ICES Journal of Marine Science: Journal du Conseil Advance Access originally published online on March 9, 2007
ICES Journal of Marine Science: Journal du Conseil 2007 64(4):718-722; doi:10.1093/icesjms/fsm011
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
When control rules collide: a comparison of fisheries management reference points and IUCN criteria for assessing risk of extinction
Department of Fisheries and Oceans, 200 Kent Street, Ottawa, Ontario, Canada K1A 0E6
Correspondence to J. C. Rice: tel: +1 613 9900288; fax: +1 613 9540807; e-mail: ricej{at}dfo-mpo.gc.ca
Rice, J. C., and Legacè, È. 2007. When control rules collide: a comparison of fisheries management reference points and IUCN criteria for assessing risk of extinction. – ICES Journal of Marine Science, 64: 718–722.The quantitative criteria used by the International Union for the Conservation of Nature (IUCN) to assess risk-of-extinction are compared with reference points used by ICES and other fisheries organizations for advising on fisheries management. Criteria based on numbers of individuals and geographic range appear to be in harmony with limit reference points and control rules used in fisheries management, with reference points indicating that fisheries should be closed well before there is any risk of extinction. However, there is huge potential for conflict between fisheries and risk-of-extinction approaches when considering the extent of population declines. Of 89 species examined, the decline criterion suggested a serious risk-of-extinction in 87%, whereas most of the stocks were still within a zone that allowed fisheries management reference points to indicate that exploitation could continue. Much of the conflict seems rooted in different types of tolerance to risk between the two disciplines. The conservation-biology community acknowledges a high tolerance for "false alarms", to keep the probability of a "miss" very low, whereas tolerance in fisheries management is comparable for both types of error.
Keywords: fisheries, harvest control rules, reference points, risk of extinction, risk tolerance
Received 17 July 2006; accepted 19 December 2006; advance access publication 9 March 2007.
 |
Introduction |
|---|
 Top Introduction The origin of risk-of-extinction...GoalsComparison of triggersDiscussionReferences |
|---|
Evaluation of management strategies in ICES has largely addressed the impact of fisheries on target species. Some fisheries with complex mixtures of fleets and target species are addressed with comparably complex control rules and strategies for estimating the triggers of these rules (ICES, 2006; Kell et al., 2007; Ulrich et al., 2007). However, all this work has been in the context of keeping fisheries impact sustainable from the perspective of the target species.
Largely independent of fisheries management agencies and their scientific advisory bodies, organizations and agencies with responsibility for conservation of biodiversity have developed quantitative criteria for classifying the risk of extinction of individual species. In recent years, the conservation-biology community has applied these increasingly to marine species. Consequently, a lively debate has developed between the two communities regarding the risk of extinction of commercially exploited marine fish (Baillie and Groombridge, 1996; FAO, 2000; Powles et al., 2000; Hutchings, 2001; Dulvy et al., 2003, 2004) and the analytical methods used for estimating the risks (Ludwig, 1996; Mace, 2004). Nonetheless, increasing numbers of jurisdictions have legislation requiring mandatory conservation measures in response to designation of a species or population as being at risk of extinction (e.g. Canadian Species-at-Risk Act—http://lois.justice.gc.ca/en/S-15.3/text.html; Australian Environmental Protection and Biodiversity Conservation Act—http://www.deh.gov.au/epbc/index.html; US Endangered Species Act—http://www.fws.gov/endangered/esaall.pdf; Annex I of CITES—http://www.ukcites.gov.uk/intro/cites_species.htm), and marine species are being considered within these frameworks. Hence, by triggering mandatory actions by jurisdictions, quantitative risk-of-extinction criteria similarly function as a trigger for a control rule, even if the terminology differs.
