Spatial variation in long-term natural recovery potential informs targeted post-fire restoration in a peri-urban forest

Abstract Post-fire recovery capacity can vary substantially within peri-urban forests, where heterogeneous environmental conditions complicate decisions about when natural regeneration is sufficient and where active intervention is warranted. We evaluated spatial variation in long-term natural recovery potential and the effects of targeted planting in a fire-affected peri-urban forest in Jinyun Mountain, Chongqing, China. Unmanned aerial vehicle (UAV) multispectral imagery, U-Net classification, and field observations were integrated to reconstruct post-fire vegetation conditions, which were used to initialize 100-year LANDIS-II simulations under natural succession (S1) and spatially targeted planting (S2). A Restoration Potential Index (RPI), derived exclusively from year-100 S1 outputs, integrated structural recovery, ecosystem productivity, and community development to identify spatial differences in modeled natural recovery potential. Recovery potential was strongly heterogeneous across the burned landscape, while bamboo showed sustained biomass expansion under natural succession. Targeted planting produced higher simulated aboveground carbon stocks and earlier biomass and aboveground net primary productivity (ANPP) breakpoints than natural succession, with estimated advances of approximately 9 and 10 years, respectively. However, scenario differences narrowed during later succession, and diversity responses were modest and spatially variable. These results indicate that post-fire restoration in heterogeneous peri-urban forests may benefit from spatially differentiated management: active planting can be concentrated in areas with relatively low natural recovery potential, whereas higher-recovery areas may rely more strongly on passive succession and continued monitoring.