Engineers at the American tech company were contacted by the U.S. National Oceanic and Atmospheric Administration for help a few years ago in trying to find the sounds of humpback whales in a huge amount of underwater audio. This became part of the AI for Social Good campaign.
The group then decided to shift its focus to a whale population that is experiencing dwindling numbers due to pollution, lack of chinook salmon, and marine traffic. The local killer whales are listed as endangered under the Canadian Species at Risk Act.
The idea is to develop a system to identify the southern resident whales by their calls. This way ships and boats in the same area can be alerted to their presence and steer clear.
Google’s engineers worked with Fisheries and Oceans Canada and identified the whales’ calls in about 1,800 hours of underwater audio. The calls and other whale sounds were then labeled to train a neural network. The underwater microphones, called hydrophones, are set up in different locations in the Salish Sea. They’re said to be sophisticated enough already to send a real-time alert back to the Marine Mammal Unit at the DFO when an orca is nearby.
According to the CBC, the group has been able to distinguish different kinds of whales from each other by studying and recording their sounds but they do not yet have enough data to separate the resident killer whales from the healthier transient populations.
“One of the primary threats that the whales are facing are entanglement and just difficulty foraging due to vessel noise,” Google software engineer Matt Harvey told the CBC.
“So they can use this [alert system] to advise traffic or the vessels themselves that orcas are in this location, consider slowing down or consider change of course.”
While Harvey uses the word “consider” it is actually law that vessels must change course and stay 400 meters away from whales. These rules took effect in June.
We’re behind anything and everything that makes an effort to save our 73 resident orcas.