| This study focuses on Qingshuihe County in Inner Mongolia as the study area.Landsat 5/8 remote sensing images and land use data from 1990,2000,2010,and 2019were selected to analyze the land use transfer characteristics over the past three decades.Alongside,essential information such as Digital Elevation Model(DEM),meteorological data,and soil type data specific to the study area were incorporated.The Revised Universal Soil Loss Equation(RUSLE)model was used as a benchmark to assess soil erosion.Leveraging remote sensing(RS),Geographic Information System(GIS),and other analytical software,the six factors of the RUSLE model were extracted and calculated,resulting in a raster database containing each factor index.This facilitated a quantitative assessment of the spatial and temporal variations in soil erosion.Moreover,by integrating multi-source data from natural,socio-economic,and geographical aspects,a random forest regression model was employed to rank the importance of 11 selected drivers and compare predicted and actual values of soil erosion across different time periods.The main findings of the thesis are as follows:(1)The main land use types in Qingshuihe County in 2000-2019 are grassland and cropland,with grassland being the most widely distributed land use type and residential and construction land being the land use type with the largest change.In terms of land use transfer patterns,the largest increase in residential and construction land and the largest decrease in cropland were observed from 1990 to 2000;from 2000 to 2010,the conversion of cropland into grassland and forest land was the main land use change pattern in Qingshuihe County during this period;from 2010 to 2019,the The main land use pattern in this period is the increase of forest land and residential and construction land,mainly from the transformation of grassland and cropland.According to the analysis of the results of land use center of gravity shift in Qingshuihe County,the center of gravity of unused land,watershed and cropland in Qingshuihe County all shifted to the southwest,the center of gravity of forest land shifted to the southeast,and the center of gravity of grassland and residential and construction land shifted to the northeast during1990-2019.(2)The average soil erosion modulus in Qingshuihe County exhibited a decreasing trend from 1990 to 2019,declining from 3459.21 t/km~2·a to 1771.18 t/km~2·a.The soil erosion area also decreased from 2812.01 km~2in 1990 to 2769.31 km~2in 2019,with a temporary increase to 2679.88 km~2in 2000.However,overall,the soil erosion area has shown a decreasing trend over the past 30 years.This indicates an improvement in the soil erosion condition and a gradual amelioration of the soil erosion situation in Qingshuihe County.In terms of the area share of erosion intensity,mild erosion dominated the soil erosion patterns from 1990 to 2010.However,in 2019,slight erosion became the predominant category.Over the past 30 years,the area of all levels of erosion experienced both increases and decreases.Notably,the area of slight erosion increased by1124.54 km~2during this period.Light,moderate,strong,very strong,and severe erosion exhibited a trend of initially increasing and then decreasing,with decreasing areas of266.47 km~2,208.81 km~2,461.55 km~2,139.07 km~2,and 91.34 km~2,respectively,between1990 and 2019.(3)The operation results of the random forest regression model showed that the model R~2=0.78,which is a good fit,indicating that the model is applicable to the prediction of soil erosion in Qingshuihe County.According to the calculation results of the factors influencing the occurrence of soil erosion,the results of five factors affecting the occurrence of soil erosion in Qingshuihe County were significant,and their importance was ranked as follows:slope>FVC>nearest distance to the settlement>nearest distance to the road>average annual precipitation.The slope is proportional to the degree of soil erosion,the greater the slope and the more treacherous the terrain,the more serious the soil erosion;followed by the nearest distance from the settlement,the nearest distance from the road and the average annual precipitation have the same trend with the predicted erosion and show a positive correlation overall;the vegetation factor has a negative correlation with the degree of soil erosion. |
