Guidelines for Using the IUCN Red List Categories and Criteria Version 15.1 (July 2022) - page 3

 

  Главная      Manuals     Guidelines for Using the IUCN Red List Categories and Criteria Version 15.1 (July 2022)

 

Search            copyright infringement  

 

 

 

 

 

 

 

 

 

 

 

Content      ..     1      2      3      4      ..

 

 

 

Guidelines for Using the IUCN Red List Categories and Criteria Version 15.1 (July 2022) - page 3

 

 

Red List Guidelines
60
4.10.8 Effect of sampling effort and detectability on estimates of AOO
Estimates of AOO may be sensitive to sampling effort, as may estimates of EOO, the number of
locations and the number of subpopulations. Inevitably, a taxon may not have been detected
everywhere that it occurs, either because it has cryptic life forms, short-lived detectable life stages,
is hard to identify (and few capable experts available), or because it occurs in inaccessible or
poorly surveyed regions. For conspicuous taxa occurring in well-sampled areas, it may be
reasonable to assume that most occurrences have been detected and AOO may be estimated by
tallying the area of 2 2 km grid cells in which observation records are located using Equation
4.1. For other taxa that may have many unrecorded occurrences, however, this assumption and the
resulting tally will underestimate AOO.
Underestimation of AOO will affect the outcome of Red List assessments under criterion B2, e.g.
if the estimated AOO is less than, or close to, 2,000 km2, the lower threshold of the VU category.
In such cases, assessors may not be able to justify the assumption that AOO is estimated accurately
from a simple intersection of current records with a standard 2 2 km grid, and an alternative
assumption must be made in support of a more accurate estimate.
Assessors should follow section 3.2 to deal with uncertainty in estimates of AOO for potentially
threatened taxa that have poorly sampled distributions. A plausible lower bound of AOO would
be no smaller than that based on an intersection of current records with a 2 2 km grid, but could
be larger. A plausible upper bound of AOO would be no larger than that based on an intersection
of potential habitat (given that it is well known) with a 2 2 km grid, but will usually be smaller
because the taxon may not occupy all of its suitable habitat. Both of these bounds must be
considered (e.g., by entering a range of values in SIS); assuming that AOO is equal to either the
lower or the upper bound is not consistent with the idea of a precautionary but realistic attitude to
uncertainty. Habitat maps and models may inform plausible estimates of AOO based on guidance
in section 4.10.7.
An important step in the approach outlined in section 4.10.7 is to estimate the proportion of
potential habitat that is occupied at the time of the Red List assessment. This should be based on
explicit assumptions referring to information on survey effort and success, and ecological factors
such as predation, competition, disease, etc. that may limit occupancy within potential habitat.
Assessors should describe this information and explain how it supports their estimate of the
proportion of potential habitat that is occupied by the taxon.
Finally, where the plausible upper and lower bounds of AOO span the full range of categories
from Least Concern to Critically Endangered, the species should be assigned to the Data Deficient
category (section 3.2), unless other criteria apply.
4.10.9 Complementarity of AOO, EOO and number of locations
It should be understood that AOO, EOO and the number of locations are all spatial metrics that
measure different (though sometimes overlapping) aspects of risk-spreading or insurance against
spatially explicit threats. Therefore, all three measures should be estimated and assessed against
the criteria where available data permit. As mentioned in section
4.9, to understand the
relationships between these spatial metrics, it may be helpful to think of species that have similar
values for one of these metrics and different values for the other. Suppose two species with similar
life histories have the same EOO, but different values for AOO, perhaps because one has more
specialized habitat requirements. For example, two species may be distributed across the same
desert (hence EOO is the same), but one is wide ranging throughout (large AOO) while the other
Red List Guidelines
61
is restricted to oases (small AOO). The species with the smaller AOO may have a higher risk of
extinction because threats to its restricted habitat (e.g., degradation of oases) are likely to reduce
its habitat more rapidly to an area that cannot support a viable population. The species with the
smaller AOO is also likely to have a smaller population size than the one with a larger AOO, and
hence is likely to have higher extinction risks for that reason.
4.11 Location (criteria B and D)
“The term ‘location’ defines a geographically or ecologically distinct area in which a single
threatening event can rapidly affect all individuals of the taxon present. The size of the location
depends on the area covered by the threatening event and may include part of one or many
subpopulations. Where a taxon is affected by more than one threatening event, location should
be defined by considering the most serious plausible threat.” (IUCN 2001, 2012b)
In the Red List criteria, “location” refers to a threat-based area, and is different from the general
notions of location and locality. The number of locations, AOO and EOO are metrics that measure
different (though sometimes overlapping) aspects of risk-spreading or insurance against spatially
explicit threats (see section 4.10.9). Fewer locations means that larger parts of a species' range are
subject to the same threat, resulting in less risk-spreading, more correlated (synchronized) declines
due to threats, and therefore greater extinction risk.
Justification for the number of locations used in Red List assessments should consider all areas
whether they are under threat or not (see below), and, for areas that are under threat, should include
reference to the most serious plausible threat(s). For example, where the most serious plausible
threat is habitat loss due to development, a location is an area where a single development project
can rapidly eliminate or severely reduce the population. The time frame should be short (e.g.,
within a single generation or three years, whichever is longer, but not any longer than is possible
to project the threats and their impacts on the species).
When there are several threats, locations should be based on the one that has the maximum product
of probability and consequence (in terms of percentage reduction in population).
Where the most serious plausible threat is habitat loss that occurs gradually and cumulatively via
many small-scale events, such as clearance of small areas for small-holder grazing, a location can
be defined by the area over which the population will be eliminated or severely reduced within a
single generation or three years, whichever is longer. Where the most serious plausible threat is
volcanic eruption, hurricane, tsunami, frequent flood or fire, locations may be defined by the
previous or predicted extent of lava flows, storm paths, inundation, fire paths, etc. Where the most
serious plausible threat is collection or harvest, then locations may be defined based on the size of
jurisdictions (within which similar regulations apply) or on the level of access (e.g., ease with
which collectors may reach different areas), as well as on the factors that determine how the levels
of exploitation change (e.g., if collection intensity in two separate areas changes in response to the
same market trends in demand, these may be counted as a single location).
If two or more subpopulations occur within an area that may be threatened by one such event, they
must be counted as a single location. Conversely, if a single subpopulation covers an area larger
than may be affected by any single event, it must be counted as more than one location.
Red List Guidelines
62
Where the most serious plausible threat does not affect all of the taxon’s distribution, other threats
can be used to define and count locations in those areas not affected by the most serious plausible
threat.
If there are two or more serious plausible threats, the number of locations should be based on the
threat that results in the smallest number of locations.
When parts of the distribution are not affected by any threat, the following options will be
appropriate under different circumstances: (a) number of locations is not used (i.e., the subcriteria
that refer to the number of locations consequently are not met), especially if the unaffected area is
more than half the taxon’s range; (b) number of locations in the unaffected areas is set to the
number of subpopulations in those areas, especially if there are several subpopulations; (c) the
number of locations is based on the smallest size of locations in the currently affected areas; (d)
the number of locations is based on the most likely threat that may affect the currently-unaffected
areas in the future. In any case, the basis of the number of locations should be documented.
In the absence of any plausible threat for the taxon, the term "location" cannot be used and the
subcriteria that refer to the number of locations will not be met.
4.12 Quantitative analysis (criterion E)
“A quantitative analysis is defined here as any form of analysis which estimates the extinction
probability of a taxon based on known life history, habitat requirements, threats and any specified
management options. Population viability analysis (PVA) is one such technique. Quantitative
analyses should make full use of all relevant available data. In a situation in which there is limited
information, such data as are available can be used to provide an estimate of extinction risk (for
instance, estimating the impact of stochastic events on habitat). In presenting the results of
quantitative analyses, the assumptions (which must be appropriate and defensible), the data used
and the uncertainty in the data or quantitative model must be documented.” (IUCN 2001, 2012b)
Quantitative analyses are used for assessing taxa under criterion E. Guidelines for applying
criterion E are discussed in section 9. It is important to note that the risk-based thresholds of
criterion E should not be used to infer an extinction risk for a taxon assessed as VU, EN and CR
under any of the criteria A to D.
5. Guidelines for Applying Criterion A
The A criterion is designed to highlight taxa that have undergone a significant reduction in the
near past, or are projected to experience a significant reduction in the near future. Methods of
calculating reductions are explained in section 4.5.
The rationale for criterion A is that, all other things being equal, the probability of extinction is
greater when the decline rate is high (Mace et al. 2008). The obvious mechanism is that if declines
are not stopped, the population will go extinct, regardless of current population size. Even if a
population is not currently declining, prior declines indicate risk of extinction. One reason is that
if a population responded to a threat with a large decline, a similar decline can happen in the future
in response to a similar threat. Further declines do not have to be immediate (criterion A does not
require continuing decline). Another reason is that having declined to densities far below those
Red List Guidelines
63
at which it existed or evolved with, the species may be vulnerable to new threats or other changes
in its environment, even if the population is not currently declining (see section 5.5 for examples).
Criterion A is based only on population reduction. The reason the IUCN criteria (except for E)
consider symptoms of endangerment (such as decline, small population, restricted distribution,
fragmentation, etc.) singly or a few in combination, instead of altogether, is that in the vast
majority of cases reliable data on all of them do not exist for the same species. For example,
although decline rates can be estimated based on an index of abundance (e.g., CPUE) and are
relatively common, unbiased estimates of population size are rare, esp. for abundant species.
Another reason criterion A considers only reduction is that, when a population is declining with a
substantial rate, extinction risk is more sensitive to the rate of decline than to the population size
(Lande et al. 2003). Finally, there are many examples of abundant species that have become
extinct or nearly extinct. Such species could have been identified as threatened only by a criterion
based only on declines (Stanton 2014). So, from both practical and theoretical points of view, it
is necessary to have a criterion based only on rate of decline, in addition to one (criterion C) that
is based on both population size and rate of decline.
Reductions under criterion A are considered over 10 years or three generations (whichever is
longer, but up to a maximum of 100 years for future reductions). Scaling reductions with
generation length is necessary because species with longer generation length recover more slowly
from declines, although they may decline just as fast (the rate of population increase is limited by
biological constraints whereas rates of human-induced declines are not). Therefore, the same
annual rate of decline would put a longer-lived species at a higher risk of extinction. Scaling with
generation length corrects this disparity.
Reductions for criterion A are calculated over 3 generations, because 1- or 2-generation reductions
