| Clarifying the relationship among vegetation cover,precipitation and sediment yield is a long-standing key issue in understanding land surface processes.The vegetation cover on China’s Loess Plateau has showed an obvious increase from the northwest to the southeast since the large-scale ecological restoration program-"Grain to Green" was launched in 1999.Importantly,the relationship between sediment yield and vegetation cover as well as precipitation has been greatly affected by the human activity.However,there still lacks quantitive research on the response of sediment yield to vegetation cover and precipitation on the Loess Plateau.Therefore,it is urgent to fill the knowledge gap to explain and predict the change of sediment yield at a large spatial scale and long-term period since the“Grain to Green”program.This study analyzed mathematical relations among vegetation cover,precipitation and sediment yield at two spatial-scale(the entire study area scale,and the typical landform types scale)based on multiple linear regression model.This paper further compared the mathematical relationsat different temporal scales(year,flood season and the four seasons)during the period of 2000-2012.Meanwhile,we quantified the relationships among vegetation cover,precipitation and sediment yield during the periods of 2000-2004 and 2005-2012.In addition,we discussed effect of the trade-off relations between vegetation cover and precipitation on sediment yield at large-spatial scale using newly developed coupling index RV.The results showed that:(1)The value of NDVI presented a largely increase trend from the northwest to the southeast since 2000.Generally,the annual and summer NDVI increased at the rates of 0.0068yr.-1 and 0.0140yr.-1(P<0.01)respectively over the entire sediment-rich region,while the annual growth rates of NDVI in Loess Middle Mountains,Loess Hills and Loess-Desert Transitional Area were 0.0070yr.-1,0.0115yr.-1 and 0.0095yr.-1(P<0.01),respectively.In addition,NDVI in Wudinghe watershed,Jialuhe watershed,Touweihe watershed,Kuyehe watershed and Huangfuchuan watershed were less than 0.3;NDVI in the lower Wudinghe watershed,Qingjianhe watershed,Fenchuanhe watershed and the upper and middle reaches of Beiluohe watershed were between 0.3 and 0.6;NDVI in Ziwuling mountains and Huanglong mountains in the southern part of the study area were more than 0.6.(2)The value of precipitation presented a remarkable increase trend from the northwest to the southeast from 2000 to 2012.Generally,the annual and summer rainfall increased at lower rates of 4.4877yr.-1 and 0.9071yr.-1(P>0.05),respectively.Furthermore,rainfall in Loess-Desert Transitional Area(in the northwest of the study area)was less than 430 mm;the rainfall in the middle of this region with Loess Hills was between 430 mm and 530 mm;while the rainfall in the Loess Sustained-Middle Mountains and Loess Tablelands(in the southern part of this region)were larger than 530 mm.(3)The annual sediment yield of 18 hydrological stations in the sediment-rich region showed a significant decrease and then tends to be stable since the "Grain to Green" was performed in 2000.More specially,the annual sediment yield decreased at a rate of 19465100 t-yr."-1.In addition,the high erosion modulus was in Yanshui watershed,Wudinghe watershed and Huangfuchuan watershed.(4)By analyzing the effect of vegetation cover on sediment yield,the nonlinear mathematical relation between NDVI and sediment yield can be detected at a large spatial scale.Also,this research found that when the NDVI value is less than 0.36(annual NDVI)or 0.54(summer NDVI),the value of sediment yield increases with the NDVI value grows;besides,when the NDVI value is bigger than(annual NDVI)or 0.54(summer NDVI),the sediment yield value shows a decrease trend and then tends to be sTab.In addition,this paper detected the nonlinear mathematical relationship between NDVI and sediment yield in the process of time:the value of sediment yield gradually decreases when NDVI is less than a certain threshold,when exceeding this threshold,however,the value of sediment yield tends to be sTab.(5)According to analyzing the effect of annual rainfall on sediment yield,this research observed a nonlinear mathematical relation between annual rainfall and sediment yield and when the precipitation is less than 490 mm,the sediment yield value increases with the rainfall grows;when the value of rainfall exceeds this threshold,shows a decrease trend and then tends to be sTab.(6)At the large-spatial scale,by analyzing the effect of multi-year average precipitation(P)and multi-year average NDVI(NDVI)on sediment yield(SY),this work built the multiple linear regression model among P,P2,NDVI,NDVI2 and SY.More specially,the mathematical relations for the period of 2000-2004 and 2005-2012 was respectively given as:SY=27.961-2.667NDVI-2.965NDVI2+0.089P-1.005P2(a=0.097,R2=0.430),SY=-12.079+25.011NDVI-10.586NDVI2-63.673P+41.906P2(a=0.042,R2=0.510).In the process of time,this paper also established the annual,flood season’ and seasonal multiple linear regression relations among P,NDVI and SY at two spatial-scale(the entire study area scale,and the typical landform types scale).However,the sigficance levels of the fitting relations had larger difference at the different time scales and spatial scales.(7)According to effect of the coupling index RV(RV)on sediment yield(SY),this work found that the correlation relationships between RV and S Y were especially better at year scale,flood season scale and summer scale.In addition,this study detected some crucial thresholds between RV and SY:when RV is less than-0.2 in the entire sediment-rich region and Loess Hills and-0.1 in Loess Middle Mountains and Loess-Desert Transitional Area,the value of sediment yield decreases with the RV increasing;when the RV exceeding the values,the value of sediment yield increases with the growth of RV. |
PDF Full Text Request