Linking landslides, land-use change and sediment connectivity: Insights from the head area of Three Gorges Reservoir

The relationship between landslides, land use, and sediment connectivity is not only a critical interdisciplinary topic, but also remains a challenging issue in assessing dynamic landslide susceptibility within reservoir areas. To explore the interactions among landslide, land use changes, and sediment dynamic, this study took Zigui Basin, the head area of the Three Gorges Reservoir, as the study area to examine this triadic relationship by single-factor detection and interactive detection. Here, we utilized Dynamic Attitude (DA) analysis to quantify land use changes and applied the Index of Connectivity (IC) to assess sediment connectivity evolution from 2018 to 2023. A multi-temporal analysis using the Landslide Susceptibility Index (LSI) was conducted to evaluate the degree of transformation in the three objects and the influence of these changes on the landslide susceptibility. According to the spatial analyst and statistics tools in ArcGIS, the results reveal that most of the landslides distributed in areas with high land use dynamic attitude, such as cultivated land transfers to forestland or garden plot, and the garden plot continuously increased across the study period with largest variation of 5% and an increment of 1.9%. Furthermore, linkage between land use and sediment transport can be effectively quantified by IC, and the resulting map indicated that garden plot increased, and catchment channel characteristics had a greater influence on the IC value than differences in vegetation cover. A comprehensive evaluation of the differences among the susceptibility maps reveals that the very high susceptibility classes are predominantly influenced by enhanced connectivity, whereas land use change has a greater effect on medium-low susceptibility region than that of sediment evolution. That is, both changes of land use and connectivity have positively correlated with landslide activity, but they exhibit differential influences on landslides susceptibility.