Machine learning unveils near-surface foraging risk for endangered North Atlantic right whale

What is it about?

By deploying high-resolution animal trackers called CATS tags, we were able to record detailed movement and underwater video from a right whales point of view. By watching these videos, we were able to collect enough data to train machine learning algorithms how to identify when a right whale opens it's mouth. Our trained algorithms were applied across all of our data, meaning we can predict when a whale's mouth is open even when the camera isn't rolling or can't see the whales mouth. We found that right whales spend almost all of their time feeding in the Gulf of St. Lawrence, but what is really cool is that we observed a change in feeding between day and night. In the day, right whales feed close to the seafloor, but at night they move closer to the surface of the water.

Featured Image

Photo by David Trinks on Unsplash

Photo by David Trinks on Unsplash

Why is it important?

The North Atlantic right whale is a critically endangered species found across the Eastern seaboard of North America. These whales face constant risks from human industry, with entanglement and ship strikes as the leading causes of mortality. Our work highlights the risk that right whales face in their summer foraging habitat in the Gulf of St. Lawrence, Canada. During the day, their near-seafloor foraging places them at risk of entanglement in bottom-set fishing gear, whereas during the night their surface feeding exposes them to potential ship-strike risk. The latter behaviour had previously been unrecorded, and is amplified by the fact that ships rely on visual detection of whales to avoid them at night, this quickly becomes irrelevant at night.

Perspectives