Sediement Yield Changes And Their Controlling Factors In The Loess Plateau Catchments

The process of soil erosion is generally slowing down and ecological environment is gradually improving,contributed to global climate change and intense human activities,as well as implementation of ecological restoration policies,such as decrease of precipitation,building of reservoirs and introduction of“Grain to Green”Project,in recent decades in Loess Plateau.Therefore,it is worth to go further clarify relationship between sediment yield and driving factors.The objectives of this study were to explore the correlation and establish relation model between sediment yield discharge and its key driving indicators—precipitation,vegetation coverage,land use types,topography and water system,as well as soil and water conservation measures.Annual sediment load series for 46 hydrological stations from 1961-2016,daily rainfall series for 231 meteorological stations from 1961-2014 and monthly maximum vegetation coverage data(NDVI)from 1981-2016 in Loess Plateau were processed at a consistent time.In addition,land use data in different periods(1991/2000/2005/2010)and measures data(terraces and check-dams)also were collected.The temporal and spatial variation characteristics of sediment yield in the Loess Plateau and its contribution to the Yellow River,as well as the change characteristics of environment indicators were analyzed by Mann-Kendall trend test,Pettitt test,Anomaly accumulation methods and Regime shift index.Main driving indicators were selected from correlation matrix between sediment transport modulus and various indicators,which was tested by Pearson correlation coefficient.Finally,the relationship modulus is confirmed by using multiple statistical regression(partial least square regression,PLSR),based on machine learning(python).The main conclusions are as follows:(1)Annual sediment transport from 1961-2016 were decreasing at a significant level of99%and occurred two change-points in rivers of Loess Plateau,which experienced high,stable and low sediment transport stages.The average annual sediment transport is 0.50×10~8t/a,0.88×10~8 t/a,5.66×10~8 t/a and 8.28×10~8t/a,respectively,in Lanzhou,Toudaoguai,Longmen Tongguan,located in main stream of Yellow River.Sediment transport of the main tributaries,such as Tao River,Huangfuchuan,Kuye River,Wuding River and Weihe River,reduced varyingly in different periods.And abrupt-change points were detected in 1970s and1990s for most of tributaries.The sediment transport modulus for each sub-basin in Loess Plateau also reduced significantly from more than 20000 t/km~2/a to 4000 t/km~2/a,with significant level of 0.01,especially in some sub-basins located in the middle of the Yellow River,(2)With the reduction of sediment transport from 1961-2016 in rivers in Loess Plateau,the sediment contribution of different sections of the main stream to the Yellow River was also reduced year by year.The greatest contribution comes from Toudaguai-Longmeng,with annual mean sediment load of 4.77 million t,which is much large than 2.62 million t in Longmen-Tongguan,and sediment load with smallest value of 0.38 million t is in Lanzhou-Toudaoguai,the decreasing rates are 0.0151,0.1848 and 6.17 million t,respectively.(3)Although there are same spatial distribution characteristics,increasing gradually from northwest to southeast,change trends are going in the opposite direction in annual rainfall and vegetation coverage in Loess Plateau—precipitation is decreasing slightly,while vegetation coverage is increasing significantly.From 1990 to 2010,the main land use types are forest land,grassland and cultivated land,accounting for about 90%of the total area,and forest land and grassland are increasing year by year,while cultivated land is decreasing year by year.(4)Key driving indicators affecting the variation of sediment yield in watershed of Loess Plateau are annual rainfall(P),vegetation coverage(NDVI),grassland area ratio(GR)and river network density(Rd).The relation modulus between sediment transport modulus and them is as follows:Sm=0.0014P+0.4648NDVI-3.8468Rd+2.1555GR-0.5476,R~2=0.848,Which has a good verification R~2,0.507,...