A comparison of cumulus parameterizations in idealized sea-breeze simulations

Four cumulus parameterizations in the fifth-generation Pennsylvania State University-National Center for Atmospheric Research (Penn State-NCAR) Mesoscale Model (MM5) are compared in idealized sea-breeze simulations, with the aim of discovering why they work as they do. Compared to simulations of real cases, idealized cases produce simpler results, which can be more easily examined and explained. By determining which features of each parameterization cause them to produce differing results, a basis for improving their formulations and assisting modelers who may design new cumulus parameterizations can be provided. The most realistic results obtained for these simulations are those using the Kain-Fritsch scheme. Rainfall is significantly delayed with the Betts-Miller scheme, due to the method of computing the reference sounding. Another version of this parameterization, which computes the reference sounding differently, produces nearly the same timing and location of deep convection as the Kain-Fritsch scheme, despite the very different physics. In applying the quasi-equilibrium closure, the Grell parameterization uses horizontal and vertical advection to compute the rate of destabilization. In the present simulation, the parameterized updraft is always derived from the top of the mixed layer, where vertical advection predominates over horizontal advection in increasing the moist static energy, instead of from the most unstable layer. By doing this, it evades the question of whether horizontal advection generates instability or merely advects an existing unstable column.