短时强降水是我国最主要的强对流灾害天气之一，易造成城市内涝和山洪、泥石流以及滑 坡等次生地质灾害。回顾了近年来我国短时强降水的主要研究进展，并简要对比了美国和欧洲的 相关研究成果，涵盖了短时强降水的时空分布特征、大气环流形势和环境条件、雷达回波特征和雨 滴谱特征以及地形和城市化对短时强降水的影响及其机制；总结了人工智能在我国短时强降水潜 势预报和短时临近预报中的应用。随着全球变暖，短时强降水频次和强度都呈增加趋势，今后需 要进一步研究其形成机制和环境条件，提升观测时空分辨率，加强新型观测资料应用，通过融合分 析多源且稠密的观测资料，提升高分辨率的快速更新循环同化数值模式预报能力，改进和发展深 度学习预报模型和算法，尤其是研发深度学习大模型来提升短时强降水的预报预警能力。

Abstract：Short-duration heavy precipitation is one type of the most important severe convective disaster weather in China, which is prone to cause urban waterlogging and secondary geological disasters such as mountain torrents, mudslides, and landslides. This paper reviews the main progress in short-duration heavy precipitation in China in recent years, and compared the relevant research findings of the United States and Europe briefly, covering the spatiotemporal distribution characteristics and diurnal variation characteristics of short-duration heavy precipitation; the atmospheric circulation situation and environmental conditions for the occurrence and development of short-duration heavy precipitation in major regions of China; the radar echo characteristics and raindrop characteristics; the impact of topography and urbanization on short-duration heavy precipitation and its mechanism; and then the application of artificial intelligence in the potential forecasting, short-term forecasting, and nowcasting of short-duration heavy precipitation in China. With global warming, the frequency and intensity of short-duration heavy precipitation are increasing. In the future, it is necessary to further study its formation mechanism and environmental conditions, improve the spatiotemporal resolution of observations, strengthen the application of new observation data, enhance the forecasting capability of highresolution rapid update cycle assimilation numerical weather prediction models through the fusion analysis of multi-source and dense observation data, optimize the deep learning model and algorithm, especially in the development of large deep learning models to enhance the forecasting and early warning capabilities for shortduration heavy precipitation.