| Taking Changting County as the research object,this study used Landsat remote sensing images from 1994,2006 and 2016 as the main data sources.Image features were extracted by using CART decision tree with texture information of object-oriented classification method.Both the dimidiate pixel model and the linear spectral mixture model were used to extract the fractional vegetation coverage.After comparing the accuracy of the two models,this study selected a model with higher precision to study the change of fractional vegetation coverage.Based on the Universal Soil Loss Equation(USLE)model,the five indicators of rainfall,topographic factor,soil type,fractional vegetation coverage and land use type were used to study the spatial and temporal differentiation of soil and water loss sensitivity in Changting County from 1994 to 2016.The main conclusions of the study are as follows:(1)The type of ground features in the study area were divided into forest land,cultivated land,construction land,unused land,water by using CART decision tree with texture information of object-oriented classification method.The overall accuracy of the three image classifications in 1994,2006,and 2016 were 80.00%,87.50%,and 86.57%,respectively,and the Kappa coefficients were 0.734,0.837,and 0.822,respectively.The overall classification effect was good.There was a large difference in reflectivity between the water area and other ground features,this shape was relatively regular and easy to distinguish,so the classification accuracy of water was high.Forest cover area was wide,vegetation spectrum was special,combined with elevation aid data,its classification effect was good.Due to the large number of types of cultivated land,construction land,and unused land,and the spectral characteristics of them were similar,those classification results in a relatively high error rate.The type of land-use cover in Changting County was mainly forest land.From 1994 to 2016,the area of forest land increased first and then decreased,but the rate of change was small,the area of construction land continued to increase,and the area of unused land continued to decrease,the area of cultivated land decreased first and then increased.Water area remained relatively stable after increasing from 1994 to 2006.(2)The fractional vegetation coverage was estimated using the dimidiate pixel model and the linear spectral mixture model,respectively.Based on the comparison of the inversion accuracy,the results showed that linear spectral mixture model by Landsat image merged with the normalized different vegetation index(NDVI)had highest estimation accuracy,because this model could increase the degree of distinction of endmembers.this model’s R2 Yalue was 0.827,the rootmeansquare error was 0.0865,and it could meet the research needs.The inversion of vegetation coverage by this model shows that vegetation coverage pattern in Changting County was dominated by high vegetation coverage,mainly in the western and northern parts of the territory,low vegetation coverage was mainly distributed in Hetian Basin.The overall vegetation coverage in the area is on the rise from 1994 to 2016,and the vegetation cover was mainly increased in the Hetian basin.(3)The quantitative assessment of soil and water sensitivity in 1994、2006、2016 was calculated by using entropy weight and multifactor weighted summation model.C1、C2、C3、C4、C5 represent the rainfall erosivity,soil erosion factor,terrain factor,fractional vegetation coverage,land use type.SS1994=0.1120 C1+0.1703 C2+0.2712 C3+0.3414 C4+0.1051C5SS2006=0.1233 C1+0.1874 C2+0.2856 C3+0.3086 C4+0.0921 C5SS2016=0.1176 C1+0.1763 C2+0.2828 C3+0.3432 C4+0.0801C5(4)Soil and water loss sensitivity was divided into five levels:micro sensitivity,slight sensitivity,moderate sensitivity,strong sensitivity,and intense sensitivity,by the method of natural break.This article combined with terrain factors such as elevation and slope to analyze the spatial and temporal differentiation of soil and water loss sensitivity.The results showed that the soil and water loss sensitivity in Changting was mainly at slight and moderate degree.The spatial pattern was characterized by the distribution features of the inner high and the outer low,and the intense sensitive areas were mainly distributed in the Hetian Basin with an elevation of<500 m and a gradient of<15°.Outside the Hetian Basin,the sensitivity level was lower and the ecological environment was better.In the change of sensitivity level of soil and water loss,from 1994 to 2016,different sensitivity levels all had different degrees of transformation,only the sensitivity level in some areas had increased,but the overall high-level sensitive areas shifted to low-grade sensitive areas.The trend was closely related to the Changting County government’s emphasis on the management of soil erosion.With the rise of elevation and slope,the areas of all sensitivity grades first increase and then decrease,which was closely related to the frequency of human activities. |
PDF Full Text Request