ECoSS: Unlocking the Marine Soundscape with an AI-Powered Curated Open Library of Underwater Sound Signatures

The levels of underwater noise, both continuous (e.g. from shipping) and impulsive (e.g. from seismic surveys), constitute the descriptor 11 of the Marine Strategy Framework Directive. A more comprehensive understanding of the diverse signatures of underwater sound sources is imperative to effectively link recorded sounds to their emitting sources and characterize the anthropogenic sound pressure within a specific area.

The ECoSS project (CINEA/CD(2022)5010/PP/SI2.899121) is dedicated to constructing and validating an open library comprising a curated and continually expanding digital repository of sound signatures from marine underwater soundscapes in shallow seas. It aligns with the objectives outlined in the EU's Open Science Policy, the INSPIRE (Infrastructure for Spatial Information in Europe) Directive, and the FAIR (Findable, Accessible, Interoperable, and Reusable) data principles. Through the utilization of Deep Learning (DL) methodologies, the project aims to streamline the process of constructing the open library while ensuring its scalability and effectiveness in addressing critical environmental challenges.

The development of an AI-powered tool to assess the acoustic impacts of human activities on marine ecosystems will greatly enhance decision-making processes regarding their management and help mitigate potential harm. The ECoSS digital catalogue will feature a robust repository of sound signatures and an AI classifier, enabling users to upload, search, and categorise sounds based on anthropogenic events. This represents a major step forward in our ability to understand and manage the impacts on shallow sea ecosystems.