WRF-3DVAR Radiance Data Assimilation Impact for Convective Rain Prediction on Jakarta

Abstract

Weather Research and Forecasting (WRF) is an open source numerical weather prediction model, but one of the main problems is the inaccuracy of initial condition that impacts the accuracy of weather predictions. Techniques that can be used to improve initial condition is to assimilate satellite radiance data using Three-Dimensional Variation (3DVAR) method. The Global Forecast System (GFS) data were used as initial conditions which were assimilated with several radiance data such as from AMSU-A sensors, MHS sensors, and ATMS sensors. The purpose of this study was to compare the effect of data assimilation on the initial condition model and the rainfall prediction for convective precipitation on February 24, 2016 in Jakarta. The results showed there is a clear flow of moisture to the northern part of Jabodetabek area that cause the model captured a high relative humidity which results in the heavy rain. The result of the study obtained different rainfall prediction between non-assimilation model and assimilation model. In general, both models have underestimated rainfall, but satellite radiance data assimilation is better able to describe the convective rain patterns and has better accuracy on the convective rainfall accumulation per hour.