|
Items in cart [0] |
Questions? Call 888-624-8373 |
| Copyright © 2009. National Academy of Sciences. All rights reserved. Terms of Use and Privacy Statement |
Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 117
Chapter 8
Making ESA Decisions in the Face of Uncertainty
The previous chapter described our ability to estimate the risk of extinction for
populations of organisms. This is a major part of estimating the degree of endangerment of a population
or species. In this chapter, we build on that information to discuss how important understanding and
assessing risk is in ESA decision making and the question of whether different levels of risk should
apply to different decisions. Finally, because decisions regarding endangered species must always be
made in the face of uncertainty regarding estimates of extinction risk and future events, we suggest ways
of improving agency decisions involving risk and uncertainty.
DECISIONS REQUIRED UNDER THE ESA
The objectives of the ESA are to conserve the ecosystems upon which endangered and
threatened species depend, provide a program for the conservation of endangered and
threatened species, and achieve the purposes of several international conservation agreements. While
these objectives are not intrinsically or philosophically conflicting, they can conflict when agencies faced
with limited budgets must decide how to apportion funds. More serious conflicts arise when the
objectives of the act conflict with other human objectives, such as private-property rights and private and
public developments. The act specifies the extent to which such conflicting objectives should be
considered when making the different decisions required under the act. Consideration of human
objectives other than those specified in the act is specifically prohibited when making decisions
regarding listing, "take," and "jeopardy," but is required when making decisions regarding critical
habitat. Initial recovery planning is to be based solely on scientific considerations, but economic effects
of the plan are to be considered before implementation.
The terminology of the act implies that many decisions regarding conservation of species should
consider estimates of extinction risk. Specific examples of such terminology include the definitions of
endangered and threatened species, the provisions for removing species from the list, and the definitions
of jeopardy on public lands and taking on private lands (see Box 8-1).
THE NEED FOR NEW APPROACHES TO DECISION MAKING
Agency decisions that have been taken under the direction of the act have been criticized by the
general public (Mann and Plummer, 1992) and the scientific community (Brownell et al., 1989,
Goodman, 1993; Wilcove et al., 1993, Tear et al., 1993) for being arbitrary. For example, Brownell et al.
(1989) pointed out that, at least for cetaceans, the list of threatened and endangered species is not
scientifically defensible. Several large whales in little danger of extinction were on the list in 1989,
although many small dolphins and porpoises with very small population sizes were not. A recent survey
of the status of species at the time of listing concluded that "the scientific rationale for listing decisions is
absent or weak" (Wilcove et se. 1993~. Population size alone is not necessarily a sensitive indicator of
extinction risk, but many species had declined to alarmingly low numbers by the time they were listed.
117
OCR for page 118
118
Science and the Endangered Species Act
Wilcove et al. (1993) found that the median population sizes of species when listed were about 1,000
individuals for animals and 100 inclivicluals for plants.
The backlog of candidate species awaiting listing and the number of legal actions taken relater! to
listing decisions indicate that the current process for making listing decisions needs review and revision.
The Fish ant! Wildlife Service (FWS) uses three categories to help make such decisions (see Box 8-2~.
Category 1 now contains about 400 species for which the agencies have substantial information to
support the proposal to list but do not have sufficient resources to complete the process. Category 2 now
contains about 3,500 species for which a petition to list might be justified but is cleemed to be lacking in
critical information. The large number of species in Category 2 indicates that the agencies have not
developed a workable process for making listing decisions when faced with limited data, uncertainty, and
limited financial and human resources. Indeed, lack of scientific information and uncertainty plague
many public anti private policy decisions.
Recovery plans often fad! to provide appropriate guidance on biologically reasonable levels of risk.
A recent survey of recovery plans concluder! that goals for species recovery were often unrealistically
low and that these plans frequently manage for extinction rather than survival (Tear et al., 1993~. For
example, of the 54 threatened ant] endangered species for which population size data were available, 15
(28%) had recovery goals set at or below the existing population size at the time the plan was written
(Tear et al., 19931. If the population was endangered at the time of listing and the review of Wilcove et
al. (1993) suggests that most listed populations are at risk ---- then such recovery goals are probably too
low.
The committee also notes that some recovered species have not been delistect or upgraded (i.e.,
from enciangered to threatened) in a timely fashion. For example, eastern Pacific gray whales
(Eschrictius robustus) were not delisted until 1994 (Fed. Reg 59:31094), although population estimates
were approaching the likely pre-exploitation sizes of 12,000-15,000 by the late 1970s ant! have continued
to increase (Reilly, 1992~. The comments about decisions to list also apply to decisions about Relisting.
ENDANGERED SPECIES: any species which is in danger of extinction throughout all or a significant
portion of its range.
THREATENED SPECIES: any species which is likely to become an endangered species within the
foreseeable future throughout all or a significant portion of its range.
CRITICAL HABITAT: The specific areas within the geographical area occupied by the species . . . on
which are found those physical or biological features essential to the conservation of the species and which
may require special management considerations or protection.
REMOVAL FROM LIST (RECOVERY): [each recovery plan shall include] objective, measurable criteria
which, when met, would result in a determination, in accordance with the provisions of this section, that the
species be removed from the list.
JEOPARDY ON PUBLIC LANDS: [each federal agency shall insure that any action authorized, funded or
carried out by such agency] is not likely to jeopardize the continued existence of any endangered species or
threatened species or result in the destruction or adverse modification of habitat of such species which is
determined to be critical.
TAKE IN PUIVATU ~ Aims
__ ~ ,. taking will not appreciably reduce the likelihood of the survival and
recovery of the species in the wild.
Box 8-l Examples of terminology in the ESA (italicized) that implies an assessment of the degree of
risk to a species.
OCR for page 119
Making ESA Decisions in the Face of Uncertainty
PROVIDING OBJECTIVE RISK STANDARDS
119
The qualitative definitions in the act provide a framework for decision-making, that is, they
provide a list of administrative and management actions and a general rationale for selecting each action.
However, qualitative clefinitions alone can be interpreter! in different ways by different people, ant!
agencies have provided no guidance on the appropriate degrees of extinction risk for making the
different decisions requires! by the act, such as listing a species as either threatened or enciangered or
clecIaring a species recovered. To ensure that ESA decisions protect endangered species as they are
intended to and do so in a scientifically defensible way requires objective methods for assessing risk of
extinction (see Chapter 7) and for assigning species to categories of protection according to their risk of
extinction. Stanclards for assigning species to categories shout! be quantitative wherever possible, and,
when this is not possible, qualitative procedures should at least be systematic and clearly defined.
Category 1. Taxa for which the Service has on file sufficient information on biological vulnerability and
threatens) to support proposals to list them as endangered or threatened species. Proposed rules have not yet
been issued because this action is precluded at present by other listing activity. Development and
publication of proposed rules on Category 1 taxa are anticipated, however, and the Service encourages other
Federal agencies to give consideration to such taxa in environmental planning.
Category 2. Taxa for which information now in the possession of the Service indicates that proposing to list
as endangered or threatened is possibly appropriate, but for which sufficient data on biological vulnerability
and threat are not currently available to support proposed rules. The Service emphasizes that these taxa are
not being proposed for listing by this notice, and that there are not current plans for such proposals unless
additional supporting information becomes available. Further biological research and field study usually
will be necessary to ascertain the status of taxa in this category. It is likely that many will be found not to
warrant listing, either because they are not threatened or endangered or because they do not qualify as
species under the definitions in the Act, while others will be found to be in greater danger of extinction than
some taxa in Category 1.
.
~ . . .. .
Category 3. Taxa in Category 3 are not current candidates for listing. Such taxa are further divided into
three subcategories to indicate the reasonks) for their removal from consideration:
Category 3A. Taxa for which the Service has persuasive evidence of extinction. If rediscovered, such taxa
might acquire high priority for listing. At this time, however, the best available information indicates that
the taxa in this subcategory, or the habitats from which they were known, have been lost.
