IUCN Orinigial site: click here
Yellowed by H.M. e.g. click here
|
The
IUCN Criteria Review: 3. Discussion of issues
a)
3D ranges It was agreed that a particular problematic range area was the 'typically less than 100 km2' annotation under criterion D2. This was a very small AOO for marine species and meant that many marine species were excluded from listing as Vulnerable D2, when this was an appropriate category. It was agreed that D2 could be reworded to make this more flexible. It was agreed that this would be helpful. RECOMMENDATION 1: Alter the definition of Vulnerable D2 by excluding the phrase, 'typically less than 100 km2'.
RECOMMENDATION 2: The terms population and sub-population in the IUCN rules for categories and criteria should be altered to refer to species and populations respectively. There
was further discussion about how sub-units should be recognised. T. Yahara
suggested that it is normal to consider genetically distinct units as
separate populations, but it was agreed that there would be difficulties in
implementing this generally given poor information. A. Vincent thought it
important to retain the ability to list threatened populations but D.
Butterworth felt that IUCN should only consider global species listings. Two options were presented by the working group: Option 1: If any two populations (or more) are 'safe', then the species cannot be listed in the global Red List. In this case, only regional listings can apply. Option 2 : The decline rate of a species should be the weighted average of the decline rates of all populations for which data are available. These were discussed. R. Akçakaya considered there were problems with option 1 because the populations might not actually be isolated. Y. Kaneko preferred option 1 because many recent disputes over IUCN listings had resulted where there were agreed to be secure subsets of listed species. There was no consensus and the discussion was continued later (see section 3.5). 3.
Issue 5: Dealing with life history variation The
discussion in this group had focused on improving some definitions used in
the criteria in order to reflect the unusual life histories of some marine
species. The group had considered the following issues, and made proposals
for amendments to the definitions of 'mature individuals' and 'generation
length' to take account of the following life history characteristics:
variation in recruitment, sex changes, clones, sessile species,
inter-dependence between species, and generation length. RECOMMENDATION 3: Alter the definition of mature individuals to read, 'The number of mature individuals is the number of individuals known, estimated or inferred to be capable of contributing to the next generation….' There
should also be two extra bullet points as follows: - Mature individuals that
cannot produce new recruits should not be counted (e.g. densities are too low
for fertilisation or reproductive conditions are not met), and; - It should
be emphasised that the sex that limits the rate of reproduction is not always
the rarer sex. RECOMMENDATION 4: An alternative definition of generation length is 'Generation time is greater than the age at first breeding, except in taxa that breed only once. It may be estimated in several ways. For example, (a) the average age of parents in the population, (b) the age at which 50% of total reproductive output is achieved' 4. Issues 3 and 6: Population
decline and extinction risk, and setting time frames for assessment Under criterion A, species can be listed as threatened simply on the basis of a past or projected decline in population size that exceeds some critical threshold. The intent here is to be able to list species that are in serious decline without the necessity to have a robust estimate of the total population size. However, in practice a number of very abundant species may have been listed inappropriately using this criterion. The problem is to find a new formulation of criterion A.. This should reflect the fact that rapid decline is a strong indicator of extinction risk, and something that can be measured or estimated even in poorly known species, but the measure needs to provide a more accurate indicator of the actual extinction risk. The working group reviewed the options for amending criterion A, and had also developed a new alternative. The plenary discussion revolved around a discussion of the various options presented by the working group. Option
1: No change Option
2: Elimination of criterion A. Option
3: Relate the decline rate directly to population size. Both these options could be expressed as opt-in or opt-out conditions. For example, an opt- in criterion would be where the species qualifies once it has lost a critical (high) % of virgin biomass. An opt-out formulation would use the current criterion A, but the species would be allowed to opt-out of listing if its loss of virgin biomass were shown to be less than some critical (low) % of the original value. This second formulation (the opt-out clause) was favoured by the working group, who had various suggestions for what the opt-out might be. They made both relative and absolute suggestions, including, for example, N>106 or an analysis showing that the probability of N<10,000 was less than 0.1 etc. A. Punt pointed out that (a) could be considered to be an example of how (b) might be defined. T. Yahara supported the general idea of linking declines to population size. J. Meeuwig observed that K was very hard to estimate and varied across time and space for most species. C.Roberts also thought that absolute population size thresholds for opting out could be non-precautionary especially if they