Independent development of different control rules for conservation and for sustainable use of fish populations presents both an opportunity and a potential complication. The opportunity arises because of lessons that can be learned from the history of the use of risk-of-extinction criteria with marine fish. The complication could arise if risk-of-extinction criteria and fisheries control rules were to provide conflicting guidance on the appropriate management actions for a specific population. We examine both these considerations.
|
The origin of risk-of-extinction criteria |
|---|
|
TopIntroductionThe origin of risk-of-extinction...GoalsComparison of triggersDiscussionReferences |
|---|
The International Union for the Conservation of Nature (IUCN) organized a series of workshops from 1991 to 1994 to develop quantitative criteria for categorizing species by risk of extinction (Mace and Stuart, 1994). One intention of the categorization was to create a close link between categories at higher risk (Vulnerable, Endangered, Critically Endangered) and mandatory conservation actions by responsible jurisdictions. These efforts culminated in the criteria and accompanying narrative and guidelines contained in the IUCN Red Book of 1996 (Baillie and Groombridge, 1996). Three major criteria, each with several subcriteria, were proposed for assessing "risk factors across a broad range of organisms and diverse life histories" (essentially all plants and animals above the unicellular level), although it was acknowledged that certain criteria would be inappropriate for some taxa. The guidelines directed that each species should be assessed against all criteria (to the extent possible with available data) and, normally, should be assigned to the highest risk category for which it qualified on any single criterion. They also specified that criteria can be applied to population units below the species level (stocks, in fisheries management terminology), when these units are important components of biodiversity.
The criteria have remained under continuous review, with workshops on many problematic issues, including applications to exploited fish populations (IUCN, 1999; Mace, 1999). Revisions to the criteria and accompanying guidelines were released in 2001 and 2005 (IUCN, 2001, 2005). As intended from the start, the criteria have been adopted by many regional and national jurisdictions with responsibility for protection of species at risk, usually with relatively minor modifications.
|
Goals |
|---|
|
TopIntroductionThe origin of risk-of-extinction...GoalsComparison of triggersDiscussionReferences |
|---|
IUCN (2001) provides specific rationales for using quantitative criteria for assessing risk of extinction: "To provide a system that can be applied consistently by different people; To improve the objectivity by providing those using the criteria with clear guidance on how to evaluate different factors which affect risk; To provide a system which will facilitate comparisons across widely different taxa; To give people using threatened species lists a better understanding of how individual species were classified".
These rationales are similar to justifications presented by fisheries management for developing limit reference points (LRP) and applying harvest control rules. Nevertheless, there are important differences between the goals of fisheries management and protection of species at risk. LRPs are intended to preserve stock productivity at levels that allow social and economic benefits to be taken sustainably, and to keep the risk of serious harm to the stock low (ICES, 2005). "Serious harm" is considered to have occurred if a stock is reduced to a size at which productivity is likely to be impaired (good recruitment rendered unlikely) or when the likelihood of poor productivity is increased substantially. To take account of uncertainties in both annual assessments and the estimate of spawning-stock biomass (SSB) associated with "serious harm" (Blim), risk is managed by basing harvest control rules on precautionary reference points (Bpa). However, because the control rules based on Bpa are intended to maintain a low probability that SSB would fall below Blim, we use Blim as the fisheries-management trigger for comparison with species-at-risk triggers.
The fisheries management goal allows some flexibility in complex multispecies or multifleet management strategies, which may permit mature biomass of a few stocks to be reduced to sizes that impair recruitment, in exchange for continued extraction of social or economic benefits from ongoing fisheries. In such cases, though, even the most impacted stocks, well below precautionary reference points and on the verge of triggering major catch reductions or closures of fisheries, are considered still sufficiently productive to support rapid and secure recovery if fishing mortality were reduced.
Conceptually, risk-of-extinction criteria are intended to flag populations that are so small that their productivity is not just diminished, but may be inadequate to allow the population to persist, with some risk of becoming extinct within some specified number of years or generations. With a worst-case scenario of extinction, management actions associated with high-risk classifications have much less flexibility: everything possible should be done to reduce threats to the population and to promote its growth.
