The Interaction between Cloud, Radiation and Turbulence and the Self-Aggregation of Convection

Self-aggregation in radiative-convective equilibrium (RCE) has been a classic scenario for the study of convection, which involves multiscale dynamics and multiple physical processes. Here we investigate the influence of microphysics and turbulence schemes on the convective circulation in RCE, with both large-eddy simulations (LES) and cloud-resolving model (CRM) simulations. We found that when the Thompson microphysics scheme is used in CRM simulations, the atmosphere in the RCE separates into moist and dry patches clearly, with deep convection dominating in the moist region. In contrast, when the microphysics is represented with the Morrison scheme, most CRM simulations do not exhibit aggregated convection; instead, they exhibit random convection randomly distributed in the entire domain. Analysis of the budget of the variance of moist static energy suggests the difference is mainly caused by longwave radiation feedback directly, which in turn is related to the difference in precipitation efficiency between the Morrison and Thompson microphysics schemes. Turbulence schemes do not alter RCE states as dramatically as microphysics, but they can modulate the timescale of the emergence of self-aggregation if that does occur. Different CRM simulations exhibit equilibrating timescales between 25 to 70 days, depending on how subgrid-scale turbulence contributes to the flow. LES runs at 100-m resolution are conducted to provide benchmark answers to the evolution of the atmosphere in RCE. Though they were conducted in smaller simulation domains, their fine resolution allows them to have reduced dependence on physical parameterizations.