Yanni Yuval, MIT
"Machine learning for parameterization of moist processes in the atmosphere"
The representation of sub-grid processes contributes to the uncertainty in climate prediction. Specifically, the parametrization of convection and clouds governs the uncertainty in rainfall distribution, severe storm frequency and temperature changes due to global warming. Increasing number of studies show that machine learning can be used to build data-driven parameterizations directly from high-resolution model output. Unfortunately, the resulting parameterizations do not always lead to stable and accurate simulations when implemented in a coarse-resolution model. Previous work suggests that a machine learning approach based on an ensemble of decision trees (random forest) can be used to robustly emulate a conventional moist convection scheme. Here we describe how random forests are used to learn a sub-grid parameterization from a coarse-grained output of a quasi-global high-resolution simulation with hypohydostatic rescaling. We discuss the performance of the parameterization when implemented at coarse resolution in the same model with a focus on statistics of mean and extreme precipitation. Furthermore, we discuss the performance of the parametrization in different coarse resolutions, and demonstrate that highly accurate parametrizations, does not necessarily lead to accurate simulations.