Study On The Dynamic Characteristics And Simulation Of Runoff From 1969 To 2015 In Upper Yellow River Based On Standard Runoff Index

Global water reserves are decreasing in the context of climate change.The Upper Yellow River Basin is the main inflow area of the Yellow River,and its hydrological change has an important impact on the natural environment and social economy of the whole basin.In recent years,the variation of runoff in the Upper Yellow River has intensified,the nonlinear changes are prominent.These changes reduce the accuracy of medium–long–term hydrological simulation and prediction.How to clarify the temporal and spatial characteristics of runoff and improve the accuracy of simulation under the realistic background of increasing nonlinear changes are urgent problems.This study used hydrological data and meteorological data from 1969 to 2015 of 7typical hydrological stations in the Upper Yellow River above Lanzhou.First,we used standard runoff index(SRI)method to quantify the degree of runoff.On this basis,we used the improved Hilbert–Huang transform(HHT)and the empirical orthogonal functions(EOF)method to analyze the multi–scale time characteristics and spatial modes of the Upper Yellow River.Then,we used the Cross–correlation,wavelet coherence and singular value decomposition(SVD)methods to discuss the temporal and spatial response mechanism of hydrological elements to meteorological elements.Last,we based on the artificial fish least squares support vector machine(AF?LS?SVM),simulated the SRI in the Upper Yellow River.It is of great significance to reveal the rule of water resources change,ensure water security and lay a foundation for medium–long–term hydrological simulation.The main conclusions of this study are listed as follows:(1)At different time scales,different degrees of low water can occur in the Upper Yellow River,mainly with slight degree.The annual runoffs had been rich or low for many years.The changes of the SRI show a downward trend.While the seasonal and monthly SRI are more complex,the temporal and spatial distributions of the SRI are quite different.In spring,winter and January,the trends of SRI are different on different hydrological stations;the spatial difference is greater than that of other seasons and months.In summer,autumn and July,the changes in the SRI are declining;the low water situation is more serious than that in spring,winter and January.(2)The change periods of the SRI have some different at different time scales inthe Upper Yellow River.The annual SRI has important cycles of 3 years,8.2 yearsand 16.8 years.The seasonal scale SRI has important cycles of 3.5,8,15.5 and 26seasons.The monthly scale SRI has important cycles of 3 months,8 month,20months and 53 months.There are some synergistic change cycles between the SRIand the standardized precipitation evapotranspiration index series,especially on theannual scale of 4,8 and 16 years,seasonal scale of 4–8,monthly scale of 4–16months.It shows that these important cycles of SRI are mainly influenced bymeteorological elements.(3)There are two main spatial distribution modes of SRI in the Upper Yellow River,the consistent spatial distribution mode and northeast–southwest spatial distribution mode.The consistent spatial distribution mode is mainly affected by meteorological factors.The northeast–southwes spatial distributions may be greatly influenced by human activities are more complicated than the annual spatial distribution.On that basis we used artificial fish swarm least square support vector machine method to improve the model parameters.On the annual scale,this method can improve the simulation accuracy.This method has a good applicability in the SRI simulation of runoff in the upper reaches of the Yellow River and can be used in future studies.(4)The most important meteorological element that affects the spatial distribution modes of the SRI in the Upper Yellow River is precipitation,followed by temperature and snow.The impact of precipitation has a time lag.Temperature and snow cover have a great seasonal influence on spatial distribution modes of SRI.The important influence areas of annual scale meteorological elements are located in the southwest(precipitation),northeast(temperature)and Maqu(snow cover)of the Upper Yellow River.The important influence areas of seasonal meteorological elements have a tendency to expand.The important influence areas of monthly meteorological elements are similar to those on annual scale.The snow cover extends from Maqu to the southern area of the Upper Yellow River.