Study On Short-term Hydrological Forecast Of Ungaged Basin In Chongqing Based On SWAT Model

Hydrological forecast has always been an important research direction in hydrology.It is an important premise and foundation for scientific allocation of water resources,prevention and control of flood disasters and protection of ecological environment.However,due to the limitations of economic development and geographical conditions,for many medium and small watershed in China,the layout of rainfall stations is uneven,the number of hydrological stations is small,and the management system is not perfect,which leads to the lack of hydrological data,especially runoff data or short sequence years,and the existing hydrological data is not enough to establish a hydrological forecasting model with practical significance.Therefore,it is urgent to solve the problem of hydrological forecast in small and medium-sized watershed without data through new technologies and methods.In this paper,Chongqing small and medium-sized watershed as the research object,using satellite precipitation data combined with ground station data fusion method to improve the accuracy of precipitation data,and in order to solve the problem of shortage of runoff data in the target watershed,this paper uses the attribute similarity method to transplant the parameters between watersheds,so as to improve the hydrological simulation effect after transplanting the parameters between watersheds,so as to improve the accuracy of hydrological forecast in the area without data.This paper intends to identify the best combination of attribute indicators to achieve the identification of the best similar watershed.The main research contents and achievements are as follows:(1)In order to solve the problem of insufficient accuracy of precipitation data in hydrological forecast,the paper introduces GPM?IMERG remote sensing precipitation data as the source of satellite precipitation data.The mean deviation method,linear regression method,geographically weighted regression method and other precipitation data fusion methods are used to fuse the GPM?IMERG remote sensing precipitation data with the measured precipitation data of the ground station,so as to realize the correction of the ground station precipitation data to the satellite precipitation data.Using the leave one validation method,through the Correlation Coefficient,Mean Absolute Error and Root Mean Square Error,the ground validation and comparative analysis of the fusion precipitation data are carried out,The results show that:1)There is a certain correlation between the precipitation data of imerg early satellite and the precipitation data of ground stations.When there is no precipitation data,the precipitation data of IMERG-early satellite can be used instead of the ground precipitation data.2)The correlation and daily average precipitation distribution between the fused and corrected precipitation data and the ground measured data are better than those of the IMERG-early satellite precipitation data,which shows that the fused method has good correction effect on the satellite precipitation data,especially the geographically weighted regression method.It provides a reliable data source for the study of hydrological forecast in areas without data.(2)Using the precipitation data as the input data of SWAT model to simulate runoff,through the comparison of simulated runoff process,the application effect of precipitation data in hydrological forecast is analyzed,and the correction effect of precipitation fusion data is indirectly supplemented and verified.The results show that:1)The daily runoff hydrograph obtained by SWAT model driven by different precipitation data is in good agreement with the measured runoff hydrograph.It shows that SWAT model has strong applicability for small and medium-sized basins in Chongqing,and can be used as the model basis for hydrological forecasting research.2)The simulation accuracy of SWAT model based on GPM?IMERG satellite precipitation data is acceptable.3)Through the comparison of runoff simulation accuracy indexes of different fusion precipitation data,on the whole,the runoff simulation accuracy of geographically weighted regression method fusion precipitation data is the highest,and the whole has reached a good level.It has a good application value for the study of regional hydrological forecast.(3)In order to solve the problem of lack of runoff data in hydrological forecasting,the paper uses attribute similarity method to transplant parameters between basins.Through the way of basin attribute permutation and combination,the best attribute index combination for identifying similar basins is studied to improve the recognition accuracy of similar basins,so as to improve the hydrological forecast accuracy of non data basins.The results show that:Terrain feature / land use is the best way to identify similar watersheds in the study area,which plays a decisive role in identifying similar watersheds.In the future,the parameter region method should be used to study the hydrological forecast of small and medium-sized watershed without data in Chongqing.These two kinds of characteristic indexes should be given priority when judging the similarity of watershed.