Bayesian spatio-temporal CPUE standardization: Case study of European sardine (Sardina pilchardus) along the western coast of Portugal

Understanding the key factors influencing population dynamics of fish stocks requires knowledge of their spatial distribution and seasonal habitat selection, but these spatio-temporal dynamics are often not explicitly included in ecological studies and stock assessment models. This study standardized the data of sardine fishery-dependent catch-per-unit- effort (CPUE) from the west coast of Portugal using Bayesian hierarchical spatio-temporal models (BHSTM) with the integrated nested Laplace approximation (INLA). Sardine CPUE was best explained by length of the vessel, vessel ID, month, year, and location (latitude, longitude). In terms of spatio-temporal distribution, sardine biomass prediction maps showed a constant pattern that changed every quarter of the year. In addition, sardine CPUE index showed a cyclical trend along the year with minimum values in July and maximum peak in November. This approach provided insights on variables and corresponding modelling effects that may be relevant in spatio-temporal fishery-dependent data standardization, and that could be applied to other fish species and areas.

Automated summary

This study standardized the fishery-dependent catch-per-unit-effort (CPUE) data of sardines from the west coast of Portugal using Bayesian hierarchical spatio-temporal models. The results showed that factors such as vessel length, vessel ID, month, year, and location influenced sardine CPUE. The spatio-temporal distribution of sardine biomass exhibited a constant pattern that changed every quarter of the year, with a cyclical trend in CPUE values over time.