EPA/600/R-09/011
February 2009
A Framework for Categorizing the Relative Vulnerability of
Threatened and Endangered Species to Climate Change
Hector Galbraith, Manomet Center for Conservation Sciences,
Dummerston, Vermont
Jeff Price, World Wildlife Fund, Washington, DC
Global Change Research Program
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC 20460
-------
DISCLAIMER
This document has been reviewed in accordance with U.S. Environmental Protection
Agency policy and approved for publication. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.
ABSTRACT
This report describes an evaluative framework that may be used to categorize the relative
vulnerability of species to climate change. Four modules compose this framework: Module 1
categorizes baseline vulnerability to extinction or major population reduction by scoring those
elements of the species' life history, demographics, and conservation status that influence the
likelihood of its survival or extinction (excluding climatic changes); Module 2 scores the likely
vulnerability of a species to future climate change, including the species' potential physiological,
behavioral, demographic, and ecological response to climate change; Module 3 combines the
results of Modules 1 and 2 into a matrix to produce an overall score of the species' vulnerability
to climate change, which maps to an adjectival category, such as "critically vulnerable", "highly
vulnerable", "less vulnerable", and "least vulnerable"; Module 4 is a qualitative determination of
uncertainty of overall vulnerability (high, medium, and low) based on evaluations of uncertainty
done in each of the first 3 modules. To illustrate the use of this framework, it was applied to five
U.S. threatened and endangered species and one species that has since been delisted. Based on
the authors' evaluation, four of those species were categorized as "critically vulnerable": the
golden-cheeked warbler (Dendroica chrysoparia), the salt marsh harvest mouse
(Reithrodontomys raviventris), the Mount Graham red squirrel (Tamiasciurus hudsonicus
grahamensis), and the Lahontan cutthroat trout (Oncorhyncus clarki henshawi). The desert
tortoise (Gopherus agassizii) was characterized as "highly vulnerable" and the bald eagle
(Haliaeetus leucocephalus) — now delisted, except for the southwest population — was
categorized as "less vulnerable". Certainty scores in Module 4 ranged between medium and high
and reflect the amount and quality of information available.
li
-------
Preferred citation:
U.S. Environmental Protection Agency (EPA). (2009) A framework for categorizing the relative vulnerability of
threatened and endangered species to climate change. National Center for Environmental Assessment, Washington,
DC; EPA/600/R-09/011. Available from the National Technical Information Service, Springfield, VA, and online at
http ://www. epa. gov/ncea.
ill
-------
CONTENTS
LIST OF TABLES V
PREFACE VI
EXECUTIVE SUMMARY VII
1. INTRODUCTION 1
2. OVERARCHING ISSUES 4
2.1. FRAMEWORK OBJECTIVES 4
2.2. IMPORTANT FRAMEWORK ATTRIBUTES 4
2.3. INFORMATION NEEDS AND SOURCES FOR FRAMEWORK 6
2.4. DIRECT AND INDIRECT EFFECTS OF CLIMATE CHANGE 11
3. FRAMEWORK GENERAL STRUCTURE 13
4. THE NARRATIVES 15
5. MODULE 1—EVALUATING BASELINE VULNERABILITY 16
5 .1. SCORING MODULE 1 VARIABLES 16
5.2. MODULE 1—CERTAINTY EVALUATION 21
6. MODULE 2—EVALUATING VULNERABILITY TO CLIMATE CHANGE 23
6.1. SCORING MODULE 2 VARIABLES 26
6.2. MODULE 2—CERTAINTY EVALUATION 29
7. MODULE 3—EVALUATING OVERALL VULNERABILITY 30
8. MODULE 4—CERTAINTY EVALUATION 32
9. SUMMARY 01 FRAMEWORK EVALUATIONS 33
10. SUMMARY AND CONCLUSIONS 36
REFERENCES 37
APPENDIX A EXAMPLE NARRATIVE FOR GOLDEN-CHEEKED WARBLER 42
APPENDIX B EXAMPLE NARRATIVE FOR BALD EAGLE 52
APPENDIX C EXAMPLE NARRATIVE FOR SALT MARSH HARVEST MOUSE 59
APPENDIX D EXAMPLE NARRATIVE FOR MOUNT GRAHAM RED SQUIRREL 66
APPENDIX E EXAMPLE NARRATIVE FOR DESERT TORTOISE 74
APPENDIX F EXAMPLE NARRATIVE FOR LAHONTAN CUTTHROAT TROUT 86
APPENDIX G 94
MODULE 1 - CATEGORIZING THE "BASELINE" VULNERABILITIES (VB) FOR
EXAMPLE SPECIES 94
APPENDIX H 101
MODULE 2 - CATEGORIZING THE VULNERABILITIES TO CLIMATE CHANGE (VC)
FOR EXAMPLE SPECIES 101
APPENDIX I 107
MODULE 3 - COMBINING BASELINE AND CLIMATE VULNERABILITY SCORES INTO
OVERALL VULNERABILITY SCORES (VO) FOR EXAMPLE SPECIES 108
APPENDIX J Ill
MODULE 4 - CERTAINTY/UNCERTAINTY ANALYSIS FOR EXAMPLE SPECIES Ill
iv
-------
LIST OF TABLES
1 Variables included in Module 1 16
2 Module 1 variables and scores used in categorizing the "baseline" vulnerabilities
(Vb) of T&E species 17
3 Components of species' potential physiological, behavioral, demographic, and
ecological sensitivity to climate change included as variables in Module 2 23
4 Module 2 variables and scores used in categorizing the vulnerabilities of T&E
species to climate change (Vc) 24
5 Module 3—Overall vulnerability best estimate scoring matrix 30
6 Module 1, 2, and certainty scores for the six test species 33
7 Summarization of results of species evaluations 34
v
-------
PREFACE
This report was prepared by Hector Galbraith of Manomet Center for Conservation
Sciences and Jeff Price of World Wildlife Fund. Review, comments and general oversight of this
work were provided by the Global Change Research Program (GCRP) in the National Center for
Environmental Assessment (NCEA), U.S. Environmental Protection Agency (U.S. EPA). This
report presents a framework for evaluating the current and future vulnerability of threatened and
endangered animal species to existing stressors and to potential future climatic changes. Results
are intended to be regarded as indications of the comparative vulnerabilities of species to climate
change, not estimates of a species' absolute vulnerability.
The report has undergone peer consultation and external peer review, including review of
the first draft in 2002 by the U.S. Fish and Wildlife Service within the Department of Interior.
Changes and edits made between the draft report and this final report posted for public comment
reflect edits made to respond to expert reviewers during the peer consultation and external peer
review process. When publishing the final report, EPA will consider any public comments
received the public comment period.
EPA's Global Change Research is an assessment-oriented program committed to
developing frameworks and tools to assist decision-makers in evaluating the impacts of climate
change to air quality, water quality and ecosystems. This framework is offered as one of a
number of potential approaches for determining species' relative vulnerability to climate change.
It is not intended to serve as a tool for determining whether a species is endangered or threatened
under the Section 4 listing process of the Endangered Species Act. It is also not intended to be
used by federal or state agencies for the determination of whether specific actions cause a
"taking" of any listed species of endangered fish or wildlife under the Endangered Species Act.
This framework is intended to provide information to ecosystem and resource managers to
support their decision making about management actions that reflect consideration of those
threatened and endangered species that are most vulnerable to climate change. This framework
also may be helpful in supporting management decisions related to species not listed as
threatened or endangered.
vi
-------
EXECUTIVE SUMMARY
Organisms listed as threatened or endangered under the Endangered Species Act (ESA)
of 1973 are at risk of extinction due to adverse effects of current natural or anthropogenic
stressors (e.g., habitat loss, contaminants, and competition with invasive species). Climate
change and variability, acting alone or exacerbating current stressors, may constitute an
important new threat for many threatened and endangered (T&E) species. Evaluative tools that
account for climate change impacts are being developed for use by resource managers as they
become more aware of the effects of climate change. This report describes the development of an
evaluative framework to categorize the vulnerability of species to climate change. The
framework is then applied to six species that were listed as threatened or endangered at the time
the framework was developed to illustrate its use in categorizing the vulnerability of these
species to climate change.
This framework for evaluating vulnerabilities to climate change comprises four modules.
Module 1, which includes 11 variables, categorizes the comparative vulnerabilities to existing
stressors, not including climate change. Likely, baseline vulnerability to extinction or major
population reduction is categorized by scoring those elements of the species' life history,
demographics, and conservation status that influence the likelihood of its survival or extinction
regardless of climate change. Module 2, consisting of 10 variables, scores the likely vulnerability
of a species to future climate change. Specifically, the species' potential response to
physiological (e.g., temperature, precipitation), behavioral, demographic, and ecological
sensitivity to climate change are the elements of this module. Additionally, each variable in
Modules 1 and 2 is assigned a "best estimate" certainty score that results in a subjective
confidence statement. Module 3 combines the results of Modules 1 and 2 into a matrix to
produce an overall evaluation and a score of the species' vulnerability to climate change. The
numerical scores are then grouped into adjectival categories: "critically vulnerable", "highly
vulnerable", "less vulnerable", and "least vulnerable". Module 4 is a qualitative scoring of
uncertainty based on the evaluations from the first 3 modules resulting in an index of certainty
(high, medium, and low) associated with the overall vulnerability score from Module 3.
The framework was applied to threatened and endangered species listed under the U.S.
ESA. The golden-cheeked warbler (Dendroica chrysoparia), salt marsh harvest mouse
(Reithrodontomys raviventris), Mount Graham red squirrel (Tamiasciurus hudsonicus
grahamensis), and the Lahontan cutthroat trout (Oncorhyncus clarki henshawi) were categorized
as "critically vulnerable." The desert tortoise Gopherus agassizii was ranked "highly
vulnerable," and the bald eagle Haliaeetus leucocephalus (no longer listed as threatened or
vii
-------
endangered, except for the southwest population) was scored "less vulnerable." Certainty scores
in Module 4 ranged between medium and high and reflect the amount and quality of information
available.
Species that are most vulnerable tend to be: restricted in their distributions, small in
population size, undergoing population reductions, habitat specialists, and found in habitats that
are likely to be most adversely affected by future climate change. Conversely, species like the
bald eagle, which are widely distributed, are flexible in their habitat preferences and are
considered to be stable or increasing, scored least vulnerable. Thus, the predictions of the model
are consistent with what might be expected based on the ecologies and demographics of the test
species. The results also indicate that major areas of uncertainty complicate any evaluations of
vulnerability. For the species tested, the greatest uncertainties are associated with our relatively
poor knowledge about the potential for direct, physiological effects on animal species;
relationships between changes in temperature and precipitation regimes and the physiologies and
behaviors of animals are apparently only poorly understood.
-------
1. INTRODUCTION
Organisms listed as Threatened or Endangered (T&E species) under the Endangered
Species Act (ESA) of 1973 (16 U.S.C. 15631 et seq.) are at risk of extinction due to the adverse
effects of current natural or anthropogenic stressors (e.g., habitat destruction, contaminants,
interactions with invasive species, etc.). Climate change, either acting alone or by exacerbating
the effects of these current stressors, may constitute an important new threat for many of these
species (Peters, 1992; Tucker and Heath, 1994; Schneider and Root, 2002; Walther et al., 2002).
If future conservation priorities, strategies, tactics, and resource allocations are to reflect these
changing circumstances, there is a need to develop new tools. In particular, tools are needed that
integrate the likely effects of both current and climate change stressors to identify those T&E
species that may face the greatest increased risks of extinction or major population reductions,
and the specific climatic, physiological, and/or ecological factors that contribute to these
increased risks. This report describes an analytical framework that is intended to rank T&E
animal species in terms of these current and future risks and potential causal factors.
The primary purpose of the ESA is to "provide for the conservation of endangered and
threatened species of fish, wildlife, and plants..." (ESA of 1973, 16 U.S.C. 15631 et seq.).
Animals typically are listed under the ESA when their population sizes or distributions become
so small or limited that their continued existence may be in jeopardy (or at least approaching
such a condition). Some T&E animals may always have had extremely restricted distributions or
small populations (e.g., some desert fish species or cave-dwelling amphibians or arachnids), and
they have been listed as a safeguard against possible future habitat destruction. Other listed
species, however, that may have been more widespread and abundant in the past have been so
reduced in range or numbers that their continued existence may be in jeopardy. In most such
cases, the population/range reduction has been due to anthropogenic stressors, particularly
habitat destruction.
Regardless of why an animal species was placed on the T&E list, its presence there
implies that its future existence may be in jeopardy. Into this already tenuous situation, a new
stressor, climate change, has now been introduced. This raises questions that are important in the
conservation, scientific, and regulatory arenas. For example, how might climate change affect
the already threatened existence of many T&E animal species; what particular aspects of climate
change may be important for individual species, and how will they affect them; which species
are likely to be most vulnerable; might some T&E species benefit from climate change; can we
mitigate the effects of climate change for any species (e.g., through habitat manipulation,
1
-------
translocation of organisms, removal of other stressors, etc.); and, lastly, do our current
conservation approaches require modification in the light of the likely effects of climate change?
As a first step toward addressing these questions, it is necessary to be able to categorize
each T&E animal species in the United States in terms of its likely relative vulnerability to
climate change, assess what its responses might be, and identify the causal factors likely to be
most important (either due to the direct effects of climate on the organisms themselves, or
indirect effects acting through their environment). This report presents the results of an attempt
to develop an evaluative framework that can be used to assess the relative vulnerabilities of T&E
animal species to climate change and address these issues. It details the structure of the proposed
framework and tests it on six species that were listed as threatened or endangered at the time the
framework was developed: the bald eagle, Haliaeetus leucocephalus (no longer listed),
golden-cheeked warbler, Dendroica chrysoparia, salt marsh harvest mouse, Reithrodontomys
raviventris, Mount Graham Red Squirrel Tamiasciurus hudsonicus grahamensis, desert tortoise,
Gopherus agassizii, and the Lahontan cutthroat trout, Oncorhyncus clarki henshawi. These
species were selected because they are very different in their natural histories, demographies,
status and distribution, population trends, and susceptibilities to different stressors, and, because
of these differences, may provide an adequate preliminary test of the framework. This report
does not provide a finalized framework but rather describes a proposed framework for discussion
and future refinement. Therefore, as the framework is developed further, it should be tested
against additional species.
At least three previous studies have attempted to categorize animal species in terms of
their population vulnerabilities: the International Union for the Conservation of Nature (IUCN)
developed a system for scoring the conservation status of organisms worldwide (Mace and
Stuart, 1994). This was the method underpinning the development of the IUCN's Red List
categorizations and was adopted by Birdlife International to assess the conservation of wild bird
species in Europe (Tucker and Heath, 1994) and to identify birds at risk worldwide (Collar et al.,
1994). In the United States, Partners in Flight have developed a framework to categorize the
conservation status and vulnerability of landbirds (Carter et al., 2000). This system was
subsequently the basis of the Watchlist of North American birds published by the National
Audubon Society. Neither of these methods attempts to predict the potential incremental effects
of climate change on species' future vulnerabilities and are, therefore, not suitable for projecting
future risks. However, Herman and Scott (1994) attempted to do so by developing a scoring
framework to evaluate the risks posed by future climate change to vertebrates in Nova Scotia,
Canada. Herman and Scott (1994) did not include risks posed by existing nonclimate stressors.
The framework developed in this study incorporates many of the concepts and components of
2
-------
these earlier studies and extends them so as to be able to predict the future risks posed by
existing stressors and climate change acting on organisms.
Section 2 identifies and discusses some overarching considerations that are relevant to
the construction of any evaluative framework. Section 3 describes the general structure of the
framework. Sections 4 through 8 describe in greater detail the specific components of the
framework. Section 9 summarizes the results of the tests of the framework on the six test species
and Section 10 presents the major conclusions of this process. Appendices A through J provide
example narratives and applications of Modules 1 through 4 to six selected species.
It should be noted that while this framework will help in evaluating the likely risks of
climate change to T&E species in the United States, the information generated is intended to be a
guide to how the future vulnerability of organisms might change. It should not be used alone to
provide a mechanism for determining whether a species is endangered or threatened under the
Section 4 listing process of the Endangered Species Act. To do so would be a misuse of the
framework's intended purpose.
3
-------
2. OVERARCHING ISSUES
In this section, some overarching issues important in evaluating species' vulnerabilities
to climate change are identified and discussed. Sections 3 through 8 demonstrate how these
issues are incorporated into the proposed framework structure.
2.1. FRAMEWORK OBJECTIVES
To provide information that will be useful in addressing the questions in Section 1 of this
report and assist conservationists and regulators in formulating conservation strategies and
policies, an effective predictive framework will have to:
(1) characterize and rank the current (i.e., non-climate change) vulnerabilities of T&E
species in a consistent fashion;
(2) characterize the potential effects of climate change on a species' vulnerability;
(3) integrate the results of (1) and (2) into an overall evaluation of potential future
vulnerability;
(4) evaluate the risks and potential magnitudes of population and distributional change;
(5) identify specific climate change causal factors that may contribute to these changes and
their relative importance;
(6) evaluate uncertainties associated with Steps 1 through 5; and
(7) identify data needs for species for which uncertainty is high.
Climate change may already have affected some T&E organisms in the United States
(Parmesan and Galbraith, 2005). To the extent that such changes are recognized and
incorporated into existing distribution, population size, and habitat estimates, they are included
and evaluated in this framework as "baseline" conditions. The primary purpose of this
framework, unlike current approaches, is to evaluate potential consequences offuture climate
change.
2.2. IMPORTANT FRAMEWORK ATTRIBUTES
Process Transparency. The intended result of this framework is to produce evaluations of
the relative vulnerabilities of T&E species to climate change and other stressors. The focus of
this framework is on evaluating vulnerabilities—not predicting risks to T&E species. T&E
species were chosen to use as examples because these species generally have sufficient data to
implement the framework. It is equally important that the process and reasoning through which
the evaluation was arrived at be well documented and transparent. This will be essential in
4
-------
modifying species evaluations if new data are gathered that cast doubt on previous assessments.
Ensuring process transparency and documenting important assumptions are as important
components of the framework as producing predictive scores.
Framework Precision and Accuracy. By their nature, the results of a predictive
framework will involve speculation (we cannot be entirely confident about how an organism will
respond to future stressors that may not be adequately understood). Thus, in the absence of a
posteriori knowledge, this framework provides approximations of species' ranked
vulnerabilities. It is not intended that results be considered completely accurate or precise
estimations of a species' absolute vulnerability—the results should be regarded as indications of
the comparative vulnerabilities of T&E species as represented by the species evaluated.
Also, not all species may be adversely affected by climate change; it is possible some
may benefit from new climatic regimes (for example, due to their habitats being expanded, or to
their competitors or predators being adversely affected). It is important, therefore, that the
evaluative framework allows for this possibility in the range of species' responses.
Treatment of Certainty/Uncertainty. Uncertainty is inevitable in any predictive
framework that attempts to anticipate specific effects of future stressors on organisms. Such
uncertainty may have many sources, including the specifics or variability of likely future
climates, the physiological sensitivity of the species, uncertainty about its demographics,
population dynamics, or habitat ecology, or about the likely responses of habitats, or critical
habitat components, to climate change. Any prediction regarding future vulnerability would be
of limited practical value without an evaluation of the certainty/uncertainty associated with it. In
this framework, the degree of certainty is assessed in two ways: first, when scoring each module
variable, "best estimate" and alternate (possible but less likely) scores are assigned. These are
intended to capture the range of responses that may occur, rather than focusing on a single "point
estimate" of responses. Second, each individual variable score is assigned a ranked certainty
evaluation (i.e., high, medium, or low level of certainty). This 3-point ranking is based on the
5-category scale developed for the Intergovernmental Panel on Climate Change (IPCC) Third
Assessment Report (Moss and Schneider, 2000). These rankings are then combined into an
assessment of the degree of certainty that should be associated with the final assessment of the
species' overall vulnerability. For most species, these certainty scores will not be based on
quantitative evidence, but on the judgment of experts in the species' ecology, conservation,
and/or demographics.
Sources of Information and Expert Opinion. Some of the scores determined in this
framework may be based on quantitative and empirical data (e.g., abundance estimates based on
actual census data) published in peer-reviewed scientific or other report literature. However, for
5
-------
many less well studied species, it is likely that many of the framework scores will be based not
on actual empirical data, but will comprise rankings (Siegel, 1956) based on expert opinion. In
this context, expert opinion is defined as the professional judgment of one or more experts in the
species or, failing that, ecologically comparable members of its taxonomic group (up to and
including the Family). If expert opinion is the main source of a score, its argument, underlying
assumptions, and the evidence that supports the opinion must be clearly stated in the species'
narrative section.
For some species there may only be a small number of experts; for others there may be
comparatively many. If expert opinion is to constitute the majority of a species' scores, and if a
number of experts can participate, some version of a Delphi approach (Linstone and Turoff,
1975; Zuboy, 1980, 1981; Crance, 1985) might be used to formalize and record their opinions.
It should be noted that the main role of the experts will be in helping to evaluate species'
framework scores, based on their expert knowledge, i.e., in the application of the finalized
framework. This paper concentrates on developing the framework. Thus, the species evaluations
provided to test this framework should not be considered definitive statements about each
species but as examples of applying the framework.
2.3. INFORMATION NEEDS AND SOURCES FOR FRAMEWORK
To meet the performance standards identified in Section 2.1 and thereby realistically
evaluate the likely responses of a T&E organism to climate change, the following categories of
information will be useful:
Physiological information
> its likely physiological vulnerability to potential changes in temperature
> its likely physiological vulnerability to potential changes in precipitation
> the likelihood of its physiological/behavioral adaptation to climate change
Demographic/life history information
> the organism's population/sub-population abundance relative to extinction risk
> the factors currently limiting its distribution/population status
> the degree to which the organism's geographical distribution is localized or
dispersed
> its past and current population/sub-population trends
> its potential dispersive ability
> its ability to recover quickly from population reductions
6
-------
> the likely vulnerabilities of populations to fluctuations in climatic variability and
severe weather events
> its interactions with competitors, predators, and pathogens
Habitat information
> the habitats needed to meet all of the organism's life history requirements
> its degree of habitat specialization
> limiting habitat components and their likely sensitivities to climate change
> current trends in the availability of preferred habitats
> the likely vulnerabilities of its main habitats to climate change
> the extent to which suitable habitats may be present within the species' new range
> the abilities of its main habitats to migrate in response to climate change
> the likely rates at which the species' habitats could migrate relative to its
physiological tolerances
> the likely vulnerabilities of habitats to climatic variability and severe weather
events
Phenological information
> the likelihood that phenological relationships between crucial events in the
species' life cycle (e.g., timing of breeding) and in its environment (e.g., snow
melt) could be disrupted
Stressor information
> the direction and magnitude of likely climate change factors that may affect the
organism
> other anthropogenic/natural stressors that may currently be affecting the organism
and how their intensities are changing, or are likely to change in the future as
humans respond to climate change
Our ability to evaluate the likely effects of climate change on T&E taxa will be a function
of the quantity and quality of the data in each of the above categories. However, in addition to
categorizing the likely vulnerabilities of taxa, the framework also must be able to identify crucial
data needs for relatively little-known T&E organisms. Thus, missing information for any one
taxon does not necessarily mean that it should not be evaluated, only that the uncertainty
associated with the conclusions should be recognized and stated.
7
-------
A number of sources exist for the above categories of information:
Information about the direct relationships between ambient temperature and the species'
physiology, and its potential ability to persist
Specifically, what are the average, transient, or maximum long- or short-term
temperatures above which the species is likely to suffer acute or chronic effects, such
as impairment of reproduction or survival, physiological malfunctions, etc.? Such
information could be gleaned from two main sources:
a) Ideally, such information should be derived from experimental physiological
studies of the species being evaluated. However, such experimental studies have
been carried out on relatively few species (especially terrestrial species), and no
such studies of the six species evaluated in this report have been found.
Furthermore, where thermal stress studies have been performed, the experimental
endpoint is most often a gross measure of the species' vulnerability, such as the
temperature that results in the death of a substantial part of the experimental
population. In the field, organisms would almost certainly begin to respond well
before such acute temperatures are reached. Thus, while acute mortality studies
may provide information on individuals' ultimate temperature tolerances, they
may only be of limited relevance regarding the temperature regimes that may
govern a species' distribution or abundance in the field.
Some experimental studies have been carried out on more subtle responses to
temperature change. These include studies of the behavioral responses and
thermal habitat choices of organisms. Such studies might provide more relevant
information for assessing the likely effects of climate change on species'
distributions. However, no such studies on T&E species have been located. If
available, information for surrogate species (i.e., organisms that are closely
related to the species under investigation and that are morphologically and
ecologically similar) could be substituted.
b) Valuable information also can be obtained by examining the current and past
ranges of organisms. For example, if the southernmost edge of an organism's
current or historical range stops substantially north of the southern limit of its
habitat type, then it could be directly climate-limited. It might, therefore, be
reasonable to assume that some climate metric at the southern edge of its range is
a limiting factor. However, what should be concluded in cases where the
organism's range matches that of its main habitat? In such cases, the species
could either be habitat- or climate-limited (or both). One way of addressing this
problem is to examine the habitats and ranges of closely-related species. For
example, the ranges of some North American Dendroica warblers suggest that
they may be climate-limited (e.g., the upland conifer forest breeding habitats of
Townsend's and hermit warblers [Dendroica townsendi and Dendroica
occidentalism respectively] extend south through the western states and into
Mexico, yet the two species do not breed any farther south than central
California). Perhaps temperature or precipitation is limiting these two species. If
the breeding habitat of the closely related golden-cheeked warbler, Dendroica
8
-------
chrysoparia, (a T&E species), extends south of its distribution (southern Texas),
it might be reasonable to conclude that this species also may be climate-limited.
In performing such analyses, it is important to consider both current and historical
distributions. For example, the current range of grizzly bears extends south from above the
Arctic Circle in Alaska and northern Canada to the northern Rocky Mountain States, and east
from the Alaska Peninsula to the western shore of Hudson's Bay (Craighead and Mitchell,
1982), covering over 25° of latitude and 70° of longitude. This may demonstrate a high degree of
overall climatic flexibility on the part of this species. However, this flexibility and tolerance of
widely different temperature and precipitation regimes becomes even more marked when the
historical range of the species is considered; in pre-Colombian times, the species' range extended
south into northern Mexico and from the Pacific coast, east to the Missouri River (Rausch,
1963), and from low-lying deserts to alpine tundra. Thus, up until 200 years before the present,
grizzly bears could be found across over 40° of latitude and 70° of longitude, and from close to
sea level to above 10,000 feet, with associated widely differing climatic conditions. From its
current and historical distribution, it could be surmised that future climate change is likely to
have relatively small direct effects on grizzly bears in areas where they still persist. The
information from this type of historical analysis should be treated with caution, however, since
species with previously wide distributions may have consisted of different genotypes each
adapted to specific climatic conditions.
When determining if a species may be climate-limited in its distribution and the extent to
which it may be directly affected by future climate change, the following procedure might be
adopted:
1. Determine whether there is evidence from experimental studies that the study species
(or closely-related and morphologically and ecologically similar species) is likely to
be affected by future climatic factors (e.g., do likely future temperature regimes
exceed those to which the species [or a surrogate] has been experimentally shown to
be sensitive?).
2. If the information required for Step 1 is not available, determine if the species'
habitat extends beyond its actual range, and into areas where the climatic conditions
exceed those within the species' actual range. If so, the extremes of the actual range
might be climatic limits on the species' distribution. In this step, care should be taken
to identify the extent to which biogeographical barriers (e.g., cities or waterbodies)
might be preventing a species from occupying the whole of its potential current
range.
3. If information for the study species is not available to perform Steps 1 or 2, carry out
Step 2 for closely-related (congeners) and morphologically and ecologically similar
species.
9
-------
Information about the species' distribution. For many vertebrate species listed under the
ESA, there is a wealth of accurate information on current distributions. This is particularly the
case where the species is terrestrial, diurnal, and restricted to relatively small areas (e.g., the
Mount Graham red squirrel \Tamiasciurus hudsonicus grahamensis] is known to be confined to
one mountain range in southern Arizona, or Kirtland's warbler [Dendroica kirtlandii\ largely
confined to a few counties in the Lower Peninsula of Michigan). Although the exact range
boundaries of more widespread T&E species may be less easy to delineate, for many taxa
(particularly birds and the larger mammals), approximate range boundaries (to about the closest
100 km) are relatively well known and published as maps or text descriptions in a large number
of sources, ranging from national distribution maps in field guides and atlases (e.g., Root, 1988;
Price et al., 1995; Kaufman, 1996; Dunn and Garrett, 1997; National Geographic Society, 1999;
Sibley, 2000; individual species accounts by various authors in The Birds of North America
series from the Philadelphia Academy of Natural Sciences [birds]; Burt and Grossenheider,
1964; Whitaker, 1980; Chapman and Feldhamer, 1982; Wilson and Ruff, 1999 [mammals]), to
state and local atlas reports (e.g., Temple and Cary, 1987; Laughlan and Kibbe, 1985; Andrews
and Righter, 1992; Bergeron et al., 1992; Veit and Petersen, 1993; Kingery, 1998 [birds]; Ingles,
1965; Baker, 1983; Merrit, 1987; Jameson and Peeters, 1988; Knox Jones and Birney, 1988;
Zevelof, 1988; Caire et al., 1989; Hoffmeister, 1989; Choate et al., 1994; Fitzgerald et al., 1994;
Whitaker and Hamilton, 1998 [mammals]).