Category 3B.- Names that, on the basis of current taxonomic understanding (usually as represented in
published revisions and monographs), do not represent distinct taxa meeting the Act's definition of
"species." Such supposed taxa could be reevaluated in the future on the basis of new information.
Category 3C. Taxa that have proven to be more abundant or widespread than previously believed and/or
those that are not subject to any identifiable threat. If further research or changes in habitat indicate a
significant decline in any of these taxa, they may be reevaluated for possible inclusion in categories 1 or 2.
Box S-2 U.S. Fish and Wildlife Service Definitions for Categories 1, 2, and 3, under the Endangered
Species Act (Federal Register 58~188~:51145, September 30, 1993~.
OCR for page 120
120
Science and the Endangerec! Species Act
Developing Quantitative Risk Standards
Agencies could achieve greater consistency in decisions if they proviclect quantitative risk
stanciarcis for such terms as "enciangered" and "threatened" to the many agency personnel involved in
implementing the act. Risk levels should be define~i as the probability of extinction within a specific
time. As an example of quantitative guidance, Shaffer (1981) definer! a minimum viable population size
as that which would have at least a 99% chance of surviving for 1,000 years. The committee is not
endorsing Shaffer's definition but noting that by providing quantitative guidance, Shaffer has Marie it
possible to discuss or disagree with his definition objectively and apply the definition in a standard
manner. As another example of a quantitative definition, Mace and Lande (1991) suggested that
enciangerec! could be clefined as a 20% chance of extinction in 20 years.
Time Frame for Estimating Risk of Extinction
When providing a quantitative stanriarc! for assigning risk categories, risk of extinction must be
defined with a specific time *ame in mind, i.e., x probability of extinction in y years. Critical levels of
extinction risk, which will trigger ESA clecisions such as listing or delisting, must be associated with a
particular number of years or generations. As the time frame increases, the probability of extinction also
increases, approaching 100% for all species if the period is long enough (see Chapter 21. The choice of a
time frame for evaluating risk of extinction for purposes of the ESA reflects scientific and societal
concerns. From a scientific standpoint, time should be long enough to guard against making
management choices that might have favorable effects over the short term (e.g., the next 5 to 10 years),
but consequent adverse effects over the longer term (e.g., the next 100 years or more). An example of
such a choice might be management that extends the lifespan of adults currently in a population
(enhancing short-term survival), but jeopardizes successful reproduction (leacling to population clecline
as the current adult population ages an(l dies). Species' life-history characteristics help determine an
appropriate time frame, with longer times being more appropriate for longer-lived species.
Another scientific consideration is the time scale for natural cycles of disturbance and regeneration
in the species' habitat. Evaluating the success of endangered species management over only a portion of
the natural habitat cycle runs the risk of confusing natural fluctuations in population size with adverse
reactions to management.
Societal considerations regarding time frame include the desire to preserve species and their
habitats over time scales meaningful to humans and their offspring, political cycles of 2-6 years,
economic cycles, and many others. Because of the large variety of societal factors, this committee
cannot specify all the appropriate time scales that should be considere(l in decision making or even the
range of time scales for which extinction probabilities should be calculated. However, it is clear that the
time scales implicit in the various calculations should be made explicit for informer! decision making.
In some cases, where there is an immediate and potentially reversible threat to species survival (such as a
propose(l clevelopment), it could make sense to analyze the probability of survival over a short period,
perhaps 5 to 10 years or less, when comparing options for species recovery over the short term. Such an
analysis should be followed, however, by another assessment of extinction risk over a longer period to
ensure that short-term gains do not become long-term losses.
Listing Systems Based on Objective Criteria and Rules
A system for making listing decisions requires a set of objective criteria for assigning species to
OCR for page 121
Making ESA Decisions in the Face of Uncertainty
121
risk categories, such as endangered! or threatened. The objective criteria most suitable for making listing
decisions would be different degrees of extinction risk. However, we rarely have sufficient data to allow
good estimates of extinction risk, so we need a system allowing the use of other criteria as well. Such a
system could be based on objective criteria, such as some combination of population size and number,
believed to represent a specific level of extinction risk.
The need to develop a listing system based on objective criteria has been recognizes! by the
principal international organization concerned with the listing ant! conservation of enciangered species,
the International Union for the Conservation of Nature (IUCN). In June 1992, the CITES (Convention
on the International Trade in Endangered Species of Flora and Fauna) Standing Committee requester!
that IUCN help to clevelop new criteria for listing species in the appendices to the CITES treaty, which
regulates trade in wilcllife and wildlife products. The resulting proposed IUCN System for assessing
ciegree of threat (Mace et al., 1992) was produced after an international effort by groups of scientists in a
series of workshops. The proposer! system is now being evaluated by members of the IUCN/SSC taxon
specialist groups, various CITES committees, and interested scientists (Gnam, 1993) and was considered
at the CITES 1994 meeting. At that meeting, a proposal was made to adopt the IUCN criteria; a
counterproposal was macle by the United States ant! a working group clevelopecl a compromise, which
was recently accepted by CITES (Rosemarie Gnam, FWS, personal communication, March 1, 19951.
Much of the substance of the IUCN criteria remains in Annex 5 of the report as definitions, rather than
the criteria for listing.
In the proposed IUCN system, a species could be listed as endangered! baser! on any of several
criteria, each of which was intendec! to represent approximately the same risk of extinction. Decisions to
list a species could be based on any of the following criteria: probability of extinction, trends in
abunclance, population size, number of populations, and geographical extent (Mace et al., 1992~. The
IUCN system was based on a combination of two previous approaches for assessing degree of threat:
population-basect criteria developer! by those working with higher vertebrates anti area-habitat criteria
cleveloped by those working with plants ant! invertebrates. Population-based criteria, known as the
"Mace-Lance criteria," were proposed in 1991 (Mace and Lande, 1991~. They suggested three
categories:
Critical: 50% probability of extinction within 5 years or two generations, whichever is longer.
Enclangerecl: 20% probability of extinction within 20 years or 10 generations, whichever is longer.
Vulnerable: 10% probability of extinction within 100 years.
Unfortunately, due to the limited data on most taxa of conservation concern, a formal population
viability analysis to estimate the probability of extinction in a particular case is often impossible.
Moreover, population viability models for plants are poorly clevelopecl (Merges, 1990~. To address this
problem, Mace and Lancie offerer! several surrogate criteria baser! on other types of information. For
example, they considered that a population would qualify as endangered if the current population
estimate was fewer than 250 mature inciivicluals, even if insufficient demographic clata were available to
calculate an estimated probability of extinction. The current IUCN criteria rely heavily on population
viability analysis, population size, and range area. Concern has been expressed that these criteria omit
historical data on a species' abundance and distribution, omit data on reproductive fitness, and ignore life
cycles (Gnam, 19931. Furthermore, the IUCN system is still incomplete, because it lacks a set of rules to
allow decisions to be macie when there is uncertainty regarding the criteria used to categorize species
according to relative risk of extinction. Nevertheless, those criteria represent the most important
scientific effort to date to reach consensus on standard criteria for assigning taxa to threat categories in a
uniform, objective manner.
The adoption of a similar system in the U.S. would make listing decisions more consistent.
OCR for page 122
122
Science and the Endangered Species Act
Unfortunately, there is as yet no evidence that the criteria used in the IUCN system do represent
comparable degrees of risk. Any system of criteria clevelopec! for use with the ESA should be thoroughly
tested against a variety of population and metapopulation ~nodels before implementation. Designing and
testing an appropriate system for listing species is a formidable scientific task best accomplishe(1 by an
independent scientific committee. For example, for a much simpler problem, management of the
commercial harvest of whales, the Scientific Committee of the International Whaling Commission took 6
years to evaluate five alternative proposed management procedures (Donovan, 1989; R. Brownell,
personal communication).
