Deep Learning in Geophysical Fluid Dynamics

Course provides a survey of the rapidly growing field of physics-informed deep learning, which integrates known physics principles with neural networks to predict the behavior of a physical system. It both introduces the background knowledge required to implement physics-informed deep learning and provides practical in-class coding exercises. Students gain experience applying this emerging method to their own research interests, including topics in geophysical fluid dynamics (atmospheric, oceanic or ice dynamics) or other nonlinear systems where the same technique applies. Students develop individual projects throughout the semester.