Effective Ecological Monitoring

EFFECTIVE ECOLOGICAL MONITORING

We are most pleased to dedicate the Second Edition of this book to Phyllis C Likens, now deceased. Her indefatigable spirit, good humour and exceptional support and encouragement for this book project are enthusiastically acknowledged and greatly missed.

EFFECTIVE ECOLOGICAL MONITORING

SECOND EDITION

David B. Lindenmayer and Gene E. Likens

© David Lindenmayer and Gene Likens 2018

All rights reserved. Except under the conditions described in the Australian Copyright Act 1968 and subsequent amendments, no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, duplicating or otherwise, without the prior permission of the copyright owner. Contact CSIRO Publishing for all permission requests.

A catalogue record for this book is available from the National Library of Australia.

Published by

CSIRO Publishing

Locked Bag 10

Clayton South VIC 3169

Australia

Telephone: +61 3 9545 8400

Email: [email protected]

Website: www.publish.csiro.au

Front cover: (clockwise from top left) Dan Florance in the field (photo by Dave Blair); Litoria fallax (photo by Damian Michael); Fungus in the Victorian Central Highlands (photo by Elle Bowd); Bull moose at Isle Royale National Park (photo by Shawn Malone)

Back cover: Eastern yellow robins (photo by Dave Blair)

Set in 11/13.5 Minion & Helvetica Neue LT Std

Edited by Karen Pearce

Cover design by James Kelly

Typeset by Desktop Concepts Pty Ltd, Melbourne

Index by Max McMaster

Printed in China by Toppan Leefung Printing Limited

CSIRO Publishing publishes and distributes scientific, technical and health science books, magazines and journals from Australia to a worldwide audience and conducts these activities autonomously from the research activities of the Commonwealth Scientific and Industrial Research Organisation (CSIRO). The views expressed in this publication are those of the author(s) and do not necessarily represent those of, and should not be attributed to, the publisher or CSIRO. The copyright owner shall not be liable for technical or other errors or omissions contained herein. The reader/user accepts all risks and responsibility for losses, damages, costs and other consequences resulting directly or indirectly from using this information.

Original print edition:

The paper this book is printed on is in accordance with the rules of the Forest Stewardship Council®. The FSC® promotes environmentally responsible, socially beneficial and economically viable management of the world’s forests.

Contents

Acknowledgments

Preface to Second Edition

Chapter 1 Introduction

Some of the ecological values and uses of long-term datasets

Time until expression

Informing policies and legislation in environmental management

Use in simulation modelling

Tests of ecological theory

Development of co-located, collaborative and multidisciplinary work

Detection of surprises

Poor record of long-term ecological monitoring

Why we wrote this book

1. Societal need

2. Correcting the record – countering the perception that long-term studies in ecology are poor quality science

3. Making sense of the vast monitoring literature

4. Providing an overview of success and failure

5. New perspectives

References

Chapter 2 Why monitoring fails

Characteristics of ineffective monitoring programs

Failure to ask the preliminary and fundamental question – Is monitoring needed at all?

Passive, mindless and lacking questions

Lack of trigger points for action

Poor experimental design

Snowed by a blizzard of ecological details

Squabbles about what to monitor – ‘It’s not monitoring without the mayflies’

Assumption that ‘one size fits all’

Big machines that go ‘bing’

Disengagement

Rush to get ‘real work’ happening on the ground and accusations of program over-engineering

Poor data management

Breaches of data integrity

Other factors contributing to ineffective monitoring programs

Lack of funding – grant myopia

The loss of a champion

Out of nowhere

Excessive bureaucracy

Summary

References

Chapter 3 What makes long-term monitoring effective?

Characteristics of effective monitoring programs

Good questions and evolving questions

The use of a conceptual model

Selection of appropriate entities to measure

Good design

Well-developed partnerships

Strong and dedicated leadership

Potential to identify key emerging issues

Ongoing funding

Frequent use of data

Scientific productivity

Maintenance of data integrity and calibration of field techniques

Little things matter a lot! Some ‘tricks of the trade’

Field transport

Field staff

Access to field sites

Time in the field

The adaptive monitoring framework

Examples of the adaptive monitoring framework

Adaptive monitoring is a general and not a prescriptive framework

Increased future role for adaptive monitoring

Summary

References

Chapter 4 The problematic, the effective and the ugly – some case studies

The problematic

PPBio Australasia

The Alberta Biodiversity Monitoring Program (ABMP)

