Distributed acoustic sensing for marine mammal monitoring

Marine mammal monitoring provides critical data to inform species management and sustainable industrial development. Passive acoustic monitoring, in particular, can provide continuous, long‐term and non‐observer‐biased data, with a subset of systems capable of delivering data in near‐real time. Acoustic monitoring is increasingly required around industrial activities to assess impacts on marine mammals. However, while effective, these systems can be costly and pose human safety risks during at sea field operations, hindering adaptive management efforts. Distributed Acoustic Sensing (DAS) is an emerging technology that converts underwater fibre optic cables into dense listening arrays providing data with high spatio‐temporal resolution. It is usually implemented from a dry end (e.g. from shore, from a substation or a platform) enabling real‐time data processing and communications. DAS implemented on the extensive existing infrastructure of telecommunication and energy industries, can detect and localize low‐frequency baleen whale vocalizations, including blue ( Balaenoptera musculus ) and fin whales ( B. Physalus ). Ongoing research seeks to extend its capabilities to broader taxonomic groups, including through custom fibre installations. DAS is a promising economic alternative or supplemental method for marine mammal monitoring. DAS is increasingly used by offshore industries for infrastructure monitoring, presenting dual‐use opportunities for simultaneous infrastructure and environmental assessment. Given the growing recognition of submarine cables as critical infrastructure, DAS implementations are likely to accelerate. The broad adoption of dual‐use DAS that allows for marine mammal monitoring requires dialogue between marine researchers, technology developers, ocean industries and policymakers. Operational DAS‐based marine mammal monitoring requires systematic advances in two critical areas to reach their full potential: (1) characterizing DAS performance across species‐relevant frequency ranges and cable configurations and (2) developing robust automated methods to process the massive spatial–temporal datasets generated. This review provides a development framework addressing these priorities through standardized protocols and algorithm benchmarking. Our recommendations emphasize open science collaboration to enable real‐time processing and adaptive management applications.