乏资料地区微地形暴雨响应机理及自适应暴雨洪水预报方法研究

暴雨洪涝是我国最严重的自然灾害之一，乏资料地区由于水文气象观测资料不完整，现有的暴雨观测手段无法分析乏资料地区微地形区域暴雨响应机理，且流域关键属性特征相似分析方法无法刻画流域微观产汇流规律，导致乏资料地区暴雨洪水预报结果难以达到预想的精度要求。本项目针对上述难题，创新性提出暴雨微地形分析新思路，通过历史观测资料分析-现场探测试验-物理机制分析，揭示乏资料地区微地形对暴雨风速、风向、温度、湿度等关键要素的影响机制，阐明微地形对局地暴雨形成的影响机理，建立微地形暴雨物理计算模型；巧妙地将深度学习框架中的图像识别技术引入乏资料地区地形、植被、气象气候特征等全空间要素微观产汇流规律分析中，形成乏资料地区典型产流微观计算标准模型库，建立乏资料地区流域自适应产汇流计算建模框架，发展乏资料地区流域水文预报理论与方法体系，提升我国乏资料地区流域水文预报水平，为抵御极端暴雨洪水灾害提供有力科学支撑。

The rainstorm and flood disaster is one of the most serious natural disasters in China. Due to the incomplete hydrologic and meteorological observation data, the existing rainstorm observation methods cannot analyze the rainstorm response mechanism of the micro terrain area in the lack of data area, and the similarity analysis method of the underlying surface characteristics of the basin cannot describe the fine runoff and confluence law of river basin. The results of rainstorm and flood forecast in the area of lack of data are difficult to meet the expected accuracy requirements. In view of the above problems, this project puts forward a new idea of rainstorm micro terrain analysis. Through the analysis of historical observation data, field detection test and physical mechanism analysis, the influence mechanism of micro topography on the key elements of rainstorm, such as wind speed, wind direction, temperature and humidity, will be revealed. This work may clarify the influence mechanism of micro terrain on the formation of local rainstorm, and then establish the physical calculation model of micro terrain rainstorm. Meanwhile, the image recognition technology in the deep learning framework is introduced into the analysis of fine runoff and confluence law of all spatial elements (including terrain, vegetation, meteorological and climatic characteristics, etc) in the area of lack of data. Thereafter, the standard model library for micro calculation of typical runoff generation in the area with lack of data is formed. Also, the framework of self-adaptive runoff generation and confluence calculation is established in the area of lack of data. It is of great scientific significance and engineering practical value to develop the hydrological forecasting theory and method system in the area of lack of data, improve the level of rainstorm flood forecasting in the basin of lack of data in China, and provide strong scientific support for resisting extreme rainstorm flood disaster.