficient of collapsibility rapidly approximates strengths even though these may dissipate rapidly
zero. The commonly observed spread of data with relatively modest amounts of deformation.
above the voids ratio threshold indicates the role When loess fails, it approximates a liquefaction
played by other factors such as cementation, state either because of saturation or, if dry, as a
shape and distribution of pores, particle size result of shear failure in cementation bonds
distribution, and zones of collapse involving (Rogers et al., 1994b; Mellors, 1995), a condition
inter-granular compaction, introducing complex that persists until pore pressures dissipate and
post-collapse permeability patterns (Derbyshire effective stresses recover. The resultant material,
et al., 2000). In more humid climates, collapsibil- sometimes referred to as ‘secondary loess’, has
ity may show greater variability depending on closer particle packing and diminished cemented
factors such as clay content, natural moisture strength. In freshly deposited loess, with its char-
content, modification by reworking in very differ- acteristically very high void ratios, relatively little
ent climates of the past, and geomorphological moisture is required to mobilise the cementing
location and history. agents that are re-deposited as bridges between
GFE-25.qxd 6/27/05 2:31 PM Page 710
the grains of the skeletal silt skeleton; this some data from North China (Table 25.1). As is
is an essentially metastable state (Derbyshire common in relatively uniform silt-rich deposits,
and Mellors, 1988). As a dipolar liquid, water is loess poses problems as a fill material because it
an efficient agent in penetrating the bonds is difficult to compact. Means of compacting or
and weakening the inter-silt cement bridges consolidating loess in order to improve its bearing
(Mellors, 1995). capacity include dynamic, chemical and water
Bearing capacity, an important parameter in compaction. Dynamic measures (including drop-
foundation design, increases consistently with ping heavy weights, piling and, rarely, blasting)
depth, the highest values being found in clay- are mainly used in foundation works for individual
enriched zones (palaeosols), as exemplified by buildings and structures. Chemical compaction is
110˚ 114˚ 118˚
^^^^ Mountain range Average thickness ^

Average δc ^ ^ ^ ^ ^h^an^ Zhangjiakou Yang Shan
^^ ^Yi^
Region boundary
km
ng
s
> 12 ^ ^^ ^ ^
0 200 Baotou 0.042 >
>
>> Beijing
>>>
<<< >>
<5
<<
> >> >

>> > > >
>
Lo 5-10 Tianjing
>>>
VII
>>> >n Shan
0.044 Bohai
n gsh << VII
>>
3
ou
Sh
3 <5
> 4
Sea
0.066 VI
>>>>>
an 10
Hela
0.022 Yinchuan 5-10 0.03

0.041
<< Qil<<< Wuwei
an
Shijazhuang
>>> >>> >> > >
< ian < 4
>>>>>lia>ng>Sh
6
10-15 << Sh
Shan
^^
Taihang Shan
Taiyuan
5-10 Baiyu 5-10 0.026
<<an
^ ^5-10
<<
an
<<
0.06
< ^^ ^ ^ ^ IV Sh
<
>
5-10 Jinan Lu
^^
n<<
Xining
Lu
0.057
Sha
^h^an^ ^^
15-20 10 5-10
<5
<<<
0.063 0.06 Yanan 0.035

pan
5-10 ^^
Lanzhou 5-10
II Shan i S 5-10
Yongjing o Ta
<<< <
Liu
15 5-10 > ngtia >> 0.016 N

>
R.
Zh >>
I 0.073 15 o
w
Bai Shan
^^
llo
0.064
^^
<<
^ ^ ^ ^^5-10 Zhengzhou
Ye
Ye Tianshui 10
Maxi 5-10
llo 7 V
<
w
R.
^^^ ^S^^^ III Xian 0.06 ^
^^^^Qingling ^ 0.05
^ Shan 0.03
Xuzhou
han ^ Hua
^ ^ ^ ^ ^^
^
102˚ 106˚ 110˚ 114˚ 118˚
Figure 25.19 Geographical variation in selected geotechnical properties of loess (mean thickness and coefficient
of collapsibility (Sc)) across the Chinese Loess Plateau, redrawn after Qian et al. (1985).
Region I: loess thickness over 300 m. Collapsible loess is over 2 m thick. Low clay contents ( 17%), low natural
moisture contents (4–10%). Subsidence 1 m common in farmland and construction sites. The sink-
holes and pipes in the ‘loess karst’ are well developed. Damage due to loess collapse is the most serious
in all China.
Region II: loess up to 150 m thick. Collapsing loess is 4–15 m thick, with mostly strong to medium collapsibility.
Region III: collapsing loess thickness 4–12 m. Clay and moisture contents higher than in regions I and II. Mostly
medium collapsibility.
Region IV: collapsing loess is 2–10 m thick, and mostly medium in scale.
Region V: scattered collapsibility ranges. Collapsible loess does not exceed 8 m (because of humid climate and
great distance from silt source regions). Clay contents very high, mostly over 50%, even in young
loess. Collapsibility in ‘weak’ range, and damage attributable to collapse is rare.
Region VI: only weakly collapsible loess, 2–6 m thick, on sites such as hill slopes and valley floors.
Region VII: the north fringe of the Loess Plateau. Transitional zone between stony desert (‘gobi’) and plateau loess.
Coarse-grained loess, with weak collapsibility, distributed sporadically on floodplains and piedmonts.
Thickness 4 m.
Some collapsible loess also occurs in Xinjiang Province (far NW) but this is less than 2 m thick and poses no threat
to structures.
GFE-25.qxd 6/27/05 2:31 PM Page 711
LOESS 711
Table 25.1 Bearing capacity of loess and palaeosols at

Gaolanshan, Lanzhou, western Chinese Loess Plateau,
after Derbyshire et al. (2000).
Lithology Depth Bearing capacity (kPa)

(m)
Minimum Maximum Average
Upper 0–4 115 127 120

Malan 4–8.6 118 147 132
Middle 8.6–13 127 155 138
Malan 13–17.8 135 173 160
Lower 17.8–22 155 185 165
Malan 22–27 160 196 170
Palaeosol 27–32 180 224 200
S1
Upper 32 166 215 192
Lishi
relatively expensive and, not least because of the

rather poor state of the art, is much less commonly
employed. Hydrocompaction is economical and
has been widely used in China in schemes ranging
from consolidation of farm land and village sites
to urban developments and individual structures.
Inadvertent hydrocompaction arising from exces-
sive addition of irrigation waters to extensive
farmlands on loess terraces in China has led to Figure 25.20 Desiccation crack system on a thin
some disastrous changes (Derbyshire et al., 2000). (2–4 cm) rain-beaten crust on the loess surface of an
extensive spreading slide in northern China, following
prolonged summer rainfall. Spreading has broken up
25.4 Key engineering the surface loess under tensional blocks, on the side of
which can be seen parallel, low-dipping shear failures
geomorphological issues (Gansu Province, China).
Loess hydrology and pseudo-karst

Meteoric water falling on a sparsely vegetated These include parts of the land surface, notably
loess soil shows variable patterns and degrees of when rainfall rates are very high, and sub-surface
penetration. This phenomenon arises as a result of zones of low permeability where the passage of
a combination of high rates but low totals of rain- joint-guided waters is impeded by bedrock or
fall in many of the world’s drylands, the diversion clay-rich buried palaeosols.
of surface flow by larger soil structures, and the Exposed loessic soils develop surface crusting
relatively fine texture of loess (despite its high void in response to rainbeat, raindrop splash and rill
ratios and only modest, sporadic cementation wash, a process often followed by desiccation
cohesion). The annual infiltration cycle in some cracking (Figure 25.20). Detailed work on the
of the drier parts of the Chinese loess region for near-surface hydrology of the thick loess in the
example, affects only a thin surface zone that may western part of the Chinese Loess Plateau
not exceed 2 m (Derbyshire et al., 1995c). Thick by Muxart et al. (1994) showed that compact
loess in the world’s drylands reaches saturation rainfall-induced crusts greatly decrease the per-
only sporadically and in certain specific conditions. meability of surface soils. Such crusts suppress
GFE-25.qxd 6/27/05 2:31 PM Page 712
infiltration into the loess pores, leaving near-

surface horizons dry even in periods following
prolonged or intense rainfall events. At the same
time, they enhance rates of overland flow (as both
rill and sheet flow). Even in the intense summer
monsoon showers that periodically affect parts of
northern China, however, a single sheet of water
rarely reaches the slope foot because of diversion
and concentration in depressions and shallow
gullies and, often after less than a linear transport
distance of 1–2 m, interception by the networks of
natural joints in the loess.
Distinctive sub-vertical jointing was one of
the earliest physical properties noted in loess
(Richtofen, 1882). Joints originate from several
causes including wetting and drying, erosive
undercutting (stream sapping), creep-induced
tension and earthquake shock. Where joints inter-
sect, the process of localised liquefaction along
the joint surfaces, suffosion (piping) and collapse
lead to joint enlargement and the development of
sub-vertical sinkholes. Features of all sizes may
be seen along some loess slopes, ranging from
holes a few centimetres across to broader features
that are loci for drainage into systems of pipes up
to 2–3 m in diameter that often run parallel to the
overlying surface slopes (Figure 25.21a). At
Figure 25.21a ‘Loess karst’ in the Upper Pleistocene
depth, an underground conduit system, variable in loess formation on the dry, western part of the Loess
size and density, may develop from fissures and Plateau of China: part of a system of subterranean
lithological discontinuities within the loess. The pipes, 2 to 3 m in diameter, that divert much of the
waters move through the piping system, intercept- surface drainage.
ing any existing shear zones until infiltration is
impeded by layers of lower permeability both
within the loess (e.g. palaeosols) and at the loess-
bedrock contact (Figure 25.22). The process may
advance to the point where complex integrated
subterranean sinks, pipes and caverns develop,
appropriately known in China as ‘loess karst’ by
analogy with the soluble morphology seen in
limestone terrains.
Variations in the efficiency of overland flow
exert strong control on the rate at which water
enters the sub-surface drainage network. The pro-
portion of the input rainfall that successfully
enters the pseudo-karst pipes is also affected
Figure 25.21b Collapse sinkhole in loess near
because the evaporation rate decreases inversely Tianshui, eastern Gansu province. Note the small land-
with runoff rate (Billard and Muxart, 2000). slide (tan-ta) on the right, induced by the sinkhole
During heavy downpours, therefore, large volumes collapse.
GFE-25.qxd 6/27/05 2:31 PM Page 713
LOESS 713
overland flow
interflow
pipeflow and
loess basal flow
palaeosol
potential basal
saturation zone
low
per loess
mea
bilit
y be
dro
ck loess
Figure 25.22 Schematic diagram of water routing in relatively thin (c. 10 m) loess draping a palaeoslope of low
permeability bedrock (after Billard et al., 1993).
of water reach zones of much lower permeability develop more rapidly in sodium-rich loess than in
within (palaeosols) or at the bedrock–loess con- loess rich in calcium and magnesium. Large
tact. This combination of overland flow and drainage tubes, local collapse zones and planar
bypass drainage made possible by the piping sys- fissures arise within intact, visually homogeneous
tems, is a more important cause of slope failure loess following leaching (Figures 25.18a and
than infiltration into the intact loess because flow 25.18b: Muxart et al., 1995). During periods of
rates through the joint systems are usually much rainfall, the plastic and liquid limits may be
higher than those through intact loess. exceeded in loess in which the clays are rich in Na
In fact, however, the physical changes that occur and the CO32
and Atterberg Limits decline when
during the drainage of intact loess have a number soil water becomes strongly charged with elec-
of features in common with those associated with trolytes. Adjacent to springs and seepages in slope-
joint-controlled drainage at the field scale. In par- foot zones, loess may be broken up by precipitation
ticular, the development of pipe and fissure of salts from supersaturated solutions, owing to
systems occurs in both, as has been shown in arti- evaporation in summer and desiccation associated
ficial leaching experiments in the laboratory with freezing in winter.
(Muxart et al., 1995). The presence of readily sol- The tendency for piping to develop is generally
uble salts (including carbonates, sulphates and enhanced by weakening of the bonds between the
chlorides of sodium and, to a lesser extent, sul- quartz particles of the fabric skeleton as a result of
phates, carbonates and chlorides of calcium and dissolution of inter-particle soluble salts and
magnesium), in combination with the typical open- dispersion of inter-particle ‘bridges’ formed by
work fabric, leads to preferential mechanical eluvi- clays and the finest silts (Derbyshire and Mellors,
ation in parts of the intact loess mass. The 1986). This occurs at two scales: along the primary
structural stability of micro-aggregates consisting macroscopic joint surfaces and as a general incre-
of clays and very fine silts is weaker in the pres- ase in porosity within intact loess. The solution of
ence of sodium, a common constituent in loess soluble salts frequently results in changes towards
especially in drylands; piping and suffosion more basic pH values, associated dispersion of
GFE-25.qxd 6/27/05 2:31 PM Page 714
aggregates and mechanical eluviation of solids. loess mass (Figure 25.21b). In such circum-
These processes, added to liquefaction of thin stances loess hydrology and erosion may involve
layers of loess along joint walls, result in joint more tunneling, collapse and mass sliding than
enlargement. In this way, an essentially homoge- rilling and gullying. Erosion can be particularly
neous material following deposition is trans- severe around some types of artificial structures,
formed into a quasi-coherent sediment divided notably bridge foundations.
into blocks by a dense network of pipes, cavities
and channels due to the process of infiltration of
water, dissolution of soluble material and piping, Mass movements in loess
to create the characteristic macroporosity. Slope failure is manifested by a range of land-
The presence of the macroporosity, known as scape features of considerable diversity and
pseudo-karst, is important for two reasons. First, involving a number of processes. These processes
notwithstanding the results of the mechanical– include fall, toppling, flow and spreading, in addi-
eluviation experiments summarised above, effective tion to sliding. While these and other processes
permeability values determined in field situations are commonly grouped together in referring to
are very much higher than those derived from failed masses on slopes as ‘landslides’, the term
measurements using blocks of intact loess in the mass movement is more comprehensive in that it
laboratory. Second, the joint and pipe systems con- emphasises mass transport (Brunsden, 1984). The
duct water to the loess–bedrock interface zone, general classification of mass movements is not
which may then reach saturation even though the considered in detail here, as important contribu-
loess above remains relatively dry. As this process tions to this topic are readily available in the work
is time dependent, the proportion of a given rainfall of Varnes (1978), Brunsden and Prior (1984),
amount infiltrating through the pipe system Hutchinson (1988), Cruden (1991) and Dikau et al.
increases with time, at least until localised collapse (1996), for example.
diverts the conduits and the process recommences Landslides in loess are very diverse as a result
in whole or in part. The metastability of some parts of the broad range of field conditions in which
of the loess always tends to increase either pro- landsliding occurs. The mode of failure of loess,
gressively or abruptly. The waters may also partly as an undisturbed, partially cemented mass of silt,
infiltrate into bedrock joints and bedding planes, a is greatly influenced by the presence of joint sys-
process that may be enhanced in its efficiency by tems; in this, it has some similarities with failure
the concentration of waters within the pseuodo- on rock slopes. At the other extreme, as a cohe-
karst conduits. The loess–bedrock contact remains sionless particulate body, loess may fail as a
a particularly critical zone where the water may flowslide given certain moisture levels at depth,
concentrate during long or heavy rainy periods or and as a mudflow when saturated. An intermedi-
following several such episodes. ate type of failure in loess involves extensional or
The role of bypass drainage varies according translated slides, or lateral spreads (Figure 25.23).
to local circumstances. Many systems are highly Creep-like behaviour also occurs.
integrated and drain freely for long periods giving Despite the extensive literature on landslides
rise to slope-foot springs. Sometimes roof falls and landslide classification, mass movements
occur, the result of which is a coincident distri- in loess have received relatively little attention,
bution of sinkholes and shallow but dry gully not appearing at all in some classifications. The
networks on some loess slopes. Some sub-surface large and catastrophic loess landslide in Gansu
distributary pipes develop and lead down into province, China, in 1920 (Zhang, 1995) is the
reduced-diameter pipes or even dead ends. In only loess mass failure mentioned (and referred
such situations, water pressure potentials may rise to as ‘typical’) in the Varnes (1958) classifica-
significantly, leading to local saturation and creat- tion, for example. Some more recent classifica-
ing temporary perched water tables; liquefaction tion schemes (e.g. Sassa, 1989) underestimate
may then be followed by collapse of the superjacent the special geotechnical characteristics of loess
GFE-25.qxd 6/27/05 2:31 PM Page 715
LOESS 715
Figure 25.23 Detail of widespread tension cracks in laterally spreading, slow, rainfall-induced landslide at Tala,
c . 40 km SW of Lanzhou city, western Loess Plateau of China, in August 1988. For context, see Figure 25.25b.
and the distinctive failure behaviour of massive and land use practices vary to such a degree that
silts (Derbyshire et al., 1990). Suggestions such no single basis for landslide classification has yet
as that the 1920 Gansu slide might have been pre- emerged from this huge region. However, a num-
vented if the loess had been artificially hydro- ber of commonly agreed criteria have been pro-
consolidated (Veder, 1981) arise from a failure to posed by practising engineers and geologists as a
appreciate the implications for the study of mass basis for damage limitation strategies in China,
movement types in loess of structure and fabric, namely:
moisture content and the geomorphology, partic-
1. the material contents of the sliding mass
ularly the slope steepness. Attention has been
2. the location of the failure plane in relation to the
drawn to the similarities of behaviour between
lithology.
loess slopes and those in classic ‘quick-clays’
(Smalley and Derbyshire, 1991; Derbyshire Additional practical considerations are the
et al., 1995b). Both loess and quick-clays are geometry of the slide and the type of movement.
sensitive materials and, although most of the
latter have higher sensitivities (16 as against c.10 Landslide types based on slide mass
for loess), both have openwork fabrics, the obvi- composition and location of
ous difference between the two being the order of failure plane
magnitude difference in scale (loess 50 m In a recent review of literature and practice in
versus 2–5 m for quickclays). The behaviour China, Meng and Derbyshire (1998) described
of both materials is dependent upon the presence three major types of landslide defined in terms
of short-range bonds. of the composition of the slide mass and the
situation of the failure plane, named ‘bedrock
Mass movement types contact landslides’, ‘palaeosol contact landslides’
Although the greatest potential database for loess and ‘mixed landslides’, to which they added a
landslides probably lies in the thick loess terrain fourth (the special case of ‘terrace landslides’)
of northern China and central Asia, local conditions (Figure 25.24).
GFE-25.qxd 6/27/05 2:31 PM Page 716
Bedrock Palaeosol Mixed

Contact Contact Landslide
Landslide Landslide
Loess
Argillic bedrock 1 Palaeosol 2 Argillic bedrock 3
Asequent Terrace Tan-ta

Landslide Landslide
4 6
Argillic bedrock 5
Figure 25.24 Main loess landslide types on the western Chinese Loess Plateau (after Meng and Derbyshire, 1998).
Bedrock contact landslides, as the name planes or joints in the bedrock. Frequent situa-
implies, involve the sliding of a mass of loess along tions are those in which argillites or mudstones
a bedrock surface. This is the commonest type underlie the loess drape, or where shallow faults
throughout the Chinese Loess Plateau. The under- or clay-lined shear zones occur within the
lying bedrock is generally of low to very low per- bedrock. A number of cases in which intermittent
meability (commonly argillites or mudstones). The sliding of mixed material occurs on slopes of
resulting impedance of groundwater movement loess overlying Cretaceous sandstone with inter-
results in a zone of saturation at the loess–bedrock calated thin mudstones are documented in the
interface. This situation is frequently exacerbated Chinese literature.
by the presence of a weathered zone on the bedrock Terrace landslides are a common and impor-
surface that contains smectitic clays (Derbyshire tant type in China. Many river terraces consist of
and Meng, 1995). Loess landslides of this type a planed rock surface with fluvial or colluvial
include some of the largest loess mass failures sediments above, upon which rests loess (both
known (Figures 25.25a and 25.25b). primary and reworked) up to thicknesses of 100 m
Palaeosol contact landslides involve the slid- or more. Given the much lower permeability of
ing of a mass of loess on or within the numerous most of the common bedrock types, the fluvial or
palaeosols found within the Chinese loess. Water colluvial gravels serve as an aquifer between rock
circulation is impeded by such relatively compact, and loess. Surface efflorescence of salts is com-
clay-enriched horizons, causing localised rises in mon along seepage lines, and eluviation of salts
soil moisture and consequent decline in soil and finer particles in moist conditions causes
strength. The probability of failure is enhanced softening of the lower parts of the loess cover,
when the basal parts of the loess slopes are under- which respond by deforming under the overbur-
cut by river incision or artificial excavation. den stresses. If the strength of the bedrock falls to
This slide type is not well documented. a critical value, cracks develop throughout the
Mixed landslides consist of mixtures of loess loess, leading to failure. Many of the thicker loess
and bedrock fragments sliding along bedding landslides in northern China are found on terrace
GFE-25.qxd 6/27/05 2:31 PM Page 717
LOESS 717
Figure 25.25a Landslide consisting of mass movement of loess on the sloping bedrock beneath a liang, in Ningxia
province, North China. The slide mass reached the valley floor and has since largely been removed by fluvial erosion.
From an original photograph by the late Wang Jingtai.
Figure 25.25b Part of an extensive (c. 4 km2) spreading slide at the loess–bedrock contact at Tala, August 1988
(cf. Figure 25.23). This was a slow landslide induced by prolonged monsoon rain; drinking water wells and roads
were all destroyed and there were some animal deaths, but no people died. The sloping contact between loess (pale)
and underlying argillite bedrock can be clearly seen on the ridge face in the middle distance.
GFE-25.qxd 6/27/05 2:31 PM Page 718
edges, some of the failures being rapid and occa-

sionally catastrophic. Many cities, towns and vil-
lages have been built on such loess-draped river
terraces, urban concentrations served by railways,
roads, and industrial plants. Thus, even quite
small examples of this type of landslide have
proved serious, especially as, in the larger slide
masses, high pore pressures may build up during
sliding, so greatly extending the sliding distance.
Slides entirely within loess: large landslides
composed exclusively of loess are relatively
uncommon in the Chinese Loess Plateau, presum-
ably because the only features within loess having
any potential to initiate slip surfaces (gravity-
induced joints and shear planes) are usually too
dry in this region, with typical ambient moisture
contents 12% (Meng and Zhang, 1989). In
such conditions, the cementation bonds remain
unbroken though brittle, providing loess with suf-
ficient shear strength to resist slope failure
(Dijkstra et al., 1994). Nevertheless, landslides
entirely within loess have been quite widely
recorded, although special conditions (including
human action in undercutting slopes and raising
the local water table) are thought to have been
present in most cases. Throughout the loess
region of North China, abundant small-scale Figure 25.26 Small slide (tan-ta) entirely in loess:
displacements (⬃ 10 m diameter) with planar slip Gaolanshan, Lanzhou, China.
surfaces from one to several metres deep show
high sliding velocities and rapid disintegration of
the slide mass (Derbyshire et al., 1993). These exacerbating the ‘closed pipe’ problem men-
small features are known as tan-ta (both singular tioned above and leading to retrogressive failure
and plural: Figure 25.26). The consistency of the (Figure 25.27).
gradients on these features is in accordance with It is important to recognise that the hydrology
long-term but episodic movement triggered by of loess in general and conditions at loess–water
precipitation and valley incision. Several factors interfaces may vary with the age of the loess unit
appear to be involved, including infiltration of under consideration because of differences in
snowmelt and rainwater, slope undercutting and bulk properties, notably particle size and bulk
natural earth tremors. Activities such as irrigation density. Loess-karst is more common in young
and slope undercutting during road and railway loess, which generally has a coarser mean particle
construction also trigger tan-ta. These small fea- size and lower bulk densities. Landslides in
tures are a recurrent nuisance, and occasionally younger loess are influenced by the distinctive
constitute a real hazard. Collapse of components physical properties of the loess but also by other
of the pseudo-karst on relatively thin loess slope factors such as its occurrence as a relatively thin
mantles may increase the volume of loess likely drape that cuts unconformably across older, sub-
to become saturated in subsequent rainfall horizontal loess units on long, steep slopes. In
events (Figure 25.27). Shallow sliding may such cases, flowslide type failures associated with
disrupt antecedent sinkhole-and-pipe systems, saturated conditions on the basal slip surface
GFE-25.qxd 6/27/05 2:31 PM Page 719
LOESS 719
Figure 25.27 Pseudo-karst collapse in loess. The car-

toon on the left shows a cross-section of loess in which
the potentially saturated loess mass increases following
collapse. The sketches on the right show the develop-
ment of loess pseudo-karst on a slope, and associated
sliding. This may be a retrogressive process, which may
begin again above the point marked ‘R’. Broken arrows
indicate water flow (overland flow and pipe flow).
Figure 25.28 Scanning electron micrograph (vertical
face) showing discrete, low-dipping shear planes and
large secondary pores arising from high porewater
following prolonged rainfall may be a critical pressures associated with rapid undrained shearing dur-
triggering factor. Failure planes develop on or ing sliding in the lowermost few centimetres of a thin
close to the loess–bedrock interface where condi- slide mass of loess overlying argillite bedrock, near
tions favour porewater fluctuations and efficient Lanzhou, northern China.
throughflow, as shown by features including dewa-
tering structures and microshearing zones in the Varnes (1958), together with lateral mass flowage
fabric of loess within 0.5 m of slide planes on low comparable to mudflows.
permeability bedrock (Figure 25.28). Mass movements in loess are often complex.
Block-type movement in the upper area of a failed
Landslide types based on their geometry mass may gradually change into a flow type of
and movement mechanism movement in the lower parts. In China, it is often
Meng et al. (2000) have described the types of the great extent of these flow lobes that causes the
landslide types in the thick loess terrain of north- greatest damage to the human environment.
ern China, classified on the basis of their geome- The best, and most fully documented case is
try and mechanism of movement. Included are the slide that occurred at 17:46 hours on 7 May
flows and complex mass movements, large rota- 1983 at Saleshan, a rural area about 80 km south
tional mass movements, planar slides and debris of Lanzhou (Cao, 1986; Zhang, 1989; Zhu, 1989;
and mudflows. Mahaney and Hancock, 1990; Cao et al., 1997).
Flows and complex mass movements. When a Here, about 150 m of loess overlies Neogene
thin layer at the base of a thick loess mantle on a argillites, thin conglomerates and clayey lime-
sloping palaeo-relief approaches saturation, it is stones. Sustained high groundwater levels in the
liable to suffer instantaneous collapse, leading contact zone between loess and argillites caused
to large-scale failure of the flowslide or ‘spread’ the argillites to soften. A mass of loess, with an esti-
type (Figure 25.25b) in the classification of mated volume of 31 million m3, and measuring
GFE-25.qxd 6/27/05 2:31 PM Page 720
1000 m in width and 170 m high, travelled a dis- and debris flows have occurred on at least 50
tance of 1600 m in less than one minute, killing occasions in Gansu province, western China,
273 people (Figure 25.29a). killing more than 2100 people. More than 50 of
Large rotational mass movements. Large rota- Gansu’s 82 counties and at least 7000 valleys have
tional landslides have frequently been triggered been affected.
by major earthquakes, as a comparison of their
distribution with the remarkable Chinese written Soil erosion and sediment yield
records describing their impact on daily life The distinctive sedimentary and geotechnical
clearly shows. The Tawa landslide (Figure properties of loess explain the very high rates of
25.29b), situated about 15 km south-west of erosion to be seen in many loess landscapes.
Lanzhou city, was triggered by an earthquake in Measurements of sediment yield in the young-
1125 AD. The slide shows a maximum vertical est (Upper Pleistocene) and most friable loess
displacement along the slip surface of 150 m, the in China by Gong and Jiang (1979) gave values
shape and gradient of its landslide scar suggesting of 14 200–34500 t/km2. They quote yields of
a single major circular failure plane, and so quite 0.7–1.0 t/m3 (48.7–61.5% by weight) during the
unlike the mode of failure of the Saleshan slide. heaviest summer rainstorms. These storm rates
Little disintegration of the slide mass is evident, are 1.10–1.76 times greater than the measured
although a stream was displaced and continues rates of loss from the interfluves, indicating that
to undercut part of the main slide mass. Some the loess slope profiles maintained by seasonal
moderate-sized mass movements have occurred undercutting in the south-central Loess Plateau
towards the toe of the slide because of river under- of China represent a rate of abstraction some 1.10
cutting, and others have been induced by highway to 1.76 times the loss rate from the interfluves.
construction. Thus, it appears that such slope profiles are
Planar slides involve the movement of an maintained largely by periodic basal corrasion. It
essentially rigid body over a shallow slip surface. has been claimed by Dai (1987) that, in present
Sliding of loess slabs up to about 5 m in thickness conditions, the volume of loess lost by erosion
appears to take place on slip surfaces that roughly from the Loess Plateau in one year is equivalent
parallel the slope surface. Planar slides are defined to 100–300 years of accumulation. This is graph-
by the shear strength conditions on the failure sur- ically illustrated by comparison of some present-
face and the cohesion of the material. Such slides day features with the remarkable descriptions to
are usually restricted to the Late Pleistocene be found in the Chinese written records. Near
(Malan) loess, but some have also been recorded Xifeng, in eastern Gansu province, the part of the
in slope deposits (reworked loess). Some degree of Loess Plateau surface known as Dungzhi Yuan
strain hardening may be involved so that, as a con- was 32 km wide in the Tang Dynasty (618–907
sequence, such planar slides are not usually asso- AD ). Thus yuan has been reduced in area from
ciated with the same acceleration and release of 1344 to 756 km2 in little more than 1000 years,
energy typical of slides that develop in the older, leaving parts of it now only 1 km wide (Figure
more brittle loess units. Usually, therefore, they 25.30).
constitute only a minor hazard. When vegetation cover is sparse or absent,
Debris and mudflows. Debris flows in loess either because of a dry climate or intensive land
terrain are usually found in locations where the use, as in many parts of China and central Asia,
loess is a thin drape over bedrock. The term loess is highly susceptible to erosion by rain-
‘debris flows’ is probably a misnomer, however, beat, sheet erosion, gullying, mass movement
because such silt-dominated flows are best and subsidence. These processes deliver pulses
described as mudflows. They commonly occur in of sediment into rivers, material that may remain
association with slope failures of other types, suspended for long distances until re-deposited
being found with debris flows and debris floods in on floodplains, deltas and in the coastal seas.
some areas. Since 1949, large-scale loessic mud The coincident distribution of major loess
GFE-25.qxd 6/27/05 2:31 PM Page 721
LOESS 721
Figure 25.29a The Saleshan landslide, about 80 km south of Lanzhou city, western Loess Plateau, China.
Photograph taken 5 years after the failure (7 May 1983). The massive area of run-out, clearly seen by its dark shade
below the slide scar, covered three villages in less than one minute, blocked the Ba Xie River and ruined two reser-
voirs used for drinking water in this semi-arid area.
Figure 25.29b The Tawa landslide (right half of photograph), its crest c. 380 m above the floor of the Xuangjia val-
ley, SW of Lanzhou in the western Loess Plateau. This is an example of a large rotational slide triggered by an earth-
quake. Written records show that this occurred in the AD 1125 Lanzhou earthquake (estimated magnitude 7).
Maximum vertical displacement along the slip surface was c. 150 m. In contrast to the Saleshan slide (above), all
evidence here points to a single, major circular failure plane, with no evidence of widespread disintegration of the
slide mass. However, the slide displaced the river, and subsequent undercutting has occurred, triggering the second-
ary slides, some of which still threaten the major roadway (bottom of picture). Photograph by Armelle Billard.
GFE-25.qxd 6/27/05 2:31 PM Page 722
Qingyang
Xifeng
Jing H
e
Niangxian
Remnants of former
plateau surface
Plateau surface in the

Tang Dynasty (618–907AD)
Gulleys and valleys
Figure 25.30 Documented progressive elimination of a loess yuan near Xifeng, eastern Gansu province (after
Derbyshire, 2001).
deposits and great rivers, and the possible role has made High Asia a region of considerable
played by the latter in the accumulation of the geodynamic energy, combining tectonic, glacial,
former, has been persuasively argued by Smalley periglacial, river and other means of particle
(1972; 1995). Geologically recent crustal overlap generation and transportation (Smalley and
GFE-25.qxd 6/27/05 2:31 PM Page 723
LOESS 723
Derbyshire, 1990). Some of the greatest rivers sand, loess is light brown in field exposures, only
on Earth drain the loess lands of Asia, including weakly plastic, with a mean particle size in the
the Yellow, the Yangtze, the Mekong, the coarse silt range. Quartz is the dominant mineral-
Irrawadi, the Ganges-Brahmaputra, the Indus, ogy, apart from certain deposits influenced, for
the Amu and the Syr-Darya. They transport huge example, by volcanic sources. Exceptionally,
volumes of silt, the most turbid major fluvial some fine sandy aeolian deposits on dryland mar-
system being the Yellow River and its tributaries, gins are referred to as loess. The airfall nature of
its suspended sediment concentration exceeding loess gives it a distinctive sedimentary fabric,
30 kg/m3 (Ferguson, 1984). Such high silt con- which is an important influence upon a range of
tents are ultimately deposited as extensive allu- bulk properties. The presence of inter-particle
vial plains, and as some of the world’s biggest cement and variations in the generally low per-
deltas. Also, there are important silt components centage of clay mineral in loess are particularly
in the deltas of the Mississippi, the Rhine and the influential factors.
Danube, and the role of river transport has been Loess terrain varies from plateaus to small
invoked in explanation of the complex silt and hills, and is frequently cut by vertical fissures,
sand facies of the Argentine pampa (e.g. Zarate especially near the surface, although fissures
and Blasi, 1988, 1993; Rabassa, 1990; Iriondo, occur in loess of all ages. Loess permeability is
1990). dictated by these systems rather than by the
The Yellow River (Hwang He), with more than porosity of the intact loess. This imposes a dis-
70 tributaries in its 750 000 km2 catchment, flows tinctive hydrology on thick loess terrain, involv-
through extensive drylands in its middle reaches. ing enlargement of the joints by suffosion,
Erosive inputs of loessic soils to the river in leading to an integrated subterranean system of
the drylands of Ningxia and Inner Mongolia pipes and chambers. The appearance and general
and, further downstream, in the Loess Plateau, handling properties of most loess are those of a
reach rates of 25 000 t/km2 per year, with some soft rock. However, though hard or firm when
individual tributary catchments yielding up to dry, much loess has a tendency to liquefy on
50 000 t/km2 per year. It has been estimated that approaching saturation, so that it may properly be
the Loess Plateau contributes some 75% of the regarded as metastable. Such hydrocompaction is
suspended sediment load of the Yellow River common and often constitutes a major geological
(Ren and Shi, 1986). In dramatic contrast, the hazard.
North China Plain and the Yellow River delta Failure of loess slopes is expressed as several
have gradients as low as 0.1–0.2%. The river, types of landslide, best documented from north-
flowing across this plain in levées up to 10 m ern China. Some of the largest, most rapid and
high, has regularly shifted channel course during most destructive landslides are those found in
recent historical time, often flooding up to areas where a thick loess mantles a hilly palaeo-
250 000 km2, and posing a major hazard for more relief consisting of low permeability rocks. These
than 120 million people. At extreme suspended distinctive sedimentary, hydrological and geot-
sediment concentrations exceeding 1300 kg/m3, echnical properties provide a key to the under-
the river enters the hyper-concentrated condition standing of the bulk behaviour of loess, and the
capable of blocking its own flow, as happened at susceptibility of loess terrain to rapid degrada-
Tongguan in 1977. tion. The major loess landscapes of Asia are sub-
ject to some of the highest rates of erosion on
Earth, their rivers being among the most turbid on
25.5 Conclusions the planet. Loess landscapes thus continue to be a
serious threat to the lives and livelihoods of
Typically consisting of accumulations of wind- millions of people. They constitute a substantial
transported and deposited silts, but sometimes challenge to developers, planners and engineers
containing a variable amount of clay and fine around the world.
GFE-25.qxd 6/27/05 2:31 PM Page 724
References Derbyshire, E. (1983b) Origin and characteristics of

