COMPARISON OF DATA ASSIMILATION USING SURFACE OBSERVATION, UPPER AIR, AND SATELLITE RADIATION DATA ON RAINFALL PREDICTION IN THE JAMBI REGION (CASE STUDY OF HEAVY RAIN OCTOBER 20TH, 2020)

Abstract

Weather Research and Forecasting (WRF) is a mesoscale numerical weather prediction model that can provide good rainfall prediction information. The accuracy of the initial conditions and the accuracy of the parameterization scheme used in the WRF model affect the quality of the resulting rainfall prediction. Therefore it is necessary to assimilate to optimize the accuracy of the initial conditions in the model using the Three Dimensional Variational (3DVAR) assimilation technique. The purpose of this study was to determine the effect of applying the 3DVAR assimilation technique with the surface, upper air, and satellite radiation observations in predicting the occurrence of heavy rain on October 20^th, 2020, in the Jambi region by first conducting a parameterization test of the cumulus and microphysical schemes. In this study, four experimental schemes were used, namely no assimilation (NON), observation data assimilation (OBS), satellite radiation data assimilation (SAT), and satellite radiation and observation data assimilation (BOTH). Each experimental model result was then verified statistically and spatially to determine the effect of the applied data assimilation. The results of this study indicate that the combination of Grell-3D and Thompson scheme shows the best performance in predicting rainfall. Then based on the spatial analysis of the SAT experiment, it is known that it can improve the model's initial conditions on the temperature and pressure parameters. Meanwhile, based on statistical verification, the SAT experiment improved the accuracy of rainfall predictions with a better forecast skill score than other experiments tested.