A library of rain-shaft simulations for calibration of microphysical processes in clouds

Mitigating and adapting to global warming and the accompanying climate changes require reliable global and local predictions of the future climate. Future climate predictions of currently available models differ widely, primarily because climate models fail to predict the cloud response to climate change reliably. It is well-accepted that the largest source of uncertainties in climate models are the parametrizations of turbulence and clouds. Although we know the equations that govern the formation and evolution of atmospheric turbulence and clouds, climate models that resolve all relevant scales in these equations are computationally impossible to realize in the foreseeable future. Consequently, turbulence and clouds are parameterized in climate models. The parametrizations used by current climate models fail to model the effect of atmospheric turbulence and the role of clouds in the climate system well. In order to improve models, we thus need to improve parametrizations of clouds and turbulence. The goal of this project is to improve parametrizations of microphysical processes leading to rain formation in earth system models. This project relies on a theory-data-simulation tripod for developing coarse-grained models and calibrating these theory-based models by using high-resolution rain-shaft simulations. We create a Bayesian framework with an error model that accounts for structural errors in the developed climate models. Calibration of climate models with this framework yields parametrizations with quantified uncertainties.