did not recognise depensatory mechanisms. J. Reynolds pointed out that depensation should be accounted for in the definition of mature individuals, which would exclude individuals that did not breed at low density, but he acknowledged that this subtlety might be lost. D. Butterworth stated that expressing population size as a fraction of some reference level was sensible and would give the system credibility with fisheries scientists who often use this kind of measure. B.Taylor noted that the opt-out clause shifts the burden of proof to those seeking to de-list species, and this was probably the right way to do it. They would need to have good data to do this and so the clause might have limited usage. R. Akçakaya cautioned against making it too easy to opt out; he cited the spotted owl example from the NW USA where K will always be 100% but all the time the habitat is declining. There is also a related problem with measuring 'virgin stock' for many species where historical levels are very different to those seen today. W. Bond commented that criterion A is most commonly used for exploited species, perhaps there should be a different criterion for these. G. Mace disagreed; criterion A is widely used for species facing habitat loss and the effects of introduced species. A. Punt emphasised again the essence of the opt-out option: that there is a 'future-safe' state that would be used for the opt-out. This could be a relative or an absolute measure. J. Musick said that relative measures would generally be problematic. S. Stuart asked if the group generally supported the principle of the future-safe opt-out. C. Roberts said he was concerned about predicting safe futures from models given the evidence of failures in the past from fisheries models. A. Vincent agreed and stated that dramatic declines should always be cause for concern. She thought that there could be a place for an opt-out but warned against over-reliance on indices (e.g. CPUE) which may be inaccurate measures of status (e.g. with improving technology appearing to reduce effort but effectively increasing it). Opt outs would need very careful analysis. H. Matsuda thought that linking population size and decline rate could be more important in criterion A2 (future projection). R. Akçakaya emphasised that whatever the safe-future clause was, it should be precautionary. S. Stuart concluded the discussion by suggesting the consensus was that a safe-future option was definitely worth further thought. Option
4: Change 'reduction' to 'continuing decline' in criterion A. B. Taylor pointed out that this involved shifting the burden of proof away from the assessor - i.e., they only have to suggest that the decline is not continuing and the species may not be listed. The burden is then on others to prove it is continuing to decline, after accounting for natural variations etc. Maybe we need evidence that the population is actually increasing and recovering. S. Stuart suggested that the discussion of this option should be folded into other 'future-safe' options. Option 5: Increase the threshold values for decline rates (i.e. increase the 20%, 50% and 80% currently used in criterion A) S. Stuart commented that there has been much discussion in particular about the 20% decline rate for VU, which is picking up too many species. These values may well need to change whatever else happens. G. Mace outlined the methods used to derive these values and agreed that they are probably too precautionary, at least at VU. J.Musick felt that these numbers are ridiculously low for fishes and it was always going to be hard to distinguish these kinds of decline rates from natural fluctuations. He proposed 95% for CR and 80% for EN. A. Punt suggested that the original derivations should be revisited; these values appear to work for some long-lived species, the problem may just be with very productive species although new formulations for estimating generation length may help with this problem. W. Bond said that 20% loss of habitat was a serious issue for may species, although 20% loss of population might not be. T. Yahara suggested that his proposal to explicitly link A to B and C would help. D. Butterworth also felt the values were absurd for fisheries although he could live with something like this if there were an opt-out clause. A. Vincent commented that the discussion should not only focus on commercial species. Lots of other marine species are more stable and for these a 20% decline is a serious issue. If these are not identified by the criteria then there is a risk that their numbers will fall so low that it will be too late. One role of the Red List is to provide adequate warning. C. Wood said that a key problem was how to measure and detect decline in species that show very wide natural fluctuations. Y. Kaneko emphasised the importance of getting this issue sorted out because of the influential nature of the Red List and the implications for CITES. S. Stuart suggested that a way forward was to review the background to the development of these values originally, and then to consult other groups for feedback and to see how they work. B. Taylor agreed and also suggested looking at the empirical evidence on what kinds of declines actually lead to extinction. G. Mace agreed that this was a good way forward but said that looking at the empirical data was problematic because the information available was strongly weighted towards species that had persisted despite declines. There are few good data sets on species that have gone extinct. A. Punt said that in any case this needed documenting. H. Matsuda made the suggestion that evidence that a decline had ceased should be taken over 1 generation length, but that evidence that it was increasing should be measured over a longer period, say 2-3 generation lengths. RECOMMENDATION 5: The formulation of the values for decline rates in A should be reviewed, and these values should be tested against real data sets. Appropriate levels for revised, more exclusive thresholds should be investigated. Option
