A Study On Soil Erosion Patterns And Their Responses To Land Use Changes In A Low Mountain And Hilly Region Of Northeast China

The northeast low mountain and hilly regions of China,as the most concentrated distribution area for black soil resources,are important for ensuring food and ecological security.In recent years,large-scale,intensive land development and utilization in these areas and the shortage of water and soil conservation measures during the reclamation process have induced serious soil erosion and infertility,restricting regional soil productivity and ecosystem services.Statistics show that the black soil area in Northeastern China is one of the areas with the fastest growth rate of soil erosion in China since the start of the 21st century.Therefore,identifying the intensity of soil erosion and its spatial distribution pattern is of far-reaching significance to prevent and control soil erosion and promote the implementation of sustainable land use and protection policies.Research on regional soil erosion patterns mostly relies on soil erosion models,but is limited by the lack of high-resolution soil information data.The quantification of soil erodibility usually relies on low-density point data,which makes it difficult to accurately characterize its spatial distribution and evolution characteristics.In addition,while unreasonable land use is one of the important factors aggravating regional soil erosion,most existing studies focused on estimating the amount of soil erosion for different land types,without exploring the relationships between land use intensity,landscape fragmentation,and soil erosion.Therefore,it is urgent to establish a method for quantification and spatial characterization of soil erodibility with high time-efficiency and high spatial resolution.In this study,through spatial characterization of soil erosion patterns and identification of erosion hotspots,the effects of land use intensity and landscape fragmentation on the spatial differentiation characteristics of soil erosion on cultivated land were identified.This work is expected to provide a theoretical and evidential basis for the implementation of major national black soil protection projects.This work took place in Jiutai District of Changchun City,which is a typical northeast low mountain and hilly region.The purpose of this research was to investigate the pattern of soil erosion and its response to land use change at the county scale,and to model soil organic carbon（SOC）inversion with high precision using multi-temporal Sentinel-2 images.The Revised Universal Soil Loss Equation（RUSLE）was introduced to spatially represent soil erosion patterns and accurately identify erosion hot spots of typical counties in the low mountain and hilly regions of Northeast China.Finally,the relationship between the soil erosion pattern and land use change factors was explored by employing Geographically Weighted Regression（GWR）,and the impact of land use intensity and cultivated land landscape fragmentation on soil erosion was also identified.The study provides a basis for the precise implementation of regional water and soil conservation measures and the formulation of macro-level land management policies.The main results are as follows:（1）The spatial characterization of soil erodibility factors based on hyperspectral remote sensing retrievalSOC content is highly correlated with soil erodibility,so it is often used as a core index to calculate soil erodibility factors in the RUSLE equation.However,efficient measurement and spatial fine characterization methods for soil erodibility factors were not possible for the study area,due to the lack of local high-resolution SOC data and the high cost of large-scale and multi-frequency SOC quantification by traditional wet chemistry method.To solve these problems,the method of SOC high-precision quantification and high-resolution spatial mapping was used in this work,based on the latest land surface soil parameters retrieved from Sentinel-2 spectral images.We aimed to provide data support for spatial visualization of soil erodibility factors.The results show that SOC content can be predicted based on the multi-temporal bare soil spectrum（R²=0.62,RMSE=0.17）,and that a 10-meter-resolution SOC distribution map of the cultivated land can be generated by certain core methods,such as pixel extraction and multi-phase synthesis of bare soil,modeling of least squares SOC inversion,uncertainty analysis of prediction value,etc.Compared with single-date remote sensing inversion,multi-temporal bare soil pixel spectral data sets provided a more robust,more extensive,and more accurate SOC prediction model.Compared with the SOC prediction model based on near-earth hyperspectral data,we found that the bands that play a decisive role in the hyperspectral inversion prediction are highly consistent（all are short-wave infrared bands）.The stability and feasibility of SOC content prediction based on Sentinel-2 data were further verified.Based on the pixel-level SOC distribution data,a new method for estimating soil erodibility factors and high-resolution spatial characterization was established,and a spatial distribution map of soil erodibility factors was generated.These results provide a solid data foundation for further application of the RUSLE model and analysis of spatial patterns of soil erosion.