Skip to main content
Computing, Environment and Life Sciences

Monitoring Biodiversity Using Acoustic Data

Long-term analysis of soundscapes using machine learning and AI will provide a baseline for animal biodiversity and how it is changing over time.

Currently, the only studies of animal biodiversity examine only a few species at a time in small areas. Large-scale studies are neither cost- nor time-efficient, greatly limiting how well researchers can understand ever-changing animal biodiversity. To keep up with the evolving nature of biodiversity, machine learning and artificial intelligence can be used to analyze acoustic data to determine the number of different species in certain areas.

Animals make sounds for a wide variety of reasons: to communicate about territory, food, predators, mating, and much more. All of these sounds contribute to what is called a soundscape. Soundscape ecology is the branch of science that analyzes acoustic data from soundscapes to understand the relationships between biodiversity, landscapes, and anthropogenic pressures.

To demonstrate the effectiveness of the acoustic data analysis approach, Argonne selected six pilot sites, including the Morton Arboretum. Bioacoustic audio recording devices have been deployed in the arboretum to collect acoustic data that will be used to understand the soundscape and biodiversity in the area.

Long-term analysis of soundscapes will provide a baseline for animal biodiversity and help determine how it is changing over time. Argonne used a variety of devices in the study, as shown below.

Battery-powered bioacoustic audio recorders (BARs) can be easily fastened to trees where they blend in and constantly record audio.

Solar BARs are charged by a solar panel. If they are deployed in a place with consistent sunlight, they can record continuously for long periods of time.

Development of a baseline for different measurements of biodiversity that cover large areas and many species would offer a number of benefits, including the ability to quantify the differences between protected and non-protected areas, better estimate the effects of land-use restriction on the conservation of certain species; and compare future data with the baseline measurements to provide information on the migration of a species and what might have caused it.