Limitations on Estimates of Risk
Our ability to estimate risk of extinction for use in assigning species to protection categories is
limited by our understanding of the factors influencing extinction. Two areas where we are acutely
aware of limitations are in moclels used to estimate risk of extinction and in our understanding of the role
of critical thresholds of risk ant! of cumulative effects of risk factors.
Limits of Models
As clescribed in Chapter 7, models for estimating risk of extinction are limited in their ability to
incorporate the full complexity of species population dynamics. Estimates of risk (lerived from these
moclels may reflect only a subset of the factors actually influencing a species' risk of extinction. As
discussed in Chapter 7, it seems likely that the simplifications and omissions of current moclels can
underestimate risk of extinction.
Poor Understanding of Cumulative Effects and Thresholds
Decisions under the ESA regarding take or jeopardy require making decisions regarding
incremental risks of extinction. Assessing the a(lde(l risk from specific human actions is usually an even
more clifficult task than estimating the overall extinction risk to a species. Individual human actions,
such as developing a few acres of habitat, pose small incremental risks of extinction. At some point,
however, the small incremental risks from numerous human actions, if not stemmed, will accumulate so
as to produce a major effect. Not enough is known about cumulative effects and threshold points to
make accurate risk predictions possible (e.g., BeanIands et al., 1986), although there has been some
theoretical work on critical threshoIcls of habitat patch size or fragmentation (e.g., Lancle, 1988 for
spotted owl habitat fragmentation). When considering the probable effects of incremental human
activities, it is reasonable to assume that adclitional activity means additional risk, but we rarely know
whether the relationship between additional activity and additional risk is linear or whether there may be
critical levels of activity above which the risk of extinction increases dramatically.
Should Different Risk Standards Apply to Different ESA Decisions?
Endangered, Threatened, Recovered
In judging whether different levels of risk should apply to different types of decisions uncler the
OCR for page 123
Making ESA Decisions in the Face of Uncertainty
123
act, the committee consitierect carefully the terminology of the act shown in Box 8-1. The definition of
an endangered species as one that is already in cIanger of extinction ant! a threatened species as one that
is likely to become an endangered species implies that a species listed as enciangerect is at greater risk of
extinction than a threatened one. The determination that a species shouIc! be remover! from the list
implies that its risk of extinction has clecreaseci to the point where it is no longer considered! threatened.
Thus, it is clear that determinations of a single species as successively enciangerect, threatened, and
recovered should represent a series of decreasing levels of risk of extinction faced by the species.
Different Taxa
Although cross-species comparisons are complicatecl by many factors specific to the biology of
inclividual species, it is appropriate to set the same degree of risk as a standard for listing any species,
whether plant or animal, as endangered and another, somewhat lower degree of risk, for listing any
species as threatened. Thus it is reasonable to expect that species determined to be endangered should,
on the average, face a greater risk of extinction titan those determined to be threatened. This seemingly
obvious ordering of risk has not always been followed in practice (Wilcove et al., 19931.
Public Versus Private Land
Controversy has arisen over whether the inclusion of habitat destruction or mollification as a form
of taking under Section 9 sets a different standard of responsibility for protection of endangered species
by private versus public entities. (We note again that this interpretation of taking is under court review at
this writing.) In particular, it has appeared! to some that the standard of responsibility might be
interpreted to be more stringent for private than for public entities. This seems, if anything, the reverse
of what was intended by the ESA. From a scientific perspective, actions resulting in a given clegree of
risk of extinction are equally hazardous to species whether they are carried out by public or private
entities on public or on private lancI. The committee sees no scientific reason for setting different
standards for categorizing risk of extinction under different sections of the act. However, because public
and private entities behave differently, achieving the same degree of biological protection on public
versus private lancis does not necessarily imply identical regulatory requirements on behalf of species
experiencing the same risk of extinction on public and on private lancis.
USING STRUCTURED APPROACHES TO DECISION MAKING
Why use structured approaches to ESA decision making? Because the issues are complex and
relevant scientific data are often fragmentary, it can be difficult to make decisions regarding enclangered
or threatened species. Decisions regarding endangered! species are often characterized by insufficient
data. probabilistic Predictions regarding future events. considerable uncertainty regarding the accuracy of
~ 1 1
,] 1 · , · ~ · , · ~ 1 · ~ · 1 · _ A _ _ _ `1_ _ 1~ _ _ _~ _ _ ~ ___
these predictions, conflicting management objectives, disagreement over the nest course or action, and
the need to justify whatever decision is made.
Given these problems, it is important that we try to make our decision-making process as explicit
as possible, especially because research on the psychology of human (recision making (Hogarth, 1980;
Kahneman et al., 1982; von Winterfel~t and Edwards, 1986) has shown that intuitive decisions exhibit
many inconsistencies and biases, particularly when probabilistic information anti multiple objectives are
involvecI. The use of an explicit framework to guide decisions can help us make choices that are
-
OCR for page 124
124
Science and the Endangered Species Act
consistent with our goals, data, ant! beliefs and facilitate compromise among those with differing views.
Several techniques clevelopeci in the fields of operations research ant! management science the
academic disciplines dealing with scientific approaches to decision making-provide helpful frameworks
for making these difficult decisions. These techniques help decision makers think about the decision in a
systematic way; break clown difficult decisions into a series of smaller, easier decisions; ant! document
the process used to reach the decision, which makes the decision easier to justify and defend.
Some applications of these structured approaches, particularly the qualitative ones, might appear
to be simple pleas for clear thinking, and they are. But clear thinking cioes not come easily or naturally
in the face of scientific complexity and uncertainty, competing objectives within recovery programs anti
beyond, and political pressures from multiple constituencies. We propose decision analysis as an
example of a structures! problem solving method, although other methods (such as the Analytic
Hierarchy Process (Saaty, 1990) or approaches based on goal programming and multiattribute ranking
(RalIs and StarfielcI, 1995) can be useful as well. We stress tile merits of using these tools as conceptual
frameworks, not just as number-crunching devices. Ant! we emphasize that using these approaches is not
necessarily a call for more information, but rather for more coherent use of existing information.
Using Subjective Probability and Expert Opinion
In some cases, there is very little "hard" information that seems relevant to estimating the risks
affecting endangered species, but some experts have accumulated experience that allows them to make
informed judgments about these risks. Such expert judgment is so often available for endangered species
decisions that it is of great benefit to have orclerly methods of eliciting ant! using it for decision making.
One of the strengths of decision analysis is its ability to estimate "subjective probabilities" and then use
them for analysis in the same way that long-term frequency estimates would be used (Behn and Vaupel,
1982; von Winterfel~t and Edwards, 1986; Maguire, 19X71. The methods allow a concrete expression of
expert opinion, facilitating scrutiny by the public ant! comparison with other views.
Another place where expert opinion can be essential is in cases when some background
information is available, but unique aspects of a situation differentiate it from similar situations that have
occurred. An example might be assessing the likelihood! of successful reproduction in a captive
population where experience has been dismal but new reproductive technologies have been clevelopect
that might prove helpful, as for example with the endangered Hawaiian crow (NRC, 1992a). In this case,
it is relevant to amend or update the historical information in light of the new techniques. There are
formal ways to combine oic! data with new opinion (and new data with oIct opinion) using Bayesian
estimation (Raiffa 1968, Clemen, 19911.