some Chinese loess at two locations in China. In
Assallay, A. M., Rogers, C. D. F. and Smalley, I. J. Brookfield, M. E. and Ahlbrandt, T. S. (eds) Eolian
(1996) Engineering properties of loess in Libya. Sediments and Processes. Elsevier, Amsterdam,
Journal of Arid Environments 32, 373–386. 69–90.
Assalay, A. M., Rogers, C. D. F., Smalley, I. J. and Derbyshire, E. (1984) Granulometry and fabric of the
Jefferson, I. F. (1998) Silt: 2–62 m, 9–4. Earth loess at Jiuzhoutai, Lanzhou, People’s Republic of
Science Reviews 45, 61–88. China. In Pecsi, M. (ed.) Lithology and Stratigraphy
Billard, A. and Muxart, T. (2000) Mineralogy, chemistry of Loess and Palaeosols. Hungarian Academy of
and elemental composition of loess and palaeosols. Science, Budapest, 95–103.
In Derbyshire, E. Meng, X. M. and Dijkstra, T. A. Derbyshire, E. (1988) Granulometry and fabric of
(eds) Landslides in the Thick Loess Terrain of Quaternary silts from Eastern Asia. Proc. of Second
Northwest China. Wiley, Chichester, 65–71. Conference of the Palaeoenvironment of East-Asia
Billard, A., Muxart, T., Derbyshire, E., Wang, J. T. and from the Mid Tertiary 1, 215–245.
Dijkstra T. A. (1993) Landsliding and land use in the Derbyshire, E. (2001) Geological hazards in loess ter-
loess of Gansu Province, China. Zeitshrift für rain, with particular reference to the loess regions of
Geomorphologie Suppl. Band 87, 117–131. China. Earth Science Reviews 54, 231–260.
Browzin, B. S. (1985) Granular loess classification Derbyshire, E. and Mellors, T. W. (1986) Loess. In
based on loessial fraction. Bull. Assoc. Eng. Geol. 22, Fookes, P. G. and Vaughan, P. R. (eds) A Handbook
217–227. of Engineering Geomorphology. Surrey University
Brunsden, D. (1984) Mudslides. In Brunsden, D. and Press and Blackie, Glasgow, 258–269.
Prior, D. B. (eds) Slope Instability. Wiley, Chichester, Derbyshire, E. and Mellors, T. W. (1988) Geological
363–418. and geotechnical characteristics of some loess and
Brunsden, D. and Prior, D. B. (eds) (1984) Slope loessic soils from China and Britain: a comparison.
Instability. Wiley, Chichester. Engineering Geology 25, 135–175.
Cao, B. (1986) The geologic characteristics of the Sala Derbyshire, E. and Meng, X. M. (1995) The landslide
Shan type of super landslide and a model for spatial hazard in North China: characteristics and remedial
prediction. Proceedings of the 5th International measures at the Jiaoshuwan and Taishanmiao slides
IAEG Conference 3, 1989–97. in Tian Shui city, Gansu Province. In McGregor, D. F.
Cao, B., Yang, Z. and Zheng, X. (1997) Movement M. and Thompson, D. A. (eds) Geomorphology and
mechanism and disaster prediction of typical high Land Management in a Changing Environment.
speed landslide in Northwest China. In Sassa, K. Wiley, Chichester, 89–104.
(ed.) International Symposium on Landslide Hazard Derbyshire, E., McGown, A. and Radwan, A. (1976)
Assessment. Kyoto University, Japan, 247–260. ‘Total’ fabric of some till landforms. Earth Surface
ISBN 4-9900618-0-2. Processes 1, 17–26.
Catt, J. A. (1978) The contributions of loess to soils in Derbyshire, E., Gregory, K. J. and Hails, J. R. (1980)
lowland Britain. In Limbrey, S. and Evans, J. G. (eds) Geomorphological Processes. Dawson-Westview
The effect of man on the landscape: the lowland zone. Press, Folkestone, UK and Boulder, USA, 312pp.
Council for British Archaeology Research Report 21, Derbyshire, E., Billard, A., van Vliet-Lanoë, B.,
12–20. Lautridou, J.-P. and Cremaschi, M. (1988) Loess and
Catt, J. A. (2001) The agricultural importance of loess. palaeoenvironment: some results of a European joint
Earth Science Reviews 54, 213–229. programme of research. Journal of Quaternany
Coudé-Gaussen, G. (1987) The presaharan loess: sedi- Science 3, 147–169.
mentological characterisation and palaeoclimato- Derbyshire, E., Wang, J. T. and Smalley, I. J. (1990)
logical significance. GeoJournal, 15, 177–183. Loess landslides and geotechnical classification.
Cruden, D. M. (1991) A simple definition of a Landslide News 4, 21–23.
landslide. Bulletin International Association for Derbyshire, E., Wang, J. T., Jin, Z. X., Billard, A.,
Engineering Geology 43, 27–29. Egels, Y., Kasser, M., Jones, D. K. C., Muxart, T.
Dai, Y. S. (1987) The engineering geological characteris- and Owen, L. (1991) Landslides in the Gansu loess
tics of the loess and soil erosion in the Middle of China. Catena Supplement 20, 119–145.
Reaches of the Huanghe River. In Liu, T. S. (ed.) Derbyshire, E., Dijkstra, T. A., Billard, A., Muxart, T.,
Aspects of Loess Research. China Ocean Press, Smalley, I. J. and Li, Y. J. (1993) Thresholds in a
Beijing, 432–436. sensitive landscape: the loess region of central
Derbyshire, E. (1983a) On the morphology, China. In Thomas, D. S. G. and Allison, R. J.
sediments and origin of the Loess Plateau of central (eds) Landscape Sensitivity. Wiley, Chichester,
China. In Gardner, R. and Scoging, H. (eds) Mega- 97–127.
geomorphology. Oxford University Press, Derbyshire, E., Dijkstra, T. A., Smalley, I. J., Rogers, C.
172–194. D. F. and Li, Y. J. (1994) Failure mechanisms in
GFE-25.qxd 6/27/05 2:31 PM Page 725
LOESS 725
loess and the effects of moisture content changes on metastable soils of east Kent. Quart. J. Eng. Geol. 2,
remoulded strength. Quaternary International 24, 103–128.
5–15. Fookes, P. G. and Knill, J. L. (1969) The application of
Derbyshire, E., Kemp, R. and Meng, X. M. (1995a) engineering geology in the regional development of
Variations in loess and palaeosol properties as indi- northern and central Iran. Engineering Geology 3,
cators of palaeoclimatic gradients across the Loess 81–120.
Plateau of North China. Quaternary Science Frye, J. C., Glass, H. D. and William, H. B. (1962)
Reviews 14, 691–699. Stratigraphy and mineralogy of the Wisconsinan
Derbyshire, E., Dijkstra, T. A. and Smalley, I. J. (eds) loesses of Illinois. Illinois State Geological Survey
(1995b) Genesis and Properties of Collapsible Soils. Circular 334, Urbana, Illinois, USA, 55pp.
NATO ASI Series C: Mathematical and Physical Gao, G. (1983) Microstructure of loess soil in China
Sciences, Vol. 468, Kluwer, Dordrecht, 413pp. relative to geological environment. Geological
Derbyshire, E., van Asch, T., Billard, A. and Meng, X. Environment and Soil Properties. Special Publica-
M. (1995c) Modelling the erosional susceptibility of tion, ASCE Geotechnology, Engineering Division,
landslide catchments in thick loess: Chinese varia- Houston, Texas, 121–136.
tions on a theme by Jan de Ploey. Catena 25, Gibbs, H. J. and Holland, W. Y. (1960) Petrographic and
315–331. engineering properties of loess. United States Dep-
Derbyshire, E., Meng, X. M. and Kemp, R. A. (1998) artment of the Interior, Bureau of Reclamation,
Provenance, transport and characteristics of modern Engineering Monograph No. 28, Denver, Colorado,
aeolian dust in western Gansu Province, China, and 37pp.
interpretation of the Quaternary loess record. Gong, S. and Jiang, D. (1979) Soil erosion and its
Journal of Arid Environments 39, 497–516. control in small watershed of the loess plateau.
Derbyshire, E., Meng, X. M. and Dijkstra, T. A. (eds) Scientia Sinica 22, 1302–1313.
(2000) Landslides in the Thick Loess Terrain of Grabowska-Olszewska, B. (1988) Engineering–
Northwest China. Wiley, Chichester, 288pp. Geological problems of loess in Poland. Enginee-
Dijkstra, T. A., Derbyshire, E. and Meng, X. M. (1993). ring Geology 25, 177–199.
Neotectonics and mass movements in the loess of Grim, R. E. (1953) Clay Mineralogy. McGraw-Hill,
north central China. Quaternary Proceedings 3, New York.
93–110. Handy, R. L. (1973) Collapsible loess in Iowa. Procs.
Dijkstra, T. A., Rogers, C. D. F., Smalley, I. J., Soil Sci. Soc. America 37, 281–284.
Derbyshire, E., Li, Y. J. and Meng, X. M. (1994) The Hutchinson, J. N. (1988) Morphological and geotechni-
loess of north-central China: geotechnical proper- cal parameters of landslides in relation to geology
ties and their relation to slope stability. Engineering and hydrology, general report. In Bonnard, C. (ed.)
Geology 36, 153–171. Landslides. Fifth Int. Symposium on Landslides
Dikau, R., Brunsden, D., Schott, L. and Ibsen, M.-L. Vol. 1, 3–35.
(1996) Landslide Recognition; Identification, Iriondo, M. H. (1990) A late Holocene dry period in the
Movement and Causes. Wiley, Chichester, 251pp. Argentine plains. In Rabassa, J. (ed.) Quaternary of
Ding, Z. L., Sun, J. M., Liu, T. S., Zhu, R. X., Yang, S. L. South America and Antarctic Peninsula. Balkema,
and Guo, B. (1998) Wind-blown origin of the Rotterdam, 197–218.
Pliocene red clay formation in the central Loess Iriondo, M. H. (1997) Models of deposition of loess and
Plateau, China. Earth and Planetary Science Letters loessoids in the Upper Quaternary of South America.
161, 135–143. J. South American Earth Sciences 10, 71–79.
Eden, D. N. and Hammond, A. P. (2003) Dust accumu- Iriondo, M. H. and Kröhling, D. M. (1995) El Sistema
lation in the New Zealand region since the Last Eólico Pampeano. Com. Mus. Prov. Ca. Naturales
Glacial Maximum. Quaternary Science Reviews, (Nueva Serie ‘Florentino Ameghino’), Santa Fe,
22, 2037–2052. Argentina (ISSN 0325-3856), 45pp.
Feda, J. (1966) Structural stability of subsident loess Jefferson, I. and Smalley, I. J. (1995) Six definable par-
soil from Praha-Dejvice. Engineering Geology 1, ticle types in engineering soils and their participa-
201–219. tion in collapse events: proposals and discussions.
Ferguson, R. I. (1984) Sediment load of the Hunza In Derbyshire, E., Dijkstra, T. A. Smalley, I. J. (eds)
River. In Miller, K. J. (ed) International Karakoram Genesis and Properties of Collapsible Soils. NATO
Project. Vol. 2, Cambridge University Press, ASI Series C: Mathematical and Physical Sciences,
581–598. Vol. 468, Kluwer, Dordrecht, 19–32.
Folk, R. L. and Ward, W. C. (1957) Brazos River bar. Karlson, A., Tauber, A. and Torres Anza, D. (1993)
A study in the significance of grainsize parameters. Tipificación mineralogical de paleosuelos del sur de
J. Sed. Pet., 27, 3–27. la provincia de Córdoba. Procs. XIV Congreso
Fookes, P. G. and Best, R. (1969) Consolidation char- Argentino de la Ciencia del Suelo, Mendoza,
acteristics of some late Pleistocene periglacial Argentina, 463–474.
GFE-25.qxd 6/27/05 2:31 PM Page 726
Kemp, R. A., Derbyshire, E. and Meng, X. M. (2001) Engineering Geology Special Publications 15,
A high-resolution micromorphological record of 141–153.
changing landscapes and climates on the western Meng, X. M. and Zhang, S. W. (1989) Investigation of
Loess Plateau of China during oxygen isotope stage 5. dangerous loess slopes in Lanzhou region. Internal
Palaeogeography, Palaeoclimatology, Palaeoecology Report, 89pp. (in Chinese).
170, 157–169. Meng, X. M., Dijkstra, T. A. and Derbyshire, E. (2000)
Keš, A. S. (1972) On spreading and forming loess Loess slope stability. In Derbyshire, E., Dijkstra, T. A.
relief in Europe. Acta Geol. Acad. Sci Hungar. 16, and Smalley, I. J. (eds) Genesis and Properties of
359–370. Collapsible Soils. NATO ASI Series C: Mathematical
Krinitzsky, E. L. and Turnbull, W. J. (1967) Loess and Physical Sciences, Vol. 468, Kluwer, Dordrecht,
Deposits of Mississippi. Geol. Soc. America, Special 173–202.
Papers No. 94, 64pp. Moss, A. J. (1966) Origin, shaping and significance of
Krinsley, D. H. and Smalley, I. J. (1973) Shape and nature quartz sand grains. J. Geol. Soc. Australia 13, 97–136.
of small sedimentary quartz particles. Science 180, Moss, A. J. and Green, P. (1975) Sand and silt grains:
1277–1279. predetermination of their formation and properties
Lautridou, J.-P. (1993) L’eau dans les loess de by micro fractures in quartz. J. Geol. Soc. Australia
Normandie. Quaternaire 4, 91–96. 22, 485–495.
Lin, Z. (1995) Variation in collapsibility and strength of Muxart, T., Billard, A., Derbyshire, E. and Wang, J. T.
loess with age. In Derbyshire, E., Dijkstra, T. A. and (1994) Variation in runoff on steep unstable loess
Smalley, I. J. (eds) Genesis and Properties of slopes near Lanzhou, China: initial results using
Collapsible Soils. NATO ASI Series C: Mathematical rainfall simulation. In Kirkby, M. J. (ed.) Process
and Physical Sciences, Vol. 468, Kluwer, Dordrecht, Models and Theoretical Geomorphology. Wiley,
413pp. Chichester, 337–355.
Lin, Z. and Liang, W. (1982) Distribution and engineer- Muxart, T., Billard, A., Andrieu, A., Derbyshire, E. and
ing properties of loess and loesslike soils in China. Meng, X. M. (1995) Changes in water chemistry
Schematic map of engineering geological zonation. and loess porosity with leaching: implications for
Bull. Int. Assoc. Eng. Geol. No. 21, 112–117. collapsibility in the loess of North China. In
Liu, T. S. (ed.) (1985) Loess and the Environment. Derbyshire, E. Dijkstra, T. A. and Smalley, I. J. (eds)
Science Press, Beijing, 215pp. Genesis and Properties of Collapsible Soils. NATO
Liu, T. S., Wang, T. M., Wang, K. L. and Wen, C. C. ASI Series C: Mathematical and Physical Sciences,
(1964) Loess in the Middle Reaches of the Vol. 468, Kluwer, Dordrecht, 313–331.
Yellow River. Science Press, Beijing, 234pp. (in Ovejero, R. (1980) Geología y geomorfología de la
Chinese). cuenca del rio San Javier, Depto. Tafi (Tucumán –
Lohnes, R. A. and Handy, R. L. (1968) Slope angles in Argentina). Unpublished thesis, Facultad de
friable loess. J. Geol 76, 247–258. Ciencias Naturales, Tucumán University, Argentina.
Lozek, V. (1965) Das Problem der Lössbildung und die Owen, L. A., Derbyshire, E., White, B. J. and Rendell, H.
Lössmollusken. Eiszetalter und Gegenwart 16, (1992) Loessic silt deposits in the western Himalayas:
61–75. their sedimentology, genesis and age. Catena 19,
Lutenegger, A. J. (1981) Stability of loess in light of the 493–509
inactive particle theory. Nature 291, 360. Phien-wej, N., Pientong, T. and Balasubramaniam, A. S.
Mahaney, W. C. and Hancock, R. G. V. (1990) (1992) Collapse and strength characteristics of loess
Stratigraphy and paleosols in the Sale terrace in Thailand. Engineering Geology 32, 59–72.
loess section, northwestern China. Catena 17, Pye, K. (1987) Aeolian Dust and Dust Deposits.
357–367. Academic Press, London, 334pp.
McTainsh, G. (1987) Desert loess in Northern Nigeria. Pye, K. (1995) The nature, origin and accumulation of
Zeitschrift für Geomorphologie 31, 145–165. loess. Quaternary Science Reviews 14, 653–667.
Mellors, T. (1995) The influence of the clay component Pye, K and Tsoar, H. (1987) The mechanics and geo-
in loess on collapse of the soil structure. In logical implications of dust transport and deposition
Derbyshire, E., Dijkstra, T. A. and Smalley, I. J. (eds) in deserts, with particular reference to loess forma-
Genesis and Properties of Collapsible Soils. NATO tion and dune sand diagenesis in the northern
ASI Series C: Mathematical and Physical Sciences, Negev, Israel. In Frostick, L. E. and Reid, I. (eds)
Vol. 468, Kluwer, Dordrecht, 207–216. Desert Sediments: Ancient and Modern. Geol. Soc.
Meng, X. M. and Derbyshire, E. (1998) Landslides and Lond. Spec. Publ. 35, Blackwell, Oxford, 139–156.
their control in the Chinese Loess Plateau: models Qian, H., Wang, J., Luo, Y., She, G., Shi, G. and Qi, W.
and case studies from Gansu Province, China. In (1985) Foundations on Collapsible Loess. Publishing
Maund, J. D. and Eddleston, M. (eds) Geohazards in House of the Chinese Architecture Industry, 470pp.
Engineering Geology. Geological Society, London, (in Chinese).
GFE-25.qxd 6/27/05 2:31 PM Page 727
LOESS 727
Rabassa, J. (1990) Late Pleistocene and Holocene loess eastern Canada. In Jones, M. and Cosgrove, J. (eds)
deposits in the upper Rio Sauce Grande basin, Geology of Neotectonic Environments. Bellhaven
Sierra de la Ventana, Argentina. International Press, Leichester, 202–219.
Symposium on Loess (Expanded abstracts), Mar del Stevens, K. F. (1988) A preliminary investigation of the
Plata, Argentina, 84–88. physical and chemical properties of the Pahiatua
Ren, M. E. and Shi, Y. F. (1986) Sediment discharge of Terrace loess beds, north-western Waisawapa, New
the Yellow River (China) and its effect on the sedi- Zealand. In Eden, D. N. and Furkert, R. J. (eds)
mentation of the Bohai and Yellow Sea. Continental Loess: Its Distribution, Geology and Soils. Balkema,
Shelf Research 6, 785–810. Rotterdam, 193–199.
Richthofen, F. von (1877–85) China. Ergebnisse Eigener Sun, J. Z. (1988). Environmental geology in loess areas
Reisen und Daran Gegruendeter Studien. Reimer, of China. Environmental Geology and Water Science
Berlin, 5 vols. 12, 49–61.
Richthofen, F. von (1882) On the origin of the loess. Tan, T. K. (1988) Fundamental properties of loess from
Geological Magazine 9, 293–305. northwestern China. Engineering Geology 25,
Rogers, C. D. F. and Smalley, I. J. (1993) The shape of 103–122.
loess particles. Naturwissenschaften 80, 461–462. Tsoar, H. and Pye, K. (1987) Dust transport and the
Rogers, C. D. F. and Smalley, I. J. (1996) The adobe question of desert loess formation. Sedimentology
reaction and the use of loess mud in construction. 34, 139–153.
Eng. Geol. 40, 137–138. Varnes, D. J. (1958) Landslide types and processes.
Rogers, C. D. F. Dijkstra, T. A. and Smalley, I. J. Landslides and Engineering Practice. Nat. Acad.
(1994a) Particle packing from an earth science Sci., Nat. Res. Council, HRB 544, 20–47.
viewpoint. Earth Science Reviews 36, 59–82. Varnes, D. J. (1978) Slope movements: type and
Rogers, C. D. F., Dijkstra, T. A. and Smalley, I. J. (1994b) processes. In Eckel, E. B. (ed.) Landslide Analysis and
The Teton Dam failure (Idaho, USA, 1976). In Control. Transport Res. Board, Special report 176.
Fookes, P. G. and Parry, R. H. G. (eds) Engineering Veder, C. (1981) Landslides and their Stabilization.
Characteristics of Arid Soils. Procs. of the First Inter- Springer-Verlag, New York, 247pp.
national Symposium. Balkjema, Rotterdam, 415–417. Vliet-Lanoë, B. van, Coutard, J.-P. and Pissart, A.
Rózycki, S. Z. (1991) Loess and Loess-like Deposits. (1984) Structures caused by repeated freezing and
Polish Academy of Sciences Publishing House, thawing in various loamy sediments: a comparison
Wroclaw, 187pp. of active, fossil and experimental data. Earth
Sassa, K. (1989) Geotechnical classification of land- Surface Processes and Landforms 9, 553–565.
slides. Landslide News 2 (June 1989). Wang, J. and Derbyshire, E. (1994) Engineering soil
Sayago, J. M. (1995) The Argentine neotropical loess. classification of the loess of Gansu, China, based on
Quaternary Science Reviews 14, 755–766. sedimentary properties: relationships to geotechni-
Sheeler, J. B. (1968) Summarization and comparison of cal behaviour. In Fookes, P. G. and Parry, R. H. G.
engineering properties of loess in the United States. (eds) Engineering Characteristics of Arid Soils.
Highway Research Record No. 212, 1–9. Balkema, Rotterdam, 153–157.
Smalley, I. J. (1966a) Formation of quartz. Nature 211, Wang, Y. and Zhang, Z. (eds) (1980) Loess in China.
476–479. Shaanxi People’s Art Publishing House, Xi’an,
Smalley, I. J. (1966b) The properties of glacial loess China, n.p.
and the formation of loess deposits. Journal of Wentworth, C. K. (1933) Fundamental limits of the
Sedimentary Petrology 36, 669–676. sizes of classic grains. Science 77, 633–634.
Smalley, I. J. (1971) ‘In situ’ theories of loess formation Zarate, M. and Blasi, A. (1988) Depositos loessicos del
and the significance of the calcium carbonate con- Pleistoceno tardio-Holoceno del flanco sudeste de
tent of loess. Earth Science Reviews 7, 67–85. Tandila. CADINQUA, 27pp.
Smalley, I. J. (1972) The interaction of great rivers and Zarate, M. and Blasi, A. (1993) Late Pleistocene–
large deposits of primary loess. Trans. New York Holocene eolian deposits of the southern Buenos
Acad. Sci. 34, 534–542. Aires Province: a preliminary model. Quaternary
Smalley, I. J. (1995) Making the material: the formation International 17, 15–20.
of silt-sized primary mineral particles for loess Zhang, L. (1989) The landslide history and Late
deposits. Quaternary Science Reviews 14, 645–651. Cenozoic environmental factors in Sale Shan area.
Smalley, I. J. and Derbyshire, E. (1990) The definition In Zhang, L. and Siwei, S. (eds) International Field
of ‘ice sheet’ and ‘mountain’ loess. Area 22, Workshop on Loess Geomorphological Processes
300–301. and Hazards. Lanzhou, China, 81–93.
Smalley, I. J. and Derbyshire, E. (1991) Flowslide-type Zhang, Z. (1995) Geological disasters in loess areas
ground failures in the airfall deposits of northern during the 1920 Haiyuan earthquake, China.
China and the shallow-marine postglacial clays of GeoJournal 36, 269–274.
GFE-25.qxd 6/27/05 2:31 PM Page 728
Zhu, H. (1989) Some types of seismic landslides in Acknowledgements

loess area in China. In Zhang, L. and Siwei, S. (eds)
International Field Workshop on Loess Geomorpho-
The authors gratefully acknowledge the support
logical Processes and Hazards. Lanzhou, China,
64–71. provided by the Centre for Quaternary Research
Zhu, Z. Y. and Ding, Z. L. (1994) The Climatic and at Royal Holloway (University of London) and
Tectonic Evolution in the Loess Plateau of China the cartographic help of Jenny Kynaston. We also
during the Quaternary. Beijing, China, Geological acknowledge the provision of sample loess mate-
Publishing House. 226pp. (in Chinese).
rial by David Boardman (Figure 25.12a) and
Sanda Balescu (Figures 25.12b, 25.12c, and
25.15), and original photographs by the late
Professor Wang Jingtai of Lanzhou (Figure 25.25a)
and Dr. Armelle Billard (Figure 25.29b).
GFE-26.qxd 6/27/05 4:47 PM Page 729
26. Chalk Landscapes

Fred Bell and Martin Culshaw
26.1 Introduction one of the reasons why most limestones are better
indurated than chalks.
Chalk has been defined as a soft, fine textured, The Chalk in western Europe continues west-
usually white or grey limestone of marine origin wards from northern Germany, Denmark and
that consists primarily of calcite that was deposited the Netherlands, to France and the British Isles
during Upper Cretaceous and early Tertiary times. (Figure 26.1). An extensive part of the North Sea is
However, hard horizons do occur in the Chalk and floored by chalk where deposition continued into
it can have a reddish or greenish colour due to the Lower Tertiary or Danian times. In central and
presence of iron oxide or glauconite respectively. southern Europe, Upper Cretaceous strata are
Most European chalks were formed by pelagic sed- represented by limestones within the Alpine moun-
imentation in a temperate to tropical shelf sea envi- tain belt. Eastwards from Poland, the Chalk extends
ronment with water depths of between 100 and to the northern slopes of the Caucasus, with exten-
300 m. sions to the Black Sea, Iraq, the Caspian Sea and
Generally, chalk is a remarkably pure micritic south-western Siberia (Figure 26.1). It also occurs
limestone, which, excluding flints and marl around the Mediterranean Sea from the Greek
bands, contains over 95% calcium carbonate, islands into Turkey, in Israel, Sinai, Egypt, Libya
which can be divided into coarse and fine frac- and Algeria. Accordingly, the Chalk of the
tions. The Chalk Marl in the Lower Chalk is an European, Asian and African basins covers a vast
exception, it containing an appreciable amount of area. Outside this region, chalk is found in North
silt and clay. The coarse fraction, which may con- America. For example, the Niobrara Chalk is pres-
stitute 20 to 30%, falls within the 40 to 100 ␮m ent in Kansas and Nebraska; the Selma Chalk in
range. This contains material derived from the Alabama, Mississippi and Tennessee; the Austin
mechanical breakdown of large shelled organ- Chalk in Texas, and the San Felipe Chalk in
isms and, to a lesser extent, from foraminifera. Mexico. Mortimore (1990) indicated that chalk also
The fine fraction, which takes the form of calcite occurs in Indonesia, Western Australia and New
particles that may be less than one micron in size, Zealand and has been found in cores obtained by
is composed almost entirely of coccoliths and the Ocean Drilling Program (ODP) (Figure 26.1).
may form up to, and sometimes over, 80% of The Chalk of England is divided into three prin-
certain horizons. Flints are common, especially cipal depositional and faunal provinces (Figure
in the Upper Chalk. 26.2); further details can be found in Mortimore,
Most of the material constituting the Chalk 1983; Bristow et al., 1997; Mortimore et al., 2001.
was deposited as low-magnesium calcite, with Table 26.1 shows the main bio- and litho-strati-
less than 5 mole per cent magnesium in the graphic sub-divisions of the Cretaceous Chalk
lattice. Although low-magnesium calcite is not Group of England and the North Sea.
uncommon in other ancient limestones, usually Chalk varies in hardness. Hard chalks may be
the carbonate was deposited as aragonite and high associated with the formation of rhombic calcite
magnesium calcite that then, because it is less crystals that had a range of sizes, pore filling micro-
stable at ordinary temperature and pressure condi- spar and cement overgrowths on coccoliths. On the
tions, reverted to a low-magnesium form. This is other hand, the individual particles in soft chalk are
729
GFE-26.qxd 6/27/05 4:47 PM Page 730
Figure 26.1 Worldwide distribution of chalks (after Mortimore, 1990). The main shaded areas are the known
extent of chalk; the small circles are sites where the Ocean Drilling Program found Chalk in cores.
The Alpine orogeny in Tertiary, primarily

Miocene, times affected southern England, and
thereby the Chalk, giving rise to structures with a
dominant east-west trend (Figure 26.3). Tectonic
compaction, pressure solution and late-stage solu-
tion have altered the engineering behaviour of
chalk (Clayton, 1983). Stresses due to Alpine earth
movements brought about mechanical disaggrega-
tion of the coccolith matrix, followed by reconsol-
idation, which then was followed by pressure
solution and reprecipitation of the calcite under
sustained pressures. Such action affected the
strength and density of chalk. For instance, in areas
where the Chalk dips at angles in excess of 30⬚, the
density of some of the weakest chalks has been
increased to that of the hardest hard-ground.
Nonetheless, the density of tectonically modified
Figure 26.2 Distribution of the Chalk outcrop in chalks varies widely over relatively short distances.
England.
bound together at their points of contact by thin 26.2 Geomorphology of the

films of calcite. Such chalk contains only minute Chalklands
amounts of cement. Diagenesis, tectonic com-
paction, pressure solution and late-stage solution The Chalk in England generally gives rise to a dis-
have all altered the properties of chalk from place to tinctive rolling country. Its outcrop normally forms
place. a prominent escarpment with hard bands within the
GFE-26.qxd 6/27/05 4:47 PM Page 731
CHALK LANDSCAPES 731
Table 26.1 English Chalk stratigraphy (after Rawson et al., 2001 and Mortimore et al., 2001).
Southern Province Northern Province
Old units Stage Sub-group Formation Member Formation Member
Portsdown Chalk Rowe Chalk
Campanian Spetisbury Chalk

Culver Chalk
Tarrant Chalk
Flamborough
Upper Newhaven Chalk
Chalk
Chalk Santonian White
Chalk Seaford Chalk
Coniacian Burnham Chalk
Lewes Nodular
Chalk
Turonian
New Pit Chalk
Middle Welton Chalk
Chalk Holywell Nodular
Chalk Plenus Marls Plenus
Marls
Lower Cenomanian Grey Chalk Zig/Zag Chalk Ferriby Chalk

