Saturday, August 9, 2014

King Crab Decline Examined

Review of “King Crabs of the World”
Review by Braxton Dew
NMFS, Ret.

The Elephant In the Room

Alaskans who pick up Brad Stevens’ book, King Crabs of the World, are likely to browse through its pages looking for answers to the question: “Whatever happened to the king crab?”. If they make it to page 589, readers will be disappointed to hear editor Stevens’ say, “…. We will never know what caused the dramatic changes in the abundance of Bering Sea RKC after 1978”. One of the reasons we will never know is that this book studiously avoids taking an honest look at the elephant in the room — the breaching of red king crab no-trawl sanctuaries by commercial groundfish trawling during the late 1970s and early 1980s.

The central theme running through the book is that adult  king crab are immeasurably more likely to be killed by natural causes, including a slight rise in sea-surface temperature, than to be killed by being dragged up in a bottom trawl (in which chances of survival are said to be good). The stated opinion of editor Stevens (p.369), is that crab bycatch mortality in groundfish trawl fisheries  is low. According to Stevens, critics’ claims to the contrary verge on the fantastic and are often tinged with prejudice. The bottom line is that bycatch “probably represents only a small fraction of overall king crab mortality” (p.xviii).

Natural Mortality Is the Only Mortality

In case there is any confusion as to the uber-dominant role of natural mortality, authors Zheng and Kruse in their management model (p.530) categorize all trawl-killed crab as having succumbed to natural mortality. Using this slight-of-hand, any fishing  mortality spikes caused by commercial trawling in the newly breached sanctuaries were interpreted as natural deaths. Zheng and Kruse offer the usual laundry list of natural-mortality elements — predation, cannibalism, senescence, and disease — but are not forthcoming as to the model’s inclusion of trawl bycatch as a source of “natural” mortality. The Zheng-Kruse Bering Sea red king crab model may be the only fisheries model in the world that so profoundly conflates man’s fishing (F) with natural mortality (M). Go to the link below to see an example of “natural” mortality according to the Zheng-Kruse model. No wonder the authors had to inflate M (by nearly 10X) during 1980-1984 (p.530).

Where Did They Go?

The beat goes on in Chap.3, p.48: “…. ocean-climate shifts in the Bering Sea from 1975 to 2001 led ….  to a shift in the center of abundance from the southwest part of the bay in the 1970s to the center of the bay in the 1980s.” The Chap.3 authors carelessly cite Loher and Armstrong (2005) for their authority. But Loher and Armstrong (2005) say: “The hypothesis that ovigerous crabs actively move to avoid cold conditions does little to explain the disappearance of the southwestern component of the population. Between 1976 and 1981, broodstock abundance declined by ~94% in the Unimak region, representing nearly 80 million crabs.”

What, then, could explain the spatially explicit disappearance of red king crab from the southwestern sector of their Bristol Bay range, as noted by Loher and Armstrong? Below is a Bering Sea trawling-distribution map that may help interested readers form their own hypothesis as to what might have happened to crab in the southwestern sector of their range.

 The trawl-density distribution of observed tows and observed joint-venture deliveries within 25-km2 grid squares during 1973–2001. The white boundary line shows the old (1959-1977) Japanese Broodstock Sanctuary (Dew 2010b). Trawling intensity is highest in the Unimak area and relatively low in the more offshore area to the northeast, where the present-day broodstock remains (Fig.3, Dew and McConnaughey, 2005)

A glance at the Bering Sea trawl distribution will enable many readers to come up with a theory more credible than that advanced by editor Stevens (p.588) that the crab, whose population had grown in an unrestrained manner to unsustainable heights, essentially did it to themselves. According to Stevens, the height from which RKC fell was an abundance level that was “anomalous”, resulting from a supportive environment and lower fishing levels, which have never reoccurred. “What we should wonder about is not so much that the abundance …. declined but that they ever reached such a high abundance level in the first place. Even without fishing it was ecologically unsustainable and would have declined due to natural mortality ….”. In sum, Stevens goes on, the decline of adult red king crab should not have been a surprise to anyone (p.587).

