NSF awards UAF $50,000 to test market for satellite depth-mapping AI

A University of Alaska Fairbanks effort to map water depth from satellite imagery drew a $50,000 award from the National Science Foundation on Aug. 19 to test its commercial potential.

The award, with Erin D. Trochim as principal investigator, runs from Sept. 1 through Aug. 31, 2027, under I-Corps, the NSF's "seven-week experiential training program" for moving lab research toward commercialization.

According to the NSF award abstract, the framework combines physics-informed neural networks, models that build known physical laws into their structure, to map bathymetry (underwater depth and terrain) across coastal, river and lake systems. River depth and velocity are derived from satellite altimetry and model-derived daily discharge data. The system returns calibrated uncertainty bounds flagging prediction reliability, and it holds sub-meter accuracy in turbid waters, where satellite-derived methods have historically degraded.

Nearly half of U.S. waters remain unmapped to modern standards, the abstract states, and ship- and aircraft-based surveys are too costly and slow to close that gap. The technology's intended uses include flood forecasting, coastal engineering and navigation, with automated updates after major storms, for government agencies, engineering firms and environmental data providers.

At the 2022 Alaska Coastal and Ocean Mapping Summit, where Trochim presented, Alaska mapping practitioners identified data processing capacity and expertise gaps as needs standing alongside the demand for new technology itself.