Automated association and localization of fin whale calls recorded with distributed acoustic sensing

Understanding the behavior and distribution of marine mammals such as fin whales is essential for conservation efforts and population recovery assessments. Distributed acoustic sensing (DAS) offers an innovative and scalable solution by transforming submarine fiber optic cables into dense arrays of acoustic sensors. Over 4 days in November 2021, a public DAS dataset was collected using the Ocean Observatories Initiative Regional Cabled Array off the coast of Oregon, recording tens of thousands of fin whale calls during their breeding season. Previous work demonstrated the feasibility of automated detection using matched filtering, Gabor filtering, and signal-to-noise ratio estimation. Building on these methods, this study aims to obtain a full set of localizations for the experiment. To achieve this, detected calls are automatically associated across DAS channels through a coarse grid search, followed by a least-squares time difference of arrival (TDOA) method to estimate call positions. These techniques are tested on representative subsets of the dataset, providing a foundation for long-term studies on fin whales using DAS. The resulting catalog of localizations and tracks can address questions such as fin whale distribution across the shelf and upper continental slope, net movement direction, and spatial organization of calling individuals. [Work supported by ONR.]