Abstract
Earth’s oceans cover 70% of the planet yet remain among the least observed regions. Underwater acoustics is a primary sensing modality, but recordings span biological, anthropogenic and natural sounds while labels remain sparse. We present an unknownness-aware monitoring framework for underwater acoustic sensing platforms: it adapts a pre-trained audio representation model to underwater soundscapes, defines an embedding-space unknownness score, and couples it with detector outputs in a Detect–Group–Promote–Union (DGPU) loop tuned under false positives per hour (FP/h) constraints: Detect surfaces candidate events from continuous audio; Group clusters unknown candidates in embedding space; Promote trains lightweight classifiers from sparse expert labels on selected clusters; and Union combines baseline and promoted detectors. Across a 56-class underwater Sound Event Detection (SED) benchmark and two cetacean field datasets, the framework attains a dataset-level bandwidth-matched out-of-distribution AUROC of 1.000 on Hawaiian survey data (HICEAS), raises right-whale upcall detection recall on continuous recordings (FRDR) from 0.027 to 0.390 at FP/h ≈ 10 with a supervised Promoter, with 300 labelled events (~ 32% of training-fold labels) reaching 0.37 (95% of the full-label recall), and classifies seven HICEAS cetacean species with the Promoter classifier alone (AUC 0.924–0.992 within-day and 0.51–0.97 cross-day with a fixed classifier), supporting stethoscope-like ocean monitoring.
Acknowledgements
We gratefully acknowledge the institutions and researchers who collected, curated and released the FRDR and HICEAS datasets. We also thank the teams and organisations that made the ocean acoustic datasets used for BEATs domain-adaptive pretraining publicly available: the NOAA and U.S. National Park Service (NPS) SanctSound programme, the U.S. Navy Undersea Warfare Training Range (USWTR), the U.S. National Park Service Glacier Bay National Park and Preserve, the NOAA Ocean Noise Reference Station Network (NOAA NRS, stations NRS01–NRS03), the NOAA Office of National Marine Sanctuaries (ONMS), the Ocean Networks Canada NEPTUNE and VENUS cabled observatories, and the PALAOA observatory and the PANGAEA data repository. Field data collection underlying the 56-class underwater sound dataset used for SED training was supported by the ATLA “Security Technology Research Promotion Programme” (Grant Number JPJ004596). We thank all collaborating institutions and team members involved in that programme for their contributions to data collection and dataset preparation. We particularly acknowledge the substantial public investment and the extensive expert annotation effort that have made these open datasets available to the community, and the careful documentation of dataset structure, sampling rate, original recording configuration and known limitations provided by the dataset stewards, which enabled the bandwidth-aware analysis presented in this work.
Funding
The annotated training data used in this study were collected under a prior project supported by the Acquisition, Technology & Logistics Agency (ATLA) of Japan (Security Technology Research Promotion Programme; Grant Number JPJ004596). This study is also associated with follow-up applied research supported by Japan’s Ministry of the Environment (MOE) SBIR Environmental Protection Research Grant (Environmental Startup Research and Development Support Programme for Innovation Creation; project ID 25-R-03) and by the New Energy and Industrial Technology Development Organization (NEDO) Small Business Innovation Research (SBIR) Promotion Programme (project number 24000810-0).
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Competing interests
T. Noda and T. Koizumi are founders and executives of Biologging Solutions Inc., which develops underwater acoustic monitoring buoys, bio-logging devices, and associated acoustic AI systems. The methods proposed in this study may be applied to future products or services of the company. Patent applications related to aspects of this work have been filed.
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Below is the link to the electronic supplementary material.
Supplementary Material 1 (download DOCX )
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Reprints and permissions
About this article
Cite this article
Noda, T., Koizumi, T. Discovery and promotion of unknown sounds into operational detection targets for underwater passive acoustic monitoring under false alarm constraints.
Sci Rep (2026). https://doi.org/10.1038/s41598-026-73751-w
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-026-73751-w
Keywords
- Passive acoustic monitoring
- Marine bioacoustics
- Sound event detection
- Out-of-distribution detection
- Open-world monitoring
- False-positive rate control
Source: Ecology - nature.com
