Bioacoustics and Machine Learning for Avian Species Presence Surveys
2021-10-01, 14:05–14:10, Group on Earth Observations

The complex realities of changing climate and biodiversity are imperfectly understood. Bioacoustics is a conservation tool, going where human ears cannot stay and listen. Locally-informed machine learning analysis leads to big data insights, empowering informed decision making. Networks of bioacoustic recorders in some of Earth’s most biodiverse and vulnerable regions (near Everest in Nepal, Madidi National Park in Bolivia, and the Chesapeake Bay watershed in the United States) are bearing witness to a changing climate. More than 1850 days of audio data already collected provide a powerful dataset for studying species distributions. Machine learning (ML) turns this data into information to understand climate change and biodiversity. MLmodels are being trained for a dozen species in Nepal, Bolivia, and USA. Analyzed data show location and time of species vocalization. Modeling can expand rapidly as labeled data is collaboratively created by local experts. Preliminary results from Nepal show that a rare bird species was identified 1,000 feet higher in elevation than previously recorded: probable proof of concept that bird species are migrating uphill with changing climate. Bioacoustics is a valuable tool for species population surveys and biodiversity monitoring.

Please see the abstract above.

Authors and Affiliations

Naomi Bates, Songs of Adaptation Project Director and Associate Professor at Future Generations University


Transition to FOSS4G


FOSS4G implementations in strategic application domains: land management, crisis/disaster response, smart cities, population mapping, climate change, ocean and marine monitoring, etc.


1 - Principiants. No required specific knowledge is needed.

Language of the Presentation