Uji Sensitivitas Skema Parameterisasi Cumulus untuk Prediksi Hujan di Wilayah Jawa Timur Apritarum Fadianika, Hariadi

Abstract

Not every process in the atmosphere can be solved explicitly by the equations of motion in numerical models. Thus, every NWP models including WRF uses a parameterization for weather parameters that cannot be calculated directly by mathematical formula. One of parameterization in WRF-ARW is a cumulus parameterization. The test of Kain-Fritsch (KF), Betts-Miller-Janjic (BMJ), Grell-Devenyi (GD) schemes are conducted to increase the accuracy of rain prediction by selecting the cumulus parameterization scheme that has the highest sensitivity. The scheme was compared spatially to determine the suitability in a scope of the district or city, and verified at seven points of BMKGs station in East Java to obtain the forecast accuration. Based on the comparison and verification between model results and observational data, generally, GD scheme is the best cumulus parameterization scheme in almost whole area of East Java, except in Bawean Island (BMJ) and Tuban, Panarukan, and Banyuwangi (KF). WRF-ARW model performance are quite good at predicting rain or not rain occurence (the average of Proportion Correct (PC) is above 0.7), but it cannot precisely predict the time of occurence and intensity of the rain. The model tend to underestimate but it approach the real prediction (Equitable
Threat score value (ETS) under 0.2 and the average of Frequence Bias Index (FBI) 0.6).