One quarter of all marine species spend part of their life on coral reefs. Those reefs occupy less than one hundredth of one percent of the ocean floor. And right now, scientists cannot study them fast enough to keep pace with their decline.
The problem is not the science. It is the scale. Sending trained divers to survey a reef is expensive, limited in coverage, and physically demanding. A diver can cover a small area in a single dive. The reef does not wait.
A robot built at the Woods Hole Oceanographic Institution has just demonstrated something fundamentally different.
What CUREE Can Do
The robot is called CUREE — Curious Underwater Robot for Ecosystem Exploration. It looks like a small torpedo with four hydrophones extending outward on arms, plus forward and downward-facing cameras. It operates entirely on its own, with no human in the loop while it works.
What makes CUREE different from a standard underwater survey drone is how it decides where to go. Most survey robots follow a pre-programmed path. CUREE listens. It picks up biological sounds in real time — the characteristic snapping of shrimp colonies, the calls of individual fish species — and navigates toward them, homing in on areas of high biological activity from as far as 80 metres away. When it gets close enough, the cameras take over, providing the kind of detailed species-level observation that sound alone cannot deliver.
The combination is the key insight. Sound travels far underwater but tells you little about what exactly is there. Vision tells you precisely what is there but only over short distances. Together they form a system that can detect, navigate toward, and then characterise a biodiversity hotspot without any prior knowledge of where it might be.
In field trials conducted over three expeditions between 2022 and 2024 at a healthy reef in the US Virgin Islands called Joel’s Shoal, CUREE consistently found the same hotspot: an area around a large pillar coral structure. Fish density in that area was nearly 25 times higher than the surrounding reef. It confirmed this not just visually but acoustically, detecting elevated biological activity across a much wider area around the same structure.
In a separate demonstration, CUREE tracked a barracuda — a top predator — as it moved through its environment. Barracudas repeatedly return to ecologically important zones. Following one is a way of finding hotspots that no sensor array could locate on its own.
What This Changes
Coral reef science has a fundamental measurement problem. The fine-scale structure of reef biodiversity — which patches are thriving, which are degrading, why some spots support 25 times more life than others nearby — has been almost impossible to map. Surveys conducted by divers cover too little area at too low a frequency to capture this variation reliably.
CUREE can operate for hours. It can cover areas that would take a team of divers days to survey. It can return to the same reef repeatedly and compare results over time. And it does this at sub-metre resolution — the scale at which the ecology actually varies.
The researchers envision fleets of such robots deployed globally to explore and monitor reefs that remain largely uncharted. That is not an exaggeration of what the system can do. It is exactly what it was designed for. The same compact platform can be packed in checked luggage and deployed by a single person anywhere in the world, without the ships and support infrastructure that conventional research expeditions require.
This matters because conservation decisions have to be made now, under time pressure, across thousands of reefs that no team of scientists will ever physically visit. Knowing where the hotspots are — the specific patches where biodiversity concentrates, feeds, shelters, and reproduces — is the prerequisite for protecting what remains.
The Honest Limitations
CUREE is not a replacement for human scientists. The robot identifies where biodiversity is concentrated and confirms it with measurements. Understanding why, and deciding what to do about it, still requires biologists. The system has been tested at one well-studied reef in the US Virgin Islands — scaling it globally will require validating performance across the full range of reef types, depths, and acoustic environments.
The technology also has a ceiling. A reef that has already collapsed into an algae-covered rubble field has no biological sounds to home in on. CUREE finds hotspots where hotspots exist.
What it does, it does at a scale and resolution that changes what reef science can actually measure. That is the point.
Sources
McCammon, S., Cai, L., Yang, D., Walsh, J., Cast, J.D., Mooney, T.A., & Girdhar, Y. et al. (2026). Autonomous seeking and mapping coral reef biodiversity hotspots with a multimodal AUV. Science Robotics, 11(114). https://doi.org/10.1126/scirobotics.adx9939
Woods Hole Oceanographic Institution. (13 May 2026). Autonomous underwater robot discovers hidden coral reef “hotspots.” https://www.whoi.edu/press-room/news-release/curee-hotspots/
CUREE project background
WHOI WARP Lab CUREE project page: https://warp.whoi.edu/curee/
WHOI Reef Solutions Initiative: https://reefsolutions.whoi.edu/
Jane holds a BSc in Biology from the University of Regina and a Master of Science in Bioscience, Technology and Public Policy from the Univesity of Winnipeg. Her reporting interests include Life Sciences, Physical Sciences and the Cosmos.