Detailed information regarding the distributions of most freshwater fish species is also
available, ranging from national atlases (e.g., Lee et al., 1982; Boschung et al., 1983), to
state-level treatises (e.g., Trautman, 1981; Cooper, 1983; Tomelleri and Eberle, 1990; Sublette et
al., 1990; Sigler and Sigler, 1996). The distributions of cold-water salmonids that are (or were)
prized quarry species such as greenback cutthroat trout or bull trout are particularly well studied.
Distributional information is generally less well known for the three remaining taxa
(reptiles, amphibians, and insects). However, good data do exist for certain of the more
"charismatic" groups such as snakes, turtles and tortoises, and salamanders (e.g., Webb, 1970;
Minton, 1972; Collins, 1982; Dixon, 1987; Lanoo, 1988; Dundee and Rossman, 1989; Ernest et
al., 1994; Harding, 1997; Conant and Collins, 1998; Petranka, 1998; Hunter et al., 1999). Among
the insects, the best distributional information is for butterflies (Scott, 1986; Shull, 1987; Opler
andMalikul, 1998; Opler and Wright, 1999).
These data are supplemented by the distributional information within the United States
(to the extent that it is known) given in the "Background" and "Distribution and Status" sections
of T&E species listing packages in the Code of Federal Regulations. Additional information may
also be available for T&E species in the Population and Habitat Viability Assessment Reports
10
-------
produced to support Recovery Plans (e.g., Beardmore et al., 1995), and in the Recovery Plans,
themselves (e.g., U.S. FWS, 1992). In general, accurate and easily obtainable data exist that
describe the distributions of many T&E species (from a number of taxa) within the United
States.
Information on the elevational distribution of species may also be of value in predicting
the effects of climate change. For example, white-tailed ptarmigan (Lagopus leucurus) and
grizzly bears (Ursus arctos horribilis) both occur in the Rocky Mountain states. However, the
ptarmigan is confined to land above about 10,000 feet, whereas the grizzly bear can be found
over a much greater range of elevations. Thus, it is likely that climate change may have a more
pronounced effect on the ptarmigan.
Information about the species 'population status. Except for a few very scarce and easily
counted organisms (e.g., grizzly bear, Kirtland's warbler), T&E species population status data
are sparse. However, in developing this predictive framework, it is sufficient to estimate
approximate population size categories such as those used by the IUCN (Collar et al., 1994;
Tucker and Heath, 1994).
Information about the species habitat preferences. General information about a T&E
species' habitat preferences may be obtainable from the sources listed in the last section (field
guides, monographs, listing packages, etc.). Most such sources will only provide information on
the ecotypes used by the organisms (e.g., deciduous forest, conifer forest, tundra, and prairie).
Nevertheless, for most cases, this level of information is sufficient for developing this predictive
framework. For some species, more detailed information may be available in individual species
accounts, monographs, or the supporting text from Habitat Suitability Index models from the
U.S. Fish and Wildlife Service (U.S. FWS) (e.g., Peterson, 1986).
Information on non-climate stressors affecting the species. The identities of the more
important stressors currently affecting T&E species are comparatively well known (e.g., habitat
destruction) and described in the materials produced by the U.S. FWS as part of the listing
process.
2.4. DIRECT AND INDIRECT EFFECTS OF CLIMATE CHANGE
The potential effects of climate change on any organism might be direct (i.e., climate
change factors, such as temperature or precipitation, might exceed the physiological tolerances
of the organism and affect its ability to persist in an area). Climate change may also indirectly
affect the organism. For example, climate change might modify the organism's habitat
composition or structure or the phenology of crucial events (e.g., ice melt or flowering seasons),
thereby affecting the ability of the organism to persist. Such trophic mismatches due to climate
11
-------
change may already be occurring as has been indicated in recent studies of European pied
flycatchers and the emergence times of arboreal caterpillars (Sanz et al., 2003). Climate change
could also indirectly jeopardize a species by conferring advantages on its predators, parasites, or
competitors.
12
-------
3. FRAMEWORK GENERAL STRUCTURE
The proposed framework for evaluating risks to a T&E species due to climate change and
other stressors comprises four connected modules and a narrative (see Figure 1). Module 1
categorizes the comparative vulnerabilities of T&E species to existing stressors (i.e., not
including climate change). This "baseline" vulnerability is subsequently combined with the
categorization in Module 2 (evaluating vulnerability to climate change) into an estimate of
overall future vulnerability in Module 3. Module 4 combines certainty scores from Modules 1
and 2 into an evaluation of the overall degree of certainty that we can assign to the framework
predictions.
The narrative that accompanies each species' evaluation details the rationales and
justifications for the assigned scores in Modules 1 and 2.
INnrrill i ve
IN/1 ocl uIe 2.
Climate Change
>/Lilnerab>il ity
IN/Ioclii le
Overnll V ii I nernl~>i I il y
1 t
IVI ocl ii le
C 'onlldence 1 C vn I mil ion
IVIoclule 1
linseline V n I nernt>i I il >
Figure 1. Framework for evaluating effects of existing stressors and climate change.
13
-------
The framework modules and their scoring categories are discussed in greater detail in
Sections 4 through 8, and examples of their application to the six species are given in
Appendices A through J. The species evaluated were chosen because it was believed that, based
on their natural histories, there is good evidence that they may cover the spectrum of potential
responses of T&E organisms to climate change (e.g., from most susceptible, to least susceptible).
It was not our intention to focus only on the most vulnerable species because that would not have
facilitated the development of a general framework.
14
-------
4. THE NARRATIVES
Most categorizations in Modules 1 through 4 will be based largely on the results of
literature reviews and expert judgment for each species being evaluated. The narrative module of
the framework reports the relevant results of those reviews and opinions, and the justifications
for the individual categorization scores in the modules. Thus, the primary aim of the narratives is
to make transparent the thought processes and assumptions that result in the scores in Modules 1
through 4.
The narratives have three additional important aims:
(1) To identify main sources of uncertainty and those areas where additional data might
reduce uncertainty.
(2) To identify and describe the roles of the main stressors (climate and nonclimate) in
the estimate of vulnerability of the study species.
(3) To qualitatively describe potential population responses of the study species to the
addition of climate change to the already existing stressors, and any resulting change
in extinction risk.
Example narratives for six species, golden-cheeked warbler, bald eagle, salt marsh
harvest mouse, Mount Graham red squirrel, desert tortoise, and Lahontan cutthroat trout are
presented in Appendices A through F, respectively.
15
-------
5. MODULE 1—EVALUATING BASELINE VULNERABILITY
In this module, the likely baseline (i.e., current) vulnerability of the study species to
extinction or major population reduction is categorized by scoring those elements of its ecology,
demographics, and conservation status that influence the likelihood of its survival or extinction
(irrespective of the potential effects of future climate change). This is based on determining
ordinal rankings for 11 Module 1 variables (see Table 1). The scoring of these is described in
greater detail below (see Section 5.1). Treatment of certainty/uncertainty is discussed in
Section 5.2. Six examples of the application of Module 1 for golden-cheeked warbler, bald eagle,
salt marsh harvest mouse, Mount Graham red squirrel, desert tortoise, and Lahontan cutthroat
trout are presented in Appendix G.
Table 1. Variables included in Module 1.
(1) Current population size
(7) Likely current stressor future trends
(2) Population trend in last 50 years
(8) Individual replacement time
(3) Current population trend
(9) Future vulnerability to stochastic events
(4) Range trend in last 50 years
(10) Future vulnerability to policy/management
changes
(5) Current range trend
(11) Future vulnerability to natural stressors
(6) Current (nonclimate) stressors
Each variable is assigned a "best estimate" certainty score, together with an "alternate"
(i.e., possible, but less likely) score(s). This will allow subjective confidence limits to be applied
to the overall framework prediction in Module 3. For some species and variables, there may be
enough confidence underlying the best estimate certainty score that no other score is considered
necessary.
5.1. SCORING MODULE 1 VARIABLES
The variable categorizations and their scores used in Module 1 are presented in Table 2.
16
-------
Table 2. Module 1 variables and scores used in categorizing the "baseline" vulnerabilities (Vb) of T&E species.
Current population size
Score
Range trend in last 50 years
Score
Individual replacement time
Score
<100
1
>80% reduction
1
>5 years
1
100-500
2
>50% reduction
2
2-5 years
2
500-1,000
3
>20% reduction
3
<2 years
3
1,000-10,000
4
Apparently stable
4
<1 year
4
10,000-50,000
5
Increasing
5
Certainty:
high (3)
>50,000
6
Certainty:
high (3)
medium (2)
Certainty:
high (3)
medium (2)
low (1)
medium (2)
low (1)
low (1)
Future vulnerability to stochastic events
Score
Current range trend
Score
Highly vulnerable
1
Population trend in last 50 yrs
Score
Rapid reduction
1
Vulnerable
2
>80% reduction
1
Slow reduction
2
Not vulnerable
3
>50% reduction
2
Stable
3
Likely to benefit
4
>20% reduction
3
Increasing
4
Certainty:
high (3)
Apparently stable
4
Certainty:
high (3)
medium (2)
Increasing
5
medium (2)
low (1)
Certainty:
high (3)
low (1)
medium (2)
Vulnerability: policy/management
change
Score
low (1)
Current stressors (narrative)
Highly vulnerable
1
Vulnerable
2
Current population trend
Score
Future non-climate stressors
Score
Not vulnerable
3
Rapid decline
1
Increase
1
Benefiting
4
17
-------
Table 2. (continued)
Current population size
Score
Range trend in last 50 years
Score
Individual replacement time
Score
Slow decline
2
Stable
2
Certainty:
high (3)
Stable
3
Reduction
3
medium (2)
Increasing
4
Certainty:
high (3)
low (1)
Certainty:
high (3)
medium (2)
medium (2)
low (1)
Future vulnerability to natural stressors
Score
low (1)
Highly vulnerable
1
Vulnerable
2
Not vulnerable
3
Certainty:
high (3)
medium (2)
low (1)
18
-------
Current population size. The importance of this variable is that, in general, species with
small populations are likely to be less resilient and more vulnerable to extinction risk than those
with larger populations.
In assessing a species' baseline vulnerability to extinction, it would be valuable to know
how close its current population size is to its minimum viable population size (the population
level below which the species may inevitably face extinction). Ideally, these categorizations
would be based on reliable long-term census data, together with estimates of minimum viable
population sizes. Unfortunately, such information exists for very few organisms, particularly for
rare or restricted species. Any approximations made for the purposes of this framework would
likely, for most species, be so conjectural as to compound rather than eliminate uncertainty. For
this reason, the population size categories in Module 1 are not intended to imply a high degree of
accuracy or precision but to delineate broad "concern categories" that reflect varying degrees of
extinction risk. Assigning a species to any category would typically be based on expert judgment
about the species (or a surrogate).
Past and current population trends. The importance of these variables is that, in general,
species with reduced and/or currently declining populations are likely to be more vulnerable to
extinction risk than those with stable or increasing populations. The greater the past population
reduction and the more rapid the current rate of decline, the more vulnerable the species is likely
to be. Thus, in assessing a species' baseline vulnerability to extinction, it is important to know to
what extent its population has been reduced in the past and its current rate of reduction.
Quantitative data on many species' populations in North America have only begun to be
gathered since about 1950. For this reason, the past reduction category focuses on this time
period. The current rate of population reduction variable focuses on the current 10-year period.
The past trend categorization scheme used (see Appendix G for examples) is similar to and based
on that used in the IUCN Red List scheme (Mace and Stuart, 1994). The current trend
categorization scheme assigns one of four categories: (1) rapid population decline, (2) slow
population decline, (3) stable populations, or (4) increasing populations.
Ideally, these population trend categorizations would be based on quantitative census
data. Unfortunately, such information exists for very few organisms, particularly for rare or
restricted species. Assigning a species to any category will, therefore, typically be based on
expert judgment about the species. For this reason, the population trend categories in Module 1
are not intended to imply a high degree of accuracy or precision, but to delineate broad "concern
categories" that reflect varying degrees of extinction risk.
Past and current range trends. As with population trends, species that have suffered
range (i.e., extent of distribution) contractions in the past, or that are currently suffering such
19
-------
contractions, are likely to be more vulnerable to extinction risk than those with stable or
extending ranges, since the change in their distribution is evidence that they are already under
stress. The greater the past range contraction and the more rapid the current contraction rate, the
more vulnerable the species is likely to be. Thus, in assessing a species' baseline vulnerability to
extinction, it is important to know to what extent its distribution has changed in the past and its
current rate of change. Similar to the population trend categories, the past range change category
focuses on the time period over the last 50 years. The current rate of range change should focus
on the current 10-year period.
Range fragmentation is another variable that was considered for inclusion in this
framework. However, without detailed information about the population viability in each of the
range fragments, or the biogeographic processes of the metapopulation, it is difficult to
determine a priori whether or how range fragmentation may affect extinction risk. A highly
fragmented range could either reduce or increase extinction risk, depending on the dispersive
capability of the organism, its subpopulation viability, and the spatial distribution of the stressor.
Future trends in the magnitude and/or extent of non-climate stressors that could affect
the species' distribution or population status. Species that are or may be affected by nonclimate
stressors that are likely to increase in the future in their intensity, frequency, or spatial extent
(e.g., habitat loss due to urban sprawl), are likely to be more vulnerable than those affected by
stressors that are reducing or stable (e.g., environmental dichlorodiphenylethylene [DDE]
concentrations). In this module component, the likely future trends in the frequencies and/or
intensities of non-climate stressors are categorized as likely to increase, remain stable, or
decrease.
Individual replacement time. k-Selected species (i.e., those with deferred maturity, slow
reproductive rates, postnatal care, etc.) may generally be more at risk of extinction than
r-selected species (i.e., fast reproducers). k-Selected species are best adapted to stable
environments with low stresses, whereas r-selected species are best able to exploit unpredictable
and stressed environments. A population of a k-selected species that is reduced by a stochastic
event has less opportunity than an r-selected species to quickly make good its losses before the
next stochastic event. Approximate individual replacement rate is a useful index for k- or r-
selection status. Thus, desert tortoises, which are k-selected, have an approximate individual
replacement time of 15 years or more and may be more vulnerable to stressors than, for example,
voles, with a replacement time of less than 1 year.
Vulnerability to stochastic events. Some species, because of their habitat preferences or
distributions, may be more at risk to stochastic events than others. For example, organisms that
20
-------
occupy habitats that are vulnerable to tropical storms, fires, tidal surges, or "red tides" may be
more vulnerable than organisms that live in more predictable environments.
Vulnerability to policy/management changes. Because their fates depend to a great extent
on societal values or policy objectives (either of which may change through time), species that
are heavily dependent on human intervention or management, or specific policies for their
continuing survival (e.g., California condors or black-footed ferrets, which are dependent on
captive breeding programs) are likely to be more vulnerable than those that depend less, or not at
all, on such interventions. All T&E species are dependent on policy (since all are listed under the
ESA). However, some, such as the species listed above, are more dependent than others.
Vulnerability to natural stressors. Some species may be more vulnerable to currently
acting natural stressors, such as disease, or invasive species than others are. Seabirds, for
example, appear to be particularly susceptible to botulism, while rodents are vulnerable to
outbreaks of sylvatic plague. A species' vulnerability to such events could affect its ability to
persist.
Each of these variables is assigned a numerical score, reflecting their ordinal rankings.
These individual scores are then combined in Module 1 into one of four baseline vulnerability
rankings:
• Critically vulnerable (Vbl)—species that are likely to be at imminent risk of extinction (a
total Module 1 score of less than 18).
• Highly vulnerable (Vb2)—species that may be close to such an extinction risk and are
likely to be recategorized as critically vulnerable if their populations or ranges are
diminished further (a total Module 1 score of 18-25).
• Less vulnerable (Vb3)—species that are not in imminent danger of extinction but that
could be so in the future if their population and range trends continue (a total Module 1
score of 26-33).
• Least vulnerable (Vb4)—species that have comparatively large and stable (or increasing)
populations or ranges (a total Module 1 score of greater than 33).
5.2. MODULE 1—CERTAINTY EVALUATION
Two methods for evaluating certainty/uncertainty have been incorporated into the
Framework:
First, where necessary, each variable in Table 1 is assigned a "best estimate" score and an
"alternate" score. The former is a professional judgment of the most likely case, whereas the
latter is a less likely, but not an unreasonably unlikely, estimate. In this, we have tried to capture
legitimate uncertainty about the individual scorings. In cases where there is very little
uncertainty, only best estimate scores are given. Summing each of these scores provides some
21
-------
indication of the accuracy or reliability of the total best estimate scores and the extent to which
they may be in error. Thus, for the bald eagle, the sum of the best estimate scores in Module 1 is
32 (see Appendix G), which translates into an overall baseline vulnerability of Vb3 (less
vulnerable). However, if the alternate scores are integrated, the overall score then becomes 28—
34; based on this range, the species is most likely to be Vb3, but could, though this is less likely,
be Vb4 (least vulnerable).
Second, each "best estimate" score in Module 1 is also assigned a numeric certainty
evaluation (high [scores 3], medium [scores 2], or low [scores 1]), which is used in Module 4 to
evaluate the overall degree of certainty that can be assigned to the framework predictions. These
are ordinal rankings, based on expert judgment about the quantity and quality of the available
data (or required but missing data) that support the "best estimate" variable scores. The three
scores should be viewed as approximately equivalent to probabilities of: high—equal to or
greater than about 70%; medium—greater than about 30% but less than 70%; or low—less than
30%.
Examples of Module 1 applied to the golden-cheeked warbler, bald eagle, salt marsh
harvest mouse, Mount Graham red squirrel, desert tortoise, and Lahontan cutthroat trout are
provided in Appendix G.
22
-------
6. MODULE 2—EVALUATING VULNERABILITY TO CLIMATE CHANGE
In this module, the likely vulnerability of a species to future climate change is assessed
and categorized by scoring those elements of its physiology, life history, and ecology that will
likely be important determinants of its responses. This is based on determining ordinal rankings
for 10 Module 2 variables (see Table 3). The scoring of these is described in greater detail below
(see Section 6.1). Treatment of certainty/uncertainty is discussed in Section 6.2. The Module 2
variables and their scores are presented in Table 4, while six examples of the application of
Module 2 are presented in Appendix H.
The scoring system used in Module 2 allows for the possibility that some species may
actually benefit from climate change, for example, species that could benefit from an increased
frequency of climate change-induced stochastic events (e.g., shrub or grassland species that may
benefit from forests being affected by an increased incidence and severity of fires).
Each variable is assigned a "best estimate" certainty score, together with an "alternate"
(i.e., possible, but less likely) score. This will allow subjective confidence limits to be applied to
the overall framework prediction in Module 3. For some species and variables, there may be
enough certainty underlying the best estimate certainty score that no other score is considered
necessary.
Table 3. Components of species' potential physiological, behavioral,
demographic, and ecological sensitivity to climate change included as
variables in Module 2.
(1) Physiological vulnerability to temperature
change
(6) Likely extent of habitat loss due to climate
change
(2) Physiological vulnerability to precipitation
change
(7) Abilities of habitats to shift at same rate as
species
(3) Vulnerability to climate change-induced
extreme weather events
(8) Habitat availability within new range of
species
(4) Dispersive capability
(9) Dependence on temporal inter-relationships
(5) Degree of habitat specialization
(10) Dependence on other species
23
-------
Table 4. Module 2 variables and scores used in categorizing the vulnerabilities of T&E species to climate
change (Vc).
Physiological vulnerability to temp,
increase
Score
Degree of habitat specialization
Score
Availability of habitat in new
range
Score
Likely highly sensitive
1
Highly specialized
1
None
1
Likely moderately sensitive
2
Moderately specialized
2
Limited extent
2
Likely insensitive
3
Generalist
3
Large extent
3
Likely to benefit
4
Certainty:
high (3)
Certainty:
high (3)
Certainty:
high (3)
medium (2)
medium (2)
medium (2)
low (1)
low (1)
low (1)
Likely future habitat loss due to
climate change
Score
Dependence on temporal
inter-relations
Score
Physiological vulnerability to
precipitation change
Score
All or most (>50%)
1
Highly dependent
1
Likely highly sensitive
1
Some (20-50%) trend
2
Moderately dependent
2
Likely moderately sensitive
2
No change
3
Independent
3
Likely insensitive
3
Some gain (20-50%)
4
Certainty:
high (3)
Likely to benefit
4
Large gain (>50%)
5
medium (2)
Certainty:
high (3)
Certainty:
high (3)
low (1)
medium (2)
medium (2)
low (1)
low (1)
Dependence on other species
Score
Highly dependent
1
Vulnerability to change in frequency
or degree of extreme weather events
Score
Ability of habitats to shift at same
rate as species
Score
Moderately dependent
2
Likely highly sensitive
1
Highly unlikely
1
Independent
3
24
-------
Table 4. (continued)
Physiological vulnerability to temp,
increase
Score
Degree of habitat specialization
Score
Availability of habitat in new
range
Score
Likely moderately sensitive
2
Unlikely
2
Certainty:
high (3)
Likely insensitive
3
Likely
3
medium (2)
Likely to benefit
4
Certainty:
high (3)
low (1)
Certainty:
high (3)
medium (2)
medium (2)
low (1)
Dispersive capability
Score
low (1)
Low
1
Moderate
2
High
3
Certainty:
high (3)
25
-------
6.1. SCORING MODULE 2 VARIABLES
The species' likely physiological sensitivity to two main aspects of climate change,
temperature, and precipitation. Some species are more likely than others to be directly affected by
climate change because their physiological tolerances may be narrower (though some species could
benefit). For example, cold-water fish species, such as some salmonids, may be affected more by
increased water temperature. These species may be more apt to avoid affected areas than warm water
fish (e.g., cyprinids or ictalurids), which are physiologically or behaviorally tolerant to increased
temperatures and/or lowered oxygen levels. Thus, it will be critical in evaluating a species' likely
sensitivity to climate change to be able to assess its intrinsic limits to physiological adaptation to
changing temperature or precipitation regimes. Ideally, such evaluations would be based on
experimental evidence for the species being evaluated, or rigorous observational data from the field.
Unfortunately, however, such data are scarce for most species, and the ordinal rankings in Module 2
will likely be based on inferences about closely related species, or from current limits to the species
distribution correlated with climate variables (e.g., Root, 1988), or recent range changes.
The sensitivity categories for this variable of Module 2 are not intended to imply a high degree
of accuracy or precision, rather, they attempt to delineate broad "response categories" that reflect
varying degrees of physiological/behavioral sensitivity. Assigning a species to any category would
typically be based on expert judgment about the species (or a surrogate).
The species' likely vulnerability to an increasedfrequency or magnitude of climate change-
induced extreme weather events. Some species (e.g., forest-nesting birds, or species confined to small
low-lying islands) may be put at greater risk of extinction or population reduction if climate change
results in an increased frequency or magnitude of stochastic events, such as lightning-caused fires or
hurricanes or storm surges. In general, species that are dependent on habitat variables that are
vulnerable to fire, wind storms, or storm surges may be most vulnerable.
Once more, the vulnerability categories for this variable of Module 2 are not intended to imply
a high degree of accuracy or precision, but to identify broad "vulnerability profiles" that reflect
varying degrees of potential sensitivity. Assigning a species to any category would typically be based
on expert judgment about the species (or a surrogate).
Dispersive characteristics that may ameliorate or exacerbate the effects of climate change.
Species with high dispersal capabilities (e.g., flying insects) may be less vulnerable to climate change
than sedentary organisms (e.g., amphibians or reptiles). In this component of Module 2, species are
ranked according to this characteristic and its likely modifying influence. This allocation is based on
the species' potential ability to disperse from the localized effects of climate change. Thus, a "low"
ranking is assigned to species that are unlikely to move more than a few or tens of kilometers from
their natal area and, hence, may be most vulnerable to the localized effects of climate change.
26
-------
Assigning a "moderate" ranking means that a species may be able to disperse as much as a few
hundreds of kilometers. A "high" ranking refers to highly mobile animals that could potentially
disperse as much as many hundreds or some thousands of kilometers. A reptile or amphibian species
may be an example of the first category, whereas a relatively mobile mammal may fit the second, and
a migratory bird the last.
The species' degree of habitat specialization. Species that have a high degree of habitat
specialization (i.e., that are not flexible in their choice of habitats), may be most vulnerable to climate
change because their "fates" are not only a function of their own responses to climate change, but also
to those of their critical habitat components. For example, golden-cheeked warblers and salt marsh
harvest mice are entirely dependent on Ashe juniper forest and salicornia flats, respectively (see
Appendices A and C). These dependencies may render these species more vulnerable than others that
are more flexible in their habitat preferences. In scoring this Module 2 variable, a species is assigned
to one of three habitat specialization categories:
• Highly specialized—species that are restricted by their behaviors or physiologies to a well
defined habitat (usually a vegetation community). Examples of such species include the
California gnatcatcher, Polioptila californica, which is restricted to the remaining fragments of
coastal sage scrub in southern California.
• Moderately specialized—species able to tolerate variability within a habitat type. Examples
might include wetland organisms that can tolerate a wide variety of wetlands from bogs to
marshes, to lakes and rivers (e.g., the bald eagle).
• Generalists—species that are able to exploit a wide variety of habitats (e.g., the European
starling, Sturnus vulgaris, or American robin, Turdus migratorius, both of which can inhabit a
wide range of habitats from native woodlands to farmlands to urban gardens).
The likely extent of habitat loss or gain due to climate change. In this variable, expert opinion
is used to judge the likely impact of climate change on the spatial extents of the T&E species' main
habitats. These classifications are necessarily speculative and should not be assumed to imply a high
degree of accuracy or precision. They are intended to be reasonable approximations.
Many, if not most, species may depend on two or more habitats during their annual or lifetime
cycles. For example, some marine mammals need both an offshore foraging habitat and a terrestrial
breeding site; some migratory shorebirds require arctic tundra breeding habitat, migration stopover
sites on mid-latitude estuaries, and southern latitude grasslands for wintering habitat. For this variable,
the species should be scored according to the largest negative effect. For example, if a species has two
or more critical habitats and the putative effects on these range between 20% and 80% loss, the latter
should determine the score.
27
-------
If the species has two or more habitats and at least one of these is a putative loss, this should
determine the score, even if the other habitats are predicted to show gains. The reasoning behind this is
that if the species is likely to be habitat limited, in the absence of any evidence to the contrary, we
should conservatively assume that the reduced habitat is likely to be the limiting factor. For example,
if, in the case of the shorebird, its southern hemisphere grassland habitat is predicted to increase in
extent under climate change but its arctic tundra habitat decrease, the latter should determine the score.
The likely ability of critical habitats to shift at same rate as species in response to climate
change. Some habitats may be able to shift in response to climate change. For example, the southern
boundary of boreal forest in northern New England may shift north into Canada, and the
corresponding northern habitat ecotone move further north in Labrador (Neilson and Drapek, 1998).
Also, montane plant communities in the European Alps are shifting upslope due to the warming
climate (Grabherr et al., 1994). In such cases, animal species dependent on these habitats could,
potentially, shift with them. However, the success with which this may occur is dependent on
synchronicity (i.e., the habitat being able to shift in approximate synchrony with the species). If a
species' physiological tolerances are exceeded and it is forced to shift its range into regions where its
optimal habitat does not already exist, its future prospects will be affected by how quickly its habitat
can also shift into that new area. For example, if the species being assessed is a songbird that breeds in
California coastal redwood forest and it is forced to move north into less optimal conifer habitat, it
may take so long for its habitat to catch up that the species' existence may be jeopardized. If, however,
the species' habitat was grassland or shrub, the habitat may be able to move in a relatively short time
frame. For this variable, expert judgment is used to score the likelihood of the critical habitat being
able to shift along with the species.
Availability of habitat within the new range. A species that is forced to track its climatic
envelope and shift its range into areas where its critical habitat already exists, may suffer less from
climate change than one that is forced to move into areas where no such habitat exists. In the latter
case, the persistence of the organisms may depend on whether or not its habitat can shift in synchrony
(see above).
The degree of dependence that the species has on other species or the temporal relationships
between species. Species that are highly dependent on another for some critical life history
requirement (for example the golden-cheeked warbler's dependence on Ashe juniper, or a species that
depends for its food supply during an energetic bottleneck on the emergence of a specific life-stage of
another species) may be more vulnerable to the effects of climate change since their likely fates are
closely dependent on those of another species.
