Journal of Ecoacoustics

Auditory signals are often used by forest species to attract mates, define and defend territories, and locate prey, and thus these signals may be monitored and used to estimate species presence, richness and acoustic complexity of a patch of habitat. We analyzed recordings from a biodiversity hotspot in the rainforests of Batang Ai National Park in Sarawak, Malaysian Borneo. Three recording sites were established in the forest understory and continuous recordings were made for an acoustic snapshot of approximately 40 h. From these recordings, the bioacoustic index (BI) and acoustic complexity index (ACI) were computed. These acoustic indices exhibited clear periodicity with periods on the order of >6 h. The ACI and BI time series also showed oscillations, with peaks separated by 12 h corresponding to the alternation between bird and frog activity during the day and night, respectively. We quantified the relationships between the acoustic index values and anuran and avian richness, and environmental variables (rainfall intensity and time of day) using correlative and information theoretic techniques. ACI and BI were moderately positively and negatively correlated with rainfall intensity, respectively. ACI and BI were also weakly-to-moderately correlated to species richness, but with mixed directions between recording sites. However, the correlations and mutual information values, a model-free estimator of the relationship strength of random variables, were highest for the relationships between ACI and BI with respect to the rate of individual frog calls, rather than species richness alone. We conclude that acoustic indices are most useful for monitoring ecological dynamics on timescales longer than 6 h. We suggest that acoustic indices are best applied to studying changes in acoustic activity at the level of ecological populations rather than for estimating species richness, as they have been commonly applied in the past. Full article