1. Department of Biology, Dalhousie University, Halifax , Nova Scotia B3H 4J1, Canada

Correspondence to: Jeffrey A. Hutchings Correspondence and requests for materials should be addressed to J.A.H. (e-mail: Email: jhutch@mscs.dal.ca).

Worldwide overfishing has raised concerns that extraordinary collapses in abundance may significantly increase the extinction probability of targeted⁴ (for example, Atlantic cod, Gadus morhua) and incidentally harvested⁸ (for example, barndoor skate, Raja laevis) marine fishes^{5, 6}. This is reflected by the work of national and international agencies responsible for assigning risk categories to potentially endangered species. Present and past fisheries for Atlantic cod, whose over-exploitation and collapse have been well documented^{2, 9, 10, 11}, provide one example. On the basis of global estimates of decline, the International Union for Conservation of Nature (IUCN) listed⁴ Atlantic cod as 'vulnerable' in 1996, the same status applied 2 years later by the Committee on the Status of Endangered Wildlife in Canada (COSEWIC)¹². Other marine fishes assigned to risk categories by these agencies include the Pacific sardine (Sardinops sagax; Clupeidae)¹², haddock (Melanogrammus aeglefinus; Gadidae)⁴, Bering wolffish (Anarhichas orientalis; Anarhichadidae)¹², and the pleuronectid flatfishes Atlantic halibut (Hippoglossus hippoglossus ) and yellowtail flounder (Pleuronectes ferrugineus)⁴.

There are, however, concerns that quantitative listing criteria based on temporal trends in abundance, such as those applied by IUCN¹³ and COSEWIC¹⁴, may significantly over-estimate extinction threats to marine fishes and should be modified to account for the great natural variability in abundance, high reproductive potential and remarkable ability to recover from population collapse perceived to be characteristic of marine fishes, relative to other taxa^{6, 7}. Theoretical analyses also suggest that there is nothing intrinsic to the population dynamics of exploited marine fishes that would prevent them from increasing at low population sizes¹, although the assumption that their per capita reproductive success increases at low population levels may not be as general as previously thought³.

I made use of the most comprehensive numerical fisheries data base available, maintained by R. A. Myers, Department of Biology, Dalhousie University, Halifax, Canada, at <http://fish.dal.ca/welcome.html>. I recorded the largest 15-year percentage decline in mature fish biomass experienced by each stock and subsequent population sizes 5, 10 and 15 years thereafter. The 15-year interval was considered short enough to obtain a reasonably large sample of populations, and long enough to be biologically meaningful, approximating the three-generation time period specified by quantitative at-risk criteria used by species-listing organizations such as IUCN and COSEWIC.

After a decline, any increase in population size, N, could be interpreted as some sort of recovery. Graphically, recovery t years after a 15-year decline can be determined from a plot of N_t+15/ N₀ on the ordinate against magnitude of population decline on the abscissa, that is, 1 - N₁₅/N₀. On such a plot, populations exhibiting no recovery, that is, N _t+15 = N₁₅, would fall on a straight line with slope of -1 extending from (0,1) to (1,0). Those continuing to decline would fall below this line, whereas those exhibiting some recovery would fall above this line. Similarly, populations falling on the line with slope of 0 extending from (0,1) to (1,1) would be said to have fully recovered, that is, N _t+15 = N₀, whereas those above this second line would have increased relative to the size from which their population declines had begun.

Among those for which data were available, 90 marine fish stocks (representing 38 species among 11 families) experienced 15-year declines of 13 to 99%, followed by 5-year changes in population size ranging from 0.3 to 178% of the size from which the declines began. Of these 90 stocks, 37 (41%) continued to decline after the 15-year period, 46 (51%) exhibited some recovery, and 7 (8%) had fully recovered (Fig. 1). Subsequent population size was negatively correlated with magnitude of population decline ( Fig. 1, Table 1; correlation coefficients will have been over-estimated slightly because of autocorrelated data).

Figure 1: Bivariate association between population decline and subsequent population size for 90 marine fish stocks.

The ordinate refers to the size of a population 5 years after the proportionately largest 15-year decline experienced by that population, relative to its size at the beginning of its 15-year decline. Populations that experienced some recovery are represented by points to the right of the solid line. Fully recovered stocks are represented by points on and above the dashed line. One datum—a 69% population decline of one clupeid followed by a 1.78 recovery—has been omitted for clarity. Slanted crosses, Engraulidae; upward triangles, Clupeidae; downward triangles, Osmeridae; filled triangles, Gadidae; stars, Scorpaenidae; upright crosses, Anoplopomatidae; filled diamonds, Sparidae; diamonds, Nototheniidae; filled squares, Scombridae; filled circles, Pleuronectidae; circles, Soleidae.

