A Reduced Complexity Framework for Evaluating Convection at High-Resolutions

Global radiative-convective equilibrium (RCE) configurations in atmospheric general circulation models (AGCMs) have shown to be useful reduced complexity frameworks for understanding model sensitivities. The role of convective parameterizations at high horizontal resolution and their impacts on clouds, circulation, and precipitation processes represent large uncertainties in current-generation AGCMs. As the statistical equilibrium in which radiative cooling is balanced by convective heating, RCE offers a simplified framework to investigate such uncertainties. The National Center for Atmospheric Research's Community Atmosphere Model 5 (CAM5) is configured in a RCE setup that consists of an ocean-covered earth with diurnally varying, spatially uniform insolation with no rotation effects. CAM5 is run with the spectral element dynamics package at a standard resolution of approximately 111 km grid spacing. A series of simulations is performed in which the planetary radius is incrementally reduced. Because of the homogeneity of the RCE setup, the effect is to bring the grid points closer together, mimicking increased resolution without increasing the number of grid points (or computational expense). These simulations have horizontal grid spacing of 56 km, 28 km, 14 km, and 7 km; the last of which begins to approach cloud-system resolving scales. When compared to high-resolution simulations with CAM5, the results suggest that the reduced Earth approach is able to reproduce the behavior of convection at reduced computational cost. At grid spacing less than 20 km, convective motions are predominantly produced by resolved scales.