Linking Science and Values
Even though estimates of risk are grounde~i in scientific information, those implementing the act
often make value judgments when making decisions about listing, jeoparcly, etc. They are deciding
which species to list quickly and which to relegate to delayed listing, which areas of habitat are worth the
socioeconomic cost or political effort to designate as critical and which are not, what degree of jeopardy
is worth altering a fecierally funded development project for and what is not. Many citizens are willing
to allow public officials to make such judgments on their behalf, but those involved might be more
comfortable if the values informing those judgments and their effects on ESA decisions were articulated
more clearly.
OCR for page 125
Making ESA Decisions in the Face of Uncertainty
125
A hallmark of formal decision analysis (Behn and Vaupel, 1982, Clemen, 1991) and other
structured problem-solving methods (RalIs and Starfield, 1995) is an emphasis on articulating clearly the
objectives for a decision and criteria for evaluating how well alternative proposals might meet those
objectives. Use of such methods can improve ESA decisions by making the connection between values,
objectives ant! decisions more transparent, helping to disarm criticisms that the government is capricious
or partisan in implementing the act (Mann and Plummer, 1992, Tear et al., 19931.
Making good use of science, as instructed in the ESA, requires making appropriate connections
between the values and objectives being pursued in a ciecision and the scientific evidence and reasoning
user! to evaluate alternative ways of meeting those objectives. Science by itself is not sufficient input to
policy decisions, apart from the objectives and values it serves. Articulating an explicit framework can
help link science and values and lead to better and more defensible decisions.
Scientific Uncertainty in ESA Decisions
For even the best-stuctied endangered species, essential pieces of information might be lacking, yet
decisions must be macle. Sometimes it is possible to delay action while gathering better information,
although that strategy carries its own risks. Sometimes important factors affecting how management
actions turn out, such as catastrophic weather conditions or pollution accidents, are inherently uncertain,
and no amount of further study could do more than improve the accuracy and precision of estimates of
their likelihoods. In any case, weighing the best choice under uncertainty about outcomes is a necessity.
This kind of probabilistic reasoning tioes not come naturally and many managers are uncomfortable with
it, resorting to shortcut heuristics to simplify information and justify their choices (von Winterfelcit and
Edwar~is, 1986~. The framework of decision analysis offers a structure for considering probabilistic
information in a coherent and consistent way, providing better use of whatever information is available
to guide decisions.
Estimating Uncertain Quantities
Several types of information bearing on ESA decisions can be uncertain, including factors
influencing how management actions turn out (such as weather events or sociopolitical events) and
measures of outcome (such as the likelihood of population persistence uncler a particular management
strategy). Such probabilistic quantities enter the analysis in different ways, but in both cases, methods of
estimating those quantities are needed. Sometimes relevant long-term tiara on the frequency of an
uncertain event can be used to estimate its probability; as an example, there are weather records for
severe storms in particular coastal areas. These could be used to estimate the likelihood! of a tropical
storm striking the Gulf Coast of Texas near the wintering ground for the migratory whooping crane
population or the probability that a hurricane would destroy the habitat of some red-cockacled
woodpeckers. In one case, population models were used to help decisions concerning conservation of
endangered sea turtles by identifying the most sensitive life stages (NRC, 1992b).
In other cases, no useful long-term data are available, but there may be moclels incorporating our
best unclerstanding of the factors affecting the likelihood of an uncertain event. Examples are stochastic
population moclels, such as are discussed in Chapter 7. These use information about demographic
parameters and environmental factors to predict probability of extinction for a population with particular
characteristics.
OCR for page 126
126
.
Science and the Endangered Species Act
Reducing Uncertainty by Gathering Information
Listing actions, recovery plans, or other ESA decisions often are clelayect clue to inadequate
information. Those implementing the act almost always believe that with aciciitional information, they
conic! make a better decision. Nevertheless, decisions to delay action pencling further information and
directives to gather additional information should be viewed critically. What kinc! of information and of
what quality collie be gathered within the time and resources available? What are the possible answers
that such investigation might reveal? What decisions would be triggered by different answers? How are
those decisions different from those that would be made using existing information? What effect will
continuing the status quo have on species status ant! on options for future action? Considering these
questions in a structures! framework can make it snore likely that a reasonable decision will be macle.
An example of such an analysis comes from B. art and Robson (19941. They analyzed the
variability in raptor population data to find out how manY Years of data woul~i be required to tell whether
, ,
the northern spotted owl population was stable or increasing, a criterion for delisting specifier! in the
spotter! owl recovery plan. Their analysis showed that it would not be possible to render a sound
judgment about clelisting basest on fewer than 8 years of population monitoring data. (See Box 8-3 on
statistical power ant] types of errors.) Such an analysis promotes realistic expectations about the time
and effort required to obtain a satisfactory answer and forestalls charges that the FWS is delaying
clelisting for nonscientific reasons.
Sometimes a qualitative but orderly consideration of questions is sufficient to guide action, giving
.
.
managers the confidence either to pursue additional Information or proceed on the basis ot the
information they aireacly have. At other times, a snore formal analysis of the value of information
(Raiffa, 1968, Clemen, 1991) Night be needed. In either case, the scientific uncertainty must be
examiner! within the context set by the objectives for a particular situation. The question of how many
years of monitoring ciata would be required to make a decision about delisting the northern spotted owl
can be answered only with reference to the level of confidence requirecl, which can be cletermined only
with reference to the objectives of spotter! owl management (Taylor and Gerrodette, 19931.
More information almost always seems better to those trainee! as cautious natural resource
scientists. Yet, too much risk aversion, or fear of making the wrong decision basest on limited
information, can be crippling. The California condor and the black-footed ferret are good examples. In
these cases, ant! in many similar ones worIdwicie, a management decision to remove inctivicluals from the
witcl to begin or to augment a captive-breecling effort would have an indisputably negative effect on the
survival of the population in the wile! (Maguire, 1986; 19891. Such a decision must be justified in terms
of the likelihood of extinction of the population even if no removals were made, and the long-term
benefit to species survival from a captive breeding program, if successful. In the condor and ferret
programs, only when it became clear that a continuation of status-quo management of the wild
population wouIcl lead to (disaster clid the negative consequences of removals become acceptable. But by
that point, it was very nearly too late to make a success of the captive-breecting programs (RalIs and
Ballou, 19921. Although both programs were ultimately successful in producing animals for
reintroduction, neither case can be cites! as a mode! of informed decision making under uncertainty. In
both cases, too much attention was given to the possible negative consequences of the novel strategy
(capturing animals from the wild) and too little to the possible negative consequences of continuing the
conservative approach (trying to protect the dwindling wild populations).
OCR for page 127
Making ESA Decisions in the Face of Uncertainty
Types of Errors
.
127
In the many cases where it is not possible to gather ant! analyze more information before making a
decision, using a formal structure like decision analysis can steep managers consider explicitly the ways
in which they might be wrong in their predictions (e.g., about whether a particular population is
genetically distinct from others or whether loss of a particular habitat area will lead to extinction) and the
biological anti socioeconomic consequences of being wrong in various ways (such as failing to predict
extinction when it will in fact occur, or declaring a population genetically ctistinct when it is not). Many
times the consequences of being wrong are highly asymmetrical; that is, one type of error is much more
serious for the species than the converse, ant! perhaps even irreversible (Box 8-3, Table 8-11. Taylor and
Gerrodette (1993) reinforced this point in their discussion of using statistical power analyses to design
and evaluate monitoring schemes for endangered species.
TABLE 8-1 Consequences of Making Two Types of Statistical Errors
When Evaluating Scientific Data on Endangered Species
Type 1 Error
Type 2 Error
Reject true null hypothesis
Claim an effect when none exists
Protect species more than necessary
Lose scientific credibility
Increase socioeconomic costs more
than necessary
Accept false null hypothesis
Claim no effect when one exists
Protect species less than necessary, even
lose species
Lose practical, and scientific, credibility
Permit activities that should not have been
approved
Source: Adapted from Noss (19921.
