Performance of Different Parameterization Configurations of WRF-ARW Model during Heavy Rainfall over Mi Oya River Basin

Abstract

Short-term intense precipitation is one of the hallmarks of climate change. Mi Oya River basin experiences severe seasonal floods annually, but the damage can be lessened by developing a numerical weather forecasting (NWF) model for the entire basin and incorporating it with effective reservoir management. Several areas in Sri Lanka have undergone NWF studies, however they are insufficient to determine the best physics schemes for the basin. This study investigates the Weather Research and Forecasting (WRF-ARW) model's predictability with varying three microphysics and two cumulus schemes to discover the optimal set of physics parameters for predicting heavy rainfall occurrences throughout the Southwest and Northeast monsoon seasons within a nested domain configuration. The WRF model's forecasting results at 3 km grid resolution were compared with four rainfall gauging stations in the basin for three rainfall events in May 2016, April 2018, and November 2015. Total Model Performance was derived for the evaluation utilizing bias, MAE, RMSE, Correlation Coefficient, and slope of each model's output data with observed rainfall data. After comparing the model output to data, WSM6 microphysics and Betts-Miller-Janjic cumulus with other default physics settings were determined to be the optimal physics combination to forecast weather across the region.