The ocean has a murmur of its own, and science can finally put a name to it. New technology allows "listen" to the fish and identify what species sounds at all times, opening a new window into the life of coral reefs without disturbing their inhabitants.
The development, led by the FishEye Collaborative together with Cornell University and Aalto University, combines spatial audio with 360° video to accurately associate each sound with the fish that produced it. The proposal, published in the journal Methods in Ecology and Evolution, represents a practical leap forward for marine conservation.
How the reef's "ear" works
The tool, a Underwater Passive Acoustic Camera (UPAC-360), integrates spatial audio hydrophones and a 360° camera. This allows researchers to locate the source of each sound signal and superimpose it onto the panoramic image to see who is "speaking" in real time.
This immersive approach transforms the reef soundscape into a synchronized visual and auditory track. The direction of the sound is represented as a trail in the video., so when a fish hits, creaks or grunts, the system allows direct evidence to be used to attribute the fish's responsibility.
Another key advantage is its autonomous deployment: the equipment is installed and remains recording for long periods, without the need for divers or boatsThis way, natural behaviors are recorded, preventing human presence from altering the results.
Overcoming background noise has been one of the biggest challenges. A reef is full of shrimp snaps, creaks and bangs, as well as multiple fish transmitting at once. Spatial audio helps separate sources, even when individuals swim in groups and signals overlap.
What have they found so far?
In the reefs of CurazaoThe system allowed sounds to be attributed to 46 species, more than half of which had never been recorded as sound. This set constitutes the largest collection of acoustic signals. de peces published to date.
The team estimates that more than 700 species in the Caribbean could produce sounds, so the data collected is just the beginning. At the same time, have released a growing library of tagged audios which will be used to train machine learning algorithms capable of detecting species in future recordings.
The idea of ​​a marine "Merlin"—like the Cornell Lab of Ornithology's bird identification app—is still a long way off, but the groundwork is already in place. Every sound attributed makes it easy to automate recognition, bringing the possibility of rapid queries on large volumes of underwater audio.
In addition to Curaçao, the approach is being applied and validated with sustained field campaigns on other reefs. Long time series allow comparison of hours, seasons and events, detecting subtle changes in the acoustic activity of communities de peces.
Conservation, management and next steps
Shallow tropical reefs cover only 0,1% of the seafloor, but are home to nearly 25% of all marine speciesTheir decline due to global warming, pollution, and overfishing makes it vital to have sensitive indicators of health and resilience.
Listening to fish provides precisely that pulse. Soundscapes offer signs of presence, reproduction and activity of key species, which helps evaluate restorations, protected areas and responses to episodes of environmental stress.
For managers and NGOs, having passive, continuous and non-invasive monitoring allows invest resources better: identify where to act, what measures work, and when to adjust efforts. The team plans to expand the Caribbean acoustic library and extend deployments to Hawaii and Indonesia in new campaigns.
The project leaders emphasize that, although there is no "universal assistant" that can identify all species instantly, the scientific utility is immediate: from detecting rare or invasive species to documenting mating periods and changes in community composition.
With this underwater "ear", research has taken a practical turn: giving voice to fish is not a simple curiosity, but a tool that allows monitor reef health with precision and in real time, key to protecting those who depend on these ecosystems and guiding data-driven policies.
