Evaluating Tag-Reliant Harvest Estimators in Chinook Salmon Mixed-Stock Fisheries Using Simulations

Abstract

Management of mixed-stock Chinook salmon fisheries requires balancing fishery access and conservation of vulnerable stocks. Although accurate, timely estimates of stock-specific harvest are crucial in achieving competing objectives, limited numbers of stock assignments (e.g., tag recoveries) can diminish the utility of estimates. We used a flexible simulation approach, applied to both a theoretical and real-world fishery case study, to compare the performance of competing monitoring alternatives and estimators for harvest. We sought to improve accuracy for point estimates of harvest and harvest trajectories over time. Bayesian models provided similarly accurate point estimates to existing models at high levels of data aggregation, generally improved estimates of harvest trajectories at intermediate aggregation, and reliable estimates of uncertainty. Incorporation of time-lagged prior information inconsistently improved estimates of harvest trajectories. Among monitoring alternatives yielding equal increases (33%) in CWT recoveries, increasing tagging rates resulted in the greatest decrease in estimate uncertainty for target stocks (37.5 to 45.3%). Variable performances of mixed-stock harvest estimators suggest their use should considered on a stock- and fishery-specific basis, potentially using a simulation-based approach.