Guidelines for the application of

IUCN Red List of Ecosystems

Categories and Criteria
Edited by Lucie M. Bland, David A. Keith, Rebecca M. Miller,
Nicholas J. Murray and Jon Paul Rodríguez
Version 1.1

INTERNATIONAL UNION FOR CONSERVATION OF NATURE

Guidelines for the Application of IUCN
Red List of Ecosystems Categories and
Criteria
Edited by Lucie M. Bland, David A. Keith, Rebecca M. Miller,
Nicholas J. Murray and Jon Paul Rodríguez
Version 1.1
The designation of geographical entities in this book, and the presentation of the material, do not
imply the expression of any opinion whatsoever on the part of IUCN concerning the legal status of any
country, territory, or area, or of its authorities, or concerning the delimitation of its frontiers or
boundaries. The views expressed in this publication do not necessarily reflect those of IUCN.

Published by: IUCN, Gland, Switzerland

Copyright: © 2017 IUCN, International Union for Conservation of Nature and Natural Resources

The Guidelines for the application of IUCN Red List of Ecosystems Categories and
Criteria is an open access publication licensed under a Creative Commons
Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0).

Reproduction of this publication for educational or other non-commercial uses is

authorized without prior written permission from the copyright holder provided the
source is fully acknowledged.

Reproduction of this publication for resale or other commercial purposes is prohibited

without prior written permission of the copyright holder.

Citation: Bland, L.M., Keith, D.A., Miller, R.M., Murray, N.J. and Rodríguez, J.P. (eds.) (2017).
Guidelines for the application of IUCN Red List of Ecosystems Categories and
Criteria, Version 1.1. Gland, Switzerland: IUCN. ix + 99pp.

Contributors to Version 1.1:

N.J. Murray, L.M. Bland, K.L. Boe, D.A. Keith, R.M. Miller, T.J. Regan, I. Zager

ISBN: 978-2-8317-1769-2
DOI: 10.2305/IUCN.CH.2016.RLE.3.en

Cover photos: © R Ferrari Legorreta, © RM Miller, © NJ Murray

All photographs used in this publication remain the property of the original copyright holder.
Photographs should not be reproduced or used in other contexts without written permission from the
copyright holder.

Available from: IUCN (International Union for Conservation of Nature)

Global Ecosystem Management Programme
Commission on Ecosystem Management
Rue Mauverney 28
1196 Gland
Switzerland
Tel +41 22 999 0000
Fax +41 22 999 0002
www.iucn.org/resources/publications

These guidelines are also freely available online on https://portals.iucn.org/library/node/45794 and the
IUCN Red List of Ecosystems website (www.iucnrle.org). The guidelines are conceived as a ‘living
document’ and will be updated periodically. Please submit your comments and suggestions to
www.iucnrle.org/work-with-us/contact-us. Numerous experts from around the world participated in the
development of these guidelines. The complete list of contributors is located in Appendix 1.
Table of Contents

Table of Contents..................................................................................................................iii
Executive Summary .............................................................................................................. v
Acknowledgements ...............................................................................................................vi
Acronyms ............................................................................................................................ viii
Glossary................................................................................................................................ix
1. Introduction ....................................................................................................................... 1
1.1 Objectives of the IUCN Red List of Ecosystems .......................................................... 1
1.2 Development of the IUCN Red List of Ecosystems ...................................................... 2
1.3 Governance of the IUCN Red List of Ecosystems ....................................................... 3
1.3.1 The Steering Committee ....................................................................................... 3
1.3.2 The Committee for Scientific Standards ................................................................ 4
1.4 Structure of the Guidelines .......................................................................................... 5
2. Categories of the IUCN Red List of Ecosystems ............................................................... 6
3. Scientific foundations ........................................................................................................ 8
3.1 Ecosystem types: the units of assessment .................................................................. 8
3.1.1 Ecosystem typologies ........................................................................................... 8
3.1.2 The influence of scale ........................................................................................... 9
3.2 Ecosystem collapse................................................................................................... 12
3.2.1 Defining ecosystem collapse ............................................................................... 12
3.2.2 Uncertainties in the endpoints for risk assessment .............................................. 15
3.3 Risk assessment protocol.......................................................................................... 18
3.3.1 Time frames ........................................................................................................ 19
3.3.2 Decline thresholds............................................................................................... 20
3.3.3 Standards of evidence and dealing with uncertainty ............................................ 21
3.3.4 Making the most of quantitative data and expert knowledge ............................... 23
4. Assessment process ....................................................................................................... 26
4.1 Area of assessment................................................................................................... 27
4.2 Describing the unit of assessment ............................................................................. 28
4.2.1 Classification ....................................................................................................... 29
4.2.2 Spatial distribution ............................................................................................... 30
4.2.3 Characteristic native biota ................................................................................... 30
4.2.4 Abiotic environment............................................................................................. 32
4.2.5 Processes and interactions ................................................................................. 33
4.2.6 Threats................................................................................................................ 37
4.2.7 Describing collapsed states ................................................................................. 40
4.3 Evaluating the criteria ................................................................................................ 40
4.4 Assessment outcome ................................................................................................ 41
4.4.1 Dealing with uncertainty ...................................................................................... 44
4.5 Documentation .......................................................................................................... 44
5. Criteria and thresholds .................................................................................................... 45
5.1 Criterion A. Reduction in geographic distribution ....................................................... 46
5.1.1 Theory................................................................................................................. 46

iii | IUCN Red List of Ecosystems

 5.1.2 Thresholds and subcriteria .................................................................................. 46
5.1.3 Application .......................................................................................................... 46
5.2 Criterion B. Restricted geographic distribution ........................................................... 52
5.2.1 Theory................................................................................................................. 52
5.2.2 Thresholds and subcriteria .................................................................................. 52
5.2.3 Application .......................................................................................................... 53
5.3 Criterion C. Environmental degradation ..................................................................... 59
5.3.1 Theory................................................................................................................. 59
5.3.2 Thresholds and subcriteria .................................................................................. 59
5.3.3 Application .......................................................................................................... 59
5.4 Criterion D. Disruption of biotic processes and interactions ....................................... 67
5.4.1 Theory................................................................................................................. 67
5.4.2 Thresholds and subcriteria .................................................................................. 67
5.4.3 Application .......................................................................................................... 68
5.5 Criterion E. Quantitative risk analysis ........................................................................ 73
5.5.1 Theory................................................................................................................. 73
5.5.2 Thresholds .......................................................................................................... 73
5.5.3 Application .......................................................................................................... 73
6. Peer review and publication ............................................................................................ 79
7. References ..................................................................................................................... 80
Appendix 1. List of contributors ........................................................................................... 93
Appendix 2. IUCN Red List of Ecosystems Criteria, Version 2.2 ......................................... 95
Appendix 3. Colour codes ................................................................................................... 98
Appendix 4. Revision history ............................................................................................... 99

iv | IUCN Red List of Ecosystems

Executive Summary
The IUCN Red List of Ecosystems is a global framework for monitoring the status of ecosystems. It is
part of the growing toolbox for assessing risks to biodiversity and aims to support conservation,
resource use, and management decisions by identifying ecosystems most at risk of biodiversity loss.
By targeting a level of biological organisation above species, the IUCN Red List of Ecosystems
complements The IUCN Red List of Threatened Species™. The IUCN Red List of Ecosystems
Categories and Criteria are designed to be: widely applicable across ecosystem types and
geographical areas; transparent and scientifically rigorous; and easily understood by policy makers
and the public.

The IUCN Red List of Ecosystems Categories and Criteria

The basis of the IUCN Red List of Ecosystems is the IUCN Red List of Ecosystems Categories and
Criteria, a set of eight categories and five criteria that provide a consistent method for assessing the
risk of ecosystem collapse. The eight categories of ecosystem risk are: Collapsed (CO), Critically
Endangered (CR), Endangered (EN), Vulnerable (VU), Near Threatened (NT), Least Concern (LC),
Data Deficient (DD), and Not Evaluated (NE).

The IUCN Red List of Ecosystems protocol comprises five rule-based criteria (A-E) for assigning
ecosystems to a risk category. Two of these criteria assess spatial symptoms of ecosystem collapse:
declining distribution (A) and restricted distribution (B). Two criteria assess functional symptoms of
ecosystem collapse: environmental degradation (C) and disruption of biotic processes and
interactions (D). Multiple threats and symptoms can be integrated in a model of ecosystem dynamics
to produce quantitative estimates of the risk of collapse (E). The Guidelines include comprehensive
sections to support application of each of the five criteria, including information on relevant theory,
thresholds and examples.

Application and documentation standards

The Guidelines assist correct implementation of the IUCN Red List of Ecosystems Categories and
Criteria by providing information on the development of the protocol and a detailed overview of the
scientific foundations supporting the categories and criteria. They define assessment units
(ecosystem types); define ecosystem collapse; discuss the influence of scale; and explain the
structure of the risk assessment protocol. The Guidelines also provide detailed definitions of the terms
used in the IUCN Red List of Ecosystems Categories and Criteria.

The Guidelines aim to support the practical implementation of the IUCN Red List of Ecosystems
Categories and Criteria from subnational to global areas of assessment. The Guidelines therefore
outline the necessary steps to: define the assessment area; define the unit under assessment; apply
the criteria; and prepare the assessment documentation for peer review and publication. All the steps
are illustrated with examples spanning a wide range of ecosystem types, geographical localities and
levels of data availability.

The future of the IUCN Red List of Ecosystems

The IUCN Red List of Ecosystems programme will assess the global status of the world’s terrestrial,
marine, freshwater and subterranean ecosystems. In addition, the programme aims to support the
development of national and regional Red Lists to inform conservation planning and sustainable
development. For more information on the IUCN Red List of Ecosystems please consult the IUCN
Red List of Ecosystems website (www.iucnrle.org).

v | IUCN Red List of Ecosystems

Acknowledgements

IUCN gratefully acknowledges the dedication and efforts of the hundreds of scientists and
practitioners who have contributed to the scientific development and practical testing of the
IUCN Red List of Ecosystems Categories and Criteria since 2008. These experts have
cumulatively participated in nearly 50 meetings and workshops in more than 20 countries
around the world, as well as submitted comments and suggestions remotely. To all we are
most indebted. Key contributors to this process are listed in Appendix 1.

The process to develop Red List criteria for ecosystems was launched with Resolution 4.020
at the Fourth IUCN World Conservation Congress in 2008, and consolidated with Resolution
5.055 adopted by the Fifth World Conservation Congress in 2012. This process culminated
in the adoption of the IUCN Red List of Ecosystems Categories and Criteria by the IUCN
Council in May 2014. Further information on the development of the IUCN Red List of
Ecosystems is available in Section 1.2.

The work on the IUCN Red List of Ecosystems has been made possible through generous
support from: Agence Française de Développement; Australian Research Council; Conoco-
Philips; Embassy of the Netherlands in Brazil; European Commission; Gordon and Betty
Moore Foundation; IUCN Netherlands National Committee; IUCN Secretariat; IUCN
Commission on Ecosystem Management; MAVA Foundation; New South Wales Office of
Environment and Heritage; and South Australian Department of Environment, Water and
Natural Resources.

vi | IUCN Red List of Ecosystems

IUCN Red List of Ecosystems

Website www.iucnrle.org

Steering Committee
David A. Keith, Chair, Committee for Scientific Standards
Emily Nicholson
Angela Andrade
Edmund G. Barrow
Thomas M. Brooks
Rebecca M. Miller
Irene Zager

Staff
Cambridge, UK Rebecca M. Miller, Red List of Ecosystems Senior Programme Coordinator
Caracas Irene Zager, Scientific Officer
Washington D.C. Kaia Boe, Social Science and Ecosystems Programme Officer
Nairobi Edmund G. Barrow, Director, Global Ecosystem Management Programme

IUCN Commission on Ecosystem Management (2017-2020)

Website www.iucn.org/about/union/commissions/cem/

Steering Committee
Angela Andrade, Chair
Madhav Karki, Deputy Chair
Mike Jones, Chair’s advisor
Bernal Herrera
Birguy Lamizana
Kelvin Passfield
Liette Vasseur

Red List of Ecosystems Thematic Group

David A. Keith, Lead
Emily Nicholson, Co-lead

About IUCN
IUCN, International Union for Conservation of Nature, helps the world find pragmatic solutions to our
most pressing environment and development challenges. IUCN’s work focuses on valuing and
conserving nature, ensuring effective and equitable governance of its use, and deploying nature-
based solutions to global challenges in climate, food and development. IUCN supports scientific
research, manages field projects all over the world, and brings governments, NGOs, the UN and
companies together to develop policy, laws and best practice. IUCN is the world’s oldest and largest
global environmental organisation, with almost 1,300 government and NGO Members and more than
16,000 volunteer experts in 185 countries. IUCN’s work is supported by almost 1,000 staff in 45
offices and hundreds of partners in public, NGO and private sectors around the world. www.iucn.org

vii | IUCN Red List of Ecosystems

Acronyms

Acronym Definition
AOO Area of occupancy
ARD Absolute rate of decline
CEM IUCN Commission on Ecosystem Management
CO Collapsed
CR Critically Endangered
DD Data Deficient
EN Endangered
EOO Extent of occurrence
GEMP IUCN Global Ecosystem Management Programme
GSP IUCN Global Species Programme
IUCN International Union for Conservation of Nature
LC Least Concern
NE Not Evaluated
NT Near Threatened
PRD Proportional rate of decline
RLE IUCN Red List of Ecosystems
RLTS The IUCN Red List of Threatened Species™
SSC IUCN Species Survival Commission
VU Vulnerable

viii | IUCN Red List of Ecosystems

Glossary

Term Definition
Area of assessment Defines the implementation bounds of the assessment.
Area of occupancy Area of occupancy (AOO) is a standardised measure of the area
that is occupied by an ecosystem type.
Characteristic native biota Biological features that define the identity of an ecosystem type
and distinguish it from other ecosystem types and/or drive
ecosystem dynamics and function, e.g. ecological processes,
ecosystem engineers, trophic or structural dominants,
functionally unique elements, species interactions.
Continuing decline A gradual or episodic decline in distribution or ecological process
that is likely to continue into the future, and is non-trivial in
magnitude and its effect on the sustainability of characteristic
native biota.
Ecosystem collapse Collapse is a transformation of identity, a loss of defining
features, and a replacement by a different ecosystem type.
Extent of occurrence Extent of occurrence (EOO) is a standardised measure of the
area within which all occurrences of an ecosystem type exist.
Ecosystem type The unit of assessment.
Geographic distribution The geographic distribution of an ecosystem type represents all
spatial occurrences of an ecosystem type.
Grain size The size of the spatial unit (e.g. grid cell, polygon segment) used
to measure a distribution.
Location A geographically or ecologically distinct area in which a single
(Threat-defined location) threatening event can rapidly affect all occurrences of an
ecosystem type.
Relative severity The estimated magnitude of past or future environmental
degradation or disruption to biotic processes, expressed as a
percentage relative to a change large enough to cause
ecosystem collapse.
Spatial extent The total area of the geographic distribution of an ecosystem
type estimated with a specified metric.
Temporal resolution The units of time over which trends are measured.
Thematic scale A measure of the similarity of features within and among
ecosystem types. May be represented by the levels of a
hierarchical classification.
Time frame The total period over which ecosystem change is assessed.

ix | IUCN Red List of Ecosystems

1. Introduction

The IUCN Red List of Ecosystems was developed to promote a consistent global framework
for monitoring the status of ecosystems (Keith et al., 2015). It is part of the growing toolbox
for assessing risks to biodiversity and aims to support conservation, resource use and
management decisions by identifying ecosystems most at risk of biodiversity loss. By
targeting a level of biological organisation above species, the IUCN Red List of Ecosystems
complements The IUCN Red List of Threatened Species™ (IUCN, 2015); together providing
simultaneous assessment of broad- and fine-scale biodiversity. A combined approach is
more likely to achieve the aim of comprehensive, effective and representative conservation
outcomes and will improve the ability to monitor the status of biodiversity on Earth.

The basis of the IUCN Red List of Ecosystems is the IUCN Red List of Ecosystems
Categories and Criteria (Appendix 2), a set of five criteria and associated thresholds that
provide a repeatable, globally consistent method for classifying the risk of ecosystem
collapse (Rodríguez et al., 2015; Keith et al., 2013). Ensuring accurate and comparable
assessments for all ecosystem types included on the IUCN Red List of Ecosystems is a key
challenge for the IUCN Red List of Ecosystems programme. These Guidelines provide the
information required to meet this challenge.

The Guidelines assist users to correctly implement the IUCN Red List of Ecosystems
Categories and Criteria by accompanying the assessor through the IUCN Red List of
Ecosystems assessment process, from understanding the scientific foundations through to
finalising assessments for publication. They provide information on the development of the
protocol (Section 1) and a detailed overview of the scientific foundations that support the
development of the categories and criteria (Sections 2 and 3). The Guidelines outline steps
required to define the area and units of assessment, and the key ecosystem processes that
will permit accurate application of the five criteria (Section 4). In addition, the Guidelines
contain comprehensive sections on each of the five criteria, including information on relevant
theory, thresholds and applications of each criterion (Section 5). Finally, the process of
preparing an assessment for peer review and publication are described (Section 6).

1.1 Objectives of the IUCN Red List of Ecosystems

The primary goal of the IUCN Red List of Ecosystems (RLE) is to support conservation in
resource use and management decisions by identifying ecosystems most at risk of
biodiversity loss (Keith et al., 2013; Keith et al., 2015). By assessing relative risks of
biodiversity loss at the ecosystem level, the RLE accounts for broad scale ecological
processes and important dependencies and interactions among species (Keith et al., 2015).
The IUCN Red List of Ecosystems also shines a light on common species, which define the
identity of many ecosystems, are involved in key interactions with large numbers of co-
occurring species, and can have major influences on ecosystem form and function (Gaston
& Fuller, 2008). To achieve the primary goal of the RLE, listing categories and criteria were
designed to be:

1 | IUCN Red List of Ecosystems

 1. A standard method for assessing and comparing risks of ecosystem collapse.
2. Easily understood by policy makers and the public.
3. Transparent, objective and scientifically rigorous.
4. Applicable to terrestrial, marine, freshwater and subterranean systems.
5. Applicable to risk assessments of local to global areas.
6. Flexible to use data of varying quality and coverage.
7. Consistent with and complementary to The IUCN Red List of Threatened
Species.

Although the primary goal of the RLE is focused on biodiversity conservation, the data
associated with the RLE may inform a wide range of other activities, including the
sustainable management of ecosystem services. Such applications will usually require
additional tools to achieve effective planning outcomes (Keith et al., 2015).

1.2 Development of the IUCN Red List of Ecosystems

Although the desire to create a global Red List of Ecosystems (RLE) is not new within IUCN
(Rodríguez et al., 2012a), the adoption of Resolution 4.020 on Quantitative Thresholds for
Categories and Criteria of Threatened Ecosystems (Fourth World Conservation Congress,
Barcelona, 2008) actively promoted the development of formal categories and criteria. The
resolution requested IUCN to “initiate a consultation process for the development,
implementation and monitoring of a global standard for the assessment of ecosystem status,
applicable at local, regional and global levels.” Over the following four years, and with
significant contributions from the scientific, government and conservation sectors, the IUCN
Red List of Ecosystems Thematic Group of the Commission on Ecosystem Management
(CEM) drafted an initial set of criteria (Version 1.0; Rodríguez et al., 2011). In subsequent
years, the criteria were disseminated and tested globally across a suite of ecosystem types
by a range of external partners and in collaboration with the IUCN Global Ecosystem
Management Programme (GEMP).

One major output of the global consultation led by the CEM was a substantial advance in the
scientific knowledge underpinning the RLE. The process resulted in a thorough review of the
relevant literature on ecosystem structure and functioning, documentation of the theoretical
basis for the RLE criteria, development of a model for ecosystem risk assessment, and
application of this new model to 20 ecosystems worldwide (Keith et al., 2013). This revised
set of IUCN Red List of Ecosystems Categories and Criteria (Keith et al., 2013) has been
refined following further application of the criteria to case studies, now spanning many
ecosystem types across all continents (Keith et al., 2015; Keith, 2015).

The Fifth World Conservation Congress (Jeju, 2012) adopted Resolution 5.055 on the
Consolidation of the IUCN Red List of Ecosystems, which acknowledged the progress of the
RLE development and requested the IUCN Council to “take the necessary steps for formal
approval of the categories and criteria as an official IUCN data analysis protocol for use by
the Members and any other stakeholder interested in ecosystem risk assessment”. Council
examined the IUCN Red List of Ecosystems Categories and Criteria documentation and on
21 May 2014 adopted them as the official global standard for assessing the risk to
ecosystems.

2 | IUCN Red List of Ecosystems

1.3 Governance of the IUCN Red List of Ecosystems

The Red List of Ecosystems is jointly coordinated by two IUCN bodies, the Commission on
Ecosystem Management (CEM) and the Global Ecosystem Management Programme
(GEMP). It is governed by two interacting committees with specific functions: (i) the Steering
Committee, and (ii) a Committee for Scientific Standards. It is supported by the Red List of
Ecosystems Thematic Group of the CEM, which is a group of volunteer experts that
undertake diverse duties in support of the objectives of the RLE. The RLE Programme
Unit—based in Cambridge, United Kingdom; Gland, Switzerland; and Nairobi, Kenya—
administers the RLE and ensures global coordination of the experts involved in research,
implementation and peer reviewing activities.

1.3.1 The Steering Committee

The RLE Steering Committee oversees the implementation of the IUCN Red List of
Ecosystems Categories and Criteria at global and sub-global levels. The Steering
Committee is composed of the Lead (and if applicable, the Co-lead) of the Red List of
Ecosystems Thematic Group of the CEM (appointed by the Chair of the CEM), the Chair of
the RLE Committee for Scientific Standards, the Chair of the CEM, the Director of the
GEMP, the Head of the IUCN Science and Knowledge Unit, as well as additional members
representing CEM, the IUCN Secretariat, and/or key RLE Partners appointed by the Chair of
the CEM because of their specific technical or organisational expertise. The GEMP Director
and the Head of the IUCN Science and Knowledge Unit represent the IUCN Secretariat.

The Steering Committee has the following functions:

1. Develop and manage the strategy and work plan for the implementation of the RLE
worldwide, to achieve the goal of assessing all ecosystems at a global level by 2025.
2. Establish a mechanism for periodically updating global assessments.
3. Identify and approach potential sources of financial support for assessments and their
dissemination.
4. Supervise a team of professional staff within the RLE Programme Unit, and build a
network of volunteers to implement the RLE work plan both within the CEM and the
IUCN Secretariat.
5. Actively engage the CEM in developing and peer reviewing assessments at the global
and sub-global levels.
6. Develop training materials and guidelines in the three official IUCN languages to support
assessments.
7. Recommend appointments to the RLE Committee for Scientific Standards.
8. Ensure that progress of the RLE is reported back to the IUCN Council and Secretariat
senior management.
9. Ensure that progress and outcomes of the RLE are well communicated in the scientific
literature and media.
10. Ensure the execution of the RLE work plan and maintain cooperation among
collaborating organisations.
11. Actively engage with others involved in the development, testing, and applications of
Knowledge Products mobilised by IUCN.

3 | IUCN Red List of Ecosystems

1.3.2 The Committee for Scientific Standards

The RLE Committee for Scientific Standards (CSS) is the principal scientific body that
provides expertise in the development, application and review of all issues related to the
RLE. The Committee consists of scientific experts with balanced expertise spanning a range
of skills, including risk assessment, ecological modelling, remote sensing, ecosystem
classification and mapping, decision theory, and ecology of terrestrial, freshwater, marine
and subterranean ecosystems. The combined expertise of the members of the Committee
for Scientific Standards covers the full diversity of ecosystem types and geographical
regions.

Members of the Committee for Scientific Standards, including the Chair and Deputy Chair,
are proposed by the RLE Steering Committee. The Chair of the CEM is ultimately
responsible for appointing members to a maximum four-year term, which expires at the
following session of the IUCN World Conservation Congress. One seat of the Committee for
Scientific Standards is reserved for a representative of The IUCN Red List of Threatened
Species designated by the Species Survival Commission (SSC) and the Global Species
Programme (GSP).

The Committee for Scientific Standards promotes the application of high scientific standards
to the implementation of the IUCN Red List of Ecosystems Categories and Criteria, and
ensures that the intent of the categories and criteria is not compromised. The specific
functions of the Committee for Scientific Standards are:

1. Develop and maintain technical guidelines in the three IUCN official languages to
support the application of the IUCN Red List of Ecosystems Categories and Criteria,
including details on implementation standards and data quality.
2. Provide scientific advice on the categories and criteria to the RLE Steering Committee
and the Programme Unit.
3. Provide scientific advice and support to the Programme Unit on the development of
databases, training materials and other resources.
4. Provide scientific advice on the design and implementation of systematic ecosystem risk
assessment projects that could contribute to the global RLE.
5. Manage a peer review process of all classifications and maps of ecosystem types
proposed for use in the global RLE.
6. Manage a peer review process for all assessments proposed for inclusion in the global
RLE and, subject to the outcomes of the review process, submit recommendations to the
Steering Committee on the inclusion or rejection of these assessments.
7. Critically review all applications of criterion E.
8. Provide scientific support and training for sub-global assessments of ecosystem types
via the RLE Programme Unit and other RLE partners.
9. Promote and undertake research to improve ecosystem risk assessment methodologies
underpinning the IUCN Red List of Ecosystems Categories and Criteria.
10. All formal decisions and recommendations of the RLE Committee for Scientific
Standards are submitted to the Steering Committee for review and formal adoption.

4 | IUCN Red List of Ecosystems

1.4 Structure of the Guidelines

The Guidelines for the application of IUCN Red List of Ecosystems Categories and Criteria
provide the information necessary to conduct a robust and repeatable ecosystem risk
assessment suitable for inclusion on the RLE. Section 1 (Introduction) offers an overview of
the motivation and history of the RLE, describing its general objectives and governing
structure. Section 2 (Categories of the IUCN Red List of Ecosystems) presents the
categories. Section 3 (Scientific Foundations) summarises the science underlying the
categories and criteria, and presents the RLE risk assessment model. Section 4
(Assessment Process) guides assessors through a full assessment suitable for submission.
The Criteria and Thresholds section (Section 5) outline the scientific theory underpinning
each criterion, the estimation of variables for assessment, and the values of the thresholds
for each category. Section 6 (Peer Review and Publication) describes the standards for
evaluating the quality of a risk assessment. Throughout, a series of worked examples and
cases studies are provided to assist assessors with the implementation of the categories and
criteria.

A summary sheet of the current version of the IUCN Red List of Ecosystems Categories and
Criteria is included as Appendix 2. More information on the IUCN Red List of Ecosystems,
links to relevant documents, and summaries of case studies are available in multiple
languages on the IUCN Red List of Ecosystems website (www.iucnrle.org).

5 | IUCN Red List of Ecosystems

2. Categories of the IUCN Red List of
Ecosystems

The IUCN Red List of Ecosystems includes eight categories: Collapsed (CO), Critically
Endangered (CR), Endangered (EN), Vulnerable (VU), Near Threatened (NT), Least
Concern (LC), Data Deficient (DD), and Not Evaluated (NE; Fig 1). The first six categories
(CO, CR, EN, VU, NT and LC) are ordered in decreasing risk of collapse. The categories
Data Deficient and Not Evaluated do not indicate a level of risk.

The categories Critically Endangered, Endangered and Vulnerable indicate threatened

ecosystems and are defined by quantitative and qualitative criteria described in Section 5
and Appendix 2. These categories are nested, so that an ecosystem type meeting a criterion
for Critically Endangered will also meet the criteria for Endangered and Vulnerable. The
three threatened ecosystem categories are complemented by several qualitative categories
that accommodate: (i) ecosystem types that almost meet the quantitative criteria for
Vulnerable (Near Threatened); (ii) ecosystems that unambiguously meet none of the
quantitative criteria (Least Concern); (iii) ecosystems for which too few data exist to apply
any criterion (Data Deficient); (iv) ecosystems that have not yet been assessed (Not
Evaluated). Following the precautionary principle (Precautionary Principle Project, 2005), the
overall status of an ecosystem type is the highest risk category obtained through any
criterion.

Figure 1. Structure of the IUCN Red List of Ecosystems categories.

6 | IUCN Red List of Ecosystems

Collapsed (CO)
An ecosystem is Collapsed when it is virtually certain (Table 3) that its defining biotic or
abiotic features are lost from all occurrences, and the characteristic native biota are no
longer sustained. Collapse may occur when most of the diagnostic components of the
characteristic native biota are lost from the system, or when functional components (biota
that perform key roles in ecosystem organisation) are greatly reduced in abundance and
lose the ability to recruit.

Critically Endangered (CR)

An ecosystem is Critically Endangered when the best available evidence indicates that it
meets any of the criteria A to E for Critically Endangered. It is therefore considered to be at
an extremely high risk of collapse.

Endangered (EN)
An ecosystem is Endangered when the best available evidence indicates that it meets any of
the criteria A to E for Endangered. It is therefore considered to be at a very high risk of
collapse.

Vulnerable (VU)
An ecosystem is Vulnerable when the best available evidence indicates that it meets any of
the criteria A to E for Vulnerable. It is therefore considered to be at a high risk of collapse.

Near Threatened (NT)

An ecosystem is Near Threatened when it has been evaluated against the criteria but does
not qualify for Critically Endangered, Endangered or Vulnerable now, but is close to
qualifying for or is likely to qualify for a threatened category in the near future.