（2）The spatial pattern of soil erosion and the law of soil organic carbon migration and redistribution on slopesThe spatial characterization of soil erosion patterns with both high precision and high time-efficiency and the identification of erosion hot spots are essential for ascertaining the degree and scope of soil erosion,and for precise positioning of regional soil and water conservation policies.Based on the RUSLE model framework,the estimation of soil erosion amounts and their spatial distribution characteristics were studied in this thesis,supported by the high-resolution soil erodibility factor data.Different erosion intensities were regarded as erosion landscapes inlaid with various erosion intensities,and the landscape pattern of soil erosion was analyzed.We also studied the spatial migration,redistribution,and transformation of SOC driven by soil erosion at the slope scale.The results show that,in 2019,the total soil erosion of cultivated land in the study area was mainly mild and slight erosion,and the proportion of cultivated land affected by extreme intensity and severe erosion was relatively small.The average soil erosion modulus was 7.09 t·hm^-2·a^-1.Based on the results of spatial concentration and hot spot analysis,the spatial distribution characteristics of soil erosion were analyzed.We found that most of the cultivated land in the study area was located in the southeast and the northeast regions of the slope farmland.With the increase of altitude and terrain slope,the proportion of mild and slight erosion areas gradually decreased,while the proportion of extreme intensity and severe erosion gradually increased.These are closely related to the increase of soil erodibility caused by the time-space migration and erosion of SOC in complex terrain.The distribution of slight and light soil erosion types was more concentrated,but the shape was more complex,the distribution of extremely intense and severe erosion was scattered,and the landscape shape was relatively simple.In order to explore the coupling mechanism of soil erosion with soil aggregate structure and SOC stability,we selected typical sloping cultivated land in hot spots of soil erosion to study SOC migration and redistribution driven by soil erosion-deposition processes.We measured particle size fractions,SOC content,and carbon stable isotope(δ¹³C)of soil aggregates in different positions of slope（stable region,erosive region,and deposition region）.We found that the preferential migration of fine soil material caused by erosion results in an increase in the percentage of clay and silt particles in the sedimentary area,and of SOC content and“young,”unstable SOC content(indicated byδ¹³C).The results indicate that soil erosion intensity and the high spatial heterogeneity of the soil carbon pool should be considered in soil management and as conservation aspects under precision farmland management.（3）The spatial response of soil erosion to land use intensity and landscape fragmentationBased on an analysis of the main characteristics of land use change from 1996 to2019 in the study area,this thesis analyzes the impact of land use change on soil erosion of cultivated land in the hilly area from the perspective of land use intensity and landscape fragmentation based on the GWR model.The results show that land use has changed greatly from 1996 to 2019 in Jiutai,especially from 1996 to 2009.The loss and replacement of cultivated land occurred alternately,as construction land increased and ecological land decreased gradually.Under the influence of natural factors and social economic factors,cultivated land had the highest change frequency.Cultivated land transformed from natural forest had the largest average soil erosion modulus.The GWR model was used to understand the influence degree and spatial distribution of external factors on soil erosion of the cultivated land.The result shows that slope had the most significant effect on soil erosion,with a strong positive effect.The increase of land use intensity and fragmentation of cultivated land both promoted soil there,especially in the main distribution areas of sloping cultivated land in these study areas.These findings are closely related to the conversion of a large amount of forestland to sloping cultivated land,resulting in increased land use intensity and the gradual fragmentation of marginal arable land.Finally,based on the current soil erosion pattern and the corresponding changes of land use,this thesis puts forward some suggestions for the prevention and control of soil erosion on the cultivated land in the low mountain and hilly regions of Northeast China.We provide a scientific basis for the sustainable utilization of land resources and the harmonious development of mankind’s relationship to the land in hilly areas.