The efficacy of acoustic indices for monitoring abundance and diversity in soil soundscapes

• Soil ecoacoustics is a rapidly emerging field. • Acoustic indices present a way to quantify soundscapes without identifying sounds. • We tested index response to simulated soundscapes with varying acoustic diversity. • Using multiple indices predicts acoustic abundance and low levels of richness well. • Indices are not effective predictors of acoustic evenness or gamma diversity. Soil ecoacoustics is a rapidly emerging field heralded as a non-invasive method for monitoring soil fauna. Ecoacoustic analysis commonly uses acoustic indices to analyse soundscapes, linking them to ‘traditional’ biodiversity value metrics such as species richness and abundance, but it is not clear if this approach is appropriate for soil soundscapes. Furthermore, there are very few controlled experiments assessing how commonly used acoustic indices respond to different sound types, and none belowground. We address this by synthesising soil soundscapes with differing levels of acoustic richness, abundance, and evenness using soil recordings from the UK, France, and Brazil. Applying 14 acoustic indices on 1-minute soundscapes, we assessed: 1) how changes in acoustic diversity impact acoustic indices and 2) how accurately combinations of indices predict biodiversity metrics. Finally we assessed 3) whether gamma acoustic richness can be predicted accurately using multiple acoustic index scores from repeated surveys, whilst experimentally altering the alpha and beta diversity components. We find that acoustic abundance strongly affects values of acoustic indices designed to quantify the number of sound events in a soundscape, and that a combination of these indices can accurately predict abundance at 1-minute timescales. Combinations of indices could predict acoustic richness when richness values were low, but were ineffective for evenness. Additionally, acoustic indices were poor predictors of gamma diversity, especially when gamma was driven solely by beta diversity. Overall, we found that acoustic indices were good predictors of acoustic abundance, but should be used with caution for other diversity metrics.