Least Concern (LC)

An ecosystem is Least Concern when it has been evaluated against the criteria and does not
qualify for Critically Endangered, Endangered, Vulnerable or Near Threatened. Widely
distributed and relatively undegraded ecosystems are included in this category.

Data Deficient (DD)

An ecosystem is Data Deficient when there is inadequate information to make a direct, or
indirect, assessment of its risk of collapse based on decline in distribution, disruption of
ecological function or degradation of the physical environment. Data Deficient is not a
category of threat, and does not imply any level of collapse risk. Listing of ecosystems in this
category indicates that their situation has been reviewed, but that more information is
required to determine their risk status.

Not Evaluated (NE)

An ecosystem is Not Evaluated when it is has not yet been evaluated against the criteria.

7 | IUCN Red List of Ecosystems

3. Scientific foundations

3.1 Ecosystem types: the units of assessment

The IUCN Red List of Ecosystems (RLE) protocol is a robust and generic risk assessment
framework that can be applied to internally consistent classifications of ecosystem types. It
has flexibility to assess risks to ecosystems that vary greatly in biological and environmental
characteristics, scales of organisation, and amounts of available data. The clear definition
and description of ecosystem types is therefore an essential first step to a RLE assessment.

Ecosystems are complexes of organisms and their associated physical environment within a
specified area (Tansley, 1935). They have four essential elements: a biotic complex, an
abiotic environment, the interactions within and between them, and a physical space in
which these operate (Pickett and Cadenasso, 1995). Guidance on how to apply these
concepts to define and describe suitable units for RLE assessment is given in section 4.2
Describing the unit of assessment.

3.1.1 Ecosystem typologies

The IUCN Red List of Ecosystems Categories and Criteria may be applied systematically to
a set of ecosystem types within a specified area of assessment (global or sub-global) or to
single ecosystem types. Standalone assessments of single ecosystem types can be useful
diagnostic tools for ecosystem management (Keith et al., 2015). So long as the unit of
assessment is clearly defined and delineated, standalone assessments are less reliant on a
classification (typology) of ecosystem types than systematic assessments of multiple
ecosystems. These systematic assessments require a typology to ensure consistent and
comparable ecosystem risk assessments across the area of assessment. The classification
may simply delineate units at a particular thematic scale, or may describe their relationships
using hierarchies or nested arrangements that span a range of thematic scales (Rodríguez
et al., 2011).

A number of jurisdictions have developed suitable typologies to support RLE assessments of

national jurisdictions (Kontula & Raunio, 2009; Lindgaard & Henriksen, 2011; Driver et al.,
2012). At the continental level, a RLE assessment of several hundred terrestrial ecosystem
types for the Americas is based on an international classification framework for terrestrial
vegetation (Faber-Langendoen et al., 2014; Rodríguez et al., 2012b). The thematic scale of
ecosystem types in this assessment corresponds to the group and macrogroup levels in the
International Vegetation Classification system (Faber-Langendoen et al., 2014), and is
appropriate for global RLE assessments.

Sub-global assessments may be based on established national or regional ecosystem

classifications, providing the units of assessment conform to the definition of ecosystem
types (see 3.1 Ecosystem types: the units of assessment, above). These units should be
justified as suitable proxies for ecological assemblages and should be cross-referenced to
national, regional or global classification systems.

8 | IUCN Red List of Ecosystems

Development of a global ecosystem typology is currently underway (led by the RLE
Committee for Scientific Standards and the CEM RLE Thematic Group). This work is guided
by recent research on classifications of terrestrial vegetation (Faber-Langendoen et al.,
2014) and marine environments (Gregr et al., 2012), and seeks to promote transparent and
repeatable crosswalks among sub-global typologies meeting certain specifications (Section
4.1). In the interim, the IUCN Habitats Classification Scheme (www.iucnredlist.org/technical-
documents/classification-schemes/habitats-classification-scheme-ver3) provides a useful
comparative framework for assessments of contrasting ecosystem types at a range of
thematic scales.

Globally recognisable ecosystem types should not be confused with biogeographic or

biophysical ecoregions (Spalding et al., 2007), or biomes (Allen & Hoekstra, 1990).
Ecoregions and biomes are areas that share common macro-environmental or
biogeographical features and contain complexes of contrasting, but co-occurring ecosystem
types (Spalding et al., 2007). The potential heterogeneity of ecoregions and biomes makes
them unsuitable for most RLE applications (Rodríguez et al., 2015; Keith et al., 2015; Keith
et al., 2013). Other terms applied in conservation assessments –such as ecological
communities, habitats, biotopes, and (largely in the terrestrial context) vegetation types – are
regarded as operational synonyms of ecosystem type (Nicholson et al., 2009) providing they
are adequately defined in accordance with the procedures described in the assessment
process (Section 4.2).

3.1.2 The influence of scale

The RLE risk assessment protocol was designed to be flexible for application at multiple
spatial scales and with a range of data types (Rodríguez et al., 2015; Keith et al., 2015; Keith
et al., 2013). However, there are practical limits to the spatial, temporal and thematic scales
of units that can be assessed, and within these limits the assessment outcomes are
sensitive to scale. Assessments of units that are too broadly or narrowly defined, or failure to
implement methods or standardisation procedures (Section 5) could lead to scale
mismatches, incomparable assessments across scales, or invalid assessment outcomes
(Keith et al., 2013). A range of measures in the RLE protocol address the influence of scale:

1. Research is underway to support the interpretation of the RLE criteria for

assessments of different geographic areas. The categories and criteria were primarily
designed for assessments at the global level, but are applicable to sub-global
assessments (Section 3.1.1). Many of these sub-global assessments will work within
ecologically arbitrary boundaries (e.g. national borders), and therefore will consider
only parts of the global distribution of some ecosystem types. Methods for
interpreting and scaling threat categories or their thresholds to account for these
scenarios are currently under investigation.

2. A growing number of national and subnational assessments provide guidance on

appropriate thematic scales (classification level or strength; Hermoso et al., 2013) for
ecosystem risk assessments (Table 1). The ecosystem typologies provide examples
of ecosystem classifications designed to support different regulatory frameworks and

9 | IUCN Red List of Ecosystems

 conservation planning applications among jurisdictions. The development of a global
typology will provide further guidance on the thematic scale of assessments for the
global RLE.

3. Standard scales for assessing geographical distribution: the grain size (e.g. pixel
resolution) at which an ecosystem distribution is mapped can greatly affect the
estimate of distribution size. To maintain consistency with the fixed thresholds for
assessing distribution size (criterion B), distributions are measured at a standard
grain size (10x10 km grid) for estimating AOO and a standard geometric method
(minimum convex polygon) for estimating EOO. This generalising process is
sufficiently broad to accommodate processes relevant to persistence in a wide range
of ecosystem types (Section 5.2). A range of tools are made available to assist with
upscaling and downscaling distribution data, and completing assessments under
criteria A and B (Section 5).

4. Standard time frames for assessment: temporally, ecosystems may develop, persist
and change over time frames that vary from hours to millennia. They appear stable at
some temporal scales, while undergoing trends or fluctuations at others (Wiens,
1989; Carpenter & Turner, 2001). The categories assess ecosystem change over
standard time frames that represent trends over present, future and historical time
scales. Present and future time frames are set at 50 years to balance the need to
diagnose trends with reasonable certainty (requiring long time frames) with the need
for timely responses to adverse trends. Historical time frames are included to
accommodate the effects of ecological lags in assessments (Section 5).

10 | IUCN Red List of Ecosystems

Table 1. Examples of ecosystem typologies and similar classifications supporting national ecosystem
risk assessments for various conservation planning and regulatory applications (adapted from Keith et
al., 2015).

Jurisdiction Application Assessment unit Reference

European Habitats Directive Habitat type. 'Plant and animal Council of the
Union 92/43/EEC (European communities as the European
Commission) characterising elements of the Commission (1992)
biotic environment, together
with abiotic factors operating
together at a particular scale.'
Germany Red List of biotopes Biotope. 'Habitat of a Riecken et al. (2009);
(Federal Environment community of fauna and flora Riecken et al. (2006)
Agency) living in the wild.'
Finland Red List of habitat types Habitat type. 'Spatially Kontula & Raunio
(Finnish Environment definable land or aquatic areas (2009)
Institute) with characteristic
environmental conditions and
biota which are similar between
these areas but differ from
areas of other habitat types.'
Norway Red List of ecosystems Habitat type. 'A homogeneous Lindgaard &
and habitat types environment, including all plant Henriksen (2011)
(Norwegian Biodiversity and animal life and
Information Centre) environmental factors that
operate there.'
Venezuela National Red List of Major vegetation types for Rodríguez et al.
ecosystems (Provita) national assessment; satellite- (2010)
derived land types for
subnational assessments.
Canada State threatened species Ecosystem. 'A dynamic Government of
and ecosystems complex of plant, animal and Manitoba (2014)
legislation (Manitoba microorganism communities
Conservation and Water and their nonliving environment
Stewardship Department) interacting as a functional unit.'
Australia Lists of threatened Ecological community. 'An Commonwealth of
ecological communities at assemblage of native species Australia (2000);
national and state levels that inhabits a particular area in Keith (2009);
(Federal Department of nature.' Nicholson et al.
Environment, state (2015)
environment agencies)
South National biodiversity Ecosystem. 'A dynamic Republic of South
Africa legislation (South African complex of animal, plant and Africa (2004); Driver
National Biodiversity micro-organism et al. (2012)
Institute) communities and their nonliving
environment interacting as a
functional unit.'

11 | IUCN Red List of Ecosystems

3.2 Ecosystem collapse

To achieve a robust application of the IUCN Red List of Ecosystems Categories and Criteria,
assessors must synthesise diverse causes, mechanisms and pathways of ecosystem
decline within the generic risk assessment framework. To estimate risk—the probability of an
adverse outcome over a specified time frame—it is necessary to define the endpoint of
ecosystem decline, the point at which an ecosystem is considered collapsed. The definition
of the endpoint to ecosystem decline must be sufficiently discrete to permit an assessment
of risk, but sufficiently general to encompass the broad range of contexts in which risk
assessments are needed. The RLE protocol has two elements to deal with this trade-off: (i) a
definition of ecosystem collapse as the endpoint to ecosystem decline; (ii) a risk assessment
model that identifies the multiple pathways to ecosystem collapse and forms the basis for the
criteria.

Within the IUCN Red List of Ecosystems Categories and Criteria, “an ecosystem is
Collapsed when it is virtually certain (Table 3) that its defining biotic or abiotic features are
lost from all occurrences, and the characteristic native biota are no longer sustained.
Collapse may occur when most of the diagnostic components of the characteristic native
biota are lost from the system, or when functional components (biota that perform key roles
in ecosystem organisation) are greatly reduced in abundance and lose the ability to recruit.”

3.2.1 Defining ecosystem collapse

Unlike species, ecosystems do not disappear; rather they transform into novel ecosystems
with different characteristic biota and mechanisms of organisation (Hobbs et al., 2006; Keith
et al., 2015; Keith et al., 2013). Many characteristic features may disappear long before the
last characteristic species disappears from the last ecosystem occurrence (assemblage
extinction; Gaston & Fuller, 2008). The novel systems may retain some of the characteristic
biota of the collapsed systems that they replace, but the abundance of those species, their
interactions or ecological functions are altered. Acknowledging the contrasts with species
extinction, the concept of ecosystem collapse is defined as the transition beyond a bounded
threshold in one or more variables that define the identity of the ecosystem. Collapse is a
transformation of identity, a loss of defining features, and/or replacement by a different
ecosystem. An ecosystem is collapsed when all occurrences lose defining biotic or abiotic
features, no longer sustain the characteristic native biota, and have moved outside their
natural range of spatial and temporal variability in composition, structure and/or function.
This can be illustrated by the familiar ‘marble’ model of state and transition theory (Fig. 2)
and by key examples such as the Aral Sea (Box 1). Ecosystem collapse may in theory be
reversible—given a long time frame, or via the reintroduction of characteristic biota and/or
the restoration of ecosystem function—but in many systems recovery will not be possible.

Transitions to collapse may be gradual, sudden, linear, non-linear, deterministic or highly

stochastic. These include regime shifts (Scheffer et al., 2001), but also other types of
transitions that may not involve reinforcing feedbacks. The dominant dynamic in an
ecosystem will depend on abiotic or external influences (e.g. weather patterns or human
disturbance), internal biotic processes (e.g. competition, predation, or epidemics), historical
legacies (e.g. climatic history, extinction debts or exploitation), and spatial context (e.g.

12 | IUCN Red List of Ecosystems

whereabouts, size and dispersion of distribution). An ecosystem may thus be driven to
collapse by different threatening processes and through multiple pathways. Trophic
cascades (Estes et al., 2011), loss of foundation species (Diamond, 2007), environmental
degradation (UNEP, 2001), and climatic forcing (Grebmeier et al., 2006) are common
pathways to ecosystem collapse. Symptoms of collapse may differ depending on the
characteristics of the ecosystem, the nature of threatening processes, and the pathways of
decline that these generate. The RLE protocol has flexibility to allow thresholds of collapse
to be expressed in appropriate terms for very different kinds of ecosystems.

The definition of ecosystem collapse may be clearest for ecosystems that have already
collapsed and for which time series data exist for relevant variables. It will often be possible
to infer characteristics of collapse from occurrences within the ecosystem distribution where
defining features have been lost, even if the majority of the ecosystem remains extant. Major
changes in functionally similar ecosystems can also provide guidance for defining the
symptoms of collapse in systems of interest. This can provide a basis for defining the spatial
and functional symptoms of ecosystem collapse.

Figure 2. Generalised schematic illustrating the interpretation of ecosystem collapse in a state and
transition framework (Keith et al., 2015). States A-G are defined by two state variables represented
on the X and Y axes. The vertical axis (Z) represents potential for change. The two broken lines
represent alternative interpretations of ecosystem collapse. For the inner line, transitions between
states A, B and C (e.g. white arrow) represent natural variability without loss of key defining
features, while transitions across broken lines (e.g. grey arrow) to states D, E, F and G represent
collapse and replacement by novel ecosystems. Progression along different pathways of collapse
is assessed with variables X and Y, or other ecosystem-specific diagnostic variables that reflect
the loss of characteristic native biota and function. The outer broken line represents an alternative
interpretation of ecosystem collapse in which state E is included within natural variation of the
ecosystem type.

13 | IUCN Red List of Ecosystems

Box 1. Ecosystem collapse in the Aral Sea

The Aral Sea – the world’s fourth largest continental water body – is fed by two major rivers, the
Syr Dar’ya and Amu Dar’ya (Aladin & Plotnikov, 1993). Its characteristic native biota includes
freshwater fish (20 species), a unique invertebrate fauna (>150 species) and shoreline reed beds,
which provide habitat for waterbirds, including migratory species (Keith et al., 2013).
Hydrologically, the sea was approximately stable during 1911-1960, with inflows balancing net
evaporation (Micklin & Aladin, 2008). Intensification of water extraction to support expansion of
irrigated agriculture lead to shrinkage and salinisation of the sea. By 2005, only 28 aquatic species
(including fish and invertebrates) were recorded, reed beds had dried and disappeared, the sea
had contracted to a fraction of its former volume and surface area, and salinity had increased
tenfold (Micklin & Aladin, 2008).

Consistent with the definition of ecosystem collapse, these changes suggest the Aral Sea has
undergone a transformation of identity, lost many of its defining features (aquatic biota, reed beds,
waterbirds, hydrological balance and brackish hydrochemistry) and has been replaced by novel
ecosystems (saline lakes and desert plains). Under this interpretation, collapse occurred before
the volume and surface area of standing water declined to zero. Although the exact point of
ecosystem collapse is uncertain, time series data for several variables are suitable for defining a
functional reference state (prior to onset of change from 1960) and a bounded threshold of
collapse, assuming this occurred sometime between 1976 and 1989 when most biota disappeared
(Keith et al., 2013).

The choice of available variables for

assessing the status of the ecosystem will
depend on how closely they represent the
ecosystem's defining features, the quantity
and quality of the data, and the sensitivity of
alternative variables to ecological change. Of
those listed above, fish species richness and
abundance may be the most proximal biotic
variable to the features that define the
identity of the Aral Sea ecosystem. Sea
volume may be a reasonable abiotic proxy,
because volume is functionally linked with
salinity, which in turn mediates persistence of
the characteristic freshwater/brackish aquatic
fauna. Sea surface area is less directly
related to these features and processes, but can be readily estimated by remote sensing and may
be useful for assessment when data are unavailable for other variables.

Collapse of the Aral Sea ecosystem may or may not be reversible. While it may be possible to
restore the hydrological regime over a small part of the former sea (Micklin & Aladin, 2008), some
components of the characteristic biota are apparently extinct (e.g. the Aral salmon, Salmo trutta
aralensis), preventing reconstruction of the pre-collapse ecosystem. Image: © NASA

14 | IUCN Red List of Ecosystems

3.2.2 Uncertainties in the endpoints for risk assessment

Risk assessment relies on the definition of an adverse outcome, typically a discrete endpoint
or event that affects the asset under evaluation. The implementation of risk assessment
confronts uncertainties in two key areas: the definition of the asset itself, and the definition of
the endpoint. The boundary which delineates an ecosystem type may be uncertain due to
imperfect knowledge of natural variability within the ecosystem, continuous patterns of
variability with other ecosystems, and changes in ecosystem classification through time, as
well as uncertainties associated with mapping distributions (Keith et al., 2013). Defining
ecosystem collapse is also subject to uncertainty which can affect the estimation of spatial
and functional symptoms of collapse (Fig. 3). All applications the IUCN Red List of
Ecosystems Categories and Criteria should consider these sources of uncertainty and
discuss them in the assessment documentation. Examples of how uncertainties can be dealt
with through the assessment process are described below, acknowledging that uncertainties
in spatial and functional systems are often related.

Uncertainty in spatial symptoms

During decline, an ecosystem may transition to collapsed state(s) in some parts of its
distribution before others. In areas where these transitions have occurred, the ecosystem
may be described as ‘locally collapsed’. Spatially, an ecosystem is considered collapsed
when all extant occurrences of the ecosystem have collapsed (i.e. area of occupancy = 0
10x10 km grid cells and extent of occurrence = 0 km2). To quantify past declines in
distribution and declines in function, assessors must identify where the ecosystem type is
currently extant, and where it was previously extant (within the time frame of assessment)
and is now in a collapsed state. Similarly, to quantify future declines in distribution and
function, assessors must project the area in which the ecosystem will collapse during the
future time frame of the assessment. All of these estimations and projections involve
uncertainties. Epistemic uncertainty (i.e. uncertainty due to a lack of knowledge, as opposed
to inherent uncertainty due to variability in the system) exists due to a range of measurement
and classification errors:

1. Thematic uncertainties caused by decisions relating to the threshold at which an

ecosystem type is considered to have moved outside of its natural bounds of
variability, and must then be considered a different ecosystem type (Payet et al.,
2013).

2. Measurement error due to imperfect measurements or mapping techniques resulting

in area estimates that are not precisely repeatable and randomly fluctuate (Elith et
al., 2002; Olofsson et al., 2014; Fuller et al., 2003).

3. Systematic error due to mapping methods that consistently produce biased area
estimates (Congalton & Green, 2008).

4. Classification errors that result in misclassification of pixels in a distribution map,

generally termed omission or commission errors (Congalton & Green, 2008; Foody,
2011).

15 | IUCN Red List of Ecosystems

 5. Errors of scale where the grain size at which an ecosystem is mapped results in area
estimates that are dependent on the scale at which they are mapped (Hartley &
Kunin, 2003; Gaston & Fuller, 2009).

Uncertainty in functional symptoms

A collapsed ecosystem may be replaced by a novel ecosystem with strongly contrasting

features. When grasslands replace forests, the change in vegetation structure is readily
detected by a range of proximal and remotely sensed methods. In other cases, ecosystems
may lose defining features and collapse, but the novel system may resemble the antecedent
one, making symptoms of collapse more difficult to detect. Burns et al. (2015) describe an
example of a forest ecosystem characterised by biota associated with large old trees. When
densities of large old trees fall below a critical level, characteristic native biota is lost from the
system. This includes birds and mammals that nest or shelter in tree hollows, and
invertebrates that live under loose bark and in deep leaf litter beds. After such transitions,
the novel ecosystem still retains a forest structure, albeit one characterised by smaller trees
and lacking biota associated with large trees. Similarly, Barrett & Yates (2015) described
collapse of a species-rich shrubland as the elimination of groups of plant species eliminated
by a soil-borne disease. The novel ecosystem replacing the antecedent one was a
structurally similar, but compositionally and functionally different shrubland. These and other
examples illustrate uncertainties in delineating extant and collapsed states, which depend on
the features of the antecedent ecosystem, the pathway of collapse, and the features of the
novel ecosystem. Sources of uncertainty include:

1. Definition of reference ecosystem states, and the natural variability within those.

2. Definition of collapsed ecosystem states, which represent critical deviations from

natural variability. Transition points from original to novel ecosystems are inherently
uncertain but can be estimated within plausible bounds (Fig. 3). The first value
represents no doubt that the ecosystem has collapsed, whereas the second is a
plausible value based on observations or inferences.

3. Variation in collapsed states caused by different threatening processes. Catastrophic

threats may cause total functional and spatial collapse of the ecosystem. Other
threats, such as environmental degradation or the spread of invasive species may
cause different functional changes in characteristic biota. These different pathways of
collapse should be reflected in the documentation (as part of the definition of
collapse; see Section 4.2.7 and Section 5).

4. Uncertainty in the measurement of variables representing ecosystem function and

collapse. As with spatial variables, measurement error in functional variables may
affect the assessment of ecosystem collapse through random errors or systematic
bias.

16 | IUCN Red List of Ecosystems

Figure 3. Probability density functions for the population and ecosystem variables that
measure proximity to the thresholds that define species extinction (a, b), species quasi-
extinction (c-e), and ecosystem collapse (f-h). For species, the population threshold that
defines extinction is known with certainty (e.g. zero abundance, described by the vertical line
in (a) and (b)). In practice, Population Viability Analyses are calibrated on a quasi-extinction
threshold higher than the extinction threshold, to account for prediction and management
uncertainty. A lower bound on the value of extinction (zero abundance), and a putative upper
bound for the value of quasi-extinction can be depicted as a dashed box (c-e). For
ecosystems (f-h) the x-axis could represent key features or processes (e.g. spatial
distribution, number of species, water quality). The bounded definition of collapse is
analogous to the definition of quasi-extinction in species. The width of the dashed box
represents uncertainty in the collapse definition. The blue area represents the probability that
the ecosystem is definitely extant, whereas the red area represents the probability that the
ecosystem may be extant (adapted from Keith et al., 2013).

17 | IUCN Red List of Ecosystems

3.3 Risk assessment protocol

The RLE protocol comprises five rule-based criteria for assessing risks to ecosystems. Risks
to ecosystems may be caused by a variety of threatening processes that are expressed
through different symptoms of ecosystem collapse (Keith, 2015). The RLE protocol groups
symptoms of ecosystem collapse into four major types and identifies the corresponding
mechanisms that link the symptoms to the risk that an ecosystem will lose its defining
features (Fig. 4). Two of the four mechanisms produce distributional symptoms: (A) declines
in distribution, which reduce carrying capacity for dependent biota; and (B) restricted
distribution, which predisposes the system to spatially explicit threats. Two other
mechanisms produce functional symptoms: (C) degradation of the abiotic environment,
reducing habitat quality or abiotic niche diversity for component biota; and (D) disruption of
biotic processes and interactions, resulting, for example, in the loss of mutualisms, biotic
niche diversity, or exclusion of some component biota by others. Interactions between two or
more of these four contrasting mechanisms may produce additional symptoms of transition
towards ecosystem collapse. Multiple mechanisms and their interactions may be integrated
into a simulation model of ecosystem dynamics to produce quantitative estimates of the risk
of collapse (E). These five groups of symptoms form the basis of the RLE criteria. An
ecosystem type under assessment should be evaluated using all of the criteria for which
data are available. The overall risk status of the ecosystem type is assigned as the highest
category of risk obtained through any criterion.

Figure 4. Mechanisms of ecosystem collapse and symptoms of collapse risk (source: Keith et
al., 2013).

18 | IUCN Red List of Ecosystems

3.3.1 Time frames

The criteria assess declines over four specified time frames: the historical past, the recent
past, any 50-year period including the recent past, present and future, and the future (Fig. 5).
The ‘recent past’ time frame encompasses the past 50 years, which is sufficiently recent to
capture current trends but long enough to distinguish directional change from natural
variability. The RLE protocol assumes that declines over this time frame are indicative of
future risk irrespective of cause.

Assessment of future declines requires predictions of changes over the next 50 years or any
50-year period including the present and future (Fig. 5). Past declines may provide a basis
for such predictions, but other information may support predictions and inferences about
rates of future decline even when the ecosystem is currently stable. Such predictions require
a defensible assumption about the pattern of future change (e.g. accelerating, constant,
decelerating). Plausible alternative models of change should be explored where appropriate,
but a constant proportional rate of decline is often a reasonable default assumption (Section
5).

Assessments of historical declines are essential for ecosystems containing biota with long
generation lengths and slow population turnover (Mace et al., 2008). They are also essential
for foundation species with short generation lengths which may have suffered extensive
historical declines (e.g. oyster reefs: Kirby, 2004; Beck et al., 2011). Even where future rates
of decline abate, historical reductions in distribution or function may predispose an
ecosystem to additional threats and reduce its ability to absorb adverse changes (Folke et
al., 2004). Historical declines are assessed relative to ecosystem status at a notional
reference date of 1750, corresponding approximately to the earliest onset of industrial-scale
exploitation of ecosystems. In parts of the world where industrial-scale exploitation of
ecosystems commenced earlier or later than 1750, it is justifiable to assess historical
declines with a different baseline. Distribution models with environmental predictors may be
used to estimate historical declines based on the difference between the current state of an
ecosystem and its expected state in the absence of industrial-scale anthropogenic effects.
Such approaches are most useful in regions where landscape-scale change did not occur
before the industrial era.

19 | IUCN Red List of Ecosystems

 Figure 5. Time frames for assessment of change under criteria A, C, and D (adapted from
Keith et al., 2013).

3.3.2 Decline thresholds

The ordinal categories of risk (Section 3) are delimited by thresholds defined in the IUCN
Red List of Ecosystems Categories and Criteria (Appendix 2). The rationale for the criteria
and ordinal categories is grounded in theory (Keith et al., 2013). However, the threshold
values that delimit categories are based partly on theoretical considerations and partly on
utilitarian considerations (Keith et al., 2015). Theory provides a qualitative basis for ordered
thresholds for decline, but offers limited guidance for setting their absolute values. The
purpose of these decision thresholds is to rank ecosystems in informative ordinal categories
of risk, rather than estimate precise probabilities of collapse. Consequently, for criteria A, C,
and D, threshold values were set at relatively even intervals for current and future declines in
ecosystem distribution or function (Vulnerable: 30%, Endangered: 50%, Critically
Endangered: 80%). The range of thresholds between 0 and 100% seeks to achieve an
informative rather than highly skewed ranking of ecosystems among categories. The lowest
threshold for a threatened ecosystem type (30%) recognises that evidence of an appreciable
decline in ecosystem distribution or function is necessary to support listing in a threatened
category. These thresholds are consistent with thresholds for population reduction in The
IUCN Red List of Threatened Species (IUCN, 2001, 2012). Thresholds for historical declines
are higher (A3, C3, D3; 50%, 70%, 90%) because times frames for assessment are longer.

Declines within 5-10% of thresholds for the Vulnerable category may warrant listing as Near
Threatened, although there are no quantitative thresholds for this category (Section 3). For
example, an ecosystem type with an extent of occurrence of 50,000 to 55,000 km2 that
qualifies for at least one of the three subcriteria of criterion B could qualify for listing as Near
Threatened. An ecosystem type with a decline in an abiotic variable of 20% to 30% relative
severity and 100% extent could qualify as Near Threatened under subcriteria C1 or C2.

20 | IUCN Red List of Ecosystems

3.3.3 Standards of evidence and dealing with uncertainty

Achieving a robust and repeatable assessment for an ecosystem type requires extensive
data, often from disparate sources. The categories and criteria were specifically designed to
allow the inclusion of various data types from a range of sources, but it is the onus of the
assessor to critically evaluate whether data quantity and quality are sufficient to support
determinate outcome of an assessment. For guidance on this evaluation, assessors are
referred to the principles adopted by the Intergovernmental Panel on Climate Change for
consistent treatment of uncertainty (Mastrandrea et al., 2010). In summary, key principles
include:

1. Evaluating the type (Table 2), amount, quality, and consistency of evidence
(summary descriptors: “limited,” “medium,” or “robust”);

2. Evaluating the degree of agreement between different sources of evidence

(summary descriptors: “low,” “medium,” or “high”);

3. Providing a traceable account describing the evaluation of evidence and agreement;

4. Evaluating the likelihood (Table 3) of alternative categories as outcomes of an

assessment;

5. Communicating the uncertainty in the outcomes of an assessment by reporting the

most likely category and as well as categories that represent plausible upper and
lower bounds of the assessment outcome (Section 4.4.1).

