Cautions on using percentile-based benchmarks of status for data-limited populations of Pacific salmon under persistent trends in productivity and uncertain outcomes from harvest management

Management strategies that track time-varying productivity may better achieve management objectives than those that assume constant productivity. In Canada, biological assessments of Pacific salmon populations use upper and lower benchmarks to delineate three zones of status, green, amber and red, representing increasing probability of extirpation, reduced production, and possible increases in management intervention. Although benchmarks that track changes in productivity based on stock-recruitment relationships have been evaluated in closed-loop simulation models, benchmarks for data-limited populations that do not use recruitment relationships have not. The primary objective of this study was to evaluate the performance of lower benchmarks for data-limited populations derived from a percentile of observed spawner time-series, against those derived from data-intensive methods, using a closed-loop simulation model that included temporal changes in stock productivity. We also evaluated the role of assumptions about metapopulation dynamics and outcome uncertainty (uncertainty from the outcomes of implementing harvest management decisions) on the relative performance of lower benchmarks. Our results suggest caution when applying percentile-based benchmarks to harvest management decision rules due to high probabilities of extirpation especially when productivity is declining and outcomes of management actions are uncertain. We further provide a risk assessment framework for evaluating probabilities of extirpation of candidate benchmarks and possible trade-offs associated with harvest opportunities.