OCR for page 132
132
Science and the Endangered Species Act
Scientists must try to avoid both types of errors. However, it is impossible to simultaneously
minimize the likelihood of making both types of errors. The more we try to avoid making the first type
of error, the more likely it is that we will make the second type. Scientists are trained to minimize the
probability of making the first error, that is rejecting a null hypothesis when it is actually true (Box S-
3~. This choice is appropriate for advancing scientific knowlecige, but it might not be the best for
making management decisions. If not examined explicitly, this asymmetric error structure can bias
decisions under the act to the detriment of endangered species, especially if they are based on analyses
that clo not take the asymmetric risk function into account. One situation where this can occur is listing
decisions for species where information on population status is limited, a common occurrence. If a
statistical analysis is performed, the trigger for listing is rejection of the null hypothesis that the species
is not endangered. Typical error rates for such statistical tests of hypotheses keep the likelihood! of
false rejection low, but at the expense of substantial risk of falsely concluding that a species is not
endangered. In the absence of conclusive evidence that a species is in fact endangerecI, uncertainty
about status can result in acceptance of the null hypothesis, whether true or not. This results in an
asymmetric risk function for the species (i.e., the probability that the species will not be protected when
protection is needed is greater than the probability that the species will be protected unnecessarily),
because the null hypothesis is usually that a population has not declined or that a specific action will
have no effect (Noss, 1992; Taylor anci Gerrociette, 19931. Furthermore, limited data often result in
inadequate statistical power. Thus, the null hypothesis of no impact on an endangered species might
not be rejected when it shouicl have been (Taylor and Gerrodette 1993~. As a result, conservation
measures that shouIct be undertaken will not be.
Burden of Proof
In the section above on types of errors that might be macle when making decisions uncler
uncertainty, we have shown that the consequences of different types of errors might be asymmetrical,
such that it is more important to avoid one type of error than another. We have suggested that ESA
decisions should explicitly recognize that the consequences of different types of errors can differ and
design decision-making procedures accordingly. An aspect of these clecision-making procedures that
we would like to emphasize here is the issue of "burcien of proof." By this, we mean presumptions
about what a party to an ESA decision should have to demonstrate to trigger protective actions. We are
concerned that some current procedures may, perhaps inadvertently, bias decision-making in ways that
are not intended uncler the act.
Statistical Errors
In the usual procedures for formulating scientific tests of hypotheses, it is customary to phrase
the null hypothesis as the "no effect" case (e.~., a proposed action will not affect the survival of a listed
- --at r ~ - - is- -Cal- 7 -- l-- -1- -a
species) and to use confidence levels that limit the probability of falsely rejecting that null hypothesis to
a known level (often, 0.05) while permitting much larger probabilities of falsely accepting the null
hypothesis. We are concerned that when such statistical procedures are followed in ESA clecision-
making, they will too often place the burden of proof (for demonstrating a significant effect) on those
who want to institute some protective action (usually the FWS or petitioners for listing of a species),
without taking into account the practical consequences of falsely conclu(ling that no effect is occurring.
This could lead to a systematic bias against species that are candidates for listing or for listed species in
need of protective actions.
OCR for page 133
Making ESA Decisions in the Face of Uncertainty
Cumulative Effects/Thresholds
133
One situation where this type of problem could arise is when those concerned with species
protection suspect that there may be a critical threshold of effect (above which the risk of extinction
may increase dramatically) or where cumulative effects might push a species past such a threshoici. We
have indicated above that our technical ability to predict such threshoIcis is very limited. If the burden
of proof is on those who must show that such a threshold exists (anct where it is and just what increase
in risk of extinction will occur), there will be few instances in which such a threshold can be
clemonstrated. As a solution to this problem, we are not advocating that such thresholds simply be
assumed unless proven otherwise (which would reverse the burden of proof), but rather that the
consequences of each type of error (failing to identify a threshold when one actually exists versus
assuming a threshold when one floes not exist) be examined to design a decision-making procedure that
properly controls the risk of errors, from the point of view of species protection and from the point of
view of avoiding unneeded constraints on other interests. In other words, it is advantageous to make
the assumptions and their pre~iicted consequences explicit.
Listing Decisions
Another area where we are concerned about asymmetric risk functions for endangered species is
in decisions to list them. Lack of information can work against species at risk at the listing stage.
Uncler current conditions, FWS resources for evaluating information on candidate species and for
gathering additional information to make a decision are severely limited. The solution is not simply to
reverse the burden of proof ant] confer protection on all species proposed for listing but to consider
explicitly the consequences of both types of error: of listing species that are not really en(langered ant!
of failing to list those that really are. Listing decisions are not one of the points in ESA decision
making where socioeconomic consequences are to be weighed against species protection, so these need
not be part of the equation for determining where the burden of proof should lie for a particular case.
However, we acknowledge that with limiter! time and money for reviewing the eligibility of species for
listing, only those species whose situations are known to be the most desperate will receive priority.
-r ~ -a ---- - ~ - ~ ~ - -a r
Reducing Asymmetry of Risk for Listed Species
In addition to concerns about risks for species at the point of listing, we are also concerned] that
similar asymmetry of risk functions can occur during (recisions regarding protection of already listed
species. The usual way of clecicling whether there is likely to be a harmful effect is to pose the null
hypothesis of no harm and set a low (usually 0.05) rate of error for falsely conclucling that there will be
harm. This way of framing the question, in combination with limited information on the effects of
habitat alteration on listec} species, is more likely deny neecleci protection than to afford unneecled
protection. If the burden of proof were to show that an action would not harm a species rather than to
show that it would harm a species, increased protection would result. The importance of shifting the
burden of proof this way has been wiclely recognized, especially in the context of marine conservation
issues, and is known as "the precautionary principle" (Cameron and Werksman, 1991; Porter and
Brown, 1991; EarIl, 1992; Norse, 1993~. This principle has aireacly been endorsed in several
international legal documents (Porter ant! Brown, 19911. Recently, the National Marine Fisheries
Service (NMFS) explicitly took this kind of asymmetry and the potential for irreversibility into account
in deciding to list the anactromous Snake River sockeye salmon (Oncorhynchus nerka) as endangered in
OCR for page 134
134
Science and the Endangered Species Act
the face of uncertainty (Waples, in press). The uncertainty concerned whether the anaciromous form,
which spawns in Redfish Lake, Idaho, was genetically identical to the landIocked form of O. nerka, the
kokanee, which is common in RecIfish Lake. The decision to listi was based in part on "the recognition
that the consequences of taking the alternate course (i.e., assuming that recent anadromous [fish] in
Redfish Lake were derived from kokanee) ant! being wrong were irreversible, since the original
sockeye salmon gene pool could easily become extinct before the mistake was realizeci."
Making Tradeoffs Among Competing Objectives
There are always too few resources for the size of the job, and the government has been
criticized for allocating funds for species protection without regard for its own stated priorities (Mann
and Plummer, 1992~. Most people recognize that not everything can be done at once, but these
limitations probably wouIct be more acceptable if there were a clearer connection between objectives
being pursued and actions taken. Although the language of the ESA suggests that the standards for
making decisions about listing, jeoparcly, etc., are to be purely scientific, analyses of ESA
implementation (e.g., Yaffee, 1982) show clearly that tradeoffs among conflicting objectives must be
made in almost every instance. In a few cases, these conDicting objectives ant! the necessity for
balancing them are macie explicit in the act anti its implementing regulations. For example, in treating
critical habitat, the act (Section 4(b)~2~) recognizes that designating critical habitat might have
socioeconomic costs and directs those implementing the act to weigh benefits to the listed species
against these costs. Similarly, the exemption process specifically directs the exemption committee to
ask whether there is an overriding benefit to society from the proposed project that would justify its
approval, despite its threat to listed species.