The standard of evidence for the RLE must be sufficient to support inferences that:

1. Some categories (LC, NT, VU, EN or CR) are ‘very unlikely’ outcomes of assessment
(i.e. probability <10%, Table 3). If no category is a very unlikely outcome of
assessment, then the status should be assigned as Data Deficient (DD);

2. The plausible bounds of assessment outcomes include all categories necessary to

ensure that collectively they are ‘very likely’ to encompass the true status (i.e.
probability >90%, Table 3). If all categories (LC-CR) are within the plausible bounds,
then the status should be assigned as Data Deficient (DD);

3. The best overall status (i.e. categorisation of an ecosystem) is more likely than any
alternative categorisation and within the plausible bounds; and

4. All categorisations of overall status in the Collapsed category (CO) are ‘virtually
certain’ (i.e. >99% certain, Table 3). Where this is not the case and CO is the most
likely category, the best overall status should be assigned to CR, and CO reported as
the upper bound of the assessment outcome.

21 | IUCN Red List of Ecosystems

 Table 2. Descriptors for types of evidence (IUCN, 2001, 2012) will typically support inferences
during an assessment. These apply to quantitative variables (such as rates of change in
distribution) and binary inferences (such as whether or not there is a continuing decline in
distribution).

Descriptor Explanation

Observed Information that is directly based on well-documented records of all known

occurrences of the ecosystem (IUCN Standards and Petitions Subcommittee,
2016).
Estimated Information that is based on calculations that may include statistical assumptions
about sampling, or biological assumptions about the relationship between an
observed variable and the variable of interest (e.g. relationship between an index
of abundance and the number of mature individuals; IUCN Standards and
Petitions Subcommittee, 2016). These assumptions should be stated and
justified in the assessment documentation. Estimation may also involve
interpolation in time to calculate the variable of interest for a particular time step
(e.g. a 50-year reduction in distribution based on observations of distribution 40
and 60 years ago).
Inferred Information that is based on indirect evidence and on variables that are indirectly
related to the variable of interest, but in the same general type of units (IUCN
Standards and Petitions Subcommittee, 2016). Inferred values rely on more
assumptions than estimated values. For example, inferring disruption of biotic
interactions from catch statistics not only requires statistical assumptions (e.g.
random sampling) and biological assumptions (about the relationship of the
harvested section of the population to the total population), but also assumptions
about trends in effort, efficiency, and the spatial and temporal distribution of
harvest in relation to the population. Inference may also involve extrapolating an
observed or estimated quantity from known ecosystem occurrences to calculate
the same quantity for other occurrences. Whether there are enough data to
make such an inference will depend on how large the known occurrences are as
a proportion of the whole distribution, and the applicability of threats and trends
observed in the known occurrences to the rest of the ecosystem.
Projected Same as estimated, but the variable of interest is extrapolated in time towards
the future (IUCN Standards and Petitions Subcommittee, 2016). Projected
variables require a discussion of the method of extrapolation (e.g. justification of
the statistical assumptions or the ecosystem model used) as well as the
extrapolation of current or potential threats into the future, including their rates of
change.

22 | IUCN Red List of Ecosystems

 Table 3. Calibrated language for describing quantified uncertainty (source: Mastrandrea et al.,
2010). It can be used to express a probabilistic estimate of a quantity, a binary inference or an
assessment outcome (e.g. a magnitude of change in distribution, whether or not there has
been a change, whether the status of an ecosystem is within a given range). Likelihood may
be based on statistical or modelling analyses, elicitation of expert views, or other quantitative
analyses. The categories defined in this table can be considered to have “fuzzy” boundaries
(Kauffman & Gupta, 1991).

Term Likelihood of outcome

(probability)
Virtually certain 99–100%
Very likely 90–100%
Likely 66–100%
More likely than not 50–100%
About as likely as not 33–66%
Unlikely 0–33%
Very unlikely 0–10%
Exceptionally unlikely 0–1%

3.3.4 Making the most of quantitative data and expert knowledge

The Red List criteria require calculations based on quantitative estimates of variables such
as areas and rates of change in biotic and abiotic features of ecosystems. Quantitative
estimates of these variables are ideally based on systematic measurements acquired in a
sampling design that permits valid statistical inferences across the geographic range of the
ecosystem type under evaluation. In reality, relevant and useful evidence on ecosystem
status includes a range of incomplete, patchy and subjective observations.

Scientific judgements are required to decide which pieces of information meet the standard
of evidence required to support an inference about the status of an ecosystem. For example,
a particular forest ecosystem may never have been mapped at an appropriate resolution to
quantify the proportional change in its distribution over the past 50 years, as required to
assess criterion A1. Despite the lack of formal data, experts are unanimous in their opinion,
based on anecdotal observations, that at least 50% of the ecosystem distribution has been
converted to pasture in the past 50 years. The high degree of certainty about the rate of
decline should inform a Red List assessment – the status of the forest ecosystem is likely to
be at least Endangered and is very unlikely to be Least Concern. Qualitative expert
knowledge may also add value to quantitative measurements. For example, data from
repeat surveys of fish in a marine reef ecosystem may indicate a 32% decline in abundance
over the past 50 years, but experts are unanimous that surveys are limited to the most
exploited reefs and, based on anecdotal observations, that fish abundance has remained
“approximately stable” on many unexploited reefs. If fish abundance was assumed to decline
by 0-20% on these unexploited reefs (a worst-case interpretation of “approximately stable”),
the overall average decline across all reefs is estimated to decline by 15-25%. In this case,
Least Concern or Near Threatened may be more likely status than Vulnerable, despite the
estimate based on formal data.

23 | IUCN Red List of Ecosystems

Both examples above show how expert knowledge can improve inferences about Red List
status compared to assessments based exclusively on measurements. However, expert
opinion is notoriously unreliable, subject to various social biases, influenced by a range of
experiential and behavioural factors and expert performance is very difficult to predict
(Burgman 2015). Use of expert opinion to estimate quantities required for Red List
assessment must therefore be subject to standards and procedures that reduce the risks of
errors and bias. The recommended standards and their rationale are given in Table 4.

Table 4. Recommended standards and procedures for expert elicitation and handling
uncertainty (based on Burgman 2015).

Step Recommended approach Rationale

Selecting People who are: i) reasonably familiar with Expertise declines dramatically outside
experts the ecosystem type, the area in which it an individual’s specialisation or
occurs, and the processes that affect it; experience. Basic familiarity is relevant
and ii) frequently seek feedback and but expert performance appears
consider uncertainty in their advice. Seek independent of experience and standing.
diversity and avoid homogeneity in Experts who seek frequent feedback on
selecting expert groups. their judgements, subdue
overconfidence and consider uncertainty
perform well.
Number of A minimum of three. More is better. Estimated values averaged across
experts multiple experts outperform individual
estimates, including those by the most
experienced experts.
Information Available data and qualitative observations Provides a common base of information
provided to relevant to the quantity being estimated, on which to base an estimate. Raises
experts including sources, contextual information awareness of cognitive factors
including definitions of terms and details of associated with accurate expert
sampling design and methods. Inform estimates and reduces linguistic
experts of the elicitation process (steps 1- uncertainties.
3) and the qualities associated with high
performance* (see Burgman 2015).
Elicitation Each expert is asked to estimate a required Independent estimates for each expert
step 1 quantity (e.g. decline in distribution over avoids social elicitation biases
past 50 years) independently of (i.e. associated with dominant personalities,
without conferring with) others. Four values seniority, perceptions of peers, etc.
are required for each estimate in the
following order:
i) a plausible upper bound
ii) a plausible lower bound;
iii) a best estimate;
iv) the probability that the true value lies
between the upper and lower bound.
Elicitation Experts are provided with all estimates Exchange of ideas and factors relevant
step 2 without names of those who made them. In for consideration, additional data and
plenary, they are given an opportunity to observations, supports more informed
discuss the reasons considered in coming estimates.

24 | IUCN Red List of Ecosystems

Step Recommended approach Rationale

to an estimate
Elicitation Each expert is given the opportunity to Reduces social biases, while
step 3 revise their estimates from step 1 incorporating additional information.
independently of other experts, in the light
of discussion in step 2
Synthesis The best estimates are averaged across all Central tendency of multiple independent
experts. Upper and lower bounds are estimates is more likely to be close to
converted to 90% confidence interval, the true values than any other expert
assuming a probability distribution and estimate. Upper and lower bounds
transformation that are appropriate to the based on means exclude extreme
quantity estimated (Speirs-Bridge et al. outlying values.
(2010), and averaged across assessors.
Assessment The Red List status is calculated for the Uncertainty (represented by upper and
against Red best estimate, upper and lower bounds, lower bounds) is propagated
List criterion producing a bounded estimate of the threat transparently through the assessment,
category for that criterion allowing reporting of the best estimate of
threat category, as well as plausibly
optimistic and pessimistic categories,
given the available information.

25 | IUCN Red List of Ecosystems

4. Assessment process

Assessing an ecosystem type against the IUCN Red List of Ecosystems Categories and
Criteria is a sequential process. All components must be completed before submission of the
assessment (Fig. 6).

Figure 6. Process for assessing the risk of collapse of an ecosystem type.

26 | IUCN Red List of Ecosystems

4.1 Area of assessment

Red List of Ecosystems (RLE) assessments may be undertaken within different geographic
areas. Global assessments consider all occurrences of an ecosystem type throughout the
world. This is essential for the set of broadly defined ecosystem types that will form the
global RLE, and for informing international biodiversity targets and conservation strategies.
Sub-global assessments are possible: they are typically defined by political (continental,
national or state assessments) or ecoregional boundaries (ocean basins or catchments).
Many sub-global lists of ecosystems already exist, such as lists of threatened ecosystems
for Germany (Blab et al., 1995), Western Australia (DEC, 2007), Finland (Kontula & Raunio,
2009), Venezuela (Rodríguez et al., 2010), Austria (Essl & Egger, 2010), Norway (Lindgaard
& Henriksen, 2011), South Africa (Driver et al., 2012), New Zealand (Holdaway et al., 2012)
and El Salvador (Crespin & Simonetti, 2015), although only the latter used Version 2.0 of the
IUCN Red List of Ecosystems Categories and Criteria.

For assessments of sub-global areas, it will usually be appropriate to assess ecosystem

types of finer thematic resolution than those for global assessments, as sub-global
assessments will usually require finer detail to support land and water use decision-making.
For example, a national RLE may have a larger number of more finely divided assessment
ecosystem types for a given area, compared to a global-level RLE assessment.

The same ecosystem type may be assigned to different risk categories in sub-global and
global assessments. Differences in status depend on the distribution of threatening
processes across the range of the ecosystem type in relation to the boundaries of the sub-
global assessment. Although regional guidelines for applying the IUCN Red List of
Ecosystems Categories and Criteria have not yet been developed, some general rules
apply:
1. Comprehensive description of the assessment unit (ecosystem type) is still required.
The area of assessment (e.g. political boundaries) must be clearly defined and
supported with maps or other spatial data.

2. No modifications of the categories or criteria A, C, D or E are required when making

sub-global assessments of ecosystems. Therefore, all thresholds, time frames,
definitions and data requirements remain unchanged for sub-global applications of
the RLE.

3. Application of subcriterion B1 and its thresholds remains unchanged (Section 5.2). A

minimum convex polygon that encloses all occurrences of an ecosystem type is
applied, regardless of whether the edges cross the bounds of the area of
assessment. No holes or cutting of the minimum convex polygon are permitted,
regardless of the bounds of the area of assessment.

4. When the area of assessment is similar to or smaller than the EOO or AOO
thresholds for the Vulnerable category, listing of ecosystem types under criterion B
will depend solely on meeting the subcriteria. Research to support specific guidelines
and tools for applying criterion B in small assessment areas is currently in progress.

27 | IUCN Red List of Ecosystems

4.2 Describing the unit of assessment

To ensure repeatable application of the IUCN Red List of Ecosystems Categories and
Criteria, detailed description and definition of the assessment units is an essential
component of the assessment process. The description and assessment is based on a
comprehensive compilation of all available information about the ecosystem type under
consideration. The description of an ecosystem type must provide contextual information on
its classification; clearly describe four elements that define the ecosystem type
(characteristic native biota; abiotic environment; key processes and interactions; and spatial
distribution); and describe the threats and collapsed states.

Assessors should use the description template for ecosystem types (Table 5) and justify why
the unit selected for assessment is recognised as a separate ecosystem type from adjacent
or similar ecosystem types. What are the key features that distinguish the focal ecosystem
type from other ecosystem types? Information supporting the description of the ecosystem
type should be included in the assessment documentation, and will be assessed by peer
review. It is expected that all submissions to the global RLE will include relevant supporting
information including a fully populated reference list, maps, geographic coordinates,
exemplar photographs and any other information that will facilitate repeatability of the
assessment. These submissions will be openly accessible on the IUCN Red List of
Ecosystems website (www.iucnrle.org).

28 | IUCN Red List of Ecosystems

 Table 5. Description template for ecosystem types.

Elements Description
Classification Cross-references to relevant ecological classifications:
a. Source classification.
b. IUCN Habitats Classification Scheme.
c. Ecoregional classifications.
Spatial distribution Describe distribution and extent:
a. Accurate spatial distribution data.
b. Estimates of area.
c. Time series, projections (past, present, future).
Characteristic native biota Identify defining biotic features:
a. Diagnostic native taxa and their relative abundance in
comparison to other ecosystem types.
b. Functional components of characteristic biota and
their roles in the focal system compared to others.
c. Limits of spatial and temporal variability in the
ecosystem biota.
d. Exemplar photographs.
Abiotic environment Identify defining abiotic features:
a. Text descriptions and citations for characteristic states
or values of abiotic variables.
b. Graphical descriptions of abiotic variables.
c. Exemplar photographs.
Processes and interactions: Describe key ecosystem drivers:
– among biota a. Text descriptions and citations.
– between biota and b. Conceptual model.
environment c. Exemplar photographs.
Threats Describe major threats and impacts on ecosystem functioning:
a. Text descriptions and citations.
b. Diagnosis based on IUCN Threats Classification
Scheme.
c. Exemplar photographs.
Collapse definition Describe ecosystem-specific collapsed state(s) and
threshold(s).

4.2.1 Classification

Ecosystem types should be cross-referenced to any relevant ecosystem classifications,

including source classifications (such as vegetation classifications for terrestrial systems),
ecoregional classifications, and the IUCN Habitats Classification Scheme
(www.iucnredlist.org/technical-documents/classification-schemes/habitats-classification-
scheme-ver3). Further guidance is available for the use of the IUCN Habitats Classification
Scheme. Cross-referencing with the global RLE typology will be required when it becomes
available (3.1.1 Ecosystem typologies).

29 | IUCN Red List of Ecosystems

4.2.2 Spatial distribution

Information on the spatial distribution of an ecosystem type is best represented by maps or

inventories of localities. They can be derived from remote sensing, biophysical distribution
models, field observations or a combination of all three (Box 2). The spatial features of some
ecosystems (such as pelagic environments) are inherently dynamic over relatively short time
frames, so spatial distributions can only be described at very coarse levels of resolution.
Given the diversity of methods and maps available, an important aspect of the description is
to justify why a particular spatial dataset is an adequate representation of the ecosystem
distribution. Further information on clearly describing the spatial distribution of an ecosystem
type is provided in Sections 5.1 and 5.2. Assessors are encouraged to deposit the
ecosystem map in a suitable online repository.

Box 2. Distribution map of the Yellow Sea tidal flat ecosystem.

The distribution of the Yellow Sea tidal

flat ecosystem was mapped by
applying a peer-reviewed remote
sensing classification method to
Landsat Archive satellite imagery
(Murray et al., 2012; Murray et al.,
2014; Murray & Fuller, 2015). The
classified map has an overall accuracy
of >94% when assessed using a
confusion matrix, a widely
implemented method for assessing the
accuracy of classified maps (Murray et
al., 2014; Congalton & Green, 2008).
Estimates of the area of the ecosystem
type for criterion A are derived from the
distribution of the ecosystem (black),
whereas the area of occupancy (AOO)
is determined by counting the number
2
of 10x10 km cells in which >1 km of the ecosystem type occurs (Murray et al., 2015).

4.2.3 Characteristic native biota

The concept of characteristic native biota is central to ecosystem risk assessment and is
therefore an important component of their description (Box 3). The characteristic native biota
include the genes, populations, species, assemblages of species and their key interactions
that: (i) compositionally distinguish an ecosystem type from others (diagnostic components);
and (ii) are central in driving ecosystem dynamics and function, such as ecosystem
engineers, trophic or structural dominants, or functionally unique elements (functional
components). The diagnostic components of characteristic native biota should demonstrate
a level a compositional uniqueness and identify functionally important elements. In general,
the description need not include exhaustive species inventories.

30 | IUCN Red List of Ecosystems

Characteristic native biota are crucial in the diagnosis of ecosystem collapse because they
define part of the ‘identity’ of the ecosystem type. Thus, the loss of characteristic native biota
or processes in which they play a functional role signals a transformation of identity, collapse
of the ecosystem type and replacement by a novel system.

Characteristic native biota may be defined in terms of taxonomy or functional traits (e.g. guild
composition, trait spectra, structural features such as architecture of trees or corals) and
excludes exotic species and uncommon or vagrant species that contribute little to ecosystem
function. Examples of characteristic native biota include species that are endemic or near-
endemic to the ecosystem type, predators that structure the animal communities, tree
species that create microclimates in their canopies or at ground level, reef-building corals
and oysters that promote niche diversity for cohabiting fish and macro-invertebrates, nurse
plants and those that provide sites for predator avoidance, burrowing animals, guilds of
nitrogen fixers, key dispersal agents responsible for movement of biota or resources, peat-
forming plants, detritivore guilds, and flammable plants that promote recurring fires.

Box 3. Describing characteristic native biota (adapted from Appendix S2 in Keith et al., 2013).

Raised Bogs, Germany

This ecosystem type is characterised by vegetation dominated by peat mosses (e.g. Sphagnum
magellanicum, Sphagnum fuscum) and insectivorous plants like sundew (Drosera sp.). The
dominance by peat mosses together with geomorphic and hydrological processes distinguishes
raised bogs from other ecosystem types. Other typical species for raised bogs in Germany are the
vascular plants bog-rosemary (Andromeda polifolia) and cranberry (Vaccinium oxycoccos), the
butterfly species Boloria aquilonaris (Cranberry Fritillary), the moth Carsia sororiata (Manchester
Treble-Bar) and the ground beetle Agonum ericeti (Blab et al., 1995).

Great Lakes Alvar, North America

This ecosystem type is characterised by a variable physiognomy, from open perennial (rarely
annual) grassland or shrubland and nonvascular pavement (5-25% herb and or shrub cover) to
dense grassland or shrubland (>25%) with scattered evergreen needleleaf (more rarely broad-leaf
deciduous) trees (Reschke et al., 1999; Catling & Brownell, 1995). Species composition contains a
mix of tallgrass prairie graminoids and forbs and sub-boreal to boreal shrubs and trees. Key
dominants and differentials include the perennials Schizachyrium scoparium, Sporobolus
heterolepis, Danthonia spicata and Deschampsia caespitosa; less commonly with Sporobolus
neglectus, Sporobolus vaginiflorus, and Panicum philadelphicum. Key shrubs, when present, are
Juniperus communis, J. horizontalis, Dasiphora fruticosa ssp. floribunda and Rhus aromatica.
Trees, when present, include Thuja occidentalis, Picea glauca, Pinus banksiana, and Abies
balsamea (in more northern sites) and Juniperus virginiana, Quercus macrocarpa or Quercus
muehlenbergii (more southern sites).

Giant Kelp Forests, Alaska

Alaskan kelp forests are structurally and functionally diverse assemblages. They are characterised
by species of brown algae in the Order Laminariales including Nereocystis luetkeana, Laminaria
groenlandica, Alaria fistulosa, Agarum fimbriatum and Thalassiophyllum sp. (Steneck et al., 2002).
These create a complex and dynamic layered forest architecture up to 15 m tall that provides
substrate, shelter and foraging resource for a diverse fauna assemblage of epibenthic invertebrate
herbivores and pelagic vertebrate predators (Steneck & Watling, 1982; Estes et al., 2009).
Characteristic invertebrates include urchins, Strongylocentrotus franciscanus, S. purpuratus and
S. droebachiensis, limpets, and starfish, Solaster spp. Fish, including the Pacific cod (Gadus

31 | IUCN Red List of Ecosystems

 macrocephalus) and rock greenling (Hexagrammos lagocephalus, are important predators that
depend directly or indirectly on the ecosystem (Reisewitz et al., 2006). Characteristic
mesopredators include sea otters, (Enhydra lutris), harbour seals (Phoca vitulina), Steller sea lions
(Eumetopias jubatus) and northern fur seals (Callhorinus ursinus). Steller's sea cow
(Hydrodamalis gigas), now extinct, was a functionally unique herbivorous member of the
vertebrate assemblage (Domning, 1972). Large pelagic predators are also important components
of the ecosystem, including killer whales (Orcinus orca) and over 15 species of great whales
including sperm (Physeter macrocephalus) and fin whales (Balaenoptera physalus). Kelp forests
are generally separated geographically by continental land masses or deep sea. The Alaskan kelp
forests are continuous with those of California, but differ compositionally in their more diverse
assemblage of macroalgae, including Macrocystis pyrifera.

Shallow under-ice benthic invertebrate communities, Antarctica (source: Clark et al., 2015)
Under-ice communities are typically composed of a mix of sessile suspension feeders and mobile
macro-invertebrates, elements of which are reminiscent of deep-sea fauna but occur at depths as
shallow as a few metres. Sessile fauna include Porifera (Demospongia, Hexactinellida, Calcaria),
Gorgonaria, Pennatularia, Alcyonaria, Stolonifera, Hydrozoa, Actiniaria, Bryozoa, Brachiopoda,
Polychaeta, and both solitary and colonial Ascidiacea (Dayton, 1990; Gili et al., 2006). Dominance
of some sessile taxa is known to occur at local scales, such as by sponges (Dayton, 1979,
McClintock et al., 2005) and ascidians (pers. obs). Fauna with fragile skeletons are distinctly
abundant, which is thought to be due to the lack of durophagous (skeleton crushing) predators
(Aronson & Blake, 2001) but may also relate to low wave energy in ice-protected coasts. Mobile
invertebrates occur with these sessile fauna or can dominate in some areas. Commonly occurring
taxa include Echinodermata (Echinoidea, Asteroidea, Ophiuroidea, Holothurioidea) and
Peracarida (Amphipoda, Isopoda, Tanaidacea, Mysidacea, Cumacea) both of which are very
successful in Antarctica and can exhibit high abundances or dominance of particular species.
Other common mobile epifauna include Pycnogonida, Ostracoda, Caridea, Teleostei,
Prosobranchia, Opisthobranchia, Polyplacophora, Bivalvia and Nemertinea (Dayton, 1990; Gili et
al., 2006). Many of these are symbionts and use sessile invertebrates as habitat, including
specialised predators such as nudibranches, asteroids, and gastropods. Some fauna such as the
pycnogonids display gigantism, where individuals grow to much larger sizes than related taxa in
non-polar regions (Chapelle & Peck, 1999).

4.2.4 Abiotic environment

Descriptions should identify salient abiotic features that influence the distribution or function
of an ecosystem type, define its natural range of variability, sustain its characteristic native
biota, and differentiate it from other systems. For terrestrial ecosystems, salient abiotic
features may include substrates, soils and landforms, as well as ranges of key climatic
variables, while those of freshwater and marine ecosystems may include key aspects of
water regimes, light regimes, tides, currents, climatic factors and physical and chemical
properties of the water column (Box 4).

32 | IUCN Red List of Ecosystems

 Box 4. Describing the abiotic environment (adapted from Appendix S2 in Keith et al., 2013)

Gnarled Mossy Cloud Forest, Lord Howe Island, Australia (source: Auld & Leishman, 2015)
The Lord Howe Island Gnarled Mossy Cloud Forest occurs on the summit plateau and ridgetops
of two mountains on Lord Howe Island. The climate is temperate, and sea level parts of the island
have a mean annual temperature of 19.2°C, ranging from 17°C–25°C in summer to 14°C–18°C in
winter (Mueller-Dombois & Fosberg, 1998). At sea level, average annual rainfall is 1,717 mm, with
a maximum of 2,886 mm and a minimum of 998 mm (Mueller-Dombois & Fosberg, 1998).
Temperature decreases with altitude in the southern mountains (0.9°C for every 100 m rise in
altitude; Simmons et al., 2012). Cloud forests on Pacific islands typically occur between 800 and
900 m a.s.l. (Meyer, 2011), and on Lord Howe Island, the Gnarled Mossy Cloud Forest ecosystem
occurs from 750 to 875 m a.s.l. The annual rainfall in Gnarled Mossy Cloud Forest is thought to be
much higher than at sea level (although this has not been quantified) and spread throughout the
year (DECC, 2007). The two southern mountains (Mounts Gower and Lidgbird) obtain significant
moisture from both rainfall and direct canopy interception of cloud water (horizontal precipitation or
cloud stripping), and their peaks are often shrouded in cloud (Auld & Hutton, 2004). Cloud forests
are characterised by increased rainfall and cooler temperatures than forest with no cloud (Jarvis &
Mulligan, 2011), and this is thought to also apply to the Gnarled Mossy Cloud Forest ecosystem
(Auld & Leishman, 2015).

Yellow Sea Tidal Flats, East Asia (source: Murray et al., 2015)
The Yellow Sea is a shallow (mean depth c. 45 m), semi-enclosed sea with surrounding
geography varying from mountain ranges in South Korea to low-elevation coastal plains across
much of the northern and western regions (Healy et al., 2002; MacKinnon et al., 2012). As such,
tidal flats in the Yellow Sea are among the largest on Earth; in areas with high tidal amplitude
(macrotidal, >4 m) they may attain a width of nearly 20 km when exposed at low tide (Healy et al.,
2002). A key feature of the Yellow Sea tidal flats is the seasonal switching from an erosion- to
accretion-dominated system in some areas, depending on the occurrence of the monsoon season
(Wang & Zhu, 1994). The ecosystem is dependent on the continuing operation of a suite of
coastal processes that are focused on sediment transport and dynamics. Sediments are
transported to tidal flats by coastal and tidal currents, where the deposition process is influenced
by factors such as sediment texture and size, occurrence of vegetation, wave dynamics, rainfall
and the composition of the benthic community, which facilitates local bioturbation, biodeposition
and biotransportation (Wang et al., 2012). Storms, wind and wave action cause seaward erosion
of tidal flats, and compaction and subsidence reduce their elevation, so sediment trapping and
replenishment are required to offset these processes and maintain tidal flat extent. However, a
feature that distinguishes tidal flats in the Yellow Sea from adjacent regions is that the tidal flat
ecosystem is largely erosion-dominated, requiring ongoing sediment replenishment and transport
to persist (Healy et al., 2002). Therefore, disruption of sediment provision via reduced supply from
sources such as rivers, and interruption of sediment transport and deposition mechanisms are
considered the primary processes that lead to degradation of the ecosystem (Wang et al., 2012).

4.2.5 Processes and interactions

A qualitative understanding of ecosystem dynamics is essential for assessing risks related to

functional declines. Generic mechanisms of ecosystem dynamics can often be inferred from
related systems if the ecosystem type under assessment lacks direct studies. For example,
pelagic marine systems are typically dominated by trophic interactions in which elements of

33 | IUCN Red List of Ecosystems

the main trophic levels are known, even if particular predator-prey relationships are not
(Estes et al., 2009). Tree and grass dynamics in savannahs across the world are influenced
by fire regimes, herbivores and rainfall, although their relative roles may vary among
savannah types (Lehmann et al., 2014). All descriptions of ecosystem types should include a
narrative account of ecosystem dynamics that addresses key ecological processes defining
the identity and behaviour of the ecosystem type and the threats that may cause their loss or
disruption.

A conceptual model of key ecosystem dynamics is required for each ecosystem type as part
of its assessment. A conceptual model is a diagram of key ecosystem processes and
threats, and serves four purposes. First, the creation of a conceptual model compels
assessors to think through and clarify their assumptions and understanding of ecosystem
processes. Second, the conceptual model provides a basis for conducting the risk
assessment, by informing selection of relevant variables for assessing criteria C and D (5.3.3
Application). Third, the conceptual model is a communication tool that effectively
summarises key features of an ecosystem type for risk managers, conservation
practitioners, peer reviewers and the wider community. Finally, the conceptual model is
useful for underpinning the development of a quantitative model for criterion E.

Two types of conceptual models are particularly useful for RLE assessments: cause-effect
models and state-and-transition models (Box 5). Cause-effect models depict the interaction
and dependencies among model components, such as characteristic biota, the abiotic
environment and threats (Box 5a). State-and-transition models depict switches between
ecosystem states due to changes in the abiotic environment or ecosystem processes (Box
5b). For example, changes in the average water level determine transitions between the
degraded hypersaline and unhealthy hypersaline states in the Coorong lagoon (Appendix S2
in Keith et al., 2013; Lester & Fairweather, 2011; Lester & Fairweather, 2009).

A standard visual repertoire can help develop consistent cause-effect models (Fig. 7).
Characteristic biota are represented by green hexagons, the elements of the abiotic
environment by blue hexagons, biotic processes by green ovals, abiotic processes by blue
ovals, and threats by red rectangles. Positive, negative and hypothesised relationships can
be represented by appropriate symbols. The use of arrows accompanied by plus and minus
signs is discouraged. Distinct ecosystem components functioning together should form part
of a compartment. For example, the Gonakier forest in Senegal (Appendix S2 in Keith et al.,
2013) can be described by two faunal and floral compartments, driven by abiotic processes
that are influenced by threats (Box 5c).

General guidelines for developing conceptual models for RLE assessments include:

1. Conceptual models of ecosystem types should be complete, unambiguous and easy

to understand. They should be consistent with the narrative description of ecosystem
processes and functions, and should not introduce elements which have not been
described in the narrative. They should focus on processes especially relevant to the
application of criteria C and D, and to the definition of the collapsed state of the
ecosystem type.