can be difficult to distinguish from fluctuations. Although the 3-generation requirement makes
calculation of reduction challenging for long-lived species, it is essential for avoiding the
underestimation of the extinction risk of these species. Ideally, reductions would be calculated
from data that span 3 or more generations, but incomplete data or data from shorter time series
can be used to calculate the 3-generation reduction (see section 4.5.1).
The criterion is split into the criteria A1, A2, A3 and A4.
Criterion A1 deals with reductions in the past 10 years or three generations (whichever is
longer) and is applicable to taxa in which the reduction is clearly reversible AND its causes
are understood AND have ceased (see discussion below), based on (and specifying) any
of (a) to (e), as discussed above.
Criterion A2 also deals with reductions in the past 10 years or three generations (whichever
is longer) but for taxa where the reduction may not be reversible OR its causes may not
have ceased OR may not be understood, based on (and specifying) any of (a) to (e) under
A1.
Criterion A3 deals with population reductions projected, inferred or suspected to be met
in the future 10 years or three generations (whichever is longer, but up to a maximum of
100 years), based on (and specifying) any of (b) to (e) under A1.
Criterion A4 deals with reductions observed, estimated, inferred, projected or suspected
over any 10 year or three-generation time period (up to a maximum of 100 years into the
future), where the time period must include both the past and the future, and where the
Red List Guidelines
64
reduction may not be reversible, OR its causes may not have ceased OR may not be
understood, based on (and specifying) any of (a) to (e) under A1.
Under criterion A, a specific quantitative threshold indicating the population reduction must be
met to qualify for one of the categories of threat. Under criterion A1, these thresholds are 90%
(CR), 70% (EN) and 50% (VU). Under criteria A2, A3 and A4, these thresholds are 80% (CR),
50% (EN) and 30% (VU). These different rates reflect the understanding that taxa in which the
reduction is clearly reversible AND its causes are understood AND have ceased are less at risk
from extinction than those where the reduction may not be reversible OR its causes may not have
ceased OR may not be understood. In order to use A1, three conditions must be met. (1) The
reduction must be reversible. For example, the population size must not be so low that factors
such as Allee effects make it impossible or unlikely to recover. It is the condition that must be
reversible, not the cause of the deteriorated state. For example, loss of habitat may be irreversible
even if the action that caused the loss has ceased. In contrast, a reduction in a forest-dependent
species caused by logging could be considered reversible if changed management practices are
leading to recovery of this species. (2) The causes of the reduction (the threatening factors) must
be identified and their actions must be understood. Thus, it is not sufficient to simply list the
threatening factors; it is also necessary to understand the scale and mechanism of their action (e.g.,
the magnitude and spatial distribution of overfishing, or the relationship between pollution and
the population reduction). (3) The threatening factors must have ceased (e.g., overfishing has
stopped). Examples of taxa that might qualify under criterion A1 are fish species that have
suffered declines under exploitation but where the cause of reduction (e.g., over-exploitation) has
ceased. This criterion may also be applicable to situations where the population is still being
exploited, at lower levels of exploitation that do not cause additional population reductions. If
any of the three conditions (reversible and understood and ceased) are not met in a substantial
portion of the taxon's population (10% or more), then A2 should be used instead of A1.
5.1 The basis of reductions
Listing a taxon under criterion A requires specifying whether the reduction is based on (a) direct
observation (A1, A2 and A4 only), (b) an index of abundance appropriate to the taxon, (c) a decline
in area of occupancy, extent of occurrence and/or quality of habitat, (d) actual or potential levels
of exploitation, and/or (e) the effects of introduced taxa, hybridization, pathogens, pollutants,
competitors or parasites.
The difference between direct observation (a) and index of abundance (b), as well as the value of
distinguishing between them, lies in the assumptions to be met to provide valid estimates of
population size. While “direct observation” requires only statistical assumptions (e.g., random
sampling), indices of abundance require assumptions related to the biology of the species. For
example, for a marine turtle species, use of “nesting females” to examine population change
assumes that the proportion of mature individuals that breeds each year, and the number of visits
to breeding sites per female per year are reasonably constant (or at least vary randomly) among
years. If these assumptions are true, then “nesting females” is an appropriate index of mature
individuals.
Direct observation (a) is the most relevant measure and, all things being equal, should be preferred.
However, other measures may be used if they result in more reliable or more consistent (i.e.,
covering the three-generation period more comprehensively) estimates of population size through
time; for example, for species that are difficult to detect, direct counts may entail large sampling
errors and be biased (i.e., systematically under or overestimate the change in population size).
Red List Guidelines
65
Alternatively, an index based on easily detectable traces (e.g., tracks, droppings, etc.) or resources
that the taxon depends on exclusively may provide more reliable estimates of population
reduction. Similarly, for a species that is censused very infrequently, or responds to habitat loss
with a time lag, habitat change may be a more comprehensive estimate of reduction than direct
observation (see section 5.8 on the relationship between habitat change and population change).
All applicable bases for reduction should be listed. Even if the reduction is calculated based on
the best available data, for example, from direct observation, if others (such as decline in area of
occupancy) are also observed, estimated, inferred or suspected, these should also be specified.
The IUCN criteria use the terms "observed", "estimated", "projected", "inferred", and "suspected"
to detail the nature of the evidence (including aspects of data quality) used for specific criteria. It
is important to note that, for a given data source, not all combinations can form the basis for use
of criterion A (Table 5.1). Examples below detail the types of data that might be used to calculate
population reduction for criterion A.
Table 5.1. The relationship between the nature of evidence (data qualifiers) and the basis of reduction for
criterion A.
Basis of reduction for criterion A:
a
b
c
d
e
introduced taxa,
actual or
hybridization,
index of
AOO,
potential
pathogens,
Nature of evidence
abundance
EOO,
exploitation
pollutants,
(see section 3.1 for
direct
(e.g.
habitat
(e.g. landings,
competitors,
detailed information)
observation
CPUE)
quality
road kill)
parasites
observed (all counted -
A1, A2, A4
n.a.
n.a.
n.a.
n.a.
census)
estimated (statistical
A1, A2, A4
A1, A2, A4
n.a.
n.a.
n.a.
assumptions)
projected (extrapolated
A4
A3, A4
n.a.
n.a.
n.a.
into future)
inferred (estimated from
A1, A2, A3,
A1, A2, A3,
indirect evidence on
n.a.
n.a.
A1, A2, A3, A4
A4
A4
variables of same type)
suspected (estimated
from indirect evidence
A1, A2,
A1, A2, A3,
n.a.
n.a.
A1, A2, A3, A4
on variables of different
A3, A4
A4
type)
n.a. : not applicable
A population reduction can be observed if the data used to deduce the decline are from a census
in which a direct count of all known individuals of a population is made. This can be used in
criteria A1 or A2. For criterion A4, where the time frame for assessing reductions spans both the
past and present, only the portion of a reduction in the past can be observed. The portion of the
population trend in the future must be under another qualifier (e.g., projected).
A population reduction can be estimated from census data, as above, or from an index of
abundance (e.g., Catch Per Unit Effort, density, number of nesting females; abundance based on
mark-recapture data). Indices of abundance rely on statistical assumptions (e.g. about how the
sampling scheme implemented relates to the number of mature individuals) and/or assumptions
Red List Guidelines
66
related to the biology of the species, i.e. how the index relates to the variable being estimated to
calculate a population reduction (mature individuals).
A population reduction can be projected if it is extrapolated from census data or an index of
abundance, either from the present into the future (criterion A3), or from past and present into the
future (criterion A4). For example, a decline may be estimated for a population over two
generations, and projected for a further generation into the future (criterion A4).
A population reduction can be inferred if it is calculated from indirect evidence of variables of the
same general type. For example, population reduction in number of mature individuals calculated
from a decline in catch data from fisheries, hunting data, or road kill (criterion A2d) could all be
classed as inferred. Inference may also involve extrapolating an observed or estimated reduction
from a known subpopulation to calculate an inferred reduction for another subpopulation of the
same species. For example, an observed decline in population size from a forest fragment could
be inferred to be the same for a subpopulation in a similar sized fragment that has not been
censused, but which is perceived to be under the same threats. Inference may also be made from
decline in EOO, or based on a reduction in habitat quality or extent. In this case we might expect
the number of mature individuals of a habitat specialist species to have a closer association to the
reduction in habitat extent than a non-habitat specialist.
A population reduction can be suspected if, based on circumstantial evidence, the relationship can
be made based on a factor related to population abundance or distribution. The relevance of the
factor as a proxy for number of mature individuals must be reasonably supported. Records of
traditional ecological knowledge or anecdotal data may, for example, be used to calculate a
suspected reduction over a given time period, if a population used to be seen regularly, but is now
rarely observed.
5.2 The use of time caps in criterion A
Generation length is used in criterion A as a way of scaling the time frame over which reductions
are measured with the life history of the taxon. Short-lived, faster-reproducing taxa have to suffer
higher annual mortality rates than long-lived, slower-reproducing taxa to meet the same
quantitative threshold (e.g., 80% reduction) over a set time period (e.g., 10 years). To put it another
way, long-lived taxa might be unlikely ever to meet quantitative decline thresholds over a fixed
time period, yet could be facing many years of population decline per recruitment opportunity.
The three-generation time period is used to scale the decline rate threshold for the species’ life
history. This important scalar allows criterion A to be applied to a wide range of taxa. A minimum
time cap of 10 years is specified because, although some taxa will have three-generation periods
of less than 10 years, 10 years is the shortest time period of relevance to conservation planning
and action. A maximum time cap has been introduced for assessments based on projections into
the future, as it is felt that the distant future cannot be predicted with enough certainty to justify
its use as a way of assessing whether a taxon is threatened. A maximum time cap is not applied
to assessments based on past reductions, as it is felt that for long-lived taxa, it is important to use
data for three generations, if it is available.
5.3 How to apply criterion A4
In order to decide whether a taxon can be listed under criterion A4, a “moving-window” reduction
must be calculated. It is not possible to determine whether criterion A4 is applicable only by
looking at the qualitative pattern of the decline, or by calculating only past or only future
reductions.
Red List Guidelines
67
To calculate a “moving window” reduction, first create a time series of past population sizes and
future projections. Then, calculate 3-generation reduction for all time frames that include at least
one past year and at least one future year. The length of all those time frames (windows) must be
three generations or 10 years (whichever is longer), but cannot extend more than 100 years into
the future. Finally, find the maximum of these reductions, which is the number to use in criterion
A4. Whether a taxon is listed under criterion A4 or not, of course, depends on whether it qualifies
under any of the other criteria.
In cases where reliable past data are available only for time periods of less than three generations,
and/or reliable future predictions can only be made for less than three generations into the future,
the 3-generation window to use in criterion A4 can be set as the time period for which reliable
data and predictions are available.
In general, if a taxon is listed under criteria A2 and A3, it will also be listed under criterion A4.
However, this is not always the case, and the category of threat determined using a “moving