Improbable Population Trend Is Suspect

Editor Stevens is right about one thing, though: we should have wondered — not about the crab but about the NMFS population estimates, which were increasingly biased during 1975-1980 by an inexplicable number of extra, non-design-based tows made at high-abundance, male-dominant stations (Dew 2010a). Too often the NMFS estimates are published and accepted as if they are free of bias or sampling error. But because of the patchy spatial distribution of podding crab, each year’s estimate of red king crab abundance is associated with a large measure of uncertainty. Figure 17.5 of Chap.17 (Zheng and Kruse) shows population estimates surrounded by confidence intervals, many of which are at least ±30-60% of the mean, with some approaching or exceeding ±100% (e.g., 1984 and 1991). The ±80-90% uncertainty surrounding the 1991 estimate, for example, was caused by a single survey tow that hit a large pod of adult legal male crab on the “bachelor grounds” (Dew 2008).

An assumption of assessment surveys is that the observed trend in annual survey estimates is reasonably well-correlated with a real abundance trend in the underlying population. However, if confidence intervals around the annual estimates are wide enough (i.e., if the uncertainty of the survey esti­mates is great enough), the survey-derived trend may bear little or no resemblance to the actual abundance changes in the population. The high level of imprecision of NMFS king crab survey estimates, combined with bias from ad hoc extra sampling at high abundance stations, raises the question whether the 1975-1980 data can be used to define a trend, or whether such a trend tells us much, if anything, about actual abundance changes in the underlying population. My own statistical examination reveals that there were no significant (< 0.05) differences among the consecutive six years, 1975 through 1980, nor were there significant differences among any of the 15 year-pairs (e.g., 1975 vs. 1978, a period during which the population was thought to increase by >100%). Thus, it is likely that the “anomalous” and “ecologically unsustainable” abundance spike reportedly achieved by Bristol Bay red king crab was nothing but a mirage — the product of an imprecise survey biased by repeated, nonrandom sampling in areas of high legal-male crab abundance. For more details, click on the link below to Podding Behavior of Adult King Crab and Its Effect on Abundance-Estimate Precision:

NMFS Warned Us

In an effort to show that the most spectacular crash in the history of U.S fisheries management did not blindside NMFS officials, Stevens (p.587) claims that, based on “warnings issued by the NMFS” (no documentation as to what warnings), the Bering Sea collapse “should not have been a surprise to anyone”. Stevens then erroneously uses Orensanz et al. (1998) as support, stating (p.588) that exploitation of the legal male stock rose to >70%, sex ratios became highly skewed in favor of females (M:F = 1:6), and the occurrence of barren females increased dramatically. Stevens goes on to say, ”Clearly, this was an unsupportable level of effort, and under such pressure it is no surprise that the [Bering Sea] crash occurred so quickly”.

The problem with all of this is that Orensanz et al. (1998) did not write about Bering Sea crab. The Orensanz warning signs reiterated by editor Stevens were for Kodiak, not Bering Sea king crab. As for Bering Sea warnings, instead of Orensanz we had Otto (1985) declaring that Bering Sea imbalances in sex ratio have been negligible and have placed no constraints on reproductive success; and Otto (1986) specifying that the “Bristol Bay catch never exceeded 60% of the estimated legal stock”. Both of these pieces of information later turned out to be incorrect (Dew and McConnaughey 2005). Moreover, because the NMFS survey is conducted during king crab spawning and spawning occurs well inshore of the survey-sampling boundary, investigators have been unable to reliably monitor mating success (barren females) or king crab sex ratios in the Bering Sea (Dew 2008). It is simply not credible to imply that NMFS or any other organization was savvy enough to warn us of the impending dismantling of Alaska’s most valuable fishery.

Ex-Managers In Denial

On p.395 Stevens says the question of whether or not trawling contributed to the decline of Bering Sea king crab stocks in the early 1980s is “debatable”. However, Stevens’ book presents little or no debate. Instead, there are denials (p.128) that the work of others is valid. For example, on p.931 of their paper, Dew and McConnaughey (2005) stated that the rate of decline of broodstock density with trawling in the Unimak area was 1.9 times that of a more lightly trawled offshore area, and that these rates of decline were statistically different (P < 0.05). This is not a startling finding — more trawling, less crab remain on the grounds. No one fabricated data — the crab densities and the number of trawl drags came from NMFS. Yet author Otto, without presenting any data or analysis of his own, flatly denies that there was any statistically discernible effect of trawling on the rates of decline in crab abundance within the two areas.