34 | IUCN Red List of Ecosystems

 2. Overly complex conceptual models should be avoided, so models will typically
include fewer details than the narrative text. Assessors are encouraged to think
carefully about the level of complexity and hierarchical organisation of the conceptual
model, revisiting the purpose of developing conceptual model described above if
necessary. Overall, the least complex model covering all ecosystem processes will
be the most appropriate (typically fewer than 12 elements).

3. The inclusion of processes relevant to other ecosystem types (but not to the
ecosystem type of interest) is discouraged.

4. Repetition of components and relationships should be avoided.

5. Assessors are encouraged to use the IUCN Threats Classification Scheme to select
appropriate threats for their system. The inclusion of generic drivers such as human
population growth or economic factors is not recommended.

Figure 7. A common visual repertoire for cause-effect models.

Development of the conceptual model may reveal uncertainties in the understanding of

ecosystem processes. It may be necessary to draft two or more alternative conceptual
model to represent this uncertainty. Refining the model multiple times may help to explore
and refine ecosystem processes and clarify the layout of the model. An effort should be
made to reach a consensus conceptual model for the ecosystem type, using the narrative
text to highlight the greatest sources of uncertainty. When assessing criterion E, it may be
useful to include a second, more complex model to describe selected indicators and
modelled relationships among components. Tools to assist in construction of conceptual
models are in development, including a computer programme to support the development of
internally consistent conceptual models. The programme will allow users to save and
retrieve conceptual models for a range of ecosystems, use a common visual repertoire and
evaluate the effects of threats on ecosystem processes.

35 | IUCN Red List of Ecosystems

Box 5. Conceptual models representing processes and interactions

(A) Cause-effect model Cause-effect model of a Caribbean

coral reef (Appendix S2 in Keith et
al., 2013). Warming, pollution,
exploitation and acidification are
direct threats. Bleaching, rugosity
and connectivity are key ecosystem
processes. The system alters
between coral and algae-dominated
patches.

(B) State-and-transition model State-and-transition model of the

Coorong lagoon in Australia
(adapted from Appendix S2 in Keith
et al., 2013). Average salinity
determines shifts between the
unhealthy marine and degraded
marine states.

(C) Cause-effect model with compartments Cause-effect model of the Gonakier

forest in Senegal (Appendix S2 in
Keith et al., 2013).). The model is
composed of two compartments.
Forest regeneration depends on
floods, and contributes to leaf litter
production. Leaf litter production in
turn provides nutrients for the
aquatic fauna.

36 | IUCN Red List of Ecosystems

4.2.6 Threats

Accompanying the description of an ecosystem type should be a full review of threatening

processes that cause ecosystem change. Describing the threats to an ecosystem type
requires two elements: (i) a brief description and explanation of the primary threats causing
ecosystem change; (ii) identification of threats with reference to the Threats Classification
Scheme, used in The IUCN Red List of Threatened Species (IUCN, 2015). When combined,
the description of threatening processes and stresses, the threat classification under the
IUCN Threats Classification Scheme, and the conceptual model for an ecosystem type will
assist in identifying collapsed states and key variables for assessing change in abiotic and
biotic function. The framework below (drivers, threats and stresses) outlines how threats
affect ecosystems.

Background

Consistent terms for drivers, threats and stresses are needed for ecosystem assessment
(Table 6). A direct threat for one ecosystem type or organism can be an indirect threat for
another or pose no threat to other organisms. For example, unsustainable fishing will directly
threaten target and by-catch species and may also have indirect effects (negative or
positive) on species that prey upon, compete with or are preyed upon by targeted species.
This complexity of effects requires careful consideration and definition of threats for each
ecosystem type.

Table 6. Definitions of threats, drivers and stresses (Salafsky et al., 2008).

Term Definition Synonyms

Driver The ultimate factors, usually social, economic, political, Contributing factors,
institutional, or cultural that enable or otherwise add to the underlying factors,
occurrence or persistence of proximate direct threats. There is root causes, indirect
typically a chain of drivers behind any given direct threat. threat, pressures
Threat Direct threats are the proximate activities or processes that have Direct threats,
impacted, are impacting, or may impact the status of the sources of stress,
ecosystem being assessed (e.g., unsustainable fishing or pressures, proximate
logging). Threats can be past (historical), ongoing, and/or likely pressures, stressors
to occur in the future. Natural phenomena are also regarded as
direct threats in some situations.
Stress Stresses are the effects on ecosystem features that are impaired Symptoms, key
directly by threats (e.g. reduced abundance of keystone species, degraded attributes.
fragmentation of habitat). A stress is not a threat in and of itself,
but rather a degraded condition or symptom of the target that
results from a direct threat. The RLE risk protocol aims to
quantify these symptoms to assess declines towards collapsed
states.

37 | IUCN Red List of Ecosystems

Description of threats

A summary of the main threats currently affecting or likely to affect the ecosystem type is
required supporting information for all ecosystem types. The description provides a brief
explanation of the major threats (past, present and future), the drivers of those threats, and
the resultant stresses or symptoms of the ecosystem. Identifying stresses is highly
informative for defining collapsed states and assessing criteria C and D. The geographic
extent of threats should also be described. Assessors can base their description on regional
and/or national threats classifications, but these cannot be used directly within the IUCN
Threats Classification Scheme. In cases where a national threats classification must be
used, assessors should report both the national designation and the IUCN Threats
Classification Scheme. Graphs, figures and exemplary photographs are encouraged to
illustrate the impact of threats on the characteristic native biota, physical environment and
interactions among them. An example of threats description is provided in Box 6.

Threats Classification Scheme

The RLE adopts the IUCN Threats Classification Scheme (www.iucnredlist.org/technical-

documents/classification-schemes/threats-classification-scheme) for consistency with The
IUCN Red List of Threatened Species. The Threats Classification Scheme is hierarchical,
consisting of three levels with increasing detail, and contains 12 main threat categories. For
a RLE assessment, the description of threats to an ecosystem type must correspond with
threats from the IUCN Threats Classification Scheme. Coding for the major threats affecting
an ecosystem type is required as supporting information for all ecosystem types except
where there are no known threats to those assigned to the Data Deficient or Least Concern
categories. Assessors should diagnose and record threats to the lowest possible level in the
Threats Classification Scheme.

Coding of timing, scope and severity for each major threat is not required but can be
provided. If assessors decide to also record minor threats (threats affecting only a very small
proportion of the distribution), then it is essential that the timing, scope and severity be
described for all of the threats recorded. This will allow major and minor threats to be clearly
identified for the ecosystem type and assist higher level analyses of the RLE. Guidance for
using the IUCN Threat Impact Scoring System is available on the IUCN Red List of
Threatened Species website (www.iucnredlist.org). The Threat Impact Scoring System for
the RLE is currently under review, so use of the current Threat Impact Scoring System is not
required within a RLE assessment at this stage.

Although recording stresses from the IUCN Stresses Classification Scheme for each threat
selected is not required, this is highly recommended supporting information for an
assessment. This information is useful for demonstrating how threats are impacting
ecosystem types listed on the RLE, and may provide useful guidance for policy makers to
address ultimate causes. It is possible to record multiple stresses, simply by selecting threat
code.

38 | IUCN Red List of Ecosystems

Box 6. Describing threats

The Coolibah - Black Box Woodlands of south-eastern Australia is a flood-dependent woodland

ecosystem type affected by five main threats (Appendix S2 in Keith et al., 2013; NSW Scientific
Committee, 2004). Expansion and intensification of agricultural land use has replaced large areas
of woodland with crops and pastures in recent decades (Keith et al., 2009). Furthermore,
extraction of water from rivers for irrigation has altered flood regimes and their spatial extent,
reducing opportunities for reproduction and dispersal of characteristic flora and fauna (Thoms &
Sheldon, 2000; Thoms, 2003; Kingsford & Thomas, 1995; Kingsford & Johnson, 1998; Kingsford &
Auld, 2005). Future climate change may also affect the spatial and temporal availability of water in
the system. Invasive plants have spread with agricultural intensification and are reducing the
diversity and abundance of native biota. Additionally, invasion of the mat-forming forb Phyla
canescens reduces the diversity of native ground layer plants (Taylor & Ganf, 2005). This species
has spread rapidly in response to altered water regimes and persistent heavy livestock grazing
(Earl, 2003). Finally, overgrazing by feral goats, rabbits and domestic livestock has altered the
composition and structure of the woodland vegetation, through selective consumption of palatable
native ground layer plants and seedlings of trees and shrubs (Reid et al., 2011; Robertson &
Rowling, 2000). These effects are most marked beneath trees and around watering points where
livestock concentrate their activities.

The threats affecting this ecosystem type correspond with five threats (underlined) and their
hierarchical categories in the IUCN Threats Classification Scheme:

2. Agriculture & Aquaculture:

2.1 Annual & Perennial Non-Timber Crops:
2.1.3 Agro-industry Farming;
2.3 Livestock Farming & Ranching:
2.3.3 Agro-industry Grazing, Ranching or Farming
7. Natural System Modifications:
7.2 Dams & Water Management/Use:
7.2.3 Abstraction of Surface Water (agricultural use)
8. Invasive & Other Problematic Species, Genes & Diseases:
8.1 Invasive Non-Native/Alien Species/Diseases:
8.1.2 Named Species – Phyla canescens
11. Climate change & severe weather
11.2 Droughts

The description of threats and stresses underpinned the selection of variables for assessing
criteria C and D and clarified their link to collapse of this ecosystem type. Under criteria A and B,
the ecosystem type was “assumed to have collapsed when its mapped distribution has declined to
zero as a consequence of clearing for agriculture”. Because flood regimes are fundamental to
ecosystem dynamics and water extraction for irrigation is a major threat, median daily river flow
was identified as a suitable variable for assessing environmental degradation under criterion C.

39 | IUCN Red List of Ecosystems

4.2.7 Describing collapsed states

Ecosystem collapse is a key concept in the RLE (Section 3.2) and underpins the application
of the IUCN Red List of Ecosystems Categories and Criteria. Assessors should describe the
collapsed state(s) of an ecosystem, based on the information summarised in the description
of the ecosystem type and the conceptual model. If multiple states of collapse are possible
(e.g. due to different threats), all of these should be described with similar levels of detail.
Descriptions should focus on the key defining features of the ecosystem type. Collapse
thresholds for the application of criteria A and B are typically defined as 100% loss of spatial
distribution of the ecosystem type (i.e. 100% decline under criterion A; EOO = 0 km2 and/or
AOO = no 10x10 km grid cells occupied under criterion B). Choosing a different collapse
threshold for criterion A or B must be thoroughly justified. Collapse thresholds for the
application of criteria C, D, and E should be identified as part of the assessment of those
criteria (5.3.3 Application). Assessors are encouraged to provide examples of locally
collapsed occurrences of the ecosystem type to support their descriptions of collapsed
states.

Box 7. Defining ecosystem collapse

The Mountain Ash Forest of south-eastern Australia is a unique ecosystem dominated by the
world’s tallest flowering plant species (Eucalyptus regnans). Mountain ash supports a wide range
of plant species and a rich array of native mammals and birds, including the Endangered
Leadbeater’s possum and the Vulnerable yellow-bellied glider (Lindenmayer, 2009). The
availability of old-growth forest and natural tree hollows is a critical factor in the survival of cavity-
dwelling animals (Keith et al., 2013; Burns et al., 2015).

Ecosystem collapse is considered to occur under any of the following (Burns et al., 2015):

1. 100% of the area where the ecosystem currently occurs is no longer

bioclimatically suitable (criterion C).
2. The abundance of hollow-bearing trees drops below one per hectare averaged
across the entire ecosystem distribution (subcriterion D2 and criterion E).
3. Less than 1% of old-growth forest remains in the ecosystem (subcriteria D1 and
D3).

4.3 Evaluating the criteria

Each ecosystem type must be assessed against all of the RLE criteria so far as the available
data permit. To assist this purpose, Section 5 provides detailed information on how to gather
data, perform an assessment, consider data quality and uncertainty, and document an
assessment outcome. At the onset of an assessment, all ecosystem types are considered
Not Evaluated (NE) for all criteria (Fig. 8). The next step is to determine whether adequate
data exist for application of the criteria, which requires data searches of the scientific
literature, unpublished reports, expert opinion, historical accounts, past and present maps,
satellite imagery or any other source of relevant data. If no adequate data exists to assess
any of the criteria, the assessment outcome is Data Deficient (DD; Fig. 8).

40 | IUCN Red List of Ecosystems

Following this initial assessment of data, assessors must systematically evaluate all of the
IUCN Red List of Ecosystems criteria. If an assessor chooses not to apply a criterion, the
risk assessment outcome for this criterion is Not Evaluated. If a reasonable search effort
indicates that adequate data are not available to assess under a criterion, the risk
assessment outcome for this criterion is Data Deficient (DD). The difference between Not
Evaluated and Data Deficient is that reporting Not Evaluated for any criterion implies that no
attempt was made to obtain relevant data and assess the ecosystem type under that
criterion. If a decent attempt was made, but data were not available or inadequate, then Data
Deficient should be used. The search effort for appropriate data should be briefly described
in documentation.

Figure 8. Process of evaluating the IUCN Red List of Ecosystems criteria.

4.4 Assessment outcome

A summary table for each ecosystem type reports the assessment outcome for all criteria
(and subcriteria) as well as the overall status (Box 8). There are a total of 20 subcriteria in
the IUCN Red List of Ecosystems Categories and Criteria, each of which can be assigned
one of the eight risk categories (Fig. 1, Fig. 9). The results for all subcriteria under criteria A,
B, C, and D, as well as which method was used to assess the subcriteria (i,ii or iii), must be
reported during the assessment process.

Some ecosystem types will be Data Deficient or Not Evaluated for some of the subcriteria;
this must be included in the summary table (Box 8). If all subcriteria are Data Deficient, the
overall outcome of the assessment is Data Deficient. If all subcriteria are Not Evaluated, the
overall outcome of the assessment is Not Evaluated. If all subcriteria are either Not
Evaluated or Data Deficient, the overall outcome of the assessment is Data Deficient.

41 | IUCN Red List of Ecosystems

Following the precautionary principle and to ensure that the most severe symptoms of risk
determine the assessment outcome, the highest risk category obtained by any of the
assessed criteria will be the overall risk status of the ecosystem. The main method currently
used for representing uncertainty in ecosystem assessment is to use bounded estimates (
4.4.1 Dealing with uncertainty). The lower bound of the overall status is the highest lower
bound across any of the subcriteria that return the same category as the overall status. The
upper bound of the overall status is the highest upper bound across any of the subcriteria
that return the same category as the overall status. For example, if an ecosystem type
qualifies for EN (plausible bounds EN-CR) under criterion B, EN (plausible bounds VU-CR)
under criterion D, and lower categories under other criteria (e.g. NT under criterion A, LC
under criterion C, and DD under criterion E), then its overall status will be EN (plausible
bounds VU-CR).

Criteria Sub-criteria Numeral

A 1
2 a
b
3
B 1 a i
ii
iii
b
c
2 a i
ii
iii
b
c
3
C 1
2 a
b
3
D 1
2 a
b
3
E

Figure 9. The IUCN Red List of Ecosystems protocol comprises a total of five rule-based
criteria (A-E) and 20 subcriteria for assigning ecosystems to a risk category.

42 | IUCN Red List of Ecosystems

Box 8. Assessment outcome (adapted from Appendix S2 in Keith et al., 2013)

Caribbean Coral Reefs

Caribbean coral reefs are primarily fringing reefs and bank barrier reefs separated from island and
mainland shorelines by reef flats, shallow waters or slightly deeper lagoons (Alevizon, 2010). Due
to the difficulties of remotely measuring the distribution of live coral and mosaic marine
ecosystems, the ecosystem is listed as DD under all subcriteria of criterion A. The ecosystem is
assessed as LC under all subcriteria of criterion B due to its large extent of occurrence, area of
occupancy and number of threat-defined locations. The data for criterion C are currently under
review; at the time of writing, the ecosystem is assessed as NE under all subcriteria of criterion C.
Data on coral cover and reef rugosity both lead to similar estimates for subcriterion D1: EN
(plausible range VU – CR). No projections are available for future disruptions to biotic interactions,
so the ecosystem is listed as DD under D2. The ecosystem is listed as EN under subcriterion D3
based on historical data. No quantitative analysis has been carried out to assess criterion E, so
the status is NE under criterion E. The most appropriate overall risk status of Caribbean coral
reefs is determined to be EN (plausible range EN – CR).

Criterion A B C D E Overall
Subcriterion 1 DD A1 LC B1a,b,c NE C1 EN(VU-CR) D1 NE E EN(EN-CR)D1,D3*
Subcriterion 2 DD A2a,b LC B2a,b,c NE C2a,b DD D2a,b
Subcriterion 3 DD A3 LC B3 NE C3 EN D3
* Overall status should specify best estimate, plausible lower and upper bounds and all criteria
and full subcriteria that support the overall status (other examples: VU (VU-CR) B1ai,iii, B3, D2a;
CR(CR-CR) A2a, B2bii, C1b)

Coastal Sandstone Upland Swamps of south-eastern Australia

The Coastal Sandstone Upland Swamps of south-eastern Australia are treeless bogs that form
relatively abrupt boundaries with surrounding eucalypt-dominated forests and woodlands that
occupy more freely draining soils (Keith & Myerscough, 1993). They are strongly associated with
high rainfall and moisture. Interactions between hydrological processes and fire regimes are
crucial to the development of upland swamps and maintenance of their diverse and characteristic
biota. To assess potential future decline due to climate change, Keith et al. (2013) used a range of
plausible bioclimatic distribution models to predict its distribution under future climate scenarios.
Based on these models and scenarios, the distribution of the ecosystem was projected to decline
by 58-90% (median 74%) over the next 50 years. The most appropriate status of the ecosystem
was therefore determined to be EN (plausible range EN – CR) under subcriterion A2. The same
distribution models used to assess future change in distribution were also used to assess trends in
climatic suitability under criterion C. From 1983 to 2009, the summed abundance of woody re-
sprouters declined by a mean of 37% at 72% of sampled sites. These are just below the severity
and extent thresholds, respectively, for VU under criterion D1, assuming that zero abundance of
re-sprouters marks the point of ecosystem collapse. No data are available prior to 1983, but if
current declines were initiated prior to that time, they may exceed the threshold for Vulnerable
status. The status of the ecosystem type is likely to be NT (plausible range NT – VU) under
subcriterion D1. The most appropriate overall risk status of the Coastal Sandstone Upland
Swamps of South-Eastern Australia is EN (plausible range EN – CR).
Criterion A B C D E Overall
Subcriterion 1 LC A1 EN B1b,c LC C1 NT(NT-VU) D1 DD E EN(EN-CR)
A2a, C2a
Subcriterion 2 EN(EN-CR)A2a EN B2b,c EN(EN-CR)C2a DD D2a,b
Subcriterion 3 LC A3 LC B3 DD C3 DD D3

43 | IUCN Red List of Ecosystems

4.4.1 Dealing with uncertainty

Uncertainty in any information used to evaluate the criteria should be propagated through
the assessment and reported as part of the outcome. Reporting both the most likely risk
category and other plausible categories, given the uncertainties in the data, is more useful
than simply reporting the most likely category. The simplest means of characterising
uncertainty is through bounded estimates. Bounded estimates represent a range of plausible
alternative values for a measure. They can take into account uncertainty in thresholds
describing collapsed states (Fig. 3 and Box 1), mapped estimates of change in distribution
(Box 9), and estimates of variables for measuring relative severity in criteria C and D (Boxes
11 and 12). The upper and lower bounds of an estimate may be propagated through an
assessment by repeating the same analysis for the best estimate, and the lower and upper
bounds. For example, if the decline in an ecosystem type’s distribution is estimated to be
between 75-85% in the last 50 years, it could plausibly be either Endangered (decline
between 50-80% based on the lower bound) or Critically Endangered (≥80% based on the
best estimate and upper bound) under subcriterion A1. Dealing with uncertainty in
ecosystem risk assessment draws largely on the experiences of The IUCN Red List of
Threatened Species (Newton, 2010; Regan & Colyvan, 2000; Akcakaya et al., 2000).

4.5 Documentation

All assessments must be accompanied by documentation and supporting information, which

should undergo peer review by appropriate experts (6. Peer review and publication), and
must be readily available when the assessment is completed (see the IUCN Red List of
Ecosystems website for examples: www.iucnrle.org). All required fields in the online RLE
database should also be completed (see the IUCN Red List of Ecosystems website). The
documentation must include the following sections:

1. Summary. A brief abstract (~200 words) that describes the complete assessment in
summarised form, including the area of assessment, the focal ecosystem type and its
defining features, threatening processes and the assessment outcome.

2. Ecosystem description. A complete description of the ecosystem type, including the

elements listed in Table 5.

3. Risk assessment. This section must include specific information on the application
and outcome of each criterion e.g. inferences, statistical analyses and spatial
analyses. It should also include a discussion of assumptions, limitations or further
data required. Further guidance is available in Section 5.

4. References. A complete reference list showing the sources of information used for
the assessment must be provided.

44 | IUCN Red List of Ecosystems

5. Criteria and thresholds

The IUCN Red List of Ecosystems (RLE) risk assessment model includes five criteria for
assessing the risk of ecosystem collapse (Table 6). This section outlines the theory,
thresholds and subcriteria relevant for the application of each criterion. A summary table of
the current IUCN Red List of Ecosystems Categories and Criteria is provided in Appendix 2.

Table 6. Purpose of the IUCN Red List of Ecosystems criteria.

Criterion Purpose
A Reduction in geographic distribution Identifies ecosystems that are undergoing
declines in area, most commonly due to threats
resulting in ecosystem loss and fragmentation.
B Restricted geographic distribution Identifies ecosystems with small distributions that
are susceptible to spatially explicit threats and
catastrophes.
C Environmental degradation Identifies ecosystems that are undergoing
environmental degradation.
D Disruption of biotic processes or Identifies ecosystems that are undergoing loss or
interactions disruption of key biotic processes or interactions.
E Quantitative analysis that estimates the Allows for an integrated evaluation of multiple
probability of ecosystem collapse threats, symptoms, and their interactions.

45 | IUCN Red List of Ecosystems

5.1 Criterion A. Reduction in geographic distribution

5.1.1 Theory

A decline in geographic distribution—defined as all spatial occurrences of an ecosystem

type—influences its risk of collapse by: (i) reducing the ability of an ecosystem to sustain its
characteristic native biota; and (ii) predisposing it to additional threats (Keith et al., 2013).
The loss of characteristic native biota due to a declining distribution typically occurs through
a combination of reduced carrying capacity, niche diversity, spatial partitioning of resources,
and increased susceptibility to competition, predation and threats (MacArthur & Wilson,
1967; Shi et al., 2010; Harpole & Tilman, 2007; Hanski, 1998; McKnight et al., 2007). The
rate of decline in an ecosystem distribution indicates its trajectory towards collapse, with
ecosystem collapse typically occurring when no spatial occurrences of the ecosystem type
remain (extent of distribution collapses to zero).

5.1.2 Thresholds and subcriteria

An ecosystem may be listed under criterion A if it meets the thresholds for any of four
subcriteria (A1, A2a, A2b or A3), quantified as a reduction in geographic distribution over the
following time frames:

Subcriterion Time frame CR EN VU

A1 Past (over the past 50 years) ≥ 80% ≥ 50% ≥ 30%
A2a Future (over the next 50 years) ≥ 80% ≥ 50% ≥ 30%
A2b Any 50 year period (including the past, present and ≥ 80% ≥ 50% ≥ 30%
future)
A3 Historical (since approximately 1750) ≥ 90% ≥ 70% ≥ 50%

5.1.3 Application

Data requirements

The rate of decline in distribution is typically estimated from time-series data appropriate for
the focal ecosystem type. Ecosystem maps — such as those derived from remote sensing
classifications, distribution models, field observations, or historical data — are a principal
data source for assessing criterion A. When more than one source of data is available, such
as different vegetation maps or estimates produced with different methods, assessors
should first critically evaluate the efficacy of the alternatives as representations of the
distribution of the ecosystem type. If more than one data source is suitable, assessors can
calculate estimates of area from each data source, and explore the sensitivity of ecosystem
status to this data uncertainty (Section 4.4.1 Dealing with Uncertainty). The net reduction in
geographic distribution will then form an interval of estimates generated from each data
source.

Remote sensing is a common approach for mapping distributions of many terrestrial and
marine ecosystems that have interpretable signatures from different sources of remote
sensing data (e.g. Fig. 10). Where regional or local data sets are lacking, global data sets,

46 | IUCN Red List of Ecosystems

such as those available for forests (Hansen et al., 2013), mangroves (Giri et al., 2011), water
cover (Pekel et al., 2016), and coral reefs (Andréfouët et al., 2006) may be suitable
templates for superimposing appropriate classifications of ecosystem types.

Figure 10. Time series maps of an ecosystem distribution inform the risk of ecosystem
collapse. Here, historical topographic maps (1954) and Landsat Archive satellite imagery
(1981, 2010) allowed a standardised time-series of the area of the Yellow Sea tidal flat
ecosystem to be developed for assessment under criterion A (Murray et al., 2014; Murray et
al., 2015; Murray et al., 2012).

In some cases, spatial proxies for ecosystem distributions may be used, such as field
observations of organism assemblages, keystone species, climate, substrate, topography,
bathymetry, ocean currents, flood regimes, water cover, aquifers or some synthesis of these
that can be justified as valid representations of the distribution of ecosystem biota or its
niche space. For example, maps of physical factors such as sea floor characteristics, ocean
currents, water temperatures and water chemistry may be appropriate for marine
ecosystems. In some subterranean, freshwater and marine ecosystems, trends in the depth
dimension may be appropriate proxies of declines in distribution, so long as they reflect
trends in carrying capacity and niche diversity for characteristic biota (Keith et al., 2013).

Spatial distribution models offer an additional opportunity to formally select and combine the
most suitable set of spatial proxies to predict ecosystem distributions. For example, Clark et
al. (2015) used bathymetric spatial data and remote sensing data on sea ice concentration to
model the distribution of suitable light conditions for under-ice marine benthic invertebrate
communities in Antarctica. Models are especially useful for projecting time series of
ecosystem distributions into the future for assessing criterion A2. Keith et al. (2014)
modelled the distribution of a mire ecosystem under future climate scenarios using a map of

47 | IUCN Red List of Ecosystems

present day mires developed from satellite imagery, in combination with hydrologically-
based climate, substrate and terrain predictor variables. In both studies, a mechanistic
understanding of the relationship between occurrence of the ecosystem and limiting
environmental factors was central to developing an adequate ecosystem map.

Methods

To apply criterion A, at least two comparable estimates of the geographic distribution of the
ecosystem type at different points in time are required. It is beyond the scope of these
guidelines to provide detailed information on the acquisition, classification and accuracy
assessment of spatial data. Nevertheless, it is assumed that spatial data used for
assessments under criterion A are suitable for the purpose in being: (i) consistent and
comparable across time periods (unbiased); (ii) sufficiently accurate (Congalton & Green,
2008); and (iii) of a suitable grain size for the ecosystem type being assessed (Murray et al.,
2017). Although assessments can be completed with just two data points (see below), efforts
should be made to ensure appropriate power in a suitable statistical model of ecosystem
change and that all model assumptions are addressed in the analysis. Good practices in
data processing and analysis (Olofsson et al., 2014; Olofsson et al., 2013; Fuller et al., 2003)
should be employed to minimise bias in estimates of areal change over a time-series of
spatial data.

Subcriterion A1 may be directly assessed if data are available for 50 years ago and the
present. However, it is rare for the raw data to be available for precisely the time frames
required by an assessment of criterion A. More typically, assessors must use methods of
interpolation, extrapolation, or prediction to calculate estimates of distribution change over
the last 50 years (A1), the next 50 years (A2), and/or since 1750 (A3). This will involve
assumptions about the nature or pattern of change (see below), as well as the quality of the
data (Alaniz et al., 2016), which must be explained and justified in the documentation.

To assist calculations, a spreadsheet tool is available on the IUCN Red List of Ecosystems
website (www.iucnrle.org). Several tools for assisting in this step are in development and will
become available on the website in the future.

Assumptions

Whether inferences are made from time series of satellite images or from other data
sources, two important aspects will fundamentally influence assessments: (i) assumptions
about the rate of decline; and (ii) the number of points in the time series. When the rate of
decline is estimated from two observations (e.g. maps) over a specified time frame,
assessors should use information about the causes and context of the decline to deduce the
likely trajectory of decline (Fig. 11).

Although criterion A can be applied acceptably with only two data points, more data enables
a more certain diagnosis of the shape of the trajectory, allows the fitting of alternative
models, and hence will result in more accurate interpolation, extrapolation or prediction to
the full time frames required by criterion A. Selection of candidate models should always be
informed by the causes and context of the decline and assessors should ensure that the
assumptions of the model are adequately met. At least two plausible alternative scenarios

48 | IUCN Red List of Ecosystems

should be explored and all sources of uncertainty in spatial data and decisions about
assessment time frames should be clearly described and justified (e.g. Alaniz et al 2016).

2
Figure 11. All distribution size trajectories in this figure have the same endpoints: 300 km in
2
1970 and 100 km in 2010. A simple interpolation between the two extremes assumes linear
decline (left panel). Addition of intermediate distribution size estimates could reveal that the
decline is not linear (middle panel). Different ecosystem types could also exhibit contrasting
trajectories with identical endpoints: future projections of distribution considering these
trajectories would clearly differ (right panel).