Chalk West Melbury
Marly Chalk
Chalk forming intermediate escarpments. The Some patches of clay-with-flints may be over 10 m
Chalk escarpment and dip slopes represent the in thickness. As the name implies, flints and clay
surviving remnants of a sub-Palaeogene erosion constitute the bulk of the material but some gravel,
surface that has been modified by late Tertiary and sand and silt may be present at some localities.
Quaternary geomorphological processes. It may be Clay-with-flints generally has been considered as
that the crest of the Chalk escarpment was nearing having been derived from the Chalk, it being a
its present position by the end of Tertiary times and residual product left after chalk was dissolved.
has undergone little further retreat subsequently. However, in some localities in the Weald of south-
In places, superficial deposits rest directly on east England it would appear to have been derived
the Chalk. For example, tills of Anglian and from Tertiary beds. It tends to be brownish in
Devensian age are found on the west and east colour and may darken with depth. The deposit was
flanks of the Lincolnshire Wolds respectively. more widely distributed than it is at present and has
Similarly, in East Anglia much of the Chalk is cov- been dissected by valleys that now are dry. Angular
ered with tills, primarily of Anglian age. These tills flint gravel occurs on the higher parts of the down-
contain fragments of chalk of varying size, some of lands, notably in south Dorset and the Isle of
which may be several cubic metres in volume. Wight. This consists of an accumulation of unworn
Another superficial deposit, which is associated and sometimes fragmentary flints in a matrix of
with the Chalk of southern England, is clay-with- quartz sand or chalk rubble. It is thought to have a
flints. These deposits tend to occur scattered in similar derivation to that of clay-with-flints.
patches over the high ground of the Chalk, and at During Devensian times much of the Chalk
times are found in sinkholes and solution pipes. outcrop was unglaciated and, as such, the surface
GFE-26.qxd 6/27/05 4:47 PM Page 732
Figure 26.3 Structural trends in the southern England Chalk (after Evans and Stubblefield, 1929).
was subject to periglacial activity. Frost shatter- occurrence of dry valleys and the scarp edge of the
ing of chalk and mass wasting of valley sides Chalk is frequently notched by wind or dry gaps,
occurred, and because of lower sea levels the that is, dry valleys beheaded by the recession of
existing drainage system was rapidly incised and the scarp (Figure 26.4). Another significant breach
accentuated. Some Tertiary deposits were eroded in the Chalk outcrop is made by the River Humber.
from above the Chalk and some superficial
deposits were removed from valley slopes by hill- Dry valleys
wash and solifluction. These materials now tend Dry valleys probably represent the remnants of
to occupy valley floors and the base of slopes in a drainage system that was developed on the
chalkland areas. They may have been affected by Palaeogene sediments covering the dip slopes of
cryoturbation. the Chalk. Birch and Griffiths (1996) suggested
that the position and orientation of dry valleys are
Drainage patterns controlled by subtle structural features that have
A trellis pattern of drainage is developed over been superimposed on the Chalk due to flexuring
much of the chalklands, which is characteristic of and possibly faulting in older Mesozoic rocks
scarpland topography, with examples of river beneath. The form of dry valleys indicates that they
piracy (where one stream has eroded to intersect were excavated by flowing water. For instance,
and capture another), obsequent streams (streams typical chalk landscape consists of convex divides,
that flow opposite to the original downslope direc- and reasonably steep-sided trough-shaped valleys.
tion following erosion at the head capturing part of However, there are numerous variations, valleys
the original stream) and wind gaps occurring in may be wider and shallower or they may be asym-
places. The Chalk escarpment of England is rarely metrical in cross-section. At times the tributaries
crossed by valleys from the west, the notable of a system hang slightly above the main valleys,
exception being the Goring Gap through which whether wet or dry. A small rise in the water table
the River Thames flows into the London Basin. often is sufficient to provide the larger valleys
A notable feature of chalkland scenery is the with streams, as is illustrated by bournes. These
GFE-26.qxd 6/27/05 4:47 PM Page 733
Figure 26.4 The Devil’s Dyke, Poynings, near Brighton, Sussex, England. The Dyke is a deep dry valley in the South
Downs escarpment formed by the capture and rejuvenation of a dip-slope valley by a scarp-facing spring and stream
(British Geological Survey photograph no. A13405, NERC (Natural Environment Research Council) copyright).
are intermittent streams that occur in larger dry to a large extent by the level of the discharge of
valleys after heavy rainfall. Bourne flows may last springs at the foot of the escarpment and that the
for a few weeks or months. There frequently is a escarpment has receded as a result of erosion. The
lag in time between the maximum heavy rainfall height of the escarpment increases as it retreats,
and the maximum flow of a bourne because of the which leads to a relative fall in the position of the
time taken for the rainwater to percolate through water table; the latter also is affected by the sur-
the Chalk and thereby cause the water table to rise. face of the clay beneath the Chalk being lowered.
The head of a bourne may extend upsteam after it
makes its first appearance. Coombes
Three explanations have been advanced for the Coombes are another topographical feature charac-
formation of dry valleys. As streams flow in some teristic of chalklands. These are bowl-shaped hol-
dry valleys after heavy rainfall, it has been sug- lows that occur on the flanks of dry valleys. Some
gested that dry valleys may have been carved by coombes that cut back into scarp slopes are similar
water during times when the climate was wetter to corries, except in size, and may have been
and the water table therefore higher. Alternatively, formed by frost action working back into snow-
it has been suggested that these valleys could have filled hollows on a hillside during Pleistocene
been formed when the ground was frozen during times. Other coombes resemble overflow channels
Pleistocene times. This would mean that the dis- in form but strike into hillsides and end blindly in
persal of precipitation, because it could barely a steep face. A number of such coombes often
infiltrate the frozen ground, was confined pre- occur together, for example, in the South Downs,
dominantly to runoff. The third explanation main- and may have been formed by spring sapping near
tains that the height of the water table is influenced the base of scarps under the periglacial conditions
GFE-26.qxd 6/27/05 4:47 PM Page 734
that existed in the south of England during the controls on the process of dissolution. Although the
Pleistocene. Springs presumably flowed more porosity of chalk may be high, the pore size is
vigorously as ice thawed and still-frozen ground small (see above). Consequently, most of the water
meant that all meltwater was concentrated in that flows through chalk is concentrated along
runoff. The material eroded from coombes formed systems of discontinuities. Dissolution along dis-
extensive fans of reworked chalk debris that accu- continuities can give rise to the formation of
mulated downhill of the coombes and over the solution pipes, which tend to develop at the inter-
slopes at the base of the escarpments. This material section of discontinuites. Enlarged discontinuity
is referred to as coombe rock. systems mean that the flow of water through chalk
increases. For instance, Atkinson and Smith (1974)
Dissolution mentioned a flow rate of 2 to 3 km per day through
Chalk being a carbonate rock, is subject to dissol- a well-developed system in chalk in Hampshire,
ution. The degree of aggressiveness of water to and Pitman (1983) referred to transmissivities of
calcium carbonate can be assessed on the basis of approximately 1200 m per day in fissured zones in
the relationship between the dissolved carbonate chalk in east Yorkshire.
content, the pH value and the temperature of the Dissolution aids disintegration of chalk by
water. At any given pH value, the cooler the water weakening its fabric and by emphasising structural
the more aggressive it is. As solution continues its weaknesses, however slight. Any enlargement of
aggressiveness decreases and eventually ceases pores brings about an increase in stress within the
when saturation is reached. Hence, solution is fabric of the rock thus reducing its strength and
greatest when the bicarbonate saturation is low. increasing stress corrosion. Accordingly, dissol-
This occurs when water is circulating, and fresh ution can weaken the bonding between grains,
supplies with low lime saturation are available producing a weaker material. In fact, such break-
continuously. down ultimately can give rise to putty chalk.
Calcium carbonate dissolves at a rate and in a
manner which is influenced by its solubility Sinkholes
(0.015 kg/m3 at 10 ⬚C in pure water) and specific Sinkholes are a part of chalkland landscape (Figure
solution rate constant (0.4 m/s ⫻ 103 at 10 ⬚C with 26.5), as well as areas where the Chalk occurs
a flow rate of 0.05 m/s). Aqueous dissolution of beneath shallow overlying deposits. Solution fea-
calcium carbonate introduces the carbonate ion tures, including cavities, occur across the outcrop
into water (Weyl, 1959; Garrels et al., 1960). of the Chalk. Although sinkhole formation is
Further slight dissociation releases H⫹ and OH⫺ occurring now, Banks et al. (1995) suggested that
ions, after which H⫹, HCO3⫺, CO32⫺, OH⫺ and it was probably more noteworthy under periglacial
Ca2⫹ become active. The relative distribution of conditions towards the end of the Pleistocene.
H2CO3, H⫹, HCO3⫺ and CO32⫺ in solution as a Culshaw and Waltham (1987) described four
function of pH is: below pH 5, H2CO3 dominates; types of sinkholes formed in more soluble rocks
from pH 6 to 10, HCO3⫺ dominates; above pH 10, (solution, collapse, subsidence and buried) while
CO32⫺ dominates. The pH of rainwater is com- Waltham and Fookes (2003) proposed a six-fold
monly 5 to 5.5 (although it may range from below classification in which collapse sinkholes were
4 to more than 7), but in contact with CaCO3, it subdivided into collapse, caprock and dropout
rises to about 8 to 8.2. In the typical range of pH sinkholes. In chalk, solution sinkholes of signifi-
7 to 9 of waters draining from carbonate material, cant size are rare, and collapse sinkholes almost
the dominant anion associated with dissolved non-existent because of the absence of large
Ca2⫹ is HCO3⫺. Thus, CaCO3 is dissolved slightly caves. Large subsidence sinkholes are known in
as CaCO3 but predominantly as Ca2⫹ with sand-covered chalk areas in Norfolk and Dorset
HCO3⫺, and less as Ca2⫹ with CO32⫺. (UK) but are inactive; active subsidence sinkholes
Both the rate of flow and the area of the are small and of limited distribution. Buried sink-
material exposed to flowing water are significant holes are commonly found near the contact of the
GFE-26.qxd 6/27/05 4:47 PM Page 735
Figure 26.5 Buried sinkholes in the Chalk of north Kent (UK) exposed in excavations for the Channel Tunnel Rail
Link (photograph courtesy of Richard Ellison, British Geological Survey, NERC copyright).
Chalk and the overlying Tertiary and superficial sinkhole, then its presence may not be seen at the
deposits. This is largely due to the concentration surface. Roof collapse of the cavity means that it
of runoff at such locations. These sinkholes are migrates upwards and eventually it may occur at the
usually small, and wide and shallow in shape, surface without warning. When streams disappear
though deep narrow pipes occur in some areas into sinkholes, then the feature is called a swallow
(for example, north of London). hole. The diameter of sinkholes varies from a few
The enlargement of pipes or cavities by continu- metres up to 20 m or more and occasionally can be
ing dissolution can lead to sudden collapse with 100 m across or even greater. On occasion numer-
associated subsidence at the surface. Moreover, ous sinkholes may occur. For example, the Mimms
voids can migrate gradually upwards through chalk Hall Brook in Hertfordshire flows over the London
due to material collapsing. Collapsed material Clay and onto the Chalk going underground via
occupying a sinkhole may possess a metastable sinks at Water End. The many sinkholes that have
structure; for instance, a zone of loosely packed developed have merged into a single depression so
ground may be present if chalk, sand and gravel that a temporary lake may form after heavy rainfall
collapsed into a sinkhole. Lowering of the Chalk because the sinkholes have been choked by mud
surface beneath overlying deposits due to solution and cannot convey water away. As this example
can occur, disturbing the latter deposits and lower- illustrates, a common location of sinkholes is near
ing their degree of packing. Hence, the Chalk the boundaries of overlying relatively impervious
surface may be extremely irregular in places. London Clay or clay-with-flints. Streams flowing
Most sinkholes can be detected at the surface but over these beds descend into sinkholes after reach-
if a cavity develops in clayey deposits overlying a ing the Chalk, the latter often having an aligned
GFE-26.qxd 6/27/05 4:47 PM Page 736
occurrence. Sinkholes also may occur in valleys, on diameter and a few metres deep (Figure 26.6).
slopes or near the tops of hills. Conversely, they Indeed, West and Dumbleton (1972) suggested that
appear to be absent from areas of bare chalk such as pipes a metre in depth could form within 10 years
found on Salisbury Plain. Some sinkholes have a where the infiltration of rainfall into the ground is
smaller steeper depression within the normal concentrated. They further suggested that such a
smooth depression, which suggests that they have concentration could be due to local variations in the
been reactivated. topography or permeability of the Chalk or shallow
As noted, sinkholes also may develop beneath a overburden, or could occur at the intersections of
shallow cover of sand or gravel, which subsides as major joints in the Chalk. West and Dumbleton con-
the sinkhole grows. These areas of sands and grav- cluded that care should be taken to avoid the risk of
els commonly possess a heathland vegetation that pipe formation or the reactivation of sinkholes due
gives rise to humic acid, which presumably aids to excessive concentration of water into chalk from
the dissolution process. Exposures in quarry faces hard surfaces such as major roads, soak-aways,
in the Chalk show that sinkholes narrow rapidly leaking water mains or sewers, and that soak-aways
downward, holes of 6 m diameter or more at the also should be located with care. Wherever possi-
surface tapering to less than a metre diameter ble, engineering structures obviously should avoid
within a shallow depth in the Chalk (Figure 26.5). areas affected by sinkholes. If they cannot be
Some of the older, better developed sinkholes may avoided, then the sinkholes will require treatment
descend to over 15 m in the Chalk. Dissolution (see below). Sinkholes, their investigation, assess-
may have taken place along bedding planes crossed ment and consequences for engineering are dis-
by such sinkholes. Normally, these sinkholes are cussed by Waltham et al. (2005).
occupied by gravelly, sandy or silty material, as are
the associated enlarged bedding planes. Cavern systems
New sinkholes often appear at the surface after a True karstic cavern systems, which are accessible
period of heavy rainfall, and may form a circular for exploration, have been recorded in the Chalk.
shaft or pipe with steep sides a metre or two in For example, Banks et al. (1995) mentioned that
Figure 26.6 Solution pipes at Seaford Head, Sussex (UK) (photograph courtesy of Andy Farrant, British
Geological Survey, NERC copyright).
GFE-26.qxd 6/27/05 4:47 PM Page 737
such systems have been encountered in south-east hollows to inverted cone-like depressions of
England during sewer construction and well significantly greater diameter than that of the
drilling. They also have been exposed in cliff associated void.
sections, perhaps the most notable being at Beachy The height to which voids can migrate depends
Head where 354 m of accessible passageway have largely on the bulking factors of the roof rocks and
been explored. Furthermore, Bradshaw et al. the height of the workings. When a fall occurs, the
(1991) referred to cavern systems up to 2 km in material involved bulks so that there comes a point
length in the Chalk of the Paris Basin. Indeed, at which upward migration is arrested unless the
Lowe (1992) suggested that sub-anthropogenic workings are near the surface. If the bulking factor
caves (that is, caves too small to be explored by of chalk is around 1.3, then the height above
man) are probably much more common in the the roof of the workings to which a void would
Chalk than previously have been appreciated. migrate is approximately 2.5 to 3.5 times the
height of the workings. However, void migrations
Subsidence in chalk due to mining in excess of this have been recorded but most of
The Chalk has been, and still is, extensively worked these are due to the near-surface weakening effects
for lime. In the past many lime workings took the of periglacial action and of dissolution.
form of surface quarries and pits often with associ- Old mine workings are entered by shafts, drifts
ated subsurface mining. These mine workings were or adits that run directly from the surface. Each
located at shallow depths and, because they lay mine could have more than one entrance. Unless
above the water table, were dry. Subsurface work- filled correctly, these also represent hazards. In the
ings took the form of, and relied on the strength of, case of shafts, many have been only partially filled,
the Chalk to form self-supporting roofs. others may be capped and the cap covered with fill.
In addition to being dug for lime, the Chalk With time, deterioration of the capping may lead to
also has been worked as a source of flint. The earl- collapse, exposing the shaft. Such an event,
iest flint mines in England date back to Neolithic although rare, can have serious consequences.
times. The presence of such cavities, galleries and Indeed, lives have been lost when individuals have
chambers at shallow depth beneath the ground, fallen into shafts that opened beneath them. For
most of which are unrecorded, represents a hazard instance, on 22 November 1967, a woman and
to surface development, as well as to existing child fell into a shaft, with a surface diameter of
property and, indeed, lives have been lost, owing 5.5 m, which opened up in West Street, Frindsbury,
to roof collapse into the workings. The timing of near Rochester, Kent (Charman and Cooper 1987).
collapse generally is difficult or impossible to The shaft was 38.7 m deep with water occurring at
predict. However, factors that influence the possi- 23.8 m. The bodies were not recovered.
bility of roof collapse, together with the migration An additional factor that must be considered
of the resultant void to the surface, include the when dealing with subsidences of workings is that
width of the unsupported roof span, the character the Chalk, as mentioned above, is subject to dis-
of the Chalk and any other cover materials, par- solution. In fact, many chalk subsidence incidents
ticularly their shear strengths and the incidence have been man induced where flows of water into
and geometry of any discontinuities, the dip of the the ground have been concentrated into small
beds, the height of the workings, the depth of areas by soak-aways, leaking drains and water
overburden and the groundwater conditions. supply pipes, and even over-zealous garden water-
A review of the geological conditions that may ing can have similar effects (Figure 26.7). Such
aid the collapse of mine workings in chalk has concentrated flows accelerate dissolution, as well
been provided by Smith and Rosenbaum (1993). as reducing the strength of chalk. Urbanisation
Where weak superficial deposits occur above can make a significant contribution in causing
workings, they may ‘flow’ into voids that have subsidence in that it radically changes the way in
reached rockhead, thereby giving rise to surface which water enters the ground compared with an
features, which may vary from gentle dish-shaped open, undeveloped site. By rendering 50 to 80%
GFE-26.qxd 6/27/05 4:47 PM Page 738
Figure 26.7 Subsidence in the Chalk at Jacqueline Close, Bury St Edmunds, Suffolk, due to dissolution and thinning
of the mine roof below leading to its collapse (photograph courtesy of the East Anglian Daily Times).
of the surface plan area relatively impermeable, which the walls between the chambers were
infiltration and percolation of precipitation occurs partially removed to create pillars (Figure 26.8).
only in the remaining areas of open ground. Other forms of workings in the Chalk include
Furthermore, much of the water draining from chalkwells, bell pits, chalkangles and pillared
built-over areas is concentrated into soak-aways. workings (Edmonds et al., 1987). Pillared work-
Backfilled service trenches also tend to act as ings in the Chalk, as noted above, are the most
preferred paths for water ingress. recent and generally date from the beginning of
Mine workings in the Chalk in south-east the nineteenth century. The pattern of the work-
England are referred to as ‘deneholes’. Some of ings varies, in particular the size of the pillars.
the deneholes have ancient origins in that they are The joint pattern and the presence of any solution
believed to have been excavated during pre- pipes tended to influence the layout of the mines.
Roman times. Others are presumed to be of The stalls were typically 2 to 3 m wide and about
Roman origin whilst the remainder could have 9 m in height and were worked by benching.
been excavated as late as the fourteenth century Subsidences resulting from the collapse of
(Bell et al., 1992). Commonly, a circular shaft, mines in the Chalk occasionally occur, one of the
10 to 20 m or so in depth, leads to two or more most notable being at Jacqueline Close, Bury
chambers. The most commonly occurring form is St Edmunds, Suffolk, in 1966 (Figure 26.7). More
the six chambered or double trefoil type of dene- recently, chalk mine workings have collapsed in
hole. Subsequent improvement in mining practice Chislehurst, south-east London (1988 but with a
led to the development of the pillared denehole in recorded history of collapse going back to 1858),
GFE-26.qxd 6/27/05 4:47 PM Page 739
Figure 26.8 Linked deneholes at Hangman’s Wood, Grays, Essex (after Le Gear, 1979).
Reading, Berkshire (2000) and Blackheath, in average rate of retreat may be between 0.05 and
south-east London (2002). 0.10 m per year. Rockfalls initiated by weathering
processes are the main agent of degradation
Erosion and mass movement (Figure 26.10). Nevertheless, landslides in the
The silty clay soils that are associated with areas form of toppling, planar and wedge failures occur
of chalklands may be prone to soil erosion by along the cliffs, the types of slide being con-
water or wind when the ground lies fallow. The trolled by the discontinuity pattern, and often are
impacts of soil erosion are confined primarily to triggered by heavy rainfall or freeze–thaw action.
the degradation of agricultural land. Nonetheless, The cliffs also are undermined by coastal erosion
roads and ditches at times have to be cleaned, that ultimately leads to a near-vertical collapse,
some siltation may occur in waterways and stand- discontinuities again affecting the character of
ing bodies of water, and surface water quality may the failure.
be adversely affected (Boardman, 1990). Yet another cause of landslides in some parts
The Chalk gives rise to some spectacular cliffs of the Chalk has been attributed to groundwater
where it meets the coast (Figure 26.9). Many of trapped in stress relief fissures behind and sub-
these are near vertical and may exceed 150 m in parallel to the cliff face. This stress causes shear-
height. The cliffs are undergoing erosion and the ing in the lower strength, less permeable marly
GFE-26.qxd 6/27/05 4:47 PM Page 740
Figure 26.9 Steep chalk cliffs north of Flamborough Head, Yorkshire (photograph: F. G. Bell copyright).
horizons, especially when they occur near the occur on the Chalk escarpment immediately north
base of the cliff. A notable example of a landslide- of the entrance to the Channel Tunnel and the
affected area is Folkestone Warren, which is a nat- possibility of their reactivation, particularly by
urally unstable stretch of coastline, 3.2 km long, construction operations, had to be investigated
between Folkestone and Dover, in south-east (Birch and Warren, 1996) (Figure 26.11). However,
England. A succession of slides affected the rail- even engineered slopes in sub-horizontally bedded
way, located at the foot of the cliff, after it was chalk are rarely subject to major instability and cut-
built in the mid-nineteenth century but the most tings with 45 ⬚ slopes in rubbly, blocky chalk rarely
notable occurred in 1915 when a series of rota- exhibit spalling (that is, the detachment usually
tional slides and chalk falls from the cliffs of individual blocks commonly by freeze–thaw
blocked the line. One of these slides was trans- action). Nonetheless, Phipps and McGinnity (2001)
formed to a flow and had a runout to sea of some concluded that a slope angle of 53 ⬚ is the steepest
400 m. Fortunately, a train was stopped by a sig- that can be maintained in unweathered chalk before
nalman before it entered the section of the line degradation processes present a risk of causing dis-
affected by the slide but subsequent movements turbance in cuttings.
derailed the train. It took five years to rectify the As would be expected, landslides are more
damage. According to Hutchinson et al. (1980) common in the superficial deposits. Phipps and
the main causes of the movements were intense McGinnity (2001) noted that, especially where
marine erosion along the toe of previously slid slopes are steepened in superficial deposits for
material and the development of high porewater example by cuttings, debris flows, washouts and
pressures along potential sliding surfaces. slumps may take place. They claimed that debris
Occasionally, landslides occur on chalk scarp- flows occur in deep heterogeneous superficial
lands. For example, several old landslide areas deposits that contain inter-layered clay soils and
GFE-26.qxd 6/27/05 4:47 PM Page 741
Figure 26.10 Rockfall at Beachy Head, East Sussex, England. Left the Devil’s Chimney in February 1999; right,
the Devil’s Chimney in August 2001 following the rockfall of 4 April 2001 (photographs courtesy of Peter Hobbs,
British Geological Survey, NERC copyright).
gravels, and that they are triggered by extreme often are due to a high water table caused by
rainfall events. Indeed, Phipps and McGinnity either intense storms or rapid snowmelt associ-
further maintained that collapse of the soil fabric ated with frozen ground, both generating rapid
on saturation of these materials might lead to runoff. An example of the first type of flash
their liquefaction. Similarly, washouts are initi- flood occurred at Louth on the margins of the
ated by heavy rainfall affecting clayey sands and Lincolnshire Wolds in 1922 when up to 153 mm
silts. Initially, slump failure occurs as a result of of rain fell in the neighbouring small catchment
rotational or sub-planar movements in clay-rich on the Chalk. This gave rise to torrents in valleys
deposits and may be followed by flow. that normally were dry and led to the level of the
River Ludd rising by 5 m in 15 minutes, the flood
Springs wave moving through Louth at approximately
A spring line occurs at the foot of the scarp slope 140 m3/s. Twenty-two people were drowned and
where the Chalk rests upon less permeable ground 1250 made homeless, with buildings being
such as the Gault Clay, from which contact springs demolished. The second type of example is pro-
issue. Similarly, where the dip slope of the Chalk vided by the Great Till flood that occurred in a
sinks beneath less permeable sediments, then chalk catchment on the downlands of Salisbury
another spring line occurs from which overflow Plain. Melting snow and frozen ground meant
springs may emerge. that the River Till burst its banks at Shrewton
and rose 2.5 m above its normal level. In this
Flooding instance, three people were drowned, 72 houses
Flash floods have been associated with stream destroyed and 200 individuals made homeless
catchments located on the Chalk. Such floods (Cross, 1967).
GFE-26.qxd 6/27/05 4:47 PM Page 742
Figure 26.11 Landslides on the Chalk escarpment immediately north of the Channel Tunnel (after Birch and
Griffiths, 1996).
Many of the features referred to above are Chalk is governed by its discontinuity pattern
illustrated in the idealized model of the chalk- rather than by inter-granular flow. Values of pri-
lands given in Figure 26.12. mary permeability range from roughly 1 ⫻ 10⫺11
to 1 ⫻ 10⫺8 m/s (Ineson, 1962; Bell et al., 1999).
Fissure permeability can be as much as 1 ⫻ 10⫺5
26.3 Chalk engineering m/s, which accounts, in part, for why the Chalk is
the most important aquifer in Britain, supplying
The Chalk has a range of distinctive engineering approximately 40% of the groundwater abstracted
properties (Table 26.2), including significant vari- (MacDonald and Allen, 2001).
ations in density (Higginbottom, 1965; Bell, 1977; A number of Chalk classifications have been
Bell et al., 1999, Clayton, 1983) and hardness developed (e.g. Ward et al., 1968, see Table 26.3;
(Mortimore and Fielding, 1990; Lord et al., 1994, Wakeling, 1970; Spink and Norbury, 1990, see
2002). Generally, the Chalk is regarded as moder- Table 26.4); the most recent (Lord et al., 1994) is
ately strong, but occasionally is moderately weak based on the hardness, or more specifically dry
to moderately strong (Anon., 1977). The uncon- density on the one hand and the nature of the
fined compressive strength of chalk undergoes a discontinuities on the other (Figure 26.13). Chalk
marked reduction when it is saturated (Table 26.2). discontinuities are an important control on the
For instance, according to Bell et al. (1999) some mass hydrogeological and mechanical behaviour,
samples of Upper Chalk from Kent suffer a dra- ranging from microscopic grain boundaries and
matic loss on saturation amounting to almost 70%. microfractures to major tectonic joint and fault
Samples from the Lower and Middle Chalk may structures (Figure 26.14; Table 26.5).
show a reduction in strength averaging over 50%. During cold weather chalk may suffer frost
In fact, the weaker the rock, the greater the loss of heave, ice lenses up to 25 mm in thickness being
strength appears to be on saturation. Porosities developed along bedding planes (Lewis and
vary between less than 25% to greater than 40%. Croney, 1965). Higginbottom (1965) suggested
Saturation moisture content has been used as an that a probable volume increase of some 20 to
indicator of hardness with a value of 15% being 30% of the original thickness of the ground might
regarded as hard whilst those chalks for which it ultimately result. The presence of permafrost
exceeds 25% are soft. The permeability of the in the surface layers of chalk during the late
GFE-26.qxd 6/27/05 4:47 PM Page 743
Figure 26.12 Conceptual block model of chalk landscape in the UK.
Pleistocene meant that it was subject to frost structures. For instance, in a review of founda-
heave. In fact, most chalk possesses a surface tions on chalk, Lord (1990) maintained that
zone of very closely jointed material in which shallow footings could be founded successfully
fractures tend to be crazed, with curvi-planar sur- in grades IV and V chalk, according to the
faces typical of frost shattering. Frost shattering Mundford classification, at bearing pressures of
has been found to extend to depths of 30 m and 250 and 300 kPa. In 1994 Lord et al. noted that
may be concentrated along certain preferred the ultimate bearing capacity of chalk could range
planes (Figure 26.15). Frost churning has led to between 1.5 MPa for low-density chalk to in
the obliteration of the macrostructure in the upper excess of 8 MPa for chalk of high density. They
metre or so of the Chalk in many places and went on to point out that even with the lowest
to the formation of a mass of pasty remoulded ultimate bearing capacity recorded and a factor of
chalk (putty chalk) enclosing angular fragments safety of 3, the minimum safe bearing capacity
of unaltered material, which increase in size with would be 500 kPa and that loads of this amount
depth. are rarely imposed by shallow foundation struc-
Chalk can present a number of problems to the tures. Consequently, the design of shallow foun-
engineer because of the presence of discontinu- dations in chalk is governed by settlement.
ities; because of the reduction in mechanical What is more, Burland et al. (1975) found that
properties such as density, strength and deform- settlements of a five-storey building constructed
ation modulus by weathering; because of the pres- on soft low-grade chalk at Reading were very
ence of dissolution features; because of the effects small, as did the results of an investigation carried
of periglaciation that has given rise, for example, out at Basingstoke by Kee et al. (1975). Previously,
to frost shattering, and because of past mining. Wakeling (1968) had indicated that the settlement
of the piers carrying the bridge over the river
Foundations Medway, Kent, amounted to some 28 and 35 mm
Fresh chalk generally has proved a satisfactory for a mean effective foundation pressure of around
foundation material for many buildings and 650 kPa.
GFE-26.qxd
Table 26.2 Some physical properties of chalk from Yorkshire, Norfolk and Kent (after Bell et al., 1999).
744
Yorkshire* Norfolk** Kent***
6/27/05
Lower Middle Upper Lower Melbourn Rock Middle Upper Upper
Specific gravity
Maximum 2.73 2.71 2.72 2.71 2.72 2.74 2.72 2.72
4:47 PM
Minimum 2.65 2.69 2.67 2.66 2.68 2.68 2.70 2.65
Mean 2.71 2.70 2.70 2.68 2.70 2.70 2.71 2.69
Dry density (Mg/m3)
Maximum 2.13 (L) 2.30 (M) 2.23 (M) 2.17 (L) 2.23 (M) 1.81 (L) 1.70 (VL) 1.61 (VL)
Page 744
Minimum 1.85 (L) 1.76 (VL) 1.77 (VL) 1.71 (VL) 2.04 (L) 1.62 (VL) 1.54 (VL) 1.35 (VL)
Mean 2.08 (L) 2.14 (L) 2.06 (L) 1.99 (L) 2.17 (L) 1.76 (VL) 1.61 (VL) 1.44 (VL)
Saturated density (Mg/m3)
Maximum 2.34 2.43 2.41 2.38 2.44 2.16 2.04 1.97
Minimum 2.16 2.11 2.05 2.06 2.27 2.01 1.93 1.82
Mean 2.31 2.34 2.29 2.24 2.34 2.14 2.00 1.91
Saturation moisture content (%)
Maximum 17.2 19.9 22.9 21.5 11.4 24.8 28.0 37.2
Minimum 10.8 7.2 7.9 9.2 7.5 18.3 21.9 25.2
Mean 11.2 10.2 11.6 13.4 9.5 19.9 25.2 32.6
Effective porosity (%)
(saturation method)
Maximum 30.2 (VH) 35.0 (VH) 36.4 (VH) 34.4 (VH) 27.0 (H) 38.2 (VH) 43.2 (VH) 45.7 (VH)
Minimum 17.2 (H) 16.2 (H) 17.7 (H) 19.9 (H) 16.1 (H) 30.2 (VH) 34.3 (VH) 29.6 (H)
Mean 20.6 (H) 21.8 (H) 23.9 (H) 26.5 (H) 19.8 (H) 34.4 (VH) 39.9 (VH) 41.7 (VH)
Absolute porosity (%)
Maximum 31.9 35.1 38.3 36.9 33.7 40.2 46.2 52.4
Minimum 20.5 18.2 17.7 19.9 18.8 33.2 37.3 38.7
Mean 26.8 24.8 24.4 26.6 29.2 35.1 40.6 47.0
Permeability (⫻10⫺9 m/s)
Maximum 1.2 2.2 37.0
Minimum 0.3 0.5 13.9
Mean 0.9 1.4 27.7
GFE-26.qxd
Dry unconfined compressive

strength (MPa)
6/27/05
Maximum 32.7 (MS) 36.4 (MS) 34.0 (MS) 30.5 (MS) 38.3 (MS) 25.1 (MS) 12.7 (MS) 6.2 (MW)
Minimum 19.1 (MS) 25.2 (MS) 18.1 14.2 22.1 (MS) 7.4 (MW) 6.9 (MW) 4.8 (MW)
Mean 26.4 (MS) 30.7 (MS) 25.6 (MS) 21.0 (MS) 29.1 (MS) 13.0 (MS) 9.5 (MW) 5.5 (MW)
Saturated unconfined compressive
4:47 PM
strength (MPa)
Maximum 16.2 20.4 15.9 13.7 17.5 10.3 5.1 2.2
Minimum 8.6 11.7 7.4 6.2 8.9 3.1 2.8 1.4
Mean 13.7 16.8 11.9 10.7 14.3 5.8 3.6 1.7
Page 745
Young’s modulus (Et50 GPa)
(tangent modulus at 50% failure)
Maximum 18.4 21.7 17.1 14.1 18.9 10.4 8.2 4.6
Minimum 7.5 9.1 7.4 6.9 7.3 5.0 4.1 4.2
Mean 12.7 15.2 11.7 8.7 13.5 8.4 6.7 4.4
Dry density VL ⫽ very low, less than 1.8 Mg/m3; L ⫽ low, 1.8 to 2.2 Mg/m3; M ⫽ moderate, over 2.2 Mg/m3.
Porosity H ⫽ high, 15 to 30%; VH ⫽ very high, over 30% (Anon., 1979).
*Yorkshire Lower — H. subglobosus (? ⫽ S. gracile) zone near Speeton; Middle — T. lata zone, Thornwick Bay; Upper — M. coranguinum zone,
Selwicks Bay.
**Norfolk Lower — S. varians (? ⫽ M. mantelli) zone, Hunstanton; Melbourn Rock and Middle — T. lata zone, Hillington;
Upper — M. coranguinum zone, Burnham Market.
***Kent Upper — M. coranguinum zone, Northfleet.
Unconfined
compressive strength: W, weak, 1.25–5 MPa; MW, moderately weak, 5–12.5 MPa; MS, moderately strong, 12.5–50 MPa (Anon., 1977).
CHALK LANDSCAPES
745
GFE-26.qxd
Table 26.3 Correlation between grades and the mechanical properties of Middle Chalk at Mundford (after Ward et al., 1968).
746
Grade Description Approx. range Approx. value Range of Bearing Creep SPT N value Rock mass
6/27/05
of Young’s of dynamic compression pressure properties (after Wakeling, factor (after
modulus Young’s wave velocities causing 1970)** Burland and
E (MPa) modulus (km/s) (after yield (kPa) Lord, 1970)
Edyn (MPa) Grainger
4:47 PM
(after Abbiss, et al., 1973)
1979)
V Structureless melange. Unweathered Below 500 Below 500 0.65–0.75 Below 200 Exhibits Below 15 0.1
and partly weathered angular chalk significant
Page 746
blocks and fragments set in a matrix creep
of deeply weathered remoulded chalk.
Bedding and jointing are absent.
IV Friable to rubbly chalk. Unweathered 500–1000 800 1.0–1.2 200–400 Exhibits 15–20 0.1–0.2
or partially weathered chalk with significant
bedding and jointing present. Joints creep
and small fractures closely spaced,
ranging from 10–60 mm apart.
III Rubbly to blocky chalk. Unweathered 1000–2000 4000 1.6–1.8 400–600 For pressures 20–25 0.2–0.4
medium to hard chalk with joints not exceeding
60–200 mm apart. Joints open up to 400 kPa creep
8 mm sometimes with secondary is small and
staining and fragmentary infillings. terminates in
a few months
II Medium hard chalk with widely 2000–5000 7000 2.2–2.3 600–1000 Negligible creep 25–35 0.6–0.8
spaced, closed joints. Joints more for pressure
than 200 mm apart. Fractures of at least
irregularly when excavated, does 400 kPa
not break along joints. Unweathered.
I Hard, brittle chalk with widely Over 5000 Over 10 000 Over 2.3 Over 1000 Negligible creep Over 35 Over 0.8
spaced, closed joints. Unweathered. for pressure
of at least
400 kPa
* Ward et al. (1968) emphasised that their classification was specifically developed for the site at Mundford and hence its application elsewhere should be made
with caution.
** The correlation between SPT (Standard Penetration Test) N value and grade may be different in the Upper Chalk (see Dennehy, 1976).
GFE-26.qxd
Table 26.4 The description and grading of Middle and Upper Chalk for engineering purposes (after Spink and Norbury, 1990).
Grade1 Structure Colloquial Definitions of grade Typical features of grade Word order for descriptions
description
6/27/05
of grade2 % comminuted % coarser Weathering Strength of
chalk matrix3 fragments3 of coarser coarser
fragments8 fragments9
4:47 PM
VI Structureless Putty chalk Greater than Less than Moderately, Very weak Structureless10 CHALK composed
chalk, bedding with small about 35% about 65% highly or or weak of: soil strength of matrix;11
and jointing lumps completely colour of matrix;12 nature
absent weathered of matrix material; amount of
Page 747
fragments; presence and nature
of flints; other features (Grade VI)
V Chalk lumps Less than Greater than Moderately Very weak Structureless10 CHALK composed of:
in comminuted about 35% about 65% or highly or weak angularity and size of fragments:
matrix weathered colour of fragments: weathering of
fragments;14 strength of fragments;15
amount of matrix;13 nature of matrix
material; presence and nature of
flints; other features (Grade V)
Fracture Fracture Material Material Colour: rock material weathering;14
spacing (mm)5,6 width (mm)5,6 weathering8 strength9 CHALK:
IV Rubbly chalk Extremely closely Open or Moderately Very weak rock material strength;15 discontinuity
to very closely infilled ⬎ 57 or highly or weak type;16 discontinuity spacing;17
spaced ⬍ 60 7
weathered discontinuity width and nature
III In situ structured Rubbly to Closely spaced Open or Slightly or Weak or of infill if appropriate; discontinuity
chalk, with blocky chalk 60–200 infilled ⬍ 3 moderately moderately orientations (in situ observations
bedding and weathered weak only);18
jointing
CHALK LANDSCAPES
II Blocky chalk At least medium Tight and Fresh or Weak or presence and nature of flints; other
spaced ⬎ 200 clean slightly moderately features (Grade)
weathered weak
I Brittle and At least medium Tight and Fresh or Moderately
massive spaced ⬎ 200 clean slightly weak or
chalk weathered moderately
strong
747
GFE-26.qxd
Notes:
748
1. Intermediate grades may be assigned to material that is intermediate 10. ‘Structureless’ may be replaced with ‘Reworked’ if evidence of reworking
between two grades e.g. III/IV. However V/IV or IV/V is not is present, such as included foreign gravel, or if the chalk is present
6/27/05
admissible. within an alluvial or glacial sequence.
2. Colloquial descriptive terms are undefined and should not be 11. Soil strengths may be assessed where the matrix is dominant, but only
used in description. for in situ observations.
3. The cohesive matrix is comminuted fragments of chalk typically 12. Although Upper and Middle Chalk are white the matrix of Grades VI and
4:47 PM
of silt and sand size. The coarser fragments of chalk are typically V is often cream or light brown.
gravel size or larger. 13. The proportion of secondary matrix or coarser fragments is quantified
4. The 35/65% cut-off between Grades V and VI is approximate and by the terms ‘little’, ‘some’, ‘much’ as defined in Table 1 of Spink and
dependent on assessed engineering behaviour. Grade VI is cohesive Norbury (1990).
Page 748
matrix dominated, Grade V is granular dominated by the coarser 14. The weathering of the chalk material or coarser fragments is described
fragments of chalk. in accordance with Anon. (1972) The preparation of maps and plans in terms
5. The grading is determined on the assessed average fracture spacing of engineering geology. Quarterly Journal Engineering Geology
(including both bedding fractures and joints) and the assessed 5, 293–382.
typical fracture width or aperture. 15. The strength of the chalk material or coarser fragments is described in
6. Other combinations of fracture spacing and fracture width are accordance with BS 5930 (1999) Code of Practice on Site
possible. Any undefined gradings have to be assessed with care, Investigations. British Standards Institution, London.
noted and discussed. 16. Discontinuity types are ‘bedding fractures’ when sub-horizontal, joints when
7. The original ‘Mundford’ definition of Grade IV is fractures 10 to non-horizontal, or ‘fractures’ to include both bedding fractures and joints.
60 mm spacing, open or infilled up to 20 mm. 17. The one dimensional fracture spacing scheme of BS 5930 (1999) is used.
8. Within each grade the degree of weathering varies, typical ranges This should be supplemented by assessed minimum/average/maximum frac-
are given. Original ‘Mundford’ terms are undefined. ture spacing in millimetres for core and in situ observations.
9. Within each grade strengths vary, typical ranges are given. Original 18. Discontinuity set orientations are given as dip direction/dip e.g. 275 ⬚/10 ⬚.
‘Mundford’ terms are undefined. If the strength is exceptionally high Ranges may be indicated e.g. 260⬚–290 ⬚/05 ⬚–15 ⬚. Major discontinuities
or low compared with the typical range, and considered to alter the are logged individually.
engineering behaviour, then the chalk may be upgraded or
downgraded respectively, but this should be noted and discussed.
GFE-26.qxd 6/27/05 4:47 PM Page 749
Figure 26.13 Engineering geological classification of chalk (redrawn from Lord et al., 1994).
Examples: White, slightly weathered, CHALK, very weak, low density. Bedding fractures very closely spaced
(20–60 mm), open up to 20 mm, infilled with powdery chalk; discontinuities widely spaced (⬎600 mm), open
⬎3 mm: Low density Grade C4.
White, moderately to highly weathered, CHALK, very weak. Structureless mélange composed of 70% sub-angular to
rounded medium to coarse gravel size fragments set in 30% soft, light brown, sandy silt size chalk matrix: Grade Dc.
Lord (1990) concluded that the most effective founded in the Upper Chalk at Bury St Edmunds,
method of estimating settlement was to carry out Suffolk, increased rapidly once the load exceeded
a series of 600 or 900 mm diameter plate load the yield stress. Hence, if settlement is to be lim-
tests at 2.5 to 5.0 m intervals vertically into the ited, then the stress applied by a foundation struc-
Chalk. Subsequently, after an extensive study of ture should not exceed the yield stress of the
data from plate load tests (Figure 26.16), Lord Chalk. Lord et al. (1994) suggested allowable
et al. (1994) concluded that the performance of bearing pressures for medium/high density, low
chalk in shallow foundations could be related to density, and clast dominated structureless chalk
the grades adopted by the CIRIA classification as of 300 kPa, 240 kPa and 225 kPa respectively.
shown in Figure 26.13. Plate load tests under- Allowable bearing pressure on matrix dominated
taken at Mundford by Burland and Lord (1970) structureless chalk should be assessed independ-
showed that at low applied pressures even grade ently during a site investigation.
IV chalk behaved elastically. They pointed out The use of pile foundations in chalk has been
that plate load tests for stresses up to 1.0 MPa discussed in detail by Lord et al. (1994). The
demonstrated that grades IV and V chalk exhib- choice of pile and its diameter depends on the
ited significant creep, and that long-term deflec- structure it has to support on the one hand and
tions may be appreciably larger than immediate the ground conditions on the other. In the case of
deflections. Creep in grade III chalk is smaller chalk, the latter includes its dry density, nature of
and terminates more rapidly whilst grades II and the discontinuities, the presence of solution fea-
I undergo negligible creep. tures and the position of the water table. If the blow
At higher pressures chalk exhibits yielding count obtained from the standard penetration test is
behaviour. Burland and Bayliss (1990) observed used to derive the ultimate base stress of piles in
that the rate of settlement beneath four silos chalk, then Lord et al. suggested a range of 2000 to
GFE-26.qxd 6/27/05 4:47 PM Page 750
Figure 26.14 Joint orientation data in the Chalk of east Yorkshire (redrawn from Bell et al., 1999).
3000 kPa for low density chalk, and from 4000 to the submerged density of the chalk leading to
5000 kPa in medium/high density chalk. disturbance of the floor due to hydrostatic uplift. In
Water can cause problems in engineering in such an instance, the groundwater pressure needs to
chalk. For instance, a problem may occur in an be relieved by the installation of wells. In addition,
excavation made in chalk that is well jointed and water entering the ground from soak-aways, leaking
overlain by structureless chalk if the water table sewers or leaking water supply pipes may give rise
occurs in the latter. In other words, as excavation to accelerated dissolution of chalk, especially in
proceeds, then the hydrostatic head may exceed relation to solution features or old mine workings,
GFE-26.qxd 6/27/05 4:47 PM Page 751
Table 26.5 Discontinuity data and RQD (Rock Quality Designation) values derived from scanlines.
Scanline Discontinuity data
Location Length (m) Orientation Number Mean spacing (mm) RQD (%)
Selswick Bay — north cliff 9.95 N 106⬚ E 29 329 95.2

Selswick Bay — wave-cut platform 4.00 N 70⬚ E 20 184 76.3
Selswick Bay — wave-cut platform 4.20 N 58⬚ E 27 147 79.6
High Stacks — horizontal line 9.00 N 135⬚ E 29 318 97.4
High Stacks — horizontal line 4.00 N 458 E 19 203 92.3
thereby leading to subsidence. Introduction of water particularly satisfactory with hard chalk since it
into the ground or a fluctuating water table can dis- usually requires some degree of crushing during
turb metastable deposits occupying sinkholes and compaction. The softer varieties are usually bro-
pipes, and so cause these deposits to collapse. ken down on excavation and can be compacted
When solution features can be located and to a low air content by most types of plant.
occur at shallow depth, then, if occupied, the mate- However, a danger often exists with soft chalk in
rial can be removed and replaced by concrete or that it may be over-compacted, thereby giving
suitably compacted fill; they can be similarly filled rise to high porewater pressures. Unfortunately,
if unoccupied. A raft may be used as a bridging the fine material forms a slurry or putty chalk at
structure to support a building if a void or disturbed higher moisture contents. Hence, fill material
ground is suspected to occur beneath deposits over- will be weak and unstable (that is, ruts are pro-
lying chalk. Piles frequently are used in such situa- duced by construction traffic) if the proportion
tions, being founded below the suspect ground. of putty chalk controls the behaviour of the fill,
However, their installation may adversely affect so that the stability of freshly placed chalk fill
such suspect ground. Grout can be injected to fill depends on both its fines content and moisture
voids that have been located or to stabilise loosely content.
packed deposits in sinkholes. Known abandoned Ingoldby and Parsons (1977) described the
mines in chalk may be filled with bulk grouts, the apparatus developed by the UK Transport and
processes involved being similar to those used to Road Research Laboratory to determine the
stabilise abandoned pillar and stall workings in old moisture condition value and chalk crushing
coal mines (Healy and Head, 1984). strength value, both of which parameters are used
in a classification to avoid or minimise instabil-
Earthworks ity of chalk used as fill. The breakdown of blocks
According to Rat and Schaeffner (1990) the of chalk within fill can lead to settlement that can
behaviour of chalk in earthworks depends on the continue for several years after construction
production of a matrix of disturbed chalk due to it operations have ceased.
being crushed by construction equipment; on the
possibility of blocks of chalk in fill being reduced Tunnelling
in size by static or dynamic stresses, and on the The Chalk is a formation in which a number of
susceptibility of chalk to frost activity. The exca- tunnels have been constructed. Generally, the
vation and compaction processes during earth- material is easily excavated by either hydraulic
works partially break down chalk into a mixture excavators or tunnel boring machines (TBMs) but
of lumps and fines. is strong enough to stand unsupported in the short
A wide range of equipment can be used quite term. However, flints in the Chalk can cause dif-
successfully to compact chalk fill. Heavy dead- ficulties for tunnelling machines, especially those
weight rollers and pneumatic-tyred rollers are of small diameter. They can produce high rates of
GFE-26.qxd 6/27/05 4:47 PM Page 752
Figure 26.15 Upper surface of the Chalk showing a high degree of fracturing due to periglacial action and solution
pipes, near Wells, Norfolk (UK) (photograph: F. G. Bell copyright).
GFE-26.qxd 6/27/05 4:47 PM Page 753
However, high water inflows were difficult to

