Novel Tools for Automated Activity Monitoring of Forest Bird Communities
Résumé
Movement data are essential for accurate conservation schemes as they shed light on species space use and timing of activity. For many small vertebrates in dense habitats, the fine-scale monitoring of activity patterns is limited. We developed a novel, automated method based on fixed VHF antenna allowing to detect small vertebrate movements in the Marburg Open Forest, Germany. We tracked a songbird community of 124 individuals and 15 species over three consecutive years and trained a machine learning model to recognize basic behavioural states (active/passive) based on VHF-signal patterns. Species showed tight temporal synchronization in the timing of their activity, although a few species covaried negatively with one another, suggesting competition may also play a role in shaping species activity patterns. Automated tracking technologies now allow to compare patterns of activity and estimate the strength of species interactions in otherwise elusive species, leading to improved conservation schemes.
Origine | Fichiers produits par l'(les) auteur(s) |
---|