window” can exceed that calculated from past and future declines. Therefore, species should
always be evaluated against criterion A4 as well as criteria A2 and A3. For a simple example of
the use of criteria A2, A3 and A4, see the worksheet
“A1-A4” in the spreadsheet
CriterionA_Workbook.xls mentioned in section 4.5.
5.4 Reduction followed by short-term stabilization or increase: The 'ski-
jump' effect
Some widespread, long-lived taxa show very large long-term declines as well as recent increases,
and their population sizes are well above the thresholds for critical population size and distribution
(under criteria B to D). This pattern has been termed the ‘ski-jump’ effect and affects any long-
lived taxa that have declined in the past and are now stable or increasing. The question often
asked is whether the long term historical declines or the more recent increases should take
precedence in the assessment of threat in such taxa. However, the question is misleading; the
IUCN criteria do not allow precedence among the criteria, or emphasizing one criterion over
another. The correct interpretation is to assess the taxon against all the criteria. The point of
criterion A is that long-term trends may indicate an underlying cause whereas recent trends may
be temporary.
When applying criterion A to taxa showing these patterns, a few points should be remembered.
(1) If the reduction is clearly reversible AND its causes are understood AND have ceased then the
higher thresholds of criterion A1 (90% for CR, 70% for EN and 50% for VU) apply, which may
lead to a down-listing of the taxon that would reflect the fact that it is currently stable or increasing.
(2) Uncertainty in the data (particularly long-term historical data) if properly incorporated into the
assessment may affect the outcome of the listing (see section 3.2). (3) If it is projected, inferred
or suspected that populations will decline to the thresholds under criterion A, the taxon can be
listed under criteria A3 or A4.
5.5 Historical reduction followed by long-term stabilization: Severely
depleted populations
Some taxa (particularly marine taxa) show persistence at very low fractions of their unexploited
equilibrium or carrying capacity. The current size of a population relative to historical levels can
be calculated by estimating the reduction from the earliest year for which data are available to the
Red List Guidelines
68
current year (see section 4.5 for methods for estimating reductions). Such estimates, or other
information, may show that a population is severely depleted relative to its unexploited
equilibrium or carrying capacity. In some cases, taxa may be severely depleted, but show no
detectable declines, so they may not qualify under criteria A1 or A2 because their declines
occurred more than three generations ago, and they may be too widespread and abundant to qualify
under any other criteria, reflecting the fact that they do not have a high extinction risk at present.
Nevertheless, they may be more cause for concern because they are more susceptible to unforeseen
catastrophic events and marine taxa may be harvested as bycatch in other fisheries. Such taxa are
not currently being assessed as threatened under the criteria A1 and A2, although they may still
qualify under criteria A3, A4, B, C, D or E.
Taxa in this situation may be assessed under criteria A3 or A4 based on projected or suspected
population declines in the future, provided there is sufficient evidence for the threats faced by the
taxon or the likely decline rate of the taxon to warrant such a listing. These range from biological
or ecological factors (e.g., depensation or sex ratio effect thresholds especially in species adapted
to high population density), to threat and detection factors (e.g. increased economic value
increasing with rarity, technological innovation, or sudden removal of management measures).
Such assessments against criteria A3 or A4 should be undertaken where the status of the species
depends on conservation or management measures that are projected, suspected or inferred to
become less effective over three generation lengths. Specific examples from marine taxa include:
Queen Conch (Strombus gigas) and abalone (Haliotis spp.), which have minimum density
requirements for reproduction (e.g., Hobday et al. 2001, Stoner et al. 2012); Gag (Mycteroperca
microlepis), which may experience sperm limitation under heavy female sex ratio skew (Coleman
et al. 1996); Nassau Grouper (Epinephelus striatus), which experienced a sudden collapse due to
hyperstability or possible depensation (Sadovy and Domeier 2005); Totoaba Croaker (Totoaba
macdonaldi), which underwent intense exploitation after a sudden increase in the value of the
swim bladder (Sadovy and Cheung 2003); and Nassau Grouper in the Bahamas, which underwent
a temporary removal of protection due to an economic downturn (Lam 2009).
The category Near Threatened could also be used if a taxon nearly qualifies as Vulnerable under
criteria A3 or A4. It must be remembered however that the IUCN Red List Criteria are designed
to identify taxa that exhibit symptoms of endangerment, and not simply depletion or conservation
priority. The problem of assessing these taxa is also related to the scaling issues discussed under
the definition of area of occupancy (section 4.10), which affects the application of criterion B. If
an appropriate taxon-specific scaling factor is used, severely depleted marine taxa may qualify as
threatened under criterion B.
5.6 Fisheries
5.6.1 Fisheries management and extinction risk
Taxa that are the targets of fisheries may show a decline in population size due to intentional
management action. Under the Red List Criteria, such taxa could be assigned a threatened status
under criterion A (declining population). Concern has been expressed that such a listing might
not reflect extinction risk, especially if the decline is a consequence of a management plan
designed to achieve a goal such as the maximisation of sustainable yield from a fishery.
It is important to note that criterion A measures declines over the last three generations, not from
the original, unexploited stock. Thus, a well-managed stock should trigger the IUCN Criterion A
thresholds only during the first three generations after the commencement of exploitation. Indeed,
a species that is sustainably fished to achieve, for example, maximum sustainable yield (which
Red List Guidelines
69
could be at a biomass that is ~90% of the original biomass for a shark through to ~30% of the
original biomass for a highly productive tuna) should have a current decline rate of zero. In
addition, fisheries that are being managed sustainably would be assessed against the higher
thresholds of criterion A1 (50% over three generations for VU), making it less likely that they will
be classified as threatened.
There should not be a large number of fish stocks for which there would be a 50% reduction in
population size over the most recent three generations due to commencement of regulated
exploitation. This is because there are few stocks that were close to their unexploited state three
generations ago. Rather, most major fisheries started more than three generations ago (Sethi et al.
2010). Even for these few stocks, a reduction of 50% should last only a few years (perhaps up to
one generation) until the population approaches the target level and the decline rate decreases. If
declines continued, there would be reason for concern; in this case a new assessment, against all
five criteria, may indicate that the taxon is still threatened.
5.6.2 Technical aspects of using criterion A for fisheries
Percentage reductions in the number of mature individuals can be estimated in a number of ways,
including ‘an index of abundance appropriate to the taxon’. In the case of exploited fishes, catch
per unit effort (CPUE) may be used. This measure should be used with caution because changes
in CPUE may underestimate population declines. This may occur, for example, if the population
aggregates even at small sizes so that catches remain high with the same level of effort, even if
the size of the population is declining. It may also occur if increases in fishing efficiency are not
fully taken into account. It is therefore preferable to assess exploited fish taxa using the results of
fishery-independent survey techniques.
Assessments of taxa under criterion A1 need to justify that the threat (e.g., overexploitation) has
ceased and the taxon is being managed sustainably. This can be based on the ratio of the average
level of fishing mortality (F) to the fishing mortality corresponding to maximum sustainable yield
(MSY), i.e., F/FMSY < 1, for the greater of one generation or five years. Other methods could be
used to justify the use of criterion A1 instead of A2. However, care needs to be taken to consider
the chance that unsustainably managed species are incorrectly judged to be sustainable.
5.7 Long-lived taxa
The generation length of some species (e.g., some trees) can exceed 100 years. It is difficult to
estimate population declines from a point in time before which the species populations or even
the species itself may have been recorded. It is important to emphasize the point that the most
significant declines, which are useful to record and which may be possible to reverse, are probably
those that have been caused over the last 100 years.
5.8 Relationship between loss of habitat and population reduction
Under criterion A, a reduction in population size may be based on a decline in area of occupancy,
extent of occurrence and/or quality of habitat. The assumptions made about the relationship
between habitat loss and population reduction have an important effect on the outcome of an
assessment. In particular, the simplest assumption, that the relationship is linear, is not often true
and may lead to over- or under-listing. For example, a bird species may not be reduced by 50% if
50% of its habitat is lost (perhaps because it will colonize new habitats). Or, reduction may happen
mostly in lower-density areas, leading to a faster decline in range than in population size.
Conversely, if reductions occur predominantly in high-density areas, population reduction will be
Red List Guidelines
70
faster than can be deducted from range contraction (decrease in EOO) (Rodríguez 2002).
Similarly, a coral reef fish may be reduced by more than 50% if 50% of its habitat is lost through
fishing with explosives (perhaps because spawning areas have been destroyed).
The sensible use of inference and projection is encouraged when estimating population reductions
from changes in habitat. For example, if a forest species' extent of occurrence has been 70% clear
cut in the last five years it might be justified to suspect a 50% decline in the population over the
past ten years. The species would therefore qualify as Endangered A2c.
In all cases, an understanding of the taxon and its relationship to its habitat, and the threats facing
the habitat is central to making the most appropriate assumptions about habitat loss and subsequent
population reduction. All assumptions about this relationship, and the information used should be
included with the assessment documentation.
Available population data may contradict habitat data (e.g., habitat seems to be declining in
quality, but population numbers are stable). This can occur because: (1) one set of data is
uncertain, biased, or dated, or (2) the population has a lagged response to loss of habitat (likely if
generation time is long). In the first case, the assessors must use their judgement to decide which
data are more certain. If it is decided that the abundance data are adequate to determine trends, the
taxon should be listed under criterion A2. The implications of a possible lagged response in
abundance to loss of habitat should, however, be considered when evaluating the taxon under
criterion A3. For example, if population reduction in the last three generations is 30% based on
abundance data, which are adequate to determine trends, then the species should be listed as VU
A2, even if habitat loss in the same period was 60%. However, if a lagged response in abundance
to loss of habitat is likely (i.e., the impact of habitat loss at present may lead to a future reduction
in the number of mature individuals), then the population may be expected to decline further in
the future (even if habitat loss has stopped), so an EN A3 or EN A4 listing should be considered
as well, if the 60% loss of habitat is suspected to lead to 50% or more reduction in the number of
mature individuals.
6. Guidelines for Applying Criterion B
Criterion B has been designed to identify populations with restricted distributions that are also
severely fragmented or have few locations, undergoing a form of continuing decline, and/or
exhibiting extreme fluctuations (in the present or near future). It is important to pay particular
attention to criterion B, as it is the most commonly misused criterion. To qualify for criterion B,
the general distributional threshold must first be met for one of the categories of threat, either in
terms of extent of occurrence (EOO) or area of occupancy (AOO). The taxon must then meet at
least TWO of the three options listed for criterion B. The options are (a) severely fragmented or
known to exist in no more than x locations, (b) continuing decline, or (c) extreme fluctuation
(Table 2.1). Therefore, if a taxon has met the distributional requirement for the Endangered
category and option (c) extreme fluctuation, but none of the other options, it would not qualify as
Endangered (or Vulnerable) under criterion B. To qualify, it would also have to meet either (a) or
