Skill of Medium-Range Forecasts over the Indian Monsoon Region Using Different Parameterizations of Deep Convection

Someshwar Das, A. K. Mitra, G. R. Iyengar, J. Singh
2002 Weather and forecasting  
The operational global spectral model of the National Center for Medium Range Weather Forecasting (NCMRWF) at T80 resolution and 18 vertical levels has been used to study the skill of medium-range forecasts using three different parameterizations of deep convection namely, a Kuo-Anthes-type cumulus parameterization scheme referred to as "KUO" scheme, the relaxed Arakawa-Schubert (RAS) scheme, and the simplified Arakawa-Schubert (SAS) scheme, during an active phase of the Indian summer monsoon.
more » ... everal medium-range forecasts (up to 5 days) have been made using the initial conditions of July and August of 1999, when the monsoon was active over the Indian region. Skill scores of predicted rainfall, rmse of wind and temperature, systematic errors, and genesis and tracks of the monsoon depressions predicted by the three schemes have been studied. Results indicate that, in general, the areas of light (heavy) rainfall are overestimated (underestimated) by KUO, which also fails to predict the rain-shadow effect observed over southern peninsular India. RAS and SAS produce fairly good forecasts of the observed rainfall; however, the best forecast is produced by SAS in most of the rainfall categories over the Indian region. The rmse of wind and temperature do not show significant differences among the three schemes over the global domain; however, they indicate considerable differences over the Indian region. The rmse of wind is slightly higher in RAS and SAS because of overestimation of the strength of the low-level westerly jet and upper-level tropical easterly jet. Errors in temperature forecasts are considerably reduced by RAS and SAS on all days. Systematic errors of the forecasts indicate that KUO tries to weaken the observed southwesterly flow and the low-level jet during the monsoon. RAS and SAS try to intensify the easterlies over the north Indian plains and to strengthen the monsoon trough. They shift the core of the tropical easterly jet stream to the south of its normal position. SAS reduces the cold bias almost everywhere over the Indian region. The improved simulation of temperature by SAS results in the reduction of rmse. The reduction of cold bias and improved simulated temperature by SAS indicate a proper redistribution of heat by deep convective clouds over the region by this scheme. Study of the lows and monsoon depressions indicated that the best forecast of the location of the genesis was produced by RAS. All three schemes were able to predict the tracks of the depression fairly well in the 24 h, but SAS produced relatively fewer errors when compared with the other two schemes. In most of the cases, SAS was also able to maintain the system up to 72 h, whereas the other two schemes weakened the systems.
doi:10.1175/1520-0434(2002)017<1194:somrfo>;2 fatcat:a6ul3jr53bh2pi2yjhcvxmu47u