Study On Runoff Variation And Forecasting Of Weihe River Basin In Baoji Section

The hydrological conditions of the Weihe River Basin in Baoji section were complicated,and it were very significance for the development and utilization of water resources in Weihe river basin to study the climate change,runoff evolution and prediction research.Linear trend method,cumulative anomaly method,Mann-Kendall method,MTT,Yamamoto method,BP neural network and other theoretical methods were used for the study of Weihe River basin in Baoji section by analyzing the data of temperature and precipitation from 1956 to 2008 and the runoff data from 1956 to 2000,and the conclusions are as follows:(1)The annual average temperature of Weihe river basin in Baoji section was 13.2?.the annual average temperature change rate was 0.244?/10 a and increasing significantly?The rising trend was the most significant in winter,while the summer was not significant increasing trend.The mutations from cold to warm occurred in 1993.(2)The annual average precipitation of Weihe river basin in Baoji section was 660 mm,which showed a decreasing trend and its fluctuation was obvious.The distribution of annual precipitation was mainly concentrated in July to September,the decreasing mutations occurred in 1992.(3)The annual average runoff of Weihe river basin in Baoji section was 24.381×108m3,showed a decreasing trend.The decreasing trend of runoff in the autumn was the most significant,and the mutation of the annual average runoff occurred in 1970 and 1993.(4)There was a significant correlation between the temperature,precipitation and runoff of Weihe river basin in Shanxi province.Runoff and temperature had significant negative correlation,and runoff and precipitation had significant positive linear correlation.(5)Based on BP artificial neural network,four kinds of annual runoff prediction models were established.The input of the first model was runoff,and the input of the second model were temperature and runoff.,the input of the third model were precipitation and runoff,the input of the fourth model were temperature,precipitation and runoff.The prediction results showed that the prediction results of the four models were qualified,and the prediction accuracy of the fourth model was the highest,followed by the third model,the second model,the first model.