(b). An example of the proper use of criterion B is Endangered: B1ab(v). This means that the
taxon is judged to have an extent of occurrence of less than 5,000 km2, the population is severely
fragmented or known to exist at no more than five locations, and there is a continuing decline in
the number of mature individuals.
Red List Guidelines
71
Subcriterion (a) requires severe fragmentation and/or limited number of locations. The numbering
in the criteria does not allow distinguishing between these two conditions. We recommend that
assessors make this distinction by explicitly specifying in their documentation: (1) whether the
taxon is severely fragmented, and (2) the number of locations.
Some of the problems encountered when applying criterion B are dealt with elsewhere in this
document, i.e. definitions of "subpopulations" (section 4.2), "location" (section 4.11), "continuing
decline" (section 4.6), "extreme fluctuations" (section 4.7), "severely fragmented" (section 4.8),
"extent of occurrence" (section 4.9) and "area of occupancy" (section 4.10). The different types
of information used in criterion B need not be based on the same area at the same time of the year.
For example, for a migratory species, AOO can be based on its distribution during the breeding
season (because, e.g., the taxon occupies a smaller area during breeding, and AOO is the "smallest
area essential at any stage to the survival of existing populations of a taxon"), and the locations
can be based on the distribution and threats in the non-breeding season (because, e.g., they are
"the most serious plausible threats" to the taxon).
7. Guidelines for Applying Criterion C
Criterion C has been designed to identify taxa with small populations that are currently declining
or may decline in the near future. For criterion C, the small population threshold must be met as
well as one of the two subcriteria that describe decline. For example, to qualify for Endangered
under criterion C, the population must be estimated to number less than 2,500 mature individuals,
and to either (1) have an estimated continuing decline in the number of mature individuals of at
least 20% within five years or two generations (whichever is longer, up to a maximum of 100
years) or (2) have a continuing decline in the number of mature individuals and either (a) a
restricted population structure or (b) extreme fluctuations in the number of mature individuals (see
Table 2.1 for details).
Few taxa have data on both population size and decline rates at the necessary resolution to apply
subcriterion C1. There is also some overlap between criteria A and C1, the difference being that
criterion C applies only to small populations, the time frame over which the decline is measured
is shorter (except for the Vulnerable category) and the decline rate thresholds are lower, because
the populations are already small.
Criterion C2a has two subcriteria (i and ii), focusing on seemingly opposite conditions. These
subcriteria take into account the fact that the distribution of a taxon's total population into either
many subpopulations, or a single (or very few) subpopulation(s) could both lead to higher
extinction risk, for different reasons. On the one hand, a taxon that is divided into many
subpopulations may be severely fragmented (as defined in section 4.8), with many of the
subpopulations having a small population size and a very high probability of extinction. On the
other hand, a single subpopulation is like putting all eggs in one basket: a single subpopulation
cannot recover from a local extinction by recolonization, or from a catastrophic decline by the
rescue effect. Which of these is more important depends on subpopulation sizes and other factors.
Criterion C2a covers both of these situations: (i) is for the first case, where even the largest
subpopulation is quite small, and (ii) is for the second case, where almost all or all individuals are
in the same subpopulation. A species that meets the general conditions for criterion C2a (i.e., has
a small, declining population) is likely to be affected by one of these two conditions if they occur.
It may seem that a species with a single subpopulation (or with almost all individuals in the largest
subpopulation) may not have increased risk of extinction, if it also has a wide range. However,
Red List Guidelines
72
this would be true only if the different parts of the range fluctuated and declined independently of
each other. But if this were the case, then these different "parts" would likely not be connected
(otherwise they would be in synchrony), so they should not be considered a single subpopulation.
Thus, in order to apply criterion C2a correctly, it is important to identify subpopulations correctly
(see section 4.2).
Some of the problems encountered when applying criterion C are dealt with elsewhere in this
document, i.e. definitions of "subpopulations" (section 4.2), "mature individuals" (section 4.3),
"continuing decline" (section 4.6), calculation of declines (section 4.5), and "extreme fluctuations"
(section 4.7).
8. Guidelines for Applying Criterion D
This criterion identifies very small or restricted populations. A taxon qualifies for criterion D if
the population of mature individuals (see section 4.3) is smaller than the threshold set for each of
the categories of threat. Under the Vulnerable category there are two options, D1 and D2. A
taxon qualifies for Vulnerable D1 if the population size is estimated to number fewer than 1,000
mature individuals (defined in section 4.3). Criterion D1 is provided for taxa that may not be
declining, but are characterized by an acute restriction in their number of mature individuals,
thereby rendering them particularly susceptible to stochastic events as well as to threats. A taxon
qualifies for Vulnerable D2 if the area of occupancy is very restricted (typically less than 20 km2)
or exists at typically five or fewer locations, and if there is a plausible natural or anthropogenic
threat. Criterion D2 is provided for taxa that may not be declining, but are characterized by an
acute restriction in their area of occupancy or in their number of locations thereby rendering them
particularly susceptible to a plausible threat.
The subcriterion D2 under Vulnerable was intended to be used for taxa with very small
distributions. However, the thresholds for area of occupancy and the number of locations, although
given as indicators (i.e., typically less than 20 km2 or typically five or fewer locations), are
frequently interpreted literally, which is not appropriate. Some people have argued that the
subcriterion is too inclusive and results in massive over-listing, while others argue that it is too
exclusive (e.g., many marine species) and so leads to under-listing. It must be emphasized that
the restricted area of occupancy under criterion D2 is defined such that the population is prone to
the effects of human activities or stochastic events in an uncertain future, and is thus capable of
becoming Critically Endangered or even Extinct in a very short time period (e.g., within one or
two generations—or within three to five years, if this is longer—after the threatening event
occurs). The numerical thresholds are given more by way of example and are not intended to be
interpreted as strict thresholds.
The focus of subcriterion D2 is not the area or the location count (for which many taxa could
qualify), but the risk that the taxon could suddenly become Critically Endangered or Extinct (i.e.,
if the plausible threat is realized, then the species will within a very short time qualify for listing
in one of these categories under, for example, criterion A or B). So, simply meeting the suggested
(or any other) threshold for AOO or number of locations is not sufficient. It is necessary that this
restriction makes the species capable of becoming CR or EX within a very short time, because of
the effects of human activities or stochastic events. There must be a substantial possibility of these
activities or events actually occurring. Thus, unlikely events
(e.g., eruption of an inactive
volcano), non-specific events that were not observed in similar species (e.g., an unspecified
disease epidemic), events unlikely to cause extinction (e.g., because the species has survived many
Red List Guidelines
73
hurricanes, or is likely to adapt to global warming, etc.), or events unlikely to take place rapidly
enough to result in a CR or EX listing in a very short time would not qualify for listing under
criterion D2. The stochastic events or human activities that lead to this listing must be specified
in the justification for listing (see example below). If the taxon is highly restricted, and there are
plausible threats that can cause the species to become VU or EN in a short time, then the taxon
should be considered for listing as NT.
8.1 Taxa known only from the type locality
If a taxon is only known from its type locality and there is no information on its current status or
possible threats, the taxon should be listed as DD. If there are no plausible threats, and the area is
relatively well known, Least Concern is appropriate, unless criteria A, B or C is met. If people
have searched for the taxon, both at the type locality and at a reasonable number of other potential
localities, and no more than 50 mature individuals are estimated, then the taxon would be listed
as Critically Endangered D (the surveys must cover an appropriate time interval for the taxon). If
any significant or plausible threats can be identified, then a full assessment will be necessary to
determine the most appropriate classification (e.g., Critically Endangered under criteria B or C, or
Vulnerable under criterion D2). If, despite searches, the taxon has not been recorded since the
collection of the type specimen, and there are threats in the area, a listing of Critically Endangered
(Possibly Extinct) or Extinct may be appropriate (see section 11 for guidance on how to make this
determination).
8.2 Example of applying criterion D
A very rare bird species is described from two female specimens collected in 1851 and an
observation in 1905 on a single island. The species was thought to be extinct in 1970, however,
islanders reported that it may still exist, and in 1972 three birds were reported by an experienced
bushman. It is thought that this unobtrusive and easily overlooked species may survive in two
separate locations popular with trekkers and bird watchers. Very little is known about this species,
but it is safe to estimate, given the limited sightings many years ago and the likelihood that bird
watchers would have seen it, that the population contains less than 50 mature individuals.
Therefore this species is listed as Critically Endangered: D.
8.3 Example of applying criterion D2
A bird species is confined to only four predator-free islands in close proximity, where it is
common and its populations are considered stable. The historical range of this species was reduced
as the result of the introduction of predators such as cats, rats Rattus spp. and a predatory bird.
Birds attempting to colonize a neighbouring island are killed by cats and the predatory bird. The
accidental introduction of alien species to the predator-free islands could easily cause local
extinction. Thus, the number of locations is estimated as four (because it is unlikely that such
introductions would occur on more than one island at any given time), and the species is classified
as VU under criterion D2.
9. Guidelines for Applying Criterion E
To qualify under the E criterion a quantitative analysis such as a Population Viability Analysis
(PVA) must be conducted to determine a species’ probability of extinction over a given time
period. For example, Critically Endangered E, would mean that the taxon has at least a 50%
probability of going extinct in the wild in the next 10 years or three generations (whichever is
longer).
Red List Guidelines
74
9.1 What is extinction?
Extinction is defined as population size reaching zero. Population size, for the purpose of defining
extinction, is the number of all individuals of the taxon (not only mature individuals). In some
cases, extinction can be defined as population size reaching a number larger than zero. For
example, if only females are modelled, it is prudent to define extinction as one female (instead of
zero) remaining in the population. More generally, an extinction threshold greater than zero is
justified if factors that were not incorporated into the analysis due to a lack of information (for
example, Allee effects, sex structure, genetics, or social interactions) make the predictions of the
analysis at low population sizes unreliable.
For criterion E, extinction risk must be calculated for up to three time periods:
10 years or three generations, whichever is longer (up to a maximum of 100 years)
20 years or five generations, whichever is longer (up to a maximum of 100 years)
100 years
For a taxon with a generation length of 34 years or longer, only one assessment (for 100 years) is
needed. For a taxon with a generation length of 20 to 33 years, two assessments (for three
generations and 100 years) are needed. For a taxon with a generation length less than 20 years, all
three assessments are needed.
9.2 Which method can be used?
One of the commonly used techniques of quantitative analysis is population viability analysis
(PVA), which is a collection of methods for evaluating the threats faced by populations of species,
their risks of extinction or decline, and their chances for recovery, based on species-specific data
and models. For an introduction to PVA, see Boyce (1992), Burgman et al. (1993), Morris and
