Global satellite analysis reveals a peak in rainfall-triggered landslide susceptibility at intermediate vegetation density, challenging assumptions that denser plant cover always stabilizes slopes.
Landslide hazard models worldwide assume that vegetation stabilizes slopes monotonically, so that denser vegetation always implies greater stability. Analyzing 12,459 rainfall-triggered landslides (2000–2024) with multi-source satellite data, we find a "vegetation paradox": landslide susceptibility does not decline monotonically with vegetation density but peaks within a critical transition zone (CTZ) at NDVI, 0.769–0.868—a unimodal, non-monotonic pattern. This zone concentrates 23.0% of the full 18,028-event inventory (33.2% of the 12,459 analyzed with reliable NDVI) and persists across biomes, climate zones, and slope classes (16.7%–44.5% CTZ prevalence across the gentle-to-steep range), and its upper boundary is stable across data-quality filtering, density smoothing, and simulated sensor noise. An independent grid-based Random Forest ( R 2 = 0.446) independently places this maximum within the CTZ (mean NDVI, 0.860) without histogram binning; its coarse grid, however, captures only the rising limb and cannot resolve the decline above the CTZ.
No takes yet. Share an insight, caveat, or question.
Yuan et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: