Estimating relative abundance from catch and effort data, using neural networks

Abstract

We develop and test a method to estimate relative abundance from catch and effort data usingneural networks. Most stock assessment models use time series of relative abundance as theirmajor source of information on abundance levels. These time series of relative abundance arefrequently derived from catch-per-unit-of-effort (CPUE) data, using general linearized models(GLMs). GLMs are used to attempt to remove variation in CPUE that is not related to the abundanceof the population. However, GLMs are restricted in the types of relationships between theCPUE and the explanatory variables. An alternative approach is to use structural models basedon scientific understanding to develop complex non-linear relationships between CPUE and theexplanatory variables. Unfortunately, the scientific understanding required to develop thesemodels may not be available. In contrast to structural models, neural networks uses the data toestimate the structure of the non-linear relationship between CPUE and the explanatory variables.Therefore neural networks may provide a better alternative when the structure of the relationshipis uncertain. We use simulated data based on a habitat based-method to test the neuralnetwork approach and to compare it to the GLM approach. Cross validation and simulation testsshow that the neural network performed better than nominal effort and the GLM approach. However,the improvement over GLMs is not substantial. We applied the neural network model toCPUE data for bigeye tuna (Thunnus obesus) in the Pacific Ocean.