Figure 12 illustrates two alternative scenarios of decline for a model ecosystem type based
on Coolibah – Black Box Woodland, an ecosystem on a semi-arid floodplain in eastern
Australia (Keith et al., 2009). As a first scenario (ARD), a constant area is lost each year,
producing a linear pattern of decline (Figure 12). An alternative scenario (PRD), where the
rate of decline is proportional, a constant fraction of the remaining distribution is lost each
year and the area lost reduces over time (Fig. 12). These scenarios may be modelled using
exponential (PRD) and linear (ARD) functions, as defined by Keith et al. (2009):

1
𝐴𝐴𝐴𝐴𝐴𝐴𝑎𝑎 (𝑌𝑌𝑌𝑌𝑌𝑌𝑟𝑟𝑡𝑡2 −𝑌𝑌𝑌𝑌𝑌𝑌𝑟𝑟𝑡𝑡1 )
Proportional rate of decline: 𝑃𝑃𝑃𝑃𝑃𝑃 = 100 × �1 − �𝐴𝐴𝐴𝐴𝐴𝐴𝑎𝑎𝑡𝑡2 � �
𝑡𝑡1

𝐴𝐴𝐴𝐴𝐴𝐴𝑎𝑎𝑡𝑡2 −𝐴𝐴𝐴𝐴𝐴𝐴𝑎𝑎𝑡𝑡1
Absolute rate of decline: 𝐴𝐴𝐴𝐴𝐴𝐴 = 𝑌𝑌𝑌𝑌𝑌𝑌𝑟𝑟𝑡𝑡2 −𝑌𝑌𝑌𝑌𝑌𝑌𝑟𝑟𝑡𝑡1

The predicted changes of these alternative models become more different the further they
are extrapolated into the future. In the absence of any other information, examining rates of
decline as proportional (PRD) or absolute (ARD) permits an assessment of ecosystem
status under these two relatively optimistic and pessimistic scenarios (Box 9). However, a
longer time series of observations — together with an understanding of the drivers of
change, the regulatory context, regional variability in land suitability, and the extent of
protected tenures across the distribution of the ecosystem — can help to select more
realistic models (Keith et al., 2009). More realistic models will produce narrower bounds of
uncertainty on the estimated change in distribution. For example, ecosystems in the early
stages of large-scale exploitation may be more likely to exhibit linear patterns of decline
(ARD) than those in an advanced stage of decline, where the area lost over time will
eventually reduce to zero with diminishing area (Puyravaud, 2003).

49 | IUCN Red List of Ecosystems

 Figure 12. Alternative scenarios for decline in distribution of a model ecosystem (Keith et al.,
2009; Keith et al., 2013). The figure shows an ecosystem with an initial area (1974) of 1,000
2
km . It declined at rate of 2% per year during the following 50 years, but the outcome was
substantially different if the decline was proportional (PRD) or absolute (ARD). In a PRD, the
decline is a fraction of the previous year’s remaining area (0.02 × last year’s area), whereas in
an ARD the area subtracted each year is a constant fraction of the area of the ecosystem at
2
the beginning of the decline (0.02 × 1000 = 20 km /year). Under a PRD scenario, this
ecosystem would be considered Endangered under A2b (50% decline over any 50 year
period including the present and future), while under an ARD scenario if would have
disappeared by 2024, and be assessed as Collapsed.

Box 9. Proportional and absolute rate of decline (criterion A)

Sierra de Perijá is the mountain range that separates north-western Venezuela from north-eastern
Colombia. The humid forests in the Venezuelan side of Perijá are threatened by the expansion of
large-scale commercial agriculture, primarily of a tuber, the arrowleaf elephant ear (Xanthosoma
saggittifolium). Using Landsat satellite images, it was estimated that in 1986 the humid forests of
2
the watersheds of the Guasare, Socuy and Cachirí rivers occupied 328 km , while in 2001 they
2
had decreased to 198 km . These two estimates allow assessment of ecosystem status under
subcriterion A2b, using 1986-2001 to first estimate an observed rate of change over 15 years, and
then extrapolating projected losses to 2036 (Portillo 2014).

2
The forests in 2001 occupied 198 km or 60.4% of their area in 1986, thus declining at a mean
proportional rate of 3.3% per year. The next step is considering how this rate may change over
time to project losses at 2036. Assuming a proportional rate of decline (PRD) between 2001 and
2036 results in a total decline of 81.5% between 1986 and 2036. Assuming an absolute rate of
decline (ARD) it is predicted to decline by 100% by 2024. Therefore, under criterion A2b PRD
leads to a classification of Critically Endangered (≥80% decline over any 50 year period including
the present and future), while ARD leads to a classification of at least Critically Endangered (≥80%
decline over any 50 year period including the present and future), although it seems unlikely to
collapse entirely if fragments of forest remain in less accessible mountain terrain. In conclusion,
the ecosystem is considered Critically Endangered (CR) under subcriterion A2b (Portillo 2014).
Information on the most likely shape of decline can help determine which of these two plausible
categories should be reported as the best estimate.

50 | IUCN Red List of Ecosystems

Documentation

Assessors should: (i) cite data repositories for time-series maps of ecosystem distributions
used in the assessment (see the IUCN Red List of Ecosystems website for a list of preferred
spatial data repositories: www.iucnrle.org); (ii) provide full bibliographic references; (iii) justify
why the spatial data used are an adequate representation of distribution of the focal
ecosystem type; (iv) justify assumptions and alternative scenarios used to interpolate,
extrapolate or predict changes in distribution from the available data; (v) explain the methods
of calculation including the assumed threshold of collapse. In addition, assessors are
encouraged to describe the source of the spatial data (such as satellite sensor type) and its
spatial resolution (grain size), and comment on the accuracy of all classified maps.

51 | IUCN Red List of Ecosystems

5.2 Criterion B. Restricted geographic distribution

5.2.1 Theory

The size of the geographic distribution of an ecosystem influences its risk of collapse when
confronted with a spatially explicit threat or catastrophe (Keith et al., 2013). In general,
ecosystems that are widely distributed or exist across multiple independent patches are at
lower risk from catastrophes, disturbance events or any other threats that exhibit a degree of
spatial contagion (e.g. invasions, pollution, fire, forestry operations, and hydrological or
regional climate change). The primary role of criterion B is to identify ecosystems whose
distribution is so restricted that they are at risk of collapse from the chance occurrence of
single or few interacting threatening events (Rodríguez et al., 2015). Criterion B also
includes an approximation for an estimate of occupied habitat for component biota, which is
positively related to population viability irrespective of exposure to catastrophic events.

5.2.2 Thresholds and subcriteria

An ecosystem may be listed under criterion B if it meets the thresholds for either of three
subcriteria (B1, B2 and B3), which indicate restricted geographic distribution as follows:

Subcriterion Measure of geographic distribution CR EN VU

2
Extent of a minimum convex polygon (km ) enclosing all
B1 ≤ 2,000 ≤ 20,000 ≤ 50,000
occurrences (extent of occurrence, EOO) is:
AND at least one of the following (a-c):
(a) An observed or inferred continuing decline in ANY OF:
i. a measure of spatial extent appropriate to the
ecosystem; OR
ii. a measure of environmental quality appropriate to
the characteristic biota of the ecosystem; OR
iii. a measure of disruption to biotic interactions
appropriate to the characteristic biota of the
ecosystem.
(b) Observed or inferred threatening processes that are
likely to cause continuing declines in geographic
distribution, environmental quality or biotic interactions
within the next 20 years.
1 ≤5 ≤ 10
threat- threat- threat-
(c) Ecosystem exists at:
defined defined defined
location locations locations
The number of 10×10 km grid cells occupied (area of
B2 ≤2 ≤ 20 ≤ 50
occupancy, AOO) are:
AND at least one of a-c above (same as for B1).
A very small number of threat-defined locations (generally fewer than 5) AND
prone to the effects of human activities or stochastic events within a very short
B3 time period in an uncertain future, and thus capable of Collapse or becoming VU
Critically Endangered (CR) within a very short time period (B3 can only lead to a
listing as Vulnerable, VU).

52 | IUCN Red List of Ecosystems

5.2.3 Application

Data requirements

The geographic distribution of an ecosystem type is assessed under criterion B with two
standardised metrics: the extent of occurrence (EOO) and the area of occupancy (AOO)
(Gaston & Fuller, 2009; Keith et al., 2013). In addition, assessment of criterion B requires a
qualitative evaluation of whether continuing declines in spatial extent, environmental quality,
or increasing disruption of biotic interactions are occurring or likely to occur as a result of
threats. Lastly, it requires an estimate of the number of threat-defined locations at which an
ecosystem occurs. Thus, accurate maps of the current distribution of an ecosystem,
information about the direction of current trends, and an understanding of the threats
influencing the ecosystem are needed (Keith et al., 2013). For further information on data
sources and the requirements of distribution maps for application in the RLE, refer to Section
5.1.3.

In some cases, spatial data may be insufficient to estimate EOO or AOO, but there is
evidence that a small number of plausible threatening events may cause an ecosystem to
become Critically Endangered or Collapsed within the near future. Such ecosystems may be
eligible for listing as Vulnerable under criterion B3 if they occupy few threat-defined locations
relative to the extent of threatening events. Distribution maps, locality records or expert
knowledge are required to determine the number of threat-defined locations in which an
ecosystem occurs.

Methods

Assessing spatial metrics for criteria B1 and B2

The two standardised measures of ecosystem distribution represent conceptually different

aspects of geographic range size for both species (Gaston, 1994; Gaston & Fuller, 2009)
and ecosystems (Keith et al., 2013; Murray et al., 2017). The EOO (subcriterion B1)
measures the spread of risk over a contiguous area that encloses all occurrences using a
minimum convex polygon. In contrast, the AOO (subcriterion B2) measures the spread of
risk among occupied patches through a count of occupied grid cells (Keith et al., 2013).

AOO and EOO have been shown to perform better than other spatial distribution metrics
(such as mean patch area, core area) for predicting the risk of ecosystem collapse in
landscapes subject to stochastic threats (Murray et al., 2017). These measurement protocols
are appropriate for all assessment units, including ecosystem types with depth dimensions
or particular distribution patterns, such as linearly occurring ecosystem types.

Ensuring standardized application of these methods is critical for objective measurement of

the size of a spatial distribution. Therefore, in no cases should AOO or EOO be measured in
ways that do not comply with the methods specified below:

53 | IUCN Red List of Ecosystems

 1. Extent of occurrence (EOO). The EOO of an ecosystem is measured by determining
the area (km2) of a minimum convex polygon – the smallest polygon that
encompasses all known occurrences of a focal ecosystem in which no internal angle
exceeds 180 degrees – fitted to an ecosystem distribution. The minimum convex
polygon (also known as a convex hull) must not exclude any areas, discontinuities or
disjunctions, regardless of whether the ecosystem can occur in those areas or not.
Regions such as oceans (for terrestrial ecosystems), land (for coastal or marine
ecosystems), or areas outside the study area (such as in a different country) must
remain included within the minimum convex polygon to ensure that this standardised
method is comparable across ecosystem types. In addition, these features contribute
to spreading risks across the distribution of the ecosystem by making different parts
of its distribution more spatially independent.

2. Area of occupancy (AOO). Measures of AOO are highly sensitive to the grain size
(pixel resolution) at which the AOO is estimated (Nicholson et al., 2009), so all
measures of AOO of an ecosystem type must be standardised to a common spatial
grain (Keith et al., in review). The AOO of an ecosystem defined in the RLE is
determined by counting the number of 10×10 km grid cells that contain the
ecosystem. This relatively large grain size is applied for four reasons: (i) ecosystem
boundaries are inherently vague (Regan et al., 2002), so it is easier to determine that
an ecosystem occurrence falls within a larger grid cell than a smaller one; (ii) larger
cells may be required to diagnose the presence of ecosystems characterised by
processes that operate over large spatial scales, or possess diagnostic features that
are sparse, cryptic, clustered or mobile (e.g. pelagic or artesian systems); (iii) larger
cells allow AOO estimation even when high resolution distribution data are limited;
and (iv) simulation studies have indicated that larger cells better predict risk in the
face of real-world threat events than finer scale cells (Keith et al. in review). A global
10×10 km gridded dataset suitable for this purpose is available via a public data
repository in raster and vector formats (Murray, 2017). Some ecosystem distributions
comprise a highly skewed distribution of patch sizes. In these cases large numbers of
small patches contribute a negligible risk-spreading effect to that of larger patches
and a correction may be applied by excluding from the AOO those grid cells that
contain patches of the ecosystem type that account for less than 1% of the grid cell
area (i.e. <1km2 of the focal ecosystem type, Box 10). Research is in progress to
support guidance on when to apply this correction.

Several spatial tools are in development to assist in measuring the EOO and AOO of an
ecosystem type. These will become available on the IUCN Red List of Ecosystems website
(www.iucnrle.org).

Assessing subcriteria B1 and B2

To be eligible for listing under subcriteria B1 or B2, an ecosystem must meet the EOO or
AOO thresholds that delineate threat categories, as well as at least one of three subcriteria
that address various forms of decline. These subcriteria distinguish restricted ecosystems at
appreciable risk of collapse from those that persist over long time scales within small stable
ranges (Keith et al., 2013). Only qualitative evidence of continuing decline is required to
invoke the subcriteria, but relatively high standards of evidence should be applied.

54 | IUCN Red List of Ecosystems

Subcriteria B1a and B2a address continuing declines in ecosystem distribution, abiotic
environment or biotic processes. To invoke this subcriterion, the declines must: (i) reduce
the ability of an ecosystem to sustain its characteristic native biota; (ii) be non-trivial in
magnitude; and (iii) be more likely than not to continue into the future (Table 3). Episodic or
intermittent declines qualify as continuing, so long as they are recurring and uncompensated
by increases of comparable magnitude. Downward phases of cyclical changes or
fluctuations do not qualify as continuing declines. These requirements imply an
understanding of the causes of decline to support a correct inference.

Subcriteria B1b and B2b do not require evidence of past or current declines, but may be
invoked by future declines inferred from serious and imminent threats. For these subcriteria,
assessors, must: (i) identify one or more specific threatening processes; (ii) present
convincing and generally agreed evidence that such threats are very likely (Table 3) to
cause continuing declines within the next two decades. These requirements imply an
understanding of how the threats affect the defining features of the ecosystem and the timing
of their effects. Speculation about generic threats with uncertain impacts or onset is
discouraged. Relevant evidence includes observations of similar threats in the past or on
similar ecosystems, as well as accumulated knowledge about the behaviour and nature of
the threat itself.

Subcriteria B1c and B2c require an estimate of the number of threat-defined locations that
are occupied relative to the extent of serious plausible threats. A threat-defined location is
defined as a geographically or ecologically distinct area in which a single threatening event
can rapidly affect all occurrences of an ecosystem type. Note that in the context of RLE
assessment, a threat-defined location is not necessarily the same as a locality or site of
occurrence; rather, a threat-defined location is defined entirely by the spatial extent of the
most serious plausible threats (this is consistent with the definition of locations for The IUCN
Red List of Threatened Species). The size of the threat-defined location depends on the
area covered by the threatening event and may include part of one or many separate
patches of the ecosystem. Where an ecosystem type is affected by more than one
threatening event, threat-defined locations should be defined by considering the most
serious plausible threat (IUCN, 2012). Where an ecosystem type is not affected by any
threatening events, the number of threat-defined locations cannot be estimated and the
subcriteria that refer to the number of locations will not be met. Box 11 contains further
guidance and examples to support the interpretation of the threat-defined location concept.

Assessing subcriterion B3

Subcriterion B3 requires only qualitative information on the distribution of an ecosystem and

threats to its persistence. To compensate for this type of evidence (cf. quantitative estimates
in other criteria), a higher standard of qualitative evidence is required and the highest
category that can be invoked by subcriterion B3 is Vulnerable. Subcriterion B3 comprises
two parts which must both be met for an ecosystem type to qualify for Vulnerable status.
First, the ecosystem type must have a very restricted distribution, generally with fewer than
five threat-defined locations (Box 11). Second, the ecosystem type must be facing severe
threats (human activities or stochastic events) within a very short time period in an uncertain
future and thus capable of collapse or becoming Critically Endangered within a very short

55 | IUCN Red List of Ecosystems

time period. In other words, the impact of the threat is very likely (Table 3) to occur in the
near future and its consequences are severe. Assessors have some flexibility to interpret the
‘very short time period’, but this generally means within the next two decades.

Documentation

For each assessment of an ecosystem type, assessors should: (i) provide the current maps
of ecosystem distributions (similar to those in Box 10) that were used to estimate the EOO
and AOO and to determine the number of threat-defined locations; (ii) provide full
bibliographic references; (iii) justify why the spatial data used is an adequate representation
of distribution of the focal ecosystem type (if not already done so for criterion A); (iv) explain
why a correction to AOO was justified if one was applied; (v) justify inferences about
continuing declines, and threats that may lead to continuing declines within the next 20
years; (vi) justify estimates of the number of threat-defined locations through reference to the
most serious plausible threats and their spatial characteristics (Box 11). As with
assessments under criterion A, description of the source of the spatial data (such as satellite
sensor type), the accuracy of all mapped data, and the spatial resolution (grain size) of all
data used in an assessment is strongly encouraged. Deposition of spatial data used for AOO
and EOO into an appropriate data repository is encouraged and should be referenced in the
documentation supporting the assessment.

56 | IUCN Red List of Ecosystems

Box 10. The extent of occurrence (EOO) and area of occupancy (AOO) of an ecosystem
(criterion B)

The distribution of the Great Fish Thicket, South

Africa (Mucina & Rutherford, 2006), is depicted
by a raster dataset with a spatial resolution of
30x30 m (A). As mapped, the area of the Great
2
Fish Thicket ecosystem type is 6,763.4 km .

A minimum convex polygon—the smallest

polygon that encompasses all known
occurrences of the ecosystem type in which no
internal angle exceeds 180 degrees—is applied
to estimate the extent of occurrence (EOO) for
assessment under criterion B1 (B). The area of
2
the minimum convex polygon is 18,359.2 km ,
meeting the initial requirements for an
Endangered classification under criterion B1.

To estimate the area of occupancy (AOO) for

assessment under criterion B2, the number of
cells covered by the ecosystem type is required
(C). The standardised measurement of AOO
ensures that distribution data mapped at varying
resolutions is generalised to a common 10x10
km grid, allowing consistent comparisons across
ecosystem types. First, a 10x10 km grid is
applied to the ecosystem type, indicating that
155 10x10 km grid cells intersect the distribution
map (shown in orange and grey). Second, when
the number of cells that contain very small
2
patches (<1km ) that negligibly contribute to risk
spreading are excluded (shown in grey), the
AOO is measured as 145 grid cells (shown in
orange). This AOO is greater than the
thresholds for classification in a threatened
category under B2.

Finally, to be eligible for listing in a threatened

category under criterion B, qualitative evidence
of continuing decline is also required. In this
case, the Great Fish Thicket ecosystem type
does not meet any of the additional subcriteria,
and is thus assigned an overall classification of
Least Concern.

57 | IUCN Red List of Ecosystems

Box 11. Determining the number of threat-defined locations (criterion B) (adapted from Appendix
S2 in Keith et al., 2013)

Coolibah - Black Box Woodland of south-eastern Australia

In its mature state, Coolibah – Black Box Woodland has an open structure with widely scattered
trees, a variable cover of shrubs and grassy ground layer. The characteristic vertebrate fauna
includes diverse assemblages of woodland and wetland bird species, many of which depend on
tree hollows, other features of large trees or standing water for breeding and foraging (NSW
Scientific Committee, 2004). The most serious plausible threats are land clearing and changes to
water regimes. Spatial patterns of land clearing show a high degree of contagion, with the best
predictor of future clearing being the proximity of a patch to land parcels already cleared of native
vegetation. A broad interpretation of threat-defined locations under subcriterion B3 identifies three
jurisdictional zones with different regulatory controls on land clearing: the leasehold Western
Division of New South Wales; the freehold Central Division of New South Wales; and Queensland;
this results in an estimate of three threat-defined locations as defined by land clearing. A more
narrow interpretation of threat-defined locations based on neighbourhoods of contagion would
produce an estimate of more than five. Small protected areas are excluded from these threat-
defined locations, as they are not threatened by land clearing. These areas were assessed by
considering the next most serious plausible threat: changes to water regimes. As protected areas
are located in at least two different sub-catchments with different water management
infrastructure, there are at least two further threat-defined locations. Hence the most precautionary
interpretation produces an estimate of five threat-defined locations, although it is likely that there
are more. Based on current rates of depletion due to land clearing (subcriterion A1) and current
rates of environmental degradation due to changes in water regime (subcriterion C1), the
ecosystem is unlikely to collapse or become Critically Endangered within the near future (c. 20
years). The ecosystem type therefore does not meet subcriterion B3, so the status of the
ecosystem type is Least Concern under this subcriterion.

Cape Flats Sand Fynbos of South Africa

Cape Flats Sand Fynbos is a species-rich, dense, moderately tall shrubland with scattered
emergent shrubs (Rebelo et al., 2006). The ecosystem type is an edaphically determined species
assemblage restricted to Tertiary acid, deep grey regic sands at low elevations (20–200 m) on flat
to undulating terrain. Cape Flats Sand Fynbos is restricted to the Western Cape province of South
Africa, almost entirely within the limits of the City of Cape Town. The most severe threat to the
ecosystem type is habitat destruction associated with urban development (Rebelo et al., 2006;
Wood et al., 1994). Occurrences that are currently within proclaimed reserves are protected from
this threat, although these stands are threatened by invasion of exotic plants (Rebelo et al., 2006).
As the entire distribution of the ecosystem type is within the City of Cape Town, the unproclaimed
remnant vegetation is subject to the same development pressures, regulatory regimes and
planning authority. The distribution is therefore interpreted as two semi-independent threat-defined
locations; one outside protected areas (threatened by habitat destruction and invasive plants) and
one within protected areas (threatened by invasive plants, but not habitat destruction). Given the
severe and immediate nature of the threats, the ecosystem type is prone to the effects of human
activity or stochastic events such that it is capable of collapse or becoming Critically Endangered
within a very short time period. The status of the ecosystem type is thus Vulnerable under
subcriterion B3.

58 | IUCN Red List of Ecosystems

5.3 Criterion C. Environmental degradation

5.3.1 Theory

The RLE risk model defines two criteria for assessing declines in ecosystem functions or
processes. Two criteria are needed to assess abiotic (environmental) and biotic degradation
because the causes, effects and mechanisms of functional decline differ fundamentally
between them (Keith et al., 2013). Abiotic degradation is the deterioration of the physical,
non-living attributes that have a defining role in ecological processes and/or the distribution
of an ecosystem type. Abiotic degradation reduces the capacity of an ecosystem to sustain
its characteristic biota. For example, declines in limiting resources (niche dimension) reduce
species diversity in a range of terrestrial, freshwater and marine ecosystems (Harpole &
Tilman, 2007).

5.3.2 Thresholds and subcriteria

An ecosystem may be listed under criterion C if it meets the thresholds for any of four
subcriteria (C1, C2a, C2b, or C3), which express different levels of environmental
degradation over the following time frames:

Subcriterion Time frame Relative severity (%)

Extent ≥ 80 ≥ 50 ≥ 30
The past 50 years based on change in an
(%)
abiotic variable affecting a fraction of the
C1 ≥ 80 CR EN VU
extent of the ecosystem and with relative
≥ 50 EN VU
severity, as indicated by the following table:
≥ 30 VU
C2a. The next 50 years, based on change in Extent ≥ 80 ≥ 50 ≥ 30
an abiotic variable affecting a fraction of the (%)
extent of the ecosystem and with relative ≥ 80 CR EN VU
severity, as indicated by the following table; ≥ 50 EN VU
OR ≥ 30 VU
C2
C2b. Any 50-year period including the past,
present and future, based on change in an
abiotic variable affecting a fraction of the
extent of the ecosystem and with relative
severity, as indicated by the following table:
Extent ≥ 90 ≥ 70 ≥ 50
Since 1750 based on change in an abiotic
(%)
variable affecting a fraction of the extent of the
C3 ≥ 90 CR EN VU
ecosystem and with relative severity, as
≥ 70 EN VU
indicated by the following table:
≥ 50 VU

5.3.3 Application

Data requirements

To assess criterion C it is necessary to select suitable abiotic variables that represent

defining features of the ecosystem type. This choice is based on a number of considerations:

59 | IUCN Red List of Ecosystems

1. There must be plausible evidence of a causal relationship between the process of
environmental degradation and the loss of characteristic native biota. For example,
an assessment of wetland degradation based on change in water quality would
require evidence that declines in water quality are associated with loss of wetland
biota, either through direct observation or inference from comparable ecosystem
types. A carefully developed conceptual model can highlight key environmental
processes and the transitions between healthy and collapsed states to be assessed
in criterion C (Section 4.2.3).

2. Assessing abiotic degradation requires suitable scalar variables for estimating the
severity of degradation, as well as suitable spatial variables for estimating the extent
of degradation. The characteristics of the ecosystem and its threats will determine
which variables are relevant. Variables with direct and clear cause-effect
relationships and the greatest sensitivity to loss of characteristic native biota will be
the most suitable.

3. The choice of a variable representing abiotic degradation should enable direct

inferences about threshold values of the variable that define ecosystem collapse (Fig.
3).

4. Applying generic indices across functionally contrasting ecosystems is unlikely to

assess degradation accurately if key processes differ among these ecosystems. The
choice of abiotic variable should be underpinned by the ecology of a particular
ecosystem (Table 7). If alternative variables representing different degradation
processes are available, they should be examined independently and the one
producing the greatest rate of decline should be used to assess status or the
outcomes should contribute to a bounded estimate of the status.

5. Aggregation of multiple variables into a single index for assessment under criterion C
can be problematic and is discouraged. Aggregation relies on statistical assumptions
which may be unwarranted, especially in data-poor ecosystems. Aggregation can
also confound different mechanisms of environmental degradation, making the index
less sensitive than individual variables due to averaging effects. Assessors should
therefore avoid aggregating variables when they are uncertain about ecosystem
dynamics and the assumptions underpinning the aggregation. In these cases, the
best effort should be made to select a variable that is relevant to ecosystem
processes and sensitive to environmental degradation. The use of aggregated
indices should be supported by critical evaluation of ecological and mathematical
assumptions. A clear link between the change in index value and proximity to
collapse must be demonstrated.

6. If the interaction between two or more variables is considered important, it is

preferable to develop robust, expert-based rules to define states that are a
combination of the variables, rather than use an index. For example, a severely
degraded example of an ecosystem type might require two variables to have crossed
a given threshold or be between two stated values, while moderate degradation may
require either one to have crossed the threshold or both to be between a different set

60 | IUCN Red List of Ecosystems

 of values. Using a rule-based method requires the assessor to explicitly state and
understand how the variables can be combined. In many indices these relationships
are submerged, poorly understood, and have unintended effects on index values.

7. Estimating the extent of abiotic degradation can be based on expert-derived

estimates, inferences or spatial data. For example, data on levels of water extraction
and surface area for each wetland were combined to assess the relative severity of
environmental degradation over the entire area of the swamps, marshes and lakes of
the Murray-Darling Basin (Keith et al., 2013).

Table 7. Examples of variables potentially suitable for assessing the severity of environmental
degradation under criterion C.

Environmental degradation Variables

Desertification of Proportional cover of bare ground, soil density, soil compaction

rangelands indices, remote sensing indices of change (Zhao et al., 2005;
Ludwig et al., 2007).
Eutrophication of soils, Levels of dissolved or soil nitrogen, phosphorus, cations, oxygen,
freshwater streams or turbidity, bioassay (Carpenter, 2003).
lakes
Dehumidification of cloud Cloud cover, cloud altitude (Pounds et al., 1999).
forests
Deforestation by acid rain Rain water chemistry (Likens, 1992).
Homogenisation of Diversity of micro-terrain features, spatial variance in inundation
microhabitats depth and duration (Cabezas et al., 2009).
Changed water regime or Field-based monitoring of stream flow volume, or piezometric water
hydroperiod table depth; remote sensing of spatial extent of surface water,
frequency and depth of inundation (Mac Nally et al., 2011).
Salinisation of soils or Field monitoring of salinity of soils or groundwater, remote sensing
wetlands of ground surface albedo (Metternicht & Zinck, 2003).
Sedimentation of streams, Sediment accumulation rates, sediment load of streams, discharge,
coral reefs turbidity of water column, frequency and intensity of sediment plume
spectral signatures (Rogers, 1990).
Structural simplification of Microrelief, abundance of benthic debris, trawling frequency and
benthic marine ecosystems spatial pattern (Watling & Norse, 1998).
Sea level rise Acoustic monitoring of sea level, extent of tidal inundation (Hannah
& Bell, 2012).
Retreat of ice masses Remote sensing of sea ice extent (Hong & Shin, 2010).

Methods

The key concept for assessing functional declines in either abiotic or biotic variables is
relative severity. Relative severity is essential for comparing risks among ecosystems
undergoing different types of degradation. Relative severity describes the proportional

61 | IUCN Red List of Ecosystems

change observed in an environmental variable scaled between two values: one describing
the initial state of the system (0%), and one describing a collapsed state (100%). Thus, if an
ecosystem type undergoes degradation with a relative severity of 50% over an assessment
time frame, this implies that that it has transformed half way to a collapsed state. Information
on relative severity is combined with information on the proportion of the ecosystem affected
(extent) to determine the risk category under criterion C. Assessors may either estimate the
extent of degradation that exceeds a threshold level of severity or estimate the average
severity of degradation across the entire ecosystem distribution (100% of extent; Fig. 13).

