洪水灾害是最常见和危害最大的一种自然灾害，洪水预报是防洪最有效的非工程措施之一。气象水文耦合方法能够延长洪水预报的预见期，为防洪减灾提供一定的响应时间。然而，气象降水预报的不确定性、水文模型的适用性、气象水文尺度匹配以及洪水预报的不确定性，成为制约提高洪水预报精度和延长预见期的主要瓶颈。本项目以GRAPES_MESO数值天气预报模式为基础，对影响洪水预报的关键方面进行研究：（1）研究GRAPES_MESO模式不确定性，研究高分辨率GRAPES_MESO预报模式；（2）以GRAPES_MESO集合预报模式与分布式水文模型耦合，研究洪水集合预报模式，并在灾害性洪水易发区的洪水预报中评估其预报性和不确定性。本项目旨在提高洪水预报精度与延长预报预见期，为防汛决策提供科学依据。

Flood disaster is the most frequent and heavily damaging natural disaster, and flood forecast is one of the most effective non-engineering measures for flood control. The coupled atmospheric-hydrological modeling for flood forecast can increase lead time, which provide the response time for flood control and disaster mitigation. However, meteorological precipitation forecast uncertainty, hydrological model applicability, scale match between meteorological model and hydrological model and flood forecast uncertainty, become a bottleneck to improve the precision and lead time of flood forecasting. Based on the GRAPES_MESO numerical weather prediction model, two key technology of flood forecast will be studied. (1) With analyzing GRAPES_MESO model's uncertainty, GRAPES_MESO ensemble prediction model with high resolution will be studied. (2) Flood forecasting with coupling GRAPES_MESO ensemble prediction model and distributed hydrological model will be evaluated and its uncertainty will be analyzed in the flood prone areas. This project aims to improve the precision and lead time of flood forecasting, and provide scientific basis for flood control decision-making.