Studying the ten years variability of Octopus vulgaris in Senegalese waters using generalized additive model (GAM)

Abstract

In the Senegalese upwelling ecosystem, Octopus vulgaris population shows high seasonal and interannual variability. Simple regression methods have been applied to investigate how octopus abundance is related to environmental change. However, the precision of such models is still unclear due to the non-linear character of the relationships. This work applied generalized additive model (GAM) to investigate ten years (1996-2005) variability of octopus recruitment in the Senegalese waters. The semi-parametric smooth function and Poisson distribution were used to build the models. The variables included in the models were sea surface temperature (SST and time (Month and Year). We show that GAM is more useful and relevant in investigating the complex relationships between octopus recruitment and SST at seasonal and interannual scales than classical regression techniques, by offering the best fits. We demonstrated that high octopus recruitment occurred at water temperature range between 17.5 and 18 °C.