Source localization in the presence of depth-varying sound speed: A test case for sperm whales

Sperm whales (Physeter macrocephalus) are the largest of the toothed whales, amongst the deepest divers and are distributed worldwide. They are the focus of intensive biological research for distribution analysis, behavioral studies and, more recently, for language processing through Project CETI. Of particular interest for the latter is the study of their mobility patterns and distribution ranges, which requires knowledge of their location. In this paper, we offer a passive acoustic monitoring (PAM) approach for the localization of Sperm whales through arrays of receivers deployed from surface moorings. Our scheme performs adaptive beamforming for localization, and takes into account the varying sound speed in water and the whale's mobility. A series of stability tests increase the accuracy of positioning. Numerical results investigate the effect of system parameters, varying sound speed, number of moorings and number of receiving elements. Results from real sea experiments replaying sperm whale vocalizations in deep water across the Dominica Island demonstrate the practicality of our approach and its robustness for varying signal to noise ratios and range. To ensure reproducibility, we provide our implementation code.