Ecosystems are listed as CR if environmental change is both extremely severe (≥80%

relative severity) and extensive (across ≥80% of the distribution). Ecosystems may be
eligible for listing in lower threat categories if they are undergoing very severe but localised
degradation or less severe degradation over extensive areas (Fig. 13). Ecosystems that fail
to meet the thresholds for the Vulnerable category may be assigned to the Near Threatened
category. For example, an ecosystem undergoing >80% decline in environmental quality
over 20-30% of its distribution, or >30% decline over 70-80% of its distribution could qualify
as Near Threatened.

In the simplest case, relative severity may be calculated by range-standardising the raw
values of the abiotic variable between its initial value and its collapse value. Assessors must:
(i) estimate the value of the abiotic variable initial state (at the beginning of the assessment
time frame); (ii) estimate the expected value in a collapsed state; (iii) measure or estimate
the present or future value of the variable (i.e. at the end of the assessment time frame).
Note that the calculated relative severity can be negative if the condition of the ecosystem
has improved.

The following equations rescale an abiotic variable to a proportional change towards

collapse suitable for assessing criterion C:

Relative severity (%) = (Observed or predicted decline / Maximum decline) × 100

where
Observed or predicted decline = Initial value – Present or future value
and
Maximum decline = Initial value – Collapse value

Next, assessors determine the extent of the degradation as a proportion of the total
distribution of the ecosystem. With these two quantities assessors assign a risk category
using the described thresholds.

62 | IUCN Red List of Ecosystems

Figure 13. Contrasting pathways of environmental or biotic degradation and their
corresponding risk classifications under criteria C1, C2, D1, D2 (A) or C3, D3 (B): (a) initially
widespread and relatively benign degradation, which increases in severity, (b) severity and
extent of degradation increase at similar rates, (c) localised but severe degradation, later
becoming more widespread.

63 | IUCN Red List of Ecosystems

Box 12. Assessing environmental degradation (criterion C) (adapted from Appendix S2 in Keith et
al., 2013)

Flooding is a key ecological process that sustains the Gonakier Forests for the Senegal River
Floodplain in Senegal-Mauritania (Keith et al., 2013). As floods occur only during the wet season
months, the maximum annual river height was assumed to be indicative of the river’s capacity to
flood each year. River height data were available for 100 years from 1904 to 2003. To assess
criterion C, mean annual maximum river height across four gauging stations was used as a proxy
for environmental degradation. River flows declined sharply, reaching a minimum during the late
3
1970s and 1980s. Floods of 2,500 m /s, which are needed for floodplain inundation, would be very
unlikely to occur based on river flows observed during 1986-1989. Extreme rates of tree mortality
were observed between the mid-1970s and the mid-1980s, corresponding to the lowest maximum
river heights (473±27 cm) observed during the 100 years of records.

Based on these observations, the collapse threshold was defined as the mean maximum river
height for a 50-year period falling below 450-500 cm, causing extensive tree mortality. To
calculate the relative severity of hydrological decline, the time series was divided into the
reference period (1904-1953) and the present period (1954-2003). Since the collapse threshold is
an interval, relative severity was estimated for the lower and upper bounds of the interval.

For the lower bound (450 mm), relative severity is:

(Observed decline) / (Maximum decline) × 100 = (712-619) / (712-450) × 100 = 35%

For the higher bound (500 mm), relative severity is

(Observed decline / Maximum decline) × 100 = (712-619) / (712-500) × 100 = 44%

Figure 14. River height in the Gonakier forest.

Since hydrological decline affects the entire ecosystem, it was assumed that the extent of the
threat was >80%, thus leading to the conclusion that the ecosystem is Vulnerable according
criterion C1 (degradation with relative severity ≥30% over an extent ≥80% in the last 50 years).

64 | IUCN Red List of Ecosystems

Assumptions

Determining an initial and a collapsed value for the abiotic variable relies on assumptions
about collapsed states of the ecosystem type. Such uncertainty in the collapse point can be
represented with bounded thresholds of the values of the variable. The calculation of relative
severity can be repeated with both values, providing a lower and upper estimate for the risk
category (Box 12). Similarly, uncertainty in the extent of degradation can be assessed with
the use of upper and lower estimates. The use of bounded values yields an estimate of the
extent and severity of abiotic degradation while clearly expressing uncertainty.

Similar to the declines of extent required for assessing under criterion A, the application of
criterion C assumes a functional form of decline. The simplest case illustrated above applies
when there is a linear relationship between the assessment variable and the trajectory
towards a collapsed state. Other scenarios are possible, for example, where collapse
proceeds more slowly or more rapidly than indicated by changes in the assessment variable.
In such cases a suitable transformation of the assessment variable should be used in the
calculation of relative severity (Fig. 15).

Figure 15. An observed value for a variable assessing degradation can be mapped to
different values of relative severity depending on the functional form considered. The red line
indicates an observed value which can be mapped to a relative severity of 20%, 60%, or 90%
depending on the functional form. This corresponds to a risk category of LC, EN, or CR if the
degradation occurs over ≥80% of the ecosystem type.

Determining whether the degradation is constant, accelerating, or decelerating can be

informed by time-series data. Assessors should evaluate whether the available data are
sufficiently representative to characterise the shape of the decline in the abiotic variable,
ideally through appropriate statistical methods (Di Fonzo et al., 2013; Connors et al., 2014).
Where time-series data are unavailable, it may be possible to infer changes in degradation

65 | IUCN Red List of Ecosystems

using expert elicitation or space-for-time substitution with appropriate reference sites
(Pickett, 1989). To overcome uncertainty due to this assumption, sensitivity analyses that
include estimates produced from multiple shapes of decline can provide a bounded estimate
for the risk assessment outcome.

Documentation

Assessors should document: (i) the selection of the abiotic variable with respect to the
conceptual model of ecosystem dynamics; (ii) the setting of a bounded collapse threshold for
the abiotic variable; (iii) the calculation of relative severity; (iv) the estimation of the extent of
degradation; (v) assumptions and appropriate sensitivity analyses (e.g. regarding the
collapse definition or shape of decline); (vi) the final risk categories and plausible bounds.
Temporal variation in degradation is best shown in a graph that depicts changes in the
variable over time, and includes any interpolation or extrapolation to match the relevant time
frame.

66 | IUCN Red List of Ecosystems

5.4 Criterion D. Disruption of biotic processes and interactions

5.4.1 Theory

The persistence of biota within ecosystems depends on biotic processes and interactions.
This includes: competitive, predatory, facilitatory, mutualistic, trophic and pathogenic
processes; mobile links (e.g. seasonal migration); and species invasions. Biodiversity loss
reduces the capacity of ecosystems to capture resources, produce biomass, decompose
organic matter and recycle carbon, water and nutrients, and also reduces the stability of
these functions through time (Cardinale et al., 2012). The identity of organisms within a
system controls its functioning as key taxa make disproportionate contributions to ecosystem
functions. The diversity of organisms is also important, because niche partitioning and
positive species interactions promote complementary contributions to ecosystem functions.

Feedback interactions are crucial for an ecosystem type to absorb environmental change
while maintaining characteristic biota and processes. Conversely, significant disruptions to
biotic processes and interactions can cause collapse, regime shifts and re-organisation into
novel ecosystems (Thébault & Loreau, 2005). Disruption of interactions through trophic
cascades is one of five major threats to biodiversity (Diamond, 1989), although non-trophic
interactions also play important roles (Fontaine et al., 2005; Goudard & Loreau, 2008).
Certain ecosystem types may be especially sensitive to disruption of biotic processes and
interactions, such as systems with strong top-down trophic regulation, with many mutualistic
or facilitation interactions that are strongly dependent on mobile links, and where positive
feedbacks operate between the biota and disturbance regimes.

5.4.2 Thresholds and subcriteria

An ecosystem may be listed under criterion D if it meets the thresholds for any of four
subcriteria (D1, D2a, D2b, or D3), which express different levels of biotic disruption over the
following time frames:

67 | IUCN Red List of Ecosystems

 Subcriterion Time frame Relative severity (%)
Extent ≥ 80 ≥ 50 ≥ 30
The past 50 years based on change in a biotic
(%)
variable affecting a fraction of the extent of the
D1 ≥ 80 CR EN VU
ecosystem and with relative severity, as
≥ 50 EN VU
indicated by the following table:
≥ 30 VU
D2a. The next 50 years, based on change in a Extent ≥ 80 ≥ 50 ≥ 30
biotic variable affecting a fraction of the extent (%)
of the ecosystem and with relative severity, as ≥ 80 CR EN VU
indicated by the following table; OR ≥ 50 EN VU
D2 ≥ 30 VU
D2b. Any 50-year period including the past,
present and future, based on change in a biotic
variable affecting a fraction of the extent of the
ecosystem and with relative severity, as
indicated by the following table:
Extent ≥ 90 ≥ 70 ≥ 50
Since 1750 based on change in a biotic
(%)
variable affecting a fraction of the extent of the
D3 ≥ 90 CR EN VU
ecosystem and with relative severity, as
≥ 70 EN VU
indicated by the following table:
≥ 50 VU

5.4.3 Application

Data requirements

Assessment of criterion D addresses the same data requirements as criterion C, except

suitable biotic variables are used. Conceptual models provide a useful framework for
selecting biotic variables linked to key ecosystem processes. A broad set of variables are
potentially useful for quantifying biotic processes and associated functional declines. This
includes changes in species richness, composition and dominance; relative abundance of
species functional types, guilds or alien species; measures of interaction diversity; changes
in identity and frequency of species movements; measures of niche diversity and structural
complexity (Table 8).

Table 8. Examples of biotic variables potentially suitable for assessing the severity of disruption to
biotic interactions under criterion D.

Variable Role in ecosystem resilience and Example

function
Species richness (the Ecological processes decline at Response of species diversity
number of species within a an accelerating rate with loss of of grasses and relative
taxonomic group per unit species (Cardinale et al., 2011). abundance to varying levels of
area). Species richness is related grazing in grassland (Walker et
indirectly to ecosystem function al., 1999).
and resilience through its
correlations with functional
diversity, redundancy and
complementarity.

68 | IUCN Red List of Ecosystems

Variable Role in ecosystem resilience and Example
function
Species composition and Shifts in dominance and Shift in diet of top predators
dominance. community structure are (killer whales) due to
symptoms of change in overfishing effects on seals,
ecosystem behaviour and caused decline of sea otters,
identity. reduced predation of kelp-
feeding urchins, causing their
populations to explode with
consequent collapse of giant
kelp, structural dominants of
the benthos (Estes et al.,
2009) (Box 13).
Abundance of key species Invasions of certain alien species Invasion of crazy ants
(ecosystem engineers, may alter ecosystem behaviour simplifies forest structure,
keystone predators and and identity, and make habitat reduces faunal diversity and
herbivores, dominant unsuitable for persistence of native ecosystem engineers
competitors, structural some native biota. Transformer (Green et al., 2011). Invasion
dominants, transformer alien species are distinguished of arid Australian shrublands
invasive species). from benign invasions that do not and grasslands by Buffel
greatly influence ecosystem Grass makes them more fire
function and dynamics prone and less favourable for
persistence of native plant
species (Clarke et al., 2005;
Miller et al., 2010).
Functional diversity High diversity of species High diversity of plant-derived
(number and evenness of functional types (e.g. resource resources sustains
types). use types, disturbance response composition, diversity and
types) promotes coexistence function of soil biota
through resource partitioning, (Eisenhauer et al., 2011). Fire
niche diversification and regimes promote coexistence
mutualisms (Allen et al., 2005). of multiple plant functional
Mechanisms similar to functional types (Keith et al., 2007).
complementarity.
Functional redundancy Functionally equivalent minor Response of bird communities
(number of taxa per type; species may substitute for loss or to varying levels of land use
within- and cross-scale decline of dominants if many intensity (Fischer et al., 2007).
redundancy; see Allen et species perform similar functional
al., 2005). roles (functional redundancy).
Low species richness may be
associated with low resilience and
high risks to ecosystem function
under environmental change
(Allen et al., 2005; Walker et al.,
1999).
Functional complementarity Functional complementarity High functional
(dissimilarity between types between species (e.g. in resource complementarity within both
or species). use, body size, stature, trophic plant and pollinator
status, phenology) enhances assemblages promotes
coexistence through niche recruitment of more diverse
partitioning and maintenance of plant communities (Fontaine et
ecosystem processes (Cardinale al., 2005).
et al., 2007).
Interaction diversity Interactions shape the Overgrazing reduced diversity
(interaction frequencies and organisation of ecosystems, of pollination interactions
dominance, properties of mediate evolution and (Vázquez & Simberloff, 2003).
network matrices). persistence of participating
species and influence ecosystem-
level functions, e.g. productivity
(Thompson, 1997).

69 | IUCN Red List of Ecosystems

 Variable Role in ecosystem resilience and Example
function
Trophic diversity (number of Compensatory effects of Diverse carnivore
trophic levels, interactions predation and resource assemblages (i.e. varied
within levels, food web competition maintain coexistence behaviour traits and densities)
structure). of inferior competitors and prey. promote coexistence of plant
Loss or reduction of some species (Calcagno et al.,
interactions (e.g. by 2011), decline of primary prey
overexploitation of top predators) precipitates diet shifts and
may precipitate trophic cascades phase shifts (Springer et al.,
via competitive elimination or 2003).
overabundance of generalist
predators.
Spatial flux of organisms Spatial exchanges among local Herbivorous fish and
(rate, timing, frequency and systems in heterogeneous invertebrates migrate into reefs
duration of species landscapes provide spatial from seagrass beds and
movements between insurance for ecosystem function mangroves, reducing algal
ecosystems). (Loreau et al., 2003). Exchanges abundance on reefs and
may involve resources, genes or maintaining suitable substrates
involvement in processes for larval establishment of
(Lundberg & Moberg, 2003). corals after disturbance
(Moberg & Folke, 1999).
Structural complexity (e.g. Simplified architecture reduces Structurally complex coral
complexity indices, number niche diversity, providing suitable reefs support greater fish
and cover of vertical strata habitats for fewer species, greater diversity (Arias-González et
in forests, reefs, remote exposure to predators or greater al., 2012), structurally complex
sensing indices). competition for resources (due to woodlands support greater bird
reduced partitioning). diversity (Huth & Possingham,
2011).

Methods

The evaluation of criterion D follows the same procedure as with criterion C, but focuses on
biotic variables rather than abiotic variables. Again, relative severity is calculated by range-
standardising the raw values of the biotic variable between its initial value and its collapse
value (Section 2). Assessors must: (i) estimate the value of the biotic variable in an initial
state; (ii) estimate the expected value in a collapsed state; (iii) measure or estimate the
present or future value of the variable. These three quantities are then used to rescale the
biotic variable to a proportional change towards collapse (Section 5.3.3). Next, assessors
must determine the extent of the disruption as a proportion of the total distribution of the
ecosystem. With these two quantities assessors proceed to assign a risk category using the
described thresholds. Similarly to criterion C, the use of generic indices is discouraged
(Section 5.3.3).

Assumptions

Assumptions in the application of criterion D are similar to assumptions for criterion C. Again,
determining an initial and a collapsed value for the biotic variable relies on assumptions
about initial and collapsed states of the ecosystem (Section 2.2). As in the application of
criterion A and C, application of criterion D relies on an assumption about the functional form
of decline. In all cases, the decisions made in relation to assumptions must be explicitly
discussed in the documentation.

70 | IUCN Red List of Ecosystems

Documentation

Assessors should document: (i) the selection of the biotic variable with respect to the
conceptual model of ecosystem dynamics; (ii) the setting of a bounded collapse threshold for
the biotic variable; (iii) the calculation of relative severity; (iv) the estimation of the extent of
disruption; (v) assumptions and appropriate sensitivity analyses (regarding the definition of
collapse or the assumed shape of decline); (vi) the final risk categories and plausible
bounds. Temporal variation in biotic disruption is best shown in a graph depicting changes in
the variable over time, and includes any interpolation or extrapolation to match the relevant
time frame (Box 13).

71 | IUCN Red List of Ecosystems

Box 13. Assessing disruption of biotic processes (criterion D)

Alaskan Giant Kelp Forests are structurally and functionally diverse assemblages, characterised
by species of brown algae in the Order Laminariales. These create complex and dynamic layered
forest architecture up to 15 m tall that provides substrate, shelter and foraging resources for a
diverse fauna assemblage of epibenthic invertebrate herbivores and pelagic vertebrate predators.

The most serious disruption to biotic interactions occurs through trophic cascades involving sea
otters, their predators (killer whales) and their prey (urchins, which consume kelp). Given that
densities of kelp are inversely related to densities of urchins, and that phase shifts between forests
and urchin barrens are related to a threshold abundance of otters (Estes et al., 2010), any of these
variables is potentially suitable for assessing criterion D. Although data are available on population
changes in great whales and pinnipeds (alternative prey for killer whales), these were not used
because: (i) data on more proximal response variables are available; (ii) the causal relationship
linking great whales and pinnipeds with otter abundance via killer whale predation is less certain
than the link between otters, urchins and kelp.

Survey data for kelp stipe densities were available between 1987 and 2000 from seven islands
(Estes et al., 2009). It was assumed that the seven islands, scattered across the Aleutian chain,
were representative of the full distribution of the ecosystem. Ecosystem collapse occurs when kelp
density is close to zero across all sites, consistent with kelp replacement by urchin barrens
throughout the distribution. Rates of change in kelp density were calculated for each island
assuming an exponential model. A weighted average across all sites indicated that kelp densities
declined on average by 49.2% between 1987 and 2000. Allowing for some decline prior to 1987 or
after 2000 suggests that the decline in kelp density over the past 50 years was at least 50%
across the full ecosystem extent.

Aerial survey data for sea otters were available for 55 islands along the Aleutian chain between
1959 and 2000 (Doroff et al., 2003). Ecosystem collapse occurs when otter populations reach zero
across all sites. The total population was estimated to be 55,000–74,000 prior to decline in the
mid-1980s. By 2000 there were a total of 3,924–13,580 animals based on extrapolation from the
aerial survey (Doroff et al., 2003). The lower and upper bounds of otter population decline are:

100 × (55000 - 13580) / 55000 = 75.3%

and 100 × (74000 - 3924) / 74000 = 94.7%

Evidence from trends in kelp density and sea otter sightings suggest a decline in biotic function of
50-95% relative severity across 100% of the ecosystem extent. The upper bound of this range
may overestimate the severity of decline because: (i) the surveys may have underestimated the
population due to detectability issues (Doroff et al., 2003); (ii) the calculations assume that otter
and kelp populations have not recovered since 2000, in spite of qualitative evidence for some
recovery. The most likely status of the ecosystem under criterion D1 is Endangered, although a
status of Critically Endangered is possible. No projections are currently available for any of the
biotic variables. The status of the ecosystem is Data Deficient under criterion D2.

The otter population in 1750 was comparable or slightly larger than its peak in the mid-1980s
(Doroff et al., 2003). Based on this assumption, the decline in otter populations throughout the
distribution of the kelp forest was 75-95% since 1750. The status of the ecosystem type under
criterion D3 is therefore Endangered (plausible range Endangered – Critically Endangered). Thus,
the Alaskan giant kelp forests ecosystem type is listed as Endangered (plausible range
Endangered – Critically Endangered).

72 | IUCN Red List of Ecosystems

5.5 Criterion E. Quantitative risk analysis

5.5.1 Theory

Criterion E serves two purposes. First it can be used to list an ecosystem type by
implementing models that integrate multiple mechanisms of decline and their interactions
into the risk assessment (as described below). Second, it provides an anchor for risk
assessment and an overarching framework for the other criteria, as its analogue does in Red
List criteria for species. Criterion E specifies the level of risk that corresponds to each
category of threat, by defining the probability of collapse and the specified time frame for
Critically Endangered (CR), Endangered (EN) and Vulnerable (VU) ecosystem types.

5.5.2 Thresholds

An ecosystem may be listed under Criterion E if it meets the thresholds for the criterion, a
quantitative analysis that estimates the probability of ecosystem collapse to be:

CR ≥ 50% within 50 years

EN ≥ 20% within 50 years
VU ≥ 10% within 100 years

5.5.3 Application

Methods

The probability of ecosystem collapse can be estimated with stochastic simulation models
incorporating key ecosystem processes. The models should:

1. Produce estimates of an ecosystem variable for which a threshold of collapse has

been estimated.
2. Produce quantitative estimates of risks of ecosystem collapse over a 50-100 year
time frame.
3. Incorporate stochasticity in key processes that determine ecosystem properties.
4. Be applied with scenarios that represent plausible future scenarios of ecosystem
dynamics.

A wide range of models can be used to apply criterion E. We provide broad

recommendations for the application of criterion E in the form of nine steps to ensure that
models are based on sound assumptions, scientifically credible and transparent (Fig. 16).

73 | IUCN Red List of Ecosystems

 Figure 16. Nine steps to apply criterion E.

1. Define model purpose

Models for criterion E should provide an adequate prediction of the risk of collapse
over a period of 50-100 years. If the model used for criterion E is being adapted from
a model with different objectives (e.g. providing guidance for management and
decision-making), it may be necessary to modify its objectives and implementation.
Although other objectives may be important in model-building, models for criterion E
will be reviewed based on the quality of their predictions for the purpose of the RLE
assessment.

2. Specify modelling context

Ecosystems are inherently scale-dependent, so the thematic, spatial and temporal
scales of ecosystem processes may affect model-building and predictions.
Adequately defining the boundaries of the ecosystem under assessment is crucial –
external forcing and external outputs should be clearly labelled as such. The model
should aim to spatially represent all occurrences of the ecosystem; if not, adequate
inferences should be made to assess the representativeness of final predictions. The
time frame of predictions for criterion E is 50-100 years, which is longer than other
subcriteria (A2a, A2b, C2a, C2b, D2a, D2b) and may therefore require a different
understanding of future threats.

74 | IUCN Red List of Ecosystems

 Modelling may involve decisions relying on the risk attitude of the assessor, i.e. the
relative costs of under or overestimating the risk of collapse. A precautionary but
realistic risk attitude is advocated when implementing criterion E. Such decisions
should be thoroughly documented within the criterion E documentation, and if
possible underpinned by quantitative measures of risk aversion.

3. Conceptualise system
Models for criterion E should rely on a sound understanding of ecosystem dynamics
and function, underpinned by data and relevant inferences from similar ecosystems.
Conceptual models can help identify key ecosystem processes and variables
indicating collapse. The conceptual model may depict cause-and-effect relationships
or transitions among reference and collapsed ecosystem states. The conceptual
model used for criterion E may differ from the general conceptual model used in the
ecosystem description (Section 6.2.3), as it may depict more complex relationships
and include measurable variables. Deciding on an appropriate level of abstraction for
key processes is a key component of conceptualisation and should consider the
model purpose, context, required resolution of output and effort required for model
building. A critical component of assessment under criterion E is the explicit definition
of collapse as it relates to the conceptual model of ecosystem dynamics and
measured variables (Section 3.2).

4. Specify data and prior knowledge

Applying criterion E requires the levels of key ecosystem variables to be predicted
over specified time frames. These variables can represent spatial distribution (as in
criteria A and B), abiotic environment (criterion C), and/or biotic interactions (criterion
D). Suitable variables can be selected by following the processes outlined in the
Application sections relevant to each criterion. The data may be quantitative
measurements (e.g. spatial data, time series) or expert-derived. At this stage the
degree of spatial and temporal aggregation of data and predictions may be revised,
to match ecosystem dynamics to the modelling context. For example, it may be
appropriate to aggregate daily or monthly data to yearly time steps. In data-poor
situations, it may be possible to infer processes and data from similar ecosystems
(Maxwell et al., 2015). This should be clearly indicated and discussed within the
model documentation.

5. Select model type

A diverse range of simulation models of ecosystem dynamics allow the probability of
ecosystem collapse to be estimated directly. Selection of an appropriate model type
will depend on: (i) ecosystem dynamics; (ii) data availability; (iii) representation of
uncertainty; and (iv) integration of stochasticity. Some models may be more
appropriate to represent specific ecosystems and their dynamics (e.g. hydrologic
models for wetlands, global vegetation models for forests). The type of input data
may also constrain model choice (e.g. some model types may be unable to handle
missing data or expert-derived data). Models should be chosen or adapted so that
appropriate uncertainty and sensitivity analyses can be conducted. Ideally, model
uncertainty should be addressed by implementing multiple models representing
alternative interpretations of ecosystem dynamics. Finally, ecosystem dynamics rely

75 | IUCN Red List of Ecosystems

 on stochastic processes, so models should be chosen or adapted so as to integrate
stochasticity (see Coorong Lagoon case study in Appendix S2 in Keith et al., 2013).

Candidate model types for the application of criterion E include:

• state-and-transition models (Lester and Fairweather, 2009, Rumpff et al.,

2011, Maxwell et al., 2015).
• mass-balance models (e.g. Ecopath, Models of Intermediate Complexity)
(Christensen and Walters, 2004, Plagányi et al., 2014).
• bifurcation plots (Holdo et al., 2013).
• network theory (e.g. Community Viability Analysis) (de Visser et al., 2011).
• dynamic Global Vegetation Models (Scholze et al., 2006).
• dynamic species distribution and population models (Midgley et al., 2010,
Keith et al., 2008).
• spatial models (e.g. cellular automata) (Soares-Filho et al., 2002).
• general ecosystem models (e.g. the Madingley model) (Harfoot et al.,
2014).

6. Model training, parameterisation, validation

Models should follow best practice recommendations for each model type, and
should be appropriately trained, parameterised and/or validated. For example, the
data-derived state-and-transition model of the Coorong Lagoon was validated
through multiple pathways, so that neither states nor transitions were determined a
priori (Lester & Fairweather, 2011). For some models full validation may not be
possible; in these cases model performance can be evaluated with relevant
performance indicators, e.g. satisfactory reproduction of observed behaviour,
absence of correlation in model residuals (Jakeman et al., 2006). Model training,
parameterisation and validation may occur in iterative steps that should be
thoroughly documented. It may be appropriate to assess the effects of data
uncertainty, parameter uncertainty and model uncertainty through sensitivity
analyses. Overall, assessors should demonstrate that the model is fit for purpose for
application in criterion E.

7. Scenarios
Future scenarios representing likely threats and changes to ecosystem dynamics
should be identified. It is important to recognise that concepts and data underpinning
scenarios may be subject to high levels of uncertainty, the effects of which may be
difficult to track in large models (e.g. climate change projections; Kujala et al., 2013).
Often, the relative likelihood of each future scenario will not be known (Peterson et
al., 2003), so the final likelihood of collapse may be expressed as a range of values
rather than a single estimate.

8. Probability of collapse
The estimate of the probability of collapse may be a single value, but in most cases
in may be expressed as a range of values representing uncertainty in model-building.
Sensitivity analyses of the probability of collapse may be done relevant to: (i) data,
model and parameters uncertainty; (ii) scenario uncertainty; and (iii) other forms of

76 | IUCN Red List of Ecosystems

 uncertainty that may affect modelling outcomes, e.g. the choice of variables to
assess ecosystem collapse. A sensitivity analysis on the threshold of collapse should
be conducted in all models, as the final outcome for criterion E may be particularly
sensitive to the definition of collapse. In simulations of the Mountain Ash forest
(Burns et al., 2015), for example, the collapse threshold would need to decrease from
an average of one hollow-bearing tree per hectare to 0.7 to change the risk
assessment outcome.

9. Interpretation
Criterion E provides an overarching framework for the application of the other criteria,
and includes ecosystem dynamics that may not be captured by other criteria. It may
therefore be useful to compare the outcome for criterion E with the outcomes of other
criteria and provide insights into possible reasons for differences in assessment
outcomes.

Documentation

A greater level of documentation is required for criterion E than for other criteria, given the
scientific nature of modelling and the effects of uncertainty. It is recommended that
assessors publish their models in the peer-reviewed literature and place their materials
(data, code) in data repositories to allow full scrutiny of models and their outcomes. Within
the RLE peer review, risk assessment and modelling experts will review models against
strict criteria and may request additional analyses. Specific guidance and examples of the
application of criterion E are currently under development, and will be made available on the
IUCN Red List of Ecosystems website (www.iucnrle.org).

77 | IUCN Red List of Ecosystems

Box 14. Developing a quantitative model of ecosystem dynamics (criterion E)

The probability of ecosystem collapse has been estimated for the Coorong Lagoon of South
Australia, through the adaptation of an empirically derived state-and-transition model (Appendix
S2 in Keith et al., 2013; Lester & Fairweather, 2011). Ecosystem collapse occurred when half of
the modelled years occurred either in degraded ecosystem states or in a period of recovery
following the occurrence of degraded states.

The quantitative assessment of the likelihood of ecosystem collapse in the Coorong was
undertaken with a chain-of-models (Lester & Fairweather, 2011). Downscaled simulations from
multiple global climate models were applied to hydrologic models for the Murray-Darling Basin to
estimate a time series of flows. Six scenarios were investigated to quantify the likelihood of
ecological collapse in the Coorong based on three climate projections for 2030 and two extraction
levels (i.e. with, and without current infrastructure and extraction). All scenarios were run for a
period of 114 years (Lester & Fairweather, 2009). Given that each scenario should be interpreted
as 114 years of possible variability due to climatic fluctuations, the proportion of years occurring in
degraded or recovery states provides an assessment of the stochasticity within the system.