In most cases, however, tradeoffs among competing objectives arise in the course of
implementing the act with too few resources, whether financial, human, or natural. For example, there
are almost 4,000 candidates for listing (Categories 1 and 2), not all of which can be acted on at once.
Those responsible for listing decisions must decide which to consider first ant! which to delay, based on
their best judgments about immeciiacy of threat, distinctiveness of the taxonomic group, etc. Again,
structured clecision-making techniques can be helpful when deciding on the best use of limited
resources. Thibodeau's (1983) application of decision analysis to choose sites for a recovery program
is an early application of a structured problem-solving technique to allocate scarce financial resources
for conservation. Maguire and Lacy (1990) used decision analysis to help allocate limited zoo space
among tiger subspecies in need of captive conservation. To make clear choices among competing
objectives and to justify those choices to interested publics, it would be helpful to follow a more explicit
framework for evaluating tradeoffs, such as that included in the repertoire of multiattribute decision
analysis (Keeney and Raiffa, 1976; Keeney, 1992), which has been applied to several
endangered-species problems (Maguire, 1986; Maguire and Servheen, 1992; RalIs and Starfield, 1995~.
Among the criteria appropriate for setting priorities among species (or other units of protection),
all of which qualify for listing or recovery, are those reflecting how protection of a particular taxon
might contribute to maintenance of biological diversity (see Chapter 3~. Distinctiveness of various sorts
is one measure of this contribution:
~Fed. Reg. 56: 58619, 20 November 1991. Information obtained after the decision indicated that the
anadromous form was indeed genetically different from the landlocked form (Waples, in press).
OCR for page 135
Making ESA Decisions in the Face of Uncertainty
Genetic. Does the taxon contain genetic material not represented! elsewhere that couIc! provide
raw material valuable for adaptation and evolution in future environments?
Phylogenetic. Does the taxon represent a branch of the evolutionary tree of life that has few or
no other living members?
Ecological. Does the taxon exhibit an unusual adaptation to its environment, particularly to a
rare habitat type (such as vernal pools or hot springs). or does it participate in an unusually close
. . ..
r - ---c7- ~ ~ ~ ~ r ~~~ -r ~~ ~ - --- ~~-- ~~--~~- -~~~~~~
interclependence with other rare (i.e., threatened, endangered, or candidate) species (such as obligate
mutualists or parasites), or does it have critical functional roles (i.e., is it a "keystone" species)?
In addition to measures of distinctiveness, other considerations in setting priorities among units
of protection include the degree to which conserving that taxon would enhance protection of overall
diversity. Higher priority would thus be given to taxa with unusually high levels of genetic diversity,
to ecosystems with high levels of endemism, and to taxa whose clemise would be likely to precipitate
further extinctions of taxa dependent on them. Finally, there are species often referred to as "umbrella
species," i.e., they are species whose protection entails the protection of habitats and ecosystems that
wouic! confer protection on other (endangered) species. Clearly, if priorities have to be set, an
umbrella species shouIc} receive a high priority.
FWS (FWS, 1983) and NMFS (NOAA, 1990) have developed hierarchical systems for
determining listing, delisting and reclassification, and species-recovery priorities; the NMFS systems
are simplified versions of the FWS systems. The FWS system for listing priorities considers the
magnitude of the threat to the species, the immediacy of the threat to the species, and the
ctistinctiveness of the species based on its taxonomy, e.g., a monotypic genus is given a higher priority
than a species, which in turn is given a higher priority than a subspecies. The system for setting
recovery priorities also uses taxonomic level as an indicator of distinctiveness. However, as explained
in the chapter on species concepts, taxonomic ranking does not necessarily reflect the same degree of
phylogenetic distinctiveness among all groups of organisms.
135
Resolving Conflicts Among Interest Groups
Parties with a stake in the outcome of ESA decisions include conservationists; developers; other
private and public industries; private individuals; acaclemics; local, state, and fecleral agencies; tribes;
and others. Public participation in ESA decisions is a part of the legislation and implementing
regulations in the form of opportunities to petition for listing of species and to offer comments on
proposed actions. Negotiated agreements among FWS and interested parties are a common part of
ESA implementation through Section 7 provisions for consultation among federal agencies and Section
10 provisions for development of habitat conservation plans by private developers. In Section 7
consultations. FWS is charged with issuing an opinion on whether a Proposed action by a federal
7 4~ ~ 4- ~ ~
agency is likely to jeopardize the continued existence of a listed species. In practice, the proposed
actions are almost always negotiated informally, with comments by FWS and changes in plans by the
developing agency, such that by the time the formal consultation occurs, jeopardy opinions are
extremely rare (Houck, 19931. Although the results may be criticized, the process of negotiating
agreement between federal agencies is entirely consistent with directives for constructive cooperation
among federal agencies (Cleland, 1991; Wondolleck et al., 19941.
The process of developing habitat conservation plans under Section 10 (Chapter 4; Bean et al.,
1991) is also a collaborative, negotiated process between FWS and individuals seeking the incidental
take permit. Again, the results of this process often have been criticized.
OCR for page 136
136
Science and the Endangered Species Act
A combination of methods from decision analysis and dispute resolution can offer disputing
parties a way out of the dilemma of how to combine the best scientific analysis with attention to the
conflicting values that often are involved in making controversial decisions regarding enciangerec!
species. Dispute resolution brings multiple parties together to develop a plan that meets the essential
needs of all (Fisher et al., 19911. Decision analysis can help facilitate this process by providing a
structure for representing the values ant! the scientific beliefs that inform each party's positions
(Maguire and Boiney, 19941. Decision analysis and dispute resolution direct attention to the objectives
and underlying interests that a decision is supposed to reflect. The parts of decision analysis that focus
on objectives, priorities, tradeoffs, and criteria can help with this values-structuring part of the analysis
(Keeney, 19921. One of the tasks of a collaborative problem-solving process is disentangling
disagreements about facts from differences in values. This process is difficult at best, and even hauler
when there is substantial scientific uncertainty, as is so often the case in ESA decisions. Decision
analysis separates each party's view of the facts of the matter from that party's value structure, which is
helpful in identifying what each party believes in terms of facts and values. Some of the benefits of
decision analysis and dispute resolution can be realizer! with a qualitative analysis. In other cases,
quantitative analysis, such as sensitivity analysis or value of information analysis, can help promote
agreement by showing where additional information might help develop a consensus plan and how
much additional information might be worth acquiring (Maguire and Boiney, 19941. Such procedures
can help ensure that negotiations un(ler Section 7 and Section 10 do a good job of incorporating both
science and values.
Implementing Structured Approaches in the Agencies
A call for a more structured approach to ESA decisions using tools such as decision analysis is
not necessarily a call for more extensive analysis or research, neither of which could be supported with
current resources. Rather, it is a way of making better use of available information to address the
problems at hand. Many of the benefits of structured analysis can be realized with a relatively quick
and qualitative application of decision and risk concepts (Behn and Vaupel, 1982~. Those concepts
include explicitly identifying objectives, setting priorities among objectives, and establishing criteria for
measuring progress toward those objectives; clearly weighing tradeoffs among conflicting objectives,
whether these tradeoffs are forcer! by the language of the ESA or by limited resources for implementing
it; and explicitly considering probabilistic information, making better use of expert opinion, anti
providing more coherent ways of combining data ant! opinion to estimate probabilities. All of these can
help provide a better connection between the values being pursued uncler the ESA and the scientific
information available to support decision making. In the long run, the decisions will improve, anti they
will be better justifier! with reference to both values and science.
