Threshold management strategies for exploited fish populations

Abstract

Under a threshold management strategy, harvesting occurs at a constant rate but ceases when a population drops below a threshold. The threshold approach seeks to enhance long-term yield of a population and to maintain population renewability. I evaluated threshold management strategies for selected herring and pollock stocks in Alaska. First, I examined stock-recruitment data from 19 major herring stocks worldwide to provide the basis for evaluating threshold management strategies. Seventy-three percent of these stocks exhibited statistically significant density-dependence. Most stocks have compensatory, dome-shaped stock-recruitment curves. Then, I simulated threshold management strategies for eastern Bering Sea (EBS) pollock and herring and Prince William Sound (PWS) herring using a single-species model. I further examined seven alternative threshold estimation methods. Cohort analysis, catch-at-age analysis, and catch and population sampling yielded estimates of population parameters. The objective function was a weighted function of increased average yield and decreased standard deviation of yield over a planning horizon. Compared to a non-threshold approach, threshold management strategies increase the long-term average yields, stabilize population abundances, shorten rebuilding times, and increase management flexibility. For a maximum yield criterion and Ricker stock-recruitment models, optimal fishing mortalities are slightly above fishing mortalities at maximum sustained yield (MSY), and optimal threshold levels range from 40% to 60% of pristine biomass for EBS pollock, from 40% to 50% for EBS herring and from 30% to 60% for PWS herring. With fishing mortality at MSY and the criterion of equal trade-off between yield and its variation, optimal thresholds range from 20% to 30% of pristine biomass for pollock. With the status quo exploitation rate of 20%, optimal thresholds range from 10% to 25% of pristine biomass for EBS herring, and from 5% to 25% for PWS herring. Of the threshold estimation methods evaluated, default percentage of pristine biomass usually performs best. Loss of yield due to errors in threshold estimation is small, generally under 10%. A bout 15 to 20 years of data are required to obtain a reliable estimate of thresholds. With single-species dynamics, the form of the stock-recruitment curve, exploitation rate and management objective are the most important factors affecting optimal thresholds.