The Importance of Cumulus Parameterization and Resolution in Simulating Rainfall over Peninsular Malaysia

Abstract

In this study, five simulations were conducted using the weather research and forecasting (WRF) model with different cumulus parameterizations schemes (CPSs) for the period from 2013 until 2018. A one-year simulation of 2013 with three different horizontal resolutions of 25, 5, and 1.6 km was also performed. The CPSs used were Kain–Fritsch (KF), Grell–Devenyi (GR), Betts–Miller–Janjic (BM), and a non-parameterized scheme (NC). In assessments of model resolutions, both the 25 and 5 km resolutions depicted a strong negative bias in the northeastern part of Peninsular Malaysia during December–January–February (DJF), with marginal differences between the two simulations. Among all 5 km experiments, the best performing scheme was the BM scheme for almost all seasons. Furthermore, the 5 km simulation did not exhibit significant differences relative to the 25 km of the diurnal cycle. The 1.6 km simulation showed significant added value as it was the only simulation that was able to simulate the high precipitation intensity in the morning and a precipitation peak during the evening. The 1.6 km resolution was also the only resolution capable of picking up the precipitation signals in the R4 region (South Peninsular Malaysia) compared to the other two resolutions. While both CPSs and resolutions are important for accurate predictions, the role of CPSs became less significant in a higher resolution simulation.