Evaluation of WRF Performance in Simulating an Extreme Precipitation Event over the South of Minas Gerais, Brazil

Abstract

Extreme precipitation events are becoming increasingly frequent and intense in southeastern Brazil, leading to socio-economic problems. While it is not possible to control these events, providing accurate weather forecasts can help society be better prepared. In this study, we assess the performance of the Weather Research and Forecasting (WRF) model in simulating a period of extreme precipitation from 31 December 2021 to 2 January 2022 in the southern region of Minas Gerais (SMG) state in southeastern Brazil. We conducted five simulations using two nested grids: a 12 km grid (coarse resolution) and a 3 km grid (high resolution). For the coarse resolution, we tested the performance of five cumulus convection parameterization schemes: Kain–Fritsch, Betts–Miller–Janjic, Grell–Freitas, Grell–Devenyi, and New Tiedke. We evaluated the impact of these simulations on driving the high-resolution simulations. To assess the performance of the simulations, we compared them with satellite estimates, in situ precipitation measurements from thirteen meteorological stations, and other variables from ERA5 reanalysis. Based on the results, we found that the Grell–Freitas scheme has better performance in simulating the spatial pattern and intensity of precipitation for the studied region when compared with the other four analyzed schemes.