Passive acoustic monitoring and audio subsampling: optimizing autonomous methods for avian biodiversity assessments
Year: 2022
Type: dissertation
Abstract: The global decline of bird populations has prompted the search for innovative tools to inventory and monitor their communities. While standard surveys require an observer in the field, autonomous sound recorders are an alternative that demands less expertise and is more scalable in time and space. The literature is not consensual on the efficiency of this method and the factors that influence it, particularly in multispecies studies, although recent attempts have yielded encouraging results towards its applicability in practical monitoring situations. In this study, we conducted a set of observer and recorder-based bird point counts in cork oak woodlands in Portugal, in winter season. We compared both methods in terms of richness values and species-by-species, and assessed the role of the observer and sampling time in recorder performance. Additionally, we compared species richness values obtained through three types of intermittent audio file subsampling, and by intermittent and continuous approaches. We found the observer detected significantly more species, but its presence did not influence the recorder’s results and the pool of species detected by both was similar. We found time of sampling to be relevant in autonomous recorders. The degree of intermittence generated different cost/benefit scenarios for audio processing. Lastly, intermittent subsampling surpassed the number of species detected through continuous subsampling by a factor of two. The results of this study showed that recorders tended to perform well in biodiversity surveys in winter, while being more flexible in scaling, especially when small portions of audio are analysed. However, they also suggest the observer should not be dismissed a priori, and highlight the complexity of factors that may influence the recorder’s performance. We encourage future studies to test this performance over a variety of different time, spatial, and species-related constraints, to maximize the universality of a future all-year autonomous method monitoring protocol.
Author: Burfin, Thomas Wojciech
Language: Portuguese
Topic: Avian ecology and behavior
Subfield: Ecology
Field: Environmental Science
Domain: Physical Sciences
SDG: Life on land
Open Access status: green
OpenAlex ID: W7042445611