Uncertainty and Fishery Management in the North American Great Lakes: Lessons from Applications of Decision Analysis

—Many fishery management decisions continue to be guided by science only through “best guess” interpretation of assessment information and deterministic models of fisheries and food webs; until very recently this was true of nearly all fishery management in the Great Lakes. However, fishery management decisions can be improved by formally considering uncertainty when evaluating management options; practical tools for doing this have become increasingly available. Accounting for uncertainty is important because acting as though the best guess is true may be substantially suboptimal if this leads to poor performance for other, less likely, but still plausible “states of the world.” For a variety of critical Great Lakes fishery management issues, including determining appropriate investments in sea lamprey Petromyzon marinus control, setting suitable levels of salmonine stocking, and establishing percid harvest policies, we have considered the importance of explicitly incorporating uncertainty. In each case, we worked closely with fishery managers to conduct a decision analysis of management options they identified, using contemporary statistical methods to formally assess uncertainty about key fishery parameters and stochastic simulation to compare management options. These decision analyses were used by fishery managers to develop policies that more objectively account for uncertainty and to garner support from stakeholders and policy makers. The approach shows considerable promise for future fishery management in the Great Lakes, but will face substantial challenges as managers seek to more effectively involve stakeholders throughout the process, foster the requisite technical expertise within their agencies, and communicate the results of highly technical analyses to both stakeholders and decision makers. Three important aspects of salmon Arctic-Yukon-Kuskokwim region management for which a decision analysis approach would be particularly valuable are (1) the evaluation of different options for assessment sampling of returning adult salmon, used to determine whether escapement targets are being met; (2) strategies for in-season management of salmon harvest; and (3) setting annual escapement goals for individual stocks.