Study On The Relationship Between The Changes Of Snow Cover Area And Runoff In Upper Reaches Of Yalong River

The Yalong River,as the maximum level river among rivers in Jinsha River in the upper reaches of the Yangtze River,is one of China's richest hydropower resources river,also is one of 13 hydropower base in China's water conservancy planning and construction.The area above Ganzi as headwaters of the Yalong River,which characteristics of runoff will be influence water resources development and utilization in the middle and lower reaches of Yalong River from the short-term to long-term,and thereby affecting the development of regional social and economic.The variation analysis and prediction of runoff in Ganzi,not only contribute to comprehend characteristics of the water resources in Yalong River Basin,and provide a theoretical basis for the rational development and utilization of water resources in the basin,but also contribute to the control and configuration in development process of Yalong River Basin,and provide reference for runoff forecasting.In this paper,the area above Ganzi in Yalong River Basin is the study area,and the using data include HJ data(2009-2014)and MODIS snow products(2000-2014),runoff data of Ganzi hydrological station,and daily precipitation and temperature data of Qingshuihe,Shiqu and Ganzi meteorological stations in the study area.First of all,the paper proposed an improved extraction method of snow using HJ data and evaluated the accuracy of the GF data results based on research comparatively of the advantages and disadvantages of existing snow identification methods using HJ data,while,assess the accuracy of MODIS snow products by the snow cover area(SCA)extraction of HJ data as a reference value;Secondly,the paper analyzed the annual and inter-annual variation characteristics and trend of runoff,temperature and precipitation and SCA in different stages;Again,the related direction and the degree of correlation between the each stage runoff and influencing factors were analyzed,the affect and response among dry season runoff,snowmelt runoff,variation of snow,changes of temperature and precipitation in the study area were analysis and the prediction of runoff were discussed using correlation analysis and multiple regression analysis method.Finally,the simulation of HBV hydrological model in the study area is realized.In order to improve the prediction accuracy of runoff simulation,the HBV hydrological model was improved by regarding SCA as one of input data according to the results of correlation analysis.The results show that the accuracy of snow identification in this paper only using HJCCD as data sources is more than 90% that verified by GF data of the artificial neural network classification results,and the SCA could be extracted quickly and effectively.The runoff curve of area above Ganzi was the unimodal,which means it is a mix-recharge type river mainly supplied by mountains snowmelt and precipitation,runoff of effect season of melting snow and flood season showed a significant increasing trend.The snow range between accumulation and melting was significant.The average annual snow melting is 43332.33 km2.The average snow area during dry season had no obvious change and of SCA in effect of snowmelt season and flood season showed a slight increasing trend.The dry season runoff mainly recharged by groundwater,and had moderate positive correlation temperature and no significant correlation with precipitation and SCA;the effect of snowmelt season runoff had significantly positive correlation with precipitation and temperature and showed significant negative correlation with SCA,the elastic coefficient of precipitation,temperature and snow cover area were 0.289 and 0.779,0.004(through the 90% confidence degree of inspection),the increase of precipitation had 58.11% influence on the runoff change,the change of SCA had the 30.85% influence on the runoff change.Flood season runoff showed moderate positive correlation with precipitation,the elastic coefficient of precipitation was 0.219,the increase of precipitation had 89.07% influence on the runoff.The regression results,including the snowmelt affect runoff with winter SCA,last month mean temperature,last month precipitation and dry season runoff with last month runoff,though five regression equation test in 95% confidence degree of inspection and goodness of fit R2 were more than 0.80,means it could be used to roughly estimated impact of the dry season runoff and effect of snowmelt season runoff.Using HBV hydrological model in daily and monthly runoff simulation of Ganzi station,regarded 2005-2007 as the validation period,regarded 2010-2013 as the verification period,in verification period Ens efficiency coefficient were 0.8153 and 0.8702,RE were-8.399 and-8.4105 that means HBV hydrological model can simulate the runoff change in general.Simulation results underestimate the actual value of the runoff.The Ens efficiency coefficient of the improved hydrological model were 0.8726 and 0.9250,RE-4.3313 and-4.3725.The model efficiency coefficient and the error range were improved,which can better predict and simulate the runoff in the study area.