Q&A: Physical scientists turn to deep learning to improve Earth systems modeling – IAM Network

Examples of typical deep learning tasks (left panel) and the corresponding Earth system science problems that they can be applied to: a, Object recognition in images relates to detection of extreme weather patterns in climate data; b, Super-resolution relates to downscaling of climate data; c, Video prediction relates to forecasting of Earth system variables; d, Language translation relates to modeling of dynamic time series. The role of deep learning in science is at a turning point, with weather, climate, and Earth systems modeling emerging as an exciting application area for physics-informed deep learning that can more effectively identify nonlinear relationships in large datasets, extract patterns, emulate complex physical processes, and build predictive models. "Deep learning has had unprecedented success in some very challenging problems, but scientists want to understand exactly how these models work and why they do the things they do," said Karthik Kashinath, a computer scientist and engineer in the Data & Analytics Services Group (DAS) at the National Energy …