1990-2023年三川河流域景观格局指数与土壤侵蚀的时空关联OACHSSCD

Taking the Sanchuan River Basin in the hilly-gully region of China's Loess Plateau as a case study,this research couples landscape ecology theory with soil erosion dynamics to establish a multi-scale analytical framework integrating"Pattern Recognition-Process Simulation-Mechanism Interpretation,"thereby revealing the impacts of landscape pattern evolution on soil erosion.Utilizing multi-source remote sensing data and universal soil loss equation(USLE)model inversion,we quantitatively characterized the spatiotemporal variations in soil erosion intensity across the Watershed from 1990 to 2023.Landscape pattern indices were analyzed through Fragstats4.2 and moving window methodology,with a Geographically and Temporally Weighted Regression(GTWR)model employed to quantitatively decipher the nonlinear relationships between these indices and soil erosion modulus.Results demonstrate:(1)An overall improvement in soil erosion modulus from 1,116 t km-2 a-1 to 652 t km-2 a-1,accompanied by a 32.3%reduction in erosion-prone areas(from 1,341.40 km2 to 908.14 km2),signifying substantial efficacy of ecological restoration projects;(2)Dynamic equilibrium in landscape patterns characterized by increasing aggregation,enhanced connectivity,and reduced fragmentation during 1990-2023,reflecting synergistic effects of human activities and ecological rehabilitation;(3)The influence of key landscape indices(PD,AWMSI,CONTAG,IJI)on erosion modulus is not static but a dynamic process driven by human intervention intensity and ecosystem function evolution,where intensified fragmentation and diminished connectivity in central-western sectors elevated erosion risks,and landscape homogenization in southern urbanized areas weakened regulatory effects.Especially after 2005,the influence of landscape patterns on soil erosion has shifted from being"natural-pattern-dominated"to"human-activity-dominated."This study elucidates landscape pattern-erosion interactions and provides quantifiable optimization paradigms for integrated"mountains-waters-forests-farmlands-lakes-grasslands"system management,while offering significant practical insights for coupled landscape ecological processes and high-quality regional ecological conservation.