The New IUCN Criteria

Birds To Watch 2

The IUCN criteria for assigning threat status and category, in use in various forms ever since the RDB programme began in the 1960s, have been a source of recurrent uncertainty and dissatisfaction, and moves have been made since the mid-1980s to replace them with a less subjective and more accountable system, using numerical criteria that reflect stepwise increases in the risk of extinction (Figure 1) based on measured or reasoned rates of decline, population levels, and range sizes (for a brief history of this development, see Collar and Stattersfield 1994, Mace and Collar 1994).

Figure 1. Extinction probabilities and the IUCN threat categories. This representation indicates the relative difficulty (represented by the relatively small, dark rectangles enclosed by their threshold lines) of qualifying as Endangered and, especially, Critical, compared with Vulnerable (large, pale rectangle).

The first outline of the new system was in Mace and Lande (1991). Although their proposals were explicitly declared provisional, they were rapidly adopted by the Captive Breeding Specialist Group of IUCN and widely deployed in its joint workshops (with taxon-specific groups) to develop Conservation Action and Management Plans (CAMPs) for particular families. This has meant that several such CAMPs for birds, on (e.g.) penguins, waterfowl, Galliformes, parrots and pigeons, have taken place using this prototype system, and their results are in some cases about to be published. We have sought to achieve consistency between these documents and ours, particularly as our work, extending into August 1994, has produced important updated status assessments of many species that were the subject of CAMPs. However, we cannot guarantee precise alignment of judgements.

Mace and Lande (1991) was replaced by Mace et al. (1992), which was in turn upgraded by the proposals that went before the IUCN General Assembly in January 1994 and which are summarized in Collar and Stattersfield (1994); and these were themselves refined (basically through the absorption of the problematic 'Susceptible' category into 'Vulnerable') in a final review made by G. M. Mace for IUCN's Species Survival Commission in late May 1994. Our categorizations and codings make use of this last version of the proposals (Mace and Stuart 1994). The definitions of the categories are given in Box 1 and their relationship to the criteria is shown in Figure 2, with their relationship to one another in Figure 3; the criteria and codings for the threatened categories are summarized in Table 1, with rather more detail given in Tables 2-5. We must emphasize, however, the importance of referring to the original document (Mace and Stuart 1994), as it contains many important qualifying remarks and points of clarification that cannot be reproduced in this outline of the system.

Figure 2. Relationship of criteria to threatened categories. The categories are decided by different thresholds in five main criteria, with Vulnerable additionally being decided by a range-size stand-alone.

Figure 3. Relationships between the IUCN categories. The dendrogram shows the path down which a species travels to arrive at a particular categorization. The line-up of categories on the right reflects the way in which the new system attempts to rank species in categories that are almost, but not entirely, nested.

Our only departure from the system is to retain (from previous BirdLife RDBs) the unofficial category of 'Near-threatened' which, although not used as such in the newest proposals, covers what those proposals refer to as 'Low Risk (type i)' taxa (see Box 1). It should be noted that in the case of Asia the Near-threatened list is disproportionately large, reflecting our wish to establish a comprehensive list of candidates for evaluation in the forthcoming Threatened birds of Asia. Moreover, the category Extinct has been used here only for species that had been included as possibly extant in the first Birds to watch but which we now feel are more appropriately placed apart from the main list, plus a few species that, since 1988, are thought to have become extinct.

Problems in the application of the new criteria

For the most part we have found it relatively straightforward to apply the new criteria, so our judgement is that they appear to work well with birds (further commentary on this is provided in the next chapter). Here we simply wish to outline certain problems that we encountered when applying them, and in some cases the decisions we took to overcome these difficulties.

Two points to emphasize are that the guidelines to the new criteria insist on the use of the precautionary principle, and the criteria themselves explicitly require the use of numbers of mature individuals, when assessing a species for categorization. Any apparently inappropriate listing of criteria, in relation to the range or level of figures we cite, will in most cases be attributable to these two latent factors.

The precautionary principle is the arbiter in borderline cases, chiefly between threatened and non-threatened status. It simply requires the use of the lowest figures on range or population and of the worst circumstance in terms of decline and threats that can reasonably be accepted. It therefore disallows 'worst-case scenarios', but urges caution (what one might call responsible pessimism) in accordance with the evidence.