28
-------
Each of the above 10 variables is assigned numerical scores. These individual scores are then
combined in Module 2 into an overall evaluation of the species' potential vulnerability to climate
change:
• critically vulnerable (Vcl)
• highly vulnerable (Vc2)
• less vulnerable (Vc3)
• least vulnerable (Vc4)
• likely to benefit from climate change (Vc5)
6.2. MODULE 2—CERTAINTY EVALUATION
Two methods for evaluating certainty/uncertainty are incorporated into the framework:
First, where necessary, each variable in Table 4 is assigned a "best estimate" score and an
"alternate" score. The former is a professional judgment of the most likely case, whereas the latter is a
less likely, but not an unreasonably unlikely, estimate. In this, we have tried to capture legitimate
uncertainty about the individual scorings. In cases where there is very little uncertainty, only best
estimate scores are given. Summing each of these scores provides some indication of the accuracy or
reliability of the total best estimate scores and the extent to which they may be in error. Thus, for the
bald eagle, the sum of the best estimate scores in Module 2 is 27 (see Appendix H), which translates
into an overall climate change vulnerability of Vc3 (less vulnerable). However, if the alternate scores
are integrated, the overall score then becomes 22 to 29; based on this range, the species is most likely
to be Vc3, but could, though this is less likely, be Vc2 or Vc4.
Second, each "best estimate" score in Module 2 is also assigned a numeric certainty evaluation
(high [scores 3], medium [scores 2], or low [scores 1]), which is used in Module 4 to evaluate the
overall degree of certainty that can be assigned to the framework predictions. These are ordinal
rankings, based on expert judgment about the quantity and quality of the available data (or required
but missing data) that support the "best estimate" variable scores. The three scores should be viewed
as approximately equivalent to probabilities of high—equal to or greater than about 70%; medium—
greater than about 30% but less than 70%; or low—less than 30%.
Examples of Module 2 applied to the golden-cheeked warbler, bald eagle, salt marsh harvest
mouse, Mount Graham red squirrel, desert tortoise, and Lahontan cutthroat trout are provided in
Appendix H.
29
-------
7. MODULE 3—EVALUATING OVERALL VULNERABILITY
In this module, the "best estimate" scores from Modules 1 and 2 are combined in a matrix to
produce an overall best estimate evaluation and score of the species' vulnerability to climate change
and important existing stressors. In doing so, species are categorized as critically vulnerable, highly
vulnerable, less vulnerable, least vulnerable, or likely to benefit from climate change. It is important to
note that these are likely approximations of each species' comparative vulnerability. They are not
measures or indices of absolute vulnerability.
The Module 3 evaluation matrix is presented in Table 5.
Table 5. Module 3—Overall vulnerability best estimate scoring matrix.
Climate change (Module 2) vulnerability
scores
Baseline (Module 1) vulnerability scores
Vbl
Vb2
Vb3
Vb4
Vcl
Vol
Vol
Vo2
Vo3
Vc2
Vol
Vol
Vo2
Vo3
Vc3
Vol
Vo2
Vo3
Vo4
Vc4
Vol
Vo2
Vo3
Vo4
Vc5
Vo2
Vo3
Vo4
Vo4
In this module, several assumptions are made (see below):
1. If a species scores Vbl in Module 1 (i.e., it is critically endangered at the present), any Module
2 score between Vcl and Vc4 will result in the overall rating of Vol. A Module 2 score of Vc5
(the species may actually benefit from climate change) will result in its overall score being
Vo2. Thus, the species continues to be critically endangered unless climate change may
actually improve its condition.
2. A species that scores Vb2 in Module 1 but Vcl or Vc2 in Module 2 will have an overall score
of Vol (i.e., its likely susceptibility to climate change will exacerbate its extinction risk).
3. A species that scores Vb2 in Module 1 but that is likely to benefit from climate change (Vc5)
will have an overall score of Vo3 (i.e., its likely benefits from climate change may ameliorate
its extinction risk).
The rationale behind these assumptions is applied throughout the matrix structure of Module 3.
30
-------
The alternate vulnerability evaluations from Modules 1 and 2 are used to develop subjective
vulnerability limits on the Module 3 estimate. Thus, for the golden-cheeked warbler, a best estimate is
that the species is overall Critically Vulnerable (see Appendix I), but there is also a less likely
possibility that it could be "only" Highly Vulnerable.
Examples of Module 3 applied to golden-cheeked warbler, bald eagle, salt marsh harvest
mouse, Mount Graham red squirrel, desert tortoise, and Lahontan cutthroat trout are provided in
Appendix I.
31
-------
8. MODULE 4—CERTAINTY EVALUATION
The approximate level of certainty for each "best estimate" score in the first two Modules is
categorized as high (approximate probability of 70% or more); medium (approximate probability of
between 30 and 70%); or low (less than approximately 30%). These qualitative certainty scores are
recorded separately in each Module and correspond to numeric scores of 3, 2, and 1, respectively. For
the most part, these categorizations will be the product of expert judgment, rather than a strictly
quantitative appraisal.
In Module 4, the "best-estimate" certainty scores assigned to each of the variables in Modules
1 and 2 are combined into an index of the certainty associated with the overall vulnerability score in
Module 3. The total minimum score (Modules 1 and 2 combined) is 20, while the maximum is 60. The
numeric range between the two is arbitrarily and approximately equally divided into three categories:
high, medium, and low certainties. A final certainty evaluation is then applied to each species. It is
important to note that these categorizations are indices of the certainty associated with the overall
"best estimate" score.
Examples of the individual variable certainty scores are provided for golden-cheeked warbler,
bald eagle, salt marsh harvest mouse, Mount Graham red squirrel, desert tortoise, and Lahontan
cutthroat trout in Appendix J. Examples of the overall certainty level assignations for these species are
also provided in the same appendix.
32
-------
9. SUMMARY OF FRAMEWORK EVALUATIONS
The results of the species evaluations are shown in Tables 6 and 7. In general, baseline
vulnerabilities were either Vb2 (highly vulnerable) or Vb3 (less vulnerable). The bald eagle is scored
as the least vulnerable species in this module (a score of 32 out of 42), and the salt marsh harvest
mouse and golden-cheeked warbler is the most vulnerable (both scored 22). The contrasting scores for
these species are largely a function of widely different population sizes and trends, and distributions,
with the bald eagle being widely distributed across the United States and with a relatively large and
increasing population, while the salt marsh harvest mouse and golden-cheeked warbler are relatively
restricted in their distributions with relatively small populations. In terms of population size and
distribution, the desert tortoise is closer to the bald eagle. However, unlike the bald eagle, it is
restricted to a relatively small area and its populations are decreasing. It, therefore, scored 26.
Table 6. Modules 1 and 2, and certainty scores for the six test species.
Species
Module 1
scorea'b
Module 2
scoreb'c
Certainty score
(Module l/Module2)d
Bald eagle
32
27
26/23
Golden-cheeked warbler
22
15
26/25
Mount Graham red squirrel
24
17
23/24
Salt marsh harvest mouse
22
18
18/19
Desert tortoise
26
19
24/21
Lahontan cutthroat trout
23
16
17/18
aThe total species evaluation score in Module 1.
bThe lower the score the greater the vulnerability.
°The total species evaluation score in Module 2.
dThe certainty scores for Modules 1 and 2.
Potential vulnerability to climate change also ranged widely among the test species, from Vc3
to Vcl. No species scored Vc4 or Vc5. This is possibly because either no species that really merits
these scores is evaluated or because, by definition, species on the T&E list are unlikely to be
successful exploiters of stressed conditions and all would be adversely affected to some degree by
climate change. Perhaps if an invasive, "weed" species were evaluated, it would score Vc4 or Vc5;
however, there are no such species on the T&E list.
33
-------
The golden-cheeked warbler has the lowest total score (or the greatest vulnerability) in Module
2 (a score of 15 out of 35). The Lahontan cutthroat trout was a close second at 16, and the Mount
Graham red squirrel third at 17. In general, species that had specialized habitat requirements that were
likely to be reduced in extent by climate change and that were vulnerable to stochastic events tend to
have the lowest scores. In contrast, the bald eagle, with its comparatively large population size,
generalized habitat requirements, widespread distribution, and low sensitivity to stochastic events
scores highest.
Table 7. Summarization of results of species evaluations.
Species
Module 1
baseline
scores
Module 2
climate change
scores
Module 3
best estimate
scores
Module 3
alternate
scores
Module 4
certainty
score
Golden-
cheeked
warbler
Vb2 (highly
vulnerable
Vcl (critically
vulnerable)
Vol
(critically
vulnerable)
Vo2 (highly)
High
Bald eagle
Vb3 (less
vulnerable)
Vc3 (less
vulnerable)
Vo3 (less
vulnerable)
Vo2, Vo4
(highly,
least)
High
Salt marsh
harvest
mouse
Vb2 (highly
vulnerable
Vc2 (highly
vulnerable)
Vol
(critically
vulnerable)
Vol, Vo2
(critically,
highly)
Medium
Mount
Graham red
squirrel
Vb2 (highly
vulnerable)
Vc2 (highly
vulnerable)
Vol
(critically
vulnerable)
Vol, Vo2
(critically,
highly)
High
Desert
tortoise
Vb3 (less
vulnerable)
Vc2 (highly
vulnerable
Vo2 (highly
vulnerable)
Vol, Vo3
(critically,
less)
Medium
Lahontan
cutthroat
trout
Vb2 (highly
vulnerable)
Vc2 (highly
vulnerable)
Vol
(critically
vulnerable)
Vol, Vo2
(critically,
highly)
Medium
Golden-cheeked warbler, salt marsh harvest mouse, Mount Graham red squirrel, and Lahontan
cutthroat trout are categorized as critically vulnerable in the overall vulnerability module (Module 3).
In contrast, with its widespread distribution, currently increasing populations, and relatively catholic
habitat preferences, the bald eagle is assessed, overall, as the least vulnerable of the species evaluated.
34
-------
Desert tortoise is intermediate. This is because of its large current population and the possibility that
the species may benefit from climate change.
Certainty scores in Module 4 range between medium and high and reflect the amount and
quality of ecological information available. The bald eagle and the golden-cheeked warbler (which are
"charismatic" species, and relatively easily studied and counted) have the highest scores. None of the
species evaluated have a "low" certainty score. However, all of the taxa evaluated (birds, mammals,
and reptiles) are relatively well studied.
The certainty evaluation identified major uncertainties in module components common to all of
the species tested. The most universal of these are the likely physiological sensitivities that organisms
may show toward changes in temperatures and precipitation. Empirical data are almost entirely
lacking for most terrestrial organisms (though data for aquatic organisms may be more plentiful). Even
where data are available, they typically are derived from studies of acute effects, which have less
relevance for most organisms (which will likely respond to changing conditions before acute levels of
change are reached). This gap in knowledge points to the need for more studies of sublethal and
behavioral temperature responses in organisms.
35
-------
10. SUMMARY AND CONCLUSIONS
This report describes a framework for evaluating the potential sensitivities of T&E animal
species in the United States to future climate change. The framework developed consists of five
separate components:
(1) A module that evaluates the current vulnerability to extinction of the species.
(2) A module that evaluates the potential incremental risk of extinction due to climate change.
(3) A module that integrates the results of 1 and 2.
(4) A module that evaluates uncertainty regarding the framework predictions.
(5) Narratives that provide the justifications for the framework scores.
The framework was preliminarily tested by applying it to six species: golden-cheeked warbler,
bald eagle, salt marsh harvest mouse, Mount Graham red squirrel, desert tortoise, and Lahontan
cutthroat trout. Scores for these species varied widely. However, the species that scored most
vulnerable are restricted in their distributions, have small population sizes, are currently undergoing
population reductions, are habitat specialists, and/or have habitats that are likely to be most adversely
affected by future climate change. Conversely, animals (for example, the bald eagle) that are widely
distributed, that are flexible in their habitat preferences, and that are stable or increasing, scored least
vulnerable. Thus, the predictions of the model accord with what might be expected based on the
ecologies and demographics of the test species. The results of these tests also indicate that major areas
of uncertainty complicate any evaluations of vulnerability. For the species tested, the greatest
uncertainties are associated with a relatively poor knowledge about the potential for direct,
physiological effects on animal species; relationships between changes in temperature and
precipitation regimes and the physiologies and behaviors of animals are, apparently, only poorly
understood.
While this framework was developed to evaluate the vulnerabilities of T&E species, there is no
reason why it could not, with some minor modification, be suitable for use with all vertebrate species.
The main limitation might be that while T&E species are often relatively well studied and the data
necessary for the framework may be available, this might not be the case for other less well known
species. However, it should still be possible to assign vulnerability scores using the framework,
although the confidence scores may be lower.
36
-------
REFERENCES
Andrews, R; Righter, R.. (1992) Colorado birds. Denver: Denver Museum of Natural History.
Baker, RH. (1983). Michigan mammals. Detroit: Michigan State University Press.
Beardmore, C; Hatfield, J; Lewis, J. (1995) Golden-cheeked warbler population and habitat viability
assessment report. Austin, Texas: U.S. Fish and Wildlife Service.
Bergeron, D; Jones, C; Genter, DL; et al. (1992) Montana bird distribution. Helena, Montana: Natural
Heritage Program.
Boschung, HT; Williams, JD; Gotshall, DW et al. (1983) The Audubon Society field guide to North
American fishes, whales, and dolphins. New York: Alfred A. Knopf.
Burt, WH; Grossenheider, RP. (1964) A field guide to the mammals. Boston: Houghton Mifflin.
Caire, W; Tyler, JD; Glass, BP; Mares, MA. (1989) Mammals of Oklahoma. Norman, Oklahoma:
University of Oklahoma Press.
Carter, MF; Hunter, WC; Pashley, DN; Rosenberg, KV. (2000) Setting conservation priorities for
landbirds in the United States: the Partners in Flight approach. The Auk 117:541-548.
Chapman, JA; Feldhamer, GA. (1982) Wild mammals of North America. Baltimore: Johns Hopkins
University Press.
Choate, JR; Knox Jones, J; Jones, C. (1994) Handbook of mammals of the South-Central states. Baton
Rouge: Louisiana State University Press.
Collar, NJ; Crossby, MJ; Stattersfield, AJ. (1994) Birds to watch 2: the world list of threatened birds.
BirdLife International.
Collins, JT. (1982) Amphibians and Reptiles in Kansas. Lawrence, Kansas: University of Kansas.
Conant, R: Collins, T. (1998) Reptiles and amphibians of Eastern and Central North America. Boston:
Houghton Mifflin.
Cooper, EL. (1983) Fishes of Pennsylvania. University Park: Pennsylvania State University Press.
Craighead, JJ; Mitchell, JA. (1982) Grizzly bear. In: Chapman, JA; Feldhamer, GA, eds. Wild
mammals of North America. Baltimore, Johns Hopkins University Press.
Crance, JH. (1985) Delphi technique procedures used to develop habitat suitability index models and
instream flow suitability curves for inland stocks of striped bass. U.S. Fish and Wild. Service,
WELUT-85/W07.
Dixon, JR. (1987). Amphibians and reptiles of Texas. College Station: Texas A&M University Press.
37
-------
Dundee, HA; Rossman, DA. (1989) The amphibians and reptiles of Louisiana. Baton Rouge:
Louisiana State University Press.
Dunn, JL; Garret, KL. (1997) A Field Guide to Warblers of North America. Houghton Mifflin, New
York
Ernest, CH; Lovich, JE; Barbour, RW. (1994) Turtles of the United States and Canada. Washington,
Smithsonian Institution Press.
Fitzgerald, JP; Meaney, CA; Armstrong, DM. (1994) Mammals of Colorado. Niwot: Denver Museum
of Natural History and University Press of Colorado.
Grabherr, G; Gottfried, M; Pauli, H. (1994) Climate effects on mountain plants. Nature 369: 448.
Harding, JH. (1997) Amphibians and reptiles of the Great Lakes Region. Ann Arbor:University of
Michigan Press.
Herman, TB; Scott, FW. (1994) Protected areas and global climate change: assessing the regional or
local vulnerability of vertebrate species. In: Pernetta, J; Leemans, R; Elder, D; Humphrey, S, eds.
Impacts of climate change on ecosystems and species: implications for protected areas. Gland,
Switzerland: IUCN.
Hoffmeister, DF. (1989) Mammals of Illinois. Urbana and Chicago: University of Illinois Press,.
Hunter, ML; Calhoun, AKJ; McCollough, M. (1999) Maine amphibians and reptiles. Orono:
University of Maine Press.
Ingles, LG. (1965) Mammals of the Pacific states. Stanford: Stanford University Press.
Jameson, EW; Peeters, HJ. (1988) Californian mammals. Berkeley: University of California Press.
Kaufman, K. (1996) Lives of North American birds. Boston: Houghton Mifflin.
Knox Jones, I; Birney, EC. (1988) Handbook of mammals of the North-Central States. Minneapolis:
University of Minnesota Press.
Kingery, HE. (ed). (1998). Colorado breeding bird atlas. Denver: Colorado Bird Atlas Partnership and
Colorado Division of Wildlife.
Lanoo, MJ. (ed). (1988) Status and conservation of midwestern amphibians. Iowa City: University of
Iowa Press.
Laughlin, SB; Kibbe, DP. (1985) The atlas of the breeding birds of Vermont. Hanover and London:
University Press of New England.
Lee, DS; Gilbert, CR; Hocutt, CH; Jenkins, RE; McAllister, DE; Stauffer, JR Jr. (1982) Atlas of North
American Freshwater Fishes. North Carolina State Museum of Natural History. 854pp.
38
-------
Linstone, HA; Turoff, M, eds. (1975) The Delphi Method. Reading, MA: Addison-Wesley.
Mace, G; Stuart, S. (1994) Draft IUCN Red List categories. Species 21-22:13-24.
Merrit, J.F. 1987. Guide to the Mammals of Pennsylvania. University of Pittsburgh Press, Pittsburgh.
Minton, SA. (1972) Amphibians and Reptiles of Indiana. Indianapolis: Indiana Academy of Science.
Moss, RH; Schneider, SH. (2000) Uncertainties in the IPCC TAR: recommendations to lead authors for
more consistent assessment and reporting. In: Pechauri, R; Tanguchi, T; Tanaka, K, eds Guidance papers
on the cross cutting issues of the third assessment report of the IPCC. Geneva: World Meteorological
Organization; pp. 33-51.
National Geographic Society. (1999). Field guide to the birds of North America. Third Edition.
Washington, DC.
Neilson, RP; Drapek, RJ. (1998) Potentially complex biosphere responses to transient global warming.
Global Change Biology 4:505-521.
Opler, PA; Malikul, V. (1998). Field guide to eastern butterflies. Boston: Houghton Mifflin.
Opler, PA; Wright, AB. (1999) Field guide to western butterflies. Boston: Houghton Mifflin.
Parmesan, C; Galbraith, H. (2005) Observed impacts of global climate change in the U.S. Report
published by: Pew Center on Global Climate Change, Arlington, VA.
Peters, RL. (1992) Conservation of biological diversity in the face of climate change. In: R.L. Peters,
RL; Lovejoy, TE, eds. Global warming and biological diversity., Yale University Press, New Haven.
Peterson, A. (1986) Habitat suitability index models: bald eagle (breeding season). U.S. Fish Wildl
Serv Biol Rep 82(10.126).
Petranka, JW. (1998). Salamanders of the United States and Canada. Washington: Smithsonian
Institution Press.
Price, J; Droege, S; Price, A. (1995). The summer atlas of North American birds. New York:
Academic Press.
Rausch, RL. (1963) Geographic variation in size in North American brown bears, Ursus arctos Z., as
indicated by condylobasal length. Can J Zool 41:33-45.
Root, T. (1988). Atlas of wintering North American birds. Chicago and London: University of
Chicago Press.
Sanz, JJ; Potti, J; Moreno, J; O Frias. (2003) Climate change and fitness components of a migratory
bird breeding in the Mediterranean region. Global Change Biology 9:461-472.
Schneider, SH; Root, TL. (2002) Wildlife responses to climate change. Washington, DC: Island Press.
39
-------
Scott, JA. (1986) The butterflies of North America. Stanford: Stanford University Press.
Shull, EM. (1987) The butterflies of Indiana. Bloomington: Indiana Academy of Science.
Sibley, DA. (2000) The Sibley guide to birds. New York: Knopf.
Siegel, S. (1956). Nonparametric statistics for the behavioral sciences. Tokyo: McGrawHill.
Sigler, WF: Sigler, JW. (1996) Fishes of Utah: a natural history. Salt Lake City: University of Utah
Press.
Sublette, JE; Hatch, MD; Sublette, M. (1990). The fishes of New Mexico. Albuquerque: University of
New Mexico Press.
Temple, SA; Cary, JR. (1987) Wisconsin birds: a seasonal and geographical guide. Madison:
University of Wisconsin Press.
Tomelleri, JR; Eberle, ME. (1990) Fishes of the Central United States. Lawrence: University Press of
Kansas.
Trautman, B. (1981) The fishes of Ohio. Ohio State University Press.
Tucker, GM; Heath, MF. (1994). Birds in Europe: their conservation status. Cambridge, U.K.: Birdlife
International (Birdlife Conservation Series No. 3).
U.S. FWS (U.S. Fish and Wildlife Service). (1992) Golden-cheeked Warbler (Dendroica chrysoparia)
recovery plan. U.S. Fish and Wildlife Service, Albuquerque, New Mexico.
Veit, RR: Petersen, WR. (1993) Birds of Massachusetts. Lincoln, Massachusetts: Massachusetts
Audubon Society.
Walther, G-R; Post, E; Convey, P;. (2002) Ecological responses to recent climate change. Nature
416:389-395.
Webb, RG. (1970) Reptiles of Oklahoma. Norman: University of Oklahoma Press.
Whitaker, JD. (1980) The Audubon society field guide to North American mammals. New York:
Alfred A. Knopf.
Whitaker, JD; W.J. Hamilton, WJ. (1998) Mammals of the Eastern United States. Ithaca: Cornell
University Press.
Wilson, DE; Ruff, S. (Eds) (1999). The Smithsonian book of North American mammals. Washington:
Smithsonian Institution Press.
Zeveloff, SI. (1988). Mammals of the Intermountain West. Salt Lake City: University of Utah Press.
40
-------
Zuboy, JR. (1980). The Delphi technique: a potential methodology for evaluating recreational
fisheries. FAO Proceed. Tech. Consultation on Allocation of Fisheries. Pp 519-529.
Zuboy, JR. (1981) A new tool for fisheries managers: the Delphi technique. North Am J Fish Mgt
1:55-59.
41
-------
APPENDIX A
EXAMPLE NARRATIVE FOR GOLDEN-CHEEKED WARBLER
A.l. INTRODUCTION
Based on a review of available information, this narrative describes the current vulnerability of
the golden-cheeked warbler to severe population reduction or extinction, and its potential future
vulnerability under climate change. Its main objectives are to
• make transparent the rationale underlying each score in Modules 1 and 2;
• identify main sources of uncertainty;
• identify and describe the roles of the main stressors in the estimate of vulnerability for this
species; and
• qualitatively describe potential population responses to climate change and other stressors.
A.2. ENDANGERED SPECIES ACT STATUS
The golden-cheeked warbler, Dendroica chrysoparia, was proposed for listing as Endangered
under the ESA in May 1990 (U.S. FWS, 1990). The Final Rule confirming this listing was published
in December 1990. The species was also listed as Endangered under State of Texas legislation in
February 1991 (Texas Parks and Wildlife Department, Executive Director Order no. 91-001).
A.3. DISTRIBUTION, STATUS, AND POPULATION TRENDS
A summer visitor and breeding bird to the United States, the golden-cheeked warbler winters
mainly in southern Mexico (Chiapas), Honduras, and Nicaragua (Curson et al., 1994; Kaufman, 1996;
Dunn and Garrett, 1997; Ladd and Gass, 1999). Its breeding range is extremely restricted—only 30
counties in central Texas. As far as is known, it breeds nowhere else. This range extends as a strip
approximately 250 miles in length and 150 miles in width from near San Antonio in the south almost
to Dallas and Ft. Worth in the north (Ladd and Gass, 1999). Even within this very limited range,
golden-cheeked warblers are localized (Dunn and Garrett, 1997) to highly fragmented areas where
suitable habitat occurs (see below).
Based on habitat delineation from aerial photographs, the total breeding population was
estimated in 1990 as somewhere between 10,000 and 30,000 individuals (Wahl et al., 1990). Wahl et
al. (1990), also estimated that the breeding population of golden-cheeked warblers declined by
approximately 25% between 1962 and 1981. Pulich (1976) estimated a decrease from 18,500 pairs in
1962 to approximately 14,750 pairs in 1974, a 20% reduction. Although both these sets of figures are
likely to have large unresolved uncertainties, it is likely that breeding numbers (and hence the world
42
-------
population) of golden-cheeked warblers has declined markedly in the last 4-5 decades and is now less
than 30,000 individuals.
A.4. HABITAT
During the breeding season, golden-cheeked warblers are habitat specialists, being confined to
juniper-oak woodlands dominated by Ashe juniper, Juniperus ashei. The close relationship with the
juniper is explained by the fact that in mature trees of this species, the bark peels in long strips, and the
warbler constructs its nest largely from this material (Ladd and Gass, 1999). Ashe juniper woodland is
confined to central Texas. Even within areas where Ashe juniper occurs, golden-cheeked warblers are
selective, preferring sites dominated by mature or old growth trees. Mature stands of Ashe juniper-oak
woodlands are extremely limited in their distribution in central Texas and are confined largely to
Cretaceous upland limestone karsts and canyon sides with shallow soils along the eastern edge of the
Edwards Plateau, the Balcones Escarpment, and the Lampasa Cut Plain (Kuchler, 1975). It is the
availability of this specialized breeding habitat that presently defines and limits the distribution and
status of golden-cheeked warblers.
In winter, golden-cheeked warblers are also dependent on conifer-oak woodlands, their main
habitat in this season being high elevation (1,500-3,000 m) pine-oak woodlands in Central America
(Ladd and Gass, 1999).
A.5. PHYSIOLOGICAL/CLIMATIC LIMITATIONS ON DISTRIBUTION AND STATUS
No experimental information has been found on the physiological sensitivity of
golden-cheeked warblers to changes in temperature or precipitation. However, field studies of four
other species of New World warblers (Wilson's warbler, Wilsonia pusilla; red-faced warbler,
Cardellina rubifrons., Virginia's warbler, Vermivora virginiae; and orange-crowned warbler,
Vermivora celata) nesting in hot, arid environments in the U.S. southwest have shown that
temperature and humidity can directly affect individuals, causing them to alter their micro-habitat
preferences (Martin, 2001). Thus, year-to-year changes in these local climatic variables, mediated
through habitat selection behavior, result in changes in distribution. It is not known whether these
climate sensitivities also apply to golden-cheeked warblers, (although information from Sexton 2007,
suggests that breeding success of golden-cheeked warblers may be depressed during extreme drought
conditions). Nevertheless, the study describes above found direct climate effects in the only species
thus far examined, and it is likely that the relationships could be more widespread within the New
World warblers.
Of the 21 species in the genus Dendroica that breed in North America, only 9 breed as far
south as the southernmost states. Of these 9, only 1 (yellow-rumped warbler, Dendroica coronata)
breeds further south than the southernmost geographical limit of the golden-cheeked warbler, and only
43
-------
by about 100 miles. The two Dendroica species that breed as far or further south than the golden-
cheeked warbler are the yellow-rumped warbler and Grace's warbler {Dendroicagraciae). However,
they breed in high-mountain forests in northern Mexico, habitats that are likely to be generally cooler
than those of the golden-cheeked warbler. Thus, the southern limit of the breeding distribution of
golden-cheeked warblers may be the hottest environment in which Dendroica warblers breed. The
absence of other congeners in this temperature zone may imply a climatic limit that follows the
southern limit of golden-cheeked warblers. If this is the case, golden-cheeked warblers may be at their
southernmost physiological climatic limit, and the 4-7° C temperature increase predicted for central
Texas by a number of General Circulation Models (GCMs [VEMAP Members, 1995]) may in the
future render currently occupied areas physiologically unsuitable.
The above considerations could mean that climate imposes a southern limit on the current
breeding range of golden-cheeked warblers. However, given the lack of experimental evidence, this
must be regarded as conjectural.
A.6. ECOLOGICAL LIMITATIONS ON DISTRIBUTION AND STATUS
The breeding distribution of golden-cheeked warblers is limited by the geographical
distribution of their preferred habitat: mature stands of Ashe juniper—dominated woodland (Pulich,
1976; Ladd and Gass, 1999). While Ashe juniper trees occur from northern Mexico to southern
Oklahoma and Arkansas, stands of woodland dominated or codominated by mature Ashe junipers are
confined to areas of central Texas where solid geology, soil characteristics, precipitation, and land use
are suitable (Diamond, 1997).
The current spatial distribution of suitable Ashe juniper woodland reflects surficial geology in
that it is largely confined to an area of upland that once was a Cretaceous marine reef. This reef now
comprises a limestone escarpment and upland area stretching north from its highest elevations at
Balcones to its lowest point just south of the Dallas-Fort Worth area (AAPG/USGS, 1973). In its
southern and central areas, the reef is close to the land surface, which is characterized by upland
limestone karst and thin skeletal soils. This is the surficial geology that best supports the development
of the Ashe juniper woodlands (Amos and Gelbach, 1988; Diamond, 1997). As the reef dips to the
north, it gradually becomes covered by deeper soils supporting grasslands and oak savannas (Kuchler,
1975), less suitable habitat for golden-cheeked warblers. To the west, on the flatter parts of the
Edwards Plateau uplands, the soils are deeper and the juniper woodlands are largely replaced by
grasslands and savannas (Diamond, 1997). To the east of the plateau escarpment, the soils are deeper
and the land is intensively cultivated. Thus, the potential breeding range of the golden-cheeked
warbler is, ultimately, largely defined by solid and surficial geology acting on plant community
development (although human land use has further contracted this range—see Section A.7).