cope with in open mode and the disintegration of
the excavated material into a slurry made control
of muck disposal via the screw conveyor difficult
in closed mode.
Frost susceptibility
Frost activity can lead to the development of seg-
regated layers of ice forming near the ground sur-
face in chalk that grow by drawing groundwater
from below. This gives rise to frost heave adversely
affecting embankments and roads, which may also
be affected by subsequent thaw settlement. Frost-
susceptible soils can be dealt with by adequate
Figure 26.16 Plate load test data for chalk (redrawn drainage, by replacement with gravels, by the
from Lord et al., 1994). addition of certain chemicals to the soil that
reduce its capacity for water absorption or by the
addition of a cement binder. D’Hem et al. (1990)
wear in cutting tools and larger pieces tend to clog described the use of cement, lime, pulverised fly
up the mucking system. ash and wet sand to enhance the engineering
A notable example is the Channel Tunnel, performance of chalk used in roadworks.
which follows the chalk syncline between
England and France. For most of its length it is in
chalk marl, the clay content being beneficial in 26.4 Summary and conclusions
reducing the permeability of the ground and
reducing water inflow. At either end, but particu- Chalk is a usually soft, fine textured, often white or
larly near France, fault zones in harder chalk grey limestone of marine origin that consists pri-
introduced problems related to high water inflow marily of calcite that was deposited during Upper
(Barthes et al., 1994). The Channel Tunnel was Cretaceous and early Tertiary times. It is a remark-
constructed using TBMs and segmental lining. ably pure limestone. For instance, excluding flints
The North Downs Tunnel, on the first stage and marl bands, the calcium carbonate content of
of the rail link between the Channel Tunnel at the English white chalk facies usually exceeds
Folkestone and London, was above the water 98%. Chalk produces a distinctive ‘rolling’ coun-
table and was constructed by open face excav- tryside. Its outcrop normally forms a prominent
ation using road-header machines and provided escarpment with hard bands within the Chalk
with a primary lining using ‘shotcrete’. Progress forming intermediate escarpments. The escarp-
was very rapid, the only problems being related ment and dip slopes represent the surviving rem-
to sinkholes in the Chalk reaching the tunnel nants of a sub-Palaeogene erosion surface that has
crown at either end where the Chalk cover was been modified by late Tertiary and Quaternary
relatively thin. geomorphological processes. These processes
A much smaller tunnel constructed for the have produced a series of geomorphological fea-
Barking Reach Power Station cooling water outlet tures that characterise chalk landscapes. These
to the River Thames, east of London, was also include dry valleys, bournes, coombes, sinkholes,
constructed using a TBM but considerable prob- solution pipes and minor cave systems. The Chalk
lems were encountered in both shaft sinking has also been excavated for lime and flints in some
and tunnelling (Evans and Hodgkins, 1997). The places, by means of surface quarries and pits and
TBM was able to work in either open or closed associated subsurface mining. Mining has left
mode (as an earth pressure balance machine). a legacy of entrances, galleries and chambers
GFE-26.qxd 6/27/05 4:47 PM Page 754
(known as deneholes) that are prone to collapse. Birch, G. P. and Warren, C. D. (1996) The cliffs behind
Chalk slopes can be prone to surface erosion, land- the Channel Tunnel workings. In Harris, C. S.,
Hart, M. B., Varley, P. M. and Warren, C. D. (eds)
slides and, on steeper slopes, rockfalls. Flooding is
Engineering Geology of the Channel Tunnel.
not uncommon and can occur rapidly in response Thomas Telford Press, London, 76–87.
to intense rainfall because of the high permeability Boardman, J. (1990) Soil erosion on the South Downs:
and porosity of the Chalk. Chalk has a distinctive a review. In Boardman, J., Foster, I. D. L. and
range of engineering properties and can present Dearing, J. A. (eds) Soil Erosion on Agricultural
Land. Wiley, Chichester.
foundation and earthworks problems.
Bradshaw, J., Caiger, N., Halpin, M., Le Gear, R.,
Pearce, A., Pearman, H., Reeve, T. and Sowan, P.
(1991) Kent and East Sussex Underground. The
References Kent Underground Research Group, Meresborough
Books, Rainham, Kent.
Abbiss, C. P. (1979) A comparison of the stiffness of Bristow, C. R., Mortimore, R. N. and Wood, C. J.
the Chalk at Mundford from a seismic survey and (1997) Lithostratigraphy for mapping the Chalk of
large scale test. Geotechnique 29, 461–468. southern England. Proceedings of the Geologists’
Anon. (1972) The preparation of maps and plans Association 108, 293–315.
in terms of engineering geology. Geological Society BS 5930 (1999) Code of Practice for Site Investigations.
Engineering Group Working Party Report. Quarterly British Standards Institution, London.
Journal of Engineering Geology 5, 293–382. Burland, J. B. and Bayliss, F. V. S. (1990) Settlement
Anon. (1977) The description of rock masses for engin- and yielding of Upper Chalk supporting foundations
eering purposes. Geological Society Engineering for a silo complex. Proceedings of the International
Group Working Party Report. Quarterly Journal of Chalk Symposium, Brighton. Thomas Telford Press,
Engineering Geology 10, 355–388. London, 365–374.
Anon. (1979) Classification of rocks and soils for engin- Burland, J. B. and Lord, J. A. (1970) The load-
eering geological mapping. Part 1: rock and soil deformation behaviour of Middle Chalk at Mundford:
materials. Bulletin of the International Association a comparison between full-scale performance and in
of Engineering Geology 19, 364–371. situ and laboratory measurements. Proceedings
Atkinson, T. C. and Smith, D. I. (1974) Rapid ground- of a Conference on In Situ Investigations in Soils
water flow in fissures in chalk. Quarterly Journal of and Rocks. British Geotechnical Society, London,
Engineering Geology 7, 197–205. 3–15.
Banks, D., Davies, C. and Davies, W. (1995) The Chalk Burland, J. B., Kee, R. and Burford, D. (1975) Short term
as a karstic aquifer: evidence from a tracer test at settlement of a five storey building on soft chalk.
Stanford Dingley, Berkshire, UK. Quarterly Journal In Settlement of Structures. British Geotechnical
of Engineering Geology 28, Supp. 1, S31–S38. Society, Pentech Press, London, 259–265.
Barthes, H., Bordas, D., Bouillot, M., Buzon, M., Charman, J. H. and Cooper, C. G. (1987) The
Dumont, P., Fermin, J., Landry, J. C., Larive, J.-P., Frindsbury area, Rochester: a review of historical
Leblond, L., Morlot, J. J., Szypura, L., Vandebrouck, data and their implication on subsidence in an
P. and Vielliard, B. (1994) Tunnels — geology. urban area. In Culshaw, M. G., Bell, F. G., Cripps,
The Channel Tunnel, Part 3: French Section. J. C. and O’Hara, M. (eds) Planning and
Proceedings of the Institution of Civil Engineers, Engineering Geology, Engineering Geology Special
Civil Engineering 102, Special Issue 1, 6–10. Publication No 4. The Geological Society, London,
Bell, F. G. (1977) A note on the geotechnical properties 115–124.
of the Chalk. Engineering Geology 11, 27–225. Clayton, C. R. I. (1983) The influence of diagenesis
Bell, F. G., Culshaw, M. G., Moorlock, B. S. P. and on some index properties of chalk in England.
Cripps, J. C. (1992) Subsidence and ground move- Geotechnique 33, 225–241.
ments in chalk. Bulletin of the International Cross, D. A. E. (1967) The Great Till floods of 1841.
Association of Engineering Geology 45, 75–82. Weather 22, 430–433.
Bell, F. G., Culshaw, M. G. and Cripps, J. C. (1999) Culshaw, M. G. and Waltham, A. C. (1987) Natural and
A review of selected engineering geological charac- artificial cavities as ground engineering hazards.
teristics of English chalk. Engineering Geology 54, Quarterly Journal of Engineering Geology 20,
237–269. 139–150.
Birch, G. P. and Griffiths J. S. (1996) Engineering geo- D’Hem, P., Fevre, A., Hiernaux, R. and Holef, J. (1990)
morphology. In Harris, C. S., Hart, M. B., Varley, P. Stabilization of chalk in north-west France.
M. and Warren, C. D. (Eds) Engineering Geology of Proceedings of the International Chalk Sympo-
the Channel Tunnel. Thomas Telford Press, London, sium, Brighton. Thomas Telford Press, London,
64–75. 449–456.
GFE-26.qxd 6/27/05 4:47 PM Page 755
Dennehy, J. P. (1976) Correlation of the SPT value with Lord, J. A., Twine, D. and Yeow, H. (1994) Foundations
chalk grade for some zones of the Upper Chalk. in Chalk. Project Report 11, Construction Industry
Geotechnique 26, 610–614. Research and Information Association (CIRIA),
Edmonds, C. N., Green, C. P. and Higginbottom, I. E. London.
(1987) Subsidence hazard prediction for limestone Lord, J. A. Clayton, C. R. I. and Mortimore, R. N.
terrains, as applied to the English Cretaceous Chalk. (2002) Engineering in Chalk. Publication C574.
In Culshaw, M. G., Bell, F. G., Cripps, J. C. and Construction Industry Research and Information
O’Hara, M. (eds) Planning and Engineering Association (CIRIA), London.
Geology, Engineering Geology Special Publication Lowe, D. J. (1992) Chalk caves revisited. Cave Science
No 4. The Geological Society, London, 283–294. 19, 2, 55–58.
Evans, J. F. and Hodgkins, D. J. (1997) Barking Reach MacDonald, A. M. and Allen, D. J. (2001) Aquifer
Power Station cooling water system. Proceedings of properties of the Chalk of England. Quarterly
the Institution of Civil Engineers, Civil Engineering Journal of Engineering Geology and Hydrogeology
120, 15–26. 34, 371–384.
Evans, J. W. and Stubblefield, C. J. (1929) Handbook of Mortimore, R. N. (1983) The stratigraphy and sedimenta-
the Geology of Great Britain: a Compilative Work. tion of the Toronian-Campanian in the Southern
Thomas Murby and Co., London, 557pp. Province of England. Zitteliana 10, 27–41.
Garrels, R. M., Thompson, M. E. and Siever, R. (1960) Mortimore, R. N. (1990) Chalk or chalk? Proceedings
Stability of some carbonates at 25 ⬚ and one atmos- of the International Chalk Symposium, Brighton.
phere total pressure. American Journal of Science Thomas Telford Press, London, 15–45.
258, 402–418. Mortimore, R. N. and Fielding, P. M. (1990) The
Grainger, P., McCann, D. M. and Gallois, R. W. (1973) relationship between texture, density, and strength
The application of the seismic refraction technique of chalk. Proceedings of the International Chalk
to the study of the fracturing of the Middle Chalk at Symposium, Brighton. Thomas Telford Press, London,
Mundford, Norfolk. Geotechnique 23, 219–232. 109–132.
Healy, P. R. and Head, J. M. (1984) Construction over Mortimore, R. N., Wood, C. J. and Gallois, R. W. (2001)
Abandoned Mine Workings. Special Publication 23, British Upper Cretaceous Stratigraphy. Geological
Construction Industry Research and Information Conservation Review Series No. 23. Nature
Association, London. Conservancy Committee, Peterborough, 558pp.
Higginbottom I. E. (1965) The engineering geology of Phipps, P. J. and McGinnity, B. T. (2001) Classification
the Chalk. Proceedings of the Symposium on Chalk and stability assessment for chalk cuttings: the
in Earthworks. Institution of Civil Engineers, Metropolitan Line case study. Quarterly Journal of
London, 1–14. Engineering Geology and Hydrogeology 34,
Hutchinson, J. N., Bromhead, E. N. and Lupini, J. F. 353–370.
(1980) Additional observations on the Folkestone Pitman, J. I. (1983) Chemical weathering of the East
Warren landslides. Quarterly Journal Engineering Yorkshire chalk. In Paterson, K. and Sweeting, M. M.
Geology 13, 1–31. (eds) New Directions in Karst. Geo Books, Norwich,
Ineson, J. (1962) A hydrogeological study of the 77–113.
permeability of Chalk. Journal of the Institution of Rat, M. and Schaeffner, M. (1990) Classification of
Water Engineers 16, 43–63. chalks and conditions of use in embankments.
Ingoldby, H. C. and Parsons, A. W. (1977) The Proceedings of the International Chalk Sympo-
Classification of Chalk for use as a Fill Material. sium, Brighton. Thomas Telford Press, London,
Laboratory Report 806, Transportation and Road 425–428.
Research Laboratory, Crowthorne, Berkshire. Rawson, P. F., Allen, P. and Gale, A. (2001) The Chalk
Kee, R., Parker, A. S. and Wehale, J. E. C. (1975) Group — a revised lithostratigraphy. Geoscientist
Settlement of a twelve storey building on pile 11, 1, 21.
foundations in chalk at Basingstoke. In Settlement Smith, G. J. and Rosenbaum M. S. (1993) Abandoned
of Structures. British Geotechnical Society, Pentech shallow mine workings in chalk: a review of the
Press, London, 275–282. geological aspects leading to their destabilisation.
Le Gear, R. F. (1979) Deneholes, Part 2. Records of the Bulletin of the International Association of
Chelsea Speleological Society 10, 1–116. Engineering Geology 48, 101–108.
Lewis, W. A. and Croney, D. (1965) The properties of Spink, T. W. and Norbury, D. R. (1990) The engineer-
chalk in relation to road foundations and pavements. ing geological description of chalk. Proceedings of
Proceedings of the Symposium on Chalk in Earth- the International Chalk Symposium, Brighton.
works. Institution of Civil Engineers, London, 27–42. Thomas Telford Press, London, 153–159.
Lord, J. A. (1990). Foundations in chalk. Proceedings Wakeling, T. R. M. (1968) Foundations on chalk. In
of the International Chalk Symposium, Brighton. Chalk in Earthworks. Institution of Civil Engineers,
Thomas Telford Press, London, 301–326. London, 15–23.
GFE-26.qxd 6/27/05 4:47 PM Page 756
Wakeling, T. R. M. (1970) A comparison of the results Norfolk, for a large proton accelerator. Geotechnique
of standard site investigation methods against the 18, 399–431.
results of a detailed geotechnical investigation in West, G. and Dumbleton, A. J. (1972) Some observa-
Middle Chalk at Mundford, Norfolk. Proceedings of tions on swallow holes and mines in the Chalk.
a Conference on In Situ Investigations in Soils and Quarterly Journal of Engineering Geology 5,
Rocks. British Geotechnical Society, London, 17–22. 171–178.
Waltham, A. C. and Fookes, P. G. (2003) Engineering Weyl, P. K. (1959) The solution kinetics of calcite.
classification of karst ground conditions. Quarterly Journal of Geology 66, 163–176.
Journal of Engineering Geology and Hydrogeology
36, 101–118.
Waltham, A. C., Bell, F. G. and Culshaw, M. G. (2005)
Sinkholes and subsidence: karst and cavernous rocks Acknowledgements
in engineering and construction. Springer-Verlag,
Berlin and Praxis Publishing Ltd, Chichester. This paper is published with the permission of the
Ward, W. H., Burland, J. B. and Gallois, R. W. (1968) Director of the British Geological Survey, Natural
Geotechnical assessment of a site at Mundford, Environment Research Council (NERC).
GFE-27.qxd 6/27/05 2:26 PM Page 757
27. Urban Geomorphology

Ian Douglas
27.1 Introduction 5 to 15 per cent to the built-up area every year.

The spectacular economic growth of China has
The rate of urban growth is accelerating throughout led to even more rapid urban spread (Figure 27.1).
the world. In North America urban sprawl is seen This transformation of peri-urban areas is becom-
as one of the greatest threats to biodiversity, ing one of the most challenging environmental
conservation and sustainable development. The problems of the twenty-first century. Even in coun-
valley around Las Vegas, the fastest growing city tries where there is relatively little growth of the
in the U.S.A, is being developed commercially at total population, urban areas are expanding, alter-
the rate of about one square kilometre every four ing the way geomorphological processes work on
weeks. The suburban expansion of cities like landforms (Figure 27.2). Even though perhaps
Jakarta, Indonesia and Delhi in India is adding only 2% of the land surface of the Earth is built-up,
Figure 27.1 The urban growth of Beijing, China.
757
GFE-27.qxd 6/27/05 2:26 PM Page 758
Figure 27.2 The urban growth of greater Manchester, England.
because over half the world’s population is in America, northern Europe and northern Asia.
urban areas, and because so much capital is Knowledge of both present-day process condi-
invested in housing, commercial premises, public tions and legacies of past geomorphological activ-
buildings and infrastructure, urban gemorphology ity is needed in urban development planning. In
is highly relevant to people’s everyday lives and to older cities a further emerging need is under-
national and regional economies. standing of past urban and industrial activities,
including mining, quarrying and waste disposal,
as urban regeneration is accelerating the re-use of
27.2 Problems of siting and planning old industrial land and other ‘brownfield’ sites.
urban developments
Periglacial (see also Chapters 14, 24, 25)
Particular environments pose specific problems Developments in active periglacial areas face four
for urban development (Table 27.1, Figure 27.3) problems: a) permafrost or perennially frozen
due to climate, geomorphology or geologic situ- ground; b) frost heave problems; c) seasonal melt
ation. In many environments, Quaternary climatic water flooding; d) river, lake and sea ice. One of
change has left a legacy of ground conditions that the best examples of urban planning in a per-
impose constraints on modern urban development. mafrost area with an active layer is the town of
Thus detailed awareness of previous periglacial Inuvik, the capital of the Northern Territories in
and glacial conditions and the resulting surface Canada. Inuvik had to be built with as little
deposits is essential for the siting and planning of disturbance of the ground as possible. The key
urban development over much of northern planning decisions were: a) to leave the natural
GFE-27.qxd 6/27/05 2:26 PM Page 759
URBAN GEOMORPHOLOGY 759
Table 27.1 Climatic, geomorphological and geological problems for urban development (based on data in Marker,
1996 and Bennett and Doyle, 1997).
Environment Chief problems
A. Climatic
Periglacial Permanently frozen ground and overlying active layer require special types of
construction and foundations for buildings and infrastructure.
Hot drylands Water supply problems; wind erosion; dune migration; flash floods; possibility of
salt weathering of building materials and foundations.
Hot wetlands Rapid weathering and decomposition of building materials; deep, uneven
weathering of most rocks in tectonically stable areas; frequent rain events
causing rapid water erosion of exposed ground surfaces; frequent landsliding.
B. Geomorphological
Mountainous Unstable slopes, soil erosion, rockfalls, large run-out landslides, landslide dams
debris flows and avalanches; potential for flash floods.
Flood plains and Liable to periodic flooding; variable foundation conditions over former, buried
estuaries river channels and alluvial deposits.
Coastal plains and deltas Storm surge, tsunami and flooding risk likely to increase with rising sea levels;
coastal erosion; complex ground conditions reflecting former shorelines and old
drainage channels; saline intrusion; possible salt penetration in groundwater
affecting foundations.
Coasts with weak rock cliffs Liable to rapid coastal erosion, cliff undercutting and collapse; eroded debris
often deposited in ports and harbours causing dredging expenditure.
Islands Storm-surge, tsunami, rising sea-level and salt water penetration risks on
low-lying atolls and coastal plains.
C. Geological
Active plate margins Major risks associated with coastal urban developments, especially on Pacific
rim, special foundation requirements on filled areas, lake sediments and other
unconsolidated materials; major earthquake triggered landslide hazards;
volcanic debris and lahar risks requiring awareness of flow pathways on lower
volcanic slopes likely to have urban settlements.
Shrink-swell clays Cracking clay problems likely to be accentuated by climate change.
Karst Buried karst a major problem for foundations of tall buildings and for sinkhole
development; need for knowledge of buried karst plains and effects of lowered
Quaternary sea levels.
moss covering intact to maintain its insulating buildings had air spaces of about 1 metre beneath
values; b) to place all permanent structures on them and all utility pipes were encased in an insu-
piles securely embedded in permafrost; c) to ban lated, raised utilidor (Cooke and Doornkamp,
all road cuts and ditches to avoid permafrost 1990).
degradation; d) to have culverts installed in gravel Past periglacial conditions pose several, usually
fill to drain seasonal surface water runoff; e) to highly localised and variable problems, the most
build gravel pads on top of the natural vegetation serious of which relate to the re-activation, during
to support all access roads and tracks and some construction on slopes, of fossil slope failures and
temporary structures. When installing the piles, solifluction lobes (Table 27.2). Slopes on stiff
steps had to be taken to avoid disturbing the per- fissured clays affected by periglacial solifluction
mafrost and to protect the mainly wooden piles some 10–13 000 years ago are widespread
against damage in the active layer. Piles were in England and Wales (e.g. Hutchinson, 1991;
sunk to a depth twice that of the active layer. All Jones and Lee, 1994). Although they may have
GFE-27.qxd 6/27/05 2:26 PM Page 760
Figure 27.3 The variety of geomorphological issues associated with urban development.
Table 27.2 Examples of geomorphological problems for urban areas found in former periglacial terrains (Based
in part on data in Bennett and Doyle, 1997).
Problem Effects Examples
Near-surface disturbances
Involutions Surface sediment diversity.
Frost shattering Disrupted and weakened rock surfaces, Weak surfaces on rocks like chalk.
liable to failure.
Solifluction Reactivation of slip surfaces below Road construction problems as mass
solifluction lobes. movements occur e.g. Sevenoaks
by-pass in UK.
Ice wedges Sand fills likely to pose ground stability Collapse of trench walls during
problems. gas pipe laying.
Deep-seated disturbances
Cambering and valley Downslope movement of clock of competent Prominent problem in the Jurassic rocks
bulging rock over weaker, more ductile rocks; of southeastern Britain. Cracks can be
formation of large cracks, known as ‘gulls’. a serious problem in housing estates.
Valley bulges
Fossil pingos Shallow craters buried by sand and gravel. Cause serious foundation problems, as
Common in lowlands where water froze. at Battersea Power Station, London.
Loess and cover sands
Deep loess Low density, high porosity, highly permeable Major problems of slope instability
material, often hundreds of metres thick, in urban areas, e.g. Lanzhou,
prone to sliding and to surface gullying. China.
Cover sands Thinner but similar material that can be Fine silt eroded from thin cover sand
easily eroded. deposits affects urban water supply
treatment in southern Pennines, UK.
GFE-27.qxd 6/27/05 2:26 PM Page 761
slopes of only 2–12⬚ and appear smooth and fea- of stratigraphic information can provide data for
tureless, the presence of many shear surfaces at adequate evaluation of the nature of glacial
residual strength and generally high groundwater deposits. Urban authorities would be well advised
levels makes such slopes easily reactivated when to keep accurate records of boreholes and tempor-
disturbed by moderate cut and fill operations. ary exposures of the urban substrate. It is most
Particular problems arise on the sides of valleys in important to evaluate a number of different explor-
southeast England, in the Pennines and South ations and interpretation of glacial deposits,
Wales. For example, construction of the A21 before making assumptions on which design and
Sevenoaks bypass, UK, had to be halted in 1966 construction programmes are based. Among the
when bulldozers cut through grass-covered lobes classic examples of failure to use existing records
that proved to be the remains of periglacial land- and to ascertain the nature of glacial deposits is
slides and solifluction lobes (Skempton and the Beacon Hill site in Boston, Massachusetts
Weeks, 1976). The affected portion of the route where workers thought they were excavating a
had to be re-routed at a (then) cost of £2M. drumlin landform, but encountered a complex of
materials, dominated by sand and gravel that were
Glacial and fluvioglacial surface part of a glacial moraine (Kaye, 1976). This costly
materials (see also Chapter 13) error delayed the project for many months as the
Glacial and fluvioglacial deposits cover large parts materials had to be excavated and treated in a dif-
of the landsurface of northern Asia, Europe, ferent way from that originally planned.
Canada and Northern USA, with less extensive Fluvioglacial materials result from the huge
deposits in parts of southern hemisphere countries volumes of sediment transported by rivers emer-
such as New Zealand, Chile, Argentina and ging from glaciers and may be classified by the
Australia. Possibly half the population of the USA types of material, mode of deposition and resulting
lives on areas underlain by such deposits. These landforms (see Chapter 13). Their extreme vari-
materials thus support the foundations of many ability poses many problems for urban develop-
urban areas and also provide an important source ment, particularly in determining the safety of
of aggregates for urban construction. foundation piers for structures such as bridges. For
Sedimentary material deposited directly by the example, in building the M60 motorway along the
action of a glacier is usually termed glacial till Mersey floodplain south of Manchester, great
and comprises a variety of material. Till is usually attention had to be given to detecting former river
classified on the basis of the processes of debris channels and buried fluvioglacial materials.
release and the position of debris deposition (see
Chapter 13). Where till has been deposited on Floodplains (see also Chapters 10, 20)
steep hillsides, urban development may lead to A floodplain is in a dynamic relationship with the
changes in water flows in the tills and in the river that flows through it, its sediments, their
pressures between the different types and sizes of stratigraphy and the form of the floodplain reflect-
material in the tills. Landsliding can occur, posing ing the river regime, peak flows and catchment
problems for landowners and supplying sediment characteristics. The concept of the hydraulic flood-
to streams, causing channel changes downstream plain refers to any surface subject to inundation by
and at the coast. a flow of a given return period (say the 1 in 100
Road construction in valleys whose slopes are year flood). This surface is not necessarily con-
till mantled has to take account of the character of fined to the alluvium deposited by the present river.
the till, so that the road cuts are stable and old Land use planning policies should ensure that
landslides are not reactivated. A major problem is development is not permitted in flood risk areas
that only part of the evidence of the nature of (e.g. Lee, 1994). Where development is permitted,
the glacial deposit is revealed at any one time the risk can be reduced by flood-proofing measures.
by excavations, boreholes and exposures. Only a The most effective building modifications include
careful recording of these temporary sources elevating a property above a prescribed flood level
GFE-27.qxd 6/27/05 2:26 PM Page 762
Figure 27.4 Example of a bridge that has suffered scour around its abutments and has required additional protec-
tive works: Pickrell Bridge, Nogales Wash, Arizona.
and, for a single storey property, incorporating a In most fans, the channel repeatedly branches
means of escape such as a dormer window into the down the fan, often through bifurcation associ-
building design. ated with avulsion, conveying the water-sediment
Active floodplains are often subject to rapid lat- mixture to all parts of the fan surface. The return
eral migration of river channels, particularly during period for complete inundation of the fan may be
major floods. In and around many urban areas, several decades or more. In dry environments, the
especially in semi-arid environments such as the effects of a major rare erosional-depositional
American southwest (Figure 27.4) and North event may persist for tens of years.
Africa, lateral migration of channels can wash Many ancient cities were cited on fans. The
away the approaches to bridges in a few hours. To Roman city at Nimes, France was built on an upper
avoid the scour of embankments and bridge Pleistocene-Holocene alluvial fan, the archaeo-
approaches, bridge wingwalls or spur dikes may be logical sedimentary record showing sequences of
constructed to deflect flows away from the channel fluvial deposits and anthropogenic debris, indicat-
banks and direct them underneath the bridge. ing that the site suffered from repeated flooding
Another alternative is to build additional bridge and deposition across the fan (Fabre and Monteil,
openings or spans, so widening the channel and 2001). Some 53 significant floods have occurred
increasing the flood capacity beneath the bridge. since the 14th century, the majority producing
some additional deposition on the fan. Now
Urban areas on alluvial fans although the surface of the fan has been made
Urban settlements on alluvial fans have long proved virtually totally impermeable, major floods, like
both attractive foci for habitation and geomorpho- that of 3rd October 1988, still deposit sediment
logically problematic. Alluvial fans are composed throughout the city and the main streets running
of debris deposited by floods, mudflows or debris down the fan surface become major torrents.
flows, often with minor channels running over them In Alpine regions, villages and towns are often
(see Chapter 16). Urban settlements tend to grow situated on alluvial cones where streams emerge
around the minor channels, ignoring the risk of the from the mountains into major valleys. The Alps
whole fan becoming active in an extreme event. has known several disasters, such as that of the
GFE-27.qxd 6/27/05 2:26 PM Page 763
Figure 27.5 The alluvial fan of the Betterwand torrent (marked stream in the diagram) and the town of Matrei in
the Austrian Tyrol.
Guil valley, where urban settlements were partly water and sediment often invade the desert fan
destroyed by debris flows (Tricart, 1961a, b). As towns.
tourist developments increase, with the growth of Flooding from the Betterwand torrent greatly
high altitude impermeable surfaces and com- affected the town of Matrei in the Austrian
pacted routes for surface runoff, the conse- Tyrol, which is built on the alluvial fan created
quences of urbanization become increasingly by the torrent as it discharges from the moun-
severe, aggravating the impact of rapid melting of tains towards the Tauern River (Figure 27.5).
deep snow packs under warm spring rains, as in The torrent is now constrained by high retaining
the French Alps in 1981. walls and numerous debris barriers in the chan-
Nimes has crucial lessons for the many mod- nel upstream of the town trap sediment
ern desert cities that flourish on or adjacent to (Verstappen, 1983). The considerable engineer-
desert fans. In such cities, much of the partially ing works offer protection from all but the most
or completely urbanised fan terrain has become extreme floods.
impermeable, creating major urban drainage
problems. Usually towns seek protection from Slope stability and mass movement
floodwaters by building walls or diversion weirs Slope failure is a natural process that in urban
to send the floodwater away from the built-up areas is often induced, accelerated or retarded by
areas. However, on fans the floods usually bring human actions, such as:
large volumes of sediment, which tend to build
up in the channels and reduce the flood con- 1. Cutting slopes for roads and other structures,
veyance capacity of the embanked waterways quarrying, removal of retaining walls, and
and diversionary channels. As in Nimes, the lowering of reservoirs.
GFE-27.qxd 6/27/05 2:26 PM Page 764
2. Adding weight by landfills, stockpiles of ore or Piping

rock, waste piles and construction of heavy The development of subterranean channels by
buildings and other structures. water moving through incoherent sediments is an
3. Water leaking from pipelines, sewers, canals, important form of subsurface discharge of water
and reservoirs. and sediment in many areas (Jones, 1981; see
4. Vibrations from explosions, machinery, road Chapter 11). Piping regularly damages bridge
and air traffic. abutments, piers, retaining walls and culvert fac-
5. Decrease of underlying support by mining, ings. Where piping is extensive, highway subsi-
loss of strength or failure and/or squeezing out dence or railway distortion can seriously disrupt
of underlying material. traffic. The dominant type of piping causing dam-
age to urban buildings and infrastructure is ‘stress
The provision of and subsequent neglect of land desiccation’, in which surface runoff, including
drainage systems has had a significant effect on road surface water, drains into a desiccation crack
some slopes (e.g. Bromhead and Ibsen, 1996). and causes erosion of silts and fine sands, the col-
One of the most serious effects is the artificial lapse of channel banks and eventually surface set-
recharge of the groundwater table. Tebbutt (1998) tlement, creating an irregular topography. Other
highlights how average water usage per person types of piping include the process of particle
has increased by 50% in the UK over the last 20 entrainment and high hydraulic gradients, and the
years. Water supply and drainage pipes can be development of subsurface passages due to vari-
vulnerable to ground movement, and it was a key able lithological permeabilities. In some cases,
factor in triggering landslides at Luccombe, Isle large voids do not develop, but seepage erosion
of Wight, in 1988 (Lee and Moore, 1989). and ‘running sand’ cause morphologically similar
Pressures of land availability and cost leading surface phenomena.
increasingly to the building of houses on or close In order to avoid piping problems, pre-
to potentially unstable slopes. A landslide of construction surveys should carefully define
some 180 000 m3 occurred in March 2000 on a materials and other environmental conditions that
steep, 50 m-high, 40-year-old cut slope between might favour piping. Remedial actions may include
two rows of residences in Millbrae, California filters or filter drains, catchment drains with stable
(McCormick, 2002). The landslide left a 20 m channels designed to carry water away from
head scarp and encroached upon 6 homes at the threatened structures, protection for abutments,
base of the slope, damaging 3 and knocking 1 off and careful avoidance of concentration of drainage
its foundations. The landslide is along the west- water where it infiltrates into natural sediments.
ern margin of an old, large quarry cut into highly
sheared and foliated shales and melange rocks. Salt (see also Chapter 16)
After quarrying ceased, houses were built at the The breakdown of rock and concrete by salt
top and the base of the 1.5 : 1 to 2 : 1 (horizontal weathering, found in many environments, is par-
to vertical) slopes. ticularly serious in arid and semiarid areas. High
The translational landslide failed as a block rates of evapotranspiration concentrate salts in
landslide (with minor rotation) along a basal fail- ground water hot, dry environments, while salt
ure surface dipping 5 to 10 degrees out of the water penetration into aquifers occurs in coastal
slope. The lower portion failed as a semi-plastic zones. Thus the risk of salt attack is often highest
flow above the toe of the slope. The failure was along arid coastlines. Salts tend to be precipitated
due to a gradient that was too steep for long-term in the upper layers of the soil. This salt causes
stability of the materials, weathering of the accelerated weathering in four ways: a) by the
bedrock over time and to localised surface insta- growth of salt crystals in cracks and fissures in
bilities and erosion that ultimately overcame rocks and building materials; b) through thermal
the marginal stability of the existing cut slope expansion of salt crystals under the diurnal heat-
(McCormick, 2002). ing cooling in hot deserts, leading to stresses in
GFE-27.qxd 6/27/05 2:26 PM Page 765
the building materials and rocks; c) breakage of topography, drainage conditions and the way
rocks through expansion caused by hydration of buildings are constructed. Drainage of such smec-
salts; d) reactions of chloride and sulphate salts titic clays alters their swell potential.
with cement. The areas where these salt problems Precautionary measures that can be undertaken
are likely to be acute need to be avoided in devel- include:
oping new urban areas. Foundations can be pro-
1. Adequate site investigations of both the soil
tected by using salt resistant cement, careful
and the substrate;
screening and washing of aggregates to avoid
2. Location of vegetation away from the house;
salts, using dense low porosity concrete, and seal-
3. Establishing strong foundations (building on
ing foundations to prevent attack by groundwater
a reinforced concrete platform (‘raft’) rather
(see Chapter 16).
than brick piers);
4. Elevating the structure so that moisture cannot
Expansive soils (see also Chapter 7)
collect near or under it;
In many parts of the world, cracking clay soils are
5. Replacing potential swelling soil with more
a major constraint on urban development. Many
stable earth material;
clay soils expand when wet and shrink and crack
6. Providing physical barriers in the ground to
during long dry periods. Such ‘shrink-swell’
prevent moisture from accumulating there;
phenomena cause differential shifts of parts of
7. Pre-wetting the terrain prior to construction to
the structure, such that floors tilt slightly and
ensure that all stresses are relieved (Coates,
windows and doors no longer close properly.
1985).
A high smectite clay content is particularly likely
to raise the shrink-swell potential. The new
Denver International Airport is built on clays with Coastal sites (see also Chapter 21)
swell potentials as high as 15 per cent and aver- Urban development on the coast alters the dynam-
aging 8 per cent. These potentials had to be taken ics of beaches, cliffs, estuaries, mangroves and salt
into account in the design of the terminal, parking marshes. Examples abound of former ports that
garages and airport office building as well as the silted up and are now well inland, such as Rye in
airfield pavements themselves. Consideration had Kent, England and Sluis in West Flanders,
to be given to the effects of heavy aircraft landing Belgium. Other instances occur of shorelines
on the runways and the compaction and compres- needing protection and beaches requiring replen-
sion effects this would produce on the underlying ishment, as at Bournemouth, England (Spencer,
clays. 1999) and Knokke-le-Zoute, Belgium (Douglas,
The US Soil Conservation Service uses the 1971). At Bournemouth, the sandy beach so essent-
coefficient of linear extensibility (COLE) as a ial for the tourist industry had beach replenish-
measure of the shrink-swell potential of soils. The ment activities in 1970, 1974–5 and 1988–90 with
COLE represents an estimate of the vertical com- the next phase due in the autumn of 2005. At
ponent of swelling in a natural soil clod. Soils Knokke, annual replenishment of beach sand has
with coefficients less than 3% are judged to rank been necessary since the late nineteen-fifties.
low in soil expansion capability; those between 3 Coastal landslides affect a wide variety of
and 6% are judged to be moderate; and those with urban settlements, some simply because houses
coefficients greater than 6% are judged to rank are built in unsuitable places, others because slope
high in this capability. The higher the coefficient, stability has been altered by urban development.
the more severe is the shrink-swell behaviour of a Undercutting of cliffs has affected residential
soil, and the greater the risk of damage to build- areas from San Diego, California to the boulder
ings. Variations in the properties and potential of clay coasts of eastern England.
expansive clays soils are related to the substrate, Around the world, large sums have been
water-table depth, antecedent soil conditions, invested in an attempt to resist the loss of cliff top
amount of new water added to the soil, vegetation, properties, services and infrastructure and to
GFE-27.qxd 6/27/05 2:26 PM Page 766
mitigate the risk to public safety (Brunsden and Vietnam exhibits a whole range of buried karst
Lee, 2004). In England, for example, there are engineering problems that can only be overcome
some 860 km of coastal protection works, with by careful geophysical investigation of subsurface
over £20 million spent each year on maintaining geomorphology.
and improving these defences, and providing new These foundations problems had to be taken
schemes. The Holderness coast, UK has retreated into account during the preparations for building
by around 2 km over the last 1000 years, leading the Petronas Twin Towers, the world’s tallest
to the loss of 26 villages listed in the Domesday building at the time (Pelli et al., 1997). Beneath
survey of 1086. 75 Mm³ of land has been eroded the level surface of the site, the pinnacled surface
in the last 100 years (Valentin, 1954; Pethick, of the karst plain sloped from 15 m depth under
1996). Rapid recession has also caused severe the northwestern corner to a depth of over
problems on the Suffolk coast, most famously at 180 metres at the southeastern edge. In order to
Dunwich where much of the city has been lost get a conventional level foundation for such a
over the last millennium (e.g. Bacon and Bacon, huge weight of concrete, the bedrock would have
1988). Gardner (1754) records that by 1328 the to be excavated at one edge, while at the other
port was virtually useless and that 400 houses piles would have had to be sunk to a far greater
together with windmills, churches, shops and depth than normal at the other. The risks and costs
many other buildings were lost in one night in were so great that the building was moved 60 m to
1347. the southeast to allow for more than 55 m of allu-
vium beneath each tower. To do this successfully,
Karst (see also Chapter 24) a foundation designed to withstand movements
Soluble rocks have subsurface conditions that are with the alluvium was required.
particularly acute in buried karst plains, such as The solution was to spread the load throughout
those underlying Florida, parts of the Caribbean the alluvium so that even movement could occur.
and parts of Southeast Asia. Around many of the A concrete mat was designed to spread the weight
prominent tower karst hills of northern Vietnam, of the building over a set of drilled 1.3 m diame-
southern Thailand, Malaysia and Indonesia, ter piles. These piles would transfer the weight of
extensive alluvial deposits bury karst surfaces the towers to the soil more gradually than a simple
that are often pinnacled, encumbered with lime- mat would. Friction between the surface of a pile
stone blocks and pitted with sinkholes that devel- and the surrounding soil would prevent the foun-
oped during Pleistocene low sea levels. The dation supports from sinking. Settlement would
buried karst now poses serious problems for civil then occur between the pile tips and the bedrock.
engineering works (Bergado and Sebanayagan Pile lengths were varied so that their tips were all
1987). New high-rise buildings require deeper at the same height above bedrock, so as to avoid
foundations than the low rise buildings that any tilting of the foundations.
sufficed until the nineteen-seventies. In Kuala This technique introduced a new concern. The
Lumpur, the low-rise structures had their founda- interlocking grains within the alluvium would
tions on the stiff clay layer within the alluvium. cease to be locked together if excavation removed
Taller multi-storey structures require piling into overlying material and allowed the alluvial mass
the underlying limestone. However, the irregular- to expand. The piles were therefore sunk from as
ity of both the karst surface and the cavities near the ground surface as possible and the con-
within the buried karst means that foundation crete mat was laid on top of them. The extremely
investigations have to be particularly careful. careful investigations and detailed precautions
Drill holes may strike limestone, unaware as to taken during the planning of the foundations of
whether it is buried rockfall material or a pinna- the Petronas Twin Towers show just how impor-
cle, while a neighbouring hole might pass through tant it is to have a full understanding of the
several more metres of alluvium before hitting ground and subsurface conditions of any major
limestone. Analogous karst terrain in northern engineering structure.
GFE-27.qxd 6/27/05 2:26 PM Page 767
Urban areas on peat urban environmental planning and management.

Many towns have expanded on to peat, building This involves: 1) Mapping the aerial distribution
houses in areas where drainage has lead to ground of permeable strata and fracture zones and effects
surface lowering (Berry et al., 1985). This is not a on the migration of ground and surface water, and
new phenomenon in human settlements. For on pollution plumes in these waters; 2) Mapping,
example on the coastal plain of Flanders in north- geophysical surveys, and hydrologic studies of
eastern France and Belgium, inversion of relief hazards, such as expansive soils, flooding, subsi-
has occurred in historic time, with the narrow dence due to ground water withdrawal, landslides
ridges on sandy materials being the former beds and faulting; 3) Providing GIS-based information
of streams between which the drained peat has critical for planning and zoning and communica-
shrunk, with the result that the oldest settlements tion and outreach to the public through the Lake
are on the low peat surface and younger villages Mead-Lake Mojave National Recreation Areas;
on the ridges. 4) Evaluation of the stratigraphic and structural
Industrial development and railway construc- setting of Las Vegas basin to determine ground
tion brought the classic engineering problems of water, petroleum, building materials, and mineral
crossing peat deposits, such as Chat Moss west of resource potential. The primary project objective
Manchester on the route of the 1829 construction is to determine the geologic, geophysical, and
of the Liverpool to Manchester railway. Now hydrologic history of Las Vegas Valley.
new housing developments on drained peat often In Umbria, Italy Antonini et al. (2000) used
require pre-loading of the peat to ensure that aerial photographs taken over 50 years for a
subsidence is artificially induced before any multi-temporal investigation of the slopes at
buildings are constructed. 80 sites to map the most active landslide hazard
areas. A simple vulnerability map showing
The role of geomorphological mapping houses, buildings, roads, railroads, power lines,
Geomorphology in city planning. If geomorpho- and related infrastructure was derived from
logical surveys are carried out sufficiently early available 1 : 10 000 topographic maps and the
in the planning of a new city, and their results are most up-to-date aerial photographs at each site.
digested into planning discussions, some of the A GIS intersecting the hazard and vulnerability
potential dissonance between environmental maps was used to derive ranked levels of risk
conditions and the city plan can be avoided (e.g. (Cardinali et al., 2002).
Bernknopf et al., 1993). Radar remote sensing has also been used to
Geomorphology in site development. Geomor- monitor landslides (e.g. Kimura and Yamaguchi,
phological contributions to site development 2000). Differential interferometric synthetic aper-
include aid in rationalising site plans to avoid ture radar (DInSAR) allows for the measurement
hazards, formulation of building and site choice of small deformations of the terrain (at a fraction
adapted, and sampling sites for ground material of the radar wavelength i.e. millimetres). DInSAR
investigations. uses two different SAR images acquired at differ-
Geomorphology and post-construction manent times plus either a third image or a digital ter-
agement. The monitoring of geomorphological rain model (DTM). The imagery and DTM are
processes and landform changes after construc- geocoded and after the removal of topographic
tion can play an important role in helping envi- and atmospheric effects, surface shift over the
ronmental managers to influence future planning period spanning the image pair is calculated.
policy and alleviate or avoid environmental DInSAR complements traditional ground survey
hazards. measurements and has a wide range of applica-
In the Las Vegas area, the US Geological tions, such as the monitoring of mining-induced
Survey is engaged in the systematic production of subsidence, urban subsidence due to tunnel con-
general-purpose 1: 24 000 and 1:100 000-scale struction and/or groundwater extraction, land-
geologic maps to provide a uniform data base for slides and mass movements.
GFE-27.qxd 6/27/05 2:26 PM Page 768
In Italy, both DInSAR and multi-temporal Ideally, development in Kuching should be