High resolution image and legend (36K)

Table 1: Linear regressions for associations between recovery and decline in marine fishes

Full table

Population sizes 5 years thereafter were significantly correlated with population declines within most of the six numerically dominant families in the analysis (representing 83 of the 90 stocks; Fig. 2, Table 1): Clupeidae (for example, herring, Clupea harengus ; sprat, Sprattus sprattus), Gadidae (for example, cod, haddock), Scombridae (for example, mackerel, Scomber spp.; tuna, Thunnus spp.), Sparidae (for example, snapper, Pagrus auratus), Scorpaenidae (for example, redfish, Sebastes spp.), and Pleuronectidae (for example, plaice, Pleuronectes spp.; Greenland halibut, Reinhardtius hippoglossoides ). The association was not significant within the clupeids or the scombrids, although the correlation was highly significant within the latter family when the outlier stock (eastern Pacific yellowfin tuna, T. albacares) was removed from the analysis. Among families, clupeids (80% of the stocks) were most likely to have experienced some level of recovery (Fig. 2).

Figure 2: Population recovery within and among six families of marine fishes 5 (triangles), 10 (filled triangles) and 15 (stars) years after the greatest proportionate 15-year decline experienced by each stock.

The dashed and solid lines represent the 'full recovery' and 'no recovery' lines, respectively.

High resolution image and legend (69K)

Data 10 and 15 years subsequent to the 15-year population declines were available for 45 and 25 stocks, respectively, comprised primarily of clupeids and gadids (69 and 84%, respectively). Among all stocks, the magnitude of decline negatively influenced population size 10 years, but not 15 years, after the declines (Fig. 3, Table 1 ). However, when clupeid data were excluded, population decline was strongly and significantly associated with both 10- and 15-year population recovery sizes among non-clupeids, notably gadids. Indeed, 15 years after their declines, 12% of marine stocks (all clupeids) had exhibited full recovery, whereas 40% (primarily gadids, but some clupeids) had experienced no recovery at all (Fig. 3). Furthermore, when clupeids were excluded from the analyses the regression coefficients for the 5- and 15-year recovery intervals (Table 1) were very close to the no-recovery line intercept and slope of 1 and -1, respectively, emphasizing the limited recoveries experienced by most marine fishes.

Figure 3: Population recovery of marine fishes 10 (upper panel) and 15 years (lower panel) after the greatest proportionate 15-year decline experienced by each stock.

The dashed and solid lines represent the 'full recovery' and 'no recovery' lines, respectively. The families corresponding to each symbol are those given in the caption to Fig. 1. a, 10 years; b, 15 years after decline.

High resolution image and legend (61K)

These data suggest that, after prolonged decline, clupeids are more likely to recover to previously experienced population sizes and are more resilient than other marine fishes. Such an increased rate of recovery may be attributable to the younger age at which clupeids mature relative to gadids, scorpaenids, scombrids, sparids and pleuronectids¹⁵, and the higher intrinsic rate of increase that earlier maturity generally effects¹⁶. Higher reproductive rates may also mitigate the negative influence of environmental stochasticity on the persistence of populations at small sizes^{17, 18}. In addition, being at a lower trophic level than other families considered here, clupeids may be better able to 'track' temporal and spatial fluctuations in primary and secondary productivity.

The apparent increased resilience of many clupeids relative to other stocks, however, may reflect different management responses to population declines and the different species selectivities of fishing gear. Clupeids are typically fished by deploying purse seines or mid-water trawls on schools identified visually or acoustically. The species uniformity of clupeid schools results in bycatches of incidentally harvested fishes one to two orders of magnitude lower than those of the bottom-deployed seines and trawls used to catch groundfish¹⁹. Thus, it may be comparatively easy to eliminate fishing mortality on affected clupeid stocks because of the high species selectivity of clupeid fishing technology. By contrast, the collapse of a groundfish stock rarely results in cessation of bottom-trawling in the affected region, meaning that fishing mortality on the collapsed stock can be reduced, but rarely eliminated, because of the comparatively low species selectivity of the fishing gear used to capture a broad diversity of demersal marine fishes.

The failure of many marine fish stocks to recover rapidly to former levels of abundance might arguably^{6, 7} be attributed to management strategies designed to maintain populations at 50% of their virgin biomass to maximize sustainable yields²⁰. But this seems a highly improbable explanation for the present observations, given the lengthy histories of direct and indirect exploitation^{21, 22} that have preceded the formal collection of data on fish numbers and biomass by fishery management agencies. In addition to the proposed influences of age at maturity, reproductive rate and fishing gear selectivity on recovery rates, it seems likely that ecosystem-level consequences of exploitation, for example, food webs altered by changes to species community structure, are also important^{23, 24}.

The suggestion that marine fishes should be exempt from existing quantitative criteria used to assign extinction risk^{6, 7} would be inconsistent with a precautionary approach to fisheries management and to the conservation of marine biodiversity. This is particularly important if, rather than providing estimates of extinction probability, the primary utility of at-risk designations lies in their reflection of the likelihood that species or populations will recover to former levels of abundance. I have found that 5–15 years after 15-year declines of 50 and 80% (the three-generation thresholds for the IUCN's 'endangered' and 'critically endangered' categories), gadid and other non-clupeid populations, on average, have increased marginally or not at all. Thus, although the effects of overfishing may indeed be generally reversible¹, the time required for population recovery in many marine fishes appears to be considerably longer than previously believed.