EMAP

The effective

Rothamsted

Ecosystem Health Monitoring Program (EHMP) for Moreton Bay in South East Queensland, Australia

The Hubbard Brook Ecosystem Study

The Central Highlands of Victoria, south-eastern Australia

Need to wait and see

NEON/TERN

The ugly

Summary

References

Chapter 5 The upshot – our general conclusions

Changes in culture needed to facilitate monitoring

The academic culture and rewards systems

Structure of organisations

Funding

Societal culture

Good things that can come from non-question based monitoring

The role of citizen science in long-term monitoring

The challenge of intellectual property and data sharing

The challenges in effective monitoring of rare, threatened and endangered species

The major challenge of keeping monitoring and long-term studies going

The big issue of integrating different kinds of monitoring

Approaches to integrate data from different kinds of monitoring

The challenges posed by differences in the kinds of entities that are monitored in different ecosystems

Using environmental and economic accounts as a way to demonstrate the value of monitoring and cement support for monitoring in place

Concluding remarks

References

Index

Acknowledgments

Since the completion of the first edition of this book, DBL would like to acknowledge the support of many organisations who have begun to or continued to support long-term work in south-eastern Australia including the Australian Research Council, the Commonwealth Department of Environment and Energy (and its various previous incarnations), Murray Local Land Services, Riverina Local Land Services, Parks Australia, Victorian Department of Environment, Water and Land Planning (and its many earlier incarnations), Parks Victoria, the now sadly defunct Land and Water Australia, the Ian Potter Foundation, the Thomas Foundation, and the Graeme Wood Foundation. Data from long-term studies have been guided in their collection and then expertly analysed by truly outstanding professional statisticians, especially Ross Cunningham, Jeff Wood, Wade Blanchard and Alan Welsh. The field research officers who expertly gathered high-quality data over many years have included David Blair, Mason Crane, Daniel Florance, Ryan Incoll, Chris MacGregor, Lachlan McBurney, Damian Michael, Rebecca Montague-Drake, Sachiko Okada, Thea O’Loughlin and Matthew Pope. Finally, many of the key insights in this book have been shaped through collaborations with the many outstanding scientists in both the Australian Government’s Threatened Species Recovery Hub and the Australian Long-Term Ecological Research Network (LTERN) that is part of the Terrestrial Ecosystem Research Network (TERN) (especially Emma Burns).

DBL would particularly like to thank Karen Viggers, Ryan Lindenmayer and Nina Lindenmayer for their understanding of a husband/father spending too much time parked in front of a computer.

GEL initiated the Hubbard Brook Ecosystem Study in 1963 with colleagues, F Herbert Bormann, Robert S Pierce and Noye M Johnson, and has continued to work within this large and complicated ecosystem study since then. Throughout the years, numerous other colleagues, students and technicians have contributed to this study. John S Eaton, Donald C Buso and the late Phyllis C Likens, three long-term support staff, and the current field and data management staff, Tammy Wooster, Brenda Minicucci and Geoff Wilson, contributed in major ways to my efforts and learning within the Hubbard Brook Ecosystem Study. Long-term financial support for these studies was provided by the National Science Foundation, including the Long-Term Research in Environmental Biology and Long-Term Ecological Research programs and The Andrew W Mellon Foundation. Additional support to GEL in the writing of this second edition was provided by the Australian Research Council (via a grant to David Lindenmayer) and The Australian National University, the Cary Institute of Ecosystem Studies and the National Science Foundation through the Hubbard Brook Long-Term Ecological Research and Long-Term Research in Environmental Biology programs. GEL also would like to thank the following people for their contributions to this book: Tammy Wooster, Brenda Minicucci, Donald C Buso, Jonathan Cole, Daniel Schindler, John Pastor, Michele Burford, Matthew Gillespie, Sylvia Lee, Stan Temple, Emily Saeck, Jerry Franklin, Adam Wilson, Amy Schuler, Deborah Fargione, Kathleen Weathers and Jim Nichols.

Tabitha Boyer completed a range of key editing and other tasks associated with the preparation of this manuscript. We thank her for these efforts.

Boxes for the text were written by Charles Krebs and John Pastor. In addition, Penny Olsen and Stephen Garnett provided an update on heroic efforts to recover the Norfolk Island boobook owl.

We would like to thank all of the photographers who contributed images, and Andrew Rankine from Atypica for drawing the figures for this book.

