UVic Biologists Harness AI to Decode Unique Fish Sounds

Researchers at the University of Victoria have made a significant breakthrough in understanding fish communication by identifying unique sounds produced by different species. Utilizing advanced machine learning techniques, the biologists determined that eight fish species native to Vancouver Island generate distinct sounds, enabling them to classify these calls with an impressive accuracy of 88 percent.

The study, led by PhD student Darienne Lancaster, employed a method known as passive acoustic monitoring. This approach involves collecting underwater audio and video data using a specialized sound localization array developed by former UVic PhD student Xavier Mouy. The research team analyzed the audio recordings to pinpoint differences in sound characteristics such as duration and frequency, which are critical for distinguishing between species.

Decoding Fish Communication

Lancaster expressed enthusiasm for the findings, stating, “We knew previously that many fish were making sounds in the wild, but we didn’t know which sounds belonged to which species, or if it was possible to tell these sounds apart.” The study revealed that the black rockfish produces a long, growling sound reminiscent of a frog’s croak, while the quillback rockfish emits a series of rapid knocks and grunts.

The research highlights specific behaviors associated with these sounds. For instance, quillback rockfish create rapid grunting noises when threatened by predators, suggesting a potential defensive mechanism. Conversely, the copper rockfish is known to make knocking sounds while pursuing prey along the ocean floor.

Global Implications of the Research

The techniques developed by Lancaster and her team could have far-reaching implications for marine biology. The AI model utilized a set of 47 different sound features to effectively differentiate between species calls, which can be adapted by scientists globally to analyze other fish sounds.

This groundbreaking research was funded by the Natural Sciences and Engineering Research Council of Canada and Fisheries and Oceans Canada. As scientists continue to explore the underwater soundscape, the ability to identify and understand fish communication may lead to deeper insights into their behavior, ecology, and conservation efforts, marking a significant advancement in marine research.