Title: Diffusion model for stochastic parameterization– a case example of numerical precipitation estimation
Lecturer: Dr. Baoxiang Pan (The Institute of Atmospheric Physics, Chinese Academy of Sciences)
Inviter：Prof. Shuguang Wang
Time: Tuesday March 20, 2023 at 9:30 AM
Venue: Lecture Hall C409, School of Atmospheric Sciences, Xianlin Campus
Abstract: Estimating the unresolved geophysical processes from resolved geophysical fluid dynamics is the key for improving numerical weather-climate predictions. While data-driven parameterization for unresolved geophysical processes shows potential, most practices fail to capture the diversity of unresolved geophysical processes that are consistent with any resolved geophysical fluid state. This pitfall undermines the likelihood or severity of simulated weather extremes, and erodes the fidelity of climate projections. We introduce diffusion model, a non-equilibrium thermodynamics inspired deep generative modeling approach for stochastic parameterization of unresolved geophysical processes. Advantage of the proposed methodology is demonstrated via comparison to other popular deep learning methods (UNet, variational autoencoder, generative adversarial net), as well as dynamical downscaling method (WRF).