The number of mature individuals is regarded as a better guide to the conservation status of a species than is a simple total of all individuals (in the case of many invertebrates and fish the importance of this distinction is much easier to appreciate). In many cases where a species has qualified under the C criterion (<10,000 and declining) we acknowledge that the total population may be several thousand higher than 10,000, but with immature birds making up the difference.

A further point to register is the difficulty surrounding the concepts of 'location' and '(sub)population'. The guidelines for the new criteria attempt definitions (Box 2), but these are by no means comprehensive and probably never can be. The problem affects many of the criteria and technically destabilizes them.

In the following paragraphs we describe various difficulties (some of which are very closely related), and in some cases how we have responded to them. While we have attempted to group them in a sequence that reflects this relatedness, and while some may be more important than others, we have given each a separate heading for the sake of clarity.

Despite the intended shift in the new criteria towards greater objectivity and accountability, the classification of species for which information is sparse remains largely intractable and hence liable to great variation between users of the system. In the guidelines to the criteria, emphasis is laid on the importance of attempting to classify poorly known species, and on allowing them threatened status (on the precautionary principle) if there are any signs from which this could be inferred. The category Data Deficient exists to accommodate only those species which cannot reasonably be assessed from the assembled evidence (see pp. 15 and 215). Typical Data Deficient species are known from one or very few specimens or sites within a habitat that remains largely intact and where no other threats (such as introduced predators) are expected; many are highly cryptic (the list includes nine nightbirds and four rails). However, when long-term absence of a species combines with suspected habitat loss, it is more or less obligatory to assign threatened status, although the role played by assumption is still enormous: Archer's Lark Heteromirafra archeri (Endangered, being the chosen precautionary compromise between the plain possibilities that it might be extinct and that it might conceivably prove locally common) is an example. Clearly there is considerable scope for variation in the categorization of such species, which in other reviewers' hands might, with some justification, appear as anything from Extinct to Data Deficient.

Conservation Dependent exists for species that would very rapidly qualify as threatened if they were not under management; it is not for species which are under management but have already met the criteria (these are threatened species, irrespective of their management). It has been applied very infrequently (11 species in total: p. 213), partly because so few species are genuinely under conservation management, and partly because so little conservation management can confidently be judged effective. This is especially true for protected areas which, despite covering most of a species' range, may suffer external pressure, or be managed inappropriately, or provide protection for only part of a life-cycle (as is the case with seabirds, where factors in the wider marine environment may be of greater conservation concern). In many cases, further research and consultation are needed to ascertain whether CD is indeed the most suitable category (e.g. some of the Australian species allocated this status). In some developed countries, e.g. New Zealand, it could be argued that all Low Risk species are to a large extent CD and will remain so forever, but, for simplicity's sake, the category has been assigned to Saddleback Philesturnus carunculatus only, a species which was in the recent past listed as threatened. The same is true for species on predator-free islands where measures to prevent the introduction of exotics render all their endemics potentially CD; we opted for threatened status in such cases only when an introduction is thought very likely to happen, using criterion A2d (i.e. future rapid decline).

Species that satisfy the numerical criteria must be treated as threatened, irrespective of the intensity and effectiveness of the management they receive; they are not Conservation Dependent. However, the situation arises in which a bird like Kirtland's Warbler Dendroica kirtlandii, which intuitively appears 'critical', being very intensively managed both to prevent breeding success dropping to zero and to maintain its highly specific breeding habitat, can only qualify as Vulnerable, precisely because that management is currently preventing any decline. There may be some merit in coding to promote an otherwise Vulnerable (and perhaps Endangered) species to a higher category, for example by a new criterion that predicts a decline in the near future based on abandonment of current conservation programmes. For Vulnerable, this could express either of the two higher degrees of threat according to the reviewer's perception of the situation, but the criterion would have to be formulated in such a way that Conservation Dependent species did not then all qualify as Vulnerable.

