Evaluation study of the Kain-Fritsch convection scheme

Cumulus clouds occur on spatial scales that are too small to be explicitly resolved by state of the art operational weather prediction and climate models. In order to take the effect on the model evolution of these sub-grid convection processes into account a parameterization in the form of a convection scheme is required. A crucial process in a convection scheme is the turbulent exchange of heat, momentum and moisture between clouds and their environment.

A relatively new convection scheme, based on buoyancy sorting, introduced by Kain and Fritsch is used to prescribe this exchange. The main idea of this convection scheme is that at the periphery of the cloud (or updraft) various mixed air parcels are formed that consist of both cloudy and clear air. These mixed parcels can have rather distinct densities. The parcels that are positively buoyant are entrained into the updraft, while the parcels that are negatively buoyant are detrained. The purpose of this study is to evaluate how well the turbulent mixing between clouds and their environment is represented by this buoyancy sorting mechanism. This is done by means of the Simplified Updraft Model, the SUM. This is a single parcel ascent model.

The performance of the SUM is tested by qualitative comparison to Large Eddy Simulation (LES) results. This is done for two distinct cases. The first test is done by means of a typical shallow cumulus case: BOMEX. The aim of this test is to investigate what happens to the updraft profiles if the frequency distribution of the mixed parcels and the amount of the mixed parcels is changed. In the second test case the sensitivity of the Kain-Fritsch scheme to relative humidity is examined. According to Cloud Resolving Model (CRM) - results this sensitivity should be huge, but it is not clearly present in most convection schemes (Derbyshire, 2003).