6: Allow criterion E to over-rule other criteria. R. Akçakaya said that E was not sufficiently precautionary to be allowed to over-rule A. E. Hudson pointed out other problems with this option. Modelling is not objective and biases can come in due to choice of parameter values and model design. This could create a large and difficult task for assessors seeking to arbitrate on alternative assessments. Criterion E could not be used for species where there was limited information. However, this option has the advantage of simplicity. D. Butterworth stated that this solution would solve the fisheries problem; fisheries managers routinely use stock assessments towards this end, and have developed a range of precautionary approaches to fishing. C. Roberts said that models give the illusion of rigour but cannot deliver confidence, he would not like to see the non-precautionary E over-ruling the precautionary A. T. Yahara said that modelling should be encouraged in conservation biology and that a simple calculation is better than a subjective judgement. J. Meeuwig observed that models should be used for heuristic analysis and for discovering more about the nature of systems affecting species and not for making projections. J. Musick felt that since more data is needed for E then there must be higher confidence than A; if it is possible to use E then it could over-rule A. A. Punt pointed out that criterion A too is a model, albeit a much simpler one. He felt that the right place for this proposal was as part of an opt-out in option 3. R. Akçakaya said that single species models are much simpler than ecosystem models. In any model output there will be a range of answers, including all the uncertainty measures. The problem is to choose between them. This need not be completely arbitrary if there are standards for models, and this would be important. H. Matsuda said that IUCN should encourage the collection of better data and should not discourage the development of models that make use of these data. G. Mace pointed out that allowing E to over-rule A has other implications. It puts E onto a different status from the other criteria. Subsequently there might be justification for allowing E to also over-rule B, C and D, so that the quantitative analyses allowed in criterion E would need to be done for all species. Given current resources for data collection and species analysis this could be detrimental to the inclusiveness of the Red List. S. Stuart concluded the discussion by commenting that there was no consensus on this issue. Option
7: incorporate the intrinsic growth rate explicitly into criterion A. A. Punt said that recovery potential would be incorporated into the new definition of generation length. The age at which maximum production is reached will be younger for a fished species and so the generation length over which declines are measured becomes shorter. The extent of this shift will increase with density dependent responses to fishing. C. Wood felt it would be difficult to estimate the extent of density dependence in natural mortality, and that in practice estimates of rMAX would depend on assumptions about the most appropriate stock recruitment model. D. Butterworth thought that generation length already accounts for variation in r. H. Matsuda pointed out that generation length and natural mortality vary widely. B. Taylor felt that this needed more thought. Some species increase their generation times under exploitation. Option
8: Simple 'unless the future is safe' opt out clause. RECOMMENDATION 6: The
current formulation of criterion A is generally agreed to be too inclusive
and has resulted in the inappropriate listing of many species. The group
discussed many options and generally agreed: (1) the threshold decline rates
are probably too low (and therefore include too many species); more work
needs to be done to set new levels. (2) the most precautionary new
formulation of criterion A would be to develop an opt-out clause of the
following form:
b)
Accumulating population status up to species status (revisited) (1)
If there are 2 or more 'safe' populations, then the species is not listed. D. Butterworth said that option 1 had to apply if risk of extinction was really being measured. C. Roberts asked how it would be determined that populations were 'safe'. J. Musick commented that just two safe populations in option 1 was reckless, A. Vincent agreed that option 1 lacked credibility. R. Akçakaya agreed that option 1 was a reckless definition of safety. B. Taylor thought that option 1 was more credible, but that the number of populations regarded as 'safe' could be increased to 3, 4 or more. S. Stuart suggested that defining 'safe' might benefit from the discussion of the opt-out clause for criterion A.. D. Butterworth and G. Webb both felt that this issue was at the heart of the credibility of the Red List; it was illogical to list a species at risk of extinction if it was represented in known 'safe' populations. A. Vincent said that in most cases where a species has lost populations, other populations are also at risk. R. Akçakaya also disagreed with option (3). He pointed out that fragmentation could increase the number of populations, and also asked what total would be used when calculating the % populations declining. There could be arbitrary consequences; a species that is continually declining would be listed when 12 out of 15 populations were in decline, but when only 1 out of the remaining three was declining it could be de-listed. J. Reynolds said that this discussion was only of population size and only referred to criterion A. There needed to be methods for assessing other aspects of populations. C. Roberts agreed; most marine species cannot be assessed by their population size, only their range. T. Yahara commented that the Japanese plant list was compiled using a system like option (2), but because they could not measure population size they weighted using the number of localities to obtain a species value. An example was presented of a species which had 3 populations meeting CR, 3 meeting EN and 3 meeting VU. Since 66% of populations (>50%) meet at least the EN criterion the species would be EN. R. Akçakaya said this only applied to criterion A. D. Butterworth thought this method could be extended to other criteria as a meta-rule. C. Roberts and J. Meeuwig were both concerned that accumulating up population data according to these formulae would lose a lot of important information about different conditions affecting different populations, and that the units needed to be biological populations and not just locations or assessment units. It was generally agreed that there is a problem related to listing species where populations face different degrees and kinds of threats, and their assessment as being globally threatened may present difficulties for management agencies. There followed a general discussion on the options that were presented, with lack of consensus. G. Mace thought this might be a case where it was very hard to generalise and that methods might be best addressed by guidelines to assessors. S. Stuart suggested that there needed to be some specific examples to work on. Participants were encouraged to send relevant examples to G. Mace. RECOMMENDATION 7: Scaling population data up to the species level can be done in a number of ways. The costs and benefits of alternative proposals should be examined using actual species data.
(1)
(2)
J. Reynolds agreed that a general definition of a clone is that it is that which regenerates when it is cut off. He thought that generation length could be calculated by weighting both kinds of reproduction and that adding in seed bank storage was also fine. He also suggested that an appropriate definition for generation length in these cases was the time from seed release to subsequent seed release in the genet. G. Cailliet said that different people used different measures of generation time, and that it was important also to look at the invertebrate literature. S. Stuart encouraged people to send relevant data and text to G. Mace. (3)
G.Webb agreed that 3 generations was too long a time period for assessment of crocodiles. J. Musick said the past declines were not now relevant and W. Bond agreed that capping made sense in terms of human time scales. A. Vincent thought that the past declines were important even if they were long ago. She said that capping the past was a problem because information about past status could rapidly be lost. She further warned about shifting baselines, such that we began to consider recent (already depleted) numbers to reflect the original or depleted numbers. J. Meeuwig agreed that this could mean the assessment period was too short for picking up non-recovery. G. Cailliet was concerned that species showing long-term fluctuations would be qualifying incorrectly if the time period was short. E. Hudson pointed out that generation time was important in the criteria as a scalar and that capping it would limit its usefulness. B. Taylor agreed that this related to the issue of recovery potential and that this proposal might act against the listing of the most vulnerable species. She thought 60 years (3 x 20 years) was too short a time to assess the status of long-lived species. There was a general discussion about the effect that the cap would have in terms of species dropping off the list. G. Mace said that it was important to know the implications and asked for examples from participants. R. Akçakaya suggested that the second reason for capping generation length (internal consistency) could be avoided by adding a generation length alternative to the definition of VU (i.e. …in 100 years, or 5 generations, whichever is longer) S. Stuart concluded that there was no general consensus here, although nobody objected to capping future projections. Many people felt that 20 years was too short for the maximum generation length for use in past, but most agreed that there should be some cap on the past. There was a more fundamental issue raised by A. Vincent about the importance of listing depleted species since population declines are important for population viability and ecosystem function, whether or not they pre-sage extinction. RECOMMENDATION 8: Generation times used for future projections should be capped. The decision about capping generation times used for measuring historical declines should be considered further, especially by looking at the kinds of species that would come off the list as a result of different maximum lengths. 20 years is too short, so another solution should be sought to establishing internal consistency in criterion E, e.g. adding a generation time alternative to VU. b)
Conservation dependent (cd) G. Webb and A. Vincent felt it was important to have a label for species that are under management; G. Webb thought the term misleading - it implied this was the most needy species. G.