SAR have been used to investigate slope move- directed towards the very suitable and suitable
ments at Caramanico (Refice et al., 2000, 2001). areas and expansion on the less and least suitable
Landsliding is the most frequently occurring geo- areas should be controlled. Nevertheless, shortage
logical hazard in the resort of Caramanico, in the of suitable land has meant that a few of the
south-central Apennines. Seasonal re-activation unfavourable areas have already begun to be used
of several mass movements threatens the town’s for development. Among the problems arising in
historic centre, cemetery and infrastructure. The the peat areas in the least suitable category is
large landslides, up to 1 km in maximum dimen- excessive settlement. Any increase in the load
sion, generally move slowly, and thus their on the peat expels pore water and leads to com-
changes are potentially detectable from SAR data. paction of the peat and lowering of the ground
An approach similar to the so-called ‘permanent surface. Any attempt to drain the peat leads to
scatterers’ (PS) technique, originally proposed by further water expulsion and more settlement, as
Ferretti et al., 1999, has also been attempted to happened in the Petra Jaya housing estate close to
monitor small displacements of objects charac- Kuching where poorly supported fences collapsed
terised by high backscattering stability. (Lam, 1992). Houses built on piles on the peat
Kuching, Sarawak, provides a good example of tend to remain stable, while those with concrete
the variety of foundation conditions experienced floor slabs laid directly on the peat generally find
in a low-lying coastal city with large areas of peat that eventually the concrete cracks and breaks.
(Figure 27.6). Detailed mapping enables four broad
categories of foundation conditions to be recog-
nised (Lam, 1992): 27.3 Impact of urban activities on
geomorphic processes
1. Very suitable areas underlain by residual
deposits and terrace deposits or both; generally Impacts during construction
hilly or undulating terrain; tens of metres During construction, the original vegetation is
above flood level; less than 10 m above disturbed and the soil and weathering profile is
bedrock; no subsidence risk; little or no filling exposed to the erosive agents of wind and rain.
required (surface stripping usually required to Road construction can lead to rapid erosion if
level the ground). cuts are too steep and roadside drains, with regu-
2. Suitable areas underlain by shallow estuarine/ lar cross drains, are not established. Temporary
deltaic deposits and riverine deposits less than access roads and building yards are particularly
5 m thick; peat deposits less than 1 m thick; liable to surface erosion. Erosion on construction
normally occurring as narrow belts around the sites results in the deposition of about 80 million
residual deposits; at or below flood or highest tonnes of sediment a year in US lakes (Harbor,
tide levels; 10 to 15 m to bedrock; minor sub- 1999).
sidence risk; some filling required. Many construction sites are levelled by cut and
3. Less suitable areas underlain by estuarine or fill operations, with the less compacted fills being
deltaic deposits more than 5 m thick; below particularly liable to be eroded by rills and gullies
flood or high tide level; 15 to 30 m above if a protective cover is not provided quickly. If the
bedrock; moderate amount of prolonged fill lies on weathered rock or colluvium, the per-
subsidence; large amount of fill required. meability of the fill is likely to be greater than that
4. Least suitable areas underlain by deep peat of of the material below and thus water will tend to
5 m or more thickness; above flood level move laterally at the base of the fill and may
but always water-logged; very poor surface induce sliding. Erosion of fill is likely to create
foundation (extremely soft and wet); 12–20 m deep gullies whose width will depend on the
to bedrock; severe subsidence risk; large cohesiveness of the material, but normally is
amount of fill required. greater than the width of gullies in weathered rock
GFE-27.qxd 6/27/05 2:26 PM Page 769
Figure 27.6 Urban land suitability classification for Kuching, Sarawak.
in situ. At places where throughflow emerges, particularly those associated with the removal of
overhanging rounded hollows, or shallow cavities underground resources, including ground water, oil,
develop in the gully walls. brines, salt, coal and other minerals. The extraction
of resources necessary to sustain the industry and
Impacts after development is completed people of a city can lead to problems for sectors of
Subsidence (see also Chapter 12). Many urban the urban population. In north-eastern Phoenix,
processes lead to geomorphological change, Arizona, groundwater abstraction resulted in up to
GFE-27.qxd 6/27/05 2:26 PM Page 770
1.6 m of subsidence in the centre of a cone of lags. In southeast London on April 7th 2002, nearly
depression of the groundwater table, caused by 100 residents from council and private properties
constant pumping between 1962 and 1982 (Péwé, were urgently evacuated from properties on
1990). Sometimes, earth fissures develop as a Blackheath Hill after subsidence caused a crater
result of such subsidence. A 130 m-long fissure in nine metres in diameter to appear in the road next
a northeast Phoenix residential construction site to their homes. The hole caused both carriageways
was reported as the first known occurrence in an of the main A2 road to be closed. The crater is
urbanising part of Arizona (Péwé, 1990), but it was thought to have been caused by collapse of old,
expected that as cities grow, more and more such and possibly illegal or undocumented, chalk work-
fissures will occur, endangering structures in the ings. Drilling revealed additional chalk workings,
affected areas. 15 metres and 30 metres below the surface of the
The Valley of Mexico is a hydrologically closed, road, at two locations. Most of the cavities had
graben structure, that had a series of shallow lakes probably been filled with quarrying spoil or loose
in the early 16th century. Groundwater recharged in sand several centuries ago, as a result of earlier
the mountains surrounding the Valley flows into the chalk tunnel roof failures. The Blackheath Hill
Basin of Mexico. Thick lacustrine clays cover the area was nearest source of chalk for lime making
Valley floor and artesian conditions once prevailed. to the centre of London and was already in use in
The water table is close to the surface over much the 1666 when the Great Fire of London destroyed the
lowland around the ancient lakes. Relatively small City. Few controls on chalk removal existed at the
rain events cause overland flow and flooding, prob- time (Smith, 2002). Similar issues arise in most
lems always faced by communities in the Valley. large urban areas. Planning departments need
Heavy pumping has caused drainage and consoli- access to inventories of past land use and former
dation of the lacustrine clays, and consequently land mineral workings so that problems of this type can
subsidence of up to 8 m in the central part of the city be avoided.
(Durazo and Farvolden, 1989). While in 1975 the Many early mine workings were not recorded
recharge of the aquifer was 38 m3 s⫺1 and abstrac- and only manifest themselves when a collapse
tions were 30.3 m3s⫺1, by 1990 recharge had occurs. The appearance of several holes along-
declined to 25 m3s⫺1 and abstractions had risen to side the UK East Coast main railway line to the
about 53 m3s⫺1, drastically lowering the water table. southeast of Edinburgh was traced to the collapse
In the centre of Mexico City, subsidence developed of mined coal seams. Overcoming the problem
following the construction of over 50 deep wells. required the realignment of the high-speed track
Subsidence averaged 8 cmy⫺1 from 1935 to 1947, onto ground, which had been treated by grouting
29 cmy⫺1 from 1947 to 1958, and 7.5 cmy⫺1 from at close centres to fill voids, with three sections
1959–1985. Since 1985 subsidence has probably of track carried on concrete slabs supported by
continued at about 10 cmy⫺1. However, the slowing mini-piles. A total of 17 000 t of grout and 162 km
of subsidence was the result of controls on pumping of drilling were required for the 1.8 km long
in the central area. New wells were then sunk in the diversion (New Civil Engineer, 2003).
southern areas of Chalco, Tláhuac and Xochimilco. Mass movements. Landsliding triggered by
Since 1995, differential SAR interferometry has urban development is found from the polar to the
shown subsidence of around 40 cmy⫺1 in some of equatorial regions. Changes to water flows caused
these areas, with a maximum of 47 cmy⫺1 southeast by construction, paving and the weight of struc-
of the International Airport (Carnec et al., 2000). tures on the ground, disrupt patterns of pore water
This case illustrates that measures successfully pressures and lead to slope failure. Lee et al.
reducing geomorphological problems in one part of (2000) note the importance of human impact in
a large urban area may have to be repeated years the artificial recharge of groundwater, and also
later in another area. the re-activation of old slides by changes in load-
Problems of former mine workings. Urban ing conditions. Gutjahr et al. (2000) discuss the
impacts of past activities may occur with long time correlation between increase in urbanisation and
GFE-27.qxd 6/27/05 2:26 PM Page 771
Table 27.3 The hydrologic impacts of urbanisation (after Leopold, 1968; see Chapter 10).
Impact Effects
Change in total runoff Impermeable surfaces alter balance between evaporation, transpiration, infiltration
and surface flows.
Alteration of peak flow More rapid rise of streamflow to peak storm discharge in totally or largely urbanised
characteristics streams, especially in short duration, intense storms. More frequent moderately
high flows in rivers flowing through cities.
Decline in water quality Initially through increased sediment flows from construction sites followed by
chemical and organic pollution from urban drainage and storm sewer overflows.
In a growing city all these sources operate at the same time but in different parts
of the urban areas.
Changes in the hydraulic Hydrogeomorphic changes, including meandering to braided state when sediment
behavior of streams yields and peak discharges are greatly increased. Channel widening downstream of
culverts and bridges. Channelisation and embankments construction. Flood controls
works. Channel change downstream of regulated reaches. Changes in channel
vegetation.
Table 27.4 Sequence of fluvial geomorphological response to land use change: Sungai Anak Ayer Batu,
Kuala Lumpur.
Land cover/land use Channel condition
Forest Narrow, meandering with low sediment load

Rubber plantation Gullying during clear weeding; Peak discharge increased; Channel slightly
widened; Later stabilized; Few cut-offs
Urban Construction High sediment yield; High peak discharge; Metamorphosis to wider, steeper,
shallower braided channel
Channelisation and stable urban Higher peak discharge; Less sediment load; Channel enlargement downstream;
built-up area Bank erosion, Minor channel incision; Loss of fine bed material by scour.
Post channelisation siltation Where large quantities or organic debris enter concrete channels and are
deposited, vegetation can become established and build up deposits that
reduce channel capacity.
higher frequency of landslides, triggered by lesser Kingston’s new development is concentrated, 866
rainfall events. different slope failures were counted prior to
Kingston, Jamaica (Gupta and Ahmad, 1999) intense rains in May 1991 that added 84 new fail-
illustrates the interplay of biogeophysical and ures and reactivated 540 others (Maharaj, 1993).
societal pressures in the slope instability: urban Fluvial channel change. Urbanisation alters
land use sequence. Deforestation and vegetation the water balance of the affected areas and
modification followed by construction by the changes the local hydrology (Table 27.3). The
affluent on the higher slopes and invasion of the changed water flows alter patterns of land sur-
lower slopes by the less wealthy leads to shallow face and channel erosion. In extreme cases,
slides and debris flows following heavy rains. increased peak runoff and sediment yields from
Failed slopes recover and acquire new vegetation, construction sites lead to channel metamorphosis
but later are used as house sites with temporary from meandering to braided state (Table 27.4,
access roads until the slope fails again in the Figure 27.7). Bank protection and channelisation
next heavy storm. The sequence is then repeated. work often leads to renewed erosion further
In the mountains of St. Andrew where much of downstream, just as below the fixed abutments
GFE-27.qxd 6/27/05 2:26 PM Page 772
Figure 27.7 Channel changes during and after construction in the Sungai Anak Ayer Batu catchment,
Kuala Lumpur.
of bridges, rivers tend to widen by scour of the has caused some subsidence of the original sea-
banks. In many ways, work to protect river banks bed, with allowance having to be made for this in
and to reduce flooding can serve to shift the the operation and maintenance of the airport. The
flood, sediment and erosion problems further problems of subsidence of the second stage run-
downstream. Thus many large cities have well way are expected to be more severe than in the
protected, highly managed river reaches through first runway, with a prediction that after 50 years
their central commercial and smart residential subsidence will have been 18 m compared to 11 m
districts, but eroding banks and degraded chan- for the first stage (Kansai Airport Authority,
nels further downstream. Barbatic et al. (1998) personal communication 25 September 2000).
describe measures involving flow velocity reduc- Detailed analyses have been made of the way
tion, energy dissipation and increased river bed landing aircraft cause small temporary depressions
stability, required to prevent potentially cata- in the runway that in turn affect the drag on
strophic failures due to erosion of a channelled aircraft moving along the runway (Endo, 2000).
river. These channel adjustments can threaten As disposal of solid waste moves from landfill
buildings, bridges and utility pipelines. Channel to land raise, new hills appear on the edges of
aggradation can often raise stream beds above the floodplains, above former gravel pits and quarries
level of adjacent urban areas and, hence, increase and on offshore islands. From the huge dumps on
flood risk. the edges of cities like Istanbul and Manila to the
managed disposal areas, such as Freshkills on
Staten Island, New York, which has been taking
27.4 Landform creation by nearly all the 17 000 tons of waste the city collects
urbanisation each day (Polan, 1992), local skylines are becom-
ing dominated by landforms of rubbish accumula-
Landforms of accumulation tion, in the way that others were marked by coal
Much modern urban development involves land waste tips in the early 20th century. As events at
reclamation and major landform modification. In the Payatas tip in Manila have shown, some of
extreme cases, huge quantities of material are these urban wastes mounds are unstable, prone to
moved, for example in the development of major massive slumps and landslides. The loss of life and
airport sites such as Kansai, Singapore and Hong property that ensues is a challenge to the manage-
Kong (Figure 27.8). At Kansai, the fill material ment of the waste disposal and the application of
GFE-27.qxd 6/27/05 2:26 PM Page 773
Figure 27.8 Transfers of earth materials involved in the creation of the new airport for Hong Kong.
geomorphology to the construction of land raise Landforms of extraction

mounds. Urban demands for building materials, stone,
In some parts of Britain, waste dumps are gravel, sand, limestone and brick clays have led to
prominent features of the landscape.The older the profound transformation of landforms around
dumps are the result of coal, slate and china clay cities. Many areas of gravel extraction along river
production. Modern land raise features dominate valleys and former floodplains have transformed
many low relief areas. Disposal to landfill of con- large areas into wetlands (Figure 27.9). As cities
trolled industrial and municipal waste in Britain have been extracting these materials for centuries,
accounts for some 108 Mt per annum (DETR many former pits and quarries have been filled in
1997). Whilst much of this waste is deposited in and re-used for other purposes. Detection of filled
disused open pit mines and quarries, land raise quarries and pits may require the use of geophys-
mounds are probably the fastest growing artificial ical techniques (McDowell, 1975).
landforms in Britain today. The greatest geomor- Sometimes there has been quarrying of clays
phological impact of landfill is in river valleys, from the surface and mining for chalk below
sections of which are being filled, raising the ground at the same site, posing problems for later
height of the ground surface well above the for- urban development. Near Reading, England,
mer floodplain level. This effectively reduces the brickmaking occurred at the former Katesgrove
flood storage capacity of the floodplain, shifting clay pit from at least the 1600s. The resu
the flood problems downstream. lting urban geomorphological hazards include
GFE-27.qxd 6/27/05 2:26 PM Page 774
Figure 27.9 Development of the Trent Valley, Nottingham, UK. Gravel extraction has left numerous flooded pits.
Photograph courtesy of Tony Waltham Geophotos.
significant thicknesses of made ground and rem- in terms of anomalous, sudden lithology, colour
nant over-steepened clay pit faces. Houses built and structure changes.
above the clay pit face have been so affected by 4. Information on the recorded occurrence of
slope instability that a number had to be domol- solution features and mining cavities can be
ished. When the clay pit floor was redeveloped for obtained by searching national databases.
housing it was first necessary to stabilise the old In the UK, a series of National Reviews
clay pit face. However, when the developer started of specific ground-related problems were
to put piles down into the underlying chalk for commissioned by the Department of the
stronger foundations, a chalk mine was discov- Environment, including: Review of Landsliding
ered below the site. The type of foundation had to (Geomorphological Services Ltd, 1986–87);
be changed. However, as the mine was not filled- Review of Mining Instability (Arup Geotech-
in, future potential risks for adjacent roads, car nics, 1991); Review of Natural Underground
parks and landscaped areas remain. Cavities (Applied Geology (Central) Ltd, 1994);
In the clays, sands and gravels that overlie the Review of Foundation Conditions (Wimpey
chalk (known as the Lambeth Group deposits in Environmental Ltd, 1994); Review of Natural
Britain) and similar materials elsewhere, the fol- Contamination (British Geological Survey,
lowing precautions should be taken during site 1994).
investigations:
Often former quarries and pits are used as
1. Always consider the potential for instability. landfill sites before being redeveloped for other
2. Carry out a desk study and look at historical urban uses. Construction over landfill introduces
maps for evidence of past quarrying in Lambeth another set of problems, unless great care is
Group deposits and similar formations. Some taken. Structures placed directly on a landfill are
evidence may be old (1500s to 1700s). subject to settlement and to the escape of methane
3. Solution features commonly affect Lambeth gas. A restaurant and shops built on a shallow,
Group and similar deposits – evidence is visible 8 metre thick, landfill south of Los Angeles, were
GFE-27.qxd 6/27/05 2:26 PM Page 775
placed on piles driven through the landfill to stop a good starting point for assessing erosion risk on
them from settling. The pathways around the construction sites.
buildings were not and a few years later had set- Construction should be carried out in phases to
tled so much that they had to be raised up to the avoid disturbing too much of the land at any one
level of the adjacent buildings. The methane gas time (Table 27.5). No unnecessary clearing
at the site was collected and used for gas lighting should be undertaken. Immediately below any
in the shopping precinct (Pipkin and Trent, 1994). cleared area, detention ponds should be con-
A useful report relating to construction over land- structed to retain any sediment washed of the site
fill was written by Charles (1993). and to hold back stormwater runoff so that peak
discharges in the stream below are not increased.
Particular attention should be paid to the design
27.5 Remediation and Prevention of construction roads and later of permanent
roads, following four basic principles and adopt-
Erosion avoidance, prevention and ing erosion control measures (Table 27.6).
control
Construction site erosion control is important for Landslide avoidance, prevention and
the control of non-point source pollution requiring control (see also Chapters 8, 9)
the combined efforts of design engineers, site The Japanese Landslide Prevention Law permits
engineers, geomorphologists and soil conserva- the designation of a landslide-threatened area. In
tion experts (Harbor, 1999). The practical guide- an urban area the landslide-threatened area to be
lines of the US Department of Agriculture and the designated must have a risk of causing damage to
New South Wales Department of Conservation are at least one of the following: a river of public
Table 27.5 Erosion and sediment control on construction sites (after Harbor, 1999).
Strategy Implication
Fit the development to the site conditions Fit development to the natural contours of land; use unstable and highly
erodible areas as open space; avoid steep andfrequently wet areas.
Retain as much existing vegetation Avoid disturbing as much ground as possible; clear small parcels at
as possible a time and re-establish permanent plant cover as quickly as possible.
Minimise bare soil exposure by managing Delay clearing and earth moving until just before construction
grading and construction timing starts; clear large sites in stages; concentrate work in dry
seasons or periods when erosion risk is least.
Establish temporary surface cover Use mulches, temporary turf, shredded plant material, geotextiles
or plastic sheeting as temporary ground cover.
Divert water away from disturbed areas Use diversion channels to take water away from construction areas.
Minimise slope length and gradient Break long slopes with intercept ditches; install cross drains
on roadways.
Minimise runoff velocity and discharge Divert unconcentrated runoff onto vegetated surfaces; line drainage
channels with vegetation or rock to increase roughness; use check
dams to slow flows.
Design drainage channels and outlets Ensure channels can cope with worst case flows; avoid risk of
to be non-erosive undercutting and erosion at channel outlets.
Trap sediment on site Use silt fences, straw bale dikes and sediment traps, ponds and
basins to hold back sediment.
Inspect and maintain Ensure all erosion control works are in good condition; maintain
sediment trapping structures, including controlled removal of
accumulated sediment where necessary.
GFE-27.qxd 6/27/05 2:26 PM Page 776
Table 27.6 Basic principles and erosion control measures for urban road construction.
Basic principles Erosion control
Minimise the amount of disturbance caused Use as narrow a road as possible commensurate with traffic speed,
by road construction by a) controlling the safety requirements and erosion hazards.
total mileage of roads, and b) reducing
the area of disturbance on the roads
that are built
Avoid construction in areas of If roads have to be built over steep and potentially unstable areas,
high erosion hazard and no alternative route is available, traffic restrictions, such as
one way working along a narrow stretch, should be considered.
Minimise erosion on areas that are Attempt to balance the volume of cut and fill material to minimise
disturbed by road construction by a excavation. Use proper layer placement and compaction techniques
variety of practices designed to wherever possible on fills to ensure their stability against mass
reduce erosion movement or erosion. In potentially unstable materials, such as
deeply weathered granite, full bench construction (with no fill
slope) may be preferable. Haul excavated material to safe disposal
areas. Stabilise the waste disposal areas as well as the road.
Minimise the off-site impacts of erosion Allow and promote vegetation establishment on unpaved areas,
especially the faces of cut and fill areas. Provide properly
designed road surfacing to prevent roadway erosion and maintain
a stable road. Provide adequate cross drains to prevent water
erosion and gully development in ruts and drains. Drains should
be located to cope with particular runoff and sediment problems.
They must be adequately maintained and their outlets to natural
channels designed to avoid downstream scour and siltation.
interest; railway tracks; prefectural and certain (e.g. Halcrow, 1986; Lee and Moore, 1991; Lee
municipal roads; other important public facilities et al., 2000). In France, production of maps of
and buildings, such as government offices, zones exposed to risks of soil movements (the
schools and hospitals. A landslide-threatened area ZERMOS system) have sometimes preceded the
can also be designated in special circumstances so production of 1 : 5000 land occupation plans and
that houses at risk may be moved. have even led to 1 : 5000 plans of risk exposure
Under the Swedish Planning and Building Act (Brand, 1988). Since 1994, plans for the prevention
of 1987, planning and granting of building permits of foreseeable risks have been required in France,
must take into account the suitability of the ground involving public participation and the application
on which development is to take place. This of the principles of sustainable development
includes soil, rock and groundwater conditions and (Dagorne and Dars, 1999). These developments
thus slope stability conditions. The responsibility give geomorphology a key role in the French urban
for ensuring that urban development does not take planning and development control process.
place in areas of insufficient stability and for
avoiding cumulative development eventually lead-
ing to increased slope instability rests firmly with 27.6 Summary
the municipalities and partly with county adminis-
trative boards (Ahlberg et al., 1988). Urban authorities and governments have to plan
Landslide hazard mapping and assessment of for multiple land uses, and sequential regeneration
landslide risks is widely used to identify the loca- and re-use of the land. Increasingly building codes
tion and magnitude of potential slope instability and planning regulations and guidance have to
GFE-27.qxd 6/27/05 2:26 PM Page 777
accommodate not only the natural geomorpho- Antonini, G., Ardizzone, F., Cacciano, M., Cardinali, M.,
logical and surface geological conditions that may Castellani, M., Galli, M., Guzzetti, F., Reichenbach, P.
and Salvati, P. (2000) A geomorphological approach
create hazards for urban activities, but the people-
to estimate landslide hazard and risk in urban and
made geomorphology and surficial sediments from sub-urban areas: examples from the Umbria region.
clay pits and landfills to made-ground and re-used EGS 2000, 25–29 April, Nice.
industrial land (Leach and Goodger, 1991). In some Applied Geology (Central) Limited (1994) Review of
cities, the present buildings are the fourth or fifth Natural Underground Cavities. Report to the
Department of the Environment.
set of structures to occupy a particular site over the
Arup Geotechnics (1991) Review of Mining Instability.
last 200 years. In urban areas, it is not only essential Report to the Department of the Environment.
to ‘know the ground you build on’ in terms of geo- Bacon, J. and Bacon, S. (1988) Dunwich Suffolk.
morphology, geology and urban sedimentation, it is Segment Publications, Colchester.
important to ‘know the land use history and people- Barbatic, Z., Langof, Z., Goluza, M., Steger, Z.,
Martinovic, D. and Lasic, M. (1998) Geotechnical
made geomorphological changes to the land you
Hazards, Balkema, Rotterdam.
build on’. Building codes and planning guidance Bennett, M. R. and Doyle, P. (1997) Environmental
have to recognise the need for a geomorphology of geology:Geology and the Human Environment,
brownfield sites, old quarries, landfills and gravel Chichester, Wiley.
pits. In England and Wales, Planning Guidance Bergado, D. T. and Sebanayagan, A. N. (1987) Pile
foundation problems in Kuala Lumpur limestone,
note PPG14 on building on unstable ground
Malaysia. Quarterly Journal of Engineering
(Department of Environment and Welsh Office, Geology 20, 159–175.
1990, 1996) makes a start in this direction. Bernknopf, R. L., Brookshire, D. S., Soller, D. R.,
Research in the UK on land instability (for McKee, M. J., Sutter, J. F., Matti, J. C. and
example Doornkamp, 1988; Lee and Moore, 1991; Campbell, R. H. (1993) Societal value of geologic
maps. U.S. Geological Survey Circular 1111.
Thomson et al., 1996; DETR, 1998) has integrated
Berry, P. L., Illsley, D. and McKay, I. R. (1985)
complex land instability, flooding and former Settlement of two housing estates at St. Annes due
industrial land use information to provide guidance to consolidation of a near surface peat stratum.
on how planning conditions in specific localities, Proceedings Institution of Civil Engineers, Part 1
including Ripon, Rotherham, St Helens, Swansea 77, 111–136.
Brand, E.W. (1988) Special Lecture: Landslide risk
and Torbay can incorporate geomorphological con-
assessment in Hong Kong. In Bonnard, C. (ed.)
siderations. Guidance forms for conveyancing Landslides: Proceedings of the Fifth International
could include questions about land instability and Symposium on Landslides Volume 2, Rotterdam:
former land uses. However, the questions are often Balkema, 1059–1074.
general in nature, and solicitors may not always be British Geological Survey (1984) Review of Natural
Contamination in Great Britain. Report to the
aware of the full range of information to seek and
Department of the Environment.
of the full range of sources of information on Bromhead, E. N., and Ibsen, M.-L. (1998) Land use and
ground conditions. Urban engineering geomor- climate-change impacts on landslide hazards in SE
phology is thus more than identifying problems Britain. In Cruden, D. and Fells, R. (eds.) Landslide
and finding solutions. Asking questions about geo- Risk Assessment, Balkema, 65–176.
Brunsden, D. and Lee, E. M. (2004) Behaviour of
morphology must be part of the total business of
coastal landslide systems: an inter-disciplinary
urban development and management. view. Zeitschrift fur Geomorphologie 134, 1–112.
Cardinali, M., Reichenbach, P., Guzzetti, F., Ardizzone,
F., Antonini, G., Galli, M., Cacciano, M., Castellani,,
References M. and Salvati, P. (2002) A geomorphological
approach to the estimation of landslide hazards and
Ahleberg, P., Stigler, B. and Viberg, L. (1988) risks in Umbria, Central Italy. Natural Hazards and
Experiences of landslide risk considerations in land Earth System Sciences 2, 57–72.
use planning in Sweden. In Bonnard, C. (ed.) Carnec, C., Huré A., Ledoux, E., Raucoules, D. and
Landslides: Proceedings of the Fifth International Rivera, A. (2000) Mapping and modelling of major
Symposium on Landslides Volume 2, Rotterdam: urban subsidence on Mexico City from SAR inter-
Balkema, 1091–1096. ferometry, EGS 2000, 25–29 April, Nice.
GFE-27.qxd 6/27/05 2:26 PM Page 778
Charles, J. A. (1993) Building on fill: Geotechnical Halcrow, W. (1986) Rhondda Landslip Potential
Aspects. BRE Report. Assessment. Department of the Environment and
Coates, D. R. (1985) Geology and Society. New York: Welsh Office.
Chapman & Hall. Harbor, J. (1999) Engineering geomorphology at the
Cooke, R. U. and Doornkamp, J. C. (1990) cutting edge of land disturbance: erosion and sedi-
Geomorphology in Environmental Management: a ment control. Geomorphology 31, 247–263.
new introduction. (Second Edition), Oxford: Oxford Hutchinson, J. N. (1991) Periglacial and slope pro-
University Press. cesses. In Forster, A., Culshaw, M. G., Cripps, J. C.,
Dagorne, A. and Dars, R. (1999) Les risques naturels. Little, J. A. and Moon, C. F. (eds) Quaternary
Paris: Presses Universitaires de France. Engineering Geology, Geological Society Publishing,
Department of the Environment, Welsh Office (1990) London, 283–334.
Planning Policy Guidance: Development on Jones, J. A. A. (1981) The nature of soil piping: a review
Unstable Land PPG 14. of research. GeoBooks, Norwich.
DETR (Department of the Environment Transport and Jones, K. D. C. and Lee, E. M. (1994) Landsliding in
Regions) (1997) Digest of Environmental Statistics Great Britain. HMSO.
19, HMSO, London. Kaye, C. A. (1976) Beacon Hill end moraines, Boston:
DETR (Department of the Environment Transport New explanation of an important urban feature. In
and Regions) (1998) Land Instability, Flooding Coates, D. R. (ed.) Urban geomorphology:
and Hazard Assessment in Ripon. Environmental Geological Society of America Special Paper 174,
Geology in Land-Use Planning Case Study No.8, 7–20.
http://www.symonds-group.com/services/ environ_ Kimura, H. and Yamaguchi, Y. (2000) Detection of land-
eng/publications/casestudies/casestudy8.pdf slide areas using satellite radar interferometry.
Doornkamp, J. C. (1988) Applied earth science back- Photogrammetric Engineering and Remote Sensing
ground: Torbay. GSL Publications. 66(3), 337–344.
Durazo, J. and Farvolden, R. N. (1989) The groundwater Lee, E. M. (1994) The Investigation and Management of
regime of the Valley of Mexico from historic evidence Erosion, Deposition and Flooding in Great Britain.
and field observations. Journal of Hydrology 112, HMSO.
171–190. Lee, E. M. and Moore, R. (1989) Landsliding in and
Douglas, I. (1971) Dynamic equilibrium in applied geo- around Luccombe Village. HMSO.
morphology. Earth Sciences 5, 29–35. Lee, E. M. and Moore, R. (1991) Coastal landslip poten-
Endo, H. (2000) The behavior of a VLFS and an air- tial assessment: Isle of Wight Undercliff, Ventnor.
plane during takeoff/landing run in wave condition. Department of the Environment.
Marine Structures 13, 477–491. Leopold , L. B. (1968) Hydrology for urban land plan-
Fabre, G. and Monteil, M. (2001) Sur 1’hydrogéomor- ning: A guidebook on the hydrologic effects of urban
phologie d’un espace à forte anthropisation urbaine: land use. US Geological Survey Circular 554.
le site de Nîmes (Languedoc, France) du Pléistocène Maharaj, R. J. (1993) Landslide processes and land-
superior à 1’ Antiquité; impacts postérieurs. Comptes slide susceptibility analysis from an upland water-
Rendus de 1’ Académie des Sciences Series IIA–– shed: a case study from St. Andrew, Jamaica, West
Earth and Planetary Sciences 333(8), 435–440. Indies. Engineering Geology 34, 53–79.
Ferretti, A. Prati, C. and Rocca, F. (1999) Monitoring Lam S. K. (1992) Progress report: Quaternary
terrain deformations using multi-temporal SAR Geological Mapping of the Kuching City area,
Images, Proceedings of FRINGE’99, Liege Belgium, Sarawak. Proceedings of the 23rd Geological
November 1999, http://esrin.esa.it/fringe99 Conference: Technical Papers. Kuala Lumpur:
Gardner, T. (1754) Historical notes on Dunwich, Geological Survey of Malaysia, 96–107.
Blythburgh and Southwold. Leach, B. A. and Goodger, H. K. (1991) Building on
Geomorphological Services Limited (1986–1987) Derelict Land, CIRIA Special Publication, CIRIA,
Review of Landsliding. Report to the Department of London.
the Environment. Lee, E. M., Jones, D. K. C. and Brunsden, D. (2000)
Gupta, A. and Ahmad, R. (1999) Geomorphology and The Landslide Environment of Great Britain. In
the Urban tropics: building an interface between Landslides in Research, Theory and Practice,
research and usage. Geomorphology 31, 133–149. Bromhead, E., Dixon, N. and Ibsen, M-L. (eds)
Gutjahr, M. R., Tavares, R., Pereira, P. R. B. and Proc. of 8th Int. Symp. on Landslides, Cardiff,
Santora, J. (2000) Natural hazards vs man-made 2000, Thomas Telford, London, 2, 911–916.
hazards: landslides in the escsarpment of Serro de Marker, B. R. (1996) The role of the earth sciences
Mar, SP, Brazil. In Bromhead, E., Dixon, N. and in addressing urban resources and constraints.
Ibsen, M-L. (eds) Landslides in Research, Theory In McCall, G. J. H., De Mulder, E. F. J. and
and Practice, Proc. of 8th Int. Symp. on Landslides, Marker, B. R. (eds) Urban Geoscience, Rotterdam,
Cardiff, 2000, Thomas Telford, London, 2, 687–692. Balkema, 163–179.
GFE-27.qxd 6/27/05 2:26 PM Page 779
McCormick, W. V. III. (2002) Urban landslide hazards, Smith, R. (2002) The Great Blackheath Hole. Greenwich
Millbrae Ca. Landslide: Case History. Geological Industrial History 5(3), http://gihs.gold.ac.uk/
Society of America, Cordilleran Section - 98th Annual gihs25. html#blackhole
Meeting (May 13–15, 2002). http://gsa.confex.com/ Spencer, T. (1999) Coastal erosion. In Pacione, M. (ed.)
gsa/2002CD/finalprogram/abstract_34955.htm Applied Geography: Principles and Practice,
McDowell, P. W. (1975) Detection of clay filled sink- London, Routledge, 109–123.
holes in the chalk by geophysical methods. Quarterly Tebbutt, T. H. Y. (1998) Principles of Water Quality
Journal of Engineering Geology 8(4), 303–310 Control. Butterworth Heinemann, Oxford.
New Civil Engineer (2003) Keep on running. Emap Thompson, A., Hine, P. D., Greig, J. R. and Peach, D. W.
Construct Ltd., 31–33. (1996) Assessment of subsidence arising from gyp-
Pelli, C., Thornton, C. and Joseph, L. (1997) The world’s sum solution. Symonds Travers Morgan Report to
tallest buildings. Scientific American 277(6), 64–73. the Department of the Environment.
Pethick, J. (1996) Coastal slope development: tempo- Tricart, J. (1961a) Les modalités de la morphogénèse
ral and spatial periodicity in the Holderness Cliff dans le lit du Guil au cours de la crue de la mi-juin
Recession. In Anderson, M. G. and Brooks, S. M. 1957. International Association of Scientific
(eds.) Advances in Hillslope Processes 2, Hydrology Publication 53, 65–73.
897–917. Tricart, J. (1961b) L’ evolution du lit du Guil au cours
Péwé, T. I. (1990) Land subsidence and earth-fissure for- de la crue de juin 1957. Bulletin de la Section
mation caused by groundwater withdrawal in Arizona; Géographique, Comité des Travaux Historiques et
A review. In Higgins, C. G. and Coates, D. R. (eds) Scientifiques 72, 169–403.
Groundwater geomorphology; The role of subsurface USGS Las Vegas Urban Corridor Geologic Mapping
water in earth-surface processes and landforms. Project http://geology.wr.usgs.gov/wgmt/lasvegas/
Boulder, Colorado, Geological Society of America lvmap.html
Special Paper 252, 219–233. Valentin, H. (1954) Der landverlust in Holderness,
Pipkin, B. W. and Trent, D. D. (1994) Geology and the Ostengland von 1852 bis 1952. Die Erde 6, 296–315.
Environment (2nd edn), Belmont, California, Vestappen, H. Th. (1983) Applied Geomorphology,
Wadsworth. Amsterdam. Elsevier.
Polan, S.M. (1992) Solid waste problems in the Big Wimpey Environmental Limited (1994) Review of
Apple. Forum for Applied Research and Public Foundation Conditions. Report to the Department
Policy 7, 62–64. of the Environment.
Refice, A., Bovenga, F., Wasowski, J. and Guerriero, L.
(2000) Use of InSAR data for landslide monitor-
ing: A case study from southern Italy, Proceedings
of IGARSS’2000, Honolulu, Hawaii USA, July Further reading
2000.
Refice, A., Bovenga, F. and Wasowski, J. (2001) Bennet, M. R. and Doyle, P. (1997) Environmental
Monitoring landslide activity in a peri-urban area by Geology and the Human Environment. Wiley,
SAR interferometry. EGS Nice. http://www.coper- Chichester.
nicus.org/EGS/egsga/nice01/programme/abstracts/ Douglas, I. (1983) Urban Environments. Edward
aai4893.pdf Arnold, London.
Skempton, A.W. and Weeks, A.G. (1976) The Eyles, N. (1997) (Ed) Environmental Geology of Urban
Quaternary history of the Lower Greensand escarp- Areas. Geological Association of Canada, St. Johns.
ment and Weald Clay vale near Sevenoaks, Kent. Gupta, A. and Pitts, J. (1992) Physical Adjustments in a
Philosophical Transactions of the Royal Society, Changing Landscape: The Singapore Story.
London A283, 493–526. Singapore University Press, Singapore.
GFE-27.qxd 6/27/05 2:26 PM Page 780
GFE-Appendix.qxd 6/27/05 4:48 PM Page 781
Appendices
Appendix A1: Visual Identification
Summary of contents Table A1.3 The Beaufort scale of wind force

Table A1.4 Earthquake intensity with equivalent
Table A1.1 Estimating percentage cover or damage
composition Table A1.5 Roughness profiles and Joint
Table A1.2 Visual assessment of pebble roundness Roughness Coefficients (JRCs)
782
APPENDICES 783
Table A1.1 Chart for estimating percentage cover or composition. From Gardiner and Dackombe (1983) after
Hodgson (1974).
Table A1.2 Images for the visual assessment of pebble roundness. From Gardiner and Dackombe (1983), after
Krumbein (1941).
APPENDICES 785
Table A1.3 The Beaufort scale of wind force.
Beaufort General Approximate Approximate Specification

number description wind speeds wave height
of wind (msⴚ1) (m) For coastal use For inland use
0 Calm 0.0–0.2 0 Sea like a mirror Smoke rises vertically

1 Light air 0.3–1.5 0.1–0.2 Ripples without Wind direction shown
appearance of by smoke drift but not
scales; no foam by wind vanes
crests
2 Light breeze 1.6–3.3 0.3–0.5 Small wavelets; Wind felt on face; leaves
crests have glassy rustle, ordinary vane
appearance but do moved by wind
not break
3 Gentle breeze 3.4–5.4 0.6–1.0 Large wavelets; crests Leaves and small twigs
begin to break; in constant motion;
scattered white wind extends light flag
horses
4 Moderate 5.5–7.9 1.5 Small waves Raises dust and loose
breeze becoming longer; paper, small branches
fairly frequent are moved
white horses
5 Fresh breeze 8.0–10.7 2.0 Moderate waves; Small trees in leaf
many white horses begin to sway; crested
and chance of wavelets form on
some spray inland water
6 Strong breeze 10.8–13.8 3.5 Large waves begin Large branches in
to form; white foam motion; whistling
crests extensive heard in telegraph
everywhere and wires; umbrellas
spray probable used with difficulty
7 Near or 13.9–17.1 5.0 Sea heaps up and Whole trees in motion;
moderate white foam from inconvenience felt
gale breaking waves when walking against
begins to be blown wind
in streaks; spindrift
begins to be seen
8 Gale or 17.2–20.7 7.5 Moderate high waves Breaks twigs off trees;
fresh gale of greater length; generally impedes
foam blown in well progress
marked streaks;
edges of crests
break into spindrift
9 Strong gale 20.8–24.4 9.5 High waves; crests Slight structural
begin to topple and damage occurs;
roll over; spray may chimney pots and
affect visibility slates removed
10 Storm or 24.5–28.4 12.0 Very high waves; long Seldom experienced
whole gale overhanging crests, inland; trees
tumbling of sea uprooted; considerable
becomes very heavy structural damage
and shocking; sea occurs
Table A1.3 (Continued).
Beaufort General Approximate Approximate Specification

number description wind speeds wave height
of wind (ms-1) (m) For coastal use For inland use
surface takes on
white appearance as
foam in large patches
is blown in very
dense streaks
11 Violent storm 28.5–32.7 15.0 Exceptionally high Very rarely experienced;
waves, sea covered widespread damage
by long patches of
foam; small and
medium-sized ships
might be lost to view
behind waves for
long periods
12 Hurricane 32.7 15 Air filled with foam
and spray; sea
completely white
with driving spray;
visibility greatly
reduced
Table A1.4 Earthquake intensity (after Waltham, 1994).
Modified Sensation Damage Area of influence, Equivalent

Mercalli distances from epicentre Richter*
intensity Magnitude
I Not Felt
II Felt at rest Social
III Felt indoors disturbance, Limited 1 to 3
IV Windows rattle no damage
V Felt outdoors
VI Frightening Slight damage Slight damage to 10 km away 4
VII Severe damage to adobe Damage to 10 km away 5
houses, little damage
to reinforced concrete
VIII Masonry damaged Damage to 50 km away
IX Foundations damaged, severe Damage to 200 km away 6
damage to many buildings
X Major damage to or Severe damage to 20 km away 7
destruction of
most buildings
XI Railways buckled
XII Total destruction 8
* Richter magnitude is given by log10 of the maximum wave amplitude in microns measured on a Wood Andersons
seismograph 100 km from the epicentre.
APPENDICES 787
Table A1.5 Roughness profiles and corresponding range of Joint Roughness Coefficient (JRC) associated with
each (ISRM, 1981). (JRC is a measure of the relative roughness of surfaces in rock joints).
References Description of Discontinuities in Rock Masses.

(ed. E.T. Brown), Pergamon Press, Oxford.
Gardiner, V. and Dackombe, R. V. (1983) Geomor- Krumbein, W. C. (1941) Measurement and geological
phological Field Manual. George Allen and Unwin, significance of shape and roundness of sedimentary
London. partices. J. Sed. Petrol. II, 64–72.
Hodgson, J.M. (ed.) (1974) Soil Survey Field Handbook. Waltham, A. C. (1994) Foundations of Engineering
Soil Survey Technical Monograph 5, Rothamsted Geology. Blackie, Glasgow, UK.
Experimental Station.
ISRM (International Society for Rock Mechanics)
(1981) Suggested Methods for the Quantitative
Appendix A2: Identification, description and

classification of soils
Summary of contents Table A2.4 BS soil classification system

(BS5930:1981)
Table A2.1 Flowchart for identification and Table A2.5 Unified soil classification system
description of soils (field tests)
Table A2.2 Identification and description of soils Table A2.6 Unified soil classification system
(BS5930:1999) (laboratory tests)
Table A2.3 Description of composite soil types Tables A2.7, A2.8, A2.9 AASHTO/ASTM soil
for very coarse soils (BS5930:1999) classification system
789
GFE-Appendix.qxd
790
Table A2.1 Flowchart for identification and description of soils (BS5930: 1999).

