Impacts of Cumulus Parameterizations on Extreme Precipitation Simulation in Semi-Arid Region: A Case Study in Northwest China

In the context of climate change, extreme precipitation in semi-arid region happens frequently. How well models simulate extreme precipitation in semi-arid region remains unclear. Based on a WRF v4.3 simulation of a rainstorm event that occurred in Qingyang, China on 21 July 2019, applying Kain-Fritsch (KF), Grell-Devenyi (GD) and Bullock-Wang (BW) schemes, the impacts of different cumulus parameterizations on extreme precipitation simulations in semi-arid region were analyzed, and the possible causes of precipitation biases were explored. The results showed that the WRF with the three schemes essentially reproduced the location and structure of precipitation, but the intensity of precipitation in the central region was underestimated. Based on the structure-amplitud-elocation (SAL) method, the KF scheme exhibited better performance in precipitation simulation than the other two schemes, while there were significant intensity and location deviations of rain band occurrence between simulations using the GD, BW schemes and observations. Convection simulation using the GD and BW schemes was less effective than that using the KF scheme, compared to the observations. As a result, the GD and BW schemes simulated a larger geopotential height at 500 hPa over Qingyang and weaker upper-level low troughs than simulations using the KF scheme. This led to simulation of less water vapor transport into the front of the trough, resulting in a deficit in simulated precipitation. The study results highlight the impacts of convection on large-scale atmospheric circulation linked to extreme precipitation in semi-arid region.