Quibbles About Survival

Why does author Otto bother trying to discredit an unsurprising revelation that trawling captures crab and more trawling captures more crab? Consider the following from Dew and McConnaughey (2005, pp.931-932):

“Given the rate and magnitude of these declines, and considering that two different levels of trawling activity in the Unimak and offshore areas each moved the local population quickly toward extinction, it is apparent that the survival of discarded, trawl-caught crab is negligible. That is, the long-term survival of red king crab affected by commercial trawling is insufficient to offset the mortality and social disruption caused by past and present levels of trawling”.

This conclusion renders irrelevant the results from myriad, ever-ongoing survival studies reported in this book and elsewhere. Moreover, it negates unwarranted optimism as to the survival of trawl-caught crab (e.g., p.128). After all, when the crab are gone, how relevant are quibbles about survival?

Male Reproductive Potential — Chronic Overestimation?

Chap.10, Reproductive Ecology.  Whenever the management of a major fishery such as Alaska’s king crab fishery is described by reputable scientists as a “failure” (Wooster 1992) or a “debacle” (Orensanz et al. 1998), it is probably time to reassess key management assumptions. A present assumption needing further scrutiny is that mature male king crab, in the wild, participate in mating shortly after molting. According to Powell et al.(1973) and Powell et al. (2002), caged males can engage in mating after a post-molt recovery period of as little as ten days. Compare this with Paul et al. (1995), who determined that the molt-recuperation period for male Tanner crab was at least 100 days; and Sainte-Marie et al. (1999, 2002), who reported that male snow crab did not mate in the same year they molted. Going further, the Powell et al. assertion rests on the shaky bedrock of yet another assumption — that investigators can accurately determine the time since last molt, in months, by visual inspection of a wild crab’s shell. Countering all this is a substantial body of evidence (link below) that wild king crab do not molt and mate in the same year

Author Webb’s chapter on reproduction could have benefitted from a more detailed discussion of this issue, given that his chapter includes sections on “Molting and Mating”, “Mating Ability” and “Male Reproductive Potential” (MRP). Although Takeshita et al. (1990) are cited for their tagging work on king crab migration, Webb omits the fact that this same tagging data led Takeshita et al. to conclude that only 50% of Bristol Bay’s mature male population, mostly old-shell skip-molts, participate in mating in any given year. The other 50% remains offshore on “bachelor grounds”, far from known spawning grounds, recuperating from the winter molt (e.g., Dew 2008).

The answer to the question of whether or not all mature males participate each year in mating has a profound effect on the management of a male-only fishery. For example, if Takeshita et al. (1990) are correct, then MRP values calculated by the model during the past 25 years are likely to have been overestimated by a factor of 2X (only half of the mature males mate each year, while the other half is molting). Furthermore, if males do not mate in the same year they molt, assurances (p.300) that males will be able to mate once or twice before their harvest are little more than wishful thinking. It is not good science to ignore reasonable, well-published, data-based hypotheses, either in the management of a stock or the writing of a book chapter. A better idea might be to provide evidence, if available, that disproves such hypotheses. Also, with regard to the presumption that there is some precise correspondence between shell condition and shell age, it would be prudent to consider the possibility that, although all new shells are clean, not all clean shells are new.

What Is Depensation?

In 1966, Kodiak’s Gulf of Alaska (GOA) fishery provided the bulk of the U.S. red king crab catch (70,000 tons vs. 450 tons from the Bering Sea). We might expect the dismaying and complete loss of the GOA fishery (commercially extinct now for more than 30 years) to have stimulated critical examination of a core management assumption — that calling a halt to fishing will permit an overfished stock to recover from low abundance levels. This assumption relies on the idea of compensatory density-dependence, which is the mechanism underlying all logistic-growth models, including the Ricker stock-recruitment curve — a mainstay of Bristol Bay king crab management. In addition to positing reduced per capita production at high stock sizes, the Ricker compensatory model incorporates the notion that at extremely low population levels — indeed, when the population is on the verge of extinction — the per capita production of recruits is highest.

The compensatory model makes no allowance for populations whose reproductive success, over some range of population size, increases with the size of its aggregations. Nor is this model relevant for populations whose per capita growth rate goes to zero well before the population size goes to zero. Editor Stevens asks (p.588) why did king crab not rebound when fishing was halted? The answer is that managers are using a model, developed for salmon, that does not work for red king crab. That is, they are using a model that does not allow for depensation, does not make provision for an Allee effect, and does not account for declining individual fitness and per capita production as the population nears extinction — all of which are earmarks of Alaska’s highly gregarious red king crab populations. Both the Bristol Bay red king crab and the North American passenger pigeon have commonalities that conform to the checklist below. For more on depensation and environmental conditioning, click on the following link to Historical Perspective On Habitat Essential To Bristol Bay Red King Crab:

Literature Cited

Dew, C.B. 2008. Red king crab mating success, sex ratio, spatial distribution, and abundance estimates as artifacts of survey timing in Bristol Bay, Alaska. North American Journal of Fisheries Management 25:1618-1637.

