Status of Stocks & Multispecies Assessment Program
Integration of Environmental Information in Fisheries Management Advice
SSMA scientists Paul Spencer and Jim Ianelli were invited to present case studies at the International Council for the Exploration of the Sea (ICES) headquarters in Copenhagen, Denmark on 18–22 June 207. The meeting was titled “Workshop of the Integration of Environmental Information into Fisheries Management Strategies and Advice” with the following objectives:
estimate the consequences of environmental variability (including “regime shifts”) for the biological reference points and other measures which are currently used to guide fisheries management
carry out analyses and formulate short-, medium-, and long-term integrated advice for the selected cases and compare this with traditional methodology
bearing in mind possible fisheries and ecosystems objectives, identify, develop, and evaluate procedures for improving fisheries management strategies and advice by including environmental information
identify future directions and needs, including operability, to bring forward the process of incorporating ecosystem advice.
Four main categories for evaluations were along productive capacities related to somatic growth, recruitment, survival, and spatial patterns. In addition to the eastern Bering Sea pollock and flatfish presentations given by REFM staff, the species/stocks covered for other case studies included: Greenland cod, North Sea cod, Northeast Atlantic sardine, North Sea herring, Northeast Arctic cod, Baltic Sea cod, Baltic Sea sprat, Bay of Biscay anchovy, California sardine, and South African anchovy.
Summaries of the case studies presented are as follows.
Case study of eastern Bering Sea Pollock: impact of environmental variability on management advice
Advice to managers is adjusted to account for environmental conditions for eastern Bering Sea pollock. Formally, the adjustments occur through the added stochastic components of environmentally driven demographic shifts (e.g., mean weight at age) and through the shape of the assumed stock-recruitment relationship. The latter is justified based on analyses of stomach content data coupled with models to show that cannibalism has an impact on subsequent recruitment levels.
The stochastic components (measurement errors together with environmentally driven process errors) affect quota recommendations by increasing the uncertainty of the Fmsy estimates. Higher uncertainty in Fmsy estimates result in lower values for quota recommendations due to analyses (shown in the original amendment to the Fishery Management Plan) that have determined that the harmonic mean value is appropriately risk-averse.
Explicit qualitative adjustments to the pollock quota recommendations are done through the process communicating apparent ecosystem changes (e.g., increased predation by arrowtooth flounder for Gulf of Alaska pollock and the lack of acoustic back scatter seen in the Bering Sea during summer survey of 2006). The latter effectively lowered the ABC recommendation by 118,000 t.
Environmentally affected components of stock productivity that were reviewed included recruitment, growth, mortality and the spatial distribution of the stock. Recruitment affects quota recommendations by the functional form and the uncertainty of that form. Information on the impact of different climate regimes on pollock recruitment appears to be relatively weak, though during the late 1950s and 1960s, recruitment (based on relatively poor data) appears to have been well below the average from 1970 through 2000.
The importance of prerecruit surveys was evaluated through simulations of likely data outcomes for 2007 and shows that without a survey, the likelihood for exploiting at higher rates was greater. The importance of direct observations on prerecruits as opposed to environmental proxies (e.g., correlative studies on factors related to recruitment success) was shown to be critical in providing timely recommendations on maximum quota levels. Stochastic components due to growth variability were shown to lower advice on quota recommendations by about 20,000 t, a relatively small percentage of the total. However, this evaluation did not account for possible trends in mean weight-at-age which may reduce the adjustment and provide a better basis for short-term projections.
Retrospective analyses are useful to evaluate errors in model structure that should be included for testing purposes. For example, retrospective patterns may be caused by a combination of factors including recruitment processes, natural mortality, somatic growth, and distribution (stock availability to fishery and surveys). Linking information on multispecies modeling results may help to determine the extent that unaccounted natural mortality is contributing retrospective patterns compared to other possibilities such as movement.
Eastern Bering Sea yellowfin sole case study
The eastern Bering Sea has exhibited substantial variability in temperature in recent decades and has shown a marked warming trend in recent years. These changes in temperature appear to significantly affect trawl survey catchability of yellowfin sole (Limanda aspera) and flathead sole (Hippoglossoides elassodon) based on analyses from stock assessment models.
These models use a “global catchability” in which estimated temperature-dependent survey catchability is a function of the annual temperature averaged over the entire survey area. However, spatial differences in temperature between survey tows may result in differential responses to the trawl gear. Gear studies indicate that eastern Bering Sea flatfish are herded into the net path from the trawl bridles, and if the flatfish swimming behavior is affected by temperature, then the degree of herding may also be affected. Thus, we evaluated a “local catchability” model in which the estimated survey catchability at a given trawl station was a function of the local temperature.
For yellowfin sole, both the global and local methods result in a statistically significant positive relationship between catchability and temperature. The increasing trend in EBS temperatures from 1999 to 2005 had a substantial effect on estimated yellowfin sole biomass when adjusting for temperature-dependent catchability. For example, the estimated 2005 total biomass declined 7% when global temperature-dependent catchability was used and the harvest recommendation dropped by nearly 10,000 t.
To assess the importance of including factors affecting survey catchability, a management strategy evaluation (MSE) was developed. The MSE analysis showed that failing to account for temperature-dependent catchability resulted in underestimates of biomass during periods of low catchability and overestimates during periods of high catchability. Our results show that current assessment methods should include factors affecting survey catchability to reduce the variability in catch.
