Scientists have developed an innovative technique to measure ocean currents with unprecedented detail and frequency, addressing a long-standing gap in ocean observation. The method, known as GOFLOW (Geostationary Ocean Flow), uses deep learning to analyze thermal imagery from existing weather satellites, eliminating the need for new hardware. By leveraging satellites already in orbit, this approach represents a major advancement in understanding how the ocean moves and interacts with Earth’s climate system.
Understanding ocean currents is essential because they regulate global climate by transporting heat, redistributing nutrients, and moving carbon between the atmosphere and deep ocean. These processes support marine ecosystems and influence weather patterns worldwide. Currents also play a critical role in practical applications such as search-and-rescue missions and tracking environmental hazards like oil spills. Despite their importance, measuring these dynamic systems across large ocean areas has been difficult, particularly at smaller scales where rapid changes occur.
Traditional methods have significant limitations. Satellite techniques that estimate currents using sea-surface height often revisit the same location only every ten days, making them too slow to capture fast-changing features. Meanwhile, ship-based instruments and coastal radar systems provide detailed data but only over limited regions. This leaves a major observational gap, especially for small-scale phenomena—often less than 10 kilometers wide—that drive vertical mixing in the ocean. These mixing processes are crucial because they bring nutrients to the surface and transport carbon dioxide to deeper layers, helping regulate Earth’s climate.
GOFLOW overcomes these challenges by transforming high-frequency weather satellite imagery into detailed maps of ocean currents. Researchers trained a neural network using high-resolution simulations of ocean circulation, teaching it to recognize how temperature patterns on the ocean surface shift under the influence of currents. By analyzing consecutive thermal images—captured as frequently as every five minutes—the system can infer water movement by tracking how temperature patterns deform over time. This allows scientists to generate hourly maps of ocean flow, capturing rapid and complex changes that were previously unobservable.
The method was validated against real-world data collected in the Gulf Stream region. GOFLOW’s results closely matched measurements from shipboard instruments and conventional satellite techniques but provided much finer detail. It revealed small, fast-moving eddies and boundary layers that older methods blurred or missed entirely. This new level of resolution enables scientists to directly observe key processes that were previously only modeled in simulations, offering fresh insights into how the ocean absorbs heat and carbon.
Beyond advancing scientific understanding, GOFLOW has practical implications. Because it relies on existing satellite infrastructure, it can be scaled globally and integrated into weather forecasting and climate models. Improved monitoring of ocean currents could enhance predictions of weather patterns, marine ecosystem changes, and the movement of debris or pollutants. However, the method does face limitations, particularly from cloud cover, which can obscure thermal imagery. Researchers are working to combine additional data sources to overcome this challenge and achieve continuous global coverage.
Overall, GOFLOW represents a significant leap forward in ocean science, providing a powerful new tool to observe and understand the complex behavior of the ocean in near real time.
https://phys.org/news/2026-04-deep-weather-satellite-thermal-imagery.html