Bayesian Inversion and LES modeling

Subgrid-scale turbulence parameterization

Bayesian inversion in weather and climate models — Project partners: IAC, SAM and MCH

Unknown model parameters used in subgrid models of modern weather and climate models are one of the main sources of uncertainties in the projections of future regional climate change. In a new project funded by an ETH project grant and in collaboration with ETH's Seminar for Applied Mathematics (Prof. Dr. Siddhartha Mishra), we investigate how recent advances in Bayesian inversion can be used to improve the representation of the small-scale flow unresolved in large-eddy resolving simulations (LES).

News

The ETH AI Center faculty workshop on "AI in Scientific Discovery" was held on 27 April 2023 and Dana Grund presented our collaborative project. [April 2023]
Dana Grund started her Ph.D. topic on machine-learning enhanced Bayesian inversion for subgrid-scale turbulence model development. [Jan 2023]

LES Simulation of a convective boundary layer — LES simulation of a convective boundary layer with varying model resolution and Smagorinsky constant (Courtesy Dana Grund).