Exploring the timing of bird chorus along an anthropization gradient through a deep learning method and geomatic solutions - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2021

Exploring the timing of bird chorus along an anthropization gradient through a deep learning method and geomatic solutions

Vincent Lostanlen

Résumé

The timing of bird chorus (i.e. the daily highest period of vocal activity in singing birds) follows a circadian pattern but can be modified by anthropogenic factors, mainly including artificial lights at night (ALAN) and traffic noise. However, the impacts of ALAN and noise are rarely dissociated each other and almost never distinguished from other anthropic pressures. The second shortcoming is that the measurement of the timing of chorus is almost systematically based on the occurrence of the first song(s) of the model species, whereas the temporal pattern of the whole vocal activity is never taken into account. Through the example of the breeding birds in the Loire Estuary (France), we first dissociated the different facets of anthropization (ALAN, traffic noise, and land-use and land cover) through a geomatic approach and we analyzed the temporal pattern of bird vocalization through a method based on the recently developed deep neural network “BirdNET”. If ALAN and traffic noise influence the timing of bird chorus, we also found both that taking either a facet or another of anthropization but also different measures of the temporal pattern of bird chorus (e.g., the median time of all the vocalizations instead of the first songs) reveals different results. We thus propose a conceptual framework to measure different facets of anthropization but also different measures of temporal pattern of bird chorus to better understand the influence of human activities on bird vocalizations.
Fichier non déposé

Dates et versions

hal-04168598 , version 1 (21-07-2023)

Identifiants

  • HAL Id : hal-04168598 , version 1

Citer

L. Godet, Vincent Lostanlen. Exploring the timing of bird chorus along an anthropization gradient through a deep learning method and geomatic solutions. International Congress for Conservation Biology 2021, Dec 2021, Kigali, Rwanda. ⟨hal-04168598⟩
20 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More