A comparison study of convective parameterization schemes in a mesoscale model

A comparison study of four cumulus parameterization schemes (CPSs), the Anthes-Kuo, Betts-Miller, Grell, and Kain-Fritsch schemes, is conducted using The Pennsylvania State University-National Center for Atmospheric Research mesoscale model. Performance of these CPSs is examined using, six precipitation events over the continental United States for both cold and warm seasons. Grid resolutions of 36 and 12 km are chosen to represent current mesoscale research models and future operational models. The key parameters used to evaluate skill include precipitation, sea level pressure, wind, and temperature predictions. Precipitation is evaluated statistically using conventional skill scores (such as threat and bias scores) for different threshold values based on hourly rainfall observations. Rainfall and other mesoscale features are also evaluated by careful examination of analyzed and simulated fields, which are discussed in the context of timing, evolution, intensity, and structure of the precipitation systems. It is found that the general 6-h precipitation forecast skill for these schemes is fairly good in predicting four out of six cases examined in this study, even for higher thresholds. The forecast skill is generally higher for cold-season events than for warm-season events. There is an increase in the forecast skill in the 12-km model, and the gain is most obvious in predicting heavier rainfall amounts. The model's precipitation forecast skill is better in rainfall volume than in either the areal coverage or the peak amount. The scheme with the convective available potential energy-based closure assumption (Kain-Fritsch scheme) appears to perform better. Some systematic behaviors associated with various schemes are also noted wherever possible. The partition of rainfall into subgrid scale and grid scale is sensitive to the particular parameterization scheme chosen, but relatively insensitive to either the model grid sizes or the convective environments. The prediction of mesoscale surface features in warm-season cases, such as mesoscale pressure centers, wind-shift lines (gust fronts), and temperature fields, strongly suggests that the CPSs with moist downdrafts are able to predict these surface features more accurately.