On April 20, 2026, a research team led by scientists from the National Center for Atmospheric Research (NCAR) and the University of Miami published a study in Nature Geoscience detailing a new artificial intelligence framework named GOFLOW. This technique utilizes thermal imagery from existing geostationary weather satellites to measure ocean surface currents with a level of detail and frequency previously unattainable through traditional satellite altimetry.
The GOFLOW system—short for Geostationary Ocean Flow—employs deep learning architectures, specifically convolutional neural networks, to track the movement of sea surface temperature (SST) patterns. By analyzing sequential images captured every 10 to 15 minutes by satellites such as the Geostationary Operational Environmental Satellite (GOES-16 and GOES-18) and Japan’s Himawari-9, the AI calculates the velocity and direction of surface water. This approach differs from conventional methods that rely on radar altimeters to measure sea surface height anomalies, which often lack the temporal resolution to capture rapidly changing coastal and sub-mesoscale features.
Technical specifications released in the study indicate that GOFLOW achieves a spatial resolution of 2 kilometers. This represents a twelve-fold improvement over standard global ocean current products, which typically operate at a 25-kilometer scale. Furthermore, the AI-driven model provides hourly updates on current vectors. According to the researchers, this high-frequency monitoring is essential for understanding sub-mesoscale eddies—oceanic swirls spanning 1 to 10 kilometers—which are responsible for a significant portion of the vertical transport of heat, carbon, and nutrients between the surface and the deep ocean.
The study authors noted that GOFLOW was validated using a comprehensive dataset from the Global Drifter Program and ship-mounted acoustic Doppler current profilers (ADCPs). In high-energy regions such as the Gulf Stream and the Agulhas Current, the model demonstrated a root-mean-square error (RMSE) of less than 0.15 meters per second, outperforming existing numerical weather prediction models. This validation confirms the model's ability to maintain accuracy across diverse oceanic conditions and thermal gradients.
Because GOFLOW utilizes the existing constellation of meteorological satellites, it provides a cost-effective solution for global ocean monitoring without the need for new orbital hardware. The researchers emphasized that the system is currently capable of processing data in near-real-time, which has immediate implications for maritime safety, oil spill trajectory modeling, and the optimization of shipping routes to reduce fuel consumption. The study concludes that the integration of GOFLOW into global climate monitoring systems will provide more accurate data for long-term climate projections by better quantifying the ocean's role as a heat sink.