John Manger from CSIRO Publishing has been a strong advocate for this book project and made it clear that a second edition of this book was needed.

Preface to Second Edition

It has been more than eight years since we wrote the first edition of this book, on which we received many comments. Many colleagues asked if we would update the book, particularly given the amount of new science associated with long-term research and monitoring that had been published in the past few years.

The task of recrafting this book was far from straightforward. A lot has happened in eight years. How to do justice to the rapidly expanding body of new science on monitoring, yet ensure that the book was relatively short but also accessible, to the point, and a reasonable representation of our own philosophies and attitudes towards long-term ecological monitoring? We have added some new sections, including those on the uses of new technology in monitoring, the role of citizen science in monitoring, the challenges created by big data, making decisions when to (and not to) monitor, the special case of threatened species monitoring, and some ways to increase the management and policy relevance of monitoring programs, such as through linking the data gathered from them to initiatives like environmental accounting, ecosystem assessments and mandated environmental reporting. We acknowledge that despite these additions, there will still nevertheless be key topics that we have not addressed or only addressed superficially. Omissions were inevitable in trading off manuscript length and readability. Perhaps we might visit missing topics in yet a third revision of this book, but the reality is that the expansion of the scientific literature is now so rapid and substantial that is quickly becoming beyond the ability of most mortals to write texts about even the most highly constrained subject areas.

The structure of this book

This book is about a framework and underlying philosophy for effective ecological monitoring. It is not a manual of field techniques – whether to do point interval counts or transects for monitoring birds or which equipment is needed to track changes in airshed quality.

Our book comprises five short chapters. Readers may wish to go to particular chapters or even read the chapters in random order. The summary and set of references at the end of each chapter are designed to give the reader both a quick overview and an opportunity to pursue particular topics in greater detail. In Chapter 1 we define three broad kinds of monitoring – mindless or passive monitoring, mandated monitoring and question-driven monitoring. We then outline some of the ecological values that can be generated from long-term datasets. We conclude Chapter 1 by discussing the poor record of long-term ecological monitoring programs as one of the primary motivations for writing this book. In Chapter 2 we discuss some of the reasons why many monitoring programs are of poor quality and often fail. We then discuss some of the features of good monitoring programs in Chapter 3. Based on that discussion, we present our adaptive monitoring framework, which incorporates many of the key features and ingredients that should characterise good and effective long-term monitoring programs. In Chapter 4, we present case studies of some good and problematic long-term ecological research and monitoring programs to illustrate the salient points in the previous two chapters. We are acutely aware that the choices for these examples represent our considered opinion and that some readers will be offended by the content of Chapter 4, and we have ventured into this territory with trepidation. Nevertheless, we have written this book in the spirit of constructive criticism in the hope that the problems we identify might be avoided in new monitoring programs or rectified in existing ones. It seems to us that constructive criticism lies at the heart of good science, our emphasis being on ‘constructive’. The final chapter, Chapter 5, not only summarises the main points of our book, but it offers some thoughts about why and how many more long-term monitoring programs might be instigated and how existing ones can be maintained and improved. We also offer some thoughts about how some issues in the future (e.g. the challenges of big data) might be resolved or used to make progress in understanding.

What is new in this version of the book?

This updated version of the book has been informed by much that has happened in the past eight years, including several new books on monitoring and long-term ecological studies. We endeavoured to maintain our focus on what we considered to be important attributes of successful and effective monitoring programs. To this end, we have included new sections on such topics as: (1) how major observation networks like the National Earth Observation Network (NEON) and similar major national and international initiatives link with effective ecological monitoring; (2) the role of citizen science in ecological monitoring; (3) when to (and not to) monitor; (4) the special case of, and particular challenges associated with, monitoring of threatened species; and (5) ways in which the place of monitoring might be better cemented into the psyche (and budget lines) of managers and decision makers. We also have tried to update particular case studies as well as modify our opinions on some topics in the light of new information and extensive discussions with colleagues.

Caveats

This book is not intended to be an exhaustive treatment of the vast amount of literature on monitoring, but adequate references are given to guide the reader to pertinent sources. The book is relatively short, drawing extensively on our personal experiences and perspectives as to what does and does not work in long-term ecological research and monitoring.

