Introducing evascape, a model-based soundscape assembler: impact of background sounds on biodiversity monitoring with ecoacoustic indices

Abstract

The development of ecoacoustics and soundscape ecology is linked to the development of field recording techniques and the production of large and diverse sound databases representative of various environments. However, an important caveat in this approach lies in the lack of reference data, limiting the possibility of developing acoustic biodiversity indices and assessing their sensitivity to environmental change. To address this issue, previous studies have attempted to generate “engineered soundscapes”, that is artificial soundscape recordings which offer some control on, among other aspects, species diversity. Unfortunately, these tools lack complexity and realism, limiting their ecological validity. Here, we introduce Evascape, a Python algorithm informed by available knowledge in soundscape ecology aiming at producing assembled soundscapes with high ecological validity. The algorithm offers the possibility to control biophony (number of species, number of individuals, time organization), geophony (wind and rain with adjustable level), anthropophony (aircraft noise) and sound propagation effects (frequency-dependent attenuation function of the habitat). The algorithm was tested using an acoustic database collected in a protected European cold forest. The generated assembled soundscapes were then used to assess a selection of commonly-used ecoacoustic indices across a large number of controlled scenarios, and to identify the indices that respond most strongly to changes in species richness. Evascape opens the path to other ecoacoustics-related research programs, offering the possibility to design empirical and computational experiments with controlled soundscape parameters.