Abstract:
Riparian corridors though essential for sustaining riverine ecosystem services, face increasing degradation due to complex interactions between natural and anthropogenic pressures. A critical barrier to effective restoration is the inability to clearly identify degraded hotspots and to understand their underlying causal factors. Existing methods often lack spatial and temporal resolution and fail to capture the integrated dynamics of hydro-ecological processes, resulting in generic and suboptimal restoration measures. This study presents a novel, data-driven framework explicitly designed to delineate degraded riparian hotspots and attribute their degradation to specific causal factors. The approach integrates artificial intelligence-based K-means clustering with Decision-Making Trial and Evaluation Laboratory (DEMATEL) analysis, applied over a 20-year (2000–2020) period using remote sensing and hydrological data and field validation. Riparian health was assessed using six key indicators: Riparian Strip Quality Index (RSQI), Normalized Difference Vegetation Index (NDVI), Rainfall Erosivity Index (R factor), sediment load, runoff risk, and nutrient load. The framework effectively captured spatial heterogeneity in riparian degradation, revealing that elevated surface runoff and sediment load, which are primarily driven by intensive agriculture and settlement expansion—are major anthropogenic contributors to riparian impairment, while higher NDVI and RSQI values characterize relatively stable and vegetated zones. By explicitly linking degraded hotspot identification with their dominant drivers, this framework provides a significant advance towards targeted, causality-informed, and spatially scalable riparian restoration planning and implementation.