Climate-adaptive prediction of wind erosion risks for earthen heritage under multi-scenario futures

• LiDAR–CFD–CMIP6 framework predicts climate-driven wind erosion. • Site-scale simulations map erosion hotspots for targeted protection. • Nonlinear thresholds reveal vulnerability under extreme wind events. • Multi-scenario projections show divergent erosion pathways to 2100. • Tools support climate-adaptive conservation and resilient planning. Earthen heritage sites are vital components of global cultural assets, yet face intensifying wind erosion driven by more frequent extreme wind events. Accurate erosion prediction is critical to preserve structural integrity and guide climate-adaptive management under UNESCO’s climate action framework. However, current methods rely on simplified geometry models or limited wind simulations, producing insufficient spatial resolution and overlooking differential mass loss among components. Temporally, projections based on linear extrapolation from typical years neglect long-term, non-stationary trends. This study developed an integrated prediction framework combining centimeter-scale LiDAR modeling, high-resolution computational fluid dynamics (CFD), and CMIP6 multi-scenario climate data. Applied to the representative Chinese earthen heritage site Xixia Imperial Tombs, the framework produces site-wide risk maps and identifies localized high-risk zones. Results quantify nonlinear erosion responses to wind speed gradients, extending conventional initiation thresholds with acceleration inflection points. Temporal resampling refines wind-speed resolution, producing annual projections of cumulative erosion metrics (mass, depth, volume) through 2100 under SSP2–4.5 and SSP5–8.5 scenarios. The intermediate SSP2–4.5 pathway shows higher interannual variability and greater mean cumulative erosion than the high-forcing SSP5–8.5 scenario, indicating elevated long-term degradation risk under moderate climate forcing. Both scenarios displayed wide uncertainty ranges, suggesting a substantial likelihood of extreme erosion outcomes beyond mean projections. By integrating high-precision simulation with scenario-sensitive climate data, this framework advances predictive modeling of climate-driven hazards in the built environment and supports resilient planning and conservation of earthen heritage in arid regions.