Study On The Variation Characteristics Of Sediment Load And Its Influencing Factors In Typical Karst Watersheds

Under the influence of climate change and human activities,the trend of sediment load is becoming more and more significant.Sediment load plays an important role in global biological and geochemical cycles.It will not only change the surface morphology of the watershed,but also affect the structure and function of the ecosystem in the watershed.Therefore,it is necessary to clarify the variation characteristics and influencing factors of sediment load at different regional and temporal scales.In recent years,sediment load change and its driving factors have produced significant spatio-temporal variation characteristics,and the study of sediment load change and its influencing factors has become one of the most urgent problems in hydrology research.Based on this,a typical karst region in southwest China was selected as the research area to analyze the annual and monthly sediment load trends and influencing factors in specific karst watersheds on a temporal scale,and to identify the main controlling factors of sediment load changes in watersheds on a regional scale.Specifically,on the temporal scale,the Mann-Kendall(M-K)trend test method is used to determine the annual and monthly sediment load trend of the karst watershed.By means of M-K mutation point analysis,the abrupt change years of annual sediment load were obtained,so as to determine the baseline period and change period.Based on the data of precipitation,runoff,evapotranspiration and sediment load in the study area,the contribution rates of climate change and human activities to annual sediment load was quantified by simple linear regression method,double mass curve,sediment identify factor analysis method and elastic coefficient method.Pearson correlation analysis was used to determine the correlation between climate factors(temperature,precipitation)and vegetation,and Partial least squares structural equation model(PLS-SEM)was used to decouple the effects of climate change and vegetation dynamics on monthly sediment load.At the regional scale,the Person correlation analysis was used to obtain variables with significant correlation of average sediment load in watersheds.The main influencing factors of sediment load in karst watershed are discussed by using the Generalized Linear Model(GLM).The main research contents are as follows:(1)At the annual scale,the effects of climate change and human activities on the annual sediment load in the Wujiang,Yujiang,Hongshui and Xijiang karst watersheds were quantified by means of simple linear regression,double mass curve,sediment identify factor analysis and elastic coefficient method.From 1955 to 2015,the annual sediment load time series of Wujiang watershed,Yujiang watershed,Hongshui watershed and Xijiang watershed all showed a significant decreasing trend(P<0.05).The decreasing range of sediment load was in the order of Wujiang watershed > Hongshui watershed > Xijiang watershed > Yujiang watershed.Compared with precipitation,runoff and evapotranspiration,the variation coefficient of sediment load is the highest,ranging from 52% to 81%.By using M-K mutation point method to test the time series mutation,the variation year of sediment load in Wujiang and Yujiang watersheds is 2003,and that in Hongshui and Xijiang watersheds is 2004.The results showed that in addition to the sediment identify factor analysis method,the results of simple linear regression method,double mass curve method and elastic coefficient method all showed that the sediment load were mainly affected by human activities,with the contribution rates ranging from 72.52% to 100.52%.According to the sediment identify factor analysis method,the sediment concentration is the main factor affecting the change of sediment load in the four watersheds,and the contribution rates to the decrease of sediment load in the four watersheds ranges from 99.87%-153.81%.(2)At the monthly scale,the PLS-SEM model was used to decoupled the contributions of climate change and vegetation dynamics to monthly sediment load in the typical Karst watershed of Liujiang.Person correlation analysis results showed that there is a significant correlation between NDVI and temperature in each lag time(T0-T3)(P<0.05),and the strongest correlation is the temperature with one month lag(T1).Both EVI and LAI were significantly correlated with temperature in T0-T2 period(P<0.05),and T0 was the most strongly correlated with temperature.There was no significant correlation between T3 and temperature(P>0.05).In the correlation between precipitation and vegetation index at different lag time,the three vegetation indexes showed significant correlation with precipitation at four lag time(P<0.05).Among them,the correlation between precipitation in T2 period and NDVI and LAI was the strongest,and the correlation between EVI and one-month lag(T1)precipitation was the strongest.The climatic factors(precipitation,evaporation)of vegetation dynamic contribution rate is as high as 85%,vegetation dynamics and climate factor has contributed 73% of monthly runoff changes,vegetation dynamic,climate change and runoff of monthly sediment load accounted for 62%,suggesting that sediment load change is closely related to climate change and the dynamics of vegetation in the watershed.Climate factors had a direct effect on vegetation dynamics,and temperature had the most significant effect(β=0.65).Precipitation was the most significant direct factor influencing runoff change(β=0.81),while temperature was the most significant indirect factor influencing runoff change(0.06).Among the factors affecting sediment load,runoff has the largest direct effect(β =0.84),and precipitation has a significant indirect effect on sediment load through runoff(0.67).The antecedent temperature,temperature,precipitation and vegetation dynamics have negative effects on sediment load.(3)At the regional scale,the main controlling factors of sediment load in karst watersheds were analyzed by using the data of 40 karst watersheds.Through Person correlation analysis,variables that are significantly correlated with sediment load are obtained,among which PH,calcium carbonate content CC,carbonate C,water flow intensity index SPI,terrain humidity index TWI and river network density Dd indicate that the correlation is more obvious when P<0.01.The generalized linear model(GLM)explained 62.63% of the sediment load variation,and water area(W)explained the largest proportion of the sediment load variation,accounting for 26.59%.The second is carbonate(C),which can account for23.26% of sediment load variation.This indicates that water area(W)and carbonate(C)are the main influencing factors of sediment load at regional scale in karst watersheds.