44
-------
A.7. EXISTING STRESSORS AND THEIR EFFECTS ON DISTRIBUTION AND STATUS
Since colonization of the area by Europeans, the geographical extent of Ashe
juniper-dominated woodlands and, therefore, the breeding distribution of golden-cheeked warblers,
has been reduced drastically. At first, the juniper forests were harvested for wood products and for
fuel; later they were cleared for grazing animals and for urbanization. In some counties where juniper
woodlands currently exist only in isolated patches, early maps (cited in Diamond, 1997) show the
landscape formerly entirely covered by woodlands. It is not possible to quantify the entire extent of
habitat loss since Europeans began manipulating the landscape (since no estimates of pre-Columbian
cover exist); however, Pulich (1976) estimated that between 1962 and 1974, 20% of the habitat was
destroyed. Keddy-Hector (1992) estimated that a further 30% was lost between 1974 and 1990. Thus,
mature Ashe juniper woodland is the main determinant of the breeding distribution of golden-cheeked
warblers, and destruction of this plant community has been a major factor responsible for the
contraction of this breeding range.
Other stressors, particularly fire and drought, may have limited the range and numbers of
golden-cheeked warblers. Ashe junipers are not fire tolerant, and wild fires and fires caused by
humans may have been a factor in determining the current distribution of the species
(Diamond et al., 1995; Diamond, 1997). The drought of the early 1950s killed many stands of junipers
(Diamond et al., 1995). During this drought, older and larger Ashe junipers (the age group preferred
by golden-cheeked warblers) suffered 90% mortality in some areas (Merrill and Young, 1959, cited in
Diamond etal., 1995).
Another more recent stressor on golden-cheeked warblers in their nesting habitat is nest
parasitism by cowbirds, Molothrus ater (Kaufman, 1996; Eckrich et al., 1999). Typically, cowbirds
are more successful at parasitising woodland birds' nests, where the woodland is fragmented
(Britingham and Temple, 1983). The current fragmentation of the golden-cheeked warblers' breeding
habitat may facilitate their parasitism by cowbirds.
A.8. POTENTIAL DIRECT (PHYSIOLOGICAL) VULNERABILITY TO CLIMATE
CHANGE
As discussed in Section A.5, the breeding range of golden-cheeked warblers may be at or close
to the southern edge of a distributional limit set by high temperatures acting directly on the physiology
and/or behavior of the species. If this is the case, the temperature increases of between 4 and 7°C that
are generally projected by Global Climate Models (GCMs) for central Texas (VEMAP Members,
1995) could result in physiological problems and pressure on individuals to shift their ranges north.
The likely feasibility of such a range shift is discussed in Section A.9.
45
-------
A.9. POTENTIAL INDIRECT (ECOLOGICAL) VULNERABILITY TO CLIMATE
CHANGE
Golden-cheeked warblers are likely to be indirectly highly vulnerable to climate change for the
following reasons:
(1) They are habitat specialists dependent on a plant community that could be drastically affected
by climate change. As discussed above, woodlands dominated by Ashe juniper are extremely
localized in their distribution. It is unlikely that this community will be able to shift its location
in response to climate change for the following reasons:
• The current distribution of the community is at least partly determined by solid geology
and soil type. Immediately north of the current range of Ashe juniper forests, the soil
becomes deeper and more fertile and more suitable for the establishment of grasslands and
deciduous trees, rather than junipers.
• Important barriers to community migration exist. Immediately north of the current range of
Ashe juniper woodlands, the land is intensively farmed for arable crops, creating habitat
that junipers would be unlikely to be able to colonize. Also, the Dallas-Ft. Worth
metropolis lies across the likely migration route. This is an urban barrier about 50 miles
wide by 30 miles deep. Even if suitable soils and land-use patterns existed to the north, it is
unlikely that Ashe juniper woodlands could cross such a barrier.
(2) Their current population is relatively small and likely to be only just maintaining itself in the
face of existing stressors (U.S. FWS, 1996). Additional important stressors (such as climate
change) could be enough to push the species closer to extinction.
Given the barriers to migration discussed above and the sensitivity of Ashe juniper woodlands
to fire and drought, it is likely that climate change will result in the further fragmentation of the
existing Ashe juniper woodlands with resulting loss of habitat for the warblers.
A.10. JUSTIFICATION FOR FRAMEWORK SCORES
The framework scores for the golden-cheeked warbler are presented in Appendices G through
J.
In Module 1 (baseline vulnerability), the species scored Vb2, indicating that it is currently in a
highly vulnerable condition. This score is based on the following subcomponents:
• Current population size and trends—based on the census data that are available (see Section
A.2), there are likely to be less than 30,000 individuals in existence. Moreover, there is
evidence that the population may be in decline. Because of this, the species has been given best
estimate scores of 5 and 2, respectively, in the population size and trends variables of Module
1 (with alternate scores of 3 in the two population trend variables).
46
-------
• Range trends—golden-cheeked warblers are restricted to a relatively small breeding range that
is being reduced in its extent. For these reasons, the species has been allocated best estimate
scores of 3 and 2 in the historic and current range trends variables of Module 1.
• Current stressors and likely stressor trends—the main stressor that has reduced the populations
of golden-cheeked warblers within their range is habitat modification (see Section A.7).
Although, not being destroyed at previous rates, the habitat of golden-cheeked warblers is still
under some threat from land use changes, in particular modification through grazing and
burning and urbanization. It is likely that this threat will persist into at least the near future.
Also, the current fragmented nature of the species breeding habitat makes it more vulnerable to
nest parasitism by cowbirds. For these reasons, the species has been allocated a best estimate
score of 1 (increasing stressors), with an alternate score of 2 (stable stressors).
• Individual replacement time—individual golden-cheeked warblers begin breeding in their first
year and may be producing young in as little as 14 months after they, themselves, fledged.
Thus potential replacement time may be less than 2 years, and the species has been allocated a
best estimate score of 3.
• Vulnerability to stochastic events—drought and fire may have important adverse impacts on
golden-cheeked warbler local populations. For this reason the species has been allocated a best
estimate score of 1 (highly vulnerable), with an alternate score of 2 (vulnerable).
• Policy/Management change vulnerability—a large component of the existing golden-cheeked
warbler population exists in protected areas (e.g., Fort Hood Military Base). Because of this, it
has been assigned a best estimate conservation dependency score of 1 (highly vulnerable to
change in policy or management), with an alternate score of 2.
• Vulnerability to natural stressors—the incidence of disease, parasitism, or other natural
stressors on this species is not known. However, given that the population is so small and
restricted in its distribution, it is feasible that such a stressor could, potentially, have important
effects on population viability. Nest parasitism by cowbirds is an important and perhaps
increasing natural stressor for this species. The species is allocated a best estimate score of 2
(vulnerable), with an alternate score of 3 (not vulnerable).
Certainty evaluations were allocated to each of the scores in Module 1. These are largely
subjective evaluations of the robustness of each of the scores and reflect the availability and quality of
information for each category, rather than rigorous evaluations. Given that the species is relatively
well studied, high certainty scores were assigned to most of the variables. The exceptions were future
vulnerability to pathogens, future stressor trends, and past and current population trends (all of which
had medium certainty scores). No variable of Module 1 had a low certainty score.
In Module 2, the species scored Vcl, indicating that it is likely to be critically vulnerable to
climate change and that its extinction risk may be increased substantially. This score is based on the
following subcomponents:
47
-------
• Physiological vulnerability to temperature and precipitation change and to extreme weather
events—the golden-cheeked warbler may nest close to the southern limit of its physiological
climate "envelope" (see Section A.5). If so, the species may be adversely affected by future
increases in temperature. Therefore, it has been allocated a best estimate score of 2 (likely
moderately sensitive) in the temperature variable of Module 2, with an alternate score of 1
(likely highly sensitive). The species may be less sensitive to precipitation change and has been
allocated a best estimate score of 3 (likely insensitive), and an alternate score of 2 (likely
moderately sensitive).
• Vulnerability to changes in the frequency/degree of extreme weather events—the species has
been allocated a best estimate score of 1 (likely highly sensitive). This has been assigned
because of the potential vulnerability of the species' habitat to increased frequencies of drought
and consequent fires. An alternate score of 2 (likely moderately sensitive) has also been
assigned.
• Dispersive capability—the dispersive capability of this mobile species is high (best estimate
score of 3 in Module 2).
• Habitat specialization—golden-cheeked warblers are extremely specialized in their breeding
habitat requirements (see Section A.4). Because of this acute dependence on one specific
habitat type, the species has been allocated a best estimate score of 1.
• Likely extents of future habitat loss due to climate change—for this variable, golden-cheeked
warblers were allocated a best estimate score of 1 (>50% likely habitat loss). This score
reflects the species' degree of habitat specialization, the already limited extent of breeding
habitat and its fragmentation, and the unlikely possibility of the habitat being able to move in
response to climate change. A secondary effect of future habitat fragmentation may be
increased rates of nest parasitism by cowbirds. An alternate score of 2 (20-50% habitat loss)
was allocated.
• Ability of habitats to shift at same rate as species in response to climate change—because the
current distribution of Ashe juniper woodland is largely determined by geology, soils, and land
use, and suitable conditions do not exist to the north of its current range, it is highly unlikely
that the breeding habitat of the species will be able to shift in response to climate change. Thus,
this variable scores (best estimate) 1, with an alternate score of 2.
• Availability of habitat within new range—as discussed in Section A.4, the breeding habitat of
this species is limited by surficial geology, and no habitat exists to the north of the species'
current breeding range because of spatial changes in the geology. It has, therefore, been
allocated a best estimate score of 1 (no habitat exists), with an alternate score of 2 (only limited
habitat exists).
• Dependence on temporal inter-relations and other species—the golden-cheeked warbler is
extremely dependent on at least one other species—Ashe juniper. Therefore, it has been
assigned a score of 1 (highly dependent) in the second variable. Most small songbirds time
their migrations and breeding to take advantage of seasonal flushes of invertebrate prey.
Accordingly, a score of 2 has been assigned to the temporal variable.
48
-------
Certainty evaluations were allocated to each of the scores in Module 2. Again, these are largely
subjective evaluations of the robustness of each of the scores and reflect the availability and quality of
information for each category, rather than rigorous evaluations. Given that golden-cheeked warbler
autoecology is relatively well known, most of the ecological components of Module 2 were assigned
high certainty scores. However, the physiological relationships between the species and climate
variables are less well understood and were generally assigned medium scores.
In Module 3, the scores from Modules 1 (Vb2) and 2 (Vcl) are combined in an integrative
matrix to give an overall vulnerability score of Vol (likely to be critically vulnerable to future changes
in existing stressors in conjunction with climate change). It is important to note that even using the
alternate scores, the species still is allocated a score of Vol in this module. This implies that the
species is likely to be particularly vulnerable.
In Module 4, the individual variable certainty scores from Modules 1 and 2 are combined in an
integrative matrix into overall evaluation of certainty of High. This implies that the vulnerability
evaluation for golden-cheeked warblers performed in Modules 1 through 3 is robust.
A.11. POTENTIAL EFFECTS OF CLIMATE CHANGE ON STATUS AND DISTRIBUTION
Golden-cheeked warblers are particularly vulnerable to climate change. They are already under
considerable stress due to habitat loss and fragmentation. A population viability analysis performed by
the U.S. FWS and the U.S. Geological Survey (U.S. GS) National Biological Service (now the U.S.
GS Biological Resources Division) in 1995 (U.S. FWS, 1996) concluded that if the current population
was to decline to below 3,000 breeding pairs (the lower end of current population estimates), the risk
of extinction would become unacceptably high. Given the likely constraints on the ability of their
breeding habitat to shift in response to climate change (plus the potential for an increased incidence of
wildfire), it is likely that increased temperatures and drought in the future will lead to further
fragmentation and loss of the species' habitat.
If, as seems likely, the current population is between 10,000 and 30,000 individuals, and the
level at which comparatively high risk of extinction would occur is about 6,000 individuals, loss of
less than 50% of the species' breeding habitat could alter its population dynamics to the extent that
extinction or near extinction may become not unlikely.
A.12. UNCERTAINTIES
There are two main areas of uncertainty inherent in predicting the likely effects of climate
change on golden-cheeked warblers:
49
-------
• Uncertainty associated with assessing the species' physiological sensitivity to increased
temperatures and changes in precipitation patterns. On the basis of the evidence presented in
Section A.4 of this narrative, it is assumed that the relationship between temperature and the
species' physiology is moderately sensitive. It is also assumed that the species is insensitive to
changes in precipitation, though uncertainty associated with this score is high.
• Uncertainty associated with predicting future changes in the frequencies of fire and drought
and the likely effects of these on the distribution and quality of the warbler's breeding habitat.
While the GCMs can predict regional changes in the main climate variables, temperature, and
precipitation, they are generally less useful in predicting climate-associated change in
stochastic events such as droughts or fires. Drought is likely an important determinant of the
extent and quality of the golden-cheeked warbler's breeding habitat (as was shown by the
effects of the drought in the 1950s [see Section A.7]). If climate change in central Texas
resulted in a substantially increased frequency of fire and/or drought, the implications for the
future survival of golden-cheeked warblers could be serious.
• Also, the effects of climate change on human populations within the range of the species are
uncertain. If it results in changes in the way that humans exploit the landscape (through, for
example, changes in water availability), anthropogenic pressure on golden-cheeked warblers
might be altered.
A. 13. SUMMARY
Golden-cheeked warblers are currently listed as Endangered under the ESA. Their current
world population is probably less than 30,000 individuals, and they are confined to a relatively small
breeding range in Central Texas, where they depend on the existence of mature Ashe juniper
woodlands for their nesting habitat. Because of their highly restricted habitat requirements, historical
and continuing losses in their breeding habitat and its current fragmentation, and the fact that the
distribution of the habitat is probably limited by surface geology (and unlikely, therefore, to shift in
response to a changing climate), golden-cheeked warblers are likely to be Critically Vulnerable to
future climate change.
A.14. REFERENCES
APPG/USGS (American Association of Petroleum Geologists/U.S. Geological Survey). (1973)
Geological highway map of Texas. Tulsa, OK: AAPG.
Amos, BB; Gehlbach, FR. (eds.). (1988) Edwards Plateau vegetation. Plant ecological studies in
central Texas. Waco, TX: Baylor University Press.
Britingham, MC: Temple, SA. (1983) Have cowbirds caused forest songbirds to decline? Bio-Science
33:31-35.
Curson, J: Quinn, D; Beagle, D. (1994) Warblers of the Americas. Boston: Houghton Mifflin.
50
-------
Diamond, DD. (1997) An old growth definition for western juniper woodlands: Texas Ashe juniper
dominated or codominated communities. USD A, Forest Service, Southern Research Station General
Technical Report SRS-15.
Diamond, DD; Rowell, GA; Keddy-Hector, DP. (1995) Conservation of Ashe juniper (.Juniperus ashei
Bucholz) woodlands of the central Texas hill country. Nat Areas J 15:189-197.
Dunn, JL; Garret, KL. (1997) A Field Guide to Warblers of North America. Houghton Mifflin, New
York.
Eckrich, GH; Koloszar; TE; Goering, MD. (1999) Effective landscape management of brown-headed
cowbirds at Fort Hood, Texas. Studies in Avian Biology 18:267-274.
Kaufman, K. 1996. Lives of North American birds. Houghton Mifflin, Boston.
Keddy-Hector, DP. (1992). Golden-cheeked warbler recovery plan. Austin, TX: U.S. Fish and
Wildlife Service.
Kuchler, AW. (1975) Potential natural vegetation of the conterminous United States. New York, NY:
American Geographical Society.
Ladd, C; Gass, L. (1999) Golden-cheeked Warbler (Dendroica chrysoparia). In: Poole, A; F. Gill, F.;
eds. The birds of North America, No. 420 Philadelphia, PA: The Birds of North America, Inc.,.
Martin, T.E. (2001) Abiotic vs. biotic influences on habitat selection of coexisting species, with
implications for climate change. Ecology 82: 175-188.
Pulich, W.M. (1976) The Golden-cheeked Warbler: a bioecological study. Austin: Texas Parks and
Wildlife Department.
Sexton, C., Balcones National Wildlife Reserve, 2007. Personal communication with Hector
Galbraith, Manomet Center for Conservation Sciences.
U.S. FWS (U.S. Fish and Wildlife Service). (1990) Endangered and threatened wildlife and plants;
proposed rule to list the Golden-cheeked Warbler as endangered. Fed Reg 55(87): 18846-18849.
U.S. FWS (U.S. Fish and Wildlife Service). (1996). Golden-cheeked Warbler population and habitat
viability assessment report. Compiled and edited by C. Beardmore, J. Hatfield, and J. Lewis in
conjunction with workshop participants. Report of a August 21-24, 1995 workshop arranged by the
U.S. Fish and Wildlife Service, Austin, TX.
VEMAP Members, (1995) Vegetation/ecosystem modeling and analysis project: Comparing
biogeography and biogeochemistry models in a continental-scale study of terrestrial ecosystem
responses to climate change and CO2 doubling. Glob Biogeochem cycles 9:407-437.
Wahl, R; Diamond, DD; Shaw, D. (1990) The Golden-cheeked Warbler; a status review. Final report
submitted to Office of Endangered Species, U.S. Fish and Wildlife Service, Albuquerque, NM.
51
-------
APPENDIX B
EXAMPLE NARRATIVE FOR BALD EAGLE
B.l. INTRODUCTION
Based on a review of available information, this narrative describes the current vulnerability of
the bald eagle to severe population reduction or extinction, and its potential future vulnerability under
climate change. Its main objectives are to
• make transparent the rationale underlying each score in Modules 1 and 2;
• identify main sources of uncertainty;
• identify and describe the roles of the main stressors in the estimate of vulnerability for this
species; and
• qualitatively describe potential population responses to climate change and other stressors.
B.2. ENDANGERED SPECIES ACT STATUS
The bald eagle was initially listed for protection under the Bald Eagle Protection Act in 1940.
Subsequently, birds in the United States were listed as Endangered in 1966 under the Endangered
Species Protection Act. The population in the contiguous states (but not Alaska) was listed as
Endangered under the Endangered Species Act (ESA) of 1973, but subsequently downgraded to
Threatened in 1995. In 1999, because of spectacular population recovery in the previous three
decades, the U.S. Fish and Wildlife Service proposed the species for de-listing from the Endangered
Species Protection Act. However, at present, the species is still listed.
B.3. DISTRIBUTION, STATUS, AND POPULATION TRENDS
At the beginning of European colonization, the bald eagle bred in Alaska and all but three
(Rhode Island, Virginia, and Vermont) of the contiguous states (Buehler, 2000). At that time, densities
of breeding eagles were locally very high; on Chesapeake Bay, for example, it is likely that between
3,000 and 8,000 pairs nested (Fraser et al., 1996; Buehler, 2000). Beginning after European settlement,
the species was persecuted to the extent that its population in the contiguous states was greatly
reduced. By the mid-1930s, only 39 nests remained in Chesapeake Bay (Tyrrel, 1936 cited in Buehler,
2000). Beginning in the mid-1940s, the effects of direct human persecution were exacerbated by the
introduction of the pesticide dichlorodiphenyltrichloroethane (DDT), the metabolite of which,
Dichlorodiphenyldichloroethylene (DDE), resulted in embryo mortality and further widespread
population decreases (Nisbett, 1989; Weimeyer et al., 1993). By the mid-1950s and 1960s, only about
400 breeding pairs remained in the entire contiguous states.
52
-------
With the banning of DDT in 1972, the North American breeding population of bald eagles
began a spectacular increase. Populations increased within areas in which the bird had persisted, while
areas from which the bird had been extirpated were recolonized. Between the mid-1960s and the mid-
1980s, the breeding population in the contiguous states increased from approximately 400 pairs to
about 2,000 pairs, and by the late 1990s, there were almost 6,000 pairs (Buehler, 2000). At present,
there are probably more than 100,000 individuals in the whole of the United States and Canada
(Buehler, 2000), and this number is still increasing.
Currently, the bald eagle breeds across Canada in the boreal forest from Newfoundland to
Vancouver. In the United States, it is a scattered breeder in the interior and Great Plains states, but it is
well distributed within coastal states from Alaska to California and Maine to Florida and Texas. There
is also a small breeding population as far south as Baja, Mexico (Buehler, 2000).
B.4. HABITAT
Within certain broad limits, the bald eagle is a habitat generalist (Kaufman, 1996). Any
forested area that has a suitable body of water close by, with fish, mammalian, or bird prey available
meets its two main requirements, nesting and foraging sites. Such areas range from reservoirs or
coastal areas with adjacent deciduous forests in the east coast states, to coniferous boreal forest on
inland lakes or rivers or on the seacoast in the west, to riparian deciduous corridors in the Great Plains
and Arizona, to forested subtropical swamps in Florida. In one area in Alaska, where trees are not
available, bald eagles even nest on the ground (Buehler, 2000).
The species' flexibility in choice of nest sites is matched by its selection of prey. Bald eagle
diets can range from exclusively fish to a mixture of fish, mammals, and birds (Buehler, 2000). In
many areas, they are largely dependent on carrion, or on prey stolen from other species (Stalmaster,
1987).
B.5. PHYSIOLOGICAL/CLIMATIC LIMITATIONS ON DISTRIBUTION AND STATUS
No data have been found that suggest that climate imposes a physiologic limit on the
distribution of bald eagles. Indeed, their wide historic and current ranges (from Sonoran deserts in the
southwest to the northernmost limit of tree line in Alaska and northern Canada) suggest that the
species may be tolerant of a wide range of temperature and precipitation regimes.
It is possible that the southern limit of the bald eagles' range in Mexico, Texas, New Mexico,
and Arizona may be at least partly a function of direct temperature or precipitation effects on the birds.
There is, however, no evidence to support this, and the range limit may equally likely be a function of
habitat availability.
53
-------
B.6. ECOLOGICAL LIMITATIONS ON DISTRIBUTION AND STATUS
The main ecological limitations on the distribution and status of bald eagles are the availability
of nest sites and prey (Stalmaster, 1987). However, the birds are flexible in their choice of both (see
Section B.3), which is why their historic and current ranges were, and are, so extensive.
B.7. EXISTING STRESSORS AND THEIR EFFECTS ON DISTRIBUTION AND STATUS
The species was historically limited by direct human persecution and the toxicological effects
of organochlorine pesticides. However, bald eagles are no longer persecuted on any large scale, and
the lingering effects of DDE are reduced to the extent that previously affected populations (such as the
Great Lakes coastal birds) may now be as productive as "clean" populations.
Some breeding sites may suffer reduced productivity due to human recreational disturbance
(Buehler, 2000). However, such effects are local rather than regional in their occurrence and do not
greatly affect the distribution and status of the species. Currently, bald eagles do not appear to be
limited by anthropogenic or natural stressors to any great extent.
B.8. POTENTIAL DIRECT (PHYSIOLOGICAL) VULNERABILITY TO CLIMATE
CHANGE
Given the bald eagle's presumed tolerance of a wide range of temperature and precipitation
regimes (see Section B.4), it is unlikely that the range of climatic changes projected in General
Circulation Models (GCMs) would exercise any great effect on bald eagles throughout most of their
current range. The only area where this might not be true may be at their current southernmost range
limits, where increased temperature might conceivably directly affect the birds. This, however, is
largely conjectural.
B.9. POTENTIAL INDIRECT (ECOLOGICAL) VULNERABILITY TO CLIMATE
CHANGE
The majority of bald eagles breeds and winters in the temperate and boreal forest zones. The
main likely climate change effect on this biome would likely be a northward shift into areas that
hitherto were arctic tundra, with a corresponding northward retraction of range in the south (Neilson
and Drapek, 1998; Bachelet and Neilson, 2009.). Thus, bald eagle habitat could be forced by climate
change to shift northward. However, although this would certainly mean redistribution, it is not certain
what, if any, effects this would have on the North American status of the species. It is unlikely,
however, that such a redistribution would put the continental population at any great risk of extinction.
Bald eagle breeding distribution in the Great Plains of the United States is fragmented,
probably by the sporadic distribution of suitable water-bodies in the largely arid landscape. Some
GCMs for this region predict increased aridity. This could result in localized loss of aquatic habitats
54
-------
and further fragmentation of the regional range of the population. However, the vast majority of bald
eagles breed in the north and in Alaska, not the Great Plains. Thus, any effect of climate change on
Great Plains habitat would affect only a minority of birds.
Other potential local or regional scale effects of climate change could be mediated through
changes in food supply. Many eagles in the Pacific Northwest and in Alaska are dependent during the
breeding season and fall on runs of salmonids (Buehler, 2000). Welch et al. (1998) predict that
increasing sea temperatures in the northern Pacific Ocean could lead to population reductions and
extinction of salmon species. This could have repercussions for the bald eagle populations in the area,
though the extent to which they would be able to switch to an alternative food supply is not known.
B.10. JUSTIFICATIONS FOR FRAMEWORK SCORES
The framework scores for the bald eagle are presented in Appendices G through J.
In Module 1 (baseline vulnerability), the species scored Vb3, indicating that it is among the
less vulnerable of T&E species. This score is based on the following subcomponents:
• Current population size and trends—based on most recent census data (see Section B.2) there
are likely to be more than 100,000 individual bald eagles in North America. Therefore, the
species scores 6 in the current population size variable of Module 1. In the past, the North
American bald eagle population was reduced by at least 50% so it has been allocated a best
estimate score of 2 in Variable 2. More recently, the species has been and is increasing
throughout its North American range, thus it has been given a best estimate score of 4 in the
current population trend variable.
• Past and current range trends—prior to the recent population resurgence, the bald eagle was
widely extirpated from much of its historical range. In Variable 4, a best estimate score of 3
(>20% reduction) is conservatively applied, with an alternate score of 2. More recently (the
past 2-3 decades), the species has been extending its range. Consequently, it has been allocated
a best estimate score of 4 (increasing) in Variable 5, with an alternate score of 3 (stable).
• Likely future stressor trends—until recently, the main stressors acting on bald eagles were
habitat destruction and the toxicological effects of organochlorine pesticides. With regulatory
protection and the banning of DDT, these stressors have been ameliorated. It is likely that these
contaminants will continue to diminish in their effects. Thus, the species has been allocated a
best estimate score of 3 (reduction in stressors) with an alternate score of 2 (stable).
• Individual replacement time—the bald eagle reproduces slowly (individuals do not breed until
they are 4-5 years old, and the maximum number of young that are reared per annum is 3, with
most pairs only producing 1 or 2). Thus, the potential replacement time for individuals is 4-5
years. It has, therefore, been allocated a best estimate score of 2 in Module 2.
• Likely future vulnerability to stochastic events—with a widespread distribution and relatively
large population, and an adult lifespan of decades, the bald eagle is relatively non-susceptible
to, and able to withstand the adverse impact of sporadic events such as temporary food
55
-------
shortage, or nest site destruction. Therefore, it has been allocated a best estimate score of 3 (not
vulnerable).
• Vulnerability to changes in policy/management—because of its relatively large and increasing
numbers and widespread range, the bald eagle is likely to be less susceptible than other T&E
species to changes in land management or future policy. It has been allocated a best estimate
score of 2, with an alternate score of 3.
• Future vulnerability to natural stressors—the incidence of disease, parasitism, or other natural
stressors on this species is not known. However, given that the population is large and
widespread in distribution, it is unlikely that such a stressor could have more than local
impacts. Therefore, the species is allocated a best estimate score of 3 (not vulnerable), with an
alternate score of 2 (vulnerable).
Certainty categories were allocated to each of the scores in Module 1. These are largely
subjective evaluations of the robustness of each of the scores and reflect the availability of information
for each category, rather than rigorous evaluations. Given that bald eagles are a well studied species, a
high certainty score was assigned to most variables, with medium scores allocated to the remainder.
In Module 2, the bald eagle scored Vc3, indicating that while it is not likely to be completely
immune to climate change, it is unlikely to be fundamentally affected to the point that its extinction
risk is greatly increased. This score is based on the following subcomponents:
• Physiological sensitivity to temperature and precipitation change and to extreme weather
events—given its ability to thrive in areas of markedly different temperature and precipitation
regimes, and the likely localized focus of extreme weather events (relative to the species'
extensive range), it is unlikely that bald eagles will be sensitive to these direct weather
variables. Thus, it has been allocated best estimate scores of 3 (likely insensitive), with
alternate scores of 2 (only moderately sensitive).
• Dispersive capability and potential rate of increase—the dispersive capability of this mobile
species is high (and scores 3 in Module 2).