Given the large difference in intended function of the two types of triggers, there should be little conflict between their use in the two management contexts. Moreover, the guidelines for use of risk-of-extinction criteria include provisions that should reduce further the potential for conflict: "Natural fluctuations will not normally count as a continuing decline, but an observed decline should not be considered to be part of a natural fluctuation unless there is evidence for this" (Guideline #10; IUCN, 2005). In the 1996 Red Book narrative, it was noted that managed declines for harvests should also be excluded, but only when there was clear evidence that the decline was indeed part of a management plan. Evolution of this provision will be discussed later. On the other hand, the different goals in the two management contexts provide very different error tolerances. Fisheries management may aim to balance "false alarms" (control rules that reduce or close fisheries when such extreme actions were not necessary to maintain stock productivity), and "misses" (control rules that do not reduce exploitation as swiftly as stock status declines), based on social and economic consequences. Protection of species at risk will be highly intolerant of "misses", because delayed action may result in extinction. Maintaining a low miss rate necessarily requires accepting a higher rate of false alarms (Piet and Rice, 2004).
|
Comparison of triggers |
|---|
|
TopIntroductionThe origin of risk-of-extinction...GoalsComparison of triggersDiscussionReferences |
|---|
Table 1 presents the IUCN (2005) quantitative criteria for Numbers of Mature Individuals, Geographic Range, and Reduction in Population Size (Decline). The IUCN actually lists five criteria, with both numbers and range modified by considerations such as fragmentation into smaller separated subpopulations, recent or continuing downward trends in abundance, and area occupied. For the present purpose, these variants are captured in the two entries in each cell in Table 1 (the larger figure is a reference point when a population also shows the other features of concern; the smaller figure is an absolute trigger, regardless of other features of the population).
|
The LRPs for SSB (Blim, in t) for 56 stocks assessed by ICES (from the 2006 working group reports) were converted into the corresponding approximate numbers (Nlim) by examining weights, numbers, and maturation vectors-at-age, taking conservative values for average weights, and assuming age compositions for relatively old fish. Although the true numbers of mature fish corresponding to Blim are almost certainly larger than estimated here, the frequency distribution of these conservatively estimated Nlims, binned by arbitrary classes, shows that none of the commercial stocks would qualify even for the criterion Vulnerable (Figure 1). All Nlims appear to be larger by at least a factor of 100 than the trigger for conservation action using the IUCN Absolute Numbers criterion.
|
ICES does not apply spatial management reference points. The area of an ICES rectangle (1° longitude; 0.5° latitude) varies with latitude, but except in the far north, each rectangle is at least 3000 km2. Therefore, if a stock were restricted to fewer than (seven) rectangles, it could fit the IUCN criterion for vulnerable. A few stocks in the English Channel and the Kattegat/Skagerrak may fall below this criterion, with potentially the Blackwater and Clyde herring stocks even below the criterion for endangered stocks. However, for most ICES stocks, geographic ranges are far greater than the IUCN criteria.
Fisheries management jurisdictions do not use reference points expressed as percentage decline, but stock assessments routinely estimate population trajectories and allow the extent of a decline to be calculated. For all stocks assessed by ICES, North Atlantic Fisheries Organization (NAFO), Department of Fisheries and Oceans (DFO, Canada), and National Marine Fisheries Service (NMFS, US) with time-series of SSB going back to 1970 or earlier, we estimated the maximum extent of decline over a 15-y period as a coarse approximation of "three generations" (assuming an average age of a spawner of 5 y). For the few stocks of Sebastes or similar species with much greater ages-at-maturity, the percentage declines over up to 30 y were rarely noticeably higher than the greatest percentage declines over 15 y. Consistently applying a 15-y window probably more often overestimates than underestimates the lifespan of three generations, particularly for stocks that have been most heavily exploited.