Remove cobbles and
boulders (>60 mm)
6/27/05
Do they weight No Is soil dark or black, Has soil been laid
more than rest of and of low density? down by man?
soil?
4:48 PM
Does soil stick
together when
Yes No wet? Yes
Does soil display low Yes
Page 790
Yes
plasticity, diltatancy,
Are most particles > 200 mm? Are most particles > 2 mm? silky touch, distintegrate
in water and dry quickly?
Yes No Yes No Yes No
BOULDERS COBBLES GRAVEL SAND SILT CLAY ORGANIC Made ground
Describe secondary and Describe secondary and Describe secondary and Describe according to Does soil comprise narural
tertiary size fractions tertiary size fractions tertiary size fractions 41.4.6 of BS5930 or man-made materials?
Natural Man-made
Describe density (by inspection) Describe density, describe bedding, Describe strength/compactness, Describe as Describe proportion,
describe bedding, describe describe colour, measure particle describe discontinuties, describe for natural type and condition
colour, measure particle size, size, describe particle shape and bedding, describe colour soils of components
describe particle shape grading
Add other information, Add other information (such as organic content, Distinguish between fill (controlled
add geological origin calcareous content) and minor constituents placement) and made ground
(uncontrolled placement)
Replace cobbles and boulders into desciption,

add geological origin
GFE-Appendix.qxd
Table A2.2 Identification and description of soils (BS5930: 1999).
Soil Density/compactness/strength Composite soil types Particle Particle PRINCIPAL

group Discontinuities Bedding Colour (mixtures of basic shape size SOIL TYPE
Term Field test soil types)
Loose By inspection of voids Scale of spacing of Scale of bedding thickness For mixtures involving BUILDERS
Very coarse soils
6/27/05
and particle packing discontinuities very coarse soils,
200
see Table A2.3 Angular
Red COBBLES
Dense Term Mean Term Mean Orange Term Approx, %c) 60
spacing mm thickness mm Yellow secondary Sub-angular
4:48 PM
Brown Coarse
Coarse soils (over about 65% sand and gravel sizes)
Borehole with SPT N-value Very Over Very thickly Over 2000 Green Slightly 5 Sub-rounded 20
widely 2000 bedded Blue (sandyd)) Rounded GRAVEL
Very 0–4 Widely 2000 to Thickly 2000 to White Medium
Page 791
loose 600 bedded 600 Cream Flat
Loose 4–10 Medium 600 to Medium 600 to Grey (sandyd)) 5 to 20b) 6
200 bedded 200 Black Tabular
Medium 10–30 Closely 200 to 60 Thinly 200 to 60 etc. Fine
dense bedded Elongated
Dense 30–50 Very 60 to 20 Very thinly 60 to 20 Very 20b) 2
closely bedded (sandyd)) Minor
constituent Coarse
Very dense 50 Extremely Under 20 Thickly 20 to 6 SAND
closely laminated type 0.6
Calcareous,
Medium
Breaks into Thinly Under 6 shelly,
Slightly Visual examination: Fissured blocks along laminated SAND AND about 50b) glauconitic, 0.2
cemented pick removes soil in unpolished GRAVEL micaceous
lumps which can be discontinuities etc. using
abraded terms such as
Un-compact Easily moulded or Sheared Breaks into Inter- Alternating Term Approx. %c) Slightly Fine SILT
crushed in the blocks along bedded layers of Light secondary calcareous,
fingers polished different types 0.06
discontinuities Prequalified by Dark calcareous,
thickness term Coarse
Compact Can be moulded or Inter- if in equal Mottled Slightly 35
crushed by strong laminated proportions. (sandye)) 0.02
pressure in the Otherwise
fingers thickness of and
GFE-Appendix.qxd
Soil Density/compactness/strength Composite soil types Particle Particle PRINCIPAL
group Discontinuities Bedding Colour (mixtures of basic shape size SOIL TYPE
Term Field test soil types)
Very soft Finger easily Spacing terms also spacing between very Medium
6/27/05
Fine soils (over about 35% silt and clay sizes)
0–20 pushed in used for distance subordinate calcareous.

(KPa) up to 25 mm between partings, layers defined 0.006
Soft Finger pushed isolated beds
20–40 in up to 10 mm or laminae, Fine
4:48 PM
Firm Thumb makes dessication cracks, (sandye)) 35 to 65a)
40–75 impression easily rootlets etc. 0.002 CLAY/SILT
Stiff Can be indented % defined
75–150 slightly by thumb on a site or
Very stiff Can be indented material
Page 792
150–300 by thumb nail specific CLAY
Hard (or Can be scratched by Very 65a) basis or
very weak thumbnail (sandyf)) subjective
mudstone see 41.2.2
Cu300 kPa of BS5930
Firm Fibres already Fibrous Plant remains Transported mixtures Colour Contains finely divided or discrete particles of organic
compressed together recognizable and Slightly organic clay or silt Grey matter, often with distinctive smell, may oxidize rapidly.
retains some Slightly organic sand as mineral Describe as for inorganic soils using terminology above.
strength Organic clay or silt Dark grey
Organic soils
Organic sand Dark grey

Spongy Very compressible Pseudo- Plant remains Very organic clay or silt Black
and open structure fibrous recognizable, Very organic sand Black
strength
Accumulated in situ
lost Predominantly plant remains, usually dark brown or black in colour,
Plastic Can be moulded in Amorphous Recognizable plant Peat distinctive smell, low bulk density. Can contain disseminated or discrete
hand and smears remains absent mineral soils
fingers
GFE-Appendix.qxd
PRINCIPAL SOIL TYPE Visual identification Minor constituents Example Stratum name Example descriptions
BOULDERS Only seen complete in pits or exposures

COBBLES Often difficult to recover whole from Shell fragments, pockets of Loose brown very sandy sub-angular fine to
boreholes peat, gypsum crystals, flint coarse flint GRAVEL with small pockets
gravel, fragments of brick, (up to 30 mm) of clay. (TERRACE GRAVELS)
6/27/05
rootlets, plastic bags etc.
GRAVEL Easily visible to naked eye; particle shape using terms such as: RECENT DEPOSITS, Medium dense light brown gravelly clayey
can be described; grading can be with rare line SAND. Gravel is fine (GLACIAL
4:48 PM
described ALLUVIUM, DEPOSITS)
with occasional
with abundant/frequent/ WEATHERED
numerous BRACKLESHAM CLAY,
SAND Visible to naked eye; Stiff very closely sheared orange mottled
Page 793
no cohesion when dry; % defined on a site or material LIAS CLAY, brown slightly gravelly CLAY. Gravel is fine
grading can be described. specific basis or subjective EMBANKMENT FILL and medium of rounded quartzite.
(REWORKED WEATHERED LONDON CLAY)
SILT Only coarse silt visible with hand lens; TOPSOIL,
exhibits little plasticity and marked
dilatancy; slightly granular or silky to
the touch: disintegrates in water,
lumps dry quickly; possesses cohesion MADE GROUND OR
but can be powdered easily GLACIAL
between fingers DEPOSITS? etc.
CLAY/ Intermediate in behaviour between clay Firm thinly laminated grey CLAY with closely
SILT and silt. Slightly dilatant spaced thick laminae of sand. (ALLUVIUM)
Dry lumps can be broken but not powdered Plastic brown clayey amorphous PEAT.
between the fingers; they also disintegrate (RECENT DEPOSITS)
under water but more slowly than silt;
CLAY smooth to the touch; exhibits plasticity
but no dilatancy; sticks to the fingers and
dries slowly; shrinks appreciably on
drying usually showing cracks.
Notes:
a d
Or described as coarse soil depending on mass behaviour. Gravelly or sandy and/or silty or clayey.
b e
Or described as fine soil depending on mass behaviour. Gravelly and/or sandy.
c f
% coarse or fine soil type assessed excluding cobbles and boulders. Gravelly or sandy.
Table A2.3 Description of composite soil types for very coarse soils (BS5930: 1999)
Main name Estimated boulder or cobble

content of very coarse fraction
Over 50% of BOULDERS Over 50% is of boulder

material is very size ( 200 mm)
coarse ( 60 mm) COBBLES Over 50% is of cobble size
(200 mm to 60 mm)
Term Composition
BOULDERS (or COBBLES) with a little finer materiala up to 5% finer material

BOULDERS (or COBBLES) with some finer materiala 5% to 20% finer material
BOULDERS (or COBBLES) with much finer materiala 20% to 50% finer material
FINER MATERIALa with many boulders (or cobbles) 50% to 20% boulders (or cobbles)
FINER MATERIALa with some boulders (or cobbles) 20% to 5% boulders (or cobbles)
FINER MATERIALa with occasional boulders (or cobbles) up to 5% boulders (or cobbles)
a
The description of ‘finer material’ is made in accordance with 41.4.2 to 41.4.6 of BS5930, ignoring the very
coarse fraction; the principal soil type name of the finer material may also be given in capital letters, e.g. sandy
GRAVEL with occasional boulders, COBBLES with some sandy CLAY.
Term Principal soil type Approximate proportion of
secondary constituent
slightly sandy or gravelly SAND up to 5%

sandy or gravelly or 5% to 20%
very sandy or gravelly GRAVEL over 20%
SAND and GRAVEL about equal proportions
GFE-Appendix.qxd
Table A 2.4 BS soil classification system (BS5930: 1981).
Soil groups Sub-groups Subdivisions Liquid Fines (%)

Limit (%) 0.06 mm)
GRAVELS ( 50% of coarse material

Slightly silty G Well graded GW 0–5
or clayey GRAVEL
6/27/05
GRAVEL Poorly graded GP Uniformly graded GPu
is of gravel size 2 mm) GRAVEL Gap graded GPg
Silty or clayey G–F Silty GRAVEL G-M Well graded GWM 5–15
4:48 PM
GRAVEL Poorly graded GPM
Clayey GRAVEL G-C Well graded GWC
Poorly graded CPC
Page 795
Very silty or GF Very silty GM As for GC GML etc. As GC
clayey GRAVEL
Coarse soils ( 35% fines)
GRAVEL Very clayey GC Low plasticity clay GCL 35

GRAVEL Intermediate ” ” GCI 35–50 15–35
High ” ” GCH 50–70
Very high ” ” GCV 70–90
Extremely high ” ” GCE 90
Slightly silty or S Well graded SW –
clayey SAND SAND
SANDS ( 50% of coarse material
Poorly graded SP Uniformly graded SPu 0–5

SAND Gap graded SPg
Silty or clayey S–F Silty SAND S-M Well graded SWM
is of sand size 2 mm)
SAND Poorly graded SPM 5–15

Clayey SAND S-C Well graded SWC
Poorly graded SPC
Very silty or SF Very silty SAND SM As for SC SML etc. As SC
clayey SAND
APPENDICES 795
Very clayey SC Low plasticity clay SCL 35
SAND Intermediate ” ” SCI 35–50 15–35
High ” ” SCH 50–70
Very high ” ” SCV 70–90
Extremely high ” ” SCE 90
GFE-Appendix.qxd
Table A 2.4 (Continued).
796
Soil groups Sub-groups Subdivisions Liquid Fines (%)

Limit (%) 0.06 mm)
Gravelly SILT FG Gravelly SILT MG As for CG MLG etc. As CG
6/27/05
or gravelly Gravelly CLAY CG Low plasticity ” CLG 35
SILTS and CLAYS CLAY* Intermediate ” CIG 35–50
Gravelly or sandy
High ” CHG 50–70

(35–65% fines)
4:48 PM
Very high ” CVG 70–90
Extremely high ” CEG 90
Fine soils ( 35% fines)
Sandy SILT or FS Sandy SILT MS As for CG MLS etc. As CG
Page 796
sandy CLAY* Sandy CLAY CS As for CG CLS etc. As CG
SILT or CLAY F SILT M As for C ML etc. As C
( 65% fines)
CLAY C Low plasticity CL 35

SILTS and
Intermediate ” CI 35–50
CLAYS
High ” CH 50–70
Very high ” CV 70–90
Extremely high CE 90
ORGANIC SOILS Letter ‘O’ suffixed to any group or sub-group symbol.
e.g. MHO organic silt of high plasticity.
PEAT Pt (Peat) soils consist predominantly of plant remains (fibrous or amorphous).
* GRAVELLY if 50% coarse material is gravel sized; SANDY if 50% is sand sized.
Notes:
1
It should always be made clear whether the group classification is based on laboratory test results or simply on inspection of the soil.
2
Silt (M-soil), M is material which plots below the A-line on a standard plasticity chart (see Table A2.6). Fine soils of this type include micaceous and diatoma-
ceous soils, pumice and volcanic soils, and soils containing halloysite, in addition to silt-sized soils and rock flour. The alternative term ‘M-soil’ is sometimes
used to describe this group of soils, to avoid confusion with purely silt-sized material.
3
The soil classification applies to material of gravel, sand, silt and clay sizes. Larger particles must be removed before classification and their proportions
recorded separately.
4
Material is generally considered to be uniformly graded if it has a uniformity coefficient of less than 6 (see Table A2.6).
APPENDICES 797
Table A2.5 Unified soil classification system (field observation), after Carter (1983).
Soil Groups Description and Subgroups Symbol

Identification
Boulders and Soils consisting chiefly Boulder gravels.

cobbles of boulders ( 200 mm)
or cobbles (75–200 m).
Identifiable by visual
inspection.
Gravel and Soils with an appreciable Well-graded gravel–sand mixtures, GW
gravelly soils gravel content (2–75 mm). little or no fines
Generally easily identifiable Well-graded gravel–sands with GC
Coarse grained soils
by visual inspection. small clay content

A medium to high dry Uniform gravel with little or no fines GU
strength indicates some clay Poorly-graded gravel–sand mixtures GP
is present. A negligible dry with little or no fines
strength indicates the Gravel–sand mixtures with excess GF
absence of clay. of lines
Sands and Soils with an appreciable Well-graded sands and gravelly SW
sandy soils sand content (0.06–2 mm). sands, little or no fines
Majority of particles can be Well-graded sand with small clay SC
distinguished by eye. Feel content
gritty when rubbed between Uniform sands with little or no fines SU
fingers. A medium to high Poorly-graded sands, little or no fines SP
dry strength indicates some Sands with excess of fines SF
clay is present.
Fine-grained Soils of low plasticity: cannot Silts (inorganic), rock flour, silty ML
soils having low be readily rolled into threads fine sands with slight plasticity
plasticity (silts) when moist. Do not feel gritty. Clayey silts (inorganic) CL
Exhibit dilatancy*. Organic silts of low plasticity OL
Fine-graiined soils
Fine-grained Can be readily rolled into Silty clays (inorganic) and MI

soils having threads when moist. Do not sandy clays
medium exhibit dilatancy*. Show some Clays (inorganic) of medium CI
plasticity shrinkage on drying plasticity
Organic clays of medium plasticity OI
Fine-grained Can be readily rolled into Highly compressible micaceous or MH
soils having threads when moist. Greasy diatomaccous soils
high plasticity to touch. Considerable Clays (inorganic) of high plasticity CH
shrinkage on drying. Organic clays of high plasticity OH
All highly compressible
Fibrous organic Usually brown or black in Peat and other highly organic Pt
soils with very colour. Very compressible. swamp soils.
high compressibility Easily identifiable visually.
* Dilatancy: If a pat of wet silt is shaken in the hand, water will appear on the surface and, if the pat is then pressed,
the water will retreat into the silt, leaving a matt surface, because of an increase in volume. This property
is known as dilatancy and is useful for distinguishing silt from clay. Fine sands also exhibit dilatancy but can be
distinguished from silts by their particle size.
Group symbols
Main soil types Sub divisions
G Gravel W Well graded
S Sand U Uniformly graded
M Silt P Poorly graded
C Clay C With clay
O Organic silts and clays F With fines
Pt Peat L Low plasticity
I Intermediate plasticity
H High plasticity
GFE-Appendix.qxd
6/27/05
Table A2.6 Unified soil classification system (laboratory test), after Carter (1983).
Soil groups Laboratory classification criteria Symbol
4:48 PM
Appreciable number of cobbles or boulders; coarser than 75 mm –
Gravels: more than
COARSE-GRAINED SOILS; more than half
D60 (D30 )2
Borderline soils (5–12% fines) require dual

half the coarse Less than 5% of whole GW
U 4 and S between 1 and 3
fraction is coarser sample finer than 0.075 mm D10 D10 D60
Page 799
than 2 mm
of material is coarser than 0.075 mm
Does not meet both the above One size predominates GU

requirements.
More than 12% of whole One size lacking GP
sample finer than 0.075 mm Below ‘A’ line or PI 4. Above ‘A’ line and PI between 4 GF
Above ‘A’ line and PI 7. and 7 is borderline and GC
requires dual symbols
Sands: more than Less than 5% of whole D60 (D30 )2

U 6 and S between 1 and 3
half the coarse sample finer than 0.075 mm D10 D10 D60 SW
symbols
fraction is finer
than 2 mm Does not meet both the above One size predominates SU
requirements.
One size lacking SP
More than 12% of whole Below ‘A’ line or PI 4. Above ‘A’ line and PI between 4 SF
sample finer than 0.075 mm Above ‘A’ line and PI 7. and 7 is borderline and requires SC
dual symbols
APPENDICES 799
GFE-Appendix.qxd
800
Soil groups Laboratory classification criteria Symbol
6/27/05
FINE-GRAINED SOILS: more han
Fine-grained soils are classified by

reference to the plasticity chart,
carried out on the fraction finer

using the results of index tests
4:48 PM
half of material is finer than
Page 800
than 0.425 mm.
0.075 mm.
* The division of the high plasticity range into very high and
extremely high applies to the British Standard classification
system only.
1
D60 is the ‘60 per cent size’ — the maximum size of the smallest 60% of the material. D30 and D10 are similarly defined. D10 is also called the ‘effective size’.
2
U is the uniformity coefficient; PI is the plasticity index.
3
Borderline soils––those possessing characteristics of two groups––are designated by combinations of symbols. For example, GW-GC; well-graded gravel-
sand mixture with clay binder.
4
The soil classification applies to material of gravel, sand, silt and clays sizes. Large particles must be removed before classification and the proportions
recorded separately.
APPENDICES 801
Table A2.7 AASHTO/ASTM soil classification, after Carter (1983).
Soils are classified into seven groups, based on the Classification procedure
results of grading and plasticity tests. The evaluation
of soils within each group is made by means of a With the required test data available, proceed from left
‘group index’ which is a value calculated from an to right in Table A2.8, and the correct group will be
empirical formula. The group classification should found by process of elimination. The first group from
be useful in determining the relative quality of the the left into which the test data will fit is the correct
soil for use in earthwork structures, particularly classification. For purposes of classification, all test
embankments, subgrades, sub-bases and bases. values should be shown as whole numbers.
The classification and group index is obtained by The group index is obtained by summing the two
using the test results and the tables and charts given in values obtained rounding to the nearest whole number.
Table A2.8. The basic classification is obtained by Group Index
using Table A2.8. If a more detailed classification is
desired, a further subdivision of the groups may be The group index is calculated from the following
made. Table A2.8 gives a classification with suggested formula:
subgroups.
Group index (F – 35) {0.2 0.005
The group index may be calculated using the
(LL – 40)} 0.01 (F – 15) (PI – 10)
empirical formula given below but is usually estimated
where F is the percentage passing the 0.075 mm sieve,
using the figures in Table A2.8. The group index is
expressed as a whole number. This percentage is
given in parenthesis after the group symbol and is
based only on the material passing the 75 mm sieve.
quoted to the nearest whole number (e.g. A-2-6 (3),
LL is the liquid limit.
A-4 (5)). The group number for A-I-a, A-I-b, A-3,
PI is the plasticity index.
A-2-4 and A-2-5 will always be zero so need not be
calculated or quoted. Notes:
1
When the calculated group index is negative, the
Descriptions and suitability of materials
result is expressed as zero.
2
Descriptions of the typical materials found in each The group index is reported to the nearest whole
group are given in Table A2.9. number.
3
In general, the lower the classification symbol and Table A2.8 may be used to estimate the group index.
the lower the group index of a soil within a given
classification, the higher will be the supporting value
of the soil as a subgrade.
Table A2.8 AASHTO/ASTM classification tables and group index charts, after AASHTO (1974).
General classification Granular materials Silt-clay materials

(35% or less passing 75 ␮m) (more than 35% passing 75␮m)
1 Classification of soils into main groups

General classification A-1 A-3* A-2 A-4 A-5 A-6 A-7
Sieve analysis, % passing:
2 mm – – – – – – –
0.425 mm 50 max 51 min – – – – –
0.075 mm 25 max 10 max 35 max 36 min 36 min 36 min 36 min
Fraction passing 0.425 mm:
Liquid limit – – – 40 max 41 min 40 max 41 min
Plasticity index 6 max N.P. – 10 max 10 max 11 min 11 min
General rating as subgrade Excellent to good Fair to poor
* The placing of A-3 before A-2 is necessary in the ‘left to right elimination process’ and does not indicate superi-
ority of A-3 over A-2.
General Granular materials Silty-clay materials

classification (35% or less passing 75 ␮m) 35% passing 75 ␮m)
(
2 More detailed classification of soils into groups and subgroups

Group A-1 A-3 A-2 A-4 A-5 A-6 A-7
classification A-1-a A-1-b A-2-4 A-2-5 A-2-6 A-2-7 A-7-5
A-7-6
Sieve analysis,
% passing:
2 mm 50 max – – –
0.425 mm 30 max 50 max 51 max – – – – –
0.075 mm 10 max 25 max 10 max 35 max 35 max 35 max 35 max 36 min 36 min 36 min 36 min
Fraction passing
0.425 mm:
Liquid limit – 40 max 41 min 40 max 41 min 40 max 41 min 40 max 41 min
Plasticity index 6 max N.P. 10 max 10 max 11 min 11 min 10 max 10 max 11 min 11 min
Usual types of Stone Fine Silty or clayey gravel Silty soils Clayey soils
significant fragments, sand and sand
constituents gravel, sand
General rating Excellent to good Fair to poor
as subgrade
* PI of A-7-5 subgroup (LL–30); PI of A-7-6 subgroup (LL–30).
Classification A-8
Highly organic soils such as peat may be classified in an A-8 group. Classification of these materials is based on
visual inspection and is not dependent on the grading or plasticity. The materials are composed primarily of decayed
organic matter and have a fibrous texture, dark brown or back colour and a smell of decay.
They are unsuitable for use in embankments or subgrades.
APPENDICES 803
Table A2.9 Description of soil types in AASHTO/ASTM groups after AASHTO (1974).
Classification of materials in the various groups applies only to the fraction passing the 75 mm sieve. The propor-
tions of boulder and cobble sized particles should be recorded separately and any specification regarding the use
of A-1, A-2 or A-3 materials in construction should state whether boulders are permitted.
Granular materials Silty–clay materials

Group A-1. Typically a well-graded mixture of stone Group A-4. Typically a non-plastic or moderately
fragments or gravel, coarse to fine sand and a plastic silty soil usually with a high
non-plastic or feebly plastic soil binder. percentage passing the 0.075 mm sieve. The
However, this group also includes stone group also includes mixtures of silty fine
fragments, gravel, coarse sand, volcanic sands and silty gravelly sands.
cinders, etc. without soil binder.
Group A-5. Similar to material described under
Subgroup A-1-a is predominantly stone fragments
group A-4 except that it is usually diatoma
or gravel, with or without binder.
ceous or micaceous and may be elastic as
Subgroup A-1-b is predominantly coarse sand
indicated by the high liquid limit.
with or without binder.
Group A-6. Typically a plastic clay soil having a high
Group A-3. Typically fine beach sand or desert sand
percentage passing the 0.075 mm sieve. Also
without silty or clayey fines or with a very
mixtures of clayey soil with sand and fine
small proportion of nonplastic silt. The group
gravel. Materials in this group have a high
also includes stream-deposited mixtures of
volume change between wet and dry states.
poorly graded fine sand with limited amounts
of coarse sand and gravel. Group A-7. Similar to material described under
group A-6 except that it has the high liquid limit
Group A-2. Includes a wide variety of ‘granular’
characteristics of group A-5 and may be elastic as
materials which are borderline between the
well as subject to volume change.
granular A-1 and A-3 groups and the
Subgroup A-7-5 materials have moderate plasticity
silty–clay materials of groups A-4 to A-7. It
indices in relation to the liquid limits and may
includes all materials with not more than 35%
be highly elastic as well as subject to volume
fines which are too plastic or have too many
change.
fines to be classified as A-1 or A-3.
Subgroup A-7-6 material have high plasticity
Subgroups A-2-4 and A-2-5 include various
indices in relation to the liquid limits and are
granular materials whose finer particles
subject to extremely high volume change.
(0.425 mm down) have the characteristics of
the A-4 and A-5 groups, respectively. Group A-8. Includes highly organic materials.
Subgroups A-2-6 and A-2-7 are similar to those Classification of these materials is based on visual
described above but whose finer particles have inspection and is not related to grading or plasticity.
the characteristics of A-6 and A-7 groups,
respectively.
References BS5930: 1999 Code of practice for site investigations.

British Standards Institution, London.
AASHTO (1974) Specification M145–73. The classifi- Carter, M. (1983) Geotechnical Engineering Handbook.
cation of soils and soil–aggregate mixtures for high- Pentech Press.
way construction purposes. American Association of
State Highway and Transport Officials, Washington.
BS5930: 1981 Code of practice for site investigations
(formerly CP 2001). British Standards Institution,
London.
Appendix A3: Identification, Description

and Classification of rocks
Summary of contents
Quantitative description of Table No.

rock masses
1. Rock material a. Rock type A3.1
description b. Wall strength A3.2
c. Weathering A3.3
2. Discontinuity BS5930 checklist A3.4
description
d. Type A3.5
e. Orientation A3.4
f. Roughness A3.4, A3.6,
A1.5 (A1)
g. Aperture A3.4, A3.7
3. Infilling h. Infilling type/width A3.4, A3.8
4. Rock mass i. Spacing A3.4, A3.9
description j. Persistence A3.4,
A3.10
k. Number of sets
l. Block size and shape A3.11
5. Groundwater m. Seepage A3.4,
A3.12
Table numbers refer to the following tables which give scales and descriptive
terms of each aspect of rock mass properties. In many cases the terms or numer-
ical scales suggested by the International Society for Rock Mechanics (ISRM,
1981) and British Standard Code of Practice for Site Investigation (1999) dif-
fer. It should therefore always be made clear which system is being used, and as
far as possible systems should not be mixed. Table A3.13 provides a rock mass
strength classification system for geomorphic purposes.
804
GFE-Appendix.qxd
Table A3.1 Aid to identification of rocks for engineering purposes (BS5930: 1999).
Grain size Bedded rocks (mostly sedimentary)
(mm)
Grain size At least 50% of grains At least 50% of grains
description are of carbonate are of grained volcanic rock
6/27/05
20 CONGLOMERATE Fragments of volcanic SALINE
RUDACEOUS
Rounded boulders, cobbles ejecta in a finer matrix. ROCKS
LIMESTONE and DOLOMITE (undifferentiated)

and gravel cemented Calcirudite Rounded grains
6 in a finer matrix AGGLOMERATE HALITE
Grain size boundaries approximate
4:48 PM
Breccia Irregular rock Angular grains
fragments in a finer matrix VOLCANIC BRECCIA ANHYDRITE
2
SANDSTONE Cemented volcanic GYPSUM
Coarse
Page 805
Angular or rounded grains ash
ARENACEOUS
commonly cemented by clay,

0.6 calcitic or iron minerals
Medium
Quartzite
Quartz grains and siliceous cement Calcarenite TUFF
0.2 Arkose
Many feldspar grains
Fine
Greywacke
Many rock chips
0.06
SILTSTONE Calcisiltite
ARGILLA-
Calcareous
mudstone
CEOUS
MUD- Mostly silt Fine-grained TUFF
CHALK
STONE
0.002 Calcilutite Very fine-grained TUFF
Amorphous Flint: occurs as bands of nodules in the Chalk COAL
or crypto Chert: occurs as nodules and beds in LIGNITE
crystalline lime stone and calcareous sandstone
Granular cemented - except amorphous rocks
APPENDICES 805
SILICEOUS CALCAREOUS SILICEOUS CARBONACEOUS
Sedimentary Rocks: Granular cemented rocks vary greatly in strength, some sandstones are stronger than many igneous rocks. Bedding may not show in hand
specimens and is best seen in outcrop. Only sedimentary rocks, and some metamorphic rocks derived from them, contain fossils.
Calcareous rocks contain calcite (calcium carbonate) which effervesces with dilute hydrochloric acid.
GFE-Appendix.qxd
Table A3.1b (Continued).
806
Igneous rocks: generally massive structure and crystalline texture

Metamorphic Rocks
Grain size
description Foliated Massive
6/27/05
GRANITE1 DIORITE1,2 GABBRO1,2 Pyroxenite GNEISS MARBLE
COARSE These rocks are sometimes porphyritic Well developed but often widely
QUARTZITE
and are then described, for example, spaced foliation sometimes with
4:48 PM
as porphyritic granite Peridotite schistose bands GRANULITE
Migmatite
Irregularly foliated; mixed schists HORNFELS
MICROGRANITE1 MICROIORITE1,2 DOLERITE3,4 and gneisses
MEDIUM AMPHIBOLITE
Page 806
These rocks are sometimes porphyritic SCHIST
increasing grain size
and are then described as porphyries Well developed undulose SERPENTINE

foliation; generally much mica
PHYLLITE
RHYOLITE4,5 ANDESITE4,5 BASALT5 Slightly undulose foliation;
FINE These rocks are sometimes porphyritic sometimes spotted
and are then described as porphyries SLATE Well developed plane
cieavage (foliation)
Amorphous MYLONITE
Crypto- OBSIDIAN5 VOLCANIC GLASS Found in fault zones, mainly in
crystalline igneous and metamorphic areas
Colour CRYSTALLINE
Pale Dark
ACID INTERMEDIATE BASIC Little or ULTRA Mainly
Much quartz Some quartz no quartz SILICEOUS SILICEOUS
BASIC
Igneous Rocks: Composed of closely interlocking mineral grains. Strong when fresh; not porous.
Mode of occurrence: 1. Batholiths, 2. Laccoliths, 3. Silts, 4. Dykes, 5. Lava flows, 6. Veins.
Metamorphic Rocks: Generally classified according to fabric and mineralogy rather than grain size. Most metamorphic rocks are distinguished by foliation
which may impart fissility. Foliation in gneisses is best observed in outcrop. Non-foliated metamorphics are difficult to recognise except by association. Most
fresh metamorphic rocks are strong although perhaps fissile.
APPENDICES 807
Table A3.2a Descriptive terms for soil and rock strength (ISRM, 1981).
Grade Description Field identification Approximate range of

compressive strength
(MPa) (psi)
R6 Extremely strong rock Specimen can only be chipped with 250 (36 000)
geological hammer
R5 Very strong rock Specimen requires many blows of 100–250 (15 000–36 000)
geological hammer to fracture it
R4 Strong rock Specimen requires more than one blow with 50–100 (7000–15 000)
a geological hammer to fracture it
R3 Medium weak rock Cannot be scraped or peeled with a pocket 25–50 (3500–7000)
knife; specimen can be fractured with single
firm blow of geological hammer
R2 Weak rock Can be peeled with a pocket knife; shallow 5–25 (725–3500)
indentations made by firm blow with point of
geological hammer
R1 Very weak rock Crumbles under firm blows with point of 1–5 (150–725)
geological hammer; can be peeled by a
pocket knife
R0 Extremely weak rock Indented by thumbnail 0.25–1 (35–150)
S6 Hard clay Indented with difficulty by thumbnail 0.5 ( 70)
S5 Very stiff clay Readily indented by thumbnail 0.25–0.5 (35–70)
S4 Stiff clay Readily indented by thumb but penetrated 0.1–0.25 (15–35)
only with great difficulty
S3 Firm clay Can be penetrated several inches by 0.05–0.1 (7–15)
thumb with moderate effort
S2 Soft clay Easily penetrated several inches by thumb 0.025–0.05 (4–7)
S1 Very soft clay Easily penetrated several inches by fist 0.025 ( 4)
Table A3.2b Descriptive terms for rock strength (BS5930: 1999).
Term Field definition Unconfined

compressive
strength (MN/m2)
Very weak Gravel size lumps can be crushed between finger and thumb. 1.25
Weak Gravel size lumps can be broken in half by heavy hand pressure. 1.25 to 5
Moderately weak Only thin slabs, corners or edges can be broken off with 5 to 12.5
heavy hand pressure.
Moderately strong When held in the hand, rock can be broken by hammer blows. 12.5 to 50
Strong When resting on a solid surface, rock can be broken by hammer blows. 50 to 100
Very strong Rock chipped by heavy hammer blows. 100 to 200
Extremely strong Rock rings on hammer blows. Only broken by sledgehammer. 200
Table A3.3a Descriptive terms for weathering grades (ISRM, 1981).
Grade Term Description
W1 Fresh Not broken easily by geological hammer; ringing sound if struck

by geological hammer; no visible signs of decomposition
W2 Slightly weathered Not broken easily by geological hammer; ringing sound if struck
by geological hammer; fresh rock colours generally retained
but stained near joint surfaces
W3 Moderately weathered Cannot usually be broken by hand, easily broken by geological
hammer; dull or slightly ringing sound if struck by geological
hammer; completely stained throughout
W4 Highly weathered Can be broken by hand into smaller pieces; dull sound if struck
by geological hammer; not easily indented by point of
geological pick; does not slake when immersed in water;
completely discoloured compared with fresh rock
W5 Completely weathered Original rock texture preserved; can be crumbled by hand and
finger pressure into constituent grains; easily indented by point
of geological pick; slakes when immersed in water; completely
discoloured compared with fresh rock
W6 Residual soil Original rock texture completely destroyed; can be crumbled by
hand and finger pressure into constituent grains
Table A3.3b Description and classification of weathered rock for engineering purpose (BS5930: 1999).
Approach 1: Factual Description of Weathering (Mandatory)
Standard descriptions should always include comments on the degree and nature of any weathering effects at
material or mass scales. This may allow subsequent classification and provide information for separating rock
into zones of like character. Typical indications of weathering includes:
- Changes in colour - Changes in fracture state
- Reduction in strength - Presence, character and extent of weathering products
These features should be described using standard terminology, quantified as appropriate, together with non-
standard English descriptions as necessary to describe the results of weathering. At the mass scale the distribution
and proportions of the variously weathered materials (e.g. corestones vs matrix) should be recorded.
Can classification be
No DO NOT CLASSIFY
applied unambiguously?
Yes
Rock is moderately strong or Rock is moderately weak or

stronger in fresh state weaker in fresh state
A B C
APPENDICES 809
Table A3.3b (Continued).

A B C
Approach 2: Classification for Approach 4: Classification Incorporating

Uniform Materials Material and Mass Features
Grade Classifier Typical characteristics Class Classifier Typical characteristics
I Fresh Unchanged from original state A Unweathered Original strength, colour,

II Slightly Slight discoloration, slight fracture spacing
weathered weakening B Partially Slightly reduced strength,
III Moderately Considerably weakened, weathered slightly closer fracture spacing,
weathered penetrative discoloration weathering penetrating in from
Large pieces cannot be broken fractures, brown oxidation
by hand C Distinctly Further weakened, much closer
IV Highly Large pieces cannot be broken weathered fracture spacing grey reduction
weathered by hand D Destructured Greatly weakened, mottled,
Does not readily disaggregate ordered lithorelics in matrix
(slake) when dry sample becoming weakened and
immersed in water disordered, bedding disturbed
V Completely Considerably weakened
E Residual or Matrix with occasional altered
weathered Slakes
reworked random or ‘apparent’
Original texture apparent
lithorelics, bedding destroyed
VI Residual Soil derived by in situ Classed as reworked when
soil weathering but retaining none foreign inclusions are present
of the original texture or fabric as a result of transportation
Is a zonal classification Approach 5: Special Cases

appropriate and is there Use Rock Mass
enough information No Classification if For rocks whose weathering state does not follow the
available? appropriate other patterns indicate here, such as karst in carbonates
and the particular effect of arid climates
Yes
Approach 3: Classification for Heterogenous Masses
Zone Proportions of material grades Typical characteristics
1 % G I–III (not necessarily Behaves as rock: apply rock mechanics

all fresh rock) principles to mass assessment and design
2 90% G I–III Weak materials along discontinuities. Shear
10% G IV–VI strength stiffness and permeability affected
3 50% to 90% G I–III Rock framework still locked and controls strength
10% to 50% G IV–VI and stiffness; matrix controls permeability
4 30% to 50% G I–III Rock framework contributes to strength; matrix
50% to 70% G IV–VI or weathering products control stiffness and
permeability
5 30% G I–III Weak grades will control behaviour.
70%G IV–VI Corestones may be significant for investigation
and construction
6 100% G IV–VI (not May behave as soil although relict fabric may
necessarily residual soil) still be significant
GFE-Appendix.qxd
Table A3.4 Terminology and checklist for discontinuity description (BS5930: 1999).
810
Spacing Orientation Persistence Type of Roughness Wall strength Aperture Filling Seepage No. of sets

termination
Extremely Dip amount Discontinuous Cannot Small scale (cm) Clean
6/27/05
wide only in normally be and intermediate
6m cores described scale (m) Schmidt Cannot Cannot be Cannot be
hammer normally Surface described described
Very wide Stepped be described staining in cores in cores
2 to 6 m in cores (colour)
4:48 PM
Rough
Wide Continuous in Smooth
600 mm to cores
2 mm Striated Soil infilling
Undulating (describe in
Page 810
Medium accordance
200 to Rough Point load Very open with 41) Moisture on
600 mm Very high Termination Smooth test 10 mm rock surface
20 m x(outside Open Mineral
Striated
Close exposure) 2.5 mm to coatings Dripping
60 to High Planar 10 mm (e.g. calcite, water
200 mm Take No. of 10 to 20 m r(within Other index Moderately chlorite
readings, rock) Rough tests open gypsum Water flow Record
Very close of dip Medium Smooth 0.5 mm to etc.) measured spacing and
20 to direction/dip 3 to 10 m d(against Striated 2.5 mm per time orientation
60 mm e.g. 015/08 discontinuity) Tight Other–– unit on an of sets to
Large scale (dm) 0.1 mm to specify individual each other
Extremely Report as Low 0.5 mm discontinuity and all
close ranges and 1 to 3 m Waviness Visual Very tight or set of details for
20 mm on stereo Curvature assessment 0.1 mm Record discontinu- each set
net if width and ities
Straightness
Take number appropriate Take number continuity Small flow
of readings Very low Record also Measure of readings of infill 0.05–0.5 l/s
state min. 1m size of amplitude and state min. Medium flow
average and exposure wavelength of average 0.5–5.0 l/s
max. feature and max. Strong flow
5 l/s
APPENDICES 811
Table A3.5a Types of discontinuities (ISRM, 1981).
Type Term Description
B Bedding The arrangement of a sedimentary rock in layers; stratification.

