PROBABILISTIC FORECAST SYSTEM OF RAIN EVENT IN METRO BANDUNG USING MODEL OUTPUT STATISTIC SCHEME

• Main Author: Noer Hadi, Anugrah
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/70316
• Institution: Institut Teknologi Bandung

Statistical Guidance is important procedure on modern weather prediction operational. Experimental prediction from WRF outputs that run by Weather and Climate Prediction Laboratory (WCPL) in Bandung Institute of Technology were still direct output model. Therefore on this research, Statistical Guidance is applied by using the model outputs. Model Output Statistic (MOS) scheme is used on Statistical Guidance approach to forecasting probability of rain event on January – March period in Metro Bandung. Development stage is executed in Bandung Institute of Technology and implemented in Metro Bandung. Forecast times are divided into three times; forecast day 1, forecast day 2, and forecast day 3 to forecasting probability of rain event for three days. To increasing the accuration, optimal point approximation is applied. The result is verified with observational data by using Brier Score and compared with time-lagged ensemble scheme by using Brier Skill Score respectively in Bandung Institute of Technology, Antapani, Cimahi, Buahbatu, Majalaya, and Antapani. Optimal point that is excercised lies in Southwestern Bandung (around Ciwidey and Rancabali). The combination between optimal point and local point can increase the accuration approximately 0.2 up to 0.3. The comparison result between MOS scheme and time-lagged scheme shows that MOS more accurate 20% up to 70% on rain event forecasting.