2 December 2025: A new study on the development of a conservative automated tonal detector with a high performance at a large temporal scale

A study by Delbosc and colleagues titled “Development of a conservative automated tonal detector with high performance at a large temporal scale”, was published in the journal Bioacoustics, in September 2025. The study provides crucial information on the use of passive acoustic monitoring (PAM), a methodology that consists in recording wildlife and environmental sounds, to monitor changes in fauna distribution in remote regions, such as in Arctic waters.

PAM is valuable for detecting vocal species over long time periods. But with the growing amount of acoustic data collected by scientists come growing challenges in analysing the data. Therefore, the authors developed an automatic tonal detector, a computer programme that can automatically detect and identify sounds made by animals. The authors tested this method on bearded seals (Erignathus barbatus), which live in remote areas in the Arctic and are highly vocal during the breeding season, making them well-suited for PAM-based monitoring. The authors assessed the detector´s performance across multiple areas and seasons.

The detector demonstrated a conservative approach and failed to detect 77.5% of the seal calls that were present in the recordings. Nevertheless, it showed high precision and, of the calls it managed to detect, it correctly identified 100% of them as seal vocalisations. Furthermore, the detector was effective in capturing seasonal presence patterns, with 80% to 96% agreement with manual annotations from recordings made in Svalbard.

In conclusion, this new tool offers a rapid, first-order estimation method for detecting bearded seals and other sounds produced by species within large acoustic datasets collected over extended time periods.

You can read the full article by Delbosc et al., here: https://doi.org/10.1080/09524622.2025.2568528

Photo credit: Ondrej Prosicky.