Multi-field Coupling Early Warning of Climate-induced Cascading Landslide Hazards: The Likan Landslides in the Upper Yellow River, China

• Three kinematic zones revealed via displacement-thickness-ecology synergy in the Likang landslides. • Evidence of 3–6-month delays in precipitation-driven motion and vegetation suppression. • Long-term creep causes road damage and surge disaster chains via feedback coupling. • Multi-source remote sensing/ground survey system enables early landslide warnings. Under a sustained warming climate, reservoir‐area landslides exhibit increasingly complex, non-linear interactions. The cascade “climate or seasonal variability-soil rheology-vegetation response/feedback-impulse-wave risk” constitutes an emerging safety threat in alpine gorge reservoirs; yet how such cascades propagate—and how their key links manifest—remains insufficiently resolved. Here, adopting a geomorphological–ecological, multi-feedback perspective, we develop a research paradigm that couples heterogeneous geological modeling with surface-ecosystem diagnostics and analyses of spatiotemporal lag effects, and apply it to an empirical study of the Likan Landslides in the Lijia Gorge Reservoir on the upper Yellow River. We show that the landslide undergoes a northward-deflecting motion as a quasi-rigid block, and we identify and quantitatively characterize three heterogeneous zones with distinct activity signatures: (1) Horizontally dominated displacement zone (0.7 km 2 ; mean thickness 10.6 m; maximum displacement rate 73 mm yr⁻ 1 ) showing thick-slope shear-creep behavior; (2) Downslope displacement-dominated zone (0.3 km 2 ; 7.9 m; 33 mm yr⁻ 1 ) characterized by thin-slope creep and brittle unloading; (3) Severe dual-motion zone (2.2 × 10⁻ 2 km 2 ; 7.03 m; 33–73 mm yr⁻ 1 ) manifesting block tensile-fracture movement. We further detect a 3–6-month lag between precipitation forcing and landslide response, coupled with a vegetation recovery suppression effect, confirming the system as a hydrologically lag-driven, progressive heterogeneous landslide. Its long-term creep–spatial differentiation–environmental feedback composite mechanism induces cumulative damage to the ground surface and poses the potential to trigger secondary wave surges. The study establishes a multi-field coupling framework of remote sensing that links process analysis with risk-warning decision-making, provides an empirical case for understanding landslide responses under climate change, and offers a transferable paradigm for cascading landslide risk assessment in reservoir regions.