Driving Factors And Trend Prediction Of The Annual Sediment Transport In The Upper And Middle Reaches Of The Yellow River

The ecological protection and high-quality development of the Yellow River Basin are major national strategies in China,but soil erosion in the upper and middle reaches of the Yellow River is still serious due to human activities and natural factors.Soil erosion is one of the important factors affecting the sediment transport capacity of the Yellow River main stream.Therefore,understanding the trend and pattern of sediment transport in the upper and middle reaches of the Yellow River,and exploring the relationship between sediment transport and natural factors and human activities can provide a scientific basis for the prevention and control of soil erosion in the Yellow River Basin.In order to reveal the pattern and causes of sediment transport in the upper and middle reaches of the Yellow River and establish a simple,feasible and accurate model to predict the annual sediment transport,this study used the Google Earth Engine(GEE)platform to study the control reaches of Tangnaihai,Lanzhou,Toudaoguai,Longmen and Tongguan,which are five gauge stations on the main stream of the Yellow River.The factors that may affect sediment transport,such as vegetation,soil moisture,population,rainfall and land use classification,were obtained.The time series of annual sediment transport in the Yellow River main stream was studied using linear fitting and the Mann-Kendall mutation test,and the driving forces were analyzed using linear regression and random forest regression models.Finally,a prediction model for annual sediment transport in the Yellow River main stream was established using the selected significant driving factors,and the prediction results were compared with the ARIMA time series model(CK).The results showed that:(I)From 2001 to 2020,in the upper and middle reaches of the Yellow River,the sediment transport of the gauge stations(Tangnaihai,Lanzhou and Toudaoguai)in the upper reaches showed an upward trend.The sediment transport of the gauge stations(Longmen and Tongguan)in the middle reaches showed a downward trend.Meanwhile,the variation at sediment transport of the gauge stations in the upper reaches has fluctuated significantly in the past 20 years.And there were 2—3 mutation points,while the gauge stations in the middle reaches had few or no mutation points.In addition,the main cycle of sediment transport changes at the five gauge stations in the upper and middle reaches of the Yellow River was about 18 months.(2)After analyzing the driving forces,it was found that the driving factors in the linear regression model that had a significant impact on the amount of interannual variation in the annual sediment transport(after the first-order difference)were the Normalized Difference Vegetation Index(ANDVI),night light(AOLS),water body(AWB),summer precipitation(ASP)and soil moisture(100—289 cm)(ASM(100—289)).In the nonlinear regression model,the driving factors that had a significant impact on the amount of interannual variation in the annual sediment transport were ΔNNDVI,AOLS,ASP,ASM(100—289)and Net Primary Productivity(ΔNPP).(3)The results of 5-fold cross-validation showed that the random forest regression model had the best fitting effect,with an R2 of 0.55.Moreover,the fitting degree of the random forest regression models for the 5 gauge stations were higher than that of the ARIMA model(CK),and the average R2 of the random forest regression prediction for the 5 gauge stations’ sediment transport was 0.78.(4)By adding categorical variables for the upper and middle reaches of the Yellow River(the gauge station located in the upper reaches of the Yellow River was denoted as 1,and the gauge station located in the middle reaches of the Yellow River was denoted as 2),an optimized random forest regression prediction model was obtained.The average fitting degree of the model is better than the original prediction model,with an average R2 of 0.84.To summarize,the trend of sediment transport in the upper and middle reaches of the Yellow River from 2001 to 2020 was not the same.The upper reaches show an upward trend,while the middle reaches show a downward trend and the mutation points in sediment transport in the upper reaches are more complex than those in the middle reaches.The five main driving factors that significantly affect the amount of interannual variation in the annual sediment transport in the upper and middle reaches of the Yellow River were ANDVI,AOLS,ANPP,ASP,and ΔSM(100—289).The optimized random forest regression model is a simple,feasible,and accurate model for predicting the annual sediment transport changes in the upper and middle reaches of the Yellow River.This study has important reference value for protecting the ecological environment of the Yellow River Basin,formulating soil and water conservation measures,and managing and utilizing water resources.