Figure 2 shows that most fish stocks would qualify as at least Vulnerable at some stage of their exploitation history, even if it could be assumed that the causes of past declines are known and have ceased, whereas more than a quarter of stocks would qualify as Critically Endangered. Only in very few stocks contributing to this tabulation did the scientific advice recommend actions as extreme as closing directed fisheries; most recommended that harvests be presented as sustainable. Clearly, there is a conflict between how fisheries management control rules and the IUCN decline criterion guide management actions.
|
35 y) from ICES, NAFO, DFO, and NMFS. Arrows represent the percentage decline corresponding to IUCN categories: V, Vulnerable; E, Endangered; CE, Critically Endangered.It is no news that this conflict exists. The debate on the application of the decline criterion to exploited marine fish started in the 1990s, and led to an IUCN Working Group focusing on the issue. The conclusion (Mace, 1999) was that, "In its current formulation, Criterion A [decline] often overestimates extinction risk". The report recommended that the decline criterion should be kept, but adapted with several qualifiers, including whether or not the decline was "managed", were the causes "understood", and had they "ceased or reversed". The debate has continued in the literature, with several authors (Hutchings, 2001; Dulvy et al., 2003, 2004; Hutchings and Reynolds, 2004) presenting meta-analyses demonstrating that fish stocks meeting the decline criterion generally have not recovered, so concluding that the criterion was sound. Others (Matsuda et al., 1998; FAO, 2000; Powles et al., 2000; Mace, 2004) have countered that most stocks meeting the decline criterion do increase when fishing mortality really is reduced, and that false alarms are a serious management issue in jurisdictions where designating a species as being at risk of extinction results in automatic and severe restrictions.
Scientific debate is healthy, and with careful research it should be possible to resolve debates on purely scientific issues. In conducting a meta-analysis on a database of more than 80 stocks, we asked the question, if a stock meets one of the decline criteria, what is the likelihood that it is actually on a trajectory to extinction? To help answer this question, we used assessment-based estimates of SSB for the years following the year in which a decline criterion was triggered. Using standard likelihood methods, we calculated the probability of observing biomasses at least as large as those estimated in the assessment, assuming a "true" trajectory to zero in 100 y. If the stock was on or below the trajectory to zero, the probability of such observations would be high, and the designation of a high risk of extinction would be a "true positive". If the stock was well above the trajectory to zero, the probability would be low, and the designation would be a false alarm.
The approach is straightforward, but its implementation is problematic. Although the IUCN guidelines discuss issues of how to choose the pre- and post-decline periods, many alternative choices may be made that are consistent with the guidelines (as exemplified in Figure 3a for one stock selected). Also, in estimating the probability of post-decline observations given the hypothetical trajectory to zero, an expected variance has to be estimated based on historical data, predominantly from a period when the population was much larger. Given the range of possibilities, the most appropriate trigger should be selected based on robustness tests, not dissimilar to the practice common in selecting harvest control rules. This is another correspondence between the control rules applied in fisheries management and those for conservation of species at risk. Unfortunately, such tests quickly become part of the debate about the reliability of the criteria. Figure 3b illustrates the problem: if the point at which the 70%-decline criterion was first exceeded (1989) is used as the starting point for the post-decline trajectory, all subsequent observations lie far below the expected values, and the criterion can be considered to have correctly indicated a serious conservation issue. If the starting point is taken as 1993 (also satisfying the 70%-decline criterion), when decisive management action was taken (a drastic quota cut followed by a moratorium), then all subsequent observations are well above the expected values, and the criterion can be considered to have issued a false alarm. This pattern is common in all the species we analysed (Figure 4).
|
|
In the example (Figure 3), the case might be made that 1993 would be an appropriate starting point, because fisheries were closed that year. However, such decisive management benchmarks are rarely available, and decisions are made by managers who are inclined to consider their responses to scientific advice as "decisive", given the information available. Experts and managers concerned about false alarms may argue that a pre-selected 70% value is arbitrary, ignoring many circumstances. They may argue also that the nadir of the population trajectory is the time when advice was sufficiently decisive. Experts concerned about misses can argue that selecting the nadir of the population has two flaws. First, it invokes circular reasoning: the action was decisive only because the population happened to increase thereafter. Second, picking an especially low starting point necessarily biases the analysis towards finding larger populations in subsequent years, or the selected starting point would not have stood out as a point of decisive management action.
