Using a spatio-temporal Bayesian approach to estimate the relative abundance index of yellow squat lobster (Cervimunida johni) off Chile

Estimating relative abundance indexes based on spatio-temporal variations in fishing effort has been one of the greatest challenges in fisheries sciences. Obtained from the catch per unit of effort (CPUE), such indexes are generally used within evaluation models as “relative” to the stock abundance. Herein, a Bayesian spatio-temporal model was used to obtain an index for yellow squat lobster (Cervimunida johni) between the III and IV regions of Chile based on CPUE data (Kg/h.a) from fishing logs. The spatial field was approximated by a GMRF using the SPDE method and posterior distributions of interest were approximated using the Integrated Nested Laplace Approximation (INLA). By taking into account the distributional assumption of the CPUE the proposed model showed a good fit to the observed data. The proposed method allowed obtaining a relative index of abundance which could be included within the classic stock assessment models.