Quantifying the Soundscape: How filters change acoustic indices

Abstract

Monitoring biodiversity can be time consuming and costly. Automated recording units (ARUs) have rapidly emerged as an efficient and cost-effective tool for measuring biodiversity. Acoustic indices are one output from recordings from ARUs that can be quantified to serve as an ecological indicator for biodiversity. However, there is a lack of guidance on what acoustic filters to apply to these indices and when. To address this gap, we collected acoustic data from study locations spanning temperate and tropical forests, agricultural grasslands and croplands, and peri-urban development. We applied filters of 80, 500, 1000, and 2000 Hz to these data when calculating the different indices. In addition, we considered the effect landscape context, road noise, season, and elevation have on seven of the most commonly used acoustic indices with different frequency filters. We found that two indices, Acoustic Diversity Index (ADI) and Acoustic Evenness Index (AEI), were most sensitive to filtering, changing significantly between an 80 and 1000 Hz filter across the different covariates. Acoustic Complexity Index (ACI), however, remained consistent with the different filters. These results suggest that when using acoustic indices, one should be cognizant of the context of the study location and the season of the study period when using ADI and AEI. ACI can be used more generously since it is not as sensitive to filtering. ARUs and acoustic indices are an effective tool for measuring biodiversity, but to ensure proper reporting and ability to compare results across studies, more guidelines on appropriate filtering of acoustic indices should be developed.

Keywords

• Acoustic indices
• Automated recording units
• Filtering
• Soundscape