No amount of robustness testing can be expected to resolve all discrepancies between advice based on the decline criterion and advice based on harvest control rules. They will continue to perform in inconsistent ways, and management and policy will have to resolve the differences. Guidelines to deal with the caveats in the IUCN guidelines can help reduce these discrepancies to some extent, but they are proving hard to operationalize. For example, the IUCN caveats assign the burden of proof such that evidence must be provided that a decline is part of a natural fluctuation rather than a reduction caused by anthropogenic activity, before the criterion does not automatically trigger management action. Such a task will rarely be possible in real time for exploited stocks, especially when exploitation rates have been sufficiently high to have contributed to the decline. The help provided by the guidelines is also countered by the clear expression (under fisheries; IUCN, 2001, p. 33) of a high tolerance for false alarms, and aversion to misses: "Such listings should not be problematic in the medium- to long-term, because, if the fishery is managed successfully, although it currently exhibits symptoms consistent with endangerment, the population will eventually stabilize at a target level and the decline will end, such that the taxon would then no longer qualify for extinction".
|
Discussion |
|---|
|
TopIntroductionThe origin of risk-of-extinction...GoalsComparison of triggersDiscussionReferences |
|---|
When contrasting the performance of the triggers for evoking management action from the species at risk and fisheries perspectives, numbers and range complement each other rather than being in conflict. Reasonable fisheries management systems guided by typical biomass reference points should have reduced fishing mortality to essentially zero well before stock size would be qualified as being at risk of extinction. The geographical extent of most exploited stocks is also substantially greater than the ranges associated with this risk-of-extinction criterion. There are some exceptions which suggest that fisheries management should be particularly risk-averse for stocks with restricted ranges, but generally, these do receive close attention.
The decline criterion often conflicts with fisheries-management practices. This is likely to be the case even when management is based on well-tested control rules, and accepting of pre-specified risk tolerances. This discrepancy has been known for at least a decade, and has stimulated substantial research and analyses, but as yet no convergence of viewpoints is in sight. There are some useful lessons from the current analyses, however:
- the problem does not arise from different management objectives: the objectives are equally similar or different for abundance and range, yet those sets of triggers tend to perform harmoniously;
- pre-specification of control rules and risk tolerances does not ensure that all scientists and managers will agree on the outcomes: controversy can still rage if different groups of experts use different triggers, each group satisfied that the triggers have been thoroughly tested and perform reliably;
- narrative guidelines facilitate the application of sets of triggers, but are not sufficiently operational for all users to agree on a single approach as consistent with the intent: transformation into contingent triggers may help;
- even when there is agreement on goals and overall error rates in probability terms, the tolerances for misses and false alarms influences these debates greatly: if different experts, managers, or sectors of the public have different tolerances, there is potential for disagreement;
- until now, the discussion within ICES and elsewhere in fisheries science on the performance of harvest control rules has not focused on the balance of potential misses and false alarms. From experience with applying risk-of-extinction criteria to marine fish, such dialogue and agreement among users might well be warranted.
|
Acknowledgements |
|---|
We thank Carl O'Brien, Kjellrun Hiis Hauge, and Niels Daan for helpful comments, and André Punt and Howard Powles for penetrating questions during preparation of the manuscript.
|
References |
|---|
|
TopIntroductionThe origin of risk-of-extinction...GoalsComparison of triggersDiscussionReferences |
|---|
-
Baillie J., Groombridge B. 1996 IUCN Red List of Threatened Animals. (1996) Gland, Switzerland: IUCN.
