Study On The Relationship Between Typical Hydrological Variables In The Source Area Of The Yellow River

Taking the source area of the Yellow River as the research object,this paper collects the hydrological time series data of daily measured rainfall,runoff and sediment at Tangnaihai hydrological station in the source area of the Yellow River.Firstly,the co-integration theory and structure change co-integration theory are used to study and analyze the long-term equilibrium and short-term fluctuation relationship between the measured rainfall,runoff and sediment before and after the abrupt change in the source area of the Yellow River.Then,the original data of rainfall,runoff and sediment in the source region of the Yellow River are decomposed into multi time scales by using CEEMDAN method,and the multi time scale co-integration modelsare constructed to study and analyze the dynamic equilibrium relationship of the three hydrological variables in different time scales.Finally,the quantile regression method is used to establish the quantile regression equation,to study the interaction and relationship changes of the three hydrological variables at different quantile levels,and to construct the quantile cointegration model,and to study the long-term equilibrium and short-ter m fluctuation relationship between the three variables at different quantile levels.At the same time,through the application,combination and improvement of the above methods,the established models realizethe simulation and prediction of the annual runoff in the source area of the Yellow River,and verifies the reliability of the relationship between the typical hydrological variables in the source area of the Yellow River by comparing and analyzing the simulation and prediction accuracy of the models.The main results are as follows:(1)There is a co-integration relationship between the measured rainfall,runoff and sediment time series of Tangnaihai hydrological station in the source area of the Yellow River from 1966 to 2013.The establishment of Rainfall-Runoff error correction model,Runoff-Sediment error correction model and Rainfall-Runoff-Sediment error correction model reflect the long-term equilibrium and short-term fluctuation relationship among rainfall,runoff and sediment,and reveals that runoff is not only affected by rainfall,but also affected by runoff deviation from equilibrium level in the previous year,the influence of rainfall on runoff is significant,but the influence of runoff on sediment is not significant.The average absolute percentage error of the Rainfall-Runoff-Sediment error correction model in the simulation period is the smallest,which is 6.93%,and the simulation accuracy is the best.The qualified rate of prediction in the prediction period is87.5%,and the accuracy grade is grade A,so the prediction effect is the best.(2)The abrupt change points of runoff series in the source area of the Yellow River are 1989 and 2002.The typical human activities like reservoir construction and climate change are the main reasons for the abrupt change of runoff series in the source area of the Yellow River.The structure change co-integration test showed that there was a structure change co-integration relationship among rainfall,runoff and sediment.After the abrupt change in 1989,the impact of rainfall on runoff decreased and that of runoff on sediment increased.After the abrupt change in 2002,the impact of rainfall on runoff increased,while the impact of runoff on sediment decreased.The abrupt change in rich year has little influence on the relationship between rainfall,runoff and sediment,while the abrupt change in dry year has great influence on the relationship between rainfall,runoff and sediment.The structure change error correction model shows that the short-term impact of rainfall on runoff and runoff o n sediment are different in the source area of the Yellow River,and the impact of rainfall on runoff is stronger than that of runoff on sediment.The average absolute percentage error of the Rainfall-Runoff-Sediment structure change error correction model in the simulation period is the smallest,which is 6.43%,and the simulation accuracy is the best.The qualified rate of prediction in the prediction period is 100%,and the accuracy grade is grade A,and the prediction effect is the best.(3)Rainfall,runoff,and sediment in the source area of the Yellow River have periodic changes on multiple time scales,all of which have the same short-period change,and the periodic year is 2 to 5 years.In the medium period,the three change cycle years are different,and the medium and long period change scales are quite different.In particular,the sediment change period spans a large scale.On the long period change scale,rainfall is 28 years,and runoff is 32 years,sediment is 41 years.The RES component of the trend item shows that the rainfall first decreases and the n increases,and both runoff and sediment show a decreasing trend.The establishment of the error correction model reveals the co-integration relationship of the three on different time scales.The Rainfall-Runoff-Sediment component error correction coupling model has the smallest average absolute percentage error in the simulation period of 6.24%,the simulation accuracy is the best.The qualified rate of prediction in the prediction period is 100%,and the accuracy grade is grade A,and the prediction effect is the best.(4)Quantile regression methods are applied to rainfall,runoff,and sediment to perform regression analysis at different quantile levels,revealing the relationship between rainfall,runoff and sediment in different rich,flat and dry years.Rainfall-Runoff quantile regression equation shows that at low quantiles,the relationship between rainfall and runoff is not obvious,and at high quantiles,rainfall has a significant impact on runoff.The regression equation of Runoff-Sediment quantile shows that the relationship between runoff and sediment is not obvious,and the relationship between the two is better in rich years.The Rainfall-Runoff-Sediment quantile regression equation shows that the high quantile level quantile equation has a better fit,and the relationship between rainfall,runoff and sediment in rich years is better.The series of rainfall,runoff,and sediment have a co-integration relationship at different quantile levels,and the quantile error correction model reflects the long-term equilibrium and short-term fluctuation relationships of the three in rich,flat,and dry years.The Rainfall-Runoff-Sediment quantile error correction model has a minimum average absolute percentage error of5.85% during the simulation period,the simulation accuracy is the best,the prediction pass rate in the prediction period is 100%,the accuracy level is A,and the prediction effect is the best.