The thresholds for the different threat categories have been set to try and reflect probabilities of extinction (see Figure 1, p. 14). However, the discovery that as many as 100 species are thereby expected to become Extinct within the next five years (see p. 22) suggests that the probability level at least for Critical is too high; it does not necessarily imply that the thresholds are set at the wrong levels. However, we certainly find that a species can move too easily from Vulnerable (under D2) to Critical (under B1+2c) as a result of any comment reflecting loss of habitat. Species with naturally small ranges, especially when their habitat is for the most part under protection, can consequently appear mistakenly classified: Ascension Frigatebird Fregata aquila, Grand Comoro Scops-owl Otus pauliani, Seychelles Scops-owl O. insularis, Puerto Rican Nightjar Caprimulgus noctitherus, Mount Karthala White-eye Zosterops mouroniensis and San Andrés Vireo Vireo caribeus are all clearly threatened, but intuitively not such that one fears for them in the short term (equivalent to a 50% chance of extinction in five years); several others, including the three Cocos (Costa Rica) and two Annobon (Equatorial Guinea) endemics, could also all instantly transfer to Critical on the slightest evidence of loss of habitat. The problem lies in the interpretation of 'continuing decline': in many cases information is poor, and thus, under the precautionary principle, the mention of the loss or deterioration of any small quantity of habitat may be enough to trigger this shift. An additional qualifying clause for the B1+2c criterion (for Critical and Endangered) indicating a rate or extent of decline may be appropriate.

The new system can theoretically disqualify species that are stable in their small ranges while accepting species with larger ranges that also occupy these areas and are stable within them, but which are suffering declines elsewhere. The equivalent problem arises with population thresholds, where a species with (say) a stable population of 1,500 mature individuals fails to be listed while one with six times as many qualifies because of a decline affecting even a small subset of its total numbers. Almost exemplifying this problem is the recently discovered Udzungwa Forest-partridge Xenoperdix udzungwensis, which (until hunting was assumed to be extensive) had proved difficult to list, since it had been found in two intact forest patches totalling almost 300 km2, although several already known species (Swynnerton's Robin Swynnertonia swynnertoni, Dappled Mountain-robin Modulatrix orostruthus, White-winged Apalis Apalis chariessa), also then encountered in these previously unexplored patches, continued to qualify on the grounds of their declines elsewhere.

Another case is the Brown Teal Anas aucklandica, which in its three races, chlorotis (2,100 birds and declining), nominate aucklandica (1,500 birds and stable), and nesiotis (60-100 birds and stable), qualifies under criterion B. However if, as is currently being proposed, these three races are treated as good species, the resulting 'new' A. aucklandica would probably lose its threatened status, since neither a decline in its numbers nor a contraction in its range (genuine influences while it remained a member of the 'old' A. aucklandica) exists, and both its population and range size are just too large to trigger a stand-alone criterion (D1 or D2). (Of course, splitting of species will tend to result in more taxa in higher categories, since they will inevitably have smaller range sizes and populations; in this case both chlorotis and nesiotis would be likely to emerge with a higher categorization than they currently do as members of A. aucklandica.)

The rapid decline criterion exists to record seemingly uncontrolled and catastrophic circumstances affecting a species. Its application becomes problematic when these circumstances are contained as a result of geographical isolation. For example, two Marianas pigeons, White-throated Ground-dove Gallicolumba xanthonura and Mariana Fruit-dove Ptilinopus roseicapilla, both went extinct on Guam, but survive in fair numbers (higher than the population criteria) on several other islands. However, the total area of these other islands is roughly the size of Guam, so that a 50% decline in 20 years has technically been registered. Nevertheless, as the decline was finite it was felt that both these species would better be placed as Near-threatened for the moment.

The choice between criteria C2a (population fragmented) and C2b (one single subpopulation) is often difficult owing to the emphasis in the proposed definition of subpopulation (Box 2, above) falling on the issue of gene-exchange rather than on the impact of discontinuity in food availability or in other resources. Although theoretically there may be genetic interchange between fragmented populations such that they can be regarded as a single population, in practice the massive fragmentation of habitats, leaving many small populations stranded, seems far more logical to register as C2a, whether genes continue to be exchanged or not. In the case of locations, scale was a problem as no size limits are given in the proposed IUCN system, so criteria relying on number of locations (B and D2) are open to application in an inconsistent way; our guiding (practical conservation) principle was to regard 'small' areas as single locations where threats (e.g. introduced predators) could have a rapid effect throughout.

