Spatial Soundscape Ecology: Application in a Paleotropical Rainforest

摘要

The world’s biodiversity is declining with much of the decrease attributed to human activity. The paleotropical rainforests of Borneo contribute 10% of the world biodiversity but are at risk of destruction due to logging and other human interests. Soundscape ecology, defi ned as the study of the composition of sounds in an environment, is a new fi eld that offers solutions for assessing biodiversity and disturbance. Spatial dynamics are an important component of an ecosystem, yet the link between spatial dynamics and soundscapes has not yet been studied. It should be possible to assess disturbance of an ecosystem by analyzing the spatial structure of the soundscape. Particularly, soundscapes in undisturbed ecosystems should exhibit more spatial autocorrelation than soundscapes in disturbed ecosystems. We took automated bihourly 10-minute recordings at a disturbed (n = 4,905) and an undisturbed (n = 8,450) transect in Brunei, Borneo, for a month-long period. Sites had identical spatial configurations, with 13 recorders at each site. Alpha acoustic indices were computed from recorded sound files, and beta acoustic indices were calculated pairwise for each 13 recorder group at each unique recording time. Moran’s I statistics were calculated to compare the resultant alpha indices from the two transects. We compared beta indices by Mantel Tests. Spatial statistics on alpha and beta indices reveal more spatial autocorrelation and correlation at the undisturbed site. Of all alpha indices, Normalized Difference Soundscape Index (NDSI) detected spatial autocorrelation to the highest extent. Beta acoustic indices detected differences between the two sites more precisely, due to their comparative nature.