Beyond the defaults: functional response parameter space and ecosystem-level fishing thresholds in dynamic food web model simulations

Abstract

Dynamic food web models are increasingly used to investigate the ecosystem effects of fishing; however, key unknown functional response parameters describing predator-prey interactions strongly influence model behavior. We explored functional response parameter uncertainty and its effect on fishing simulation results using a dynamic food web model of the Gulf of Alaska (GOA) with14 fishing fleets, 104 consumer groups, four primary producer groups, and five detritus pools. After generating millions of potential ecosystems with randomly selected functional response parameters, we assigned groups of these randomly parameterized systems to one of five increasingly intense ecosystem-wide fishing treatments. For each fishing treatment, we counted and compared resulting ecosystems with no extinctions. Surprisingly, the model GOA ecosystems were robust to a wide range of functional response parameters. However, we found an abrupt threshold effect between moderate and heavy exploitation rates, beyond which a much lower proportion of model ecosystems persisted. Beyond this fishing threshold, extinction was more likely, and system attributes differed greatly from moderately fished model ecosystems. Fishing thresholds were not found with default functional response parameters, implying that model simulations should include a wide range of parameterizations to reflect ecological uncertainty and to support sustainable ecosystem-based fishery management.