P-357 Accounting for Indirect Effects and Non-Commensurate Values in Ecosystem Based Fishery Management

Kate Richerson , Department of Applied Mathematics and Statistics, Center for Stock Assessment Research, University of California Santa Cruz, Santa Cruz, CA
Phillip S. Levin , Conservation Biology Division, NOAA Northwest Fisheries Science Center, Seattle, WA
Marc Mangel , Center for Stock Assessment Research & Department of Applied Mathematics and Statistics, University of California, Santa Cruz, Santa Cruz, CA
Ecosystem-based fishery management (EBFM) requires taking account of indirect effects (such as habitat destruction, incidental mortality, and competition between the fishery and marine mammals or birds) and dealing with non-commensurate values (such as yield from the fishery and production of offspring by the birds or mammals competing for the same resource). The perspective of EBFM requires that the rate of fishing mortality is less than the value that provides maximum sustainable yield (MSY), but the question is how far below this level should the fishery operate. For this problem in multiobjective programming, simple method of solution was developed and illustrated with the fishery for sandeels (Ammodytes spp.) in the Shetland Islands. The yield from the fishery at a given fishing mortality F is scaled by MSY (so that this quantity increases as fishing mortality increases from 0 to that giving MSY) and the breeding success of predators (black-legged kittiwakes Rissa tridactyla and Arctic terns Sterna paradisaea) at a given fishing mortality is scaled by that in the absence of fishing. The result is two non-dimensional quantities that can be combined into a single value function, which can then be explored or optimized. It is shown that a reduction of only about 20 percent in yield can nearly double the breeding performance of the more sensitive predator. Extensions of the method are discussed with particular application to the Antarctic krill fishery, using state dependent life history, as implemented by stochastic dynamic programming.