Doak (2003). Types of models used in a PVA will be discussed below.
In some cases, criterion E can be used without a full PVA, using instead a quantitative analysis
that does not necessarily include demographic information. For example, if a species is restricted
to a small area, it may be possible to estimate the probability of the destruction of its entire
remaining habitat. Such estimations may be based on past weather records, or other information
about trends and locations of past habitat loss. It is important to remember, however, that such
estimates can only be considered as lower bounds on the risk of extinction as it would have been
estimated using a PVA. This is because a PVA incorporates such stochastic effects on habitat as
well as other factors such as demographic variability, and other threats such as direct exploitation.
Whatever the method used, the analysis must be numerical (i.e., a qualitative assessment such as
“high probability of extinction” is not sufficient).
Which method is appropriate depends on the availability of data and the ecology of the taxon. The
model structure should be detailed enough to use all the relevant data, but no more detailed.
Assessments that use all the available and relevant data are more reliable than those that ignore
part of the relevant information. However, including more detail than can be justified by the
quality of the available data may result in increased uncertainty.
If the only available data are presence-absence information from a number of locations, occupancy
models can be used (see Sjögren-Gulve and Hanski 2000, Mackenzie et al. 2017). If census
information from a number of years is available, then a scalar (unstructured; count-based) dynamic
model can be used (see Dennis et al. 1991, Burgman et al. 1993, Morris and Doak 2003). If data
Red List Guidelines
75
are available for various age classes or stages (e.g., juvenile and adult), then a structured model
can be used (see Akçakaya 2000). If detailed data are available at the individual level (for
example, pedigree data), then an individual-based (agent-based) model can be used (see Lacy
2000, DeAngelis and Diaz 2019). If data on the spatial distribution are available, a metapopulation
model or other spatially explicit model should be considered (note that scalar, structured and
individual-based models can all be spatially structured).
The second important consideration in selecting a model is the ecology of the species. The model
structure and assumptions should be realistic with respect to the ecology of the species. The
documentation should list all the assumptions (even the most obvious ones) related to model
structure, parameters and uncertainties. In cases where the available data and the ecology of the
species allow more than one type of model, comparative modelling (e.g., Brook et al. 2000,
Kindvall 2000) and other types of validation
(McCarthy et al. 2001) may strengthen the
conclusions.
9.3 Are there sufficient data?
The types of data that can be used in an assessment include spatial distributions of suitable habitat,
local populations or individuals, patterns of occupancy and extinction in habitat patches, presence-
absence data, habitat relationships, abundance estimates from surveys and censuses, vital rate
(fecundity and survival) estimates from censuses and mark-recapture studies, as well as temporal
variation and spatial covariation in these parameters. Not all of these types of data are required
for any one model. For more information about data needs of particular types of PVA models,
see the references mentioned above.
When there is not sufficient data, or when the available information is too uncertain, it is risky to
make a criterion E assessment with any method, including PVA. In order to decide whether the
available data are sufficient to make a criterion E assessment, we suggest the following procedure.
First, select a model structure based on the discussion in the previous section. Then, estimate the
model parameters (see below), incorporating the uncertainties in the data. A simple way to do
this is to make a best estimate for each parameter, as well as an “optimistic” and a “pessimistic”
estimate. The more uncertain a parameter is, the wider the difference will be between the
“optimistic” and the “pessimistic” estimates. Use these estimates to create a range of models,
which should give a range of extinction risk estimates. The range of these estimates indicates
whether the results are useful (and, hence, whether there is enough data). See also “Incorporating
uncertainty” (section 9.5) below.
Remember that criterion E does not require very specific predictions. Even very uncertain results
may be useful. For example, if the minimum estimate for the risk of extinction in 100 years is
10%, then the taxon is at least Vulnerable, regardless of the most pessimistic predictions. The
criteria also allow incorporating uncertainty in the form of a range of categories presented in the
documentation, while a single category should always be specified in the Red List (see Annex 1
of IUCN 2001, 2012b). So, for example, if the generation length is 10 years, and the extinction
risk is 20-60% in 100 years, 10-30% in 50 years, and 5-10% in 30 years, the taxon could be
classified as (VU-EN) in the documentation, while either has to be chosen for the Red List.
9.4 Model components and parameters
It is very important that model parameters are estimated without bias. However, it is difficult to
provide detailed guidelines on parameter estimation because the components and parameters of a
Red List Guidelines
76
model depend on its structure. Thus, although we provide some general guidelines and specific
examples in this section, these are not comprehensive.
9.4.1 Density dependence
Density dependence is the relationship between demographic parameters (such as survival,
fecundity, population growth rate, etc.) and the size or density of the local population. The
relationship can be negative (also called compensation), with demographic parameters decreasing
as density increases, or it may be positive (also called depensation), with demographic parameters
decreasing as density decreases. The former type of density dependence may result, for instance,
from overcrowding and interspecific competition, and the latter may result from Allee effects,
social structure, and inbreeding depression. Both types of density dependence have important
effects on extinction risks, so models should address both. In other words, whether the model
includes or excludes these types of density dependence, the choice should be justified.
Compensation is especially important to include in cases where habitat loss is a threat.
Depensation can be incorporated by setting an extinction threshold greater than zero (see above).
Because density dependence affects demographic parameters such as survival and fecundity,
estimates of these rates should include description of the population sizes or densities during the
time period when the data for these estimates were obtained.
9.4.2 Temporal variability
Because the criteria are in terms of probabilities, it is essential that all relevant forms of variability
are included in the assessment. Thus, the following types of variability should be considered:
environmental fluctuations (in the form of random changes in one or more model parameters),
demographic stochasticity, expected future trends in the average values of model parameters (e.g.,
as a result of deteriorating habitat), genetic stochasticity, random changes in the sex ratio, and
low-probability, high-impact events (disturbances or catastrophes).
In modelling environmental fluctuations, the estimates of the variances of model parameters
should include only temporal variation; variation due to demographic stochasticity, measurement
error, spatial variation, etc. should be subtracted. For example, if survival rates are based on
census data, binomial variance representing demographic stochasticity can be subtracted from
total observed variance (Akçakaya 2002); if the survival rates are based on a mark-recapture
analysis, methods described by Gould and Nichols (1998) and White et al. (2002), or in the help
file of Program MARK can be used to remove demographic/sampling variance.
If catastrophes are included in the model, only data from non-catastrophe years should be used
when estimating the mean and variance of the model variable (such as survival, fecundity, or
carrying capacity) that the catastrophe affects.
When probabilistic results are based on simulations, the number of replications or iterations
determines the precision of these results. In most cases, the randomly sampled model parameters
are statistically representative if the number of replications is in the 1,000 to 10,000 range.
9.4.3 Spatial variability
If different subpopulations of the taxon are spatially separated or have different demographic rates,
these should be incorporated by making the model spatially explicit. Modelling such a taxon with
a single-population model may underestimate the extinction probability. When multiple
Red List Guidelines
77
populations are included in the model, the correlation among the different populations is an
important factor; ignoring it (i.e., assuming all populations to be independent) may underestimate
the extinction probability.
9.5 Incorporating uncertainty
We suggest that all parameters be specified as ranges (or as distributions) that reflect uncertainties
in the data (lack of knowledge or measurement errors). In addition, uncertainties in the structure
of the model can be incorporated by building multiple models (e.g., with different types of density
dependence). There are various methods of propagating such uncertainties in calculations and
simulations (Ferson et al. 1998). One of the simplest methods is to build best-case and worst-case
models (e.g., Akçakaya and Raphael 1998). A best-case (or optimistic) model includes a
combination of the lower bounds of parameters that have a negative effect on viability (such as
variation in survival rate), and upper bounds of those that have a positive effect (such as average
survival rate). A worst-case or pessimistic model includes the reverse bounds. The results from
these two models can be used as upper and lower bounds on the estimate of extinction risk, which
in turn can be used to specify a range of threat categories (see Annex 1 of IUCN 2001, 2012b).
9.6 Documentation requirements
Any Red List assessment that relies on criterion E should include a document that describes the
quantitative methods used, as well as all the data files that were used in the analysis. The document
and accompanying information should include enough detail to allow a reviewer to reconstruct
the methods used and the results obtained.
The documentation should include a list of assumptions of the analysis, and provide explanations
and justifications for these assumptions. All data used in estimation should be either referenced to
a publication that is available in the public domain, or else be included with the listing
documentation. The uncertainties in the data should be documented.
Methods used in estimating model parameters and in incorporating uncertainties should be
described in detail. Time units used for different model parameters and components should be
consistent; the periods over which parameters are estimated should be specified.
10. Guidelines for Applying the Categories DD, NT and NE
10.1 When to use the category Near Threatened
To qualify for the Near Threatened category, the taxon should be close to qualifying for the
Vulnerable category. The estimates of population size or range size should be close to the
Vulnerable thresholds, especially when there is a high degree of uncertainty, or possibly meet
some of the subcriteria. This may be combined with biological susceptibility and threat.
The category Near Threatened is not specified by its own criteria, but instead by the proximity of
a species to the criteria for the category Vulnerable. One way of determining whether the taxon is
close to qualifying for Vulnerable is to follow the uncertainty guidance given in section 3. If the
range of plausible categories include both LC and VU (or EN), the taxon can be classified as NT,
unless the best estimate is VU (or EN). (If all categories from LC to CR are equally plausible, the
taxon should be classified as DD.)
Red List Guidelines
78
For taxa listed as Near Threatened on the IUCN Red List, assessors are asked to indicate as part
of the justification, which criteria were nearly met. For example, NT listing would be justified in
the following cases (in each case, any criteria not specifically mentioned are not met and are not
nearly met):
Based on data uncertainties, LC and VU are equally plausible categories.
Based on data uncertainties, LC, VU, and EN are equally plausible categories (in this case,
both NT and VU can be considered as the listing category).
Population has declined by an estimated 20-25% in the last three generations.
The taxon meets the area requirements under criterion B for threatened (EOO <20,000 km2
and/or AOO <2,000 km2) and is declining, but the population is not severely fragmented,
occurs at many more than 10 locations, and there are no extreme fluctuations.
The taxon meets the area requirements under criterion B for threatened (EOO <20,000 km2
and/or AOO <2,000 km2) and is severely fragmented, but the population is not declining,
occurs at many more than 10 locations, and there are no extreme fluctuations.
The taxon is declining and occurs at ten locations, but has an EOO of 30,000 km2 and/or
an AOO of 3,000 km2, which are uncertain estimates.
