Evaluation of a buoyancy sorting scheme for cumulus convection with Large Eddy Simulation results

Observations and recent experiments with cloud resolving models suggest that the depth and intensity of cumulus convection does not only depend on vertical stability but also on relative humidity in the environment. However, most cumulus convection parameterizations are completely insensitive to environmental humidity conditions. The buoyancy sorting scheme proposed by Kain and Fritsch is perhaps the mechanism in principle that determines fractional entrainment and detrainment as a function of environmental conditions of both temperature and humidity and is being widely used now in many operational numerical weather prediction and climate models. In this paper we will use Large Eddy Simulation (LES) model results to critically evaluate the underlying principles of this scheme. It will be shown that the mixing principle behaves correctly only in a very qualitative sense for a suite of different shallow cumulus convection cases. The main flaw in the Kain-Fritsch mixing mechanism turns out to be the assumption that the amounts of cloud and environmental air mass that mix with each other are constant. LES results show that these amounts of air masses actually scale with cloud depth. The impact of this generalisation of the Kain-Fritsch scheme in a single column model will be demonstrated for a number of cumulus convection cases.