It wouicl be possible to provide many of the managers responsible for administering the ESA with
the tools to conduct qualitative decision analyses and viability analyses themselves, augmenting the
informal methods of analysis they use now. Several short courses ant! seminars for the Fish and
WiTcllife Service, U.S. Forest Service, Bureau of Land Management, and state wildlife agencies, all of
which have ESA responsibilities, have included decision analysis and population viability analysis as
tools for endangered species management.
In some cases, more thorough, quantitative analyses will be needed, with consultation from a
decision analyst or population biologist. That input could be handled in the same way as subject matter
input from species experts when preparing listing documents and recovery plans. Quantitative analysis,
such as sensitivity analysis or analysis of the value of information (Maguire, 1986; 1987; Maguire and
OCR for page 137
Making ESA Decisions in the Face of Uncertainty
Servheen, 1992; Rails and Starfield, 1995), can help build confidence in the robustness of a particular
course of action or direct further research to critical parts of a problem.
CONCLUSIONS AND RECOMMENDATIONS
· Major advances in both theory anct methods of estimating risk of extinction allow us to base
listing and recovery decisions on scientific principles.
- Many previous ESA decisions did not meet the guidelines suggested by current scientific
thinking, listing species as endangered only after populations tract dropped to the point where the
risk of extinction was very high ant} proposing recovery goals that left the species still at high
risk of extinction.
- Where natural history anti demographic data are available, analytical and simulation moclels
can be used to provide quantitative estimates of risks of extinction.
- General results from these extinction moclels have been user! to develop rating systems based
on objective criteria (such as population size, number of populations, and other demographic and
environmental characteristics) to categorize species according to relative risk of extinction.
Rating systems for use in situations where detailed data are not available should be developed
and tested with simulation and observational methods.
- Because current extinction models clo not consider the interactions of all factors promoting
extinction, estimates of extinction risk may underestimate the true risk.
· Setting levels of risk to trigger listing and recovery decisions entails scientific and public
policy considerations.
137
We can find no scientific basis for setting different levels of risk for different taxonomic
groups, such as plants and animals, or for public versus private actions that might affect listed
species. However, it is critical to understand that achieving the same biological risks for listed
species might well entail different management policies on public and private lancIs, because
public and private entitities behave differently from each other. No implementation of the
Endangered Species Act can be fully successful without recognition of these differences.
- To the degree that they can be quantified, the levels of risk associated with endangered status
should be higher than those for threatened status. Once a species no longer qualifies as
threatened, it should be considered recovered and clelisted.
- Levels of risk to trigger ESA decisions must be framed as a probability of extinction during a
specified period (e.g., x% probability of extinction over the next y years). Although some crises
may call for short time horizons (on the order of tens of years), ordinarily it will be necessary to
view extinction over longer periods (on the order of huncirecis of years) so that short-term
solutions do not create long-term problems.
- The selection of particular degrees of risk associated with particular periods to trigger ESA
decisions reflects scientific knowledge and societal values.
- When implementing the ESA with limited resources, it will probably be necessary to allocate
effort among species, all of which qualify for protection according to the risk level that has been
adopted. Scientific considerations, such as whether a species or its habitat possesses unusually
distinctive attributes or whether protection of a taxon would confer protection on other candidate
taxa or their habitats, should help set priorities for action.
OCR for page 138
138
Science and the Endangered Species Act
· There will always be uncertainty in the estimates of risk user! to trigger decisions under the
ESA, requiring policies and processes for making decisions with incomplete and uncertain data.
Making decisions uncler uncertainty poses the possibility of errors of various types, such as delisting
when a species has not actually recovered or listing when a species is not really endangered or
threatened. For a variety of statistical reasons, including those pertaining to availability of data,
protection would be more likely if the burden were to show that a proposed action would not harm a
listed species rather than to show that it would.
· Because ESA decisions are often clifficult and controversial, the procedures used to make
them should be explicit and well documented. Structured methods can improve the substance of these
decisions and the justification for them. Structured methods can be particularly appropriate to ESA
decisions when:
- scientific risk assessments anti societal values must be integrated;
- tradeoffs among conflicting objectives must be made or negotiations among disputing parties
must be conducted;
- the costs and benefits of delaying decisions while gathering additional information to reduce
uncertainty must be weighed; ant!
when empirical data are lacking anti information clerived from expert opinion should be used.
REFERENCES
Ashford, W. 1995. The Economy of Nature. Houghton-Mifflin, Boston.
Barlow, I., I. Sisson, and S. B. Reilly. 1993. Status of California cetacean stocks: a summary of the
work shop held on March 31 to April 12 1993. National Marine Fisheries Service, Southwest
Fisheries Science Center, Administrative Report LI-93-20.
Bart, I., and D. Robson. 1994. Design of a monitoring program for northern spotted owls. In Ralph,
_ _ _
C.~., I.R., Sauer, and S. Droege, technical coordinators. Proceedings of the Symposium on
Monitoring Bird Population Trends by Point Counts. General Technical Report PSW - 000.
Albany, CA: Pacific Southwest Research Station, Forest Service, USDA. In press.
Bean, M. I., S. G. Fitzgerald, and M. A. O'Connell. 1991. Reconciling Conflicts Under the
Endangered Species Act: The Habitat Conservation Planning Experience. World Wildlife Fund,
Washington, DC.
BeanIands, G. E., W. J. Erckmann, G. H. Orians, J. O'Riorcian, D. Policansky, M. H. Sadar, ant] B.
Sadler. 1986. Cumulative Environmental Effects: A Binational Perspective. Canadian
Environmental Assessment Council, Ottawa, Ontario; ant} the U.S. National Research Council,
Washington, D.C.
Behn, R. D. and I. W. Vaupel. 1982. Quick Analysis for Busy Decision Makers. Basic Books, New
York, NY.
Brown, G. M., Jr. i990. Valuation of Genetic Resources. Pp 203-229 in G. H. Orians, G. M.
Brown, Ir., W. E. Kunin, and J. E. Swierbinski, eds. The Preservation ant! Valuation of
Biological Resources. University of Washington Press, Seattle.
Brownell, R. L. Jr., K. Rails, and W. F. Perrin. 1989. The plight of the "forgotten" whales.
Oceanus 32~1~:5-11.
Cameron, I. and I. D. Werksman. 1991. The Precautionary Principle: A Policy for Action in the
OCR for page 139
Making ESA Decisions in the Face of Uncertainty
Face of Uncertainty. Centre for International Environmental Law, London.
Clelanci, I.C. 1991. Application of alternative dispute resolution to Endangered Species Act
139
interagency consultations. Master of Environmental Management Project, School of Forestry
ant} Environmental Studies, Duke University, Durham, NC.
Clemen, R.T. 1991. Making Harc!Decisions. AnIntrocluctionto Decision Analysis. PWS - Kent,
Boston MA.
Donovan, G. P. ed. 1989. The Comprehensive Assessement of Whale Stocks: The Early Years. The
International Whaling Commission, Cambridge, U. K.
EarIl, R. C. 1992. Commonsense ant! the precautionary principle-an environmentalist's perspective.
Marine Pollution Bulletin 24: 182-~86.
Fisher, R., W. Ury, and B. Patton. 1991. Getting to Yes. 2nd. ed. Penguin Books, New York, NY.
FWS. 1983. Endangered and threatened species listing and recovery priority guidelines. Federal
Register 48~184~:43098-43105.
Given, D. R. and D. A. Norton. 1993. A multivariate approach to assessing threat and for priority
setting in threatener! species conservation. Biological Conservation 64:57-66.
Gnam, R. 1993. Comments invited on species' risk. BioScience 43:430.
Goodman, D. 1993. Scientific stanclarcis for endangereci species management. Appendix ~ to Annual
Report No. i, Research on methods of biodiversity management. Environmental Protection
Agency Office of Research and Development, Cooperative Agreement No. CR-8200- 8601.
