An evaluation of a mass-flux cumulus parameterization scheme in the KMA global forecast system

A mass-flux type cumulus parameterization scheme is evaluated in the Korea Meteorological Administration (KMA)'s Global Data Assimilation and Prediction System (GDAPS). Two sets of parallel assimilation runs, with a horizontal resolution corresponding to the spectral truncation of T213, approximately 50 km, are designed, one with the operational Kuo convection scheme, and another with a simplified Arakawa-Schubert (SAS) convection scheme, for the month of July 2001, which was a typical summer monsoon period in Korea. An additional sensitivity experiment, that excludes the moist down-drafts in the SAS scheme, is conducted. A 10-day forecast cycle is run, within a data assimilation window, beginning at every 1200 UTC in July 2001. A heavy rainfall event, that occurred in the Korean peninsula on 14-15 July 2001, is investigated in detail. For the heavy rainfall case, the SAS scheme tends to reduce the precipitation amount, as compared to the Kuo scheme. It is found that the exclusion of moist downdrafts in the SAS scheme, increases drying and heating rates, resulting in enhanced rainfall, as shown in the Kuo scheme. The monthly verification demonstrates that the systematic warm bias in the lower troposphere in the operational model is nearly absent when the SAS scheme is employed. In terms of statistical measure of the skill for the predicted sea-level pressure, and that for the upper level patterns, the experiment with the SAS (Kuo) scheme outperforms in the northern (southern) hemisphere. In the case of precipitation forecasts, the SAS scheme effectively corrects the spurious excessive precipitation over the tropical land area, due to the inclusion of moist downdrafts. However, the reduction in precipitation in the SAS scheme, deteriorates the forecast skill for high precipitation categories.