MIT-linked project uses AI to help rebuild beaches, plans expand to US

In short: A climate-tech effort rooted at MIT is using AI and ocean data to place underwater structures that help waves and currents pile up sand, with expansion planned beyond the Maldives to places like Boston and Miami.

What's going on

Researchers at MIT’s Self-Assembly Lab have been testing an idea that sounds simple: the ocean can rebuild a beach if you guide the water in the right way. In large wave tanks, they used water, sand, and controlled waves to learn how sand naturally forms sandbars and small islands.

The team’s approach uses submerged or partly submerged shapes on the seafloor. These shapes do not pump sand or move it with machines. Instead, they act like a gently placed ramp or barrier that changes how water flows, so sand drops in certain spots instead of being carried away.

AI, meaning computer systems that learn patterns from lots of examples, helps decide what to build and where to put it. The project uses satellite images and ocean-current data, including about a decade of records in some cases, to predict where sand is being lost and where it could collect. This lets the team test different designs on a computer first, like trying a few layouts on a map before doing construction.

Full-scale tests in the Maldives have shown measurable sand gains. One early setup reported about 0.5 meters of new sand in roughly four months, and another reported a new sandbar that extended a beach about 90 feet.

What to watch

A company called Coastal Assembly has formed from this MIT-rooted work and is aiming for wider coastal projects, including in US cities such as Boston and Miami. The key question is whether results seen on low-lying tropical islands will hold up on busy urban coastlines with different storms and wave patterns.

Source: NYTimes