Research Of Large Scale Distributed Hydrological Model In The Upper Reaches Of The Yangtze River

Hydrological forecasting,as a core research area of hydrology,is the basis and guarantee of reasonable and efficient use of water resource,well planned and constructed of water project as well as safe and stable operation of hydropower-system.However,the constructions of large-scale reservoir group in recent years changed the hydrological situation significantly,which pose greater challenge and higher requirement on hydrological forecasting.In this paper,Xiluodu-Yichang interval in the upper Yangtze River basin is selected as the study area.A land-atmosphere coupling method is established,using the distributed hydrological model(Variable Infiltration Capacity model)driven by the results of the Weather Research and Forecasting model,to achieve the goals of this paper: extending the leading time for the flood forecast and improving the forecasting accuracy in the Xiluodu-Yichang interval.The main research results are as follows:(1)Firstly,a daily grid-based distributed hydrological model for Xiluodu-Yichang interval is developed and applied to the land surface process simulation of Land-atmosphere coupling model.At the same time,Shuffled Complex Evolution-University of Arizona algorithm and Particle swarm optimization algorithm are applied with available meteorological and hydrological data to calibrate and optimize a best model parameter.The result of simulation demonstrates that the parameter shows good representative.(2)Secondly,WRF model is introduced into the study area to optimize a parameterized combination.Multi-index evaluation method based on TS Score,Mean Absolute Error and Blair Score bring out a combination scheme with the best parameter scheme of microphysical process(WSM 3-class scheme),land surface layer(Noah scheme)and cloud convection parameterization(New KF scheme).Using the above result to make a verification and analyze the space-time distribution characteristic of rainfall,the parameterized combination shows preferable applicability.(3)Finally,a land-atmosphere coupling model is established using the VIC model driven by the results of the WRF model.Regarding the temporal and spatial matching problem between meteorological model and land surface mode of land-atmosphere coupling,the coupling model proposes an efficient solution.After using adequate rains and floods observation to verify,it indicates that the coupling model effectively extends the leading time for the flood forecast and obtains a higher forecasting accuracy than ignoring rains and floods,consequently the model contributes to the joint disposal of reservoir group in this interval by proving accurate prediction of incoming water with a certain leading time.