Dew, C.B. 2010a. Podding behavior of adult king crab and its effect on abundance-estimate precision. Pages 129-151 in Biology and Management of Exploited Crab Populations under Climate Change. Alaska Sea Grant, AK-SG-10-01, Fairbanks. 562 pp.

Dew, C.B. 2010b. Historical perspective on habitat essential to Bristol Bay red king crab. Pages 377-402 in Biology and Management of Exploited Crab Populations under Climate Change. Alaska Sea Grant, AK-SG-10-01, Fairbanks. 562 pp.

Dew, C. B., and R. A. McConnaughey. 2005. Did trawling on the broodstock contribute to the collapse of Alaska’s king crab? Ecological Applications 15:919–941.
Liermann, M., and R. Hilborn. 2001. Depensation: evidence, models and impli­cations. Fish. 2:33-58.
Loher, T., and D.A. Armstrong. 2005. Historical changes in the abundance and distribution of ovigerous red king crabs (Paralithodes camtschaticus) in Bristol Bay (Alaska), and potential relationship with bottom temperature. Fisheries Oceanography 14:292-306.

Orensanz, J.M., J. Armstrong, D. Armstrong, and R. Hilborn. 1998. Crustacean resources are vulnerable to serial depletion: the multifaceted decline of crab and shrimp fisheries in the greater Gulf of Alaska. Rev. Fish Biol. Fish. 8:117-176.

Otto, R.S. 1985. Management of Alaskan king crab stocks in relation to the possible effects of past policies. Pages 447–481 in B. R. Melteff, coordinator. Proceedings of the
international king crab symposium. Report 85-12. University of Alaska Sea Grant Program, Anchorage, Alaska, USA.

Otto, R.S. 1986. Management and assessment of eastern Bering Sea king crab stocks. Pages 83-106 in G.S. Jamieson and N. Bourne (eds.), North Pacific workshop on stock assessment and management of invertebrates. Canadian Special Publication of Fisheries and Aquatic Science 92.

Paul, A.J., J.M. Paul, and W.E. Donaldson. 1995. Shell condition and breeding success in Tanner crabs. Journal of Crustacean Biology15(3):476-480.

Powell, G. C., B. Shafford, and M. Jones. 1973. Reproductive biology of young adult king crabs, Paralithodes camtschatica (Tilesius), at Kodiak, Alaska. Proceedings of the National Shellfisheries Association 63:77–87.

Powell, G.C., D. Pengilly, and S.F. Blau. 2002. Mating pairs of red king crabs (Paralithodes camtschatica) in the Kodiak Archipelago, Alaska, 1960-1984. Pages 225-245 in Crabs in cold water regions: biology, management, and economics. University of Alaska Sea Grant, AK-SG-02-01, Fairbanks. 876 pp.

Sainte-Marie, B., J.-M. Sevigny, and M. Carpentier. 2002. Interannual variability of sperm reserves and fecundity of primiparous females of the snow crab (Chionoecetes opilio) in relation to sex ratio. Canadian Journal of Fisheries and Aquatic Sciences 59:1932–1940.

Sainte-Marie, B., N. Urbani, J.-M. Sevigny, F. Hazel, and U. Kuhnlein. 1999. Multiple choice criteria and the dynamics of assortive mating during the first breeding season of female snow crab, Chionoecetes opilio (Brachyura, Majidae). Marine Ecology Progress Series 181:141–153.

Takeshita, K., H. Fujita, and S. Matsuura. 1990. A note on population structure in the eastern Bering Sea adult red king crab, Paralithodes camtschatica. Pages 427–433 in Proceedings of the International Symposium on King and Tanner Crabs. University of Alaska Sea Grant, AK-SG-90-04, Fairbanks.

Wilson, E.O. 1975. Sociobiology: the new synthesis. Belknap Press, Cambridge. 697 pp.

Wooster, W.S. 1992. King crab dethroned. Pages 14-30 in M.H. Glantz, editor. Climate, variability, climate change, and fisheries. Cambridge University Press, New York.