Analyzing the Arctic's Underwater Noise with AI: Evaluating Correlation between Climate Change, Sound, and Marine Ecology

(US and Canada) Arctic sea ice is a keystone indicator of greenhouse-gas induced global climate change, which is expected to be amplified in the Arctic. The underwater acoustic environment in the Arctic Ocean contains a great deal of unique insights, such as behavioral changes of its marine species and sea ice melting. Through collaborations with Cochl, KOPRI, a research institute under the Korean Ministry of Oceans and Fisheries for scientific research of the polar regions, is exploring the application of sound AI to automate some of its repetitive tasks and to derive meaningful insights from AI-based quantification of its time-series audio data for analyzing climate change, biodiversity, and other environmental metrics.

KOPRI has been actively carrying out collection of ecological data from the Arctic Sea. Its recent studies include validation of the effects of the sea-ice conditions on the ambient noise level, where KOPRI analyzed its data collected from underwater microphones placed off the East Siberian Sea.

Analyzing the terabytes of acoustic data, however, is oftentimes a very daunting and complex task, as there is no automated system to "listen" to the sounds and extract key information such as mammal vocalizations from hundreds of hours of recordings. The complexity of the unique acoustic characteristics of the Arctic Ocean due to its unique geological features with sea ice makes it even more difficult to apply traditional acoustic monitoring techniques. Therefore, existing analysis tasks have inevitably involved biologists manually sorting out the sounds.

Cochl is a deep-tech startup advancing sound AI technology, founded by researchers in the audio and machine learning field. Cochl's core AI model has been applied in a wide range of applications to address industry-specific problems, from smart cities, manufacturing, and automotive to entertainment. For the initial phase in this collaborative research, Cochl will adapt its deep learning methods to clustering and categorizing underwater noises.

"We hope our sound AI technology could help bring a paradigm shift in the underwater acoustics studies, and we are excited for this unique opportunity to contribute to sophisticated studies of acoustic habitats and climate change," says Yoonchang Han, co-founder & CEO of Cochl.