Evaluation of Parametric Assumptions for Shallow Cumulus Convection

Abstract

A large-eddy simulation (LES) model has been utilized to study nonprecipitating Shallow Convective clouds such as observed during the undisturbed BOMEX period in the trade wind areas. By choosing a realistic large-scale forcing the authors have been able to simulate shallow convective clouds under quasi-steady-state conditions over a long period of 7 hours. This is a necessary condition to investigate diagnostic cumulus parameterization schemes since such schemes usually assume steady-state conditions. The response of the model to the applied large-scale forcing compares well with budget study results of BOMEX. In addition, the LES model delivers detailed information concerning the dynamics of shallow convective clouds. This is used to verify basic parameterizations of turbulent fluxes and entrainment and detrainment rates used in large-scale models. The most important conclusions are (i) the fractional entrainment and detrainment rates used in present large-scale atmospheric models are one order of magnitude too small, confirming previous results obtained by Esbensen, and (ii) estimates of turbulent fluxes by bulk cloud updrafts and environmental downdrafts give an underestimation of 20% to 50% depending on the variable that is transported. Implications of these results for cumulus parameterizations will be discussed.