Standardization of the catch rates for swordfish (Xiphias gladius) using fixed and random effects

Abstract

Management decisions in most cases are based on models of stock assessment require estimates of the rates of standardized Catch per Unit Effort (CPUE). Often standardization CPUE are performed using three approaches: A) the year included in the model only as main fixed effect; B) the year included in the model as fixed effect, but also in interactions of fixed effect, and C) the year included as main fixed effect and also in the interactions of random effect. This work proposes to calculate the standardized CPUEs using the three approaches and to accomplish a comparison of results based on the nonparametric Kruskal-Wallis test at 5% level of significance. Generalized Linear Models (GLM) and Generalized Linear Mixed Models (GLMM) were used to obtain estimates of the standardized CPUEs of swordfish for each approach. The results obtained indicated the approach choice may lead to erroneous results about the situation of the stock when analyzed only those years alone.