Changes In The Middle And Lower Reaches Of Ganjiang River Hydrological Factors And Midium And Long Term Hydrological Forcasting

Ganjiang river as jiangxi province's largest river is frequently threatened by floods and suffered serious losses.Therefore,it is of great significance to analyze and study the variation regularities of hydrological elements and actively improve the medium-and long-term hydrological forecast accuracy in the middle and lower reaches of the Ganjiang River for the development,utilization,protection,planning and deployment of water resources,flood control and disaster mitigation.Based on the middle and lower reaches of Ganjiang river as the study area,this paper through the Waizhou hydrologic station and in the middle and lower reaches of Ganjiang River Basin rainfall and evaporation stations to collect a series of data on precipitation,runoff and evaporation.Through using the method of linear regression method,cumulative departure curve,Mann Kendall nonparametric statistical tests,wavelet analysis method to reveal the variation of hydrological factors in the middle and lower reaches of the Ganjiang river basin in a changing environment.Then Based on the forecasting factors screened by three different methods,the BP neural network model was established to predict and compare the annual average flow and annual maximum flow of Wazhouzhou Station.The conclusion is as follows:(1)According to the middle and lower reaches of Ganjiang river analysis of discharge data from 1950 to 2012,has found the conversion of abundant and rare year is fairly balanced and runoff annual distribution more and more uniform,nearly 60 years the average flow rate showed a trend of weak increase.The middle and lower reaches of Ganjiang river of mean annual discharge had mutations in 1989.The establishment of Wanan reservoir and other water conservancy projects and operation may be the cause.On the frequency statistics of the amply flow and the low water of basin,find the probability of the the continual low water in the middle and lower reaches of Ganjiang river basin in greater than even the continual amply flow,Through the wavelet analysis in the evolution of the average flow rate is 2 ~ 4 a,6 ~ 10 a,14 ~ 18 a,25 ~ 30 a cycle of four kinds of time scale,and the main cycle is28 years.(2)According to the middle and lower reaches of Ganjiang river analysis of rainfall data from 1950 to 2012,has found that the increase of annual rainfall also present weak micro trend and the trend of increase is not obvious.In the 1990 s flood disasters mainly because rainfall is larger.Floods in April,may and June likely to happen because the middle and lower reaches of Ganjiang river in April,may and June have a large number of average monthly rainfall and changes smaller interannual variability of rainfall of the month.The middle and lower reaches of Ganjiang river of annual rainfall mutation starting time for the 1968,since 1968 basin rainfall began to showe increasing trend.Annual rainfall is 2 ~ 4 a,5 ~ 10.A,12 ~ 18 a,25 ~ 30 a cycles of a total of four kinds of time scale,the main cycle is 28 years,and this is in line with the annual average flow evolution trend.(3)According to the middle and lower reaches of Ganjiang river evaporation data analysis from 1979 to 2012,has found presents the significant decline in annual evaporation,evaporation of the four seasons are all shows the tendency of decrease.This can be proved that the middle and lower reaches of Ganjiang river obvious paradox "evaporate" phenomenon.The middle and lower reaches of Ganjiang river of annual evaporation mutation starting time for the 1984,since 1968 basin evaporation began to drop dramatically.The evolution of the annual evaporation is 10 ~ 14 a,16 ~20 a cycles change rule for 2 kinds of time scale.The main cycle is 15 years.(4)By comparing the prediction results,it is found that under the same neural network model,the correlation probability method combined with principal component analysis has higher fitting accuracy and accuracy than the other two methods.In the same case of the same forecasting factor screening method and model,The forecast accuracy of annual maximum flow is obviously lower than the forecast accuracy of annual average flow,which shows that it is more difficult to predict the extreme value of medium and long-term hydrological elements.