- Scientists have combined passive acoustic monitoring, machine-learning tools and aerial surveys to estimate the population of North Atlantic right whales in Cape Cod Bay.
- Using the method, researchers from Cornell University in the U.S. were able to estimate the daily population of the whales over a period of four months.
- While passive acoustic monitoring has helped scientists around the world detect the presence of whales, it’s often challenging to estimate population numbers from the data, especially for species like North Atlantic right whales that have highly variable call rates.
What better way to track whales than listening in on them?
Passive acoustic monitoring, in which microphones are placed underwater to pick up any sounds, has long helped scientists detect the presence, or absence, of whales in oceans. More often than not, however, the method isn’t that great at estimating the population of whales.
To fill that gap, scientists have combined audio data with aerial surveys and machine-learning tools to count whale populations.
A new study published in the journal Endangered Species Research describes how researchers at Cornell University in the U.S. used a combination of these technologies to estimate the abundance of North Atlantic right whales (Eubalaena glacialis) in Cape Cod Bay in the U.S. northeast. While underwater microphones and machine-learning tools were deployed to estimate population, aerial surveys were used to calibrate and verify the data.
“It has been challenging to make the leap of faith for using whale calls to signify information beyond presence or absence in an area,” Marissa Garcia, lead author of the study and researcher at the K. Lisa Yang Center for Conservation Bioacoustics at the Cornell Lab of Ornithology, told Mongabay in a video interview. “With this research, we are seeking to advance passive acoustic monitoring to address questions beyond just whale presence.”
North Atlantic right whales are one of the most threatened whale species in the world, with only 356 individuals estimated to be left, according to a 2022 survey. As they migrate along the east coasts of the U.S. and Canada, they often face threats of vessel strikes and entanglement in fishing gear. That makes estimating their populations and identifying how they’re moving crucial to enforcing effective conservation strategies.
Aerial surveys can be expensive and are often contingent on weather conditions. And acoustic monitoring on its own can really only deal with population estimates for species that produce “rhythmic, metronomic and highly predictable songs or clicks,” Garcia said. North Atlantic right whales, on the other hand, have highly variable calling rates, making it challenging to deduce their numbers just from the acoustic data. “We need the calling rates to convert data from the sounds we hear to the numbers that exist,” Garcia said.
By combining acoustic data with aerial surveys, the team managed to find a way around the problem. Thanks to decades of aerial whale surveys conducted by the nonprofit organization Center for Coastal Studies, they were able to cross-reference the visual sightings with the acoustic data. Garcia said this then helped them ground truth (or verify), calibrate, and make sense of the acoustic data.
Starting in February 2019, the team deployed five underwater microphones, or hydrophones, across Cape Cod Bay. The microphones, which the team called marine acoustic recording units (MARU), were designed and engineered specifically for the purpose.
Four months later, they retrieved the recorders and synchronized the data to match the audio from the five microphones. This was followed by the arduous task of manually going through the data and annotating a set of “up” calls, whale vocalizations that sound like a “whoop” and are most commonly used to detect the presence of right whales. The annotations of these calls formed the basis for creating a machine-learning model.
“After validating the model, we were able to quantify its performance and apply this model across all of the data which gave us hundreds of thousands of these calls,” study co-author Irina Tolkova, a postdoctoral fellow at the Yang Center, told Mongabay in a video interview.
After accounting for duplicate calls (the same audio captured by different recorders), the team was able to deduce the total number of whale calls per day in Cape Cod Bay. Over the four months, this came to 250,000 unique up calls.
With the data set in hand, the team matched it against the data gathered from 16 aerial surveys that coincided with the course of the deployment. “Then we could see, as the plane is flying over the bay and the number of whales is being counted from an airplane, how many calls are being made during that same time interval,” Tolkova said. “Those two values gave us a statistical relationship that we could quantify and apply across the whole acoustic survey to give us a population estimate for that specific day.”
The method isn’t without its challenges. The limited amount of aerial survey data, which serves as the ground-truth data, often makes it difficult to derive more information from acoustic data beyond the presence or absence of the animals. “It’s still subject to things like surveys being cut short because of poor weather conditions,” Tolkova said. “Combining acoustic data with aerial surveys makes it a more fine-scale, long-term abundance estimate.”
Garcia and Tolkova said there are still uncertainties in their estimates that they plan to work on and address in the future. Nonetheless, they said the method, once further refined, could potentially be applied across the entirety of the whales’ migration route.
“This is just one of several critical grounds for North Atlantic right whales in their journey along the U.S. East Coast,” Garcia said. “We were hoping to identify a method that is relatively low-effort and ultimately scalable across this migratory corridor.”
Banner image: North Atlantic right whales (Eubalaena glacialis). Image courtesy of NOAA. (CC BY 2.0).
Abhishyant Kidangoor is a staff writer at Mongabay. Find him on 𝕏 @AbhishyantPK.
Citation:
Garcia, M., Tolkova, I., Madhusudhana, S., Rahaman, A., Baker, C., Mayo, C., … Klinck, H. (2025). Acoustic abundance estimation for critically endangered North Atlantic right whales in Cape Cod Bay, Massachusetts, USA. Endangered Species Research, 56, 101-115. doi: 10.3354/esr01384
Credits