J Joint A break of geological origin in the continuity of a body of rock along which
there has been no visible displacement. A group of parallel joints is called a
set and joint sets intersect to form a joint system. Joints can be open, filled
or healed.
SL Slickenside A fracture along which there has been recognisable displacement, from a few
centimetres to a few kilometres in scale. The walls are often striated and
polished (slickensided) resulting from the shear displacement.
F Fault A fracture zone along which there has been recognisable displacement, from a
few centimetres to a few kilometres in scale. The walls are often striated and
polished (slickensided) resulting from the shear displacement. Frequently rock
on both sides of a fault is shattered and altered or weathered, resulting in
fillings such as breccia and gouge. Fault widths may vary from millimetres to
hundreds of metres.
Table A3.5b Types of discontinuities (BS5930: 1999). Table A3.6 Descriptive terms for discontinuity
roughness (ISRM, 1981).
Type of Description
discontinuity Level Description
Joint A discontinuity in the body of rock I Rough, stepped
along which there has been no visible II Smooth, stepped
displacement. III Slickensided, stepped
Fault A fracture or fracture zone along IV Rough, undulating
which there has been recognisable V Smooth, undulating
displacement. VI Slickensided, undulating
Bedding A fracture along the bedding VII Rough, planar
fracture (bedding is a surface parallel to the VIII Smooth, planar
plane of deposition). IX Slickensided, planar
Cleavage A fracture along a cleavage (cleavage
fracture is a set of parallel planes of weakness
often associated with mineral Table A3.7 Descriptive terms for discontinuity aper-
realignment). ture and thickness (ISRM, 1981).
Induced A discontinuity of non-geological
Aperture Description
fracture origin, e.g. brought about by coring,
(mm)
blasting, ripping etc.
0.1

Incipient A discontinuity which retains some Very tight
fracture tensile strength, which may not be 0.1–0.25 Tight  ‘Closed’ features
fully developed or which may be 0.25–0.5 Partly open 

partially cemented. Many incipient 0.5–2.5 Open
fractures are along bedding or 2.5–10 Moderately wide  ‘Gapped’ features
cleavage. 10 Wide 
10–100 Very wide 
100–1000 Extremely wide  ‘Open’ features
1m Cavernous 
Table A3.8 Terms for discontinuity spacing (ISRM, Table A3.10b Descriptive terms for block size and
1981). shape (BS5930: 1999).
Description Spacing (mm) First term Dimension
Extremely close spacing 20 Very large Greater than 2 m

Very close spacing 20–60 Large 600 mm to 2 m
Close spacing 60–200 Medium 200 mm to 600 mm
Moderate spacing 200–600 Small 60 mm to 200 mm
Wide spacing 600–2000 Very small Less than 60 mm
Very wide spacing 2000–6000
Second term Nature of block
Extremely wide spacing 6000
Blocky Equidimensional
Tabular One dimension much less than the
other two
Table A3.9 Persistence dimensions (ISRM, 1981). Columnar One dimension much greater than
Persistence Dimensions (m) the other two
Very low persistence 1

Low persistence 1–3
Medium persistence 3–10 Table A3.11a Seepage quantities in unfilled disconti-
High persistence 10–20 nuities (ISRM, 1981).
Very high persistence 20
Seepage Description
rating
Table A3.10a Descriptive terms for block size (joint
I The discontinuity is very tight and dry,
density) (ISRM, 1981).
water flow along it does not appear
Description Jv (joints/m3) possible.
II The discontinuity is dry with no evidence
Very large blocks 1.0 of water flow.
Large blocks 1–3 III The discontinuity flow is dry but shows
Medium-sized blocks 3–10 evidence of water flow, i.e. rust
Small blocks 10–30 staining etc.
Very small blocks 30 IV The discontinuity is damp but no free
water is present.
V The discontinuity shows seepage,
occasional drops of water, but no
continuous flow.
VI The discontinuity shows a continuous
flow of water/estimate litres/min and
describe pressure i.e. low, medium, high).
APPENDICES 813
Table A3.11b Seepage quantities in filled discontinu- Table A3.11c Seepage quantities in rock mass (e.g.
ities (ISRM, 1981). tunnel wall) (ISRM, 1981).
Seepage Description Seepage Description

rating rating
I The filling materials are heavily I Dry walls and roof, no detectable
consolidated and dry, significant flow seepage.
appears unlikely due to very low II Minor seepage, specify dripping
permeability. discontinuities.
II The filling materials are damp, but no III Medium inflow, specify discontinuities
free water is present. with continuous flow (estimate litre/min/
10 m length of excavation).
III The filling materials are wet, occasional
IV Major inflow, specify discontinuities with
drops of water.
strong flows (estimate litre/min/10 m
IV The filling materials show signs of length of excavation).
outwash, continuous flow of water V Exceptionally high inflow, specify
(estimate litre/min). source of exceptional flows (estimate
V The filling materials are washed out litre/min/10 m length of excavation).
locally, considerable water flow along
outwash channels (estimate litre/min and
describe pressure, i.e. low, medium, high).
VI The filling materials are washed out
completely, very high water pressures
experienced, especially on first
exposure (estimate litre/min and
describe pressure).
GFE-Appendix.qxd
Table A3.13 Geomorphic rock mass strength classification and ratings, after Selby (1980).
814
I 2 3 4 5

Very strong Strong Moderate Weak Very weak
Intact rock strength 100–60 60–50 50–40 40–35 35–10
6/27/05
(N-type Schmidt Very strong Strong Moderate Weak Very weak
hammer ‘R’) r 20 r 18 r 14 r 10 r5
Weathering Unweathered Slightly Moderately Highly weathered Completely
weathered weathered weathered
4:48 PM
r 10 r9 r7 r5 r3
Spacing of 3m 3–1 m 1–0.3 m 300–50 mm 50 mm
discontinuities Solid Massive Blocky/seamy Fractured Crushed or shattered
r 30 r 28 r 21 r 15 r8
Page 814
Joint orientations Very favourable; Favourable; moderate Fair; horizontal dips, Unfavourable; Very unfavourable;
steep dips into slope, dips into slope or nearly vertical moderate dips out steep dips out of slope
cross joints interlock (hard rocks only) of slope
r 20 r 18 r 14 r9 r5
Width of joints 0.1 mm 0.1 mm 1–5 mm 5–20 mm 20 mm
r7 r6 r5 r4 r2
Fracture continuity None continuous Few continuous Continuous, no infill Continuous, thin infill Continuous, thick infill
r7 r6 r5 r4 r1
Outflow of None Trace Slight Moderate Great
groundwater ( 25 l /min (25–125 l /min ( 125 l /min
from 10 m2) from 10 m2) from 10 m2)
r6 r5 r4 r3 r1
Total rating 100–91 90–71 70–51 50–26 26
APPENDICES 815
References Selby, M. J. (1980) A rock mass strength classification

for geomorphic purposes: with tests from Antarctica
BS5930: 1999 Code of practice for site investigations. and New Zealand. Z. Geomorph. 24, 31–51.
British Standards Institution, London. ISRM (International Society for Rock Mechanics)
Geological Society Engineering Group Working Party (1981) Suggested Methods for the Quantitative
(1977) The description of rock masses for engineer- Description of Discontinuities in Rock Masses.
ing purposes. Q. J. Eng. Geol. 10, 355–388. (ed. E. T. Brown), Pergamon Press, Oxford.
Appendix A4: Geomorphological mapping
The Geological Society Working Party Report on objective of the exercise and the complexity of the
maps and plans (Anon., 1972) identified examples landscape; often, however, it is driven more by the
of geomorphological mapping that could be of use availability of suitable base maps and the con-
to engineers. Lee (2001) presents a brief introduc- straints of the fieldwork programme.
tion to the preparation of engineering geomor- There is no single approach to geomorphologi-
phology maps. Stylised symbols are often used cal mapping. The method chosen will generally
(see Demek and Embleton, 1978; Gardiner and reflect the nature of the problem to be solved, the
Dakcombe, 1983; Cooke and Doornkamp, 1990). resources available and, not least, the training and
The style and format of an engineering experience of the mapper. However, all maps should
geomorphological map needs to reflect the nature seek to subdivide the landscape into units with
of the environment and the problems to be similar surface form, materials and process charac-
addressed. Amongst the more common types of teristics. At the smaller scales these units will,
map are: regional surveys of terrain conditions, inevitably, be terrain models or land systems
either to provide a framework for land use plan- (Figure A4.1; see Phipps, 2001). Individual land-
ning or as part of the baseline studies for environ- forms or terrain units (e.g. escarpments, dune
mental impact assessment; general assessments of ridges, river channels, landslides) might be
resources or geohazards within an area or along a recorded on medium-scale maps. Landform ele-
route corridor; specific-purpose surveys to delin- ments or geomorphological units (e.g. individual
eate and characterise particular landforms, e.g. landslide blocks, within-channel bars, gullys) might
the investigation of pre-exising landslide prob- be recorded on large-scale maps.
lems in and around the Channel Tunnel terminal The method of recording geomorphologi-
area, Folkestone (Griffiths et al., 1995); the delin- cal information may change with map scale, but
eation of gypsum-related subsidence problems in the basic approach to data collection should
the Ripon area of the UK (Thompson et al., remain the same: mapping of surface form,
1996). description of materials and recording evidence
Map scale is an important issue. It will not of process.
only determine the level of detail that should be A key stage in map creation is the identifica-
sought, but it will also dictate what features tion of suitable mapping units to reflect the scale
should be recorded. For example, at a large scale of mapping and the objectives of the study.
(i.e. 1 : 500 to 1 : 5000) it will be practical to Ideally, each unit should have consistent geomor-
record individual landslides and their components phological characteristics, although internal vari-
(e.g. the head, track and accumulation zone of a ability of materials or rate of process may be a
mudslide), whereas at a medium scale (i.e. feature of the unit (e.g. in areas mantled by glacial
1 : 10 000 to 1 : 50 000) it will generally only be tills). Three broad categories of geomorphological
possible to show the limits of larger landslides. At unit can be recognised:
small scales (i.e. 1 : 100 000 and smaller) it may
only be practical to note ‘areas prone to landslid- 1. Units reflecting the control of the underlying
ing’ or to show the larger features by a stylised geology (e.g. plateau surfaces, lithological
map symbol. Ideally map scale should reflect the benches)
816
APPENDICES 817
Land system Gentle hills with broad terraced

river valleys
Land unit
3. Terrace 1. Hill
2. Minor valley
4. Major valley
Land elements
3b. Bluff 3a. Surface 2a. Floor 2b. Stream 4a. Abandoned 1a. slope
channel
1b. Gully side
4c. Stream 1d. Stream
4c. Floodplain 1c. Gully floor
Figure A4.1 Diagram to show the relationship between land system, land facet (or land unit), and land element
(after Lawrence, 1972).
2. Units reflecting the activity of active or relict mapping is the most convenient and efficient way
surface processes (e.g. landslide, fluvial, aeolian of recording the surface morphology and allows
features etc.) later interpretation of form and process (Savigear,
3. Units reflecting modification of the landscape 1965; Waters, 1958). Breaks or changes of slope
by man (e.g. areas of cut and fill, quarries, are identified from aerial photographs or in the
made ground etc.). field, and recorded using standard symbols
(Figures A4.2 and A4.3).
At the smaller scales, topographic map contours Geomorphological map unit boundaries should
can be used to subdivide the landscape into units, generally follow morphological boundaries.
based on slope steepness (i.e. from the spacing of Depending on the map scale these units can be
contour lines) and slope form (i.e. from the shape portrayed in blocks of contrasting colour or shad-
of the contours). However, at larger scales con- ing, or as stylised symbols (Figures A4.4 to A4.16)
tours reveal little about landforms and their or a combination of both.
assemblages. The technique of morphological
Figure A4.2 A morphological mapping system (after Savigear, 1965).

APPENDICES 819
Figure A4.3 Different uses of geomorphological mapping data and styles of presentation for the same land surface
(from Cooke and Doornkamp, 1990).
Figure A4.4 Bedrock lithology: geomorphological mapping symbols (from Cooke and Doornkamp, 1990).
Figure A4.5 Geological structure: geomorphological mapping symbols (from Cooke and Doornkamp, 1990).
Figure A4.6 Features resulting from bedrock structure: geomorphological mapping symbols (from Cooke and
Doornkamp, 1990).
APPENDICES 821
Figure A4.7 Features of volcanic origin: geomorphological mapping symbols (from Cooke and Doornkamp,
1990).
Figure A4.8 Superficial unconsolidated materials: geomorphological mapping symbols (from Cooke and
Doornkamp, 1990).
Figure A4.9 Slope instability features: geomorphological mapping symbols (from Cooke and Doornkamp, 1990).
Figure A4.10 Aeolian features: geomorphological mapping symbols (from Cooke and Doornkamp, 1990).
Figure A4.11 Coastal features: geomorphological mapping symbols. (from Cooke and Doornkamp, 1990).
Figure A4.12 Forms of permafrost areas, glacial and periglacial features: geomorphological mapping symbols
(from Cooke and Doornkamp, 1990).
Figure A4.13 Forms of fluvial origin: geomorphological mapping symbols (from Cooke and Doornkamp, 1990).
APPENDICES 823
Figure A4.14 Karst landscape features: geomorphological mapping symbols (from Cooke and Doornkamp, 1990).
Figure A4.15 Major features not included in previous figures: geomorphological mapping symbols (from Cooke
and Doornkamp, 1990).
Figure A4.16 Man-made features: geomorphological mapping symbols (from Cooke and Doornkamp, 1990).
References methods. Report LR 506, Transport and Road

Research Laboratory, Crawthorne, UK
Anon (1972) The preparation of maps and plans in Lee, E. M. (2001) Geomorphological Mapping. In
terms of engineering geology. Quarterly Journal of Griffiths, J. S. (ed.) Land Surface Evaluation for
Engineering Geology 5, 293–381. Engineering Practice. Geological Society Special
Cooke, R. U. and Doornkamp, J. C. (1990) Geomorph- Publication 18, 53–56.
ology in Environmental Management. Oxford Phipps, P. (2001) Terrain systems mapping. In Griffiths,
University Press. J. S. (ed.) Land Surface Evaluation for Engineering
Demek, J. and Embleton, C. (eds) (1978) Guide to Practice. Geological Society Special Publication
Medium-scale Geomorphological Mapping. 18, 59–64.
International Geographical Union, Brno. Savigear, R. A. G. (1965) A technique of morpho-
Gardiner, V. and Dackombe, R. (1983) Geomophological logical mapping. Ann. Assoc. Am. Geog. 53,
Field Manual. George Allen and Unwin. 514–538.
Griffiths, J. S. Brunsden, D. Lee, E. M. and Jones, Thompson, A., Hine, P. D., Greig, J. R. and Peach, D. W.
D. K. C. (1995) Geomorphological investigation for (1996) Assessment of Subsidence Arising from
the Channel Tunnel and Portal. Geographical Gypsum Solution. Symonds Travers Morgan Report
Journal 161, 257–284. to the DoE.
Lawrence, C. J. (1972) Terrain evaluation in West Waters, R. S. (1958) Morphological mapping.
Malaysia: Part 1: Terrain classification and survey Geography 43, 10–17.
Appendix A5: Soil erosion assessment
Chapter 11 describes a range of approaches avail- Tables A5.1 to A5.3 provide a guide for
able for assessing the erosion hazard within an Manning’s n for different plant covers and stream
area, ranging from simple measures derived from bed types.
readily available climatic data (these consider Figure A5.3 presents a useful nomogram for
erosion hazard in terms of the erosivity and not the evaluation of the Manning equation (from
soil erodibility) to more detailed mapping based Chow, 1959). To use this nomogram, a line is set
methods that consider field evidence of erosion. out from the value of the hydraulic radius (R)
Field mapping or aerial photograph interpreta- through the estimated value of Manning’s n to its
tion of the evidence of soil erosion activity (e.g. intersection with the pivot line. A second line is
sheet wash, rills, gullies) can provide a measure of then set out from this point on the pivot line to the
the erosion hazard across different terrain units. approximate value of slope (S). The velocity is
Soil erosion survey approaches include geomor- then read off from the velocity axis.
phological mapping and field observation. The most important equation for predicting
general soil loss from rainsplash and runoff is
Geomorphological mapping
the Universal Soil Loss Equation (USLE: see
A legend for mapping soil erosion features is pre-
Chapter 11):
sented by Morgan (1995) and is included as
Figure A5.1.
E R K LS C P
Field observation
A proforma for recording evidence of soil erosion where E average annual soil loss (t/ha) and R is
and other site characteristics is presented as the rainfall erosivity factor, based on the mean EI30.
Figure A5.2. K is the soil erodibility index i.e. the soil loss per
The velocity of flow is critical in initiating soil unit of EI30 (as measured on a standard bare soil
erosion, as the flow must attain a threshold value plot, 22 m long and on a 5 slope). LS is the com-
before entrainment of particles begins. The veloc- bined slope length (L) and slope steepness (S) fac-
ity (v) of fully turbulent flow can be determined tor. C is the crop management factor and represents
using Manning’s equation: the ratio of soil loss under a given crop to that from
bare soil (see Table A5.4). P is the erosion control
r 2/3 S1/2
v factor. With no erosion control in place P 1.0
n (see Table A5.5). Table A5.6 provides a worked
where r is the hydraulic radius, S is slope (m/m) example of the use of the USLE for predicting
and n is the Manning roughness coefficient. mean annual soil loss, from Morgan (1995).
825
Symbol Feature Colour
Perennial water course blue

Seasonal water course blue
Crest line brown
Contour line brown
Major escarpment brown
Convex slope break brown
Concave slope break brown
Waterfall blue
Rapids blue
Edge of flood plain blue
Edge of river terrace blue
Back of river terrace blue
Swamp or marsh blue
Active gully red
Stable gully blue
Active rills red
Sheetwash/rainsplash (inter-rill erosion) red
River bank erosion red
Landslide or slump scar red
Landslide or slump tongue red
Small slides, slips red
Colluvial or alluvial fans brown
Sedimentation brown
Landuse boundary (landuse denoted by letter
e.g. Rrubber; F forest; Pgrazing land; Larable land.) green
Roads and tracks black
Railway black
Cutting black
Embankment black
Buildings black
Terrace black
Waterway black
Slopes
0–1˚
2–3˚
4–8˚
9–14˚
15–19˚
over 19˚
Figure A5.1 Legend for mapping soil erosion (from Morgan, 1995).
APPENDICES 827
Figure A5.2 Proforma for recording soil erosion in the field (from Morgan, 1995).
Table A5.1 Guide values for Manning’s n. Table A5.2 Values of Manning’s n for major steams
and streams with coarse bed materials or rock-cut chan-
Land use or cover Manning’s n nels (Chow, 1959).
Bare soil Manning’s n
roughness depth 25 mm 0.010–0.030
roughness depth 25–50 mm 0.014–0.033 MAJOR STREAMS (bank full
roughness depth 50–100 mm 0.023–0.038 width 30 m)
roughness depth 100 mm 0.045–0.049 (1) Regular sections with no 0.025–0.060
Bermuda grass––sparse to good cover boulders or brush
very short ( 50 mm) 0.015–0.040 (2) Irregular and rough sections 0.035–0.100
short (50–100 mm) 0.030–0.060
CHANNELS WITH RELATIVELY
medium (150–200 mm) 0.030–0.085
COARSE BED MATERIALS AND
long (250–600 mm) 0.040–0.150
ROCK-CUT CHANNELS
very long ( 600 mm) 0.060–0.200
gravel, 4–8 mm in diameter 0.019–0.020
Bermuda grass––dense cover 0.300–0.480
Other dense sod-forming grasses 0.390–0.630
Dense bunch grasses 0.150
pebbles and shingle, 60–110 mm 0.027–0.030
Kudzu 0.070–0.230
in diameter
Lespedeza 0.100
pebbles and shingle, 110–250 mm 0.030–0.035
Natural rangeland 0.100–0.320
in diameter
Clipped rangeland 0.020–0.240
smooth and uniform rock-cut 0.025–0.035–0.040
Wheat straw mulch
channels
2.5 t/ha 0.050–0.060
jagged and irregular rock-cut 0.035–0.040–0.050
5.0 t/ha 0.075–0.150
channels
7.5 t/ha 0.100–0.200
10.0 t/ha 0.130–0.250
Chopped maize stalks
2.5 t/ha 0.012–0.050
5.0 t/ha 0.020–0.075
10.0 t/ha 0.023–0.130
Cotton 0.070–0.090
Wheat 0.100–0.300
Sorghum 0.040–0.110
Concrete or asphalt 0.010–0.013
Gravelled surface 0.012–0.030
Chisel-ploughed soil
0.6 t/ha residue 0.006–0.170
0.6–2.5 t/ha residue 0.070–0.340
2.5–7.5 t/ha residue 0.190–0.470
Disc-harrowed soil
0.6 t/ha residue 0.008–0.410
0.6–2.5 t/ha residue 0.100–0.250
2.5–7.5 t/ha residue 0.140–0.530
No tillage
0.6 t/ha residue 0.030–0.070
0.6–2.5 t/ha residue 0.010–0.130
2.5–7.5 t/ha residue 0.160–0.470
Bare mouldboard-ploughed soil 0.020–0.100
Bare soil tilled with coulter 0.050–0.130
After Petryk and Bosmajian (1975), Temple (1982) and
Engman (1986).
APPENDICES 829
Table A5.3 Manning’s n for minor streams and flood plains (after Chow, 1959).
Manning’s n
MINOR STREAMS (bank full width 30 m)

(a) Streams on plain
(1) Clean, straight, full-stage, no rifts or deep pools 0.025–0.030–0.033
(2) As (1) but more stones and weeds 0.030–0.035–0.040
(3) Clean, winding, some pools and shoals 0.033–0.040–0.045
(4) As (3) but with some weeds and stones 0.035–0.045–0.050
(5) As (4) but lower stages, less efficient slopes and sections 0.040–0.048–0.055
(6) As (4) but with more stones 0.045–0.050–0.060
(7) Sluggish reaches, weedy, deep pools 0.050–0.070–0.080
(8) Very weedy reaches with deep pools 0.075–0.100–0.150
(9) Floodways with heavy stands of timber and vegetation 0.075–0.100–0.150
(b) Mountain streams with no vegetation in channel, usually with steep banks and 0.030–0.040–0.050
bank vegetation submerged at high stages
(1) Bed of gravel, cobbles and a few boulders
(2) Bed of cobbles with large boulders 0.040–0.050–0.070
FLOOD PLAINS
(a) Pasture, no brush
(1) Short grass 0.025–0.030–0.035
(2) Long grass 0.030–0.035–0.050
(b) Cultivated areas
(1) No crops 0.020–0.030–0.040
(2) Mature row crops 0.025–0.035–0.045
(3) Mature field crops 0.030–0.040–0.050
(c) Brush
(1) Scattered brush, heavy weeds 0.035–0.050–0.070
(2) Light brush and trees, in winter 0.035–0.050–0.060
(3) As (2), in summer 0.040–0.060–0.080
(4) Medium to dense brush, in winter 0.045–0.070–0.110
(5) As (4), in summer 0.070–0.100–0.160
(d) Trees
(1) Dense straight willows, summer 0.110–0.150–0.200
(2) Cleared land with stumps, not sprouting 0.030–0.040–0.050
(3) As (2), but heavy sprout growth 0.050–0.060–0.080
(4) Heavy stands of timber, a few fallen trees, little undergrowth, 0.080–0.100–0.120
flood stage below branches
(5) As (4), but floods reaching branches 0.100–0.120–0.160
Figure A5.3 Nomogram for evaluation of the Manning equation (after Chow, 1959).
Table A5.4 C-factor values for the Universal Soil Loss Equation (from
Morgan 1995).
Practice Average annual

C-factor
Bare soil 1.00

Forest or dense shrub, high mulch crops 0.001
Savanna or prairie grass in good condition 0.01
Overgrazed savanna or prairie grass 0.10
Maize, sorghum or millet: high productivity 0.20–0.55
conventional tillage
conventional tillage
Maize, sorghum or millet: low productivity 0.02–0.10
no or minimum tillage
chisel ploughing into residue
Maize, sorghum or millet: low productivity 0.30–0.45
chisel ploughing into residue
Cotton 0.40–0.70
Meadow grass 0.01–0.025
Soya beans 0.20–0.50
Wheat 0.10–0.40
Rice 0.10–0.20
APPENDICES 831
Groundnuts 0.30–0.80
Palm trees, coffee, cocoa with crop cover 0.10–0.30
Pineapple on contour: residue removed 0.10–0.40
Pineapple on contour: with surface residue 0.01
Potatoes: rows downslope 0.20–0.50
Potatoes: rows across-slope 0.10–0.40
Cowpeas 0.30–0.40
Strawberries: with weed cover 0.27
Pomegranate: with weed cover 0.08
Pomegranate: clean-weeded 0.56
Ethiopian tef 0.25
Sugar cane 0.13–0.40
Yams 0.40–0.50
Pigeon peas 0.60–0.70
Mungbean 0.04
Chilli 0.33
Coffee: after first harvest 0.05
Plantains: after establishment 0.05–0.10
Papaya 0.21
Table A5.5 P-factor values for the Universal Soil

Loss Equation (from Morgan 1995).
Erosion-control practice P-factor value
Contouring: 0–1 slope 0.60*

Level bench terrace 0.14
Reverse-slope bench terrace 0.05
Outward-sloping bench terrace 0.35
Level retention bench terrace 0.01
Tied-ridging 0.10–0.20
* Use 50% of the value for contour bunds or if contour
strip cropping is practised.
Table A5.6 Prediction of soil loss using the Universal Soil Loss Equation (from Morgan 1995).
Problem
Calculation of mean annual soil loss on a 100 m long slope of 7 on soils of the Rengam Series under maize cultivation
with contour bunds spaced at 20 m intervals, near Kuala Lumpur.
Equation
Mean annual soil loss R K LS C P
Estimating R (Rainfall erosion index)
Method 1:
Mean annual precipitation (P) 2695 mm
From Roose (1975), mean annual rainfall erosion index (R) in US units 0.5 P
0.5 2695
1347.5
Conversion to metric units 1347.5 1.73
2331.18
Method 2:
From Morgan (1974), mean annual erosivity (KE 25) 9.28 P – 8838
(9.28 2695) – 8838
16 171.6 J/m2
Multiply by I30 (use 75 mm/h; maximum value recommended by
Wischmeier and Smith 1978) 16171.6 75
1 212 870
Divide by 1000 to give R value in metric units 1212.87
Method 3:
From Foster et al. (1981), mean annual EI30 (kg.m.mm)/(m2.h) 0.276 P 75
0.276 2695 75
55 786.5
With these units, divide by 100 to give R value in metric units 557.9
Best estimate: discard result from Method 3 which is rather low. Take average
value of Methods 1 and 2:
R 1772
Estimating K (Soil erodibility index)
The soils have a 43% clay, 8% silt, 9% fine sand and 40% coarse sand; organic content about 3%.
Using the nomograph (Figures A5.3; 11.13) gives a first approxiamation K value 0.05
Estimating LS (Slope factor)
For slope lenght (l) and slope steepness (s) in meters and per cent respectively,
LS (l/22)0.5 (0.065 0.045s 0.0065s2)
With contour bunds at 20 m spacing, l 20 m and s 12% (approximation of 7)
LS (20/22)0.5 (0.065 (0.045 12) (0.0065 122)
LS 0.95 1.54
LS 1.46
Estimating C (Crop management factor)
According to Table A5.4, the C value for maize ranges between 0.2 and 0.9, depending on the productivity. For many
tropical farming conditions, C for maize lies between 0.4 and 0.9 (Roose, 1975), depending on the cover.
During the three-month period from seeding to harvest, the cover is likely to vary from 9 to 45 per cent in the
first month, to 55 to 93 per cent in the second month, and 45 to 57 per cent in the third month. Therefore, we might
assume C values of 0.9, 0.4 and 0.7 for the three respective months.
APPENDICES 833
Maize can be planted at any time of year in Malaysia but assume planting after the April rains, allowing growth,
ripening and harvesting in June and July which are the driest months. Land is under dense secondary growth prior
to planting (assume C 0.001) and allowed to revert to the same after harvest (assume C 0.1).
Of the mean annual precipitation, 32 per cent falls between January and April inclusive, 10 per cent in May,
6 per cent in June, 7 per cent in July, and 45 per cent between August and December. Assuming that erosivity is
directly related to precipitation amount, these values can be used to describe the distribution of the R factor
throughout the year.
From these data, the following table is constructed.
Months C value Adjustment factor Weighted C value

(% R value) (col. 2 ⴛ col. 3)
January–April 0.001 0.32 0.000 32

May 0.9 0.10 0.09
June 0.4 0.06 0.024
July 0.7 0.07 0.049
August–December 0.1 0.45 0.045
Total 0.208 32
C factor value for the year 0.028

Estimating P (Erosion-control practice factor)
From Table A5.5, P value for contour bunds 0.3
Soil loss estimation
Mean annual soil loss 1772 0.05 1.46 0.208 0.3
8.07 t/ha
References Petryk, S. and Bosmajian, G. (1975) Analysis of flow

through vegetation. Transactions of American
Chow, V. T. (1959) Open-channel Hydraulics. McGraw- Society of Civil Engineers 101, 871–884.
Hill, New York. Roose, E. J. (1975) Erosion et ruissellement en Afrique
Morgan, R. P. C. (1995) Soil Erosion and Conservation. de l’ouest: vingt années de measures en petites par-
Longman, London. celles expérimentales. Cyclo, Orstom, Ivory Coast.
Engman, E. T. (1986) Roughness coefficients for rout- Temple, D. M. (1982) Flow retardance of submerged
ing surface runoff. Transactions of American grass channel linings. Transactions of American
Society of Civil Engineers 112, 39–53. Society of Agricultural Engineers 25, 1300–1303.
Foster, G. R., Lane, L. J., Nowlin, J. D., Laflen, J. M. Wischmeier, W. H. and Smith, D. D. (1978) Predicting
and Young, R. A. (1981) Estimating erosion and rainfall erosion losses. USDA Agricultural Research
sediment yield on field-sized areas. Transactions of Service Handbook 537.
American Society of Agricultural Engineers 24,
1253–1263.
Morgan, R. P. C. (1974) Estimating regional variations
in soil erosion hazard in Peninsula Malaysia.
Malayan Nature Journal 28, 94–106.
Appendix A6: Units, Scales and conversions
Table A6.1 Derived units in the SI and CGS systems (after Tennent, 1971).
Quantity and Dimensions SI unit CGS unit Ratio CGS: SI units

recommended symbol
Mass, m m kilogram (kg) gram (g) 103

Length, l l metre (m) centimetre (cm) 102
Time, t t second (s) second (s) 1
Area, A, S l2 m2 cm2 104
Volume, V l3 m3 cm3 106
Density, r ml3 kg m3 g cm3 103
Velocity, u, y lt1 m s1 cm s1 102
Acceleration, G lt2 m s2 gal 102
Momentum, P mlt1 kg m s1 g cm s1 103
Moment of inertia, I, J ml 2 kg m2 g cm2 107
Angular momentum, L ml 2 t1 kg m2 s1 g cm2 s1 107
Force, F mlt2 newton (N) dyne (dyn) 105
Energy of work, E, W ml 2t2 joule (J) erg 107
Power, P ml 2t3 watt (W) erg s1 107
Pressure or stress, p ml1t2 pascal (Pa) dyn cm2 101
Surface tension, ml2 N m1 dyn cm1 103
Viscosity, h ml1 t1 kg m1 s1 poise 101
Frequency, , f t1 hertz (Hz) s1 1
Table A6.2 Conversion factors.
To convert A A B To convert B
to B multiply by to A multiply by
Length
2.54 102 inch (in) metre (m) 39.37
0.3048 foot (ft) metre (m) 3.2468
0.9144 yard (yd) metre (m) 1.0936
1.8288 fathom (fm) metre (m) 0.5468
20.1168 chain (ch) metre (m) 4.97 102
201.168 furlong (fl) metre (m) 4.97 103
1609.34 mile (mi) metre (m) 6.214 104
1853.2 nautical mile (UK) metre (m) 5.3961 104
1852.0 nautical mile (Int.) metre (m) 5.3996 104
834
APPENDICES 835
1853.25 nautical mile (US) metre (m) 5.3959 104

100.0 cable fathom (fm) 102
1.6094 mile (mi) kilometre (km) 0.6214
8.0 mile (mi) furlong (fl) 0.125
1760.0 mile (mi) yard (yd) 5.6818 104
5280.0 mile (mi) foot (ft) 1.8939 104
Area
6.4516 104 sq. inch (in2) sq. metre (m2) 1.55 103
9.2903 102 sq. foot (ft2) sq. metre (m2) 10.764
0.8361 sq. yard (yd2) sq. metre (m2) 1.196
2 589 988.0 sq. mile (mi2) sq. metre (m2) 3.861 107
4046.856 acre sq. metre (m2) 2.4711 104
2.590 sq. mile (mi2) sq. kilometre (km2) 0.3861
0.4047 acre hectare (ha) 2.471
640.0 sq. mile (mi2) acre 1.5625 103
10.000 hectare (ha) sq. metre (m2) 104
3.861 103 hectare (ha) sq. mile (mi2) 259.0045
247.105 sq. kilometre (km2) acre 4.047 103
Volume
1.6387 105 cubic inch (in3) cubic metre (m3) 6.1024 104
16.387 cubic inch (in3) cubic centimetre (cm3) 6.1024 102
2.8317 102 cubic foot (ft3) cubic metre (m3) 35.314
28 316.8 cubic foot (ft3) cubic centimetre (cm3) 3.5315 105
28.3168 cubic foot (ft3) litre (l) 3.5315 102
0.7646 cubic yard (yd3) cubic metre (m3) 1.3079
3.785 103 gallon (US) cubic metre (m3) 264.2
4.546 103 gallon (UK) cubic metre (m3) 219.97
0.21998 litre (l) gallon (UK) 4.546
0.26418 litre (l) gallon (US) 3.7853
8.0 bushel gallon (UK) 0.125
9.608 bushel gallon (US) 0.104
231.0 gallon (UK) cubic inch (in3) 4.329 103
1233.482 acre foot cubic metre (m3) 8.1071 104
0.8326 gallon (US) gallon (UK) 1.2011
2.8413 102 fluid ounce (fl. oz) litre (l) 35.195
Mass
2.835 102 ounce (oz) kilogram (kg) 35.273
0.4536 pound (lb) kilogram (kg) 2.2046
6.3503 stone (st) kilogram (kg) 0.1575
50.8023 hundredweight (cwt) kilogram (kg) 1.9684 102
1016.04 ton kilogram (kg) 9.8421 104
907.20 short ton kilogram (kg) 1.1023 103
1.016 ton metric tonne (t) 0.9842
2204.6 metric tonne (t) pound (lb) 4.536 104
14.5939 slug kilogram (kg) 6.852 102
Weight
4.448 pound force (lbf) newton (N) 0.2248
0.1383 poundal (pdl) newton (N) 7.233
105 newton (N) dyne (dyn) 105
32.17 pound-force (lbf) poundal (pdl) 0.031 08
980.7 gram-force (lgf) dyne (dyn) 1.0197 103
Pressure or stress
15.44 106 ton-force per sq. in pascal (Pa) 6.4767 108
157.47 ton-force per sq. in kilogram-force per sq. cm 6.3504 103
107.3 103 ton-force per sq. ft pascal (Pa) 9.3197 106
1.0936 104 ton-force per sq. ft kilogram-force per sq. m 9.1441 105
6.895 103 pound-force per sq. in pascal (Pa) 1.4503 104
7.03 102 pound-force per sq. in kilogram-force per sq. m 14.225
47.9 pound-force per sq. ft pascal (Pa) 2.0877 102
4.882 pound-force per sq. ft kilogram-force per sq. m 0.204 83
101.325 103 standard atmosphere pascal (Pa) 9.869 106
1.033 standard atmosphere kilogram-force per sq. m 9.6805
14.697 standard atmosphere pound-force per sq. in 6.804 102
760 standard atmosphere millimetres of mercury 1.315 103
(mm Hg)
33.901 standard atmosphere feet of water (ft H2O) 2.9498 102
3.05 102 foot of water (ft H2O) kilogram-force per sq. cm 3.2787 103
2.989 103 foot of water pascal (Pa) 3.3456 104
105 bar pascal (Pa) 105
133.322 millimetre of mercury pascal (Pa) 7.5 103
9.80665 144 kilogram-force per sq. cm pascal (Pa) 1.0197 105
Energy or work
1.3558 foot pound-force joule (J) 0.7376
1.3558 107 foot pound-force erg 7.3757 108
0.1383 foot pound-force metre kilogram force 7.2307
4.2140 102 foot poundal joule (J) 2.373 103
4.2140 109 foot poundal erg 2.373 1010
1.055 103 BTU joule (J) 9.4787 104
107 joule erg 107
4.1855 calorie at 15C joule (J) 0.2389
Power
550 horse-power foot pound-force per second 1.8182 103
7.457 109 horse-power erg per second 1.341 1010
7.457 102 horse-power watt (W) 1.341 103
1.3405 kilowatt (kW) horse-power 0.74599
1.3558 foot pound-force per second watt (W) 0.7376
Density
16.019 pound per cubic foot kilogram per cubic metre 6.243 102
1.6019 102 pound per cubic foot gram per cubic centimetre 6.243 103
1.0012 ounces per cubic foot gram per litre 0.9988
APPENDICES 837
Unit weight
16.019 pound-force per cubic foot kilogram-force per cubic metre 6.243 102
1.571 102 pound-force per cubic foot newton per cubic metre 6.3654 103
27.68 pound-force per cubic inch gram-force per cubic 3.613 102
centimetre
271.4 103 pound-force per cubic inch newton per cubic metre 3.6846 106
Compressibility
1.45 104 sq. in per pound-force sq. m per newton 6.897 103
14.22 sq. in per pound-force sq. cm per kilogram-force 7.032 102
9.324 106 sq. foot per ton-force sq. m per newton 1.0725 105
0.914 sq. foot per ton-force sq. cm per kilogram-force 1.0941
Speed
2.54 inch per second centimetre per second 0.3937
30.48 foot per second centimetre per second 3.281 102
0.447 mile per hour metre per second 2.2371
0.5144 knot (Int.) metre per second 1.944
1.0973 foot per second kilometre per hour 0.9113
0.61818 foot per second mile per hour 1.61765
0.5925 foot per second knot (Int.) 1.6878
3.6 metre per second kilometre per hour 0.2778
0.9659 108 foot per year metre per second 1.0353 108
Rate of flow and discharge
2832 cubic foot per second cubic centimetre per second 3.53 104
2.832 102 cubic foot per second cubic metre per second 35.311
76 464 cubic yard per second cubic centimetre per second 1.3078 105
0.7646 cubic yard per second cubic metre per second 1.3078
101.941 cubic foot per second cubic metre per hour 9.8096 103
2446.57 cubic foot per second cubic metre per day 4.0874 104
28.3161 cubic foot per second litres per second 3.5315 102
11.573 75 cubic metre per day litres per second 8.6402 102
0.408 735 cubic metre per day cubic foot per second 2.4466
4.3813 102 million gallons (US) per day cubic metre per second 22.824
5.261 102 million gallons (UK) per day cubic metre per second 19.008
0.2713 acre-feet per day million gallons (UK) per day 3.686
0.3259 acre-feet per day million gallons (US) per day 3.0684
Yield
0.699 725 cubic feet per acre cubic metres per hectare 1.4291
1.120 85 pounds per acre kilograms per hectare 0.892 18
Coefficient of consolidation
0.1075 sq. in per minute sq. cm per second 9.3023
1.075 105 sq. in per minute sq. m per second 9.3023 102
2.94 105 sq. ft per year sq. cm per second 3.4014 104
2.94 109 sq. ft per year sq. m per second 3.4014 108
Concentration
1.0012 103 ounce per cubic foot gram per cubic centimetre 9.988 102
1.0012 ounce per cubic foot kilogram per cubic metre 0.9988
16.019 103 pound per cubic foot gram per cubic centimetre 62.426
16.019 pound per cubic foot kilogram per cubic metre 6.2426
Dynamic viscosity
47.8803 pound-seconds per square foot newton-seconds per sq. metre 2.0885 102
47.8803 slugs per foot-second newton-seconds per sq. metre 2.0885 102
103 centipoise (cP) newton-seconds per sq. metre 103
Kinematic viscosity
9.290 30 104 sq. foot per second centistoke (cSt) 1.0764 105
9.290 30 102 sq. foot per second sq. metre per second 10.764
106 centistoke (cSt) sq. metre per second 106
Table A6.3 Conversion tables for slope data.
Degrees to percentages Percentage to degrees Natural tangent to degrees
Degrees Minutes Per cent Degrees and Ratio (tan)1 Degrees (rounded)
0ⴕ 30ⴕ minutes
1 : 0.25 76
0 0 0.87 1 034
1 : 0.5 63
1 1.75 2.62 2 109
1:1 45
2 3.49 4.37 3 143
1 : 1.5 34
3 5.24 6.12 4 218
1:2 27
4 6.99 7.87 5 252
1 : 2.5 22
5 8.75 9.63 6 326
1:3 18
6 10.51 11.39 7 400
1 : 3.5 16
7 12.28 13.17 8 434
1: 4 14
8 14.05 14.95 9 509
1
The first figure refers to the height, the
9 15.84 16.73 10 543
second to the horizontal distance; e.g. on

10 17.63 18.53 11 617
a 1 : 2 smooth slope there is a climb (or

11 19.44 20.35 12 651
descent) of 5 m for every 10 m traversed

12 21.26 22.17 13 724
horizontally along a line parallel to the

13 23.09 24.01 14 758
maximum gradient.
14 24.93 25.86 15 832
15 26.80 27.73 16 905
16 28.68 29.62 17 939
17 30.57 31.53 18 1012
18 32.49 33.46 19 1045
pe
19 34.43 35.41 20 1119
2 slo
n
20 36.40 37.39 25 1402
1i 5m
21 38.39 39.39 30 1642
22 40.40 41.42 35 1945
23 42.45 43.38 40 2148
24 44.52 45.57 45 2414
25 46.63 47.70 50 2634

10 m
27 (rounded)
APPENDICES 839
Degrees to percentages Percentage to degrees
Degrees Minutes Per cent Degrees and