• Habitat specialization—bald eagles are largely dependent on one habitat type (wooded coastal,
lake, or river margins). Within these broad habitat types, they are flexible in their habitat use.
For these reasons, the species scores 3 (best estimate) and 2 (alternate estimate).
• Likely extents of future habitat loss due to climate change—bald eagles score 2 (20-50%
habitat loss). This score reflects the possibility of habitat change in the arid west areas of the
species' range. Such changes are less likely elsewhere. Given the species' habitat flexibility,
this score may be over-conservative, and an alternate score of 3 (no change) has also been
applied.
• Ability of habitats to shift at same rate as the species in response to climate change—it is
assumed in this scoring that the species' main habitat (conifer and deciduous forest) will be
able to shift northward in response to climate change but only slowly (relative to the rate at
which eagles may shift).
56
-------
• Availability of habitats within new range—bald eagles currently occupy all forested habitat
north to the tree line. Therefore, this entire extensive boreal forest habitat is suitable for
colonization by individuals forced to move north by climate change, and the species is
allocated a best estimate score of 3.
• Dependence on temporal inter-relations and other species—except in the limited case of
salmon runs, bald eagles are unlikely to be dependent on such factors and, accordingly, score 3
(best estimates) in each of these variables.
Certainty categories were allocated to each of the scores in Module 2. Again, these are largely
subjective evaluations of the robustness of each of the scores and reflect the availability of information
for each category, rather than rigorous evaluations. Given that bald eagle autoecology is relatively
well known, the ecological components of Module 2 were generally assigned high or medium
certainty scores. However, the physiological relationships between bald eagles and climate variables
are less well understood and were generally assigned medium scores. No low certainty scores were
assigned.
In Module 3, the scores from Module 1 (Vb3) and Module 2 (Vc3) are combined in an
integrative matrix to give an overall vulnerability score of Vo3 (likely to be among those T&E species
that are less vulnerable to climate change).
In Module 4, the individual variable certainty scores from Modules 1 and 2 are combined in an
integrative matrix into an overall evaluation of certainty of High. This implies that the vulnerability
evaluation for bald eagles performed in Modules 1 through 3 is robust.
B.11. POTENTIAL EFFECTS OF CLIMATE CHANGE ON STATUS AND DISTRIBUTION
Among T&E species, bald eagles are likely to be among the least vulnerable to climate change.
This is due to their currently burgeoning populations and range extension (since their main
anthropogenic stressors were reduced), the flexibility of their habitat preferences, and the likely ability
of their main habitats to survive climate change. Except in the Great Plains, where their habitat is
already limited and fragmented by surface water distribution, radical population reductions and/or
extinctions due to climate change are not expected.
B.12. UNCERTAINTIES
The main areas of uncertainty in the bald eagle analysis are associated with their potential
direct/physiological sensitivity to climate change. If the species is at or close to abundance and
distributional limits set by temperature of precipitation patterns, then it may be more vulnerable, at
least in the southern and interior components of its range, than this analysis suggests.
57
-------
B.13. SUMMARY
Bald eagles are currently listed as Threatened under the ESA. Their current North American
(and world) population is probably more than 100,000 individuals, and they breed and winter in a
variety of habitats throughout the contiguous states and Alaska. Because of their relatively flexible
habitat and diet requirements and their currently increasing populations, bald eagles have been scored
as Less Vulnerable (the second-least vulnerable of the potential scores) to future climate change.
B.14. REFERENCES
Bachelet, D; Neilson, RP. (1999) Biome redistribution under climate change. Gen. Tech. Rep. PNW-
GAR. USD A PNW Res. Stat. Oregon
Buehler, DA. (2000). Bald eagle (Haliaeetus leucocephalus). In The birds of North America, No. 506
(A. Poole and F. Gill, eds.). The birds of North America, Inc., Philadelphia, PA.
Fraser, JD; Chandler, SK; Buehler, DA; Seegar, JKF. 1996 The decline, recovery and future of the
Bald Eagle population of the Chesapeake Bay, USA. In Eagle Studies (B.U. Meyburg and R.D.
Chandler, eds) World Working Group for Birds of Prey, Berlin, Germany
Kaufman, K. 1996. Lives of North American birds. Houghton Mifflin Company, Boston, MA.
Neilson, RP; Drapek, RJ. (1998) Potentially complex biosphere responses to transient global warming.
Global Change Biology 4:505-521.
Nisbet, ICT. (1989) Organochlorines, reproductive impairment, and declines in bald eagle, Haliaeetus
leucocephalus, populations: mechanisms and dose relationships. In: Meyerberg, BU; Chancellor, RD,
eds. Raptors in the modern world. Proceedings of the 3rd world conference on birds of prey and owls,
Berlin, London and Paris, ISBN 3-9801961-0-0,. Pp. 483-489.
Stalmaster, M.V. 1987. The bald eagle. Universe Books, New York.
Weimeyer, SN; Burick, CM; Stafford, CJ. (1993). Environmental contaminants in bald eagle eggs -
1980-1984- and further interpretations of relationships to productivity and shell thickness. Arch
Environ Contam Toxicol 24:213-227.
Welch, DW; Ishida, Y; Nagasawa, K. (1998) Thermal limits and ocean migrations of sockeye salmon
(Oncorhynchus nerka): long-term consequences of global warming. Can J Fish Aquat Sci.
55:937-948.
58
-------
APPENDIX C
EXAMPLE NARRATIVE FOR SALT MARSH HARVEST MOUSE
C.l. INTRODUCTION
Based on a review of available information, this narrative describes the current vulnerability of
the salt marsh harvest mouse, Reithrodontomys raviventris, to severe population reduction or
extinction, and its potential future vulnerability under climate change. It is intended to provide support
for the scores listed in Modules 1 through 4 with the specific objectives of
• making transparent the rationale underlying each score in Modules 1 and 2;
• identifying main sources of uncertainty;
• identifying and describing the roles of the main stressors in the estimate of vulnerability for
this species; and
• qualitatively describing potential population responses to climate change and other stressors.
C.l. ENDANGERED SPECIES ACT STATUS
The salt marsh harvest mouse was listed as Endangered under the Federal Endangered Species
Preservation Act in 1970. The species was later listed as Endangered under that act's successor—the
Endangered Species Act (ESA), in 1973. The State of California also extended Endangered status to
the species under its ESA in 1971.
C.3. DISTRIBUTION, STATUS, AND POPULATION TRENDS
The salt marsh harvest mouse is endemic to salt marshes in San Francisco Bay and occurs in
the southern Bay, the central Bay, San Pablo Bay, and Suison Bay (U.S. FWS, 1984). Because the
species is difficult to census, there have been few attempts to estimate its population size. However,
based on live-trapping results, the entire population at its midsummer peak has been estimated as
probably no larger than a few thousand individuals (U.S. FWS, 1984). There may be less than this
once postbreeding juvenile mortality has occurred.
Historically, there have likely been large reductions in the population status of the salt marsh
harvest mouse due to habitat loss (U.S. FWS, 1984). Between 1850 and the 1960s, approximately 80%
of its salt marsh habitat in San Francisco Bay was converted to agricultural or urban use (see Section
C.7). The rate of habitat loss has slowed but continues in some areas due to land subsidence and
inundation of tidal lands.
59
-------
C.4. HABITAT
Within their restricted range and throughout their life cycle, salt marsh harvest mice are largely
confined to intertidal salicornia flats. This is the tidal zone that is intermediate between the lowest
vegetated intertidal community (spartina beds) and the upper, intermittently flooded high marsh, a
community dominated by fewer obligate halophytic plant species. They have their highest densities in
salicornia flats that have minimal densities of other plant species (i.e., monocultures of salicornia).
Salicornia plants and grasses comprise the greatest part of the harvest mouse diet.
During extreme high tides, salt marsh harvest mice leave the salicornia flats and take refuge
from the rising water by moving up into the high marsh or, beyond that, into upland vegetation. Thus,
the salt marsh harvest mouse is a habitat specialist, and its future status and population viability are
largely a function of the fate of its salicornia habitat, and access to its high tide escape habitat above
the salicornia zone.
C.5. PHYSIOLOGICAL/CLIMATIC LIMITATIONS ON DISTRIBUTION AND STATUS
Salt marsh harvest mice are relatively slow moving and inactive rodents (Fisler, 1968). Perhaps
because of this, and their consequent vulnerability to predators, they avoid open areas where they
cannot find cover. This physiological/behavioral characteristic may limit their ability to colonize new
areas and may be one of the reasons why their distribution is highly limited and patchy (U.S. FWS,
1984).
C.6. ECOLOGICAL LIMITATIONS ON DISTRIBUTION AND STATUS
The distribution of salt marsh harvest mice is confined to tidal salt marshes in the
San Francisco Bay complex. The species has never been found outside of this area, despite the fact
that its salicornia habitat is extensive along the Californian coast. Even within its range, the species is
patchily distributed, with highest densities in salicornia marshes that are largely monocultures and that
are connected to higher elevation high tide escape habitat. Areas where the salicornia marsh has been
invaded by brackish or fresh water plants tend to have fewer or no mice, as do areas where access to
escape habitat has been closed (e.g., by diking the salt marsh). Thus, harvest mice are indicators of
comparatively pristine and functioning salt marshes.
C.7. EXISTING STRESSORS AND THEIR EFFECTS ON DISTRIBUTION AND STATUS
The current distribution and status of the salt marsh harvest mouse is mainly a function of its
own habitat specialization and habitat destruction and modification by humans. Although there is no
evidence that it has ever occurred outside of San Francisco Bay, it was once much more widely
distributed and abundant within that area. Prior to post-Columbian colonization, the salt marshes of
60
-------
the bay were much more extensive than they are now, covering about 730 km2 (U.S. FWS, 1984; San
Francisco Estuary Institute: http://www.sfei.org/ecoatlas/index.html). Beginning in the 1850s, much
of this marshland was diked and reclaimed for agriculture or urbanization. In addition, mining in the
Sierra Nevada in the second half of the 19th Century resulted in large quantities of silt being washed
downstream and deposited in the Suison and San Pablo bays. Later, in the 20th Century, groundwater
withdrawals for human use triggered land subsidence in the southern part of the bay, with consequent
increased inundation and habitat loss. At the same time that this was happening, large areas of the
southern bay were diked and used as salt pannes. The net result of this was that by the middle of the
20th Century, approximately 80% of the historical marshes had been lost or highly modified (San
Francisco Estuary Project, 1992; U.S. FWS, 1984).
While the habitat destruction and modification that occurred prior to the 1960s has slowed (and
even been reversed in some areas through habitat management and restoration), there continue to be
anthropogenic effects on the salt marshes that are detrimental to the harvest mice: management of
some areas for waterfowl populations has resulted in the replacement of salicornia flats with plant
species preferred by waterfowl but unacceptable to harvest mice (U.S. FWS, 1984). Also, land
subsidence in the southern bay continues. This, together with freshwater outflows from the cities that
line the southern bay, have changed the salinity patterns of inshore habitats and the vegetation
communities, away from the salicornia flats preferred by the harvest mice.
C.8. POTENTIAL DIRECT (PHYSIOLOGICAL) VULNERABILITY TO CLIMATE
CHANGE
No data have been found on the likely physiological tolerances of salt marsh harvest mice to
changes in temperature or precipitation. Therefore, it is not possible to comprehensively assess their
potential direct vulnerability to changes in climate. However, perhaps because it is relatively slow
moving, the species avoids crossing open spaces. This physiological trait may limit its ability to move
to and colonize new areas when their present habitats are affected or modified by climate change.
C.9. POTENTIAL INDIRECT (ECOLOGICAL) VULNERABILITY TO CLIMATE
CHANGE
The climate change component to which the salt marsh harvest mouse is likely to be most
vulnerable is sea level rise, with consequent inundation of their current habitats. Titus and Narayanan
(1995) have estimated that there is a 50% probability that sea level in the southern bay will rise by
approximately 2 m by the year 2100. Even without factoring climate change into this calculation, Titus
and Narayanan (1995) estimate that current rates of land subsidence will result in a 1.5-m rise in sea
level in the southern bay by 2100. In the northern part of San Francisco Bay (where such drastic land
subsidence is not occurring), Titus and Narayanan (1995) estimate a 50% probability of about 0.4-m
61
-------
rise in sea level by 2100. The extensive network of dikes and salt pannes in the southern bay limits the
ability of the estuary to simply move inland in response to sea level rise.
Galbraith et al. (2002) used the Titus and Narayanan (1995) projections to model changes in
the extents of intertidal habitats in the southern and northern parts of the bay. They project that a 2-m
rise in sea level in the southern bay will result in a 63% reduction in the current area of salt marsh (a
1.5-m sea level rise will result in a 50% loss). By the year 2200, Galbraith et al. (2002) project that salt
marsh habitat loss in the southern bay will exceed 90%. In the northern bay, they project no loss of salt
marsh because in that area, the salt marshes are buffered by the conversion of intertidal mud and sand
flats to subtidal habitats.
These extents of habitat loss in the southern part of the San Francisco Bay could have
extremely important consequences since they are in an area where a large part of the salt marsh
harvest mouse population currently exists.
C.10. JUSTIFICATION FOR FRAMEWORK SCORES
The framework scores for the salt marsh harvest mouse are presented in Appendices G through J.
In Module 1 (baseline vulnerability), the species scored Vb2, indicating that it is currently in a
highly vulnerable condition. This score is based on the following subcomponents:
• Current population size and trends—based on the data that are available (see Section C.2),
there are likely to be, at most, a few thousand individual harvest mice at any one time, perhaps
substantially less. Much of the historical habitat of the species was lost between colonization of
the area by Europeans and the mid part of the 20th Century, and there is also good evidence that
the population may be reducing further as habitat continues to be lost. Because of this, the
species scores 4 (best estimate) in the population size variable and 2 in the population trend
variable.
• Range trends—the U.S. range of salt marsh harvest mice has always been confined to salt
marshes in San Francisco Bay. The majority of this habitat was lost in the 19th and
20th centuries, and it continues to be lost, though at a slower rate. Thus, the species has been
allocated best estimate scores of 2 in each of these variables.
• Likely future stressor trends—the main stressor that has reduced salt marsh harvest mice
populations in the past, and that continues to do so, is anthropogenic habitat destruction or
modification. Although the rate of loss is slower than in the past, it still continues. Also, land
subsidence in the southern bay (caused by aquifer depletion) is likely to continue, if not
increase, as the area becomes more developed. This will result in yet more marshes becoming
inundated. For these reason, the species scores 1 (best estimate) in this variable.
• Individual replacement time—small rodents tend to have replacement times in the order of 1-2
years. No data were found on the population dynamics and reproductive biology of the species,
but it is assumed that individual replacement time is less than 2 years (i.e., a best estimate
score of 3).
62
-------
• Likely future vulnerability to stochastic events—This species is likely to be particularly
vulnerable to high tidal surges. While in the past, individuals could simply move upslope to
avoid drowning, the diking of much of their habitat makes this less feasible. For this reason,
the species scores 1.
• Likely future vulnerability to policy/management changes—many, perhaps most, salt marsh
harvest mice exist on preserves owned by the Federal government (e.g., San Francisco
National Wildlife Refuge) or by the State of California. They are, therefore, somewhat
conservation dependent and have been assigned a score of 2.
• Likely future vulnerability to natural stressors—no information has been found on the
susceptibility of this species to natural stressors. It has been assumed that it is not vulnerable
and allocated a score of 3.
Certainty evaluations were allocated to each of the scores in Module 1. These are largely
subjective evaluations of the robustness of each of the scores and reflect the availability and quality of
information for each category, rather than rigorous evaluations. Most of the scores allocated to this
species were medium, reflecting the lack of information about the species. Some variables scored low,
particularly those pertaining to the extent of past and current population and range trends. This reflects
the lack of information on the species' population status and trends.
In Module 2, the salt marsh harvest mouse scored Vc2, indicating that it is likely to be
critically vulnerable to climate change and that its extinction risk may be increased substantially. This
score is largely based on the extreme habitat specialization of the species, the potential scale of habitat
loss due to sea level rise (>50%), and the likely inability of the species to move to and colonize new
areas.
Certainty scores were allocated to each of the scores in Module 2. Again, these are largely
subjective evaluations of the robustness of each of the scores and reflect the availability and quality of
information for each category, rather than rigorous evaluations. Given that little is known about the
physiological tolerances of the species, its sensitivity to extreme weather events, or its dependencies
on other species or conservation actions, many of these scores are only medium or low.
In Module 3, the scores from Modules 1 (Vb2) and 2 (Vc2) are combined in an integrative
matrix to give an overall vulnerability score of Vol (likely to be critically vulnerable to future trends
in stressors in conjunction with climate change). It should be noted that the alternate estimates
(possible though less likely) are, at best, Vo2. This emphasizes the vulnerability of the species.
In Module 4, the individual variable certainty scores from Modules 1 and 2 are combined in an
integrative matrix into overall evaluation of certainty of Medium. This implies that the vulnerability
evaluation for the salt marsh harvest mouse performed in Modules 1 through 3 is reasonably robust,
though not entirely dependable.
63
-------
C.11. POTENTIAL EFFECTS OF CLIMATE CHANGE ON STATUS AND DISTRIBUTION
The projected extent of salt marsh habitat loss in southern San Francisco Bay due to sea-level
rise could have extremely serious consequences for the population viability of the salt marsh harvest
mouse and its likelihood of extinction. If the southern bay population is catastrophically reduced (as
would be expected from a >50% habitat loss), the future viability of the species would then be solely
dependent on the population in the northern part of the bay, thus reducing still further the species
ability to survive future stochastic events and continued climate change.
C.12. UNCERTAINTIES
There are four main areas of uncertainty inherent in predicting the likely effects of climate
change on salt marsh harvest mice:
• Uncertainty associated with assessing the species' physiological sensitivity to increased
temperatures or altered precipitation patterns. There is no information on the species'
physiological tolerances or sensitivity, and the evaluation of Vol is based largely on the
projected indirect effects of climate change (i.e., acting through habitat modification). If the
species is also physiologically sensitive to climate change, the population effects and risks
could be more profound than assessed.
• Uncertainty associated with our understanding of current population trends. This species is
extremely difficult to census, and its current population trends are not well known. The
assumption for this analysis is a slow decline in population numbers. However, this is not
based on rigorous evidence, and it is also possible that the species is responding to current
stressors at a much faster rate. If so, the extinction risk could be higher than assumed.
• Uncertainty associated with predicting future frequencies and severities of extreme weather
events. If rising sea levels are also accompanied by an increased frequency of on-shore storms
and tidal surges, the risks posed to the salt marsh harvest mice could be greater than anticipated
(especially those segments of the population that have little or no escape habitat. General
Circulation Models do not provide reliable predictions of the likely frequency of extreme
weather events.
• Future trends in human land use in the areas surrounding the bay. Much of the current risk
posed to the harvest mice is due to anthropogenic depletion of the underground aquifer
underlying the southern bay (with resultant land subsidence), the rate of urban development in
the area, and concomitant increases in the rate at which wastewater and sewage is released into
the bay (thereby affecting the floristics of the salt marshes). It is likely that the Californian
human population may approximately double by the year 2100 (Landis 2006). This could
result in accelerated rates of habitat loss even without factoring climate change into the
equation.
C.13. SUMMARY
64
-------
The salt marsh harvest mouse is currently listed as Endangered under the Federal and State of
California ESAs. It is a habitat specialist confined to fragmented salt marshes surrounding San
Francisco Bay, and its current population is likely to be, at most, a few thousand individuals. Its
essential habitat (salicornia marshes) is likely to be at great risk to sea level rise (induced by climate
change and land subsidence). Conservative estimates project that more than 50% of the mouse's
habitat could be lost by 2100 (U.S. FWS 1984). Given its current restricted range, its habitat
specialization, and the potential degree of habitat loss, this analysis has concluded that the future
extinction risk for this species may be high.
C.14. REFERENCES
Fisler, GF. (1968) Adaptations in movement patterns of two species of salt marsh rodents. Bull So
Calif .Acad Sci67:96-103.
Galbraith, H: Jones, R; Park, R; et al.. (2002) Global climate change and sea level rise: potential losses
of intertidal habitat for shorebirds. Waterbirds 25:173-183.
San Francisco Estuary Project. (1992) State of the Estuary. San Francisco Estuary Project, Oakland,
California.
Titus, JG; Narayanan, V. 1995. The probability of sea level rise: U.S. Environmental Protection
Agency. 186 pp. EPA 230-R95-008. Washington DC.
U.S. FWS (Fish and Wildlife Service). (1984) Salt marsh harvest mouse and California clapper rail
recovery plan. U.S. Fish and Wildlife Service, Portland, Oregon.
65
-------
APPENDIX D
EXAMPLE NARRATIVE FOR MOUNT GRAHAM RED SQUIRREL
D.l. INTRODUCTION
Based on a review of available information, this narrative describes the current vulnerability of
the Mount Graham red squirrel, Tamiasciurus hudsonicus grahamensis, to severe population reduction
or extinction, and its potential future vulnerability under climate change. It is intended to provide
support for the scores listed in Modules 1 through 4 with the specific objectives of
• making transparent the rationale underlying each score in Modules 1 and 2;
• identifying main sources of uncertainty;
• identifying and describing the roles of the main stressors in the estimate of vulnerability for
this species; and
• qualitatively describing potential population responses to climate change and other stressors.
D.2. ENDANGERED SPECIES ACT STATUS
The Mount Graham red squirrel was listed as Endangered under the Federal Endangered
Species Act in 1987.
D.3. DISTRIBUTION, STATUS, AND POPULATION TRENDS
The Mount Graham red squirrel is endemic to the Pinalenos Mountains in southern Arizona
(U.S. FWS, 1993). Since the Pinalenos Mountains are the southernmost extreme of the range of
Tamiasciurus hudsonicus, the Mount Graham red squirrel is the southernmost outlier of the species
(Flyger and Gates, 1982; U.S. FWS, 1993). Within these mountains, it is confined to two successive
life zones: the mixed conifer forest from about 8,500 ft in elevation to about 10,000 ft, and the
spruce-fir forest from 10,000 ft to the highest peaks at 10,700 ft. They apparently do not occur in the
open ponderosa pine forests below 8,500 ft. Thus, the subspecies is confined to a fairly narrow
altitudinal zone of about 2,000 vertical ft. Based on fieldwork and review of aerial photographs, the
U.S. Forest Service (1988) estimated that within this zone there were probably approximately 50 km2
of habitat suitable for the squirrels. Thus, not only is the subspecies restricted elevationally, it is also
restricted to a very small horizontal range.
The Mount Graham red squirrel has been censused each year since 1986. During this time, the
entire population has varied from a low of about 150 individuals to about 570. U.S. FWS (1993)
speculates that these fluctuations may be due to variability in the quantity and quality of the cone crop,
66
-------
with the lowest populations during years of cone crop failure (perhaps tied to climatic variability).
During the last decade, numbers have apparently increased, and there have typically been between 300
and 500 individuals counted (http://medusa.as.arizona.edu/graham/envir.html).
D.4. HABITAT
Within their restricted range, and throughout their life cycle, Mount Graham red squirrels are
entirely confined to two main plant associations: mixed conifer and spruce-fir forests. The pine seeds
are the squirrels' main food item. Prior to the 1990s, it was believed that most squirrels lived in the
high elevation spruce-fir association and that the mixed conifer zone was less important for the
survival of the species. This belief was the basis for the designation by the U.S. Forest Service of
1,700 acres of spruce-fir forest above 10,000 ft as critical habitat for the subspecies. More recently,
however, census data indicate that the mixed forest zone may be more important for the species than
hitherto believed (http://medusa.as.arizona.edu/graham/envir.html). Irrespective of the relative
importance of the two associations, the squirrels are confined to a narrow zone of conifer habitat
between 8,500 and about 10,000 ft.
Within their area of occurrence, Mount Graham red squirrels apparently prefer conifer forests
that are relatively dense, with a closed canopy, and composed of mature or old-growth trees (U.S.
FWS, 1993). This may reflect their need for relatively cool and moist food storage sites ("middens"),
where their stored food will not decay and where fungal growth (another important food item) can
occur. It might also indicate a physiological requirement on the part of the squirrel, itself.
Mount Graham red squirrels do not occur much below 8,500 ft in the ponderosa pine forests,
although red squirrels use this habitat further to the north within their range. This avoidance may be
due to the high degree of solar insolation at the low latitude of the Pinal enos (U.S. Forest Service,
1988). Insulation could act either as a direct climatic limitation on elevational range of the squirrel
(i.e., acting through its physiology), or indirectly (by limiting the availability of cool, moist midden
sites).
D.5. PHYSIOLOGICAL/CLIMATIC LIMITATIONS ON DISTRIBUTION AND STATUS
The extent to which the current range of red squirrels, and Mount Graham red squirrel, in
particular, is limited by relationships between physiology and climate is not known. However, they are
adapted to moist, cool forests, and the Pinalenos Mountains are the southernmost outpost of the
species. It is feasible that they are thermally limited from existing any further south. Also, at this
southern extreme of their range, they do not inhabit the lower ponderosa pine association, though they
do so further north in their range. This also could indicate a direct thermal constraint on their
distribution. However, it is also feasible that their range and habitat preferences in the Pinalenos
67
-------
Mountains might be an indirect effect of climate, acting through some attribute of habitat quality, such
as latitudinal and elevational patterns in canopy cover, limiting the availability of cool, moist midden
sites. Thus, red squirrels might not occur any further south because they cannot physiologically
tolerate higher temperatures and/or because their habitat is unsuitable.
D.6. ECOLOGICAL LIMITATIONS ON DISTRIBUTION AND STATUS
Red squirrels probably colonized southern Arizona during the last glaciation, when contiguous
stretches of spruce-fir forest extended farther south. With subsequent warming temperatures, it is
likely that the conifer and spruce-fir forests on individual mountain ranges retracted and became
isolated by intervening tracts of either Sonoran or Chihuahuan desert (Gelbach, 1981) that the
squirrels were unable to cross. It was this that led to the reproductive isolation of the Pinalenos
population of red squirrels and the eventual evolution of the Mount Graham subspecies.
It is likely that the current distribution of the Mount Graham red squirrel is at least partly due
to habitat limitation. They are at the southernmost limit of the spruce-fir forest vegetation complex,
which is the main habitat of red squirrels in general. This habitat does not extend further south into
Mexico, where it is replaced by mixed forest or drier pine forests (Barbour and Billings, 1988). The
fact that Mount Graham red squirrels do not occur further to the north in other mountain ranges is
explained by the isolation of the mountain ranges in the region. Within their restricted range, they are
also limited elevationally by habitat: the lower slopes of the Pinalenos Mountains support much more
open ponderosa pine woodland, which apparently, is unsuitable for the squirrel at these latitudes
(though not further north). Thus, the Mount Graham red squirrel is confined to a relatively small area
of Arizona by their strict habitat requirements, by the fragmented distribution of that habitat, and
perhaps by their thermal tolerances (see Section D.5).
D.7. EXISTING STRESSORS AND THEIR EFFECTS ON DISTRIBUTION AND STATUS
The main current limitations on the status and distribution of Mount Graham red squirrels are
all factors that affect the spatial extent and productivity of the spruce-fir and mixed conifer forests in
southern Arizona. Thus, logging (prior to the Pinalenos Mountains being declared a refuge area by the
U.S. Forest Service) probably exerted a limiting effect on the squirrel's distribution. Also, fire,
although not frequent in the high, cool, and damp spruce-fir forests, probably also has limited their
distribution and population status. The main factor that currently affects their numbers within their
small range seems to be the size of the annual cone crop. In years of cone shortage, the squirrel
population is reduced (probably through reproductive failure or mortality of juvenile animals). Thus,
any factors that caused their preferred conifer habitats to shrink, that increased the risk of catastrophic
68
-------
fires, or that increased the frequency of poor cone crops could have detrimental effects on the
population viability of the squirrel.
D.8. POTENTIAL DIRECT (PHYSIOLOGICAL) VULNERABILITY TO CLIMATE
CHANGE
No data have been found on the likely physiological tolerances of Mount Graham red squirrels
to changes in temperature or precipitation. However, there is evidence that squirrels avoid areas of
high solar insolation. Also, their distribution may suggest a thermal constraint acting on their
physiology (see Section D.5). Alternatively, or in addition, these attributes could be due to increased
levels of insolation further south, or in the more open ponderosa pine forests, rendering the squirrels'
habitat less suitable for food storage. If the former explanation is true, increased temperatures due to
global climate change could directly affect the ability of the squirrels to persist in these southernmost
areas of their range (especially if precipitation patterns also change, increasing the "droughtiness" in
the squirrels' habitat).
D.9. POTENTIAL INDIRECT (ECOLOGICAL) VULNERABILITY TO CLIMATE
CHANGE
As noted above (Section D.7), any factors that reduce the extent of the mixed conifer and
spruce-fir habitat of Mount Graham red squirrels, or that reduce the cone crop could adversely affect
the status and distribution of the subspecies.
