Research/
Neural Networks Map Complex Ocean Dynamics from Satellite Data
Researchers have developed a machine learning framework to reconstruct submesoscale ocean currents using sparse satellite observations. By training neural networks on high-resolution simulations, the model can accurately predict fluid dynamics in regions where physical sensors are absent or data is noisy.