The taxon is declining and severely fragmented, but has an EOO of 30,000 km2 and/or an
AOO of 3,000 km2, which are uncertain estimates.
The taxon is declining and severely fragmented, but has an EOO of 22,000 km2 and/or an
AOO of 2,200 km2, which are highly certain estimates.
Population has declined by an estimated 10% in the last three generations, and is
continuing to decline, and has about 15,000 mature individuals.
The taxon exists in a single subpopulation of about 15,000 individuals and is declining.
The population has about 1,500 mature individuals.
The best estimate of population size is 2,000 mature individuals, but this estimate is very
uncertain, and as low as 1,000 mature individuals cannot be ruled out.
The taxon exists at three sites, occupying an area of 12 km2; the population is being
harvested but is not declining; there are no current threats, but there are plausible events
that may cause the species to decline, but these are unlikely to make the species Extinct or
Critically Endangered in a short time.
Population has declined by 40% in the last three generations, but the decline has stopped,
and the causes of the decline have been understood.
The following are examples of species that should not be listed as NT (or any of the categories of
threat), unless other criteria apply:
Based on data uncertainties, LC is the only plausible category.
Population has declined by an estimated 10% in the last three generations, and there are
more than 20,000 mature individuals.
Population has declined by an estimated 30% as part of fluctuations.
The taxon meets the area requirements under criterion B for CR (EOO <100 km2 and/or
AOO <10 km2), but is not declining, not severely fragmented, there are no extreme
fluctuations, and there are no obvious threats.
The taxon is long-lived and slow growing, but does not meet any criteria A-E.
The population has more than 2,000 mature individuals.
The taxon exists at three sites, occupying an area of 30 km2; the population is not
declining; there are no current threats, and the species is very unlikely to become Extinct
or Critically Endangered in a short time.
Red List Guidelines
79
Previously (prior to 2021), dependence on conservation measures had been used to categorize taxa
as NT that otherwise did not meet NT guidelines (see IUCN 2001; Annex 3). However, taxa in
any Red List Category can be conservation-dependent. Conservation dependence (or conservation
reliance) of taxa is more appropriately assessed as part of an IUCN Green Status of Species
assessment (IUCN 2021), and should no longer be used to assess taxa as NT on the IUCN Red
List.
10.2 Not Evaluated and Data Deficient
Listing in the categories of Not Evaluated (NE) and Data Deficient (DD) indicates that no
estimation of extinction risk has been made, though for different reasons. NE indicates that no
attempt to evaluate the current status of the taxon has been made. DD indicates that the taxon was
evaluated using available data, which were found to be insufficient to place the taxon into a
category. Taxa listed in these categories should not be treated as if they were not threatened.
10.3 When to use Data Deficient
If a taxon is known, but there is no direct or indirect information about its current status or possible
threats, then it is obviously Data Deficient (DD). A Data Deficient listing does not imply that a
taxon is not threatened.
The issue becomes more complex when there is very little information known about a taxon, but
the available information indicates that the taxon may be threatened. The question then becomes
how far is it acceptable to take inference and projection? This is discussed in greater detail in
sections 3.1 and 3.2 (Data availability, inference and projection, and uncertainty).
When data are very uncertain, the category of Data Deficient may be assigned. However, in this
case the assessor must provide documentation showing that this category has been assigned
because data are inadequate to determine a threat category. If the data are so uncertain that both
CR and LC are plausible categories, the taxon can be listed as DD. If plausible categories range
from NT to threatened categories, DD is not the appropriate category; in this case, see section 3.2
about guidance to select the most plausible category while documenting the uncertainty. It is
important to recognize that taxa that are poorly known can often be assigned a threat category on
the basis of background information concerning the deterioration of their habitat and/or other
causal factors; therefore the liberal use of Data Deficient is discouraged.
Data Deficient species may be flagged with one or both of the following tags, although most DD
species would not need either:
1. Unknown provenance. The taxon is known only from one or more specimens with no or
extremely uncertain locality information, so that it is not possible to make any further inference
about its status.
Examples:
A hypothetical hummingbird known from a single trade-skin purchased in the 1900s in Bogotá, and
speculated to have been collected on the East Andes or possibly the Central Andes of Colombia, within
a few hundred kilometres of the capital. However, some "Bogotá" specimens came from as far away
as Ecuador. Since no other specimen is known, it is assumed to be (or have been) a relict species of
restricted range.
Red List Guidelines
80
A hypothetical freshwater fish known only from the type collection made in 1930 one day west of
Andapa which is somewhere along the northeast coast of Madagascar. This species has not been
collected again since the 1930s, largely because its exact type locality is not known. There are therefore
no data available upon which to base statements about the quality of its habitat or the size of its
population, but it is assumed to be (or have been) a relict species of restricted range.
A hypothetical hemi-epiphytic aroid plant is endemic to Ecuador. It is known only from the type
collection made in an unknown locale by a well-known botanist a century ago. The lack of information
prevents any evaluation of the species' conservation status and this is further compounded by
taxonomic problems with many species of the same genus described by the same botanist.
2.
Taxonomic uncertainty explains lack of information. The paucity of data may be a
consequence of taxonomic uncertainty, i.e. the lack of information on distribution, status,
ecology and threats is because there are very few specimens and/or records, and this may be
because the taxon represents aberrant individuals, hybrids, rare colour morphs, or subspecies
of other species. This explanation is as or more likely than the possibility that the taxon is
genuinely rare, threatened or has been inadequately searched for. It is important to note that
this tag should not be used for taxa that simply have uncertainty around their taxonomy. Such
taxa should not be classified as Data Deficient simply because of this uncertainty: they should
either be regarded as good species and assessed against the Red List Criteria, or not assessed
for the Red List. The process of determining the list of taxa to be assessed should be separated
from the process of assessing extinction risk (see section 2.1 on taxonomy).
Examples
A hypothetical island bird species was named relatively recently on the basis of two specimens
collected in the 1930’s in a single location. The specimens are juveniles, and it was speculated that
they may refer to juveniles of a related species, although differences in some morphological features
make this unlikely. Nevertheless, the lack of any further information on distribution, population size,
trends, ecology and threats, mean that the IUCN Red List Criteria cannot be applied, and the species
is consequently classified as Data Deficient.
A hypothetical bird species is known from one specimen collected in northeast Kalimantan in the early
1900’s and another from Sumatra in the 1930’s, plus reports in 1992 in Brunei. It has been speculated
to be of hybrid origin, or a rare morph, although it is possible that it may be a genuinely rare habitat
specialist that is occasionally forced to search other areas for food. With no further information, this
uncertainty makes Data Deficient the most appropriate category.
For further discussion and examples, see Butchart and Bird (2009).
Where a species name is widely accepted as containing multiple taxa that may deserve species-
level recognition (a ‘species complex’) AND there is insufficient information (direct or indirect)
to apply the Red List Categories and Criteria, the ‘species complex’ should be listed as Data
Deficient. If the complexity and uncertainty of the taxonomic status plausibly explains the lack of
information, then the assessment should be tagged as ‘Taxonomic uncertainty explains lack of
information’.
10.4 When not to use Data Deficient
Data Deficient classification implies that the taxon has been assessed against all criteria. All DD
assessments require documentation of available data, sources of uncertainty and justification for
why each of the five criteria cannot be applied (and, if applicable, the tags discussed in the
Red List Guidelines
81
previous section). If all of the five criteria have not been considered, DD cannot be used (the taxon
must be categorized as NE).
In many cases, uncertainty in the data precludes placing the taxon in one of the other categories
(LC to EX). However, not being able to place the taxon into a single category is, by itself, not a
sufficient reason for a DD assessment. As discussed above, if the data are so uncertain that both
CR and LC are plausible categories, the taxon can be listed as DD. If, however, plausible
categories range from NT to threatened categories, DD is not the appropriate category. In this
case, the assessor must select the most plausible category. If it is not possible to identify the most
plausible category, the assessor must select one of the categories, based on their level of risk
tolerance. For example, if LC, NT, and VU are considered to be equally plausible categories, the
taxon may be categorised as NT. In all cases, the justification text must specify all categories that
were considered plausible, as well as the degree of risk tolerance (see section 3.2.3). If assessors
cannot decide on the level of risk tolerance, the mid category should be selected. It is important
to note that, if uncertainty is specified at the parameter level (using the Red List Criteria Calculator
in SIS), then the range of plausible categories and the most plausible category would be
automatically selected, in accordance with the specified level of risk tolerance. See also section
3.2 about guidance to select the most plausible category while documenting the uncertainty;
section 3.1 on data availability, inference and projection, and section 5.8 on inferring population
reduction based on habitat loss.
In some cases, the data uncertainty has a spatial component; for example, there may be some data
from one part of the range, but none or little from the other parts. In such cases, the assessors
should try to avoid a DD listing by considering different plausible assumptions about how
representative the threats are from known areas, and use these assumptions to form uncertainty
intervals for the parameters used (such as mature individuals, locations, subpopulations, etc.).
In other cases, the uncertainty may have a temporal component: the information may be more
uncertain in the more distant past and/or about the more distant future. In such cases, the assessors
should try to avoid a DD listing by using criterion A4 to minimize uncertainty. Considering a 3-
generation window that includes both the more recent past and the more near future would focus
the assessment to a period where data uncertainties are smaller.
11. Guidelines for Applying the Extinct Categories and Tag
11.1 The extinct categories (EX and EW)
The category of Extinct is used when ‘there is no reasonable doubt that the last individual has
died’. However, extinction—the disappearance of the last individual of a species—is very difficult
to detect. Listing of a species as Extinct requires that exhaustive surveys have been undertaken in
all known or likely habitat throughout its historical range, at appropriate times (diurnal, seasonal,
annual) and over a timeframe appropriate to its life cycle and life form. Thus, a key aspect of the
definition of Extinct is "exhaustive surveys" (further guidance on this is in section 11.3).
Listing as Extinct has significant conservation implications, because protective measures and
conservation funding are usually not targeted at species believed to be extinct. Therefore, a species
should not be listed in the Extinct (EX) or Extinct in the Wild (EW) categories if there is any
reasonable possibility that they may still be extant, in order to avoid the ‘Romeo Error’ (Collar
1998), where any protective measures and funding are removed from threatened species in the
mistaken belief that they are already extinct. This term was first applied to the case of Cebu
Red List Guidelines
82
Flowerpecker Dicaeum quadricolor, which was rediscovered in 1992 after 86 years without a
record (Dutson et al. 1993), having been written off as extinct at least 40 years earlier on the
presumption that none of its forest habitat remained on the island of Cebu (Magsalay et al. 1995).
An evidentiary approach to classifying extinctions is appropriate to encourage continuing
conservation efforts until there is no reasonable doubt that the last individual of a species has died.
However, if assessments of EX or EW are too evidentiary, then extinction rates based on the Red