References

1. Myers, R. A., Barrowman, N. J., Hutchings, J. A. & Rosenberg, A. A. Population dynamics of exploited fish stocks at low population levels. Science 269, 1106–1108 ( 1995). | ISI | ChemPort |
2. Cook, R. M., Sinclair, A. & Stefànsson , G. Potential collapse of North Sea cod stocks. Nature 385, 521–522(1997). | Article | ISI | ChemPort |
3. Leirmann, M. & Hilborn, R. Depensation in fish stocks in a hierarchic Bayesian meta-analysis. Can. J. Fish. Aquat. Sci. 54, 1976–1984 (1997). | Article |
4. The International Union for Conservation of Nature(IUCN). IUCN Red List of Threatened Animals (IUCN, London, 1996).
5. Carlton, J. T., Geller, J. B., Reaka-Kudla, M. J. & Norse, E. A. Historical extinctions in the sea. Annu. Rev. Ecol. Syst. 30, 515–538 (1999). | Article | ISI |
6. Hudson, E. & Mace, G. (eds) Marine Fish and the IUCN Red List of Threatened Animals (Zoological Society of London, London, 1996).
7. Musick, J. A. Criteria to define extinction risk in marine fishes. Fisheries 24, 6–12 (1999 ). | ISI |
8. Casey, J. M. & Myers, R. A. Near extinction of a large, widely distributed fish. Science 228, 690– 692 (1998). | Article |
9. Hutchings, J. A.& Myers, R. A. What can be learned from the collapse of a renewable resource? Atlantic cod, Gadus morhua, of Newfoundland and Labrador. Can. J. Fish. Aquat. Sci. 51, 2126–2146 (1994). | ISI |
10. Myers, R. A., Hutchings, J. A. & Barrowman, N. J. Hypotheses for the decline of cod in the North Atlantic. Mar. Ecol. Prog. Ser. 138, 293– 308 (1996). | ISI |
11. Sinclair, A. F.& Murawski, S. A. in Northwest Atlantic Groundfish: Perspectives on a Fishery Collapse (eds Boreman,J., Nakashima, B. S., Wilson, J. A. & Kendall, R. L.) 71– 93 (American Fisheries Society, Bethesda, Maryland, 1997).
12. Shank, C. C. The Committee on the Status of Endangered Wildlife in Canada (COSEWIC): a 21-year retrospective. Can. Field-Nat. 113, 318–341 (1999). | ISI |
13. IUCN Species Survival Commission. IUCN Red List Categories (IUCN, Cambridge, 1999).
14. Campbell, R. R. Rare and endangered fishes and marine mammals of Canada: COSEWIC Fish and Marine Mammal Subcommittee status reports: X. Can. Field-Nat. 110, 454–461 (1996). | ISI |
15. Froese, R. & Pauly, D. (eds) FishBase 99 [online] (cited18 Feb 00) http:www.fishbase.org (1999).
16. Roff, D. A. The Evolution of Life Histories: Theory and Analysis (Chapman &Hall, New York, 1992).
17. Lande, R. Risks of population extinction from demographic and environmental stochasticity and random catastrophes. Am. Nat. 142, 911 –927 (1993). | Article | ISI |
18. Sæther, B. -E., Engen, S., Swenson, J. E. , Bakke, Ø. & Sandegren, F. Assessing the viability of Scandinavian brown bear, Ursus arctos, populations: the effects of uncertain parameter estimates. Oikos 83, 403– 416 (1998).
19. Alverson, D. L., Freeberg, M. H., Murawski, S. A. & Pope, J. G. A global assessment of fisheries bycatch and discards. FAO Fish. Tech. Pap. 339, 1–233 ( 1994).
20. Hilborn, R. & Walters, C. Quantitative Fisheries Stock Assessment. (Chapman & Hall, New York, 1992).
21. Hutchings, J. A.& Myers, R. A. in The North Atlantic Fisheries: Successes, Failures, and Challenges (eds Arnason,R. & Felt, L.) 38–93 (Institute of IslandStudies, Charlottetown, Canada, 1995).
22. Smith, T. D. Scaling Fisheries: The Science of Measuring the Effects of Fishing, 1855–1955 (Cambridge Univ. Press, Cambridge, UK, 1994).
23. Pauly, D., Christensen, V., Dalsgaard, J., Froese, R. & Torres, F. Fishing down marine food webs. Science 279, 860–863 ( 1997). | Article | ISI |
24. Jennings, S. & Kaiser, M. J. The effects of fishing on marine ecosystems. Adv. Mar. Biol. 34, 201– 352 (1998). | ISI |

Acknowledgements

This work was supported by a Natural Sciences and Engineering Research Council of Canada Research Grant. I thank B. Hall, I. McLaren, J. Shepherd and D. Swain for their helpful comments on previous versions of the manuscript.