Abstract: The paper presents a high-resolution numerical simulation of an extreme rainstorm event occurring under the complex topographic conditions in Southern Xinjiang, utilizing the Weather Research and Forecasting model. It also carries out water vapor sensitivity experiments to investigate the impact of water vapor on extreme rainstorms in this arid zone. These experiments unveil the initial understanding of how water vapor dynamically and thermally affects rainstorms in South Xinjiang. Further investigation into the characterization ability of the generalized wet vortex and convective vorticity vector reveals their impact on heavy rainfall in arid zones. The study concludes that changes in water vapor significantly influence the intensity of extreme precipitation. Specifically, an increase in water vapor results in a significantly higher change in precipitation intensity than the change in water vapor itself. This effect is most pronounced when low-level water vapor is enhanced. When mid-level water vapor increases, the extremity of precipitation also notably intensifies. Furthermore, the study finds that the change in precipitation intensity is linear within the first hour of changing water vapor. This is attributed to the fact that varying water vapor significantly affects the convective available potential energy (CAPE) in the precipitation atmosphere. The stronger the CAPE, the more powerful the convective trigger becomes, leading to stronger vertical motion that can further pump lower-level water vapor. This positive feedback mechanism results in extreme precipitation. Sensitivity experiments reveal that increasing water vapor significantly enhances upward motion, intensifies the low-level easterly rush, and strengthens water vapor convergence in the precipitation zone. This leads to an increase in water vapor transportation to the mesosphere and an increase in the release of latent heat of condensation, further amplifying upward motion and increasing precipitation intensity. Conversely, decreasing water vapor results in a significant reduction in upward motion, the strength of the low-level easterly rush, and the water vapor convergence in the precipitation zone. As a result, the water vapor transported to the middle layer decreases, leading to a reduction in the release of latent heat of condensation, resulting in a decrease in upward motion and precipitation intensity. The simulated precipitation strongly correlates with the spatial distribution and temporal evolution of both the generalized wet vortex and convective vorticity vectors. This correlation is particularly evident following changes in water vapor, as the features of both vectors display similar changes. These results suggest that the rich dynamic and thermal information contained within these physical factors can effectively capture the key features of extreme precipitation in arid regions. Moreover, by integrating these physical factors with precipitation data, they could potentially prove useful for early warning and prediction of rainfall, particularly in arid areas. Therefore, coupling these physical factors with precipitation data could greatly improve the accuracy of precipitation forecasting in arid zones.