Feeding habitat of the whale shark Rhincodon typus in the northern Gulf of Mexico determined using species distribution modelling

Whale shark Rhincodon typus is a globally distributed species, but there is a lack of knowledge pertaining to their biology, seasonal occurrence, and distribution in the northern Gulf of Mexico (NGOM). Understanding critical habitat for whale sharks is essential on both a regional and global basis for proper management because of their large migratory range. The goal of the present study was to describe the regional distribution of whale shark feeding aggregations in the NGOM by exploiting a presence-only dataset collected as a part of a volunteer sighting survey. Whale shark aggregations have been documented in large numbers in the NGOM since 2003, and species distribution models provide a unique approach to analyzing these presence data. We used maximum entropy and ecological niche factor analysis, 2 algorithms designed for predicting species distribution based only on presence data, to analyze data for the summer period in 2008 and 2009. Cohen's kappa (kappa) and the 'area under the receiver operating characteristic curve' (AUC) were used to evaluate model performance with an external testing dataset. Kappa values ranged from 0.28 to 0.69, and AUC values ranged from 0.73 to 0.80, indicating that the predicted distribution had a fair to substantial agreement with the testing data. Distance to continental shelf edge, distance to adjacent petroleum platforms, and chlorophyll a were the variables most strongly related to whale shark sightings, likely due to an association of these variables with high food availability. Suitable habitat was predicted along the continental shelf edge, with the most suitable habitat predicted south of the Mississippi River Delta. The spatial distribution of suitable habitat is dynamic; therefore, a multi-year study is underway to better delineate temporal trends in regional whale shark distribution and to identify consistent areas of high suitability. Presence-only habitat models are a powerful tool for delineating important regional habitat for a vulnerable, highly migratory species.