Batchelet and Neilson (1999), using the MAPSS model and 4 different GCM outputs, showed
that in all cases temperate evergreen forest (which includes mixed conifer and spruce-fir associations)
was eliminated in southern Arizona and replaced by either mixed forests or shrub woodland (e.g.,
pinon-juniper associations). Moreover, assuming an elevational lapse rate of about 1°C for every 120
m (the measured temperature lapse rate at Niwot Ridge, Colorado), it would take only a 5°C annual
average temperature increase, which is within the range expected to occur under a CO2 doubling, to
entirely eliminate the squirrel's preferred habitats in the Pinalenos Mountains. Even if the annual
average temperature did not increase by as much as 5°C, increases that are more modest could still
result in the extinction of the Mount Graham red squirrels, as their distribution is reduced and
fragmented to the tops of the highest peaks.
Global warming could also adversely affect the habitat of the squirrel short of eliminating it
entirely. Southern Arizona is an area where a number of GCMs (e.g., the Hadley, Canadian Climate
Center, and Geophysical Fluid Dynamics Laboratory (GFDL) models predict warming. If, as may be
likely, this warming resulted in an increased frequency, intensity, or duration of drought, increased
forest fire frequency could also adversely affect the squirrels. Increased temperatures could also
potentially affect the overwinter survival of insect pests and, thereby, result in more frequent
69
-------
outbreaks and tree mortality. This also could reduce the amount of habitat available to Mount Graham
red squirrels. The environmental mechanisms responsible for local pine cone crop failures are little
understood and may vary from tree species to tree species and area to area. However, there is evidence
that summer drought and low soil moisture can reduce cone productivity in some conifers (Barbour
and Billings, 1988). Thus, warming in southern Arizona could also affect the cone crop and, thereby,
the squirrel's food supply.
D.10. JUSTIFICATION FOR FRAMEWORK SCORES
The framework scores for the Mount Graham red squirrel are presented in Appendices G
through J.
In Module 1 (baseline vulnerability), the species scored Vb2, indicating that it is currently in a
highly vulnerable condition. This score is based on the following subcomponents:
• Current population size and trends—based on the data that are available (see Section D.2),
there are likely to be less than 600 Mount Graham red squirrels in existence. This small
population size confers a best estimate score of 2 in the population size variable. Since some
habitat loss has occurred in the past, it is assumed that the population has been reduced, but
only by a relatively small amount (a best estimate score of 3).
• Range trends—the range of the Mount Graham red squirrel is confined to a few tens of km2 of
conifer forests in the Pinalenos Mountains. This range may have contracted somewhat (though
probably by less than 20%) in the past few decades due to development, although it is likely
stable now. Thus the species scores 4 and 3 in the past and current range trend variables.
• Likely future stressor trends—the main "stressor" that limits Mount Graham red squirrels is
habitat availability, which is currently stable. Thus, the species scores 2 in this variable.
• Individual replacement time—red squirrels begin breeding when about 1-year old (Flyger and
Gates, 1982). For this analysis, an individual replacement time of 2-5 years has been assumed.
• Future vulnerability to stochastic events—Mount Graham red squirrels are likely to be highly
vulnerable to future stochastic events. In particular, catastrophic forest fires could potentially
eradicate a large part of the population. With the increase in the human population in Arizona
and the enhanced access to the squirrel's habitat, the likelihood of such fires is increased. For
this reason, the species scores 1.
• Future vulnerability to policy/management change—all Mount Graham red squirrels exist
within a refuge area owned by the Federal government. They are, therefore, entirely
conservation dependent and have been assigned a score of 1 in this variable.
• Likely future vulnerability to natural stressors—no information has been found on the
susceptibility of this species to non-climate natural stressors. It has been assumed that it is not
vulnerable and allocated a score of 3.
70
-------
Certainty evaluations were allocated to each of the scores in Module 1. These are largely
subjective evaluations of the robustness of each of the scores and reflect the availability and quality of
information for each category, rather than rigorous evaluations. Most of the certainty scores allocated
for this species were medium or high.
In Module 2, the Mount Graham red squirrel scored Vc2, indicating that it is likely to be
highly vulnerable to climate change and that its extinction risk may be increased substantially. This
score is largely based on the extreme habitat specialization of the species, the potential scale of habitat
loss due to warming temperatures, the increased likelihood of catastrophic fires, and the isolation of
the species and its likely inability to move to and colonize new areas.
Certainty evaluations were allocated to each of the scores in Module 2. Again, these are largely
subjective evaluations of the robustness of each of the scores and reflect the availability and quality of
information for each category, rather than rigorous evaluations. Given that little is known about the
physiological tolerances of the species, its sensitivity to extreme weather events, or its dependencies
on other species or conservation actions, many of these scores are only medium.
In Module 3, the scores from Modules 1 (Vb2) and 2 (Vc2) are combined in an integrative
matrix to give an overall vulnerability score of Vol (likely to be critically vulnerable to climate
change).
In Module 4, the individual variable certainty scores from Modules 1 and 2 are combined in an
integrative matrix into overall evaluation of certainty of High. This implies that the vulnerability
evaluation for the Mount Graham red squirrel performed in Modules 1 through 3 is likely to be robust.
D.ll. POTENTIAL EFFECTS OF CLIMATE CHANGE ON STATUS AND DISTRIBUTION
The extremely limited range of Mount Graham red squirrels, together with their high degree of
habitat specialization and the potential effects of global warming on their habitats, makes the species
vulnerable to a high risk of extinction. Even if their habitat was not eliminated completely, the
subspecies could still suffer extinction due to its fragmentation and the further fragmentation and
relative isolation of subpopulations. For these reasons, the Mount Graham red squirrel should be
considered one of the most highly vulnerable T&E organisms. It is conceivable that a mitigation
strategy could involve animals being introduced into less threatened habitat further north. However,
there they would likely interbreed with the resident red squirrels (unless they were first eradicated),
and the unique genetic identity of Mount Graham red squirrels would be lost.
D.12. UNCERTAINTIES
There are two main areas of uncertainty inherent in predicting the likely effects of climate
change on Mount Graham red squirrels:
71
-------
• Uncertainty associated with assessing the species' physiological sensitivity to increased
temperatures. There is some evidence from the species habitat use patterns that they may be
directly sensitive to insolation and temperature, and we have assumed in Module 2 a moderate
degree of sensitivity. However, they could be more sensitive and, therefore, more vulnerable
than assessed.
• Uncertainty associated with our understanding of how climate change may affect factors that
influence the extent and quality of the red squirrel's habitat. Warming temperatures and
increased drought frequencies could conceivably increase the risk of catastrophic forest fires or
pest attacks. This variable of Module 2 has been scored as moderately sensitive. However, the
ecosystem may be more sensitive than assessed. Also, we have little information or theoretical
basis for projecting how warming temperatures might affect the pinecone crop, the squirrel's
main food supply.
D.13. SUMMARY
The entire population of Mount Graham red squirrels consists of a few hundred individuals
confined to a small area (probably less than 50 km2) of the Pinalenos Mountains in southern Arizona.
Within their range, they are habitat specialists, being confined to higher elevation mixed conifer or
spruce-fir forests. Because of their small population, their habitat requirements and the potential
eradication by climate change of their main habitats, this subspecies should be considered critically
endangered to the effects of current stressors and future climate change and at high risk of extinction.
D.14. REFERENCES
Bachelet, D; Neilson, RP. (1999) Biome redistribution under climate change. Gen. Tech. Rep. PNW-
GAR. USD A PNW Res. Stat. Oregon.
Barbour, MG; Billings, WD. (1988) North American terrestrial vegetation. Cambridge, U.K
Cambridge University Press.
Flyger, V; Gates, JE. (1982) Pine squirrels. In: Chapman, JA; Feldhamer, GA; eds. Wild mammals of
North America., Baltimore: Johns Hopkins Press.
Gelbach, FR. (1981). Mountain islands and desert seas: a natural history of the U.S.-Mexican
borderlands. College Station: Texas A&M University Press.
Landis, J. University of California, (2006) Personal communication with Hector Galbraith, Manomet
Center for Conservation Sciences.
U.S. FWS (U.S. Fish and Wildlife Service). (1993) Mount Graham red squirrel recovery plan. U.S.
FWS, Phoenix, Arizona.
72
-------
U.S. Forest Service. (1988) Mount Graham red squirrel: an expanded biological assessment. Coronado
National Forest, Tucson, Arizona.
73
-------
APPENDIX E
EXAMPLE NARRATIVE FOR DESERT TORTOISE
E.l. INTRODUCTION
Based on a review of available information, this narrative describes the current vulnerability of
the desert tortoise, Gopherus agassizii, to severe population reduction or extinction, and its potential
future vulnerability under climate change. It is intended to provide support for the scores listed in
Modules 1 through 4 with the specific objectives of
• making transparent the rationale underlying each score in Modules 1 and 2;
• identifying main sources of uncertainty;
• identifying and describing the roles of the main stressors in the estimate of vulnerability for
this species; and
• qualitatively describing potential population responses to climate change and other stressors.
E.2. ENDANGERED SPECIES ACT STATUS
The Utah population of the desert tortoise was listed as Threatened under the Endangered
Species Act (ESA) in 1980. The Mohave population (the population to the north and west of the
Colorado River) was listed as Endangered in 1989, but upgraded to Threatened in April 1990.
The species is also protected by legislation at the State level: it is listed as Threatened under
the California ESA and is protected under the Revised Statutes of both Arizona and Nevada.
E.3. DISTRIBUTION, STATUS, AND POPULATION TRENDS
The desert tortoise is resident in the deserts of the American southwest. Within the United
States, its current range includes the Mohave and Sonoran deserts of southwest Utah, southern
Nevada, southern California, and south and west Arizona. This is an area of approximately 60,000
miles2 (extrapolated from a range map developed by the U.S. Geological Survey
http://geochange.er.usgs.gov/sw/impacts/biology/tortoisel/tortmap.html), though not all of this area
may be suitable habitat, or occupied by desert tortoises. South of the U.S.-Mexico border, the desert
tortoise's range extends through Sonora to northern Sinaloa.
No previous attempts to estimate the total U.S. population of desert tortoises have been found.
However, densities in 14 proposed Desert Wildlife Management Areas (DWMAs) range between
approximately 5-10 and 100+ animals per square mile (U.S. FWS, 1994). With a total area of
approximately 12,500 miles2 (U.S. FWS, 1994), and conservatively assuming an average density of
74
-------
about 7-8 animals per mile2, the DWMAs, alone, currently support at least 100,000 desert tortoises.
Consequently, the current total population throughout the entire U.S. range is likely to be at least in
the low hundreds of thousands of individuals.
Beginning in the 1970s, reductions were noticed in a number of desert tortoise populations.
The reduction rates varied between 3 and 59% per annum, with the highest rates among those
populations most exposed to human disturbance (U.S. FWS, 1994). In some areas that previously
supported healthy tortoise populations, these reductions have led to apparent local extinctions (e.g., in
Antelope Valley in Kern and Los Angeles counties, California [Berry and Nicholson, 1984; Tierra
Madre Consultants, 1991]). These rapid negative population trends were the main reason for the listing
of the Utah and Mohave populations of the species.
E.4. HABITAT
In the wild, desert tortoises occur exclusively in deserts, particularly in scrub and cactus
deserts, where the shrubs provide shelter from the summer sun. They prefer habitat where the shrubs
are widely spaced, since this discontinuous canopy facilitates the growth of the desert annuals on
which the tortoises feed. Preferred shrubs include creosote bush, bursage, blackbush, saguaro cactus,
Joshua tree, palo verde, and yucca. Thus, they are fairly flexible in their habitat requirements within
the scrub desert association (Ernst et al., 1994).
The diet of desert tortoises comprises mainly ephemeral forbs and their flowers, and the
population densities of the tortoises, in some areas, may be determined by the biomass production of
these species after spring rains (U.S. FWS, 1994). Therefore, ideal habitat consists of scrub deserts
with relatively dense seasonal developments of ephemeral plants. Grasses are eaten but may be
secondary food items (Ernst et al., 1994).
The tortoises spend much of the drier part of the year between November-March (when plant
growth is limited) below the surface of the ground in burrows that they dig themselves. They,
therefore, prefer areas of loose and penetrable soils.
E.5. PHYSIOLOGICAL/CLIMATIC LIMITATIONS ON DISTRIBUTION AND STATUS
Not surprisingly for an ectothermic organism, ambient temperature plays an important role in
the autecology and behavior of desert tortoises. Much of their behavior is apparently aimed at
avoiding excessively low and high temperatures. Desert tortoises hibernate in burrows during the
cooler months (generally October-April) and emerge only in the warmer summer months (Ernst et al.,
1994). When active, their activity patterns are highly influenced by diel temperature cycles, remaining
in their burrows until the ambient air temperature exceeds about 20°C (Woodbury and Hardy, 1948).
There may also be a critical maximum temperature that desert tortoises can tolerate of about 43°C
75
-------
(Hutchison et al., 1966). Spending most of their lives in burrows is probably an adaptation to avoiding
low surface temperatures, and to conserve water and avoid desiccation (Ernst et al., 1994). In
abnormally hot weather, the desert tortoise may remain in their burrows most of the day, emerging
only in the cooler mornings and evenings (Ernst et al., 1994).
In a number of turtle species, sexual differentiation has been shown to be affected by egg
incubation temperature (Gans, 1985). In these species, incubation at temperatures above 31°C usually
results in an abnormally high proportion of phenotypic females, while incubation at temperatures
below about 27°C yields mostly phenotypic males. It is not known whether a temperature-based
sexual differentiation relationship exists in desert tortoises. However, if it did, it could potentially
contribute to the current range limits.
Desert tortoises are k-selected species. That is, their adult survival and longevity is high (they
may live for 50 or more years), but their reproductive rates are low. Individuals do not reach breeding
age until they are about 15 years old, and embryonic and juvenile mortality is high, with more than
90% of juveniles dying before reaching adulthood (U.S. FWS, 1994). This, coupled with relatively
low dispersal ability, means that desert tortoises have only a limited ability to compensate for
population reductions caused by anthropogenic or natural factors. U.S. FWS (1994) estimates that the
normal population growth rate (in favorable environmental circumstances) could not exceed 0.5% per
year. Thus, if a population were halved in size by a stressor, it would require at least 140
comparatively stress-free years to return to its previous size. This highlights the fact that any increased
mortality among adult tortoises could fundamentally affect population viability.
E.6. ECOLOGICAL LIMITATIONS ON DISTRIBUTION AND STATUS
The current range of desert tortoises in the United States is confined to the Mohave and
Sonoran warm scrub deserts (Barbour and Billings, 19889). This is also the likely extent of the
species' range prior to European colonization of North America (U.S. FWS, 1994). There is no fossil
evidence that the species ever successfully colonized the colder Great Basin deserts further to the
north. The species' current and historical distribution suggests that the northern extent of its range may
be limited, either directly or indirectly, by temperature. The ranges of many reptiles reach their
northernmost extents at the northern extremes of the warm deserts (e.g., western banded gecko, desert
night lizard, desert iguana, long-tailed brush lizard, and Gila monster [Stebbins, 1985]), suggesting a
general climatic limitation on many members of the taxon.
The absence of the species from the warm Chihuahuan deserts of eastern Arizona, New
Mexico, and Texas is less easily explained. However, climate may also play a role: since the
Chihuahuan desert generally lies at higher elevations than the Mohave or Sonoran deserts, it is
generally cooler, with an annual mean temperature of 18.6°C, compared with 20°C or higher in the
76
-------
Sonoran and Mohave deserts (Bailey, 1979; Schmidt, 1986; Barbour and Billings, 1989). Furthermore,
precipitation amounts and patterns differ between the Mohave/Sonoran and the Chihuahuan deserts:
at most sites in the former, the majority of the precipitation falls in winter, while in the latter,
precipitation is higher, and it falls mainly in summer (Barbour and Billings, 1989). The generally
cooler and moister conditions of the Chihuahuan desert have resulted in marked differences in the
structure and composition of its vegetation communities, when compared to the Mohave and Sonoran.
The flora of the former has a much higher representation of grass species, to the extent that grasses
may dominate, while the latter are mainly shrub and forb dominated. Thus, the desert tortoise might be
excluded from the Chihuahuan desert either directly by climate or indirectly through its effect on
habitat.
It is not likely that competition with other chelonians plays a role in excluding the desert
tortoise from the Chihuahuan desert since no other species are found there, at least in the western part
abutting the desert tortoise range.
E.7. EXISTING STRESSORS AND THEIR EFFECTS ON DISTRIBUTION AND STATUS
Desert tortoise populations in the southwestern U.S. are currently under a high degree of
anthropogenic stress. Humans are directly responsible for much of the premature mortality, either
through shooting, crushing by off-road vehicles, mortality on the roads, facilitating the spread of
disease (e.g., upper respiratory tract disease in pet tortoises being released into wild populations), or
by collecting the animals either to eat or to keep as pets. Humans are also indirectly responsible for
declining desert tortoise populations through urbanization leading to habitat loss, increased
recreational disturbance and modification of habitat, the introduction of alien predators (e.g., dogs)
into the tortoise habitat or increasing the density of native predators (e.g., ravens), crushing by
livestock, livestock-induced habitat modification, fires, or the introduction of alien plant species into
the deserts.
Since there is a common denominator to all of these effects—humans—the effects are often
spatially correlated in their occurrence. For example, the development of desert subdivisions results in
direct effects, in that the increased human population and traffic leads to increased mortality in
tortoises. Furthermore, the subdivision itself results in habitat loss, while increased use of the desert by
the new human residents and their pets leads to elevated mortality rates and further, more widespread,
habitat modification. Thus, the greatest rates of population decline in desert tortoises have occurred in
areas that have been developed or that are undergoing development (U.S. FWS, 1994).
Projections of future urbanization in California hold out little relief for desert tortoises. Some
of the areas that are projected to encompass the greatest growth in urbanization are in the arid scrub
desert in the southwest, particularly in San Bernardino County in the Twentynine Palms area. This
77
-------
area is adjacent to the proposed Joshua Tree DWMA (Landis 2006). Also, rapid urbanization and
sprawl in the neighborhoods of Tucson and Phoenix, Arizona are likely to affect the quantity and
quality of desert tortoise habitat. The footprint of the City of Phoenix doubled in size between 1970
and 1990, and it continues to grow: the city's population increased by 22% in the years between 1990
and 1995 (http://www.sierraclub.org/sprawl/report98/phoenix.html).
Natural factors also may affect tortoise population processes. During drought conditions, their
breeding success may be low, due, possibly, to the limited growth of their food plants and poorer
maternal condition (U.S. FWS, 1994). Fires may also cause increased mortality among desert tortoises
(U.S. FWS, 1994).
E.8. POTENTIAL DIRECT (PHYSIOLOGICAL) VULNERABILITY TO CLIMATE
CHANGE
It is unlikely that increases in temperature due to global climate change will have a direct
physiological effect on desert tortoises, resulting in further population reduction or local extinctions.
The species is already adapted to some of the hottest, most arid environments on Earth. Furthermore,
its Mexican distribution extends 600 miles south of the U.S.-Mexican border into areas where annual
ambient air temperature is 3-5°C higher than in its U.S. range (http://www.cdc.noaa.gov/cg).
Alternatively, it is feasible that temperature increases in the northern part of the desert tortoise range
may assist it to extend its distribution into the Great Basin deserts, an area from which it is currently
excluded, probably by low ambient temperatures (annual ambient air temperature approximately
3-5°C lower than in Sonoran and Mohave deserts: http://www. cdc.noaa.gov/cg).
While it may be possible that desert tortoise reproductive biology may be adversely affected by
increased temperature through disruption of gender differentiation, it is not certain whether the species
is sensitive to this factor, and the thermal thresholds at which it this might become important are not
known.
E.9. POTENTIAL INDIRECT (ECOLOGICAL) VULNERABILITY TO CLIMATE
CHANGE
Galbraith et al. {inprep) have modeled the likely spatial responses of southern Californian
vegetation communities to future climate change scenarios. The future climate scenarios that they
78
-------
evaluated are shown below in Table E-l.
Table E-l. Climate scenarios used to model change in spatial distribution of
southern California vegetation communities.
Scenario
Temp. Change
Precipitation change
Hadley
+3°C
+ 100%
T3P0
+3°C
0%
T5P0
+5°C
0%
T3P18
+3°C
+18%
These scenarios shown in Table E-l bracket the likely range of changes that the current Global
Circulation Models (GCMs) project for the area.
The potential future (by 2100) spatial distributions of the vegetation communities were
projected using the Mapped Atmosphere-Plant-Soil system MAPPS model (VEMAP members, 1995)
and the above climate scenarios. The projected future distributions of the vegetation communities were
delineated after current agricultural land and land that is urban now or predicted to be urban by the
year 2020 was masked out (Landis 2006). The projected changes for southern Californian subtropical
arid shrubland (the scrub desert required by desert tortoises) are shown in Table E-2. Approximately
1.4% of the loss of desert acres will be due to development (from data supplied by John Landis), the
remainder to climate change.
Table E-2. Acre and percent changes in the spatial extent of subtropical arid
shrubland projected by the MAPPS model (to 2100) and using the climate change
scenarios from Table E-l. Agricultural and developed (by 2020) land masked out.
Scenario
Current acres
2100 acres (% change)
Hadley
17,346,799
7,610,846 (-56%)
T3P0
17,346,799
14,579,218 (-16%)
T5P0
17,346,799
13,837,902 (-20%)
T3P18
17,346,799
12,379,980 (-29%)
In general, MAPPS projects that the subtropical arid shrubland will be invaded and replaced by
other vegetation communities, particularly grasslands dominated by C4 species. The extent of this
replacement will vary with the temperature and precipitation assumptions in the scenario but will
typically range between about 20 and 50% habitat loss.
79
-------
The data in Table E-2 could be viewed as "best-case estimates" of the likely extents of habitat
loss since the urbanization projections only go as far as 2020. Nevertheless, all the climate change
scenarios project substantial changes in the extent of desert tortoise habitat in southern California. It is
likely that these losses will be matched with corresponding losses in Arizona, Utah, and Nevada.
Also, the model projections reported above focus on relatively gross measures of vegetation
change (community replacement). Climate change could result in habitat change that might be more
subtle but just as fundamentally important for desert tortoises. For example, facilitated invasion by
nonnative plant species could result in alterations to the amount and quality of food provided by
habitat. Invasive grass species are already a problem for tortoises in some parts of their range (U.S.
FWS, 1994). Also, changes in vegetation or woody biomass could lead to an increased frequency of
fires, which desert tortoises are not well adapted to withstand.
E.10. JUSTIFICATION FOR FRAMEWORK SCORES
The framework scores for the desert tortoise are presented in Appendices G through J. In
Module 1 (baseline vulnerability), the species scored Vb3, indicating that it is currently in a less
vulnerable condition. This score is based on the following subcomponents:
• Current population size and trends—based on the distribution and density data that are
available (see Section E.2), there are likely to be more than 100,000 individual desert tortoises
in existence in the U.S. However, there is good evidence that the U.S. population has and is
declining in many parts of its range (U.S. FWS, 1994). Because of this, the species scores 6 in
the population size variable; but only 3 in the past population trend variable, and 2 in the
current population trend variable.
• Range trends—the U.S. range of desert tortoises extends over a relatively large area (>100,000
km2). However, its range is contracting as local populations are reduced or become extinct. For
these reasons, the species scores 3 and 2, respectively, in these variables.
• Likely future stressor trends—the main stressors that have reduced desert tortoise populations
in the U.S. have been anthropogenic habitat loss or modification, and human-induced
mortality. Much of this is connected to increasing urbanization in the tortoises' habitat, with
concomitant increases in recreational use. Growth of desert communities and the establishment
of new communities are continuing and expected to continue for the foreseeable future. Thus,
the resulting stressors are expected to increase in the future and to reach areas that may be
currently less affected. For this reason, desert tortoise scores 1 in this variable.
• Individual replacement time—desert tortoises do not begin breeding until they have reached
about 15 years of age. For this reason, the species scores 1.
• Likely future vulnerability to stochastic events—since the desert tortoise is widespread in its
distribution and individuals live for many decades, the species has the potential to withstand
localized and short-term stochastic events such as sporadic droughts. Therefore, it scores 3.
80
-------
• Future vulnerability to policy/management change—most desert tortoise populations in the
U.S. are not on protected land, and few populations are being actively managed or conserved.
However, State and Federal protection is important in the viability of local populations.
Because of this, it has been assigned a policy/conservation dependency score of 2.
• Likely future vulnerability to natural stressors—no information has been found on the
susceptibility of this species to non-climate change natural stressors. It has been assumed that it
is not vulnerable and allocated a score of 3.
Certainty evaluations were allocated to each of the scores in Module 1. These are largely
subjective evaluations of the robustness of each of the scores and reflect the availability and quality of
information for each category, rather than rigorous evaluations. Given that the species is relatively
well studied, medium-to-high certainty scores were assigned to most of the variables.
In Module 2, the desert tortoise scored Vc2, indicating that it is likely to be highly vulnerable
to climate change and that its extinction risk may be increased substantially. This score is based on the
following subcomponents:
• Physiological sensitivity to temperature and precipitation change and to extreme weather
events—desert tortoises are believed to be sensitive to drought, to the extent that their
reproductive success is reduced during drought years. Also, the availability of their preferred
food plants is dependent on seasonal precipitation. Thus, future climate scenarios in which
temperature increases and the rainfall pattern changes, resulting in less soil moisture at critical
times of the year, could have adverse effects on tortoises. While we are not able to predict
changes in precipitation in the tortoise's range with any certainty, it is highly likely that
temperature will increase by several degrees Centigrade. For this reason, it scores 2 in each of
these variables.
• Dispersive capability and potential rate of increase—the dispersive capability of this species is
low. Also, given that its potential reproductive replacement rate is low (see Section E.5), it
scores only 1 in each variable.
• Habitat specialization—desert tortoises are only moderately specialized in their habitat
requirements (see Section E.3). However, they are restricted to one major plant association—
warm deserts. Because of this, the species scores 2 in the habitat specialization and diversity
variables.
• Likely extent of future habitat loss due to climate change—vegetation modeling suggests that
future climate change in southern California could result in the replacement of between 20 and
50% of the tortoises scrub desert habitat by C4 grasslands (this does not include additional
habitat that will be lost due to urbanization). Thus, desert tortoise scores 2 in this variable.
• Ability of habitats to shift in response to climate change—It is possible that increasing
temperatures could result in the northward extension of the Mohave and Sonoran deserts into
what is currently the southern range of the Great Basin deserts. No apparent geological or
81
-------
anthropogenic barriers oppose this migration. If this occurred, new habitat could eventually
become available for colonization by the tortoise. Desert tortoise scores 3 in this variable.
• Dependence on temporal inter-relations and other species—desert tortoises are dependent to a
great extent on winter and spring rains triggering the growth of the desert annual plant species
on which they feed. Without this spring growth, it is unlikely that the tortoises could reproduce
(in drought years when the annual plants are less abundant, tortoise breeding success is low).
Accordingly a score of 1 has been assigned to this variable of Module 2. Desert tortoises are
assigned a score of 2 for Variable 10 (dependence on other species) because of their reliance
on the spring growth of annual plants.
Certainty evaluations were allocated to each of the scores in Module 2. Again, these are largely
subjective evaluations of the robustness of each of the scores and reflect the availability and quality of
information for each category, rather than rigorous evaluations. Given that desert tortoise autoecology
is relatively well known, the ecological components of Module 2 were assigned high or medium
certainty scores. However, the physiological relationships between the species and climate variables
are not as well understood and were therefore assigned low-to-medium scores.
In Module 3, the scores from Modules 1 (Vb3) and 2 (Vc2) are combined in an integrative
matrix to produce an overall vulnerability score of Vo2 (likely to be highly vulnerable to climate
change).
In Module 4, the individual variable certainty scores from Modules 1 and 2 are combined in an
integrative matrix into overall evaluation of certainty of Medium. This implies that the vulnerability
evaluation for desert tortoises performed in Modules 1 through 3 is reasonably robust.
82
-------
E.11. POTENTIAL EFFECTS OF CLIMATE CHANGE ON STATUS AND DISTRIBUTION
Desert tortoises, though they are abundant in comparison to most other T&E species, are
currently under considerable anthropogenic stress. This has already resulted in population reductions
and local extinctions. It is likely that, since these stressors (urbanization and associated human
activities) are unlikely to be ameliorated in the foreseeable future, additional population reductions
and extinctions will follow. Thus, even without the added complication of climate change, there is a
strong possibility that the U.S. population of desert tortoises could be reduced markedly in the
relatively near future, with associated range contractions. However, the vegetation modeling that has
been performed thus far indicates that most climate scenarios project losses of the warm desert scrub
habitat on which the species depends. Thus, the habitat losses due to future climate change may
greatly accelerate the overall rate of habitat loss for the species. Also, the vegetation models fail to
capture more subtle potential effects such as the invasion of native desert communities by alien plant
species. These also could have adverse impacts on tortoises as preferred forb food plants are replaced
by non-native species. Conversely, although uncertain, temperature increase could open up new areas
of habitat for the species to the north of its current range. The extent to which this may occur is
conjectural (see below).
