Choosing fisheries harvest policies: when does uncertainty matter?

Abstract

Fisheries harvest policies are formulated under uncertainty because estimates of stock abundance and biological parameters are imprecise. However, allowable harvests are often based solely on point estimates of these quantities, essentially ignoring uncertainty. Some fishery scientists have advocated adjusting harvest levels downward to account for uncertainty, but few formal methods have been developed to determine how large these "uncertainty adjustments" should be. We describe an age-structured simulation model that explicitly incorporated uncertainty in parameters of the stock–recruitment relationship, errors in abundance estimates, and year-to-year variability in recruitment and calculated which uncertainty adjustment was optimal in terms of expected discounted yield. The optimal adjustment varied considerably, depending on the stock and harvest policy simulated. The increase in expected value from incorporating the adjustment into the harvest policy was usually small, except when we modeled a biological "threshold", where overharvests could lead to an irreversible stock collapse. Therefore, while our analysis suggests that basing harvest decisions solely on the best point estimates may often be an approximately optimal strategy, it also indicates that large adjustments may sometimes be appropriate. Consequently, fishery managers should avoid making arbitrary adjustments for uncertainty, and instead derive the optimal adjustment for each situation.