List are likely to be under-estimated. In other words, there are costs to making both types of errors
(listing an extant species as EX and failing to list an extinct species as EX) and benefits of making
the correct listing (Akçakaya et al. 2017). These issues are addressed by:
i) defining a 'Possibly Extinct' tag for species listed as CR but that are likely to be extinct
(section 11.2);
ii) using methods that calculate the probability that the species is extinct, and comparing this
probability to recommended thresholds (section 11.3); and
iii) using the probability that a species is extinct in calculating the number of extinct species
and extinction rates (section 11.4).
It is strongly recommended that the methods and thresholds described in section 11.3 are applied
to any species that has not been recorded since the last assessment, or is suspected to have become
extinct.
Extinct in the Wild is defined as existing only in cultivation, in captivity or as a naturalized
population (or populations) well outside the past range. "Cultivation" and "captivity" are not
necessarily restricted to confinement. To be consistent with the definition of a "wild"
subpopulation (see section 2.1.4 on managed subpopulations), EW should also be used if none of
the subpopulations are wild. Thus, if the only surviving subpopulations of a taxon are not
confined, but are nonetheless subject to intensive, individual-level management interventions as
discussed in section 2.1.4, that taxon should be listed as EW. This category can also be applied
when plant or fungal taxa are represented only by viable propagules (e.g., seeds or spores) in
adequate storage facilities, if effective protocols have been developed for the taxon to ensure there
is the potential for these propagules to develop into viable reproductive offspring and to undertake
species recovery in situ.
11.2 ‘Possibly Extinct’ tags for Critically Endangered taxa
Although an evidentiary approach to classifying extinctions is appropriate, this approach biases
analyses of recent extinctions when based only on those species classified as Extinct or Extinct in
the Wild (when individuals survive only in captivity). For example, the number of recent
extinctions documented on the IUCN Red List is likely to be a significant underestimate, even for
well-known taxa such as birds. The tag of ‘Possibly Extinct’ has therefore been developed to
identify those Critically Endangered species that are, on the balance of evidence, likely to be
extinct, but for which there is a small chance that they may be extant. ‘Possibly Extinct in the
Wild’ correspondingly applies to such species known to survive in cultivation or captivity. Note
that ‘Possibly Extinct’ and ‘Possibly Extinct in the Wild’ are tags, and not Red List Categories.
Relevant types of evidence supporting a listing as Extinct include (Butchart et al. 2006):
for species with recent last records, the decline has been well documented.;
severe threatening processes are known to have occurred (e.g., extensive habitat loss, the
spread of alien invasive predators, intensive hunting, etc.);
Red List Guidelines
83
the species possesses attributes known to predispose taxa to extinction, e.g. flightlessness
(for birds); or
recent surveys have been apparently adequate and appropriate to the species’ detectability,
but have failed to detect the species.
Such evidence should be balanced against the following considerations (Butchart et al. 2006):
recent field work has been inadequate
(any surveys have been insufficiently
intensive/extensive, or inappropriately timed; or the species’ range is inaccessible, remote,
unsafe or inadequately known);
the species is difficult to detect (it is cryptic, inconspicuous, nocturnal, nomadic, silent or
its vocalisations are unknown, identification is difficult, or the species occurs at low
densities);
there have been reasonably convincing recent local reports or unconfirmed sightings; and
suitable habitat (free of introduced predators and pathogens if relevant) remains within the
species’ known range, and/or allospecies or congeners may survive despite similar
threatening processes.
Similar considerations apply when assigning a taxon to either the Extinct in the Wild or Critically
Endangered (Possibly Extinct in the Wild) categories. These considerations are implemented in
the methods for calculating the probability that a species is extinct, and comparing this probability
to recommended thresholds (as discussed in section 11.3). All assessments of taxa that might be
extinct should follow the approach described in sections 11.3 and 11.4.
The documentation for each taxon assessed as Extinct, Extinct in the Wild, Critically Endangered
(Possibly Extinct) and Critically Endangered (Possibly Extinct in the Wild) should explicitly
justify the application of the Extinct categories and ‘Possibly Extinct’ tags. The documentation
must summarize the lines of evidence for and against extinction, describe surveys carried out to
search for the species and specify the date and relevant details of the last confirmed record. A
completed data template (described below) can be used for this purpose. The status of all taxa
assigned ‘Possibly Extinct’ tags should preferably be reviewed at five-year intervals.
There is sometimes difficulty in choosing the correct criteria for species listed as CR(PE) or
CR(PEW). If the species disappeared from known sites within the last ten years or three
generations (whichever is longer), then listing under criterion A2 is the preferable option. If the
species is known from a single location with EOO less than 100 km2 or AOO less than 10 km2,
then listing as CR B1ab(i,ii,v) or B2ab(i,ii,v) are possibilities. However, there are many species
for which extinction is a possibility, but for which the declines or disappearances took place more
than 10 years or three generations ago (whichever is longer), and for which the EOO and AOO
are too large for listing as CR, and/or at least two subcriteria for CR B are not met. In such
instances, the species should be listed as CR C2a(i), CR C2a(ii), and/or CR D, whichever seems
more plausible. Such an assessment therefore implies an estimated population size of fewer than
250 mature individuals (for C2) or 50 mature individuals (for D). Even though it is impossible to
know whether or not such an assumption is correct, it is a reasonable one for a species that could
be Extinct.
11.3 Assigning taxa to EX or CR(PE)
Extinction of a taxon is often difficult to confirm, yet there are costs associated with the wrong
listing (listing an extant taxon as EX, or failing to list an extinct taxon as EX) as well as benefits
Red List Guidelines
84
to making the correct listing. This section describes an approach to making these listings as
consistently as possible, by quantifying how "exhaustive" surveys for the taxon have been, which
is a key aspect of the definition of Extinct. The approach involves two models (the Threats Model
and the Records and Surveys Model) that calculate the probability that a taxon is extinct, P(E),
and comparing this probability to thresholds that were determined based on a cost-benefit
framework (Akçakaya et al. 2017). The following sections describe these two models, their
parameters, and recommendations for interpreting their results. In these sections, everything
mentioned about EX also applies to EW, and everything mentioned about CR(PE) also applies to
CR(PEW).
To use the models described here, download the data template EX_data.xlsx,
the instructions document EX_instructions.pdf, and the R script
RecordsSurveysModel.R,
which
are
available
at
https://www.iucnredlist.org/resources/ex-probability.
Assessors can use other approaches to estimate P(E), as long as the alternative approaches
incorporate the factors and parameters about threats, records, and surveys that are discussed
below, and therefore quantify the extent to which the surveys had been exhaustive.
11.3.1 The Threats Model
The Threats Model (Keith et al. 2017) estimates the probability that the taxon is extinct, P(E),
based on qualitative and, where available, quantitative information about the severity, duration
and scope of threats and their interaction with the life history traits that determine the species'
susceptibility to these threats.
To use this model, estimate two subjective probabilities, based on expert knowledge of the threats
faced by the species:
1. P(local), the probability that the combination of threats affecting the species occurred for
a sufficient duration and was sufficiently severe that they caused local extinction;
2. P(spatial), the probability that the threats occurred over the entire range of the species.
Estimating P(local) requires assessors to draw on the history of the impacts of threats on
populations of the target taxon. A relevant historical observation, for example, would be that the
taxon disappeared from an area shortly after the introduction of an invasive alien predator. It may
also draw on examples where the threats have caused ecologically similar or phylogenetically
related taxa to become extinct at a particular location. Inferences about which taxa are
‘ecologically similar or related’ may be based on life history (e.g., life cycle structure, dependence
on hosts, body size, diet), habitat ecology
(e.g., microhabitat type, breeding sites) and/or
phylogeny.
Estimating P(spatial) requires assessors to evaluate two components: (i) the likelihood that the
threats (with sufficient severity and duration to have caused local extinction) operated throughout
the entire range of the taxon (i.e., distribution of habitat and/or individuals, as appropriate); and
(ii) the certainty with which the range limits are known. Relevant considerations for the first
component include whether the threats operated in such a pattern as to have caused extinction
throughout the taxon's range. This may be influenced by the spatial occurrence of different threats,
dispersal dynamics, migration patterns and patch dynamics, as well as species life-history traits
and cultural factors that influence species susceptibility to threats (see Keith et al. 2017 for further
discussion). Relevant factors to consider for the second component
(range limits) include
Red List Guidelines
85
taxonomic uncertainty, reliability of records and whether potential habitat outside the confirmed
range has been adequately searched. These uncertainties are incorporated into estimates of
P(spatial) by setting upper and lower bounds taking into account plausible maximum and
minimum extents of the taxon’s range.
For both P(local) and P(spatial), estimate a plausible lower bound (minimum), an upper bound
(maximum) and a mid-value (best estimate). See the instructions document for a general guide in
estimating these probabilities.
11.3.2 The Records and Surveys Model
The Records and Surveys Model (Thompson et al. 2017) is an iterative model to estimate the
probability that the taxon is extinct, P(E), based on a time series of records of the taxon, and the
timing, comprehensiveness and adequacy of any targeted surveys designed to detect the taxon
after the last known record. A record is any evidence that the taxon was extant in a given year.
Surveys are dedicated or passive (opportunistic) but unsuccessful efforts to find the taxon (i.e.,
surveys do not result in a record). For each year, enter a maximum of 1 record or 1 survey.
For each record, estimate p(ci), the probability that the taxon is correctly identified as extant. This
probability depends on the type and quality of evidence, similarity of the individual recorded to
taxa with which it could potentially be confused, circumstances of the record and the skill and
experience of the recorder. Before estimating p(ci), it may be helpful to create a default table of
probabilities for each of the common types of records available for the taxa you are assessing, to
act as a guide and to ensure consistency, rather than being prescriptive (see the instructions
document for an example).
For this and any other quantity described in this section, estimate a plausible lower bound
(minimum), a plausible upper bound (maximum) and a mid-value (best estimate).
For each survey, estimate the following three quantities:
(1) ε (epsilon), the proportion of the taxon's habitat within its likely entire range that was
surveyed (or covered by passive surveillance). If there were several dedicated surveys
within a year in different areas of the range, make only one entry, with the total proportion
of the taxon's habitat surveyed across all the surveys. Even when the range of a species is
very uncertain, it may be possible to estimate ε with sufficient certainty to allow estimation
of P(E). If, for instance, the northern and southern range limits of a species' range are
unknown, but the survey intensity is similar regardless of latitude, it would be possible to
estimate ε with a higher certainty than the range itself. The full range of potential plausible
habitat should be considered when estimating ε. For example, outlier records (e.g., at
greater depths, altitude, or drier climates) may give insights into the potential occurrence
of the taxon in a broader range of environments than indicated by the majority of previous
records, particularly if some of these environments are hitherto poorly explored. The
reliability and precision of outlier records should also be considered when estimating ε.
(2) p(r), the probability that the taxon, or recent evidence of it, would have been recorded in
the area that was surveyed, if it were present. This depends on aspects of detectability,
including body size, behaviour (e.g., activity and movement patterns, shyness, tendency