Gregory, R., S. Lichtenstein, and P. SIovic. 1993. Valuing environmental resources: a constructive
approach. Journal of Risk and Uncertainty 7: 177-197.
Hogarth, R. 1980. Judgment and Choice: The Psychology of Decisions. Wiley, Chichester, UK.
Houck, O. A. 1993. The Enciangerec! Species Act and its implementation by the U.S. Departments of
Interior and Commerce. University of Colorado Law Review 64~2~:277-370.
Kahneman, D., P. SIovic, ant! A. Tversky. 1982. Judgement Under Uncertainty: Heuristics and
Biases. Cambridge University Press, Cambriclge, UK.
Keeney, R. L. 1992. Value-Focusec! Thinking. Harvard University Press, Cambridge, MA.
Keeney, R. L., ant! H. Raiffa. 1976. Decisions with Multiple Objectives: Preferences and Value
Tradeoffs. Wiley, New York, NY.
Lande, R. 1988. Demographic models of the norther spotter] owl gastric occidentalis caurina).
Oecologia 75:601-607.
Mace, G., N. Collar, I. Cooke, K. Gaston, I. Ginsberg, N. Leacler Williams, M. Maunder, and E. I.
MiIner-GulIand. 1992. The development of new criteria for listing species on the IUCN Red
List. Species 19:16-22.
Mace, G. M. and R. Lande. 1991. Assessing extinction threats: toward a reevaluation of IUCN
threatened species categories. Conservation Biology 5: 148-157.
Maguire, L`.A. 1986. Using decision analysis to manage endangered species populations. Journal of
Environmental Management 22: 345-360.
Maguire, L.A. 1987. Decision analysis: A too! for tiger conservation anti management. Pp. 75-486
in Tigers of the World, R.L. Tilson and U.S. Seal, ecis. Noyes Press, Park Ridge, NI.
Maguire, L.A. 1989. Managing black-footed ferret populations under uncertainty: An analysis of
capture anti release decisions. Pp. 268-292 in Conservation Biology and the Black-footed Ferret.
U.S. Seal, M. Bogan, T. Thorne, and S.F. Anderson, eds. Yale University Press, New Haven,
CN.
Maguire, L.A. and L. G. Boiney. 1994. Resolving environmental disputes: a framework
incorporating decision analysis and dispute resolution techniques. Journal of Environmental
Management 42:31-48.
OCR for page 140
140
Science and the Endangered Species Act
Maguire, L.A. and R.C. Lacy. 1990. Allocating scarce resources for conservation of endangered
species: partitioning zoo space for tigers. Conservation Biology 4: 157-166.
Maguire, L.A., and C. Servheen. 1992. Integrating biological and sociological concerns in
endangered species management: Augmentation of grizzly bear populations. Conservation
Biology 6:426-434.
Mann, C.C. and M. L. Plummer. 1992. The butterfly problem. The Atiantic Monthly 269~: 47-70.
Menges, E. S. 1990. Population viability analysis for an endangered plant. Conservation Biology
4:52-62.
Mitchell, R. C., and R. T. Carson. 1988. Using Surveys To Value Public Goods: The Contingent
Valuation Method. Resources for the Future, Washington, D.C.
Molloy, I. and A. Davis. 1992. Setting priorities for the conservation of New Zealand's threatened
plants and animals. Department of Conservation, Wellington, New Zealand. 44 pages.
NOAA. 1990. Enclangereci and threatened species; listing and recovery priority guidelines. Federal
Register 55~16~:24296-24298.
Norse, E. A. (edge. 1993. Global Marine Biological Diversity: A Strategy for Building Conservation
into Decision Making. Island Press, Washington D. C., 383 pp.
Noss, R. F. 1992. Biodiversity: many scales and many concerns. Pp. 17-22 in H. F. Kerner (ed.),
Proceedings of the Symposium on Bioctiversity of Northwestern California, Oct. 28-30, 1991,
Santa Rosa, CA.
NRC (National Research Council). 1992a. The Scientific Bases for the Preservation of the Hawaiian
Crow. National Acaciemy Press, Washington, D.C.
NRC (National Research Council). 1992b. Decline of the Sea Turtles: Causes and Prevention.
National Academy Press, Washington, D.C.
Nunney, L. and K. A. Campbell. 1993. Assessing minimum viable population size: demography
meets population genetics. Trends in Ecology and Evolution S:234-239.
Porter, G. and I. W. Brown. 1991. Global Environmental Politics. Westview Press, Boulder,
Colorado.
Raiffa, H. 1968. Decision analysis: Introductory lectures on choices under uncertainty.
Addison-Wesley, Reading, MA.
RalIs, K., and J. D. Ballou. 1992. Managing genetic diversity in captive breeding and reintroduction
programs. Trans. 57th N. A. WildI. & Nat. Res. Conf. (1992~: 263-282.
Rails, K., ant} A. M. Starfield. 1995. Two decision analysis methods for conservation problems.
Conservation Biology 9: 175- ~ ~ ~ .
Reilly, S. 1992. Population biology and status of eastern Pacific gray whales: recent developments.
Pp. 1062-1074 in D. R. McCullough and R. H. Barrett, eds. Wildlife 2001: Populations.
Elsevier Applied Science, London.
Roberts, T. 1994. Mitigation: the Judeo-Christian concept of expiating evil deeds by good works or
sacrifice or: just how much is a k-rat worth, anyway?/part one of two parts. The Wildlife
Society, Western Section Newsletter 39 (4~:2.
Saaty, TV. 1990. How to make a decisions The Analytic Hierarchy Process. European Journal of
Operational Research 48: 9-26.
Shaffer, M. L. 1981. Minimum viable population sizes for species conservation. BioScience 31:131-
i34.
Soule, M. E. 1987. Viable populations for conservation. Cambridge University Press.
Taylor, B.~. and T. Gerroclette. 1993. The uses of statistical power in conservation biology: the
vaquita and northern spotted owl. Conservation Biology 7:489-500.
Tear, T. H., I. M. Scott, P. H. Hayward, and B. Griffith. 1993. Status and prospects for success of
OCR for page 141
Making ESA Decisions in the Face of Uncertainty
the Endangered Species Act: a look at recovery plans. Science 262: 976-977.
Thibodeau, F.R. 1983. Endangered species: deciding which species to save. Environmental
Management 7: 101-107.
Tversky, A., S. Sattah, and P. SIovic. 1988. Contingent weighting in judgement anti choice.
Psychological Review 95:371-384.
vonWinterfelcit, D., anctW. Ec~warcis. 1986. Decision Analysis and Behavioral Research.
Cambridge University Press, Cambridge, UK.
141
Waples, R. S. In press. Evolutionarily Significant Units and the Conservation of Biological Diversity
Under the Endangerec! Species Act. In I. L. Nielsen and D. A. Powers, eds. Evolution and the
Aquatic Ecosystem. American Fisheries Society, Bethesda, Mct.
Whitten, A. I. 1990. Recovery: a proposed programme for Britain's protected species. Nature
Conservancy Council, CSD Report, No. 1089.
Wilcove, D. S., M. McMilIan, anti K. C. Winston. 1993. What exactly is an endangered species? An
analysis of the endangered species list, 1985-1991. Conservation Biology 7:87-93.
Wondolleck, I. M., S. L. Yaffee, and I. E. Crowfoot. 1994. Applying the principles of alternative
dispute resolution to enciangerec! species conservation. In Improving Endangered Species
Conservation: Reviewing the Experience and Learning the Lessons. T. Clark and A. Clarke,
eds. Island Press, Covelo, CA.
Yaffee, S.L. 1982. Prohibitive Policy: Implementing the Federal Endangered Species Act. MIT
Press, Cambridge, MA.
OCR for page 142
Representative terms from entire chapter:
null hypothesis
