GOFLOW AI Tracks Ocean Currents

Scientists, led by Luc Lenain of Scripps Institution of Oceanography and Kaushik Srinivasan of University of California, Los Angeles, have introduced GOFLOW, a new deep learning technique to track ocean surface currents with unprecedented detail. Published in Nature Geoscience, this method utilizes thermal images from existing weather satellites like GOES-16, eliminating the need for new hardware. GOFLOW provides hourly maps of currents by analyzing how temperature patterns shift, offering a major breakthrough in understanding ocean dynamics, vertical mixing, and the global transport of heat and carbon. The project received funding from the United States — Office of Naval Research, United States — NASA, and the Australia — Australian Research Council. This innovation is expected to improve weather forecasting, climate models, and predictions related to marine debris and ecosystem dynamics, despite current limitations due to cloud cover.