to skulk, phenology, vocalization, sociality), degree of crypsis, local abundance, and
accessibility to or searchability of its habitat and microhabitat.
(3) p(i), the probability that the taxon, or recent evidence of it, could have been reliably
identified in the survey if it had been recorded. This depends on the verifiability of the
Red List Guidelines
86
record; that is, the likelihood that the recorded taxon could be distinguished from a similar
taxon (e.g., a congener) given its distinctiveness
(e.g., in appearance, morphology,
vocalizations, behaviour), and the identification skill of the observers. Assessors must
consider all signs of recent evidence (e.g., scat, spoor, nests, owl pellets, woodpecker bark
peelings, shells, etc.) and all life-stages at the time of the survey; for example, the mature
life-form may be highly distinctive, but the juvenile/seed/ larval/dormant life-stages may
be extremely difficult to distinguish from similar taxa.
For p(r) and p(i), it may be helpful to create a default table of probabilities for taxa with different
characteristics in the taxonomic group you are assessing
(see examples in the instruction
document).
11.3.3 Interpreting the model results
After completing the data entry, follow the instructions in the files mentioned above (the data file,
the instructions document, and the R script). The results of both models will be displayed in a
graph such as Figure 11.1, which includes P(E) estimated by the two models (the square marker),
the bounds of the estimates (the error bars), and lines indicating the thresholds of P(E) for
considering a species CR(PE) or EX (the red lines). The following thresholds are recommended:
CR(PE), if P(E) ≥ 0.5 and <0.9
EX,
if P(E) ≥ 0.9
Figure
11.1. Graphical
display of P(E), the
probability that the species
is extinct, based on the two
models.
The square
marker shows the best
estimates and the error
bars show the uncertainty
bounds, based on the
Threats Model (y axis) and
the Records and Surveys
Model (x-axis). The thick
red lines indicate the
thresholds of P(E) for
listing a species as EX, and
the thin red lines indicate
the thresholds for listing as
CR(PE).
The recommended thresholds are based on considerations of the costs of making the wrong call
and benefits of making the right one, and on consideration that the costs are not the same for the
different types of errors (e.g., listing an extant species as EX versus not listing an extinct species
as EX). Akçakaya et al. (2017) presents a detailed discussion of these considerations. In addition,
these thresholds were tested for birds (Butchart et al. 2018) and for a small number of amphibians
and invertebrates.
Red List Guidelines
87
Nevertheless, the recommended thresholds should be considered as guidelines, rather than rules.
For example, Butchart et al. (2018) give examples of species for which they judge the probability
of extinction to have been under-estimated by these methods. One reason this could happen is that
the records may not be independent as assumed by the Records and Surveys Model. A possible
mechanism of dependence among records might be that, if a record is known publicly, it may bias
the judgment of less-experienced observers, or increase their likelihood of claiming a record.
When interpreting the results for possible listing of the species as CR(PE) or EX, results of both
models, as well as the uncertainties of the results should be considered. For instance, if both
methods give P(E) estimates with lower bounds above 0.9, then there is strong indication that the
species should be listed as EX. Conversely, if both methods give P(E) estimates with upper bounds
below 0.5, then there is strong indication that the species should be considered extant.
When the two methods give substantially different results, but have similar amounts of
uncertainty, we recommend that the decision is based on the method that gives the lower value for
P(E). In other words, listing as EX, for instance, requires both methods to give P(E)>0.9. This
corresponds to "Method 1" in Butchart et al. (2018; see Figure 1).
If the two models have substantially different amounts of uncertainty, the assessors may consider
giving more weight to the model with narrower uncertainty bounds. Such consideration can be
guided by calculating a weighted average of the two P(E) estimates, where the weights are the
reciprocal or complement of the uncertainty range (i.e., 1/range or 1-range, where range is
P(E)max-P(E)min; see the 'Results' worksheet).
The spreadsheet of input estimates and the output extinction probabilities calculated using these
methods should be documented and referenced (if published) or submitted (as Supplementary
Information) as part of Red List assessments for the relevant taxa.
11.4 Calculating the number of extinct species and extinction rates
Analyses that calculate the number of extinct species (globally, in a region, or in a taxonomic
group) or extinction rates (proportion of species that have gone extinct) should consider estimates
of P(E), the probability that a species is extinct. If P(E) can be estimated for all species, the number
of extinct species should be estimated as the sum of these probabilities— rather than simply
summing the numbers of species listed as EX or CR(PE)—so that the estimated number of extinct
species is independent of the thresholds of P(E) for EX and CR(PE). See Akçakaya et al. (2017,
Table 3) for a demonstration of this calculation.
If P(E) is not calculated for some species that are listed as EX or CR(PE), then the above
calculation should be made by assigning a weight to each such species, and summing those
weights. The weights should be the average P(E) for species in the same taxonomic group with
the same listing, for which P(E) has been calculated. If there are no (or very few) species in the
group for which P(E) has been calculated, then the weights should be 0.95 and 0.70, for species
listed as EX and CR(PE), respectively. These weights are based on the midpoint of the range of
P(E) for each category based on the recommended thresholds (see above).
Red List Guidelines
88
12. Guidelines for Threatening Processes
As discussed in an earlier section (2.3), the criteria aim to detect symptoms of endangerment rather
than causes (see also Mace et al. 2008). Consequently, they are applicable to any threatening
process that results in symptoms such as population decline, small population sizes, and small
geographic distributions. A taxon may be classified as threatened even if a threatening process
cannot be identified. Regardless of the nature of threats, assessments must follow IUCN (2001,
2012b) and these guidelines to ensure valid application of the criteria. However, different threats,
especially new or poorly understood processes such as global climate change may require further
guidance in the application of definitions and criteria.
The purpose of this section is to provide such specific guidance. In this version, we focus on
global climate change; future versions will provide further guidance on how the criteria may be
interpreted to assess taxa affected by other threats. It is important to note that the guidance in this
section is not an alternative to previous sections.
One aspect of a Red List assessment involves listing the major threats in the required
documentation, as described in IUCN (2001, 2012b; Annex 3), using a standard classification
scheme available at www.iucnredlist.org/technical-documents/classification-schemes. The
guidance given here does not relate to this process; instead the focus is on the application of the
Red List Categories and Criteria.
12.1 Global climate change
There has been concern that the Red List Criteria may not be adequate for assessing species
threatened with climate change. This is because many species that are projected to undergo
substantial range contractions in the future have short generation lengths. Consequently, there are
concerns that the assessment time frames are too short for the inferred population declines to
trigger the relevant IUCN Red List Criteria, which consider declines over a three-generation
period (see section 12.1.1). However, recent studies show that the IUCN Red List Criteria can
identify species vulnerable to extinction due to climate change. In a study involving North
American reptiles and amphibians, Pearson et al. (2014) showed that extinction risk due to climate
change can be predicted by information available now, such as current occupied area and
population size, much of which is used in the IUCN Red List Criteria.
Stanton et al. (2015) defined "warning time" as the time between when a species is first identified
as threatened and when it goes extinct, assuming no conservation action. Using the same species
and climate projections as Pearson et al. (2014), they showed that IUCN Red List Criteria can
identify species that would go extinct because of climate change without conservation action, and
can do so with decades of warning time. In an independent study, Keith et al. (2014) reached the
same conclusion for a short-lived Australian amphibian. Although these studies show the ability
of the IUCN Red List Criteria to identify species vulnerable to extinction because of climate
change, they also show that warning times may be short in data-poor situations, and if conservation
action is started only when a species is listed at the highest IUCN threat category (Critically
Endangered). Therefore, there is a need to develop further guidance for using the IUCN Red List
system, especially in data-poor situations and for timely policy responses to exploit the maximum
warning time available for species on extinction trajectories in response to changes in climate. As
new research increases understanding of the impacts of climate change on species, the results will
be used to improve these guidelines. Below, guidance is provided on a number of relevant issues,
based on research available in 2015.
Red List Guidelines
89
12.1.1 Time horizons
An important issue in the application of the criteria to species impacted by climate change
concerns the time horizons over which the assessments are made.
The time horizons used in the criteria serve several purposes. First, the generation time is used as
a surrogate for turnover rates within populations and as a biologically relevant scaling factor that
corrects for the variation in rates at which different taxa survive and reproduce. Second, the time
horizon is set to a minimum of 10 years because measuring changes over shorter time periods is
difficult and does not reflect time scales for human interventions. Third, the time horizon is set to
a maximum of 100 years into the future, because of the uncertainties in predicting population sizes
for a long time from the present day (Mace et al. 2008).
The global climate is projected to continue to change for several centuries (IPCC 2013; Chapter
12). The effects on biological systems will certainly continue for a long time. Thus, for many
species, especially short-lived ones, Red List assessments are based on time horizons much shorter
than the long periods over which we now expect the world's climate and its effect on species will
change. This by itself may not make climate change fundamentally different: other threats, such
as habitat loss may also continue for a long time.
However, the nature of change in biological systems caused by climate change is thought to be
different than changes caused by other threats. Thuiller et al. (2005), for example, argued that,
"the recognized time scales for assigning species IUCN Red List Categories are not suited to
evaluating the consequences of slow-acting but persistent threats," suggesting that the projected
climate change impacts are thought to be of a more deterministic nature than other threats. In
addition, some amount of climate change-related impact is irreversible (already committed)
because of the lag between greenhouse gas emissions and climate change (and subsequent
biological change).
While stochastic events (catastrophic fires, ENSO events, etc.) that contribute to the variability
and hence the risk of extinction of populations clearly operate at different time scales than climate
change, there are other processes that also are slow-acting and persistent. For example, it is
debatable whether threats such as habitat loss and fragmentation are any less persistent or any
more uncertain than climate change. Although climate change may be persistent, the predictions
are also very uncertain. For example, IPCC (2013) makes most of its predictions only until 2100
because general climate models tend to produce very different outputs towards the end of the 21st
century.
The criteria recognize that some threats may be irreversible (as explicitly noted in criterion A).
For example, in many cases, habitat loss brought about by urban sprawl is not reversible. Various
threats may involve time lags similar to that of climate change. For example, human populations
have a momentum, and thus there is often a lag between a change in the human population growth
rate and resulting changes in human pressures on natural systems.
Thus, the assessment of species with short generation times is not fundamentally different under
climate change and under other threats. Although short-lived species may not be listed under
criterion A initially, if they are affected by climate change they will be listed (likely under criteria
B or C) as their ranges and populations change as a response to climate change. They can also be
listed under criterion E (see below).

 

 

 

 

 

 

 

 

Content      ..     1      2      3      4      ..