Projections of global road risk exposed to landslides under climate change

Climate change-driven rainfall extremes are increasingly threatening global road infrastructure through amplified landslide hazards. Here we quantified global landslide susceptibility by coupling seven machine learning models with ensemble projections from the Coupled Model Intercomparison Project Phase 6 under four shared socioeconomic pathways. Results indicate a general upward trend in global road risk, with the most severe rise (approximately 30.6%) occurring under the high-emission scenario by 2100. Currently, approximately 8.3% of the global land area of the road is at high and very high risk of landslides, mainly concentrated in eight hotspots, including the middle west of North America, eastern United States, western Europe, South and East Asia, and southeastern Asia. Across countries with different income levels, high-income countries’ road risk exceeds others by 2–5-fold. These projections highlight the urgent need for climate adaptation strategies in high-risk regions, particularly where aging infrastructure intersects with intensifying hazards in high-income countries. Under high-emission scenarios, a projection of ~30.6% greater global landslide risk to roads is likely by 2100, with wealthier countries facing disproportionately elevated risk, according to findings from coupled model intercomparison project phase 6 and machine learning models