Limitations of α-acoustic diversity indices in assessing invertebrate sounds in coral reefs

anglais. peer reviewed

anglais. Acoustic indices were developed to monitor aerial/terrestrial soundscapes. However, recent studies have revealed they can be unsuitable for inferring marine fish sound diversity. Less is understood about their use for benthic invertebrate sounds. This study aims to investigate the relationship between 13 indices and snap rates in altiphotic and mesophotic coral reefs. Data from five Polynesian islands were used. We assessed the impact of odontocetes sounds, inter-site variations, and ambient noise levels (partly attributable to distant benthic invertebrate sounds) on the indices. Most indices exhibited correlations with odontocetes vocal activity. Several indices (e.g. ACI, AEI, and H) exhibited different behaviours depending on the FFT used. Furthermore, their application may vary based on the targeted spatial scale. Specifically, the temporal entropy and the bioacoustics index, which were primarily affected by the number of snaps with a high signal-to-noise ratio, are suitable for analysing sounds from specimens near the hydrophone. For global biophony, spectral entropy and Gini-Simpson spectral entropy are the most reliable, as they were predominantly influenced by the ambient sound level. Several indices were influenced by depth and inter-site variations. We recommend avoiding such indices and discouraging using indices that yield conflicting results based on settings to facilitate cross-study comparisons.

14. Life below water