Frogs croaking, crickets chirping, and birds singing from the trees – these sounds endear us to nature and may also be a powerful method for monitoring biodiversity. Affordable, automated acoustic recorders are now widely available, allowing researchers to examine ‘soundscapes’, or the collection of sounds on a landscape. From coral reefs to soybean farms, studies have linked the complexity of the acoustic environment to the health of a biological community. In this way, biodiversity can be assessed remotely over long time periods, with less disturbance to a landscape than surveys by human observers.
Many organizations have begun to amass enormous acoustic recording data sets. Because listening to these recordings in their entirety would be impractical, researchers have begun to use acoustic indices to measure diversity in the acoustic environment. Acoustic indices are based on the principle that each species uses a specific part of the acoustic environment to avoid overlapping sound signals; thus, in habitats with more species, more ‘acoustic niches’ will be full.
Our study summarized the different types of indices and how effective they were at indicating different types of biological information in previous studies. We then gathered the most successful indices and calculated them for acoustic recordings from 43 marine and terrestrial sites around the US. We found that by combining indices using machine learning, we were able to build models that accurately predict the diversity of birdsong within terrestrial recordings. The ability of models to predict bird song was hampered by the presence of constant sounds, including insects, wind, and anthropogenic noise. Models were less effective in marine recordings, likely because of the constant presence of snapping shrimp and wave noise.
With more research, soundscapes could be effectively used to monitor biodiversity and document its response to conservation efforts.
Further Reading:
Buxton RT, McKenna MF, Clapp M, Meyer E, Angeloni L, Crooks K, and Wittemyer G. In press. Efficacy of extracting indices from large-scale acoustic recordings to monitor biodiversity. Conservation Biology.
