Remote-Sensing-Informed AHP Prioritisation of Downstream Risk after the Matai'an River Landslide-Dam Overflow Event in Hualien, Taiwan

Landslide-dam overflow events in steep mountain catchments can rapidly transfer upstream geomorphic instability into downstream flood-sediment risk, especially where settlements, roads, and bridges occupy narrow river corridors. This study develops a remote-sensing-informed analytic hierarchy process (AHP) framework to prioritise downstream risk after the Matai'an River landslide-dam overflow event in Hualien, Taiwan. Multi-source evidence—including DEM-derived terrain information; pre- and post-event optical imagery; GIS layers of roads, bridges, settlements and critical facilities; hydro-meteorological records; event chronology and expert judgements—was integrated into a hazard-exposure-vulnerability hierarchy comprising 12 subcriteria. Pairwise comparisons were aggregated using the geometric mean, and all consistency ratios were below 0.10. Eight downstream analytical units were scored on a five-point relative scale and ranked using the resulting global weights. Hazard received the highest criterion weight (0.481), followed by vulnerability (0.314) and exposure (0.205). The leading global priorities were landslide-dam overflow potential (0.159), upstream landslide and sediment supply (0.124), channel overbank-flow likelihood (0.108), terrain and relative elevation (0.097), rainfall persistence (0.089) and embankment and protection capacity (0.085). G02, G01 and G03 were classified as high-priority units, with composite scores of 4.725, 4.238 and 4.100, respectively. The findings show that heterogeneous Earth-observation, geospatial and expert evidence can be converted into an interpretable downstream priority structure. The framework provides a rapid screening tool for monitoring, bridge and road protection, warning verification and evacuation planning when detailed hydrodynamic simulation is not yet available.