Rapid assessment of avian species richness and abundance using acoustic indices

Tom Bradfer-Lawrence,Daisy H Dent,Stephen G Willis,Nick Gardner,Nils Bunnefeld

doi:10.1016/j.ecolind.2020.106400

Tom Bradfer-Lawrence, Daisy H Dent + Show 3 more

Open Access

https://doi.org/10.1016/j.ecolind.2020.106400

Copy DOI

Abstract

• Monitoring environmental sound presents an opportunity for rapid biodiversity assessment. • Environmental sound can be readily described using acoustic indices. • Higher bird species-richness and abundance associated with patterns in indices values. • Patterns were coherent across multiple acoustic indices, and across habitat types. • We solve methodological issues and provide conceptual clarity for using indices. Accelerating global shifts in climate and land use change are altering natural habitats and species assemblages, making management interventions crucial to halt the biodiversity crisis. Management decisions must be informed by accurate biodiversity assessments. However, such assessments are often time consuming, expensive, and require specialist knowledge. Monitoring environmental sound may offer a novel method for rapid biodiversity assessment. Changes in species assemblages at a given location are reflected in the site’s acoustic energy, termed the soundscape. Soundscapes can be readily described using acoustic indices; metrics based on objective features of recordings such as pitch and amplitude. Changes in acoustic indices values may therefore reflect changes in species assemblages, alerting land managers to shifts in wildlife populations. However, thus far, evidence supporting the use of acoustic indices in biodiversity monitoring has been equivocal. Here, we test the practical application of acoustic indices for biodiversity monitoring while solving methodological issues and providing conceptual clarity. Using 84 h of audio recordings covering 315 dawns from 43 sites, coupled with bird assemblage and vegetation data collected in the field, we demonstrate strong relationships between acoustic indices and avian species richness and abundance. In contrast with many previous studies, we found that sites with high bird species-richness and abundance had less even soundscapes (i.e. acoustic energy was less evenly distributed among frequencies) compared with sites with low species richness and abundance. Crucially, these patterns were coherent across multiple acoustic indices, and across habitat types, emphasising their utility for monitoring. Acoustic indices sensitive to the frequencies at which birds sing are most useful for monitoring avian communities; the Acoustic Evenness Index, Biophony Index, and the biophony component of the Normalised Difference Soundscape Index exhibited the strongest relationship with species richness. Land managers can use acoustic indices for biodiversity monitoring, complementing other, more established, assessment methods.

Highlights

Biodiversity assessment is an increasingly urgent task in the face of global environmental change (Pereira et al, 2013)
In contrast with many previous studies, we found that sites with high bird speciesrichness and abundance had less even soundscapes compared with sites with low species richness and abundance
The principal question we address is: can acoustic indices be used as effective biodiversity monitoring tools, reflecting species richness and abundance? We hypothesise that greater avian species richness, avian abundance, and vegetation structural complexity will be reflected in acoustic indices values indicating greater soundscape complexity

Summary

Introduction

Biodiversity assessment is an increasingly urgent task in the face of global environmental change (Pereira et al, 2013). Ecoacoustics, the study of environmental sound, may offer a more rapid and economical means of terrestrial biodiversity appraisal than traditional approaches (Burivalova et al 2019a). Open-source audio recorders that can be deployed in the field for weeks or months at a time have made it relatively straightforward to collect tens of thousands of hours of sound recordings (Sueur and Farina 2015; Bradfer-Lawrence et al 2019). Acoustic indices can be calculated from audio recordings, rather than manually categorising species composition, which is prohibitively timeconsuming (Pijanowski et al 2011). Acoustic indices are derived from features of the recordings such as amplitude and frequency, with individual indices typically describing different characteristics of the soundscape (Sueur et al 2014). The soundscape is comprised of the acoustic energy at a given location and has three components: biophony, sounds produced by animals; anthrophony, sounds produced by humans or machinery; and geophony, sounds from natural processes such as wind or rain (Pijanowski et al 2011)

Methods

Results

Conclusion