Spatial and temporal prediction of fin whale distribution in the northwestern Mediterranean Sea

Understanding the distribution of the cetaceans is crucial to improving their conservation. Therefore, a prediction model of fin whale's (Balaenoptera physalus) summer distribution was developed from data collected between May and August, in the Pelagos Mediterranean Marine Mammals Sanctuary. Explanatory variables were selected by multiple logistic regression, among several physiographic and oceanographic parameters. Depth, chlorophyll (Chl a) concentration, and sea surface temperature (SST) were selected for characterizing fin whale presence. Remote sensing imagery (Chl a and SST) was used at an 8-d resolution to capture short-term environmental variability. With the selection of a presence/absence threshold by the receiver operating characteristic curve, a correct classification of 70% (49% for presence, 85% for absence) was achieved for the initial dataset. Model reliability was also tested on an independent dataset, collected in the northwestern Basin; a correct classification of 71% (41% for presence prediction, 86% for absence prediction) was obtained. This study contributes to an understanding of where fin whales might concentrate to feed in summer. Weekly predictions of their distribution represent a valuable conservation tool in a marine protected area, for example to prevent collisions with ships.