The likely net result of climate change and other stressors is that a considerable part of the
current desert tortoise range in the U.S. may be converted to unsuitable habitat. This loss could
approach or exceed 50%. The consequences for tortoises are likely to be further major population
reductions and local extinctions among already fragmented populations. Thus, climate change could
act to exacerbate the effects of other stressors.
E.12. UNCERTAINTIES
There are three main areas of uncertainty inherent in predicting the likely effects of climate
change on desert tortoises:
• Uncertainty associated with assessing the species' physiological sensitivity to increased
temperatures. While desert tortoises are adapted to extremely hot and arid climates, they do
apparently have maximum temperature thresholds (during extremely hot periods, they may
withdraw to their burrows). Also, like some other chelonians, their reproductive biology may
be sensitive to high temperatures (acting through sexual differentiation). If these are the case,
the species may be more sensitive to the direct effects of climate change than we have
assumed.
• Uncertainty associated with predicting future climate-induced changes in extent and quality of
desert tortoise habitat. It is possible that increased temperatures could result in the northward
extension of the tortoise range. However, it is not certain that all of the habitat factors that are
important to tortoises will survive this migration. For example, if the migration is accompanied
83
-------
by an increasing dominance of invasive plant species and a decrease in the native forbs
preferred by tortoises, they may not be able to exploit the new habitat. Also, it is possible that
there may be new constants in the new landscape that may be detrimental to tortoises. For
example, if the new and existing habitat has an increased fire frequency (due to higher
temperatures), tortoise populations may be put at risk.
• Uncertainty associated with the ability of desert tortoise populations to withstand an altered
frequency of environmental perturbation. Although long-lived, desert tortoises are slow to
make good losses due to environmental stochasticity. Thus, they are adapted to extreme, but
relatively predictable habitats. If the predictability of their habitat is altered (a higher frequency
of unpredictable extreme events such as fires, droughts, or storms), tortoise population viability
may be put at risk. While GCMs can predict regional changes in the main climate variables,
temperature and precipitation, they are generally less useful in predicting climate-associated
change in stochastic events such as droughts.
E.13. SUMMARY
The Mohave population of the desert tortoise is currently listed as Threatened under the ESA.
The Sonoran population is listed and protected by Arizona State regulations. The current U.S.
population is probably a few hundreds of thousands of individuals, distributed across about 60,000
miles2 of scrub desert in California, Utah, Nevada, and Arizona. Over the last three or four decades,
these populations have come under high degrees of stress due largely to human activity (particularly
urbanization and recreational intrusion). This has resulted in population reductions and local
extinctions. Climate change may be a significant new stressor, causing even more habitat loss and
exacerbating an already difficult situation. Together, existing stressors and the direct and indirect
effects of climate change could result in desert tortoises being put at even greater risk of population
reduction and extinction in their U.S. range.
E.14. REFERENCES
Bailey, HP. (1979) Semi-arid climates: their definition and distribution. In: Hall, AE; Cannell, GH;
Lawton, HW (eds) Agriculture in semi-arid environments. New Your, Springer-Verlag, pp 73-96.
Barbour, MG; Billings, WD. (1988). North American terrestrial vegetation. Cambridge University
Press, New York, 434 pp.
Berry, K.H. and L.L. Nicholson. 1984a. A summary of human activities and their impacts on desert
tortoise populations and habitat in California. In Berry, K.H., ed. The status of the desert tortoise
(Gopherus agassizii) in the United States. U.S. Department of the Interior, Bureau of Land
Management. Riverside, California.
Ernst, CH; Barbour, RW; Lovich, JE. (1994) Turtles of the United States and Canada. Washington,
DC: Smithsonian Institution Press.
84
-------
Gans, C. (1985) Biology of the reptilia. New York: John Wiley and Sons.
Hutchison, VH; Vinegar, A; Kosh, RJ. (1966) Critical thermal maxima in turtles. Herpetologica
22:32-41.
Landis, J. University of California, (2006) Personal communication with Hector Galbraith, Manomet
Center for Conservation Sciences.
Schmidt, RH. (1986) Chihuahuan climate. In: Barlow. JC; Powell, AM; Timmermann, BN eds.
Chihuahuan Desert - U.S. and Mexico, vol II: Proceedings of the 2nd symposium on resources of the
Chihuahuan Desert region; 1983 October 20-21; Alpine, TX: Sul Ross State University, Chihuahuan
Desert Research Institute: 20-39.
Stebbins, RC. (1985) Western reptiles and amphibians. Boston: Houghton Mifflin.
Tierra Madre Consultants. 1991. Biological assessment for Lancaster City and Planning Area:
Relative density surveys for desert tortoises and cumulative human impact evaluations for Mohave
ground squirrel habitat. Rpt. for City of Lancaster. Tierra Madre Consultants, Riverside, CA.
U. S. Fish and Wildlife Service. 1994. Desert tortoise (Mojave population) Recovery Plan. U.S. Fish and
Wildlife Service, Portland, OR. 74 pgs plus appendices.
Woodbury, AM; Hardy, R. (1948) Studies of the desert tortoise Gopherus agassizii. Ecol Monogr
18:145-200.
VEMAP Members.(1995) Vegetation/ecosystem modeling and analysis project: Comparing
biogeography and biogeochemistry models in a continental-scale study of terrestrial ecosystem
responses to climate change and CO2 doubling. Glob Biogeochem Cycles 9:407-437.
85
-------
APPENDIX F
EXAMPLE NARRATIVE FOR LAHONTAN CUTTHROAT TROUT
F.l. INTRODUCTION
Based on a review of available information, this narrative describes the current vulnerability of
the Lahontan cutthroat trout, Oncorhynchus clarki henshawi, to severe population reduction or
extinction, and its potential future vulnerability under climate change. Its main objectives are to
• make transparent the rationale underlying each score in Modules 1 and 2;
• identify main sources of uncertainty;
• identify and describe the roles of the main stressors in the estimate of vulnerability for this
species; and
• qualitatively describe potential population responses to climate change and other stressors.
F.2. ENDANGERED SPECIES ACT STATUS
The Lahontan subspecies of the cutthroat trout was listed as Endangered by the U.S. Fish and
Wildlife Service (U.S. FWS) in 1970 [35 FR 16047 16048] but was reclassified under the Endangered
Species Act (ESA) in 1975 as Threatened [40 FR 29863 29864], This status applies to all populations
throughout its range in California, Nevada, and Oregon.
F.3. DISTRIBUTION, STATUS, AND POPULATION TRENDS
Lahontan cutthroat trout originally evolved in Lake Lahontan, which until about 14,000 years
b.p., covered approximately 8,500 miles2 of present day Nevada, California, and Oregon (Benson and
Thompson, 1987). With the drying up of the lake, beginning about 12,000 years ago, the trout's
distribution was fragmented into a number of disconnected drainage basins. Many of these were later
rendered unsuitable as trout habitat by anthropogenic modifications (water diversions, dams, pollution,
or over-harvesting) or the introduction of non-native fish species (U.S. FWS, 1995), causing further
fragmentation of the fish's distribution. Impacts caused by dams, water diversions, and non-native
species introductions still continue.
Currently, self-sustaining populations of Lahontan cutthroat trout are restricted to about 11%
of their ancestral riverine habitats and to less than 1% of lake habitats (U.S. FWS, 1995). They occur
in about 160 streams and 6 lakes in three major areas: the Truckee/Carson/Walker river basins in
eastern California/western Nevada; the Quinn River/Black Rock Desert/Coyote Lake basins in
northern Nevada/southern Oregon; and the Humboldt River basin in north-central Nevada. Survey
86
-------
data assembled by U.S. FWS (1995) indicate that within this range approximately 56% of existing
trout subpopulations comprise less than 500 individual fish and only about 29% have more than 1,000
individuals. Thus, the Lahontan cutthroat trout currently is distributed in a number of isolated,
relatively small, and, in some cases, declining populations.
F.4. HABITAT
The Lahontan cutthroat trout is a cold-water fish that can inhabit a wide variety of riverine and
lake habitats. They generally live in relatively small streams that provide gravel spawning areas and
deep, shaded pools. Lake habitats range from small alpine lakes to large, saline basin lakes. However,
if a lake is to provide suitable habitat, it must provide access to stream spawning areas.
F.5. PHYSIOLOGICAL/CLIMATIC LIMITATIONS ON DISTRIBUTION AND STATUS
Being a fish that is adapted to existence in relatively cold waters, the distribution of Lahontan
cutthroat trout, like many salmonid species, is ultimately limited by ambient water temperatures.
Above maximum temperature thresholds, the physiological processes of the fish began to deteriorate:
growth in juvenile fish slows, egg viability decreases, behavioral avoidance occurs, and fish may
expire (Scott and Crossman, 1973; McGinnis, 1984; Crisp, 2000). The water temperature threshold
that imposes these physiological limitations varies depending on the endpoint and life stage being
assessed. However, exceedances of about 19°C may cause increased lethality in eggs (Crisp, 2000),
while temperatures in excess of 25°C will cause excessive fry mortality (Scott and Crossman, 1973).
U.S. EPA (1995) report that the upper 95th percentile of the maximum weekly average water
temperatures at all stations where the species has been reported is 22.8°C. This, they postulate, is a
good indicator of the upper thermal limit on the species' distribution. U.S. FWS (1995) report that
Lahontan cutthroat trout can withstand short-term water temperatures exceeding 27°C and daily
maxima of 20°C. Thus, it is likely that water temperatures that exceed 20-23°C on a regular basis may
set physiological limits to the distribution of the species.
F.6. ECOLOGICAL LIMITATIONS ON DISTRIBUTION AND STATUS
The main ecological limitations on the distribution of Lahontan cutthroat trout are likely to be
habitat availability and contiguity, and competition with other fish species. Their distribution is
confined to a relatively arid area of the Great Basin where suitable coldwater streams and lakes are
few and isolated. Approximately 90% of their potential riverine habitat and more than 99% of their
potential lake habitat has been rendered unsuitable by human activities including dewatering,
pollution, dams, increased sediment loads, and destruction of shading riparian vegetation by livestock
(U.S. FWS, 1995). Overfishing and the introduction of non-native competitors (e.g., kokanee salmon,
87
-------
Oncorhynchus nerka, brook trout, Salvelinus fontinalis, and shiners, Richardsonius egregious), have
also been responsible for eliminating cutthroat trout from many areas of their ancestral habitat. During
the 19th and early 20th centuries, important commercial fisheries existed at Lake Tahoe, California
and Pyramid Lake, Nevada. These were wiped out by fish population crashes due to overharvesting
and competition with introduced species.
F.7. EXISTING STRESSORS AND THEIR EFFECTS ON DISTRIBUTION AND STATUS
The main ancestral stressor that resulted in a large contraction of the range of the Lahontan
cutthroat trout was the drying of Lake Lahontan. However, since then, the main stressors have been
anthropogenic. During the mining booms of the 19th Century, much habitat was lost as a result of
releases of mining-related contaminants and sediments into the streams and lakes. The increase in
agriculture during the late 19th and early 20th Centuries resulted in habitat impacts as water was
diverted for irrigation and streams and lakes dewatered. In the 20th Century, and continuing into this
century, the main stressors have been competition with nonnative fish species and habitat destruction
due to overgrazing by domestic livestock (U.S. FWS, 1995).
F.8. POTENTIAL DIRECT (PHYSIOLOGICAL) VULNERABILITY TO CLIMATE
CHANGE
Short-term maximum temperatures in water bodies inhabited by Lahontan cutthroat trout are
close to the subspecies' likely physiological maximum. U.S. EPA (1995) projects that maximum
weekly average water temperatures in Nevada range between about 18 and 23°C, while in California
and Oregon, the corresponding data are 32 and 19°C and 24 and 15°C, respectively.
General circulation models (GCMs) such as the Hadley, the Canadian Climate Center, and the
Oregon State University models project substantial warming over the next century in the areas
occupied by Lahontan cutthroat trout. Assuming a doubling of atmospheric CO2 by 2100, these GCMs
project mean annual temperature increases of between 2 and 5.5°C or between 3.6 and 10°F (U.S.
EPA, 1995; National Assessment Synthesis Team, 2000). Assuming that these air temperature
increases are directly translated into water temperature increases (U.S. EPA, 1995), ambient water
temperatures in the habitat of the cutthroat trout could rise by up to 5.5°C (10°F). Such rises in
temperature could have fundamental effects on habitat suitability and distribution. In a modeling
exercise, U.S. EPA (1995) predicted that similar temperature increases could result in cutthroat trout
habitat loss in California and Oregon of more than 50%. Thus, global warming over the next century
could result in a drastic reduction in the habitat available to Lahontan cutthroat trout.
88
-------
F.9. POTENTIAL INDIRECT (ECOLOGICAL) VULNERABILITY TO CLIMATE
CHANGE
In aquatic organisms, the distinction between direct physiological effects and indirect
ecological effects is not as clear as it is in terrestrial organisms. Water is both the habitat for the
organism and the main interface with the changing climatic parameters. Therefore, increased water
temperature results in both direct and ecological effects. The main ecological vulnerability of
Lahontan cutthroat trout to global temperature increases is likely to be through rising water
temperatures rendering hitherto suitable habitat less so.
GCMs predict increased temperatures within the trout habitat over the next century (see
Section F.8). However, there are also likely to be changes in the amount and timing of precipitation
(National Assessment Synthesis Team, 2000). Both the Canadian Climate Center and the Hadley
models predict increased precipitation in the trout's range (80-100% in the former and 20-40% in the
latter). Not only might the amount of precipitation change, but its seasonal pattern is also likely to be
affected: given the higher temperatures, much more may fall as rain rather than snow. This could
mean that high spring high flows due to snow melt may be less marked. This could have important
implications for the seasonal availability of trout spawning habitat: if there is less of a spring peak in
flow, trout may be less able to move upriver to spawn.
F.10. JUSTIFICATION FOR FRAMEWORK SCORES
The framework scores for the Lahontan cutthroat trout are presented in Appendices G through J.
In Module 1 (baseline vulnerability), the subspecies scored Vb2, indicating that it is currently
in a highly vulnerable condition, when compared to other T&E species. This score is based on the
following subcomponents:
• Current population size and trends—U.S. FWS (1995) collated population data from 92 of the
approximately 160 extant river populations. These data indicate that between 68,000 and
140,000 individual fish exist at these sites, alone. The numbers of fish at the approximately 70
sites not included in this analysis are unknown, neither are the numbers in lakes inhabited by
the fish. However, based on these data, it is very likely that an excess of 100,000 Lahontan
cutthroat trout exist within their range. Because of this population status, the subspecies scores
6 in this variable.
Given the alleviation of many of the historical major stressors (pollutants, etc.) and the
protected status extended to the species, it is unlikely that population trends are as steeply
negative as they once were. However, it is still possible, and perhaps likely, that the subspecies
is in slow decline. For this reason, it has been assigned a score of 2.
• Range trends—extrapolating from a range map in U.S. FWS (1995), the current range of
Lahontan cutthroat trout has three main centers extending over an approximate total area of
8,000 miles2. U.S. FWS (1995) indicate that only about 485 miles of stream are currently
89
-------
inhabited, plus six lakes. Thus, the distribution of this species is limited to a comparatively
small area. Also, some subpopulations probably are continuing to decline through interactions
with invasive species and human modification of habitat. Because of these data, the subspecies
has been given scores of 2 and 2, respectively.
• Current stressors and stressor trends—the main stressor that has reduced Lahontan cutthroat
trout populations since post-Colombian colonization has been habitat modification and
destruction. These include releases of pollutants, dewatering for agriculture, riparian vegetation
destruction by livestock, overharvesting, and the introduction of exotic fish species. Some of
these stressors have been reduced in the last few decades (e.g., pollution and overharvesting),
however, others (e.g., livestock vegetation modification and exotic species) continue to exert
deleterious effects on trout populations. The intensity and spatial pattern of these stressors is,
however, probably stable. Because of this, the subspecies scores 2 in this variable.
• Individual replacement time—Lahontan cutthroat trout breed in the first few years of life and
have been assigned a replacement time of 2-5 years.
• Likely future vulnerability to stochastic events—While they inhabit a comparatively stable
freshwater environment, this species could be susceptible to an increased frequency, or degree,
of climate change-induced droughts. For this reason, it has been assigned a score of 2 in this
variable.
• Likely future vulnerability to policy/management change—Significant portions (i.e., >50%) of
streams and lakes inhabited by Lahontan cutthroat trout are on lands owned by the Bureau of
Land Management or the U.S. Forest Service. However, there is only sporadic active
conservation occurring on these lands, and stressors, such as overgrazing, continue to affect
habitat. Thus, the potential for active conservation is relatively unexploited, and the trout's
current status is relatively independent of conservation. For this reason, the subspecies scores 3
in this variable.
• Likely future vulnerability to natural stressors—the species is susceptible to salmonid
pathogens such as whirling disease, though the incidence is low. It has been assigned a score of
2.
Certainty evaluations were allocated to each of the scores in Module 1. These are largely
subjective evaluations of the robustness of each of the scores and reflect the availability and quality of
information for each category, rather than rigorous evaluations. Given that uncertainties exist
regarding the subspecies' current population and habitat trends, low-to-medium certainty scores were
assigned to a number of the variables.
In Module 2, the species scored Vc2, indicating that it is likely to be highly vulnerable to
climate change and that its extinction risk may be increased substantially. It should be noted that the
species almost merits a score of Vcl (in fact, the low alternate score is Vcl). These scores are based
on the following subcomponents:
90
-------
• Physiological sensitivity to temperature and precipitation change and to extreme weather
events—the Lahontan cutthroat trout is likely to be physiologically highly sensitive to
increased water temperatures (see Section F.8). It, therefore, scores 1 in this variable of
Module 2. The subspecies may also be sensitive to altered precipitation patterns (see Section
F.9). However, these effects are likely to be expressed in changes in the trout's ecology and
habitat and dealt with in Variable 4.
• Dispersive capability—the dispersive capability of this subspecies, which lives in isolated
subpopulations, is low (scores 1 in Module 2).
• Habitat specialization—coldwater species cutthroat trout are relatively specialized in their
habitat requirements. However, within this cold-water regime, they are flexible in that they can
inhabit streams, rivers, or lakes with a wide variability of water chemistries. Therefore, it
scores 2 in this variable.
• Likely extents of future habitat loss due to climate change—in this variable, Lahontan
cutthroat trout scored 2 (20-50% likely habitat loss). This score reflects the species' degree of
dependency on coldwater habitats and the potential effects of ambient temperature increases.
However, this score may underestimate the potential degree of habitat loss: considerable
uncertainty exists regarding how altered precipitation patterns may affect habitat quality and
availability. While the assigned score is 2, further information and analysis could elevate it to
1.
• Ability of habitats to shift at same rate as species in response to climate change—because the
current distribution of the habitats of the cutthroat trout are severely circumscribed and limited
by topography and human activities, it is highly unlikely that they will be able to shift much in
response to climate change. Thus, this variable scores 1, only.
• Availability of habitats within new range—given the enclosed nature of the species'
watersheds, it is likely that the only way in which it can naturally (i.e., without human
intervention) colonize new habitats is by moving upstream. However, most such habitats are
probably already colonized, and the potential for this shift is low. Therefore, this variable
scores 2.
• Dependence on temporal inter-relations and other species—the timing of spawning in cutthroat
trout is dependent on the seasonal pulse of water from spring snowmelt. Changes in the amount
and seasonality of precipitation predicted by GCMs could mean that the size of this pulse may
be reduced. This could adversely affect the trout, and this variable has been assigned a score of
2. The cutthroat trout is currently under considerable stress from introductions of non-native
fish species. If increased climate stress were to give these invasives even more of a competitive
advantage, the cutthroat trout could be harmed further. For this reason, a score of 2 has been
assigned to this variable.
Certainty evaluations were allocated to each of the scores in Module 2. Again, these are largely
subjective evaluations of the robustness of each of the scores and reflect the availability and quality of
information for each category, rather than rigorous evaluations. Given that some aspects of Lahontan
cutthroat trout autoecology are relatively well known (e.g., numbers, distribution, habitat preferences),
91
-------
some of the ecological variables of Module 2 were assigned high or medium certainty scores.
However, the potential relationships between climate change and physiological and ecological effects
are uncertain, and in many of these categories, the subspecies scored only medium or low.
In Module 3, the scores from Modules 1 (Vb2) and 2 (Vc2) are combined in an integrative
matrix to give an overall vulnerability score of Vol (likely to be critically vulnerable to climate
change).
In Module 4, the individual variable certainty scores from Modules 1 and 2 are combined in an
integrative matrix into overall evaluation of certainty of Medium. This implies that the vulnerability
evaluation for Lahontan cutthroat trout performed in Modules 1 through 3 is reasonably robust.
F.ll. POTENTIAL EFFECTS OF CLIMATE CHANGE ON STATUS AND DISTRIBUTION
Climate change, particularly temperature increases and changes in precipitation patterns, could
have important implications for the distribution and population status of Lahontan cutthroat trout. Like
other coldwater salmonids, the subspecies is sensitive to water temperatures, and the increases
projected by GCMs could result in a large portion of its current range being made unsuitable. Also, the
trout is probably sensitive to the timing of river flow patterns. Precipitation changes forecast by the
GCMs could potentially disrupt its spawning activities and render currently suitable habitat unsuitable.
Taken together, these potential effects could result in population reductions among local
subpopulations of the fish and render its overall status and stability problematic.
F.12. UNCERTAINTIES
The greatest uncertainties associated with this analysis concern potential future changes in
precipitation amount, seasonality, and the likely responses of the trout. GCMs are relatively imprecise
in their predictions of future precipitation patterns. This, together with uncertainty about the resulting
changes in river flow patterns complicates the projection of effects on the cutthroat trout. However, it
is at least feasible that the effects of changing flow patterns could be severe, perhaps as much as
temperature change. For this reason, the score allocated in Module 2 may underestimate the actual
degree of risk.
F.13. SUMMARY
Lahontan cutthroat trout are currently listed as Threatened under the ESA. Though their
current world population probably exceeds 100,000 individuals, they have declined greatly since
Europeans settled the Great Basin and their population is now highly fragmented into relatively small,
isolated, and vulnerable subpopulations. They are also under considerable stress from invasive species
and human modification of their habitats. This, and their sensitivity to temperature and precipitation
92
-------
regimes, make this species one of the most vulnerable of the T&E species evaluated with regard to
climate change. It is likely that climate change could result in the reduction and extinction of, at least,
some of the subpopulations.
F.14. REFERENCES
Benson, LV; Thompson, RS. (1987) Lake-level variation in the Lahontan Basin for the past 50,000
years. Quaternary Res 28:69-85.
Crisp, DT. (2000) Trout and salmon ecology, conservation, and rehabilitation. Blackwell, Oxford.
McGinnis, SM. (1984) Freshwater fishes of California. Berkeley: University of California Press.
National Assessment Synthesis Team. (2000)Climate change impacts on the United States. U.S.
Global Change Research Program, U.S. EPA, Washington, DC.
Scott, WB; Crossman, EJ. (1973) Freshwater fishes of Canada. Fisheries Res. Board of Canada. Bull
No. 1884, Ottawa.
U.S. FWS (U.S. Fish and Wildlife Service). (1995) Recovery plan for the Lahontan cutthroat trout.
U.S. Fish and Wildlife Service, Region 1, Portland, Oregon.
U.S. EPA. (1995). Ecological impacts from climate change: an economic analysis of freshwater
recreational fishing. EPA 220-R-95-004, Washington, DC: U.S. Environmental Protection Agency.