Mace said this issue was a trade off between complexity and accuracy. S.
Stuart said that revising cd could be a big task, and he asked whether anyone
thought it of very high priority, He suggested that this issue was better
addressed by clearer guidelines and documentation, and there was general
agreement with this. J. Musick said the shark group used nt where there was not enough information to place a species in a threatened category. C. Roberts thought it played a useful role in assessing biodiversity status, even in its present form. B. Taylor pointed out that the level of risk in VU is actually quite a serious issue and that therefore nt was important. W. Bond was concerned that the relationship between the Lower Risk categories was often misunderstood, and that nt and cd were very different from lc, which was actually 'safe'. There was agreement that the cladogram in IUCN rules needs clarification and that there needed to be better explanation of the roles of the sub-categories of LR. S. Stuart suggested that more guidelines were needed but that this was not a very important issue to deal with immediately. RECOMMENDATION 9: Better explanation of the role of the sub-categories of LR is needed and a re-drafted figure. The development of criteria for nt and the expansion of cd are not a high priority.
e)
Uncertainty There was general agreement that this was a helpful approach, especially for providing explicit information about how each assessor was dealing with risk and uncertainty. This topic would be explored more at a later workshop. This method could also provide a clear role for nt.
J. Paxton, C. Wood and others considered that the main difference was poor detectability and difficulties with obtaining measures in the oceans. A. Vincent also thought there was massive potential for habitat degradation from the effects of pollutants etc. C. Wood thought there might be longer auto-correlation in environmental variables affecting recruitment in the oceans. In general, the differences were thought to be of scale only.
R. Akçakaya said that criterion E could be used although there might need to be higher extinction thresholds to make it more precautionary. J. Reynolds said that it should be dealt with under criterion A and there was no uniqueness in this effect for marine species. S. Stuart thought that addax, oryx and desert tortoises might be terrestrial examples. W. Bond thought there were plant examples, but he did not think that this was worth a new criterion. J. Musick thought that people were over-emphasising the probability of depensatory effects. C. Roberts said that sessile marine invertebrates might suffer more serious depensation. S. Stuart concluded by saying that this might justify more thought, but at present there seemed to be no basis for a whole new criterion. 9.
Communication R. Froese reiterated that it was very important to redesign the figure in the red booklet describing the nature of the categories, especially to distinguish least concern more clearly from other LR categories. He also stressed the importance of clarifying the terminology used for populations and subpoulations (see section 3.2a). Given the wide application of IUCN categories, A. Vincent was worried about the consequences of species being de-listed. S. Stuart said that a number of high profile species had been de-listed last time with little impact, e.g. wolf, tiger, white rhino. There was also some discussion about improving the effectiveness of the nt category. C. Wood suggested that if the opt-put clause were adopted for criterion A, then nt could be used to flag concerns about declining trends in opted-out species. There needs to be more detail and clarity in the red book; S. Stuart agreed this was a high priority. |