How Dark Fiber Is Tracking Waves in the Frozen North

Sandia National Laboratories tapped an active Alaska subsea cable to track Arctic waves, proving telecom infrastructure can double as a smart sensor

22 May 2026

Submarine cables carry the vast majority of global internet traffic across ocean floors. Now, researchers at Sandia National Laboratories have proven these same telecom lines can pull double duty as advanced ocean wave sensors. By applying machine learning to a 27-kilometer stretch of fiber-optic cable in the Arctic, the team successfully extracted precise wave data without dropping a single piece of new hardware into the sea.

The experiment tapped into an existing cable offshore Alaska using distributed acoustic sensing, a technique that measures tiny vibrations and strain on the fiber. Scientists fed this movement data into three distinct machine learning models, training them to translate the cable's shifting tension into seafloor pressure and surface wave patterns. The top-performing model mapped the ocean surface with remarkable accuracy, turning standard utility infrastructure into a highly sensitive radar.

Measuring Arctic waters usually requires deploying specialized marine equipment into brutal, freezing environments that test the absolute limits of technology. Utilizing pre-existing fiber infrastructure bypasses those logistical nightmares entirely, opening up a continuous stream of real-time climate data for a fraction of the cost. The project holds immediate value for coastal hazard tracking, shipping logistics, and national security infrastructure strategy in rapidly changing northern waters.

Shallow Arctic coastlines offer ideal conditions for this technique, meaning the same crisp data might be harder to capture in the deep, open ocean. Accessing the unused "dark fiber" required for these tests also depends on the cooperation of private telecom companies. Even with those hurdles, prior tests on this same Alaska cable have successfully tracked moving sea ice and thawing permafrost.

As machine learning software grows more sophisticated, the line between global communications and planetary science continues to blur. These underwater data highways are no longer just for handling internet traffic. With the right software, the global cable network could soon become one of the largest scientific instruments on Earth.