CSAS. The northern Gulf of St Lawrence cod (3Pn4RS). In: CSAS Stock Status Report 2004/016 (2004) http://www.meds-sdmm.dfo-mpo.gc.ca/csas/applications/publications/publication_e.asp?year_selected=2004andseries=SSR.
Dulvy N. K., Sadovy Y., Reynolds J. D. Extinction vulnerability in marine fish populations. Fish and Fisheries (2003) 4:25–64.[CrossRef][Web of Science]
Dulvy N. K., Ellis J. R., Goodwin N. B., Grant A., Reynolds J. D., Jennings S. Methods of assessing extinction risk in marine fish. Fish and Fisheries (2004) 5:255–276.[CrossRef][Web of Science]
FAO. An appraisal of the suitability of the CITES criteria for listing commercially exploited marine species. FAO Fisheries Circular (2000) 954:66.
Hutchings J. A. Conservation biology of marine fish: perceptions and caveats regarding assignment of extinction risk. Canadian Journal of Fisheries and Aquatic Sciences (2001) 58:108–121.
Hutchings J. A., Reynolds J. D. Marine fish population collapses: consequences for recovery and extinction risk. Biosciences (2004) 53:297–309.
ICES. Report of the Advisory Committee on Fisheries Management. (2005) 1. Advisory Committee on Ecosystems, and Advisory Committee on the Marine Environment. 140. ICES Advice 2005.
ICES. Report of the Study Group on Fisheries Management Strategies. (2006) 19. ICES Document CM 2006/ACFM.
IUCN. IUCN Red List Criteria review provisional report: draft of the proposed changes and recommendations. Species (1999) 31/32:43–57.
IUCN. The IUCN Red List Categories and Criteria (Version 3.1). (2001) Gland, Switzerland and Cambridge, UK: IUCN Species Survival Commission. IUCN. 30.
IUCN. Guidelines for Using the IUCN Red List Categories and Criteria. (2005) Gland, Switzerland and Cambridge, UK: IUCN Species Survival Commission. IUCN. 52.
Kell L. T., Mosqueira I., Grosjean P., Fromentin J-M., Garcia D., Hillary R., Jardim E., et al. FLR: an open source framework for the evaluation and development of management strategies. ICES Journal of Marine Science (2007) 64:640–646.
Ludwig D. M. Uncertainty and the assessment of extinction probabilities. Ecological Applications (1996) 6:1067–1076.[CrossRef][Web of Science]
Mace G. M. IUCN Criterion Working Group: Report of the Marine Workshop. (1999) http://app.iucn.org/webfiles/doc/SSC/RedList/CriteriaReview_Marine_Workshop.pdf.
Mace P. M. In defense of fisheries scientists, single-species models, and other scapegoats: confronting the real problems. Marine Ecology Progress Series (2004) 274:285–291.[Web of Science]
Mace G. M., Stuart S. N. Draft IUCN Red List Categories. Species (1994) 21/22:13–24. version 2.2.
Matsuda H., Yahara T., Uozumi Y. Extinction risk assessments of declining wild populations: the case of the southern bluefin tuna. Research in Population Ecology (1998) 40:271–278.
Piet G. J., Rice J. C. Performance of precautionary reference points in providing management advice on North Sea fish stocks. ICES Journal of Marine Science (2004) 61:1305–1312.
Powles H., Bradford M. J., Bradford R. G., Doubleday W. G., Innes S., Levings C. D. Assessing and protecting endangered marine species. ICES Journal of Marine Science (2000) 57:669–676.
Ulrich C., Andersen B. S., Sparre P. J., Nielsen J. R. TEMAS: a fleet-based bio-economic simulation software for management strategies accounting for fishers' behaviour. ICES Journal of Marine Science (2007) 64:000–000.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||