A curated bioacoustics dataset of European bat vocalizations

A. Leonie Baier, Panagiotis Georgiakakis, Karl-Heinz Frommolt, Signe M.M. Brinkløv, Simeon Q. Smeele, Bob Planqué, Willem-Pier Vellinga, Burooj Ghani, Dan Stowell, Vincent J. Kalkman

Acoustic surveys represent the primary method for monitoring bat populations and assessing their habitat use, yet reference recordings and associated metadata remain fragmented across repositories, recording programs, and private collections. We present ChirosetEurope, a curated dataset of European bat vocalizations (echolocation and social calls) assembled from multiple open repositories and expert-curated collections. The dataset integrates recordings from diverse geographic regions, habitat types, recording devices, and species groups, together with harmonized taxonomic labels, standardized metadata, and group-aware train/validation/test partitions optimized for machine-learning applications. To respect the licensing terms under which recordings were originally published, it is distributed as a metadata resource providing provenance information and persistent links to the original recordings, rather than redistributing all audio files. ChirosetEurope is designed to support a broad range of applications, including bioacoustic research, biodiversity monitoring, species identification, ecological analyses, and machine-learning development. By consolidating previously dispersed resources into a unified and versioned framework, the dataset provides a community resource for research on European bat bioacoustics and automated acoustic analysis.