No document available.
Abstract :
[en] In the last decades, passive acoustic monitoring has become an effective tool to study a great variety of vocal species. This method is most effective for species that are difficult to observe due to their remote location or the great deal of time they spend out of sight. One of those species is the minke whale, Balaenoptera acutorostrata. As one of the smallest baleen whales, the size of its blow is inconspicuous and makes it hard to spot. As a result, it can be uncertain whether this species inhabits certain areas or not. Common minke whale, B. acutorostrata acutorostrata, is a subspecies known to occur in the Northern Atlantic, where migratory patterns have been recorded. While their presence is greatly studied in the northern part of this region, little is known about their winter distribution, further south. During this study, a Convolutional Neural Network detector was used on previously collected datasets from Bermuda, the Caribbeans and the Southeast U.S. coast in order to detect low-frequency pulse trains (< 1 kHz) emitted by common minke whales. The detected calls were then examined to confirm whether they belonged to common minke whales or not. The results have revealed the presence of common minke whales off Florida’s coast, Netherlands Antilles, Martinique, and Virgin Islands. Off Florida’s coast, calls were detected during December and January by recorders placed at 100 to 400 m
depths. Recorders located in the same area, but at lesser depth (around 50 m), did not record as many calls in comparison. These results provide new insight on the species’ distribution in the area during the winter months. This is of crucial importance given the current expansion of offshore economic development worldwide.