AI system maps ocean currents hourly using existing weather satellites
Scientists have developed Geostationary Ocean Flow, or GOFLOW, a deep learning system that converts images from existing weather satellites into hourly ocean current maps. This approach captures fast-moving small-scale eddies previously hard to observe. The method appears in a study published Monday in Nature Geoscience.
The system, codirected by oceanographer Luc Lenain at Scripps Institution of Oceanography at UC San Diego and Kaushik Srinivasan now at UCLA, trains a neural network on high-resolution computer simulations of ocean circulation. It learns to detect how sea surface temperature patterns shift under underlying currents. Fed consecutive thermal images from the GOES-East geostationary satellite, it infers velocities producing those changes and generates detailed hourly current maps without new space instruments.
Lenain noted that weather satellites have long observed ocean surfaces. The key advance extracts hourly current maps from time-lapse images by tracking how temperature patterns curve, stretch, and move hour to hour.
Researchers validated GOFLOW against shipboard instruments in the 2023 Gulf Stream region and standard satellite estimates from ocean surface height. It matched large-scale data but showed far greater small-scale detail, resolving currents at about 10 kilometers with hourly updates. This surpasses the 21-day revisit cycle of the Surface Water and Ocean Topography satellite or daily altimetry averages from AVISO.
For the first time from real observations, the team measured statistical signatures of intense submesoscale currents driving vertical ocean mixing, previously seen only in models. Lenain said this will test long-held ideas on how oceans absorb heat and carbon.
GOFLOW relies on existing satellites, so it could integrate into weather forecasts and climate models. It would better depict fast-changing currents affecting air-sea exchanges, marine debris transport, and ocean ecosystems. The framework applies to all ocean basins under geostationary satellites, paving the way for global high-resolution current monitoring.
-
15:20
-
15:15
-
15:01
-
15:00
-
14:45
-
14:40
-
14:30
-
14:20
-
14:15
-
14:00
-
13:50
-
13:45
-
13:33
-
13:20
-
13:17
-
13:05
-
13:00
-
12:40
-
12:30
-
12:20
-
12:15
-
12:00
-
11:50
-
11:45
-
11:40
-
11:30
-
11:20
-
11:17
-
11:15
-
11:09
-
11:00
-
11:00
-
10:45
-
10:34
-
10:33
-
10:30
-
10:15
-
10:00
-
10:00
-
09:45
-
09:40
-
09:30
-
09:20
-
09:17
-
09:15
-
09:01
-
09:00
-
08:45
-
08:40
-
08:30
-
08:20
-
08:15
-
08:00
-
07:50
-
07:45
-
07:30
-
07:15
-
07:06
-
17:20
-
17:00
-
16:40
-
16:20
-
16:00
-
15:40