Our discussions are based on our respective expertise in long-term population studies and biodiversity monitoring (David B Lindenmayer, DBL) and long-term ecosystem and biogeochemical studies and environmental monitoring (Gene E Likens, GEL). We also have targeted terrestrial and aquatic ecosystems. However, we believe that the general principles we outline are also relevant to marine ecosystems.

We have selected case studies to provide a range of long-term research and monitoring programs. However, we readily acknowledge that this book contains a bias to our own work. We also admit to a bias towards the science of long-term monitoring rather than strictly a policy perspective on this topic. This bias arose because we are both scientists and not policy makers, policy implementers or resource managers. Nevertheless, we readily acknowledge (in Chapter 3) that partnerships among people with different roles and sets of skills are one of the fundamental ingredients of successful monitoring programs.

We have sought to focus more philosophically and conceptually on monitoring programs, and we do not make specific recommendations or give precise ‘recipes’ about ‘how to do monitoring’. We do not present a detailed review of field methods to guide monitoring programs – there is a truly enormous amount of literature on methods that is readily accessible to those initiating new work or revising existing programs. We believe that a monitoring program generally needs to be tailored to a particular question or site, although we acknowledge there sometimes can be a role for surveillance monitoring. Also, we do not examine the equally extensive literature on statistical and analytical methods and experimental designs for monitoring programs. Statistical challenges will often need to be resolved, often in consultation with a professional statistical scientist.

Finally, we aimed to present our opinions without religious-like ‘fundagelical’ fervour. We believe that too much of science is characterised by the attitude that ‘my way is right and yours is wrong’. We certainly recognise that we have made mistakes in our own research and monitoring. As we point out in Chapter 5, there will be occasions where the main ‘rules’ we propose to guide monitoring programs will be breached and yet valuable outcomes are still produced.

David B Lindenmayer and Gene E Likens

May 2017

1

Introduction

Organisms, including humans, depend upon the integrity of ecosystems for their wellbeing and survival. High-quality ecological information collected over long periods provides valuable insights into changes in ecosystem structure, ecological processes, and the services ecosystems provide. Without this information, we would have no knowledge about the changing status of the life support system of the planet (Muir 2010; Lindenmayer et al. 2015b). Nor will we know how effective management has been for actions like weed control, restoration programs, recovery efforts for endangered species, or abatement of nutrient enrichment. For example, a lack of effective monitoring means that Australia is unable to determine the effectiveness of the millions of dollars it invests in biodiversity management (State of Environment 2017). Similarly, there is no credible monitoring program for bees anywhere in the world, despite the serious global declines in this group and the critical role they play in pollination (Inoye et al. 2017). Indeed, the public has every right to ask what scientists and resource managers have been doing in spending of public money over many decades! For the purposes of this book, we consider such information, collected conscientiously and continuously for at least 10 years, and then analysed, to be long-term monitoring (see also Box 1.1).

We see the broad area of monitoring falling into three types:

•Curiosity-driven or passive monitoring . This type is monitoring, sometimes also called surveillance monitoring (e.g. see Wintle et al . 2010) is typically devoid of specified questions, with little or no purpose other than curiosity and no underlying experimental design. It may be done out of inquisitiveness, but often has limited usefulness in addressing environmental problems or in discovering how the world works (but see Olsen et al. 1993). Passive monitoring has these limitations because: (1) it is not hypothesis-driven, and (2) it lacks management interventions or different experimental treatments that facilitate scientific understanding about ecosystem responses to natural or human disturbance. Curiosity-driven, passive or surveillance monitoring could be purely mindless or it could be based on other motivations (see Chapter 4).

Box 1.1. What is not long-term monitoring?

Although one can easily do runs from a simulation model and make projections for thousands of years into the future, we do not consider simulation modelling per se to be long-term work. Studies that substitute space for time (e.g. snapshot investigations (Diamond 1986) or retrospective investigations (Cambridge et al. 2007)) can provide long-term perspectives but we do not consider them to be long-term work per se. We also do not consider haphazard revisits to a site after a prolonged absence to be programs of long-term research or monitoring (e.g. Currie and Parry 1999; Smith et al. 2007). Contrary to widely held beliefs in some organisations, we also do not consider that simply measuring something in the environment constitutes monitoring. Rather, as we outline in the following chapters, monitoring needs to be driven by questions, an experimental design, a conceptual framework, and data integrity through repeatable application of appropriate field protocols (usually the same ones employed through time) (Lindenmayer and Likens 2010).

