Effect of Construction Restrictions on Future Landslide Exposure Under Land-Use Simulation in Southwest China

Natural hazards significantly influence landscape patterns and ecological processes, yet their integration into land-use simulation remains limited. This study incorporates landslide susceptibility into the Patch-generating Land Use Simulation (PLUS) model to evaluate the effect of construction restrictions on future landslide exposure in Nanbu County, Sichuan Province, China. Landslide susceptibility was assessed using the certainty factor–logistic regression (CF-LR) model based on multi-period historical data, while future land-use changes (2020–2050) were projected with the PLUS model under two scenarios: a baseline scenario (without landslide constraints) and a susceptibility-constrained scenario (where high-susceptibility zones are restricted for construction). The results indicate that landslide-prone areas are spatially heterogeneous, with non-high-susceptibility zone concentrated in the west and northwest, and high-susceptibility zones clustered in the central area near the county seat. Under the baseline scenario, construction land within high-susceptibility zones increases by 107.47% over the 2020–2050 period. By contrast, the susceptibility-constrained scenario reduces construction land in these zones by 56.6% relative to the baseline 2050 projection, and by 10.01% relative to its own 2020 level. Ecosystem service value (ESV) assessment reveals that the two scenarios achieve nearly identical total ESV (with only a 0.09% difference), but the susceptibility-constrained scenario preserves 10.36 km2 of high-value ecosystems within high-susceptibility zones compared to the baseline scenario. These findings demonstrate that incorporating landslide susceptibility as a spatial constraint reduces future landslide exposure without compromising ecosystem service value, providing a practical, science-based framework for decision makers to inform land-use planning in mountainous regions prone to landslides.