0ⴕ 30ⴕ minutes
26 48.77 49.86 55 2849
27 50.95 52.06 60 3038
28 53.17 54.30 65 3301
29 55.43 56.58 70 3500
30 57.74 58.91 75 3652
31 60.09 61.28 80 3840
32 62.49 63.71 85 4022
33 64.94 66.19 90 4200
34 67.45 68.73 95 4332
35 70.02 71.33 100 4500
36 72.65 74.00
37 75.36 76.73
38 78.13 79.54 Percentage slope natural tangent 100 thus 1 : 2 50%.
39 80.98 82.43
40 83.91 85.41
41 86.93 88.47
42 90.04 91.63
43 93.25 94.90
44 96.57 98.27
GFE-Appendix.qxd
Table A6.4 Guide to map scales and mapping resolution.
840
Map scale Feet on ground Metres on ground Inches on map Centimetres on True ground width of map lines of various

per inch on per centimetre per mile on map per kilometre thickness (mm). Ground widths are given in
map on map ground on ground metres
6/27/05
0.10 0.13 0.25 0.35 0.50 1.00
1 : 500 41.67 5.00 126.72 200.00 0.05 0.07 0.13 0.18 0.25 0.50
1 : 600 50.00 6.00 105.60 166.67 0.06 0.08 0.15 0.20 0.30 0.60
1 : 1000 83.33 10.00 63.36 100.00 0.10 0.13 0.25 0.35 0.50 1.00
4:49 PM
1 : 1250 104.17 12.50 50.68 80.00 0.13 0.16 0.31 0.44 0.63 1.25
1 : 1500 125.00 15.00 42.24 66.67 0.15 0.20 0.38 0.53 0.75 1.50
1 : 2000 166.67 20.00 31.68 50.00 0.20 0.26 0.50 0.70 1.00 2.00
1 : 2500 208.33 25.00 25.34 40.00 0.25 0.33 0.63 0.88 1.25 2.50
Page 840
1 : 3000 250.00 30.00 21.12 33.33 0.30 0.39 0.75 1.05 1.50 3.00
1 : 5000 416.67 50.00 12.67 20.00 0.50 0.65 1.25 1.75 2.50 5.00
1 : 6000 500.00 60.00 10.56 16.67 0.60 0.78 1.50 2.10 3.00 6.00
1 : 10 000 833.33 100.00 6.34 10.00 1.00 1.30 2.50 3.50 5.00 10.00
1 : 10 560 880.00 105.60 6.00 9.47 1.06 1.37 2.64 3.70 5.28 10.56
1 : 12 000 1000.00 120.00 5.28 8.33 1.20 1.56 3.00 4.20 6.00 12.00
1 : 20 000 1666.67 200.00 3.17 5.00 2.00 2.60 5.00 7.00 10.00 20.00
1 : 24 000 2000.00 240.00 2.64 4.17 2.40 3.12 6.00 8.40 12.00 24.00
1 : 25 000 2083.33 250.00 2.53 4.00 2.50 3.25 6.25 8.75 12.50 25.00
1 : 40 000 3333.33 400.00 1.58 2.50 4.00 5.20 10.00 14.00 20.00 40.00
1 : 48 000 4000.00 480.00 1.32 2.08 4.80 6.24 12.00 16.80 24.00 48.00
1 : 50 000 4166.67 500.00 1.27 2.00 5.00 6.50 12.50 17.50 25.00 50.00
1 : 63 360 5280.00 633.60 1.00 1.58 6.34 8.24 15.84 22.18 31.68 63.36
APPENDICES 841
Table A6.5 Map and AP scales: relation to actual lengths and areas (metric units).
Scale Distance represented Distance representing Area represented Area representing

by 1 cm (m) 1 km (cm) by 1 cm2 (ha)1 1000 ha (cm2)
1 : 1000 10 100 0.01 105

1 : 2500 25 40 0.0625 1.6 104
1 : 5000 50 20 0.25 4 103
1 : 7500 75 13 0.56 1786
1 : 10 000 100 10 1.00 103
1 : 15 000 150 6.7 2.25 444
1 : 20 000 200 5.0 4.00 250
1 : 25 000 250 4.0 6.25 160
1 : 30 000 300 3.3 9.00 111
1 : 40 000 400 2.5 16.0 62.5
1 : 50 000 500 2.0 25.0 40
1 : 100 000 1000 1 km 1.0 100 1 km2 10
1 : 250 000 2.5 km 0.4 6.25 km2 1.6
1 : 500 000 5 km 0.2 25 km2 0.4
1 : 750 000 7.5 km 0.1 56.25 km2 0.18
1 : 1 000 000 10 km 0.1 100 km2 0.1
1
Effectively the minimum mappable area.
Table A6.6 Imperial units.
Scale in/mi mi/in mi2/in2 ac/in2
1 : 1 000 000 0.063 4 15.782 8 249.097 159 422.09

1 : 500 000 0.126 7 7.891 4 62.274 39 855.36
1 : 253 400 0.250 0 4.000 0 16.000 10 240.00
1 : 250 000 0.253 4 3.945 7 15.569 9 964.16
1 : 126 720 0.500 0 2.000 0 4.000 2 560.00
1 : 125 000 0.506 9 1.972 8 3.892 2 490.88
1 : 100 000 0.633 6 1.578 3 2.491 0 1 594.22
1 : 63 360 1.000 0 1.000 0 1.000 0 640.00
1 : 62 500 1.013 8 0.986 4 0.973 0 622.72
1 : 50 000 1.267 2 0.789 1 0.622 7 398.56
1 : 31 680 2.000 0 0.500 0 0.250 0 160.00
1 : 30 000 2.112 0 0.473 5 0.224 2 143.49
1 : 25 000 2.534 4 0.394 6 0.155 6 99.64
1 : 10 560 6.000 0.166 6 0.027 7 17.728
1 : 10 000 6 .336 0.157 8 0.024 9 15.942
1 : 5 000 12.672 0.078 9 0.006 23 3.986
1 : 2 534 25.000 0.040 0 0.001 60 1.024
1 : 2 500 25.344 0.039 5 0.001 56 0.996
1 : 1 250 50.688 0.019 7 0.000 389 0.249
References
Tennent, R. M. (ed.) (1971) Science Data Book. Oliver
and Boyd, Edinburgh.
GFE-Index.qxd 6/27/05 1:57 PM Page 843
Index
active tectonic environments, engineering properties, 742, rockfalls, 739

218–62 744, 745 slumps, 740–41
see also seismic hazards England, 103 washouts, 740–41
aerial photograph, 233–34, 236, foundations on, 743, 749–51 change
242 mechanical properties, 746–48 cyclical, 4
avalanche physical properties, 744–45 episodic, 4
classification, 526 tunnelling, 751, 753 periodic, 4
ice, 527–28 chalk components, 729–30 secular, 5
snow, 524–27 aragonite, 729 channel changes, 263–66
calcite, 729 arroyos, 263
calcium carbonate, 734 channel scour, 265
bank erosion, 273–74
coccoliths, 729 gravel extraction, 266
bank failure, 274, 279
chalk geomorphology, 730–42 gulley erosion, 263
barred estuaries, 536–37
bournes, 732–33 Huang He, 264
basalt, 645
cavern systems, 736–37 urbanisation, 266
bearing capacity, 166–67
cavities, 735 channel form, 266–70
bed scour, 274–77
coombes, 733–34 bankfull discharge, 269
biostratigraphy, 112–13
deneholes, 738 classification, 275–76
Bond cycles, 127
dissolution, 734 hydraulic geometry equations,
buttes, 98
drainage patterns, 732 268
dry valleys, 732–33 stream power, 270
Cainozoic, 124–26 erosion and mass movement, clay, loading compaction of,
catchment, 263, 412 739–41 322–24
network, 10 see also chalk mass Acceptable Bearing Pressures,
catchment runoff estimation, movement 322
293–95 flash floods, 741 Leaning Tower of Pisa, 323
runoff curve number, 294 frost churning, 743 Mexico City, 324
time of concentration, 294 frost shattering, 732, 743 smectite-rich, 323
unit peak discharge, 294 mass wasting, 732 clay subsidence, 318–24
caves, 332, 670–72 obsequent streams, 732 due to fluid abstraction,
engineering implications, pipes, 735–36 319–22
672–75 river piracy, 732 Santa Clara Valley, 320
geophysical surveys, 674 shafts, 737 Venice, 321
gypsum, 675 sinkholes, 734–36 climate and flooding, 42–44
Chalk solution pipes, 734, 738 ENSO, 42
associated deposits, 731 spring line, 741 climate change
classifications, 742, 746–48 spring sapping, 733 Bond cycles, 127
distribution, 730 subsidence due to mining, Cainozoic, 124–26
flint, 729, 731, 737 737–39 coastal zone, 567
pillared workings, 738 void migrations, 737 Dansgaard-Oeschger events,
stratigraphy, 731 water table, 732–33 126
structural trends, 732 chalk landscapes, 729–56 El Niño Southern Oscillation,
chalk conceptual block model, 743 127
Channel Tunnel, 753 chalk mass movement, 739–41 four orders of, 401
earthworks in, 751 cliffs, 739 global, 44
engineering, 742–53 debris flows, 740 Henrich events, 127
engineering classification, 749 landslides, 739–40, 742 ice coring, 126–27
843
844 INDEX
climate change (cont.) wave surge currents, 608 sediment sequence, 555
interglacials, 125 wind drift currents, 608 depositional environments in
Milankovitch cycles, 401 continental shelves, 603–32 deltas, 556–62
regolith transformation, 458 Amazon shelf, 619–20 abandoned distributaries,
sediment coring, 127 geotechnical problems, 558–60, 561
variability, 41 613–14 bay-filling episodes, 557,
climate classification, 37–41 hydraulic regime, 608 560
effective precipitation, 38 hydrodynamic regime, 607–11 channel-mouth bars, 560–61
thermal efficiency, 38 morphology, 603 crevasse splay, 556–57, 558
morphoclimatic zones, 38–40, organism-sediment interaction, inter-distributary bays, 556–58
42 614 sub-aqueous slopes, 561–62,
Thornthwaite’s, 39 property of shelf sediments, 563
climate records 611–14 see also shelf desert
analysis of, 20 sediments aggregates, 439, 440, 441
Gumbel distribution, 20 sedimentation, 605–7 see also aggressive salty ground, 446–50
moisture balance index, 20 shelf sedimentation aridity index, 419
coastal cell analysis, 12, 578–79 tectonically active margins, erosion and instability, 437–38
cell delimitation, 578 603–4 excavation, 438
coastal geomorphology, 579–81 tectonically buttressed margins, desert base level plains, 434–37
relative sea-level change, 604–5 depressions, 437
581–83 tectonically passive margins, enclosed basins, 435–37
sediments sources/sinks, 578 604 playa, 436
coastal morphology tide-dominated sandy shelves, sabkhas, 436–37
coastal barriers, 588 622–25 desert footslopes and fans,
controls, 587 conversions, units, 838–45 alluvial fans, 425–28
fringing beaches, 587, 588–91 creep, 59 fan types, 427
macro-tidal, 585 cryogenic processes, 359, 376 piedmont plains, 425
wave shoaling slope, 585 pediments, 424–25
wetland areas, 589 dating sedimentary processes,
coastal zone, 566 amino acid, 26 425–27
barriers, 592–93 caution-ratio rock varnish, 26 segmented fan, 427
behaviour systems, 584–83 cosmogenic series, 25 desert landscape, 419–21
development on, 765–66 dendochronology, 25 Algeria, 420
engineering intervention, fission-track, 113 pediments, 424–25
591–92, 593–98 lichenometric, 25, 113 terrain model, 422
future possible scenarios, 567 luminescence, 25 desert plains, 428–34
inherited engineering radioactive, 24 alluvial plains, 428–31
structures, 598 radiometric, 110–11 dune mobility, 443–46
landslides, 765 uranium series, 25 plains and plateaus, 431–32
sediment losses, 575 deforestation, 414 sand dunes, 432–34
sensitivity and response, 570 delta dynamics, 554–56 desert uplands, 421–23
storm surge, 575 conditions for building, 554 backslopes, 421
time and space scales, 568–69 progradation, 554 flatirons, 423
wave energy, 575 delta switching, 555, 557 pediments, 421
collapsing soils, hydrocompaction deltaic deposits, 555 rampart, 423
of, 326–27 facies changes, 556 scarp slopes, 421–23
San Joaquin Valley, 327 deltaic plain, 552–54 talus slopes, 423
complex behaviour, 14 components, 552–54 drainage basin, 412–13
continental shelf hydrodynamic deltas, 552–64 catchment, 412
regime, 607–11 classic, 552 channelization, 412, 414
Agulhas Current, 608 components, 552–54 hydrological cycle, 412
Coriolis effect, 608, 610 deposition in channels, 561 water balance equation, 412
density currents, 610–11 depositional environments, dust storm, 689–90
meteorological currents, 608–9 56–62
oceanic currents, 607–8 dynamics, 554–56 Earth movements, 93–95
storm surge, 609 engineering works, 558, 562 crustal sag, 93
tidal currents, 609–10 morphology and denudation rates, 93
wave drift current, 609 sedimentation, 553 isostatic rebound, 93
INDEX 845
earthquake erosivity of rainfall, 293 feedback mechanisms

engineering studies, 232 kinetic energy of the rain, 293 negative, 14
location, 232 rainfall intensity, 293 positive, 14
magnitude, 222–23, 238–44 estuaries, 535–52 floods, 281–83
mechanics, 220–22 and engineering, 551–52 floodplain accretion, 277–78
occurrence, 220–23 barred, 536–37 fold structures, 88–90
paleoliquefaction features, 244, definition, 535 fold mountains, 90
248 depositional environments, 541 monocline, 90
recurrence, 244 fjords, 537–38 schistosity, 90
recurrence intervals, 244–51 fresh/salt water mixing, 540 fractures, joints and faults,
risk, 218 morphology, 536 90–93
surface deformation, 240 processes and forms, 541–44 bedding planes, 92
engineering salinity, 540–41, 549 fault breccia, 92
basalt, 645 salinity gradient, 540–41 fault zone, 92
caves, 672–75 salt wedge, 540 frost
coastal zone, 566, 568 sediment sources, 546–47 churning, 743
fossil periglacial forms, sediment transport, 540 shattering, 732, 743
392–94 tectonic, 537–38
geomorphological mapping, tidal cycle, 540 geohazards, 17
18–19 types, 536–37 mountain, 531
geomorphology, 1, 2, 415 unmodified drowned river savanna, 465
glacial environments, 361–72 valleys, 536–37 subsidence, 341
intervention in coastal zones, estuarine classification, 539 temperate zone, 416
567, 591–92 estuarine components, 539 volcanic eruptions, 638–39,
karst, 668–70 estuarine dynamics, 539–41 650–53
permafrost environments, estuarine sedimentation, 544–51 geological maps, 121–22
394–95 depositional dynamics, 541 solid, 121
savanna, 464 depositional facies, 546–48 superficial, 122
soft, 284, 416 flocculation rates, 541 geomorphic analysis, 238
temperate environments, 402 mixing processes, 541 geomorphological
time, 6, 10 mudflats, 547–48, 550 mapping, 767–68, 820–28
environmental management, saltmarshes, 547–48, 550 maps, 203, 213, 767
416 sources, 546–47 geomorphological events, 8–10
equilibrium tidal effects, 545–46 estimates, 9
conditions, 13–14 estuary extreme, 41–42
dynamic, 14 deltas, 543 frequency, 9, 12
metastable, 14 depth, 542 magnitude, 12
stable, 13 land reclamation within, 544 probability, 9
static, 13 mouth, 542, 543 return period, 9
steady-state, 14 polderisation, 574 triggering events, 15, 23
unstable, 14 sedimentation, 543 triggering factors, 9, 15
erodibility, 287, 297–301 tidal prism, 542–43 geomorphology
erodibility index, 299, 301 applied, 1
Universal Soil Loss Equation, fault engineering, 1, 7
307 activity, 243 savanna, 454–55
erosion hazard assessment, 301–6 characterisation, 253 glacial erosion, 349–51
dust hazard assessment, 306 characteristics, 220 debris, 349, 350
erosivity, 287, 301 development, 88 friction cracks, 349
potential sand drift, 304 displacement, 223 regelation ice, 349
rainfall aggressiveness, 302 earthquakes, 246 rouche moutonnées, 349
rainfall erosion index, 302 identification and mapping, sediments, 350–51
erosion features, 288–93 232–38 striae, 349
badlands, 291 normal, 224–26 subglacial meltwater, 349
gullying, 290 reverse, 227–28 glacial hazards, 361–62, 364
pedestal, 288 rupture associated with past glacier depositional processes,
piping, 289 slip rate, 250 351
rill, 288 strike-slip, 224–26 lodgement, 351
sheet, 288 types, 219, 224 meltout, 351
846 INDEX
glacier morphology, 345–47 gully erosion, 310–15 dayas, 680

cirque glacier, 347, 348 check dams, 314 dolines, 664–65
facies, 347 gully protection, 313 fengcong, 681, 682
ice cap, 348 savanna, 468–70 fenglin, 681, 682
ice sheet, 348, 353–54 gorges, 667
landsystem model, 347, historical karren, 664
353–61 analysis of map sources, 23 polje, 667
moraine, 362, 364 research, 22 rockhead, 664
nivation hollow, 347, 348 history sinkholes, 664–65
streamlined landform, geological/geomorphological, suffosion, 665
353–54 16 tower, 681
valley glacier, 347, 348, Holocene, 130–31 travertine, 668
354–55 hot wetlands, 473–500 see also tufa, 668
glacier physics, 348–49 tropical karst terrain, 679–83
mass balance, 348, 349, 358 rainforest environment buried plain, 495, 766
pressure melting point, 348 extreme events, 496–97 fluviokarst, 681
glacier sediments, 350–51, 370 geomorphological problems, glaciokarst, 679
geotechnical properties, 364, 496–97 hot wetlands, 493–4
367 types, 476 humid tropics, 492–6
grain/particle size, 350–351, hydrological cycle, 32, 44–47, 412 Kentucky’s Sinkhole Plain, 680
365 groundwater table, 44, 670 palaeokarst, 680
sampling methods, 367 pinnacle, 683
glacier till ice shilin, 683
classification, 350, 351, 353, ground, 382–84 temperate, 680
368–69 intrusion, 383 tropical, 681
geotechnical properties, 364, needle, 382 types, 680
366 pore, 382
glacioaeolian, 359–61 segregated, 383 lag time, 11
coversands, 360 vein, 383 land reclamation within an
glacial flour, 360 igneous rock estuary, 544
katabatic winds, 359 classification, 634 landforms, 7–8
glaciofluvial, 357–58 granite, 104 arches, 101
eskers, 358 interglacials, 130, 345 buttes, 98
kettle holes, 358 duricrust, 51
sandurs, 358 karst, 492–96, 766 escarpment, 97, 104
glaciolacustrine, 358–59 dolomite, 662 etchplain, 51
proglacial lakes, 359, 361, 362, gypsum, 663 fault scarps, 98–100
364 halite, 663 fins, 101
subglacial lakes, 359 limestone, 662 fossil, 8
supraglacial, 358–59, 364 marble, 662 graben, 100
glaciomarine environments, quartzite, 663 karst, 52
355–57 karst engineering hazards, 495, mesas, 98
glaciomarine sediments, 668–70, 766 plateaus, 98
355 buried sinkholes, 669 relict, 8
mass movements, 355 pinnacled rockhead, 668 stratimorphs, 97–98, 104
ground ice, 382–84 subsidence sinkholes, 669 landscape
frost mounds, 383 karst engineering implications, adjustment delay, 415
intrusion ice, 383 676–79 inherited, 404
needle ice, 382 foundations, 686 sensitivity, 16, 416
pore ice, 382 karstic springs, 678 zones, 405
segregated ice, 383 pollution, 678 landslides, 60, 173–217, 496–97,
vein ice, 383 reservoir impoundment, 678 714–20, 764
ground model, 16 safe bearing pressure, 683 activity and behaviour, 180–86,
groundwater flow, 675–76 terrain classification, 683 188, 189
hydrogeological properties, well yields, 678 avoidance, 775–76
676 karst landforms, 663–70 causes, 186
karst aquifer, 675 collapse, 665 classifications, 179–80, 188
INDEX 847
coastal, 765 potential, 254–56 joints, 702, 708, 712, 713, 714
destructive significance, 190 susceptibility, 256–58 particle shape, 698
dimensions, 176, 178 susceptibility map, 258 particle size, 698, 702
distribution and occurrence, lithification, 74–78 shear strength, 709–11
186–89 cementation, 77–78 soil fabric, 699–711
frequency-magnitude, 183 compaction, 74, 700 tensile strength, 705
hazard, 198, 200, 204 grain packing, 76 loess, landslides in, 714–20
hazard mapping, 201, 204, 776 grain size, 76, 78 classification, 714, 715
high magnitude events, 174–75 porosity, 76 debris flow, 720
Hong Kong, 33 loess, 65, 688–728 flowslides, 718, 719
in chalk, 739–40, 742 bulk properties, 693, 697, 723 main types, 716–19
in loess, 714–20 clay minerals, 705, 706 mass movement types, 715,
infrastructure damage clayey, 688, 707 719
classification, 206 collapsibility, 696–97, 708–9 mudflows, 719–20
investigation, 205–6 deposition, 688 spreading slide, 717
morphoclimatic regions, distribution in Eurasia, 692 types based on geometry and
189–98 engineering geomorphological mechanism, 719–20
mountains, 508–13 issues, 711–23 (and see loess regions, 690
Philippines, 33 separate entry) loess, soil fabric, 699–711
processes causing, 185–86 engineering properties, 690, Atterberg limits, 706, 713
rainfall thresholds, 487 697–711 bulk density, 706
recognition, 180, 187 facies, 701–2 cementation bonds, 718
risk, 198, 202 formation, 690 clay coatings on grains, 700,
risk assessment, 200 grain size, 689, 690 702
savanna, 466 landscape, 696 clay mineral bonding, 709
scars, 409 limon à doublets, 701, 703 cohesion, 709, 711
site investigation, 207 mineral dust, 688 collapsibility, 696–97, 708–9
‘situations’, 198 morphology, 693 compaction, 700, 710
stabilisation, 206 see landslide particle size, 690, 693, 695 consolidation, 700, 709, 711
stabilisation sandy, 688, 696, 699, 707 grain shape, 700
susceptibility, 198, 200, 204 subsidence, 697 hydrocompaction, 711, 723
terrain prone to, 189–98, see terrain, 690–97 liquid limit, 706, 707
landslide prone terrains loess, engineering moisture content, 706, 709
triggered by urban geomorphological issues, packing, 700, 701
development, 770–1 711–23 particle size, 701, 702, 709,
types and mechanisms, 176, desiccation, 713 723
177, 179–80 desiccation cracks, 711 particle size distribution, 700,
volume estimation, 181 hydrology and pseudo-karst, 709
world’s largest, 102 711–14 permeability, 703, 706, 708
landslide prone terrains, 189–98 landslides, 714–20 see also permeability rates, 703, 711,
cliff systems, 193, 197 loess, landslides in 716
eroding coastlines, 192 liquefaction, 712 plastic limit, 706
landslide stabilisation, 206–14 mass movement, 714–20 plasticity index, 706, 707, 709
case studies, 211–14 overland flow, 712, 713 pore pressure, 709
investigation, 208 pseudo-karst, 712, 714, 718, porosity, 706, 713
remediation, 206, 209–10 719 quick clays, 709
lava sinkholes, 712 soil strength, 704
basalt, 660 slope failure, 714 void ratios, 703, 709, 711
flows, 636–37 soil erosion and sediment loess terrain, 690–7
tephra, 638 yield, 720–23 accretion, 696–7
tubes, 637, 638, 646, 647 loess, engineering properties of, adobe, 698
limestone and gypsum, subsidence 697–711 alluvial fans, 695
on, 329 bearing capacity, 710, 711 bypass drainage, 713, 714
liquefaction, 253 bulk properties, 697, 723 Chinese Loess Plateau, 693,
causes, 254 comminution limit, 697 696, 698, 701, 704, 706,
failure modes, 256 geotechnical properties, 710 711, 720
hazards, 257 glacial comminution, 697–98 dust falls, 691
848 INDEX
loess terrain (cont.) periglacial landforms, 386–89 probabilistic seismic hazard

geomorphology of, 692 altiplanation terraces, 386 assessment, 251–53
gobi, 695 beaded drainage, 389
major morphological features block fields, 386 Quaternary, 119–20, 131, 236
in, 692 frost mounds, 389 Dating, 120
palaeosols, 694 (and see ice-wedge polygons, 386, ‘drift’maps, 121
separate entry) palsas, 389 Re. engineering excavation,
pedogenesis, 695 patterned ground, 387–88 120
piping, 708, 713 pingos, 389
river terraces, 695 solifluction stripes, 386, radiometric dating, 110–111
loessification, 689 periglacial processes, 378–82 rates of denudation, 591
cambering, 394 relative sea-level change, 570–72
mining, subsidence due to, 335–36 frost action, 378–80 Brunn’s rule, 572
areal subsidence, 336 frost cracking, 380 cannibilisation, 572
longwall mining, 335 solifluction, 380 coastal cell analysis, 581–83
mine shafts, 339 periglacial soils, 389–91 eustatic, 570
pillar-and-stall mining, 336, gleization, 390 geodetic, 570
338 permafrost, 129, 376–78, 384–86, isostatic, 570
stope working, 336 758 periodic effects, 571
mine workings, 770 active layer, 384 regressive response, 571
mountain dry permafrost, 384 return period, 571
avalanches, 524–28 permafrost table, 384 transgressive response, 571
belts, 501 thaw-bulb, 385 risk, 198, 202, 219
classification, 507 wet permafrost, 384 assessment, 200
climate, 503–05 permafrost, subsidence on, flood, 761
construction materials, 528–29 327–28 river terraces, 278–81
debris flows, 513–14 active layer, 327 rock
flash floods, 514–17 adfreeze piles, 327 bursts, 88
glacial lake outburst floods, conservation of, 328 classification, 804–819
519–23 talik, 327 cleavage, 95–96
hazards, 531 Trans-Alaska Oil Pipeline, deformation, 87–88
landslide dams, 517–19 329, 396 identification, 804–819
terrain model, 505–6 utilidors, 329, 395 mean fracture spacing, 95–96
types, 501, 502 physical systems, 10–14 Rock Mass Factor, 95–96
mountain landslides, 508–13 piping failures, 332 Rock Quality Designation,
Bairaman, 511 planform types, 270–73 95–96
catastrophic, 510–12 active meanders, 270 Safe Bearing Pressure, 96, 104
Huascaràn, 511 anastomosed channels, 272 slope failures, 10, 96
Mount Zandila, 511 braided channel, 272 unconfined compression
Nikawa, 512 meander wavelength, 270 strength, 149
rock spreading and sagging, passive meanders, 270 runoff, 293–96
508–10 straight alluvial channels, 270 curve number, 297, 298, 299
mountains, weathering in, 502–5 plate convergence, 635 infiltration capacity, 293
andesite, 635, 660 Manning’s equation, 295
needle ice, 382 convergent boundaries, 85 Reynold’s number, 296
nuées ardentes, 73, 649 plate divergence, 634–35
basalt lavas, 635, 660 salt subsidence, 329–31
palaeohydrology, 407 divergent boundaries, 85 brine-pumping, 330
palaeosols, 694, 696, 699, 701, ophiolite sequences, 634 Cheshire Plain, 329
704, 708, 716 plate tectonics, 85–87 flashes, 330
paraglacial environments, 124 orogenesis, 85–87 linear subsidences, 329
peat polar climates, 35–37 wet rockhead, 329
head decline, 326 porosity, 76 saltation, 64–65
Holme Post, 325 precipitation saltmarsh, 574
subsidence on, 324–326 debris slides, 409 reclamation, 574
urban development on 767 global mean annual, 33 savanna
Wastage, 325 high intensity, 409 climate history, 455
INDEX 849
distribution, 455 cross-bedding, 69, 70, 71 river-mouth bypassing, 616

geomorphology, 454–55 cross-lamination, 69, 71 sea-level change, 607
savanna duricrusts, 456–58 from oscillatory movements, shoal-retreat massifs, 627
calcrete, 457 70–72 shoreface bypassing, 616
detrital ferricrete, 457 grain size, 69–70 tectonics, 605
psuedo-karst, 457 migration, 68 textural gradients, 616
silcrete, 457 nuées ardentes, 73 tide-dominated, 622–25
tufa, 457 ripples, 67 types, 606
void ratio, 457 sequence, 68 wave and storm dominated,
savanna regolith, 455–60 terminology, 66 625–26
bioturbation, 459 turbidity currents, 72–73, 4 Yellow Sea, 631–32
change, 458 sedimentary rock, 74–83 shelf sediments, properties,
dambos, 458, 460 chemical and organic, 82 611–14
features and landforms, 460 chert, 82 aggregation, 613, 619
materials, 458–60 classification of limestones, 81 bed shear stress, 611
toposequences, 458 clay, 105 bioturbation, 615
transformation due to climate, dolomites, 80 carbonate sediments, 612–13
458 facies, 82–83 cohesive, 611, 613
vertisols, 458 limestones, 80–81 compaction, 613
weathering, 461–64 limestones, components, 81 fluid muds, 619
savanna weathering, 456, 465 lithification, 74–78 muddy shelves, 617–22
collapsible soils, 456 maturity, 80 non-cohesive, 611
duricrust, 456–57 sandstones, 78 shear velocity, 611
kaolinite clays, 456 siliciclastic, 78–80 sinkholes
laterite, 456 sorting, 79 buried, 669
saprolite, 456 texture, 79 karst, 664–65
weathered mantle, 455 sedimentology, 57–58 subsidence, 669
sea level seismic hazards, 218–62 see also site investigation, 207, 774
changes, 116, 131–34, 567, earthquake slope failure, 60, 714, 763
570 assessment, 251–53 block glides, 60, 61
eustacy, 131, 567, 572 earthquake magnitude, 222–23 debris flow, 60, 61
Flandrian transgression, 131 earthquake mechanics, 220–22 instability, 496
glacio-isostasy, 131, 133 earthquake occurrence, 220–23 landslides, 60
mean temperatures, 34 earthquake risk, 218 pore pressure, 61
sediment, 4, 400–411 faults, 220 see also fault rock, 10, 96
bed forms, 66–74 liquefaction and ground savanna, 466
budgets, 411 movements, 253–58 see slumps, 60, 61
cohesive properties, 65–66 also liquefaction subduction stabilisation, 489
conservation, 410 zones, 228–32 see also slope stability, 137, 149, 763–64
density flows, 72–74 subduction factor of safety, 149
erosion, 65–6, 410 seismic stratigraphy, 114 plane failure, 152
mass movement, 59–62, 408 shelf sedimentation, 605–7 pore pressure ratio, 150
movement, 62 Allochthonous regime, 622 stability number, 150
movement, physics of, 62–65 Amazon shelf, 619–20 toppling failure, 152
sources, 410 autochthonous regime, 617, wedge failure, 152
storage, 408–11, 464 622 soil compression, 145–47
transfer, 407–08, 417 cape-associated bars, 628 coefficient of compressibility,
transport and deposition, climate, 606–7 166
58–74, 75 Flandrian transgression, 607 compression index, 146
sediments in the coastal zone Inlet-associated bars, 627 consolidation, 145
offshore losses, 575 palimpsest, 615 over-consolidation ratio, 146
weathering, 576 patterns, 614–15 Virgin Compression Line, 146
sediment transport, 283–84 preservation of nearshore soil erosion
bed load, 283 marine patterns, 627–8 assessment, 829–37
sediment rating curve, 283 preservation of relict features, control methods, 307
suspended sediment yield, 283 626–27 Ethiopian Highlands, 287
sedimentary bed forms, 66–74 relict, 615 Maryland, USA, 287
850 INDEX
soil erosion (cont.) earthquakes along zones, 232, tropical rainforest environment,
rate of, 287 241 473–78
San Gabriel Dam, USA, 288 -related submergence and flood events, 489
savanna, 466 emergence, 228 hillslope profiles, 480
South Downs, UK, 288 subsidence, 769 karst terrain, 492–96
soil identification/classification, areal, 336 landsliding, 496
140, 788–803 classification of damage, 337 rainfall intensities, 474–75
Casagrande plasticity chart, clay, 318–24 rivers and alluvial landforms,
142 environments, 318 487–90
soil loss equation, 20 in loess, 697 rivers in alluvial lowlands, 490
soil profile development, 236–37 mining, 335–36, 737–39 sediment yields, 475, 477, 498
soil properties on peat, 324–26 slope instability, 496
Atterberg limits, 141 salt, 329–31 slope stabilisation, 489
bulk density, 163 soil characteristics, 479–87
degree of saturation, 139 tectonic provinces, 117 terrain, 473
effective stress, 143 subducting margin, 119 types, 476
expansive, 765 tectonic subsidence, 339–41 volcanic terrain, 490–92
fabric, 140 compaction, 339 weathering, 479–85, 486
liquid limit, 141 crustal sag, 340 weathering grades, 479, 481
microstructure, 140 soil liquefaction, 340 weathering products, 483
permeability, 163–65 temperate weathering profiles, 485
plastic limit, 141 climates, 33–35 tropical soil phases, 482
porosity, 139 deposits, 403 turbidity currents, 72–73, 4
shear failure, 143–45 environments, 400, 402 Bouma sequence, 74
specific gravity, 163 hillslopes, 407
specific volume, 139 karst, 680 units, scales and conversions,
strength, 165 landscape systems, 401, 405 838–45
undrained strength, 148 terrain evaluation, 17 urban activities
void ratio, 139 CSIRO, 17 erosion and sediment control
water content, 139 land systems, 17 on construction sites, 775
soil types, 156–62 MEXE, 17 erosion control measures for
clay, 138 models, 17 urban road construction, 776
coarse, 138 units/sub-units, 17 fluvial channel change, 771–72
duricrust, 153 thresholds former mine workings, 770
fine, 138 external, 14 hydrological impacts, 771
peat, 140 internal, 14 impact after development, 769
statis, 10 tidal impact during constriction,
stratigraphical column, 108 delta structures, 13 768
stratigraphy, 107–23 prisms, 13 impact on geomorphic
bedding, 109–10 tidal activity, 572–75 processes, 768–72
biostratigraphy, 112–13 gravitational effects, 572 landsliding, 770–71
chronostratigraphy, 113 surge probabilities, 572 subsidence, 769
correlating sequences, 116–19 tidal range, 572 urban developments
correlation, 109 tide-dominated sandy shelves, alluvial cones, 762
geological time scale, 114–16 622–25 alluvial fans, 762–63
lithostratigraphic correlation, bed load convergence, 622 chief problems, 759
118 bed load partings, 622 coastal sites, 765–66
lithostratigraphy, 111–12 mud zones, 623 contributory factors to slope
non-radiometric chronology, sand patches, 623 failure, 763–64
113–14 sand ribbons, 622 expansive soils, 765
Quaternary, 119–20 and see sand ridges, 623–24 flood risk areas, 761
separate entry sand waves, 623 floodplains, 761–62
seismic, 114 tracers foundation conditions, 768
subducting margin, 119 Chesil Beach, 23 geomorphological issues,
tectonic provinces, 117 radionuclides, 23 760
Walther’s Law, 112 stable isotopes, 23 glacial and fluvioglacial
subduction tropical climates, 31–33 deposits, 761
INDEX 851
in periglacial areas, 758–61 visual identification, 782–88 engineering implications,

land suitability, 769 volcanic eruption 655–58
landslide, 764 hazards, 638–39, 660 hazard zoning, 652
on alluvial fans, 762–63 jökulhlaup, 639 ignimbrites, 654, 657
on peat, 767 prediction, 639 Katmai, 654, 656
permafrost, 758 styles, 636, 637 lahars, 650
piping, 764 volcanic landforms, 636–48 monitoring, 653
road construction, 761 aa, 640 Mt St Helens, 651, 652
salt problems, 764–65 lava flows, 636–37 nuées ardentes, 649
siting and planning, 758 pahoehoe, 640 pyroclastic flows, 649
slope stability, 763–64 pillow, 641 styles, 648
till, 761 tephra, 638, 643, 648 Vesuvius, 649
urban geomorphology, 757–79 tuya, 643
growth of Beijing, 757 volcanic subsidence, 331–32 wave and storm dominated sandy
growth of Manchester, 758 deflation, 332 shelves, 625–6
remediation and prevention, volcanic terrain, 490–92 Bering Sea, 625
775–76 volcanoes, 633 Middle Atlantic Bight, 625,
siting developments, 758 andesite, 660 628
urbanisation, landforms created Azerbaijan, 658 sand ridges, 626
by, 772–75 engineering implications, wave energy, 575
accumulation, 772–73 645–48 weathering, 47
erosion avoidance, 775–76 Etna, 639 and landscape, 47–52
extraction, 773–75 explosive, 648 coastal sediments, 576
gravel extraction, 774 intra-plate plumes, 635 grades, 479, 481
landslide avoidance, 775–76 mud, 658–59 process classification, 49
precautions during site plateau basalts, 644–45 processes, 58
investigation, 774 shield, 641–43 profiles, 485
remediation and prevention, smectite-rich, 648 products, 52–53, 483
775–76 vent, 643 salt, 764–65
transfer of material, Hong volcanoes, explosive, 648–58 zones, 50
Kong Airport, 773 calderas, 654, 655 wind erosion, 296–97, 308–10
waste dumps, 773 hazards, 650 wind erosivity index, 297

Geomorphology For Engineers

Uploaded by

Copyright:

Available Formats

Geomorphology For Engineers

Uploaded by

Document Information

Original Title

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

Geomorphology For Engineers

Uploaded by

Copyright:

Available Formats

What are some of the main topics covered in the document?

What are some of the main topics covered in the document?

What are some of the techniques discussed for engineering geomorphology?

What are some of the techniques discussed for engineering geomorphology?

GFE-Prelims.

qxd 6/27/05 2:35 PM Page i

Geomorphology for Engineers

Geomorphology for Engineers

Distributed in North America by

© 2005 P. G. Fookes, E. M. Lee and G. Milligan

Printed and bound in Poland, EU

Foreword xi 2.14 Weathering and geohazards 53

PART I CONTROLS 4.1 Geological structures 85

5.9 Geological maps 121 8.7 Landslides: susceptibility,

11. Soil Erosion 13.3 Glacial erosion and entrainment

15.1 Deposits, sediments and the

16. Hot Drylands 18.8 Slope instability and landsliding 496

18.1 The tropical rainforest

21.5 Coastal organisation 577 24. Karst Terrains

22. Continental Shelves 25. Loess

22.1 Introduction 603 25.1 Introduction 688

23. Volcanic Landscapes 27.1 Introduction 757

APPENDICES Appendix A4 Geomorphological

Denys Brunsden OBE, DSc (Hons), FKC

Contemporary description of the 1575 ‘Wonder Landslide’ in Herefordshire, written by Fuller

1. Introduction to Engineering Geomorphology

1.1 Preamble Engineering and geomorphology

2 GEOMORPHOLOGY FOR ENGINEERS

INTRODUCTION TO ENGINEERING GEOMORPHOLOGY 3

4 GEOMORPHOLOGY FOR ENGINEERS

INTRODUCTION TO ENGINEERING GEOMORPHOLOGY 5

On the coast, the Holocene sea level rise and

6 GEOMORPHOLOGY FOR ENGINEERS

INTRODUCTION TO ENGINEERING GEOMORPHOLOGY 7

8 GEOMORPHOLOGY FOR ENGINEERS

INTRODUCTION TO ENGINEERING GEOMORPHOLOGY 9

10 GEOMORPHOLOGY FOR ENGINEERS

INTRODUCTION TO ENGINEERING GEOMORPHOLOGY 11

12 GEOMORPHOLOGY FOR ENGINEERS

Open boundary Cliff

Variable Sediment Onshore

INTRODUCTION TO ENGINEERING GEOMORPHOLOGY 13

Figure 1.7 Types of equilibrium (after Chorley and Kennedy, 1971).

14 GEOMORPHOLOGY FOR ENGINEERS

feedback operations within the system follow- Complex behaviour

INTRODUCTION TO ENGINEERING GEOMORPHOLOGY 15

coastal cliffs due to the superimposition of trigger-

A. Interaction between declining stability and triggering events

B. Event sequences: redundant events and variable epochs

Epoch 1 Epoch 2 Epoch 3

16 GEOMORPHOLOGY FOR ENGINEERS

INTRODUCTION TO ENGINEERING GEOMORPHOLOGY 17

18 GEOMORPHOLOGY FOR ENGINEERS

Engineering geomorphological mapping 2. The characterisation of these units in terms

INTRODUCTION TO ENGINEERING GEOMORPHOLOGY 19

20 GEOMORPHOLOGY FOR ENGINEERS

INTRODUCTION TO ENGINEERING GEOMORPHOLOGY 21

22 GEOMORPHOLOGY FOR ENGINEERS

7 Recorded landslide events