Research integrates physics-based models and machine learning to reduce uncertainty in weather and climate predictions. Emphasis on data-driven parameterizations, cross-scale consistency, and using observations and simulations to better represent cloud and precipitation processes.