The criterion D2, which at one stage in the formulation of the new criteria was the category Susceptible (see Collar and Stattersfield 1994), was intended to indicate the risk of extinction inherent in species with highly restricted ranges, such that they can easily disappear before a threat to them has even been recorded. We use D2 with some discretion: with seasonally gregarious species, such as geese and seabirds, we have tended to take number of flock areas or colonies to equal locations (bearing in mind the point above); but with species which do not congregate we have allocated this criterion to those which have a range of less than 100 km2.

No E coding (whose criteria correspond to the probabilities of extinction depicted in Figure 1, p. 14) was attempted because the formally published record of the results of population viability analyses is virtually non-existent. Moreover, there is real difficulty both in interpreting these analyses (since they provide a variety of results depending on combinations of anticipated futures) and even simply in knowing the status of the exercises (since to date most appear entirely informal and provisional). This is not, of course, to deny the value of such simulations; the entry for Nene Branta sandvicensis cites an interesting multiple-outcome PVA (which nevertheless illustrates the point that these outcomes are not easy to match to the new criteria).

The new system has been developed to minimize subjectivity, so that as far as possible assessments of risk are comparable between taxa. However, considerable emphasis is laid in the criteria on future prospects, and the future is an area where speculation and assumption inevitably play major roles. Moreover, the distribution of knowledge concerning broader types of key information such as on habitat destruction is inevitably very patchy. This means that while a specialist can be expected to assemble all the relevant species-specific material for an assessment (distribution, population, ecology), he or she cannot be expected to have access to general data relevant to a species' environment. Thus future (as well as present) assessments can be seriously influenced by knowledge or ignorance of key pieces of information. For the process of assessment to approach comparability between assessors, the distribution of knowledge on habitat loss and general conservation measures (e.g. the existence, condition and precise location of protected areas within a species' range) has to be as even as possible. This is one very good reason for publishing findings in detail, for example in Red Data Books and species action plans, as it discloses what knowledge has informed the judgements made and renders that knowledge accessible to other assessors of species with similar ranges and problems.

One problem in risk assessment is that the probability of a particular event may be very low while the probability of its outcome, unless immediately countered, is very high. Hybridization with vigorous invasive species is a threat that perhaps can only be neutralized at the moment the invader establishes itself in or near the range of the invaded (as in the case of the White-headed Duck Oxyura leucocephala and the Ruddy Duck O. jamaicensis). The introduction to or escape onto islands of certain predators or competitors is another predictably 'final' event, although the time it may take to have an impact (and indeed which taxa it may affect) may be impossible to predict. After the unforeseen but astonishingly rapid loss of the Guam avifauna to the brown tree snake Boiga irregularis it has been tempting to code as A2d every bird species on every island, irrespective of its size, wherever ships from Guam commonly dock. The establishment of primates on an island even as large as New Guinea-let alone Halmahera and other parts of the Moluccan archipelago-could, unless remedied by action within a few months, seal the fate of an unknown number of avian (and other) frugivores-pigeons, parrots, birds of paradise-that have speciated in the absence of competition from such a source.

It is impossible to devise a set of criteria that respond helpfully and consistently to this problem. The D2 criterion exists to highlight a particular condition under which the process of extinction can occur very rapidly and often without detection until after its completion; yet, as the examples above show, highly restricted range is only one of several evolutionary circumstances that permanently expose certain species to seemingly irreversible events. Understanding all such circumstances and guarding against all such events is a hydra-headed challenge in biodiversity conservation, connected ultimately to wider problems in the realm of global environmental management (such as carbon emission, global warming and sea-level rise) with which conservation biologists have had, to date, relatively little practical involvement.

Box 1. Definitions of proposed IUCN categories (as given in Mace and Stuart 1994).

EXTINCT (EX)  A taxon is Extinct when there is no reasonable doubt that its last individual has died.

EXTINCT IN THE WILD (EW)  A taxon is Extinct in the Wild when it is known only to survive in cultivation, in captivity, or as a naturalized population (or populations) well outside the past range. A taxon is presumed extinct in the wild when exhaustive surveys in known and/or expected habitat, at appropriate times (diurnal, seasonal, annual), throughout its historic range, have failed to record an individual. Surveys should be over a time frame appropriate to the taxon's life cycle and life form.

CRITICALLY ENDANGERED (CR)  A taxon is Critically Endangered1 when it is facing an extremely high risk of extinction in the wild in the immediate future, as defined in any of the criteria (A-E)2.