Figure 17. Likelihood of collapse of the Coorong Lagoon under six scenarios of climate
change and water extraction. The three climate scenarios are: historical sequence since
1895; the median future climate projection based on three climate change scenarios from
15 global climate models; and a dry future climate projection based on the 10th percentile
of the same models.

Of the six scenarios investigated, ecological collapse occurred in four. Water extraction will not
cease in the Murray-Darling Basin, so the ‘without development’ scenarios can be discounted from
the overall calculation of risk of collapse. The likelihood of ecological collapse ranges from 30% to
100% across three scenarios representing current levels of development. The Coorong Lagoon is
thus listed as Critically Endangered (plausible range Endangered – Critically Endangered) under
criterion E.

78 | IUCN Red List of Ecosystems

6. Peer review and publication

The IUCN Red List of Ecosystems Committee for Scientific Standards will coordinate
independent peer reviews of risk assessments for the global IUCN Red List of Ecosystems.
Reviews of sub-global assessments will be the responsibility of project managers, though
they are encouraged to seek advice from the Committee for Scientific Standards.
Assessments will be reviewed by at least two experts: one with expertise in the ecology of
the ecosystem type under assessment, and another familiar with the IUCN Red List of
Ecosystems Categories and Criteria.

The review criteria include:

1. Whether the ecosystem type is consistent with the conceptual definition of an

ecosystem, and hence a valid unit for assessment using the IUCN Red List of
Ecosystems criteria.

2. Whether documentation includes or references an adequate description of the

ecosystem type. This includes crosswalks to relevant classifications, an account of
key ecological processes and threats, a graphical conceptual model, and a
quantitative evaluation of each criterion for which data exist.

3. Whether all accessible data and information relevant to IUCN Red List of
Ecosystems assessment of the ecosystem type have been addressed.

4. Whether the quality of underlying data has been evaluated and found to be
adequate.

5. Whether definitions and concepts in the Guidelines have been correctly interpreted
and applied.

6. Whether methods and calculations have been validly applied, and whether
alternative methods are more suitable.

7. Whether estimates of variables for past, present, future, and collapsed states are
complete and supported by evidence.

8. Whether inferences related to the IUCN Red List of Ecosystems criteria are justified
and transparently communicated.

9. Whether uncertainties have been adequately incorporated in the assessment.

79 | IUCN Red List of Ecosystems

7. References
Akcakaya, H.R., Ferson, S., Burgman, M.A., Keith, D.A., Mace, G.M. and Todd, C.R. (2000).
‘Making Consistent IUCN Classifications under Uncertainty’. Conservation Biology
14:1001–1013.
Aladin, N. & Plotnikov, I. (1993). ‘Large saline lakes of former USSR: a summary review’.
Hydrobiologia 267:1–12.
Alaniz, A.J., Galleguillos, M., and Perez-Quezada, J.F. (2016) Assessment of quality of input
data used to classify ecosystems according to the IUCN Red List methodology: The
case of the central Chile hotspot. Biological Conservation 204: 378-385.
Alevizon, W.S. (2010). Introduction To Caribbean Coral Reefs. Gainsville, USA: William
Alevizon.
Allen, C., Gunderson, L. and Johnson, A.R. (2005). ‘The Use of Discontinuities and
Functional Groups to Assess Relative Resilience in Complex Systems’. Ecosystems
8: 958–966.
Allen, T.F. and Hoekstra, T.W. (1990). ‘The confusion between scale‐defined levels and
conventional levels of organization in ecology’. Journal of Vegetation Science 1:5–12.
Andréfouët, S., Muller-Karger, F.E., Robinson, J.A., Kranenburg, C.J., Torres-Pulliza, D.,
Spraggins, S.A. and Murch, B. (2006). ‘Global assessment of modern coral reef
extent and diversity for regional science and management applications: a view from
space’. In: Suzuki, Y., Nakamori, T., Hidaka, M., Kayanne, H., Casareto, B.,
Nadaoka, K., Yamano, H. and Tsuchiya, M. (eds.) Proceedings of 10th International
Coral Reef Symposium, Okinawa, Japan, 28 June-2 July 2004, pp. 1732–1745.
Tokyo, Japan: Japanese Coral Reef Society.
Arias-González, J.E., Acosta-González, G., Membrillo, N., Garza-Pérez, J.R. and Castro-
Pérez, J.M. (2012). ‘Predicting spatially explicit coral reef fish abundance, richness
and Shannon–Weaver index from habitat characteristics’. Biodiversity and
Conservation 21:115–130.
Aronson, R.B. and Blake, D.B. (2001). ‘Global climate change and the origin of modern
benthic communities in Antarctica’. American Zoologist 41:27–39.
Auld, T.D. and Hutton, I. (2004). ‘Conservation issues for the vascular flora of Lord Howe
Island’. Cunninghamia 8:490–500.
Auld, T.D. and Leishman, M.R. (2015). ‘Ecosystem risk assessment for Gnarled Mossy
Cloud Forest, Lord Howe Island, Australia’. Austral Ecology 40:364–372.
Barrett, S. and Yates, C.J. (2015). ‘Risks to a mountain summit ecosystem with endemic
biota in southwestern Australia’. Austral Ecology 40:423–432.
Beck, M.W., Brumbaugh, R.D., Airoldi, L., Carranza, A., Coen, L.D., Crawford, C., Defeo, O.,
Edgar, G.J., Hancock, B., Kay, M.C., Lenihan, H.S., Luckenbach, M.W., Toropova, C.
L., Zhang, G. and Guo, X. (2011). ‘Oyster Reefs at Risk and Recommendations for
Conservation, Restoration, and Management’. BioScience 61:107–116.
Blab, J., Riecken, U. and Ssymank, A. (1995). ‘Proposal on a criteria system for a National
Red Data Book of Biotopes’. Landscape Ecology 10:41–50.
Burgman, M.A. (2015) Trusting judgements: how to get the best out of experts. Cambridge
University Press, Cambridge.

80 | IUCN Red List of Ecosystems

Burns, E.L., Lindenmayer, D.B., Stein, J., Blanchard, W., McBurney, L., Blair, D. and Banks,
S.C. (2015). ‘Ecosystem assessment of mountain ash forest in the Central Highlands
of Victoria, south-eastern Australia’. Austral Ecology 40:386–399.
Cabezas, A., Comín, F.A., Beguería, S. and Trabucchi, M. (2009). ‘Hydrologic and
landscape changes in the Middle Ebro River (NE Spain): implications for restoration
and management’. Hydrology and Earth Systems Science 13:273–284.
Calcagno, V., Sun, C., Schmitz, O.J. and Loreau, M. (2011). ‘Keystone Predation and Plant
Species Coexistence: The Role of Carnivore Hunting Mode’. The American Naturalist
177(1):E1–E13.
Cardinale, B.J., Duffy, J.E., Gonzalez, A., Hooper, D.U., Perrings, C., Venail, P., Narwani,
A., Mace, G.M., Tilman, D., Wardle, D.A., Kinzig, A.P., Daily, G.C., Loreau, M.,
Grace, J.B., Larigauderie, A., Srivastava, D.S. and Naeem, S. (2012). ‘Biodiversity
loss and its impact on humanity’. Nature 486:59–67.
Cardinale, B.J., Matulich, K.L., Hooper, D.U., Byrnes, J.E., Duffy, E., Gamfeldt, L.,
Balvanera, P., O’Connor, M.I. and Gonzalez, A. (2011). ‘The functional role of
producer diversity in ecosystems’. American Journal of Botany 98:572–592.
Cardinale, B.J., Wright, J.P., Cadotte, M.W., Carroll, I.T., Hector, A., Srivastava, D.S.,
Loreau, M. and Weis, J.J. (2007). ‘Impacts of plant diversity on biomass production
increase through time because of species complementarity’. Proceedings of the
National Academy of Sciences 104:18123–18128.
Carpenter, S.R. (2003). Regime Shifts in Lake Ecosystems: Patterns and Variation.
Excellence in Ecology Series, Volume 15. Olderdorf/Luhe, Germany: International
Ecology Institute.
Carpenter, S.R. and Turner, M.G. (2001). ‘Hares and Tortoises: Interactions of Fast and
Slow Variables in Ecosystems’. Ecosystems 3:495–497.
Catling, P.M. and Brownell, V.R. (1995). ‘A review of the alvars of the Great Lakes region:
distribution, floristic composition, biogeography and protection’. Canadian Field
Naturalist 109:143–171.
Chapelle, G. and Peck, L.S. (1999). ‘Polar gigantism dictated by oxygen availability’. Nature
399:114–115.
Christensen, V. and Walters, C.J. (2004). ‘Ecopath with Ecosim: methods, capabilities and
limitations’. Ecological modelling 172:109–139.
Clark, G.F., Raymond, B., Riddle, M.J., Stark, J.S. and Johnston, E.L. (2015). ‘Vulnerability
of Antarctic shallow invertebrate-dominated ecosystems’. Austral Ecology 40:482–
491.
Clarke, P.J., Latz, P.K. and Albrecht, D.E. (2005). ‘Long-term changes in semi-arid
vegetation: Invasion of an exotic perennial grass has larger effects than rainfall
variability’. Journal of Vegetation Science 16:237–248.
Commonwealth of Australia (2000). Environment Protection and Biodiversity Conservation
Regulations, REG 7.02 Criteria for listing threatened ecological communities.
Commonwealth of Australia, Canberra, ACT.
Congalton, R.G. and Green, K. (2008). Assessing the Accuracy of Remotely Sensed Data:
Principles and Practices, 2nd Edition. Boca Raton, Florida, USA: CRC Press/Taylor &
Francis Group.

81 | IUCN Red List of Ecosystems

Connors, B.M., Cooper, A.B., Peterman, R.M. and Dulvy, N.K. (2014). ‘The false
classification of extinction risk in noisy environments’. Proceedings of the Royal
Society B 281:20132935.
Council of the European Commission (1992). ‘Council directive 92/43/EEC of 21 May 1992
on the conservation of natural habitats and of wild fauna and flora’. Official Journal of
the European Communities, Series L 206:7–49.
Crespin, S.J. and Simonetti, J.A. (2015). ‘Predicting ecosystem collapse: Spatial factors that
influence risks to tropical ecosystems’. Austral Ecology 40:492–501.
Dayton, P. (1979). ‘Observations of growth, dispersal and population dynamics of some
sponges in McMurdo Sound, Antarctica’. In: Levi, C. and Boury-Esnault, N. (eds.)
Biologie des Spongiares (Sponge Biology). Colloques Internationaux du Centre
National de la Reserch Scientifique 291:271–282.
Dayton, P.K. (1990). ‘Polar benthos’. In: W.O. Smith (ed.) Polar oceanography, pp. 631-685.
London: Academic Press.
de Visser, S.N., Freymann, B.P. and Olff, H. (2011). ‘The serengeti food web: empirical
quantification and analysis of topological changes under increasing human impact’.
Journal of Animal Ecology 80:484–494.
DEC (2007). Definitions, categories and criteria for threatened and priority ecological
communities. Perth: Department of Environment and Conservation (DEC),
Government of Western Australia.
DECC (2007). Lord Howe Island Biodiversity Management Plan. Sydney: Department of
Environment and Climate Change (DECC), NSW.
Di Fonzo, M., Collen, B. and Mace, G.M. (2013). ‘A new method for identifying rapid decline
dynamics in wild vertebrate populations’. Ecology and Evolution 3:2378–2391.
Diamond, J. (2007). ‘Easter Island revisited’. Science 317:1692–1694.
Diamond, J.M. (1989). ‘The present, past and future of human-caused extinctions’.
Philosophical Transactions of the Royal Society of London B 325:469–477.
Domning, D. (1972). ‘Steller’s sea cow and the origin of North Pacific aboriginal whaling’.
Syesis 5:187–189.
Doroff, A.M., Estes, J.A., Tinker, M.T., Burn, D.M. and Evans, T.J. (2003). ‘Sea otter
population declines in the Aleutian archipelago’. Journal of Mammalogy 84:55–64.
Driver, A., Sink, K.J., Nel, J.N., Holness, S., Van Niekerk, L., Daniels, F., Jonas, Z., Majiedt,
P.A., Harris, L. and Maze, K. (2012). National Biodiversity Assessment 2011: An
assessment of South Africa’s biodiversity and ecosystems. Synthesis Report’.
Pretoria: South African National Biodiversity Institute and Department of
Environmental Affairs.
Earl, J. (2003). The Distribution and Impacts of Lippia (Phyla canescens) in the Murray
Darling System. Guyra: Murray Darling Basin Lippia Working Group.
Eisenhauer, N., Milcu, A., Sabais, A.C.W., Bessler, H., Brenner, J., Engels, C., Klarner, B.,
Maraun, M., Partsch, S., Roscher, C., Schonert, F., Temperton, V.M., Thomisch, K.,
Weigelt, A., Weisser, W.W. and Scheu, S. (2011). ‘Plant Diversity Surpasses Plant
Functional Groups and Plant Productivity as Driver of Soil Biota in the Long Term’.
PLoS ONE 6:e16055.
Elith, J., Burgman, M.A. and Regan, H.M. (2002). ‘Mapping epistemic uncertainties and
vague concepts in predictions of species distribution’. Ecological Modelling 157:313–
329.

82 | IUCN Red List of Ecosystems

Essl, F. and Egger, G. (2010). Lebensraumvielfalt in Österreich – Gefährdung und
Handlungsbedarf. Zusammenschau der Roten Liste gefährdeter Biotoptypen
Österreichs. Klagenfurt: Naturwissenschaftlicher Verein für Kärnten und
Umweltbundesamt GmbH.
Estes, J.A., Doak, D., Springer, A. and Williams, T. (2009). ‘Causes and consequences of
marine mammal population declines in southwest Alaska: a food-web perspective’.
Philosophical Transactions of the Royal Society B 364:1647–1658.
Estes, J.A., Terborgh, J., Brashares, J.S., Power, M.E., Berger, J., Bond, W.J., Carpenter, S.
R., Essington, T.E., Holt, R.D. and Jackson, J.B. (2011). ‘Trophic downgrading of
planet Earth’. Science 333:301–306.
Estes, J.A., Tinker, M. and Bodkin, J.L. (2010). ‘Using ecological function to develop
recovery criteria for depleted species: sea otters and kelp forests in the Aleutian
archipelago’. Conservation Biology 24:852–860.
Faber-Langendoen, D., Keeler-Wolf, T., Meidinger, D., Tart, D., Hoagland, B., Josse, C.,
Navarro, G., Ponomarenko, S., Saucier, J.-P., Weakley, A. and Comer, P. (2014).
‘EcoVeg: a new approach to vegetation description and classification’. Ecological
Monographs 84:533–561.
Fischer, J., Lindenmayer, D.B., Blomberg, S.P., Montague-Drake, R., Felton, A. and Stein,
J.A. (2007). ‘Functional Richness and Relative Resilience of Bird Communities in
Regions with Different Land Use Intensities’. Ecosystems 10:964–974.
Folke, C., Carpenter, S., Walker, B., Scheffer, M., Elmqvist, T., Gunderson, L. and Holling,
C.S. (2004). ‘Regime shifts, resilience, and biodiversity in ecosystem management’.
Annual Review of Ecology, Evolution, and Systematics 35:557–581.
Fontaine, C., Dajoz, I., Meriguet, J. and Loreau, M. (2005). ‘Functional diversity of plant–
pollinator interaction webs enhances the persistence of plant communities’. PLoS
Biol 4(1), e1. [doi:10.1371/journal.pbio.0040001]
Foody, G.M. (2011). ‘Impacts Of Imperfect Reference Data On The Apparent Accuracy Of
Species Presence-Absence Models And Their Predictions’. Global Ecology And
Biogeography 20(3):498–508.
Fuller, R.M., Smith, G.M. and Devereux, B.J. (2003). ‘The characterisation and
measurement of land cover change through remote sensing: problems in operational
applications?’. International Journal of Applied Earth Observation and
Geoinformation 4:243–253.
Gaston, K.J. (1994). Rarity. London, UK: Chapman and Hall.
Gaston, K.J. and Fuller, R.A. (2008). ‘Commonness, population depletion and conservation
biology’. Trends in Ecology & Evolution 23:14–19.
Gaston, K.J. and Fuller, R.A. (2009). ‘The sizes of species’ geographic ranges’. Journal of
Applied Ecology 46:1–9.
Gili, J.-M., Arntz, W.E., Palanques, A., Orejas, C., Clarke, A., Dayton, P.K., Isla, E., Teixidó,
N., Rossi, S. and López-González, P.J. (2006). ‘A unique assemblage of epibenthic
sessile suspension feeders with archaic features in the high-Antarctic’. Deep Sea
Research Part II: Topical Studies in Oceanography 53:1029–1052.
Giri, C., Ochieng, E., Tieszen, L., Zhu, Z., Singh, A., Loveland, T., Masek, J. and Duke, N.
(2011). ‘Status and distribution of mangrove forests of the world using earth
observation satellite data’. Global Ecology and Biogeography 20:154–159.

83 | IUCN Red List of Ecosystems

Goudard, A. and Loreau, M. (2008). ‘Nontrophic interactions, biodiversity, and ecosystem
functioning: an interaction web model’. The American Naturalist 171:91–106.
Government of Manitoba (2014). The Endangered Species and Ecosystems Act.
<https://web2.gov.mb.ca/laws/statutes/ccsm/e111e.php?ccsm=e111>. Accessed 15
November 2015.
Grebmeier, J.M., Overland, J.E., Moore, S.E., Farley, E.V., Carmack, E.C., Cooper, L.W.,
Frey, K.E., Helle, J.H., McLaughlin, F.A. and McNutt, S.L. (2006). ‘A major
ecosystem shift in the northern Bering Sea’. Science 311:1461–1464.
Green, P.T., O'Dowd, D.J., Abbott, K.L., Jeffery, M., Retallick, K. and Mac Nally, R. (2011).
‘Invasional meltdown: Invader-invader mutualism facilitates a secondary invasion’.
Ecology 92:1758–1768.
Gregr, E.J., Ahrens, A.L. and Perry, R.I. (2012). ‘Reconciling classifications of ecologically
and biologically significant areas in the world’s oceans’. Marine Policy 36:716–726.
Hannah, J. and Bell, R.G. (2012). ‘Regional sea level trends in New Zealand’. Journal of
Geophysical Research: Oceans 117:C01004.
Hansen, M.C., Potapov, P.V., Moore, R., Hancher, M., Turubanova, S.A., Tyukavina, A.,
Thau, D., Stehman, S.V., Goetz, S.J., Loveland, T.R., Kommareddy, A., Egorov, A.,
Chini, L., Justice, C.O. and Townshend, J.R.G. (2013). ‘High-Resolution Global Maps
of 21st-Century Forest Cover Change’. Science 342:850–853.
Hanski, I. (1998). ‘Metapopulation dynamics’. Nature 396:41–49.
Harfoot, M.B.J., Newbold, T., Tittensor, D.P., Emmott, S., Hutton, J., Lyutsarev, V., Smith,
M.J., Scharlemann, J.P.W. and Purves, D.W. (2014). ‘Emergent Global Patterns of
Ecosystem Structure and Function from a Mechanistic General Ecosystem Model’.
PLoS Biol 12:e1001841.
Harpole, W.S. and Tilman, D. (2007). ‘Grassland species loss resulting from reduced niche
dimension’. Nature 446:791–793.
Hartley, S. and Kunin, W.E. (2003). ‘Scale dependency of rarity, extinction risk, and
conservation priority’. Conservation Biology 17:1559–1570.
Healy, T., Wang, Y. and Healy, J. (eds.) (2002). Muddy Coasts of the World: Processes,
Deposits, and Function, Amsterdam, The Netherlands: Elsevier Science.
Hermoso, V., Januchowski-Hartley, S.R. & Pressey, R.L. (2013). ‘When the suit does not fit
biodiversity: loose surrogates compromise the achievement of conservation goals’.
Biological Conservation 159:197–205.
Hobbs, R.J., Arico, S., Aronson, J., Baron, J.S., Bridgewater, P., Cramer, V.A., Epstein,
P.R., Ewel, J.J., Klink, C.A., Lugo, A.E., Norton, D., Ojima, D., Richardson, D.M.,
Sanderson, E.W., Valladares, F., Vilà, M., Zamora, R. and Zobel, M. (2006). ‘Novel
ecosystems: theoretical and management aspects of the new ecological world order’.
Global Ecology and Biogeography 15:1–7.
Holdaway, R.J., Wiser, S.K. and Williams, P.A. (2012). ‘Status assessment of New
Zealand's naturally uncommon ecosystems’. Conservation Biology 26:619–629.
Holdo, R.M., Holt, R.D. and Fryxell, J.M. (2013). ‘Herbivore–vegetation feedbacks can
expand the range of savanna persistence: insights from a simple theoretical model’.
Oikos 122:441–453.
Hong, S. and Shin, I. (2010). ‘Global Trends of Sea Ice: Small-Scale Roughness and
Refractive Index’. Journal of Climate 23:4669–4676.

84 | IUCN Red List of Ecosystems

Huth, N. and Possingham, H.P. (2011). ‘Basic ecological theory can inform habitat
restoration for woodland birds’. Journal of Applied Ecology 48:293–300.
IUCN (2001). IUCN Red List Categories and Criteria: Version 3.1, Gland, Switzerland and
Cambridge, UK: IUCN Species Survival Commission.
IUCN (2012). IUCN Red List Categories and Criteria: Version 3.1. Second edition. Gland,
Switzerland and Cambridge, UK: IUCN.
IUCN (2015). The IUCN Red List of Threatened Species. Version 2015-4.
<www.iucnredlist.org>. Accessed 19 November 2015.
IUCN Standards and Petitions Subcommittee (2016). Guidelines for Using the IUCN Red
List Categories and Criteria. Version 12. Prepared by the Standards and Petitions
Subcommittee. <www.iucnredlist.org/documents/RedListGuidelines.pdf>
Jakeman, A.J., Letcher, R.A. and Norton, J.P. (2006). ‘Ten iterative steps in development
and evaluation of environmental models’. Environmental Modelling & Software
21:602–614.
Jarvis, A. and Mulligan, M. (2011). ‘The climate of cloud forests’. Hydrological Processes
25:327–343.
Kauffman, A. and Gupta, M.M. (1991). Introduction to fuzzy arithmetic: theory and
application. New York: VanNostrand Reinhold.
Keith, D. and Myerscough, P. (1993). ‘Floristics and soil relations of upland swamp
vegetation near Sydney’. Australian Journal of Ecology 18:325–344.
Keith, D.A. (2009). ‘The interpretation, assessment and conservation of ecological
communities’. Ecological Management & Restoration 10:S3–S15.
Keith, D.A. (2015). ‘Assessing and managing risks to ecosystem biodiversity’. Austral
Ecology 40:337–346.
Keith, D.A., Akçakaya, H.R. and Murray N.J. (in review) Scaling range sizes to threats for
robust predictions of risks to biodiversity. Conservation Biology.
Keith, D.A., Akçakaya, H.R., Thuiller, W., Midgley, G.F., Pearson, R.G., Phillips, S.J., Regan,
H.M., Araújo, M.B. and Rebelo, T.G. (2008). ‘Predicting extinction risks under climate
change: coupling stochastic population models with dynamic bioclimatic habitat
models’. Biology Letters 4:560–563.
Keith, D.A., Elith, J. and Simpson, C.C. (2014). ‘Predicting distribution changes of a mire
ecosystem under future climates’. Diversity and Distributions 20:440–454.
Keith, D.A., Holman, L., Rodoreda, S., Lemmon, J. and Bedward, M. (2007). ‘Plant functional
types can predict decade-scale changes in fire-prone vegetation’. Journal of Ecology
95:1324–1337.
Keith, D.A., Orscheg, C., Simpson, C.C., Clarke, P.J., Hughes, L., Kennelly, S.J., Major,
R.E., Soderquist, T.R., Wilson, A.L. and Bedward, M. (2009). ‘A new approach and
case study for estimating extent and rates of habitat loss for ecological communities’.
Biological Conservation 142:1469–1479.
Keith, D.A., Rodríguez, J.P., Brooks, T.M., Burgman, M.A., Barrow, E.G., Bland, L., Comer,
P.J., Franklin, J., Link, J., McCarthy, M.A., Miller, R.M., Murray, N.J., Nel, J.,
Nicholson, E., Oliveira-Miranda, M.A., Regan, T.J., Rodríguez-Clark, K.M., Rouget,
M. and Spalding, M.D. (2015). ‘The IUCN Red List of Ecosystems: Motivations,
Challenges, and Applications’. Conservation Letters 8:214–226.

85 | IUCN Red List of Ecosystems

Keith, D.A., Rodríguez, J.P., Rodríguez-Clark, K.M., Nicholson, E., Aapala, K., Alonso, A.,
Asmussen, M., Bachman, S., Basset, A., Barrow, E.G., Benson, J.S., Bishop, M.J.,
Bonifacio, R., Brooks, T.M., Burgman, M.A., Comer, P., Comín, F.A., Essl, F., Faber-
Langendoen, D., Fairweather, P.G., Holdaway, R.J., Jennings, M., Kingsford, R.T.,
Lester, R.E., Nally, R.M., McCarthy, M.A., Moat, J., Oliveira-Miranda, M.A., Pisanu,
P., Poulin, B., Regan, T.J., Riecken, U., Spalding, M.D. and Zambrano-Martínez, S.
(2013). ‘Scientific Foundations for an IUCN Red List of Ecosystems’. PLoS ONE
8(5):e62111.
Kingsford, R. and Johnson, W. (1998). ‘Impact of water diversions on colonially-nesting
waterbirds in the Macquarie Marshes of arid Australia’. Colonial Waterbirds
21(2):159–170.
Kingsford, R.T. and Auld, K.M. (2005). ‘Waterbird breeding and environmental flow
management in the Macquarie Marshes, arid Australia’. River Research and
Applications 21:187–200.
Kingsford, R.T. and Thomas, R.F. (1995). ‘The Macquarie Marshes in arid Australia and their
waterbirds: a 50-year history of decline’. Environmental Management 19:867–878.
Kirby, M.X. (2004). ‘Fishing down the coast: Historical expansion and collapse of oyster
fisheries along continental margins’. Proceedings of the National Academy of
Sciences of the United States of America 101:13096–13099.
Kontula, T. and Raunio, A. (2009). ‘New method and criteria for national assessments of
threatened habitat types’. Biodiversity and Conservation 18:3861–3876.
Kujala, H., Moilanen, A., Araújo, M.B. and Cabeza, M. (2013). ‘Conservation planning with
uncertain climate change projections’. PloS ONE 8:e53315.
Lehmann, C.E.R., Anderson, T.M., Sankaran, M., Higgins, S.I., Archibald, S., Hoffmann,
W.A., Hanan, N.P., Williams, R.J., Fensham, R.J., Felfili, J., Hutley, L.B., Ratnam, J.,
San Jose, J., Montes, R., Franklin, D., Russell-Smith, J., Ryan, C.M., Durigan, G.,
Hiernaux, P., Haidar, R., Bowman, D.M.J.S. and Bond, W.J. (2014). ‘Savanna
vegetation-fire-climate relationships differ among continents’. Science 343:548–552.
Lester, R.E. and Fairweather, P.G. (2009). ‘Modelling future conditions in the degraded
semi-arid estuary of Australia's largest river using ecosystem states’. Estuarine,
Coastal and Shelf Science 85:1–11.
Lester, R.E. and Fairweather, P.G. (2011). ‘Ecosystem states: Creating a data-derived,
ecosystem-scale ecological response model that is explicit in space and time’.
Ecological Modelling 222:2690–2703.
Likens, G.E. (1992). The Ecosystem Approach: Its Use and Abuse. Excellence in Ecology
Series, Volume 3. Olderdorf/Luhe, Germany: Ecology Institute
Lindenmayer, D. (2009). Forest pattern and ecological process: a synthesis of 25 years of
research. Melbourne, Australia: CSIRO Publishing.
Lindgaard, A. and Henriksen, S. (eds.) (2011). The 2011 Norwegian Red List for
Ecosystems and Habitat Types. Trondheim: Norwegian Biodiversity Information
Centre.
Loreau, M., Mouquet, N. and Gonzalez, A. (2003). ‘Biodiversity as spatial insurance in
heterogeneous landscapes’. Proceedings of the National Academy of Sciences
100:12765–12770.
Ludwig, J.A., Bastin, G.N., Chewings, V.H., Eager, R.W. and Liedloff, A.C. (2007).
‘Leakiness: A new index for monitoring the health of arid and semiarid landscapes

86 | IUCN Red List of Ecosystems

 using remotely sensed vegetation cover and elevation data’. Ecological Indicators
7:442–454.
Lundberg, J. and Moberg, F. (2003). ‘Mobile Link Organisms and Ecosystem Functioning:
Implications for Ecosystem Resilience and Management’. Ecosystems 6:87–98.
Mac Nally, R., Cunningham, S.C., Baker, P.J., Horner, G.J. and Thomson, J.R. (2011).
‘Dynamics of Murray-Darling floodplain forests under multiple stressors: The past,
present, and future of an Australian icon’. Water Resources Research 47:W00G05.
MacArthur, R.H. and Wilson, E.O. (1967). The theory of island biogeography. Princeton,
New Jersey: Princeton University Press.
Mace, G.M., Collar, N.J., Gaston, K.J., Hilton-Taylor, C., Akcakaya, H.R., Leader-Williams,
N., Milner-Gulland, E.J. and Stuart, S.N. (2008). ‘Quantification of extinction risk:
IUCN's system for classifying threatened species’. Conservation Biology 22:1424–42.
MacKinnon, J., Verkuil, Y.I. & Murray, N.J. (2012). IUCN situation analysis on East and
Southeast Asian intertidal habitats, with particular reference to the Yellow Sea
(including the Bohai Sea). Occasional Paper of the IUCN Species Survival
Commission No. 47. Gland, Switzerland and Cambridge, UK: IUCN.
Mastrandrea, M.D., Field, C.B., Stocker, T.F., Edenhofer, O., Ebi, K.L., Frame, D.J., Held,
H., Kriegler, E., Mach, K.J. and Matschoss, P.R. (2010). Guidance Note for Lead
Authors of the IPCC Fifth Assessment Report on Consistent Treatment of
Uncertainties. Intergovernmental Panel on Climate Change (IPCC).
Maxwell, P.S., Pitt, K.A., Olds, A.D., Rissik, D. and Connolly, R.M. (2015). ‘Identifying
habitats at risk: simple models can reveal complex ecosystem dynamics’. Ecological
Applications 25:573–587.
McClintock, J.B., Amsler, C.D., Baker, B.J. and Van Soest, R.W. (2005). ‘Ecology of
Antarctic marine sponges: an overview’. Integrative and Comparative Biology
45:359–368.
McKnight, M.W., White, P.S., McDonald, R.I., Lamoreux, J.F., Sechrest, W., Ridgely, R.S.
and Stuart, S.N. (2007). ‘Putting beta-diversity on the map: broad-scale congruence
and coincidence in the extremes’. PLoS Biol 5:e272.
Metternicht, G.I. and Zinck, J.A. (2003). ‘Remote sensing of soil salinity: potentials and
constraints’. Remote Sensing of Environment 85:1–20.
Meyer, J. (2011). ‘Montane cloud forests on remote islands of Oceania: the example of
French Polynesia (South Pacific Ocean)’. In: Bruijnzeel, L.A., Scatena, F.N. and
Hamilton, L.S. (eds).Tropical Montane Cloud Forests: Science for Conservation and
Management, pp. 121–129. Cambridge, UK: Cambridge University Press.
Micklin, P. and Aladin, N.V. (2008). ‘Reclaiming the Aral Sea’. Scientific American 298:64–
71.
Midgley, G.F., Davies, I.D., Albert, C.H., Altwegg, R., Hannah, L., Hughes, G.O., O'Halloran,
L.R., Seo, C., Thorne, J.H. and Thuiller, W. (2010). ‘BioMove–an integrated platform
simulating the dynamic response of species to environmental change’. Ecography
33:612–616.
Miller, G., Friedel, M., Adam, P. and Chewings, V. (2010). ‘Ecological impacts of buffel grass
(Cenchrus ciliaris L.) invasion in central Australia – does field evidence support a fire-
invasion feedback?’. The Rangeland Journal 32:353–365.
Moberg, F. and Folke, C. (1999). ‘Ecological goods and services of coral reef ecosystems’.
Ecological Economics 29:215–233.

