Super Droplet Cloud Microphysics in the Energy Research and Forecasting Model

We implement the Super Droplet Method (SDM) in the Energy Research and Forecasting Model (ERF) and probe its physical fidelity and computational performance. A benchmarking exercise against PySDM verifies that ERF-SDM produces similar rain and cloud profiles in a 1D rain shaft configuration. A strong scaling study demonstrates that ERF-SDM scales well on both CPUs and GPUs with the latter almost completely ameliorating the computational penalty of running SDM in Large Eddy Simulations. To validate the new model, we contrast rain size distributions collected during the Rain in Cumulus over the Ocean (RICO) campaign against model simulated distributions and find excellent agreement in the 10 mm to 100 mm size range. We further demonstrate the readiness of ERF-SDM for cloud system level investigations by tackling aerosol impacts on drizzling shallow cumulus clouds. ERF-SDM’s rain initiation is far less sensitive to the underlying aerosol compared to a traditional bulk scheme, which leads it to exhibiting weaker precipitation and albedo susceptibilities in simulations of shallow cumuli. This underscores the relevance of retackling longstanding paradigms (such as aerosol-cloud interactions) with advanced simulation tools like ERF-SDM.