ENDANGERED (EN)  A taxon is Endangered when it is not Critical but is facing a very high risk of extinction in the wild in the near future, as defined in any of the criteria (A-E)2.

VULNERABLE (VU)  A taxon is Vulnerable when it is not Critical or Endangered but is facing a high risk of extinction in the wild in the medium-term future, as defined in any of the criteria (A-E)2.

CONSERVATION DEPENDENT (CD)  Taxa which do not currently qualify as Critical, Endangered or Vulnerable may be classified as Conservation Dependent. To be considered Conservation Dependent, a taxon must be the focus of a continuing taxon-specific or habitat-specific conservation programme which directly affects the taxon in question. The cessation of this conservation programme would result in the taxon qualifying for one of the threatened categories above.

LOW RISK (LR)  A taxon is Low Risk when it has been evaluated and does not qualify for any of the categories Critical, Endangered, Vulnerable, Conservation Dependent or Data Deficient. It is clear that a range of forms will be included in this category, including: (i) those that are close to qualifying for the threatened categories3, (ii) those that are of less concern, and (iii) those that are presently abundant and unlikely to face extinction in the foreseeable future. It may be appropriate to indicate into which of these three classes taxa in Low Risk seem to fall. It is especially recommended to indicate an appropriate interval, or circumstance, before re-evaluation is necessary for taxa in the Low Risk class, especially for those indicated in (i) above.

DATA DEFICIENT (DD)  A taxon is Data Deficient when there is inadequate information to make a direct, or indirect, assessment of its risk of extinction based on its distribution and/or population status. A taxon in this category may be well studied, and its biology well known, but appropriate data on abundance and/or distribution are lacking. DD is therefore not a category of threat or Low Risk. Listing of taxa in this category indicates that more information is required. Listing a taxon as DD acknowledges the possibility that future research will show that threatened classification is appropriate. It is important to make positive use of whatever data are available. In many cases great care should be exercised in choosing between DD and threatened status. If the range of a taxon is suspected to be relatively circumscribed, if a considerable period of time has elapsed since the last record of a taxon, or if there are reasonable chances of unreported surveys in which the taxon has not been found, or that habitat loss has had an unfavourable impact, threatened status may well be justified.

NOT EVALUATED (NE)  A taxon is Not Evaluated when it has not yet been assessed against the criteria.

1. 'Critically Endangered' is the formal term under consideration; it is referred to throughout this book as 'Critical'.
   
2. See Tables 1-5.
   
3. This is the distinction which in this book bears the title 'Near-threatened'.

Table 1. The new IUCN threatened category thresholds at a glance. 'Extent of occurrence' and 'area of occupancy' are defined in the caption to Table 3.

Table 2. Criteria for the IUCN threatened categories: species undergoing rapid decline. 'Extent of occurrence' and 'area of occupancy' are defined in the caption to Table 3.

Table 3. Criteria for the IUCN threatened categories: species with a small range and declining. 'Extent of occurrence' is the area contained within the shortest continuous imaginary boundary which encompasses all known, inferred or projected sites of present occurrence. 'Area of occupancy' is the area within the extent of occurrence which is occupied by a taxon (this measure is often applicable to species with highly specific habitats).

Small range, plus any two of: fragmented, declining, fluctuating

Table 4. Criteria for the IUCN threatened categories: species with a small population and declining.

Table 5. Criteria for the IUCN threatened categories: species with a very small population and/or a very small range.

Box 2. Definitions of 'Location', 'Population' and 'Subpopulation' in the proposed IUCN criteria.

This information comes from Birds to Watch 2: (The World List of Threatened Birds) NJ Collar, MJ Crosby and AJ Stattersfield (1994) BirdLife, Cambridge. This is the official source for birds on the IUCN Red List.

BirdLife is interested in hearing about new information or references on threatened species. See contact details at the end of this page.

Birds to Watch 2 and other publications are available for purchase from Natural History Book Service, email: nhbs@nhbs.co.uk, http://www.nhbs.com.

For further information please write to: Wellbrook Court, Girton Road, Cambridge, UK CB3 ONA Tel: +44 1223 277318 Email: redlist@birdlife.org Document URL: http:// www.unep-wcmc.org /species/data/birdlife/btw2.htm Revision date: -September-2007 Current date: -May-2008 Comments © BirdLife