In Support of Sound Science: Bioacoustics Training at Duke Farms

If a tree falls in the forest and no one is around to hear it, does it make a sound?

Well, that depends on how you’ve configured your field recorder.

Perception unfettered–or, perhaps more importantly, unaltered–by human presence: that’s the power of bioacoustics, a rapidly growing technology which, thanks to advances in machine learning, allows practitioners to record and listen to soundscapes, efficiently process large audio datasets, and identify animals through their vocalizations.

Its potential to expand how ecological monitoring happens (and who gets to do it) has stirred up excitement in the scientific community in recent years, and some fascinating studies have begun to bear that excitement out. Whether they’re told in chirps, calls, croaks, echolocation, underwater gurgling, infrasound rumblings across the savannah or serenades that span an ocean–these days, someone somewhere is listening in on these stories of planetary change.

While the range of bioacoustics targets may be expansive, the adoption of this emerging technique is less encompassing. Plenty of conservation NGOs, researchers, and government agencies recognize the promise of bioacoustics, and many have already begun deploying ARUs and collecting acoustic data from lands they aim to protect. But programming a recorder to listen several times a day, sometimes for months at a time, unsurprisingly gives rise to vast amounts of data that are both a blessing and a curse. The trends or correlations contained within may be of immense empirical value, but figuring out which statistical analyses can reliably unearth them? That’s the hard part.

Since we first started listening to the land here several years ago, we’ve fielded questions about our bioacoustics program from colleagues across the conservation field who also seek to inform land management decisions or conduct research using these tools. Last week’s course was our answer to these requests.

Over three days, a dedicated group of conservation professionals representing land trusts, county government, nonprofits, and consulting firms–each of them wrangling their own large datasets–gathered at Duke Farms to learn from the best: a trio of researchers from Cornell’s Lisa K. Yang Center for Conservation Bioacoustics who visited our campus to share their expertise.

Under the guidance of scientists Ben Gottesman, Tim Boycott, and Trifosa Simamora, participants split their time between farm fields and coding environments. They marveled at bobolinks before training algorithms on their song. They worked through case studies drawn from the Cornell team’s research, then developed or refined individual projects relevant to their own work. The result of this collective work was a cohort of professionals ready to bring all the potential of bioacoustic technology to bear on the landscapes they steward.

This workshop, along with January's "GIS for Conservation" course and even more to come this summer, reflects a core aspect of the Center for Conservation's mission: opening pathways from scientific insight to on-the-ground implementation. Expanding access to emerging tools like these is an essential step toward broader adoption in applied conservation and more enduring protections for our land.