We acknowledge there are many nuances to discussions about what constitutes ‘long term’ in ecological research and monitoring. The review by Stankey et al. (2003) is instructive. They note that some workers consider long-term studies to be those that continue beyond the generation time of dominant organisms in an ecosystem or sufficiently long to quantify the key processes that structure the ecosystem under investigation. This definition would mean that studies of bacterial assemblages with very rapid generation times would be long-term investigations if they were to persist for a year, a month or even a week. Conversely, a 300-year study of stands of giant sequoia (Sequoiadendron giganteum) in which the dominant trees may live for over 1000 years would not qualify as a long-term study. These considerations are important because they emphasise the variable lifespan of different organisms, but they are not feasible to use for many ecosystem analyses. For the purposes of this book, we use a practical, operational definition and consider long-term as monitoring efforts that continue beyond 10 years.

•Mandated monitoring . This is monitoring for which environmental data must be gathered as a stipulated requirement of government legislation or a political directive, such as monitoring of weather, air quality or river flow funded by governments and conducted by government agencies. For example, populations of many threatened species listed under the US Endangered Species Act must be subject to monitoring, making such species unusual among those rare and/or threatened globally that are typically not subject to rigorous monitoring. The challenges associated with monitoring of rare and endangered species are discussed in Chapter 5. Similarly, Mexico has developed a mandated monitoring and reporting framework for assessing land condition and degradation in that country (Garcia-Alaniz et al. 2017). Rigid quality assurance/quality control protocols are usually strictly stipulated in this type of monitoring. Mandated monitoring is driven by the potential to answer broad and often practical questions. These questions almost always are posed post-hoc in regard to some environmental problem and are not derived from a conceptual model of a particular system or process. Mandated monitoring does not attempt to identify or understand the mechanism influencing a change in an ecosystem or an entity. Rather, the focus is usually to identify trends in a given entity (e.g. whether environmental conditions are getting ‘better or worse’).

•Question-driven monitoring . This type of monitoring is guided by a conceptual model of an ecosystem or some other entity (e.g. a population of organisms) and guided by a rigorous experimental design. The use of a conceptual model will typically result in a priori predictions that are then tested as part of the monitoring program. This kind of monitoring might be undertaken by a single investigator or it might fall under a program like the US Long-Term Research in Environmental Biology (LTREB) program ( https://www.nsf.gov/funding/pgm_summ.jsp?pims_id=13544 ). In question-driven monitoring, mechanisms can be discovered whereby prospective scenarios of trends can be calculated and modelled (see Chapters 3 and 4 ). Often such learning is informed by strongly contrasting management interventions (Carpenter et al . 1995) and in statistical parlance such studies might be best termed ‘longitudinal studies with interventions’ (Cunningham and Lindenmayer 2017). This approach can lead to robust predictive capacity and enable an investigator to pose new questions – an advantageous part of the adaptive monitoring paradigm ( sensu Lindenmayer and Likens 2009) that we outline in Chapter 3 . This predictive capacity can be of immense value for ecologists, resource managers and decision makers (Lindenmayer et al . 2015b). In contrast, in curiosity-driven or passive monitoring, and in mandated monitoring to a lesser extent, the predictive and prospective capability of monitoring can be limited such as simply extending trend lines without supporting data.

Obviously, there can be overlap between broad categories of monitoring. For example, a rigorous statistical framework can characterise both mandated and question-driven monitoring (Cunningham and Lindenmayer 2017), although many long-term studies were never initially designed to be long-term studies (see Box 1.2). There can be advantages and disadvantages of mandated monitoring programs and question-driven research and monitoring. As we discuss later in this chapter, mandated monitoring is often coarse scale, leading to assessments of resource condition, but providing limited understanding of ecological mechanisms that have given rise to that condition. Question-driven monitoring is often the converse. It is finer scaled and often process based (see examples across Australian ecosystems in Lindenmayer et al. 2014c), but it is very difficult to make valid spatial extrapolations to larger scales (e.g. at the state, province or national level). Thus, in Chapter 5, we discuss the challenges posed by the integration of data, insights and management recommendations from these broad kinds of monitoring programs.

Some of the ecological values and uses of long-term datasets

Countless scientific articles, books, management plans and other documents have been written about the need to do long-term monitoring (Strayer et al. 1986; Likens 1989, 1992; Goldsmith 1991; Spellerberg 1994; Thompson et al. 1998; Lovett et al. 2007; Krebs 2008; Muller et al. 2011;