Résumé
Tropical tuna purse-seine fisheries spatially co-occur with various megafauna species, such as whale sharks, dolphins and baleen whales in all oceans of the world. Here, we analyzed a 10-year (2002-2011) dataset from logbooks of European tropical tuna purse-seine vessels operating in the tropical Eastern Atlantic and Western Indian Oceans, with the aim of identifying the principle environmental variables under which such co-occurrence appear. We applied a Delta-model approach using Generalized Additive Models (GAM) and Boosted Regression Trees (BRT) models, accounting for spatial autocorrelation using a contiguity matrix based on a residuals autocovariate (RAC) approach. The variables that contributed most in the models were chlorophyll-a concentration in the Atlantic Ocean, as well as depth and monsoon in theIndian Ocean. High co-occurrence between whale sharks, baleen whales and tuna purse-seine fisheries were mostly observed in productive areas during particular seasons. In light of the lack of a full coverage scientific observer on board program, the large, long-term dataset obtained from logbooks of tuna purse-seine vessels is highly important for identifying seasonal and spatial co-occurrence between the distribution of fisheries and megafauna, and the underlying environmental variables. This study can help to design conservation management measures for megafauna species within the framework of spatial fishery management strategies.