93
-------
APPENDIX G
EXAMPLE APPLICATIONS OF MODULE 1 - ESTIMATING "BASELINE"
VULNERABILITIES
94
-------
MODULE 1 - CATEGORIZING THE "BASELINE" VULNERABILITIES (Vb) OF T&E SPECIES
Species - Golden-cheeked Warbler (bold scores represent "best estimates" others are alternate scores)
1) Current population size:
Score
7) Likely future non-climate stressor trends:
Score
<100
1
increase
1
1
100-500
2
stable
2
2
500-1,000
3
reduction
3
1,000-10,000
4
10,000-50,000
5
5
Certainty: high (3)
>50,000
6
medium (2)
2
low (1)
Certainty: high (3)
3
medium (2)
low (1)
8) Replacement time for individuals:
Score
> 5 years
1
2) Population trend in last 50 years:
Score
2-5 years
2
>80% reduction
1
<2 years
3
3
>50% reduction
2
2
<1 year
4
>20% reduction
3
3
Apparently stable
4
Certainty: high (3)
3
Increasing
5
medium (2)
low (1)
Certainty: high (3)
medium (2)
2
low (1)
9) Likely future vulnerability to
stochastic events:
Score
Highly vulnerable
1
1
3) Current population trend:
Score
vulnerable
2
2
rapid decline
1
not vulnerable
3
slow decline
2
2
benefiting
4
stable
3
3
increasing
4
Certainty: high (3)
3
medium (2)
Certainty: high (3)
low(1)
medium (2)
2
low (1)
10) Likely future vulnerability to
policy/management changes
Score
4) Range trend in last 50 years
Score
Highly vulnerable
1
1
>80% reduction
1
vulnerable
2
2
>50% reduction
2
2
not vulnerable
3
>20%reduction
3
3
benefiting
4
apparently stable
4
increasing
5
Certainty: high (3)
3
medium (2)
Certainty: high (3)
3
low (1)
medium (2)
low (1)
11) Likely future vulnerability to
natural stressors:
Score
5) Current range trend:
Score
Highly vulnerable
1
rapid reduction
1
vulnerable
2
2
slow reduction
2
2
not vulnerable
3
3
stable
3
increasing
4
Certainty: high (3)
medium (2)
2
Certainty: high (3)
3
low (1)
medium (2)
low (1)
6) Main current stressors (narrative)
TOTAL SCORE
CUMULATIVE CERTAINTY SCORE:
26
Max. score 30
Min. score 10
Species score: Vb2 (Vb3)
95
-------
MODULE 1 - CATEGORIZING THE "BASELINE" VULNERABILITIES (Vb) OF T&E SPECIES
Species - Bald Eagle (bold scores represent "best estimates" others are alternate scores)
1) Current population size:
Score
7) Likely future non-climate stressor trends:
Score
<100
increase
1
100-500
2
stable
2
2
500-1,000
3
reduction
3
3
1,000-10,000
4
10,000-50,000
5
Certainty: high (3)
>50,000
6
6
medium (2)
2
low (1)
Certainty high (3)
3
medium (2)
low (1)
8 Replacement time for individuals:
Score
2) Population trend in last 50 years:
Score
> 5 years
1
>80% reduction
2-5 years
2
2
>50% reduction
2
2
<2 years
3
>20% reduction
3
3
<1 year
4
Apparently stable
4
Increasing
5
Certainty: high (3)
3
medium (2)
Certainty: high (3)
3
low (1)
medium (2)
low (1)
9) Likely future vulnerability to
3) Current population trend:
Score
stochastic events:
Score
rapid decline
Highly vulnerable
1
slow decline
2
vulnerable
2
stable
3
not vulnerable
3
3
increasing
4
4
benefiting
4
Certainty: high (3)
3
Certainty: high (3)
3
medium (2)
medium (2)
low (1)
low (1)
4) Range trend in last 50 years
Score
>80% reduction
10) Likely future vulnerability to
>50% reduction
2
2
policy or management changes
Score
>20%reduction
3
3
Highly vulnerable
1
apparently stable
4
vulnerable
2
2
increasing
5
not vulnerable
3
3
benefiting
4
Certainty: high (3)
medium (2)
2
Certainty: high (3)
low (1)
medium (2)
2
low (1)
5) Current range trend:
Score
rapid reduction
slow reduction
2
11) Likely future vulnerability to
stable
3
3
natural stressors:
Score
increasing
4
4
Highly vulnerable
1
vulnerable
2
2
Certainty: high (3)
3
not vulnerable
3
3
medium (2)
low (1)
Certainty: high (3)
medium (2)
2
low (1)
6) Main current stressors (narrative)
TOTAL SCORE 32
CUMULATIVE CERTAINTY SCORE:
26
Max. score 30
Min. score 10
Species score: Vb3 (Vb3. Vb4)
96
-------
MODULE 1 - CATEGORIZING THE "BASELINE" VULNERABILITIES (Vb) OF T&E SPECIES
Species - Salt marsh harvest mouse (bold scores represent "best estimates" others are alternate scores)
1) Current population size:
Score
7) Likely future non-climate stressor trends:
Score
<100
increase
1
1
100-500
2
stable
2
500-1,000
3
3
reduction
3
1,000-10,000
4
4
10,000-50,000
5
Certainty: high (3)
3
>50,000
6
medium (2)
low (1)
Certainty: high (3)
medium (2)
low (1)
1
8) Replacement time for individuals:
Score
> 5 years
1
2) Population trend in last 50 years:
Score
2-5 years
2
>80% reduction
<2 years
3
3
>50% reduction
2
2
<1 year
4
>20% reduction
3
3
Apparently stable
4
Certainty: high (3)
Increasing
5
medium (2)
2
low (1)
Certainty: high (3)
medium (2)
low (1)
1
9) Likely future vulnerability to
stochastic events:
Score
Highly vulnerable
1
1
3) Current population trend:
Score
vulnerable
2
2
rapid decline
not vulnerable
3
slow decline
2
2
benefiting
4
stable
3
3
increasing
4
Certainty: high (3)
medium (2)
2
Certainty: high (3)
low (1)
medium (2)
low (1)
1
10) Likely future vulnerability to
policy or management changes
Score
4) Range trend in last 50 years
Score
Highly vulnerable
1
>80% reduction
vulnerable
2
2
>50% reduction
2
2
not vulnerable
3
3
>20%reduction
3
3
benefiting
4
apparently stable
4
increasing
5
Certainty: high (3)
medium (2)
2
Certainty: high (3)
low (1)
medium (2)
2
low (1)
11) Likely future vulnerability to
natural stressors:
Score
5) Current range trend:
Score
Highly vulnerable
1
rapid reduction
vulnerable
2
2
slow reduction
2
2
not vulnerable
3
3
stable
3
increasing
4
Certainty: high (3)
medium (2)
2
Certainty: high (3)
low (1)
medium (2)
2
low (1)
6) Main current stressors (narrative)
TOTAL SCORE 22 (
CUMULATIVE CERTAINTY SCORE:
18
Max. score
Min. score
Species score: Vb2 (Vb2 - Vb3)
97
-------
MODULE 1 - CATEGORIZING THE "BASELINE" VULNERABILITIES (Vb) OF T&E SPECIES
Species - Mount Graham red squirrel (bold scores represent "best estimates" others are alternate scores)
1) Current population size:
Score
7) Likely future non-climate stressor trends:
Score
<100
1
increase
1
1
100-500
2
2
stable
2
2
500-1,000
3
reduction
3
3
1,000-10,000
4
10,000-50,000
5
Certainty: high (3)
>50,000
6
medium (2)
2
low (1)
Certainty: high (3)
3
medium (2)
low (1)
8) Replacement time for individuals:
Score
> 5 years
1
2-5 years
2
2
2) Population trend in last 50 years:
Score
<2 years
3
>80% reduction
1
<1 year
4
>50% reduction
2
>20% reduction
3
3
Certainty: high (3)
3
Apparently stable
4
4
medium (2)
Increasing
5
low (1)
Certainty: high (3)
medium (2)
2
9) Likely future vulnerability to
low (1)
stochastic events:
Score
Highly vulnerable
1
1
vulnerable
2
2
3) Current population trend:
Score
not vulnerable
3
rapid decline
1
benefiting
4
slow decline
2
2
stable
3
3
Certainty: high (3)
3
increasing
4
medium (2)
low (1)
Certainty: high (3)
medium (2)
2
low (1)
10) Likely future vulnerability to
policy or management changes
Score
Highly vulnerable
1
1
4) Range trend in last 50 years
Score
vulnerable
2
2
>80% reduction
1
not vulnerable
3
>50% reduction
2
benefiting
4
>20%reduction
3
apparently stable
4
4
Certainty: high (3)
3
increasing
5
medium (2)
low (1)
Certainty: high (3)
medium (2)
2
low (1)
11) Likely future vulnerability to
natural stressors:
Score
Highly vulnerable
1
5) Current range trend:
Score
vulnerable
2
2
rapid reduction
1
not vulnerable
3
3
slow reduction
2
stable
3
3
Certainty: high (3)
increasing
4
medium (2)
low (1)
1
Certainty: high (3)
medium (2)
2
low (1)
6) Main current stressors (narrative)
TOTAL SCORE 24
CUMULATIVE CERTAINTY SCORE:
23
Max. score
Min. score
Species score: Vb2 (Vb2. Vb3)
98
-------
MODULE 1 - CATEGORIZING THE
"BASELINE" VULNERABILITIES (Vb) OF T&E SPECIES
Species - Desert tortoise (bold scores represent "best estimates" others are alternate scores)
1) Current population size:
Score 7) Likely future non-climate stressor trends:
Score
<100
1
increase
1
1
100-500
2
stable
2
2
500-1,000
3
reduction
3
1,000-10,000
4
10,000-50,000
5
5 Certainty: high (3)
>50,000
6
6 medium (2)
2
low (1)
Certainty: high (3)
3
medium (2)
low (1)
8) Replacement time for individuals:
Score
2) Population trend in last 50 years:
Score > 5 years
1
1
>80% reduction
1
2-5 years
2
2
>50% reduction
2
2 <2 years
3
>20% reduction
3
3 <1 year
4
Apparently stable
4
Increasing
5
Certainty: high (3)
3
medium (2)
Certainty: high (3)
3 low (1)
medium (2)
low (1)
9) Likely future vulnerability to
stochastic events:
Score
Highly vulnerable
1
vulnerable
2
2
3) Current population trend:
Score not vulnerable
3
3
rapid decline
1
1 benefiting
4
slow decline
2
2
stable
3
Certainty: high (3)
3
increasing
4
medium (2)
low (1)
Certainty: high (3)
3
medium (2)
low (1)
10) Likely future vulnerability to
policy or management changes
Score
Highly vulnerable
1
1
4) Range trend in last 50 years
Score vulnerable
2
2
>80% reduction
1
not vulnerable
3
>50% reduction
2
2 benefiting
4
>20%reduction
3
3
apparently stable
4
Certainty: high (3)
increasing
5
medium (2)
2
low (1)
Certainty: high (3)
medium (2)
2
low (1)
11) Likely future vulnerability to
natural stressors:
Score
Highly vulnerable
1
5) Current range trend:
Score vulnerable
2
2
rapid reduction
1
1 not vulnerable
3
3
slow reduction
2
2
stable
3
Certainty: high (3)
increasing
4
medium (2)
2
low (1)
Certainty: high (3)
3
medium (2)
low (1)
6) Main current stressors (narrative)
TOTAL SCORE
CUMULATIVE CERTAINTY SCORE:
24
Max. score
Min. score
Max. score 30
Min. score 10
Baseline vulnerability scores:
Vb1
=16
Critically vulnerable
Vb2
17-22
Highly vulnerable
Vb3
23-29
less vulnerable
Vb
-29
Least vulnerable
Species score:
Vb3 (Vb2. Vb3)
99
-------
MODULE 1 - CATEGORIZING THE "BASELINE" VULNERABILITIES (Vb) OF T&E SPECIES
Species - Lahontan cutthroat trout (bold scores represent "best estimates" others are alternate scores)
1) Current population size:
Score 7) Likely future non-climate stressor trends:
Score
<100
1
increase
1
1
100-500
2
stable
2
500-1,000
3
reduction
3
1,000-10,000
4
10,000-50,000
5
5 Certainty: high (3)
>50,000
6
6 medium (2)
2
low (1)
Certainty: high (3)
medium (2)
low (1)
1 8) Replacement time for individuals:
Score
>5 years
1
2-5 years
2
2
2) Population trend in last 50 years
Score <2 years
3
3
>80% reduction
1
1 <1 year
4
>50% reduction
2
2
>20% reduction
3
Certainty: high (3)
Apparently stable
4
medium (2)
2
Increasing
5
low (1)
Certainty: high (3)
3
medium (2)
9) Likely future vulnerability to stochastic events:
Score
low (1)
highly vulnerable
1
vulnerable
2
2
not vulnerable
3
3
3) Current population trend:
Score benefiting
4
rapid decline
1
1
slow decline
2
2 Certainty: high (3)
stable
3
3 medium (2)
2
increasing
4
low (1)
Certainty: high (3)
10) Likely future vulnerability to policy or
medium (2)
management changes:
Score
low (1)
1 highly vulnerable
1
vulnerable
2
2
not vulnerable
3
4) Range trend in last 50 years
Score benefiting
4
>80% reduction
1
1
>50% reduction
2
2 Certainty: high (3)
>20% reduction
3
3 medium (2)
2
Apparently stable
4
low (1)
Increasing
5
Certainty: high (3)
11) Likely future vulnerability to
Score
medium (2)
natural stressors:
low (1)
1 highly vulnerable
1
vulnerable
2
2
not vulnerable
3
3
5) Current range trend
Score
rapid decline
1
1 Certainty: high (3)
slow decline
2
2 medium (2)
2
stable
3
low (1)
increasing
4
Certainty: high (3)
medium (2)
low (1)
1
6) main current stressors (narrative)
TOTAL SCORE 23
CUMULATIVE CERTAINTY SCORE:
17
Max. score 30
Min. score 10
Species score: Vb2 (Vb2. Vb3)
100
-------
APPENDIX H
EXAMPLE APPLICATIONS OF MODULE 2 - ESTIMATING VULNERABILITIES TO
CLIMATE CHANGE
101
-------
MODULE 2 - CATEGORIZING THE VULNERABILITIES OF T&E SPECIES TO CLIMATE CHANGE (Vc)
Species - Golden-cheeked warbler (bold scores represent "best estimates" others are alternate scores)
1) Physiological vulnerability to temperature increase:
Score
6) Likely extent of habitat loss
Likely highly sensitive
1
1
due to climate change
Likely moderately sensitive
2
2
all or most (>50%)
Score
Likely insensitive
3
some (20-50%)
1
1
likely to benefit
4
no change
2
2
some gain (20-50%)
3
Certainty: high (3)
large gain (>50%)
4
medium (2)
5
low (1)
1
Certainty: high (3)
medium (2)
low (1)
2
2) Physiological vulnerability to precipitation change:
Score
Likely highly sensitive
1
Likely moderately sensitive
2
2
7) Ability of habitats to shift at same rate
Likely insensitive
3
3
as species:
Likely to benefit
4
highly unlikely
Score
unlikely
1
1
Certainty: high (3)
likely
2
2
medium (2)
2
3
low (1)
Certainty: high (3)
medium (2)
3
low (1)
3) Vulnerability to changes in frequency/degree
of extreme weather events:
Score
Likely highly sensitive
1
1
8) Availability of habitat within new range:
Likely moderately sensitive
2
2
none
Score
Likely insensitive
3
limited extent
1
1
likely to benefit
4
large extent
2
3
2
Certainty: high (3)
3
Certainty: high (3)
medium (2)
medium (2)
3
low (1)
low (1)
4) Dispersive capability:
Score
9) Dependence on temporal inter-relations:
Low
1
highly dependent
Score
Moderate
2
moderately dependent
1
1
High
3
3
independent
2
3
2
Certainty: high (3)
3
Certainty: high (3)
medium (2)
medium (2)
low (1)
low (1)
2
5) Degree of habitat specialization:
Score
10) Dependence on other species:
Highly specialized
1
1
highly dependent
Score
Moderately specialized
2
moderately dependent
1
1
Generalist
3
independent
2
3
2
Certainty: high (3)
3
Certainty: high (3)
medium (2)
medium (2)
3
low (1)
low (1)
TOTAL
CUMULATIVE CERTAINTY SCORE:
25
Species score: Vc1 (Vc1.Vc2)
102
-------
MODULE 2 - CATEGORIZING THE VULNERABILITIES OF T&E SPECIES TO CLIMATE CHANGE (Vc)
Species - Bald Eagle (bold scores represent "best estimates" others are alternate scores)
1) Physiological vulnerability to temperature increase:
Score
6) Likely extent of habitat loss
Likely highly sensitive
1
due to climate change
Score
Likely moderately sensitive
2
2
all or most (>50%)
1
Likely insensitive
3
3
some (20-50%)
2
2
likely to benefit
4
no change
3
3
some gain (20-50%)
4
Certainty: high (3)
large gain (>50%)
5
medium (2)
2
low (1)
Certainty: high (3)
medium (2)
2
low (1)
2) Physiological vulnerability to precipitation change:
Score
Likely highly sensitive
1
Likely moderately sensitive
2
2
7) Ability of habitats to shift at same rate
Likely insensitive
3
3
as species:
Score
Likely to benefit
4
highly unlikely
1
1
unlikely
2
2
Certainty: high (3)
likely
3
medium (2)
2
low (1)
Certainty: high (3)
3
medium (2)
low (1)
3) Vulnerability to change in frequency/degree
of extreme weather events:
Score
Likely highly sensitive
1
8) Availability of habitat within new range:
Score
Likely moderately sensitive
2
none
1
Likely insensitive
3
3
limited extent
2
likely to benefit
4
large extent
3
3
Certainty: high (3)
3
Certainty: high (3)
medium (2)
medium (2)
2
low (1)
low (1)
4) Dispersive capability:
Score
9) Dependence on temporal inter-relations:
Score
Low
1
highly dependent
1
Moderate
2
moderately dependent
2
2
High
3
3
independent
3
3
Certainty: high (3)
3
Certainty: high (3)
medium (2)
medium (2)
2
low (1)
low (1)
5) Degree of habitat specialization:
Score
10) Dependence on other species:
Score
Highly specialized
1
highly dependent
1
Moderately specialized
2
2
moderately dependent
2
2
Generalist
3
3
independent
3
3
Certainty: high (3)
Certainty: high (3)
medium (2)
2
medium (2)
2
low (1)
low (1)
TOTAL
CUMULATIVE CERTAINTY SCORE:
23
Species score: Vc3 (Vc2. Vc4)
103
-------
MODULE 2 - CATEGORIZING THE VULNERABILITIES OF T&E SPECIES TO CLIMATE CHANGE (Vc)
Species - Salt marsh harvest mouse (bold scores represent "best estimates" others are alternate scores)
1) Physiological vulnerability to temperature increase:
Score
6) Likely extent of habitat loss
Likely highly sensitive
1
due to climate change
Score
Likely moderately sensitive
2
2
all or most (>50%)
1
1
Likely insensitive
3
3
some (20-50%)
2
2
likely to benefit
4
no change
3
some gain (20-50%)
4
Certainty: high (3)
large gain (>50%)
5
medium (2)
low (1)
1
Certainty: high (3)
medium (2)
2
low (1)
2) Physiological vulnerability to precipitation change:
Score
Likely highly sensitive
1
Likely moderately sensitive
2
2
7) Ability of habitats to shift at same rate
Likely insensitive
3
3
as species:
Score
Likely to benefit
4
highly unlikely
1
1
unlikely
2
2
Certainty: high (3)
likely
3
medium (2)
low (1)
1
Certainty: high (3)
medium (2)
2
low (1)
3) Vulnerability to change in frequency/degree
of extreme weather events:
Score
Likely highly sensitive
1
1
8) Availability of habitat within new range:
Score
Likely moderately sensitive
2
none
1
Likely insensitive
3
limited extent
2
2
likely to benefit
4
large extent
3
Certainty: high (3)
Certainty: high (3)
3
medium (2)
2
medium (2)
low (1)
low (1)
4) Dispersive capability:
Score
9) Dependence on temporal inter-relations:
Score
Low
1
1
highly dependent
1
Moderate
2
moderately dependent
2
High
3
independent
3
3
Certainty: high (3)
3
Certainty: high (3)
medium (2)
medium (2)
low (1)
low (1)
1
5) Degree of habitat specialization:
Score
10) Dependence on other species:
Score
Highly specialized
1
1
highly dependent
1
Moderately specialized
2
moderately dependent
2
Generalist
3
independent
3
3
Certainty: high (3)
3
Certainty: high (3)
medium (2)
medium (2)
low (1)
low (1)
1
TOTAL
CUMULATIVE CERTAINTY SCORE:
19
Species score: Vc2 (Vc2)
104
-------
MODULE 2 - CATEGORIZING THE VULNERABILITIES OF T&E SPECIES TO CLIMATE CHANGE (Vc)
Species - Mount Graham red squirrel (bold scores represent "best estimates" others are alternate scores)
1) Physiological vulnerability to temperature increase:
Score
6) Likely extent of habitat loss
Likely highly sensitive
1
due to climate change
Score
Likely moderately sensitive
2
2
all or most (>50%)
1
1
Likely insensitive
3
3
some (20-50%)
2
2
likely to benefit
4
no change
3
some gain (20-50%)
4
Certainty: high (3)
large gain (>50%)
5
medium (2)
2
low (1)
Certainty: high (3)
medium (2)
2
low (1)
2) Physiological vulnerability to precipitation change:
Score
Likely highly sensitive
1
Likely moderately sensitive
2
2
7) Ability of habitats to shift at same rate
Likely insensitive
3
3
as species:
Score
Likely to benefit
4
highly unlikely
1
1
unlikely
2
2
Certainty: high (3)
likely
3
medium (2)
2
low (1)
Certainty: high (3)
3
medium (2)
low (1)
3) Vulnerability to change in frequency/degree
of extreme weather events:
Score
Likely highly sensitive
1
1
8) Availability of habitat within new range:
Score
Likely moderately sensitive
2
2
none
1
1
Likely insensitive
3
limited extent
2
likely to benefit
4
large extent
3
Certainty: high (3)
Certainty: high (3)
3
medium (2)
2
medium (2)
low (1)
low (1)
4) Dispersive capability:
Score
9) Dependence on temporal inter-relations:
Score
Low
1
1
highly dependent
1
Moderate
2
moderately dependent
2
2
High
3
independent
3
3
Certainty: high (3)
3
Certainty: high (3)
medium (2)
medium (2)
2
low (1)
low (1)
5) Degree of habitat specialization:
Score
10) Dependence on other species:
Score
Highly specialized
1
1
highly dependent
1
Moderately specialized
2
moderately dependent
2
2
Generalist
3
independent
3
3
Certainty: high (3)
3
Certainty: high (3)
medium (2)
medium (2)
2
low (1)
low (1)
TOTAL
CUMULATIVE CERTAINTY SCORE:
24
Species score: Vc2 (Vc1. Vc2)
105
-------
MODULE 2 - CATEGORIZING THE VULNERABILITIES OF T&E SPECIES TO CLIMATE CHANGE (Vc)
Species - Desert tortoise (bold scores represent "best estimates" others are alternate scores)
1) Physiological vulnerability to temperature increase:
Score
6) Likely extent of habitat loss
Likely highly sensitive
1
due to climate change
Score
Likely moderately sensitive
2
2
all or most (>50%)
1
Likely insensitive
3
3
some (20-50%)
2
2
likely to benefit
4
no change
3
3
some gain (20-50%)
4
Certainty: high (3)
large gain (>50%)
5
medium (2)
2
low (1)
Certainty: high (3)
medium (2)
2
low (1)
2) Physiological vulnerability to precipitation change:
Score
Likely highly sensitive
1
Likely moderately sensitive
2
2
7) Ability of habitats to shift at same rate
Likely insensitive
3
3
as species:
Score
Likely to benefit
4
highly unlikely
1
unlikely
2
2
Certainty: high (3)
likely
3
3
medium (2)
low (1)
1
Certainty: high (3)
medium (2)
2
low (1)
3) Vulnerability to change in frequency/degree
of extreme weather events:
Score
Likely highly sensitive
1
8) Availability of habitat within new range:
Score
Likely moderately sensitive
2
2
none
1
1
Likely insensitive
3
limited extent
2
2
likely to benefit
4
large extent
3
Certainty: high (3)
Certainty: high (3)
medium (2)
2
medium (2)
2
low (1)
low (1)
4) Dispersive capability:
Score
9) Dependence on temporal inter-relations:
Score
Low
1
1
highly dependent
1
1
Moderate
2
moderately dependent
2
2
High
3
independent
3
Certainty: high (3)
3
Certainty: high (3)
medium (2)
medium (2)
2
low (1)
low (1)
5) Degree of habitat specialization:
Score
10) Dependence on other species:
Score
Highly specialized
1
highly dependent
1
Moderately specialized
2
2
moderately dependent
2
2
Generalist
3
independent
3
3
Certainty: high (3)
3
Certainty: high (3)
medium (2)
medium (2)
2
low (1)
low (1)
TOTAL
CUMULATIVE CERTAINTY SCORE:
21
Species score: Vc2 (Vc2. Vc3)
106
-------
MODULE 2 - CATEGORIZING THE VULNERABILITIES OF T&E SPECIES TO CLIMATE CHANGE (Vc)
Species - Lahontan cutthroat trout (bold scores represent "best estimates" others are alternate scores)
1) Physiological vulnerability to temperature increase:
Score
6) Likely extent of habitat loss
Likely highly sensitive
1
1
due to climate change:
Score
Likely moderately sensitive
2
2
all or most (>50%)
1
1
Likely insensitive
3
some (20-50%)
2
2
Likely to benefit
4
no change
3
some gain (20-50%)
4
Certainty: High (3)
large gain (>50%)
5
medium (2)
2
low (1)
Certainty: High (3)
medium (2)
low (1)
1
2) Physiological vulnerability to precipitation change:
Score
Likely highly sensitive
1
1
7) Ability of habitats to shift at same rate
Likely moderately sensitive
2
2
as species:
Score
Likely insensitive
3
highly unlikely
1
1
Likely to benefit
4
unlikely
2
2
likely
3
Certainty: High (3)
medium (2)
2
Certainty: High (3)
low (1)
medium (2)
2
low (1)
3) Vulnerability to change in frequency/degree
8) Availability of habitat within new range:
Score
of extreme weather events:
Score
none
1
Likely highly sensitive
1
limited extent
2
2
Likely moderately sensitive
2
2
large extent
3
Likely insensitive
3
3
Likely to benefit
4
Certainty: High (3)
medium (2)
2
Certainty: High (3)
low (1)
medium (2)
low (1)
1
9) Dependence on temporal inter-relations:
Score
highly dependent
1
4) Dispersive capability:
Score
moderately dependent
2
2
Low
1
1
independent
3
Moderate
2
High
3
Certainty: High (3)
medium (2)
Certainty: High (3)
3
low (1)
1
medium (2)
low (1)
10) Dependence on other species:
Score
highly dependent
1
5) Degree of habitat specialization:
Score
moderately dependent
2
2
Highly specialized
1
independent
3
3
Moderately specialized
2
2
Generalist
3
Certainty: High (3)
medium (2)
2
Certainty: High (3)
low (1)
medium (2)
2
TOTAL ' '
CUMULATIVE CERTAINTY SCORE:
18
Species score: Vc2 (Vc1. Vc2)
Inerable
107
-------
APPENDIX I
EXAMPLE APPLICATIONS OF MODULE 3 - ESTIMATING OVERALL
VULNERABILITY SCORES
108
-------
MODULE 3 - COMBINING BASELINE AND CLIMATE VULNERABILITY SCORES
INTO OVERALL VULNERABILITY SCORE (Vo)
Species: Golden-cheeked Warbler (bold and italics show "best estimate" and "alternate" scores, respectively
Vb1
Vb2
Vb3
Vb4
Vc1
Vol
Vol
Vol
Vo2
Vc2
Vol
Vol
Vo2
Vo3
Vc3
Vol
Vo2
Vo3
Vo4
Vc4
Vol
Vo2
Vo3
Vo4
Vc5
Vo2
Vo3
Vo4
Vo4
Vol - Critically Vulnerable
Vo2 - Highly Vulnerable
Vo3 - Less Vulnerable
Vo4 - LeastVulnerable
Species score: Vol (Vo2)
MODULE 3 - COMBINING BASELINE AND CLIMATE VULNERABILITY SCORES
INTO OVERALL VULNERABILITY SCORE (Vo)
Species: Bald eagle (bold and italics show "best estimate" and "alternate" scores, respectively
Vb1
Vb2
Vb3
Vb4
Vc1
Vol
Vol
Vo2
Vo3
Vc2
Vol
Vol
\/o2
Vo3
Vc3
Vol
Vo2
Vo3
Vo4
Vc4
Vol
Vo2
Vo3
Vo4
Vc5
Vo2
Vo3
Vo4
Vo4
Vol - Critically Vulnerable
Vo2 - Highly Vulnerable
Vo3 - Less Vulnerable
Vo4 - LeastVulnerable
Species score: Vo3 (Vo2, Vo4)
MODULE 3 - COMBINING BASELINE AND CLIMATE VULNERABILITY SCORES
INTO OVERALL VULNERABILITY SCORE (Vo)
Species: Salt marsh harvest mouse (bold and italics show "best estimate" and "alternate" scores, respectively
Vb1
Vb2
Vb3
Vb4
Vc1
Vol
Vol
Vo2
Vo3
Vc2
Vol
Vol
Vo2
Vo3
Vc3
Vol
Vo2
Vo3
Vo4
Vc4
Vol
Vo2
Vo3
Vo4
Vc5
Vo2
Vo3 Vo4
Vo4
Species score: Vo1(Vo1,Vo2)
Vol - Critically Vulnerable
Vo2 - Highly Vulnerable
Vo3 - Less Vulnerable
Vo4 - LeastVulnerable
109
-------
MODULE 3 - COMBINING BASELINE AND CLIMATE VULNERABILITY SCORES
INTO OVERALL VULNERABILITY SCORE (Vo)
Species: Mount Graham red squirrel (bold and italics show "best estimate" and "alternate" scores, respectively
Vb1
Vb2
Vb3
Vb4
Vc1
Vol
Vol
Vo2
Vo3
Vc2
Vol
Vol
Vo2
Vo3
Vc3
Vol
Vo2
Vo3
Vo4
Vc4
Vol
Vo2
Vo3
Vo4
Vc5
Vo2
Vo3 Vo4
Vo4
Vol - Critically Vulnerable
Vo2 - Highly Vulnerable
Vo3 - Less Vulnerable
Vo4 - LeastVulnerable
Species score: Vol (Vol, Vo2)
MODULE 3 - COMBINING BASELINE AND CLIMATE VULNERABILITY SCORES
INTO OVERALL VULNERABILITY SCORE (Vo)
Species: Desert tortoise (bold and italics show "best estimate" and "alternate" scores, respectively
Vb1
Vb2
Vb3
Vb4
Vc1
Vol
Vol
Vo2
Vo3
Vc2
Vol
Vol
Vo2
Vo3
Vc3
Vol
Vo2
Vo3
Vo4
Vc4
Vol
Vo2
Vo3
Vo4
Vc5
Vo2
Vo3 Vo4
Vo4
Vol - Critically Vulnerable
Vo2 - Highly Vulnerable
Vo3 - Less Vulnerable
Vo4 - LeastVulnerable
Species score: Vo2 (Vol, Vo3)
MODULE 3 - COMBINING BASELINE AND CLIMATE VULNERABILITY SCORES
INTO OVERALL VULNERABILITY SCORE (Vo)
Species: Lahontan cutthroat trout (bold and italics show "best estimate" and "alternate" scores, respectively
Vb1
Vb2
Vb3
Vb4
Vc1
Vol
Vol
Vo2
Vo3
Vc2
Vol
Vol
\/o2
Vo3
Vc3
Vol
Vo2
Vo3
Vo4
Vc4
Vol
Vo2
Vo3
Vo4
Vc5
Vo2
Vo3 Vo4
Vo4
Species score: Vol (Vol, Vo2)
Vol - Critically Vulnerable
Vo2 - Highly Vulnerable
Vo3 - Less Vulnerable
Vo4 - LeastVulnerable
110
-------
APPENDIX J
EXAMPLE APPLICATIONS OF MODULE 4 - CERTAINTY/UNCERTAINTY ANALYSIS
111
-------
MODULE 4
- CERTAINTY/UNCERTAINTY ANALYSIS
Species:
Golden-cheeked warbler
Max. Scores
Min. Scores
Module 1
30
10
Module 2
30
10
Both Modules
60
20
Total Score
Certainty Evaluation
20-32
Low
33-45
Medium
>45
High
Module 1
Module 2
Both
Total score
26
25
51
Certainty Score - High
MODULE 4 - CERTAINTY/UNCERTAINTY ANALYSIS
Species: Bald eagle
Module 1
Module 2
Both Modules
Total Score
20-32
33-45
45
Max. Scores
30
30
60
Certainty Evaluation
Low
Medium
High
Min. Scores
10
10
20
Module 1 Module 2 Both
Total score
26
23
49
Certainty Score - High
MODULE 4
- CERTAINTY/UNCERTAINTY ANALYSIS
Species:
Salt marsh harvest mouse
Max. Scores
Min. Scores
Module 1
30
10
Module 2
30
10
Both Modules
60
20
Total Score
Certainty Evaluation
20-32
Low
33-45
Medium
>45
High
Module 1
Module 2
Both
37
Total score
18
19
Certainty Score - Medium
112
-------
MODULE 4
-CERTAINTY/UNCERTAINTY ANALYSIS
Species:
Mount Graham red squirrel
Max. Scores
Min. Scores
Module 1
30
10
Module 2
30
10
Both Modules
60
20
Total Score
Certainty Evaluation
20-32
Low
33-45
Medium
>45
High
Module 1
Module 2
Both
47
Total score
23
24
Certainty Score - High
MODULE 4
-CERTAINTY/UNCERTAINTY ANALYSIS
Species:
Desert tortoise
Max. Scores
Min. Scores
Module 1
30
10
Module 2
30
10
Both Modules
60
20
Total Score
Certainty Evaluation
20-32
Low
33-45
Medium
>45
High
Module 1
Module 2
Both
45
Total score
24
21
Certainty Score - Medium
MODULE 4 - CERTAINTY/UNCERTAINTY ANALYSIS
Species: Lahontan cutthroat trout
Vlodule 1
Vlodule 2
Both Modules
Total Score
20-32
33-45
>45
Max. Scores
30
30
60
Certainty Evaluation
Low
Medium
High
Min. Scores
10
10
20
Total score
Module 1
17
Module 2
18
Both
35
Certainty Score - Medium
113
------- |