87 | IUCN Red List of Ecosystems

Mucina, L. and Rutherford, M.C. (eds.) (2006). The Vegetation of South Africa, Lesotho and
Swaziland. Strelitzia 19. Pretoria: South African National Biodiversity Institute.
Mueller-Dombois, D. and Fosberg, F.R. (1998). Vegetation of the Tropical Pacific Islands.
New York: Springer-Verlag.
Murray, N.J. (2017) Global 10 x 10-km grids suitable for use in IUCN Red List of Ecosystems
assessments (vector and raster format). figshare
https://doi.org/10.6084/m9.figshare.4653439.v1
Murray, N.J., Clemens, R.S., Phinn, S.R., Possingham, H.P. and Fuller, R.A. (2014).
‘Tracking the rapid loss of tidal wetlands in the Yellow Sea’. Frontiers in Ecology and
the Environment 12:267–272.
Murray, N.J. and Fuller, R.A. (2015). ‘Protecting stopover habitat for migratory shorebirds in
East Asia’. Journal of Ornithology 156:217-225.
Murray, N.J., Keith, D.A., Bland, L.M., Nicholson, E., Regan, T.M., Rodriguez, J.P.,
Bedward, M. (2017) The use of range size to assess risks to biodiversity from
stochastic threats. Diversity and Distributions. doi:10.1111/ddi.12533
Murray, N.J., Ma, Z. and Fuller, R.A. (2015). ‘Tidal flats of the Yellow Sea: A review of
ecosystem status and anthropogenic threats’. Austral Ecology 40:472–481.
Murray, N.J., Phinn, S.R., Clemens, R.S., Roelfsema, C.M. and Fuller, R.A. (2012).
‘Continental Scale Mapping of Tidal Flats across East Asia Using the Landsat
Archive’. Remote Sensing 4:3417–3426.
Newton, A.C. (2010). ‘Use of a Bayesian network for Red Listing under uncertainty’.
Environmental Modelling & Software 25:15–23.
Nicholson, E., Keith, D.A. and Wilcove, D.S. (2009). ‘Assessing the threat status of
ecological communities’. Conservation Biology 23:259–274.
Nicholson, E., Regan, T.J., Auld, T.D., Burns, E.L., Chisholm, L.A., English, V., Harris, S.,
Harrison, P., Kingsford, R.T., Leishman, M.R., Metcalfe, D.J., Pisanu, P., Watson,
C.J., White, M., White, M.D., Williams, R.J., Wilson, B. and Keith, D.A. (2015).
‘Towards consistency, rigour and compatibility of risk assessments for ecosystems
and ecological communities’. Austral Ecology 40:347–363.
NSW Scientific Committee (2004). Coolibah – Black Box Woodland of the northern riverine
plains in the Darling Riverine Plains and Brigalow Belt South bioregions –
endangered ecological community listing. Final Determination. Sydney: NSW
Scientific Committee.
Olofsson, P., Foody, G.M., Herold, M., Stehman, S.V., Woodcock, C.E. and Wulder, M.A.
(2014). ‘Good practices for estimating area and assessing accuracy of land change’.
Remote Sensing of Environment 148:42–57.
Olofsson, P., Foody, G.M., Stehman, S.V. and Woodcock, C.E. (2013). ‘Making better use of
accuracy data in land change studies: Estimating accuracy and area and quantifying
uncertainty using stratified estimation’. Remote Sensing of Environment 129:122–
131.
Payet, K., Rouget, M., Esler, K.J., Reyers, B., Rebelo, T., Thompson, M.W. and Vlok, J.H.
(2013). ‘Effect of Land Cover and Ecosystem Mapping on Ecosystem‐Risk
Assessment in the Little Karoo, South Africa’. Conservation Biology 27:531–541.
Pekel, J.F., Cottam, A., Gorelick, N. and Belward, A.S. (2016) High-resolution mapping of
global surface water and its long-term changes. Nature 540:418-422.

88 | IUCN Red List of Ecosystems

Peterson, G.D., Cumming, G.S. and Carpenter, S.R. (2003). ‘Scenario planning: a tool for
conservation in an uncertain world’. Conservation Biology 17:358–366.
Pickett, S.T.A. (1989). ‘Space-for-time substitution as an alternative to long-term studies’. In:
Likens, G.E. (ed.) Long-term Studies in Ecology: Approaches and Alternatives,
pp.110–135. New York: Springer-Verlag.
Pickett, S.T.A. and Cadenasso, M.L. (1995). ‘Landscape ecology: spatial heterogeneity in
ecological systems’. Science 269:331–334.
Plagányi, É.E., Punt, A.E., Hillary, R., Morello, E.B., Thébaud, O., Hutton, T., Pillans, R.D.,
Thorson, J.T., Fulton, E.A. and Smith, A.D. (2014). ‘Multispecies fisheries
management and conservation: tactical applications using models of intermediate
complexity’. Fish and Fisheries 15:1–22.
Portillo, CA. (2004). Cuantificación del riesgo de extinción de los hábitats terrestres de la
Zona Protectora de la Cuenca Baja de los Ríos Socuy, Guasare y Cachirí. Trabajo
Especial de Grado. La Universidad del Zulia, Maracaibo. 88+xi pp.
Pounds, J.A., Fogden, M.P.L. and Campbell, J.H. (1999). ‘Biological response to climate
change on a tropical mountain’. Nature 398:611–615.
Precautionary Principle Project (2005). Guidelines for Applying the Precautionary Principle to
Biodiversity Conservation and Natural Resource Management. Cambridge, UK: The
Precautionary Principle Project – a joint initiative of Fauna & Flora International,
IUCN, ResourceAfrica and TRAFFIC.
Puyravaud, J.P. (2003) Standardizing the calculation of the annual rate of deforestation.
Forest Ecology and Management 177:593-596.
Rebelo, A.G., Boucher, C., Helme, N., Mucina, L., Rutherford, M.C. (2006). ‘Fynbos biome’.
In: Mucina, L. and Rutherford, M.C. (eds.) The Vegetation of South Africa, Lesotho
and Swaziland. Strelitzia 19, pp.52-219. Pretoria: South African National Biodiversity
Institute.
Regan, H.M. and Colyvan, M. (2000). ‘Fuzzy Sets and Threatened Species Classification’.
Conservation Biology 14:1197–1199.
Regan, H.M., Colyvan, M. and Burgman, M.A. (2002). ‘A taxonomy and treatment of
uncertainty for ecology and conservation biology’. Ecological Applications 12:618–
628.
Reid, M., Ogden, R. and Thoms, M. (2011). ‘The influence of flood frequency, geomorphic
setting and grazing on plant communities and plant biomass on a large dryland
floodplain’. Journal of Arid Environments 75:815–826.
Reisewitz, S.E., Estes, J.A. and Simenstad, C.A. (2006). ‘Indirect food web interactions: sea
otters and kelp forest fishes in the Aleutian archipelago’. Oecologia 146:623–631.
Republic of South Africa (2004). ‘National Environmental Management: Biodiversity Act,
2004’. Government Gazette Republic of South Africa Cape Town 467:1–84
Reschke, C., Reid, R., Jones, J., Feeney, T., and Potter, H. (1999). Conserving Great Lakes
Alvars. Chicago, USA: The Nature Conservancy.
Riecken, U., Finck, P., Raths, U., Schröder, E. and Ssymank, A. (2009). German Red Data
Book on endangered habitats (short version, July 2009). Bonn-Bad Godesberg,
Germany: German Federal Agency for Nature Conservation (BfN).
Riecken, U., Heinzel, K., Schlumprecht, H., Boedeker, D. and für Naturschutz, B. (2006).
Rote Liste der gefährdeten Biotoptypen Deutschlands: zweite fortgeschriebene
Fassung 2006. Bonn, Germany: Bundesamt für Naturschutz.

89 | IUCN Red List of Ecosystems

Robertson, A.I. and Rowling, R.W. (2000). ‘Effects of livestock on riparian zone vegetation in
an Australian dryland river’. Regulated Rivers: Research & Management 16:527–
541.
Rodríguez, J.P., Keith, D.A., Rodríguez-Clark, K.M., Murray, N.J., Nicholson, E., Regan,
T.J., Miller, R.M., Barrow, E.G., Bland, L.M., Boe, K., Brooks, T.M., Oliveira-Miranda,
M.A., Spalding, M. and Wit, P. (2015). ‘A practical guide to the application of the
IUCN Red List of Ecosystems criteria’. Philosophical Transactions of the Royal
Society B 370: 20140003.
Rodríguez, J.P., Rodríguez-Clark, K.M., Baillie, J.E.M., Ash, N., Benson, J., Boucher, T.,
Brown, C., Burgess, N.D., Collen, B., Jennings, M., Keith, D.A., Nicholson, E.,
Revenga, C., Reyers, B., Rouget, M., Smith, T., Spalding, M., Taber, A., Walpole, M.,
Zager, I. and Zamin, T. (2011). ‘Establishing IUCN Red List Criteria for Threatened
Ecosystems’. Conservation Biology 25:21–29.
Rodríguez, J.P., Rodríguez-Clark, K.M., Keith, D.A., Barrow, E.G., Benson, J., Nicholson, E.
and Wit, P. (2012a). ‘IUCN Red List of Ecosystems’. S.A.P.I.EN.S [Online] 5.2
<http://sapiens.revues.org/1286>.
Rodríguez, J.P., Rodríguez-Clark, K.M., Keith, D.A., Barrow, E.G., Comer, P. and Oliveira-
Miranda, M.A. (2012b). ‘From Alaska to Patagonia: the IUCN Red List of the
continental ecosystems of the Americas’. Oryx 46:170–171.
Rodríguez, J.P., Rojas-Suárez, F. and Giraldo Hernández, D. (eds.) (2010). Libro Rojo de
los Ecosistemas Terrestres de Venezuela. Caracas, Venezuela: Provita, Shell
Venezuela y Lenovo (Venezuela).
Rogers, C.S. (1990). ‘Responses of coral reef s and reef organisms to sedimentation’.
Marine Ecology Progress Series 62:185–202.
Rumpff, L., Duncan, D., Vesk, P., Keith, D. and Wintle, B. (2011). ‘State-and-transition
modelling for adaptive management of native woodlands’. Biological Conservation
144:1224–1236.
Salafsky, N., Salzer, D., Stattersfield, A.J., Hilton-Taylor, C., Neugarten, R., Butchart,
S.H.M., Collen, B., Cox, N., Master, L.L., O'Connor, S. and Wilkie, D. (2008). ‘A
standard lexicon for biodiversity conservation: Unified classifications of threats and
actions’. Conservation Biology 22:897–911.
Scheffer, M., Carpenter, S.R., Foley, J., Folke, C. and Walker, B.H. (2001). ‘Catastrophic
shifts in ecosystems’. Nature 413:591–596.
Scholze, M., Knorr, W., Arnell, N.W. and Prentice, I.C. (2006). ‘A climate-change risk
analysis for world ecosystems’. Proceedings of the National Academy of Sciences
103:13116–13120.
Shi, J., Ma, K., Wang, J., Zhao, J. and He, K. (2010). ‘Vascular plant species richness on
wetland remnants is determined by both area and habitat heterogeneity’. Biodiversity
and Conservation 19:1279–1295.
Simmons, C.L., Auld, T.D., Hutton, I., Baker, W. J. and Shapcott, A. (2012). ‘Will climate
change, genetic and demographic variation or rat predation pose the greatest risk for
persistence of an altitudinally distributed island endemic?’. Biology 1:736–765.
Soares-Filho, B.S., Coutinho Cerqueira, G. and Lopes Pennachin, C. (2002). ‘DINAMICA—a
stochastic cellular automata model designed to simulate the landscape dynamics in
an Amazonian colonization frontier’. Ecological Modelling 154:217–235.
Spalding, M.D., Fox, H.E., Halpern, B.S., McManus, M.A., Molnar, J., Allen, G.R., Davidson,
N., Jorge, Z.A., Lombana, A.L., Lourie, S.A., Martin, K.D., McManus, E., Recchia,

90 | IUCN Red List of Ecosystems

 C.A. and Robertson, J. (2007). ‘Marine ecoregions of the world: A bioregionalization
of coastal and shelf areas’. BioScience 57:573–583.
Springer, A.M., Estes, J.A., van Vliet, G.B., Williams, T.M., Doak, D.F., Danner, E.M.,
Forney, K.A. and Pfister, B. (2003). ‘Sequential megafaunal collapse in the North
Pacific Ocean: An ongoing legacy of industrial whaling?’. PNAS 100:12223–12228.
Steneck, R. and Watling, L. (1982). ‘Feeding capabilities and limitation of herbivorous
molluscs: a functional group approach’. Marine Biology 68:299–319.
Steneck, R.S., Graham, M.H., Bourque, B.J., Corbett, D., Erlandson, J.M., Estes, J.A. and
Tegner, M.J. (2002). ‘Kelp forest ecosystems: biodiversity, stability, resilience and
future’. Environmental Conservation 29:436–459.
Tansley, A.G. (1935). ‘The use and abuse of vegetational concepts and terms’. Ecology
16:284–307.
Taylor, B. and Ganf, G.G. (2005). ‘Comparative ecology of two co-occurring floodplain
plants: the native Sporobolus mitchellii and the exotic Phyla canescens’. Marine and
Freshwater Research 56:431–440.
Thébault, E. and Loreau, M. (2005). ‘Trophic Interactions and the Relationship between
Species Diversity and Ecosystem Stability’. The American Naturalist 166:E95–E114.
Thompson, J.N. (1997). Conserving Interaction Biodiversity. In: Pickett, S.T.A., Ostfeld, R.S.,
Shachak, M. and Likens, G.E. (eds.) The Ecological Basis of Conservation, pp.285-
293. Springer US.
Thoms, M. and Sheldon, F. (2000). ‘Water resource development and hydrological change in
a large dryland river: the Barwon–Darling River, Australia’. Journal of Hydrology
228:10–21.
Thoms, M.C. (2003). Floodplain–river ecosystems: lateral connections and the implications
of human interference. Geomorphology, 56, 335-349.
UNEP (2001). Partow, H. The Mesopotamian marshlands: demise of an ecosystem. Early
Warning and Assessment Technical Report, UNEP/DEWA/TR.01-3 Rev. 1. Division
of Early Warning and Assessment. Nairobi, Kenya: United Nations Environment
Programme.
Vázquez, D.P. and Simberloff, D. (2003). ‘Changes in interaction biodiversity induced by an
introduced ungulate’. Ecology Letters 6:1077–1083.
Walker, B., Kinzig, A. and Langridge, J. (1999). ‘Plant Attribute Diversity, Resilience, and
Ecosystem Function: The Nature and Significance of Dominant and Minor Species’.
Ecosystems 2:95–113.
Wang, Y. and Zhu, D. (1994). ‘Tidal flats in China’. Oceanology of China Seas 2:445–456.
Wang, Y.P., Gao, S., Jia, J., Thompson, C.E.L., Gao, J. and Yang, Y. (2012). ‘Sediment
transport over an accretional intertidal flat with influences of reclamation, Jiangsu
coast, China’. Marine Geology 291–294:147–161.
Watling, L. and Norse, E.A. (1998). ‘Disturbance of the seabed by mobile fishing gear: a
comparison to forest clearcutting’. Conservation Biology 12:1180–1197.
Wiens, J.A. (1989). ‘Spatial scaling in ecology’. Functional Ecology 3:385–397.
Wood, J., Low, A.B., Donaldson, J.S. and Rebelo, A.G. (1994). ‘Threats to plant species
diversity through urbanization and habitat fragmentation in the Cape Metropolitan
Area, South Africa’. In: Huntley, B.J. (ed.) Botanical Diversity in Southern Africa.
Strelitzia 1, pp.259-274. Pretoria: South African National Biodiversity Institute.

91 | IUCN Red List of Ecosystems

Zhao, H.L., Zhao, X.Y., Zhou, R.L., Zhang, T.H. and Drake, S. (2005). ‘Desertification
processes due to heavy grazing in sandy rangeland, Inner Mongolia’. Journal of Arid
Environments 62:309–319.

92 | IUCN Red List of Ecosystems

Appendix 1. List of contributors
Name Organisation Country
Kaisu Aapala Finnish Environment Institute Finland
Alfonso Alonso Smithsonian Conservation Biology Institute USA
Marianne Asmüssen Instituto Venezolano de Investigaciones Venezuela
Científicas
Steven Bachman Royal Botanic Gardens, Kew UK
Jonathan Baillie Zoological Society of London UK
Edmund G. Barrow IUCN Kenya / Ireland
Alberto Bassett University of Lecce Italy
John Benson Royal Botanic Gardens and Domain Trust Australia
Melanie J. Bishop Macquarie University Australia
Lucie M. Bland University of Melbourne Australia
Kaia Boe IUCN Switzerland
Ronald Bonifacio South Australian Department of Environment Australia
and Natural Resources
Tim Boucher The Nature Conservancy USA
Thomas M. Brooks IUCN Switzerland
Claire Brown UNEP-WCMC UK
Neil Burgess UNEP-WCMC UK
Mark A. Burgman University of Melbourne Australia
Ben Collen Zoological Society of London UK
Patrick Comer NatureServe USA
Francisco A. Comín Pyrenean Institute of Ecology-CSIC Spain
Franz Essl Environment Agency Austria
Don Faber-Langendoen NatureServe USA
Peter G. Fairweather Flinders University Australia
Ariany García Rawlins Provita Venezuela
Robert J. Holdaway Landcare Research New Zealand
Michael Jennings University of Idaho USA
David A. Keith University of New South Wales Australia
Richard T. Kingsford University of New South Wales Australia
Tytti Kontula Finnish Environment Institute Finland
Rebecca E. Lester Deakin University Australia
Ralph Mac Nally Monash University Australia
Michael A. McCarthy University of Melbourne Australia
Rebecca M. Miller IUCN UK
Justin Moat Royal Botanic Gardens, Kew UK

93 | IUCN Red List of Ecosystems

Name Organisation Country
Nicholas J. Murray University of New South Wales Australia
Emily Nicholson Deakin University Australia
María A. (Tina) Oliveira- Provita Venezuela
Miranda
Phil Pisanu South Australian Department of Environment Australia
and Natural Resources
Brigitte Poulin Tour du Valat Research Center France
Tracey J. Regan Arthur Rylah Institute Australia
Belinda Reyers Council for Scientific and Industrial Research South Africa
Uwe Riecken German Federal Agency for Nature Germany
Conservation
Jon Paul Rodríguez Instituto Venezolano de Investigaciones Venezuela
Científicas
Kathryn M. Rodríguez- Instituto Venezolano de Investigaciones Venezuela
Clark Científicas
Mark D. Spalding The Nature Conservancy and University of UK
Cambridge
Andrew Taber The Wildlife Trust USA
Irene Zager Provita Venezuela
Sergio Zambrano- Instituto Venezolano de Investigaciones Venezuela
Martínez Científicas
Carlos Zambrana-Torrelio EcoHealth Alliance USA
Tara Zamin Monash University Australia

94 | IUCN Red List of Ecosystems

Appendix 2. IUCN Red List of
Ecosystems Criteria, Version 2.2
A. Reduction in geographic distribution over ANY of the following time periods:
A2b. any 50 year
A1. the past 50 years A2a. the next 50 years period including the A3. since 1750
past, present and future
CR ≥ 80% ≥ 80% ≥ 80% ≥ 90%
EN ≥ 50% ≥ 50% ≥ 50% ≥ 70%
VU ≥ 30% ≥ 30% ≥ 30% ≥ 50%

B. Restricted geographic distribution indicated by ANY OF B1, B2 or B3:

B1. Extent of a minimum convex polygon enclosing all occurrences (extent of occurrence, EOO) is no larger than:
(a) An observed or inferred continuing decline in ANY OF:
i. a measure of spatial extent appropriate to the ecosystem; OR
ii. a measure of environmental quality appropriate to characteristic biota of the
AND at ecosystem; OR
least one of iii. a measure of disruption to biotic interactions appropriate to the characteristic
CR 2,000 km2 the biota of the ecosystem.
following
(b) Observed or inferred threatening processes that are likely to cause continuing
(a-c):
declines in geographic distribution, environmental quality or biotic interactions within
the next 20 years.
(c) Ecosystem exists at 1 threat-defined location
(a) An observed or inferred continuing decline in ANY OF:
i. a measure of spatial extent appropriate to the ecosystem; OR
ii. a measure of environmental quality appropriate to characteristic biota of the
AND at ecosystem; OR
least one of iii. a measure of disruption to biotic interactions appropriate to the characteristic
EN 20,000 km2 the biota of the ecosystem.
following
(b) Observed or inferred threatening processes that are likely to cause continuing
(a-c):
declines in geographic distribution, environmental quality or biotic interactions within
the next 20 years.
(c) Ecosystem exists at ≤ 5 threat-defined locations
(a) An observed or inferred continuing decline in ANY OF:
i. a measure of spatial extent appropriate to the ecosystem; OR
ii. a measure of environmental quality appropriate to characteristic biota of the
AND at
ecosystem; OR
least one of
iii. a measure of disruption to biotic interactions appropriate to the characteristic
VU 50,000 km2 the
biota of the ecosystem.
following
(a-c): (b) Observed or inferred threatening processes that are likely to cause continuing
declines in geographic distribution, environmental quality or biotic interactions within
the next 20 years.
(c) Ecosystem exists at ≤ 10 threat-defined locations

95 | IUCN Red List of Ecosystems

 B2. The number of 10 × 10 km grid cells occupied (area of occupancy, AOO) is no more than:
(a) An observed or inferred continuing decline in ANY OF:
i. a measure of spatial extent appropriate to the ecosystem; OR
ii. a measure of environmental quality appropriate to characteristic biota of the
AND at ecosystem; OR
least one of iii. a measure of disruption to biotic interactions appropriate to the characteristic
CR 2 the biota of the ecosystem.
following
(b) Observed or inferred threatening processes that are likely to cause continuing
(a-c):
declines in geographic distribution, environmental quality or biotic interactions within
the next 20 years.
(c) Ecosystem exists at 1 threat-defined location
(a) An observed or inferred continuing decline in ANY OF:
i. a measure of spatial extent appropriate to the ecosystem; OR
ii. a measure of environmental quality appropriate to characteristic biota of the
AND at ecosystem; OR
least one of iii. a measure of disruption to biotic interactions appropriate to the characteristic
EN 20 the biota of the ecosystem.
following (b) Observed or inferred threatening processes that are likely to cause continuing
(a-c): declines in geographic distribution, environmental quality or biotic interactions within
the next 20 years.
(c) Ecosystem exists at ≤ 5 threat-defined locations

(a) An observed or inferred continuing decline in ANY OF:

i. a measure of spatial extent appropriate to the ecosystem; OR
ii. a measure of environmental quality appropriate to characteristic biota of the
AND at ecosystem; OR
least one of iii. a measure of disruption to biotic interactions appropriate to the characteristic
VU 50 the biota of the ecosystem.
following (b) Observed or inferred threatening processes that are likely to cause continuing
(a-c): declines in geographic distribution, environmental quality or biotic interactions within
the next 20 years.
(c) Ecosystem exists at ≤ 10 threat-defined locations

B3. The number of threat-defined locations is:

Very small (generally fewer than 5 threat-defined locations) AND prone to the effects of human activities or
VU stochastic events within a very short time period in an uncertain future, and thus capable of Collapse or becoming
Critically Endangered within a very short time period (B3 can only lead to a listing as VU).

C. Environmental degradation over ANY of the following time periods:

Relative severity (%)

Extent (%) ≥ 80 ≥ 50 ≥ 30
C1. The past 50 years, based on change in an abiotic variable affecting
a fraction of the extent of the ecosystem and with relative severity, as ≥ 80 CR EN VU
indicated by the following table:
≥ 50 EN VU

≥ 30 VU

C2a. The next 50 years, based on change in an abiotic variable affecting Relative severity (%)
a fraction of the extent of the ecosystem and with relative severity, as Extent (%) ≥ 80 ≥ 50 ≥ 30
indicated by the following table; OR
≥ 80 CR EN VU
C2b. Any 50-year period including the past, present and future, based on
change in an abiotic variable affecting a fraction of the extent of the ≥ 50 EN VU
ecosystem and with relative severity, as indicated by the following table: ≥ 30 VU

Relative severity (%)

Extent (%) ≥ 90 ≥ 70 ≥ 50
C3. Since 1750, based on change in an abiotic variable affecting a
fraction of the extent of the ecosystem and with relative severity, as ≥ 90 CR EN VU
indicated by the following table:
≥ 70 EN VU
≥ 50 VU

96 | IUCN Red List of Ecosystems

D. Disruption of biotic processes or interactions over ANY of the following time periods:

Relative severity (%)

Extent (%) ≥ 80 ≥ 50 ≥ 30
D1. The past 50 years, based on change in a biotic variable affecting a
fraction of the extent of the ecosystem and with relative severity, as ≥ 80 CR EN VU
indicated by the following table:
≥ 50 EN VU

≥ 30 VU

D2a. The next 50 years, based on change in a biotic variable affecting a Relative severity (%)
fraction of the extent of the ecosystem and with relative severity, as Extent (%) ≥ 80 ≥ 50 ≥ 30
indicated by the following table; OR
≥ 80 CR EN VU
D2b. Any 50-year period including the past, present and future, based on
change in a biotic variable affecting a fraction of the extent of the ≥ 50 EN VU
ecosystem and with relative severity, as indicated by the following table: ≥ 30 VU
Relative severity (%)
Extent (%) ≥ 90 ≥ 70 ≥ 50
D3. Since 1750, based on change in a biotic variable affecting a fraction
of the extent of the ecosystem and with relative severity, as indicated by ≥ 90 CR EN VU
the following table:
≥ 70 EN VU

≥ 50 VU

E. Quantitative analysis that estimates the probability of ecosystem collapse to be:

CR ≥ 50% within 50 years
EN ≥ 20% within 50 years
VU ≥ 10% within 100 years

97 | IUCN Red List of Ecosystems

Appendix 3. Colour codes

98 | IUCN Red List of Ecosystems

Appendix 4. Revision history

Version Number Revisions Date

1.0 Guidelines v1.0 launched 27/10/2016
1.1 Updated guidelines to include new research 16/06/2017
published between v1.0 and v1.1
New section: 3.3.4 Making the most of
quantitative data and expert knowledge
New Fig. 8 and Fig. 9
Incorporated recent research on performance
of range size measures
Change of language from ‘location’ to ‘threat-
defined location’ to avoid ambiguity with locality

99 | IUCN Red List of Ecosystems

INTERNATIONAL UNION
FOR CONSERVATION OF NATURE

WORLD HEADQUARTERS
Rue Mauverney 28
1196 Gland, Switzerland
Tel +41 22 999 0000
Fax +41 22 999 0002
www.iucn.org