PERFORMANCE OF TWO GRAY-ZONE WRF-MODEL SIMULATIONS FOR EXTREME PRECIPITATION IN THE SICHUAN BASIN

Abstract

The Sichuan Basin (SB) in the Southwest region of China is a low-lying area surrounded by mountains, with complex regional climate and topography. The region frequently experiences heavy precipitation. One such event with anomalously high precipitation occurred in the summer of 2020, which resulted in flooding, waterlogging and mudslides, and lead to both economic losses and human suffering. Several large-scale processes affect precipitation in the area, including the subtropical westerly jet stream and the Western Pacific Subtropical High (WPSH). To predict and study such weather events, theWeather Research and Forecasting (WRF) model is widely used. High resolution models are especially important in areas with complex topography. Models in the gray-zone resolution (4-15 km) could provide a good balance between model performance and computing costs. The aim of this study is therefore to evaluate if two gray-zone resolution model simulations (WRF9km and WRF9km35E) can reproduce an extreme precipitation event in the SB, using GPM IMERG and ERA5 as reference data, where Spatial Correlation Coefficient (SCC), BIAS and Fraction Skill Score (FSS) are evaluated. The study additionally aims to see if one of the simulations can reproduce the event better than the other, and if differences in performance can be connected to certain atmospheric circulation features, with the expectation that WRF9km35E will perform better because it was created to increase the accuracy of simulating the jet stream. The result shows that none of the simulations were able to realistically reproduce the timing, location and amount of the event. Both simulations failed to reproduce the most extreme days of the event, but they both performed better on the days with less extreme amounts. Generally WRF9km35E performed better than WRF9km in all tests. Differences in large-scale circulation were found, where WRF9km35E simulate the location and strength of the jet stream as well as the WPSH better than WRF9km. This could affect the performance, although the differences in circulation alone might not account for the differences in performance of the simulations, as there are several differences between them. It is important to understand and accurately simulate atmospheric processes in order to provide reliable forecasts and identify key processes that are crucial in the development of extreme precipitation events. To be able to simulate future climate, the models need to be able to realistically represent the current conditions, which is why further research regarding regional climate in areas with complex terrain, especially precipitation processes, is needed.