The Study Of Runoff Variation And Forecasting Of Weihe River Basin In Shaanxi Province

Since the 1990 s,the runoff of Weihe river basin in Shaanxi province reduced constantly.The study of runoff variation and forecasting of Weihe river basin in Shananxi province is particularly important for the development and utilization of water resource in Shaanxi province.In this paper,Linear trend method,departure method,Kendall non-parametric rank correlation test method,Mann-Kendall method,wavelet analysis,RBF neural network and other theoretical methods were used for the study of the climate data(temperature,precipitation)and runoff data of Weihe river basin in Shaanxi province from 1960 to 2005.The following conclusions were obtained:(1)The annual average temperature of Weihe river basin in Shaanxi province was 6.8?.The reduction rate of annual temperature was0.5?/10 a and the coefficient of variationvC was 0.2.The annual average temperature was significantly decreased,mutations occurred in 1981 and1994.(2)The annual average precipitation of Weihe river basin in Shaanxi province was 662 mm.On the seasonal distribution,precipitation was(3)concentrated in the summer,the spring and autumn followed and the least in winter.The mutation of annual precipitation occurred in 1965,1967,1984 and 1986.(4)The annual average runoff of Weihe river basin in Shaanxi province was 82.57×108m3.The annual extreme value was 5.57 and annual runoff variation coefficient was 0.39,The annual runoff mutation occurred in1982 and 1994.The annual runoff change was significant,the annual distribution non-uniform coeffcient was 1.73,which showed the annual runoff distribution of Weihe river was uneven.(5)There was a significant correlation between the temperature,precipitation and runoff of Weihe river basin in Shaanxi province.Temperature and runoff had significant negative correlation and precipitation and runoff had significant positive linear correlation,(6)Wavelet analysis theory was used to obtain the conclusion of the runoff had variation period of about 2a,6a and 18 a.The forecast was goven that the runoff would remain in flood season and showed a increasing trend,which were consistent with the observed data.(7)RBF neural network model was established for the forecast of runoff of Weihe river basin in Shaanxi province from 2003 to 2005.The fitting accuracy was 0.14% and precision accuracy was 5.78%,The precision accuracy was high and can be used to the runoff forecast of Weihe river basin.