Impact Of Potential Evapotranspiration To The Simulation Of CREST Model In Qinhuai River Basin

The water resources imbalance between supply and demand becomes to be a big problem,so how to accurately grasp the hydrological characteristics of water resources and hydrological cycle process is particularly important.Hydrological model is a possible solution to this problem.The hydrological model provides practical support for mastering the hydrological characteristics and water dynamic simulation.Qinhuai River Basin locates in the Yangtze River Delta,and which is the middle flat basin surrounded by mountains.The fast and fierce upland water,the Yangtze River flood downstream and abundant precipitation easily cause this basin to be drowned lengthy and widely by the flood,which brings great pressure to Qinhuai River Basin flood control work.So mastering the Qinhuai River Basin Hydrological Characteristics and searching for an accurate watershed hydrological model have become a top priority.This work will be beneficial to watershed's water resources management and planning in the future,and also can provide a scientific basis and theoretical support for basin storage flood control and irrigation.Climate variables largely affect the accuracy of the model,hydrology researchs focus on impacts of precipitation or underlying surface data source on the model in recent years,however potential evapotranspiration,as one of the main input variables has been less researched.Taking the Qinhuai River basin as the research area,using GIS technology and RS technology to deal with DEM,FAC,FDR and the drainage map which are required by the CREST model;using IDW to interpolate the precipitation,and evaluating the interpolation accuracy;the potential evapotranspiration of the Qinhuai River basin was calculated by Blaney-Criddle method and Hargreaves method,then analyzing potential evapotranspiration ofQinhuai River Basin temporally and spatially.The three potential evapotranspiration sources calculated by Blaney-Criddle method and Hargreaves method and downloaded from FEWS were used as input data of CREST model to simulate annual runoff and flood season runoff,in order to find out which one is most suitable for CREST model,and finally evaluating the applicability of CREST model in the Qinhuai River Basin.The conclusions of the thesis are as followed:(1)The absolute error,relative error between the measured precipitation from rainfall station and precipitation interpolated by IDW in Qinhuai River Basin were small and the correlation coefficient were larger,also pass the F significant test,so the interpolation result is reliable,which can meet the accurate data requirement of the following CREST model.Qinhuai River Basin is small,and the difference among rainfall stations' precipitation is small with little spatial distribution difference which can be described as "North low and south high,around high and middle low".(2)Potential evaporation calculated by Blaney-Criddle method and Hargreaves method and download from Famine Early Warning Systems Network website(FEWS)were all suitable for CREST model in Qinhuai River Basin,especially the accuracy was better by Hargreaves method and FEWS.Peak season of potential evaporation in Qinhuai River Basin was summer,the valley season was the winter;space distribution of annual potential evapotranspiration can be described as "higher in the West and lower in the East",and the highest value appeared in the South-West,the lowest value appeared in the South-East,(3)The CREST model can be applicable in the Qinhuai River basin,and it also can accurately simulate the hydrological characteristics of annual runoff and the flood season's runoff.Potential evapotranspiration downloaded from Famine Early Warning Systems Network website(FEWS)could simulate annual runoff better than potential evapotranspiration calculated by Hargreaves method,and the later could simulate flood season's runoff better than the former,however the results based on Blaney-Criddle method were both poorer.This meant potential evapotranspiration calculated by Hargreaves method and downloaded from FEWS were fitful in different time scale of CREST model.In general,the CREST model based on these two sources could have a good result.So considering about the costs of time and computation,Potential evapotranspiration downloaded from Famine Early Warning Systems Network website(FEWS)could be the best choose for CREST model,and from the point of view of flood management,potential evapotranspiration calculated by Hargreaves method could be better.(4)The spatial resolution of the potential evapotranspiration in the small watershed region,like the Qinhuai River basin,had little effect on the simulation results of CREST model.Taking Potential evapotranspiration calculated by Hargreaves method and downloaded from Famine Early Warning Systems Network website(FEWS)as input value of CREST model to simulate runoff and flood season runoff.The data quality of these two sources potential evapotranspiration were close,and spatial resolution of the former is higher than the later,but the results were close and both of them were good.Potential evapotranspiration calculated by Blaney-Criddle method as CREST model's input value simulated the runoff,that accuracy is slightly worse than the former two,indicating the quality of potential evapotranspiration data can affect the accuracy of CREST model simulation's results.Therefore,it is more important to select the appropriate calculation formula of evapotranspiration than simply to improve the spatial resolution of evapotranspiration data in the hydrological simulation of the Qinhuai River Basin.In addition as per the simulation results of time scale for month,the accuracy of the result was improved,it meant that as the time scale became greater,the potential evaporation data would have less influence on the model.