Abstract

Catch per unit effort (CPUE) data from the jack mackerel fishery operating in central-south Chile and covering the period 1987-1999 are analyzed in order to detect changes in relative abundance. Two units for fishing effort were used to compute the catch rate: number of days outside port and the number of trips with capture. The fleet was classified in four homogeneous strata by using a principal component analysis (PCA) based on the correlation matrix of technical-operational variables of the purse-seine fleet. The first component explained 87.8% of the variation because the highly correlationed variables. The four homogeneous strata were used as categorical variables to standardize CPUE through a General Linear Model (MLG), previous log-transformation of CPUE data. The first model, based on catch per days outside port, explained more than 64% of the observed variability. In this model, the stratum explains an important fraction of the total variability (24.6%). Similarly, in the case of the CPUE based on trips with capture, a model explained (42.6%) of the total variability and particularly the stratum. However, the two time series of CPUE indicate different tendencies in the annual coefficients for the 1987-1993 period. From 1994 to 1997, the two series exhibit a negative tendency in the annual coefficients. The analysis of catch rate can be improved when a multivariate approach is applied to a heterogeneous fleet, but spatial component should be take into account in future studies. Furthermore, the fishing effort unit can impact remarkably the tendencies of standardized CPUE, particularly if this is considered as a relative index of abundance.