Predicting potential distributions of dominant species and assessing future afforestation risks in Guangdong...

Climate change is reshaping species distributions across karst landscapes, yet large-scale ecological restoration in southern China remains hampered by species and site mismatch, rooted in inadequate macro-scale assessment of climate-vegetation equilibrium. Existing species distribution model (SDM) studies in karst regions are dominated by single-species, locally calibrated analyses. Here we assembled occurrence records spanning the full Chinese range of 12 dominant tree species (1,376 records) and combined them with bioclimatic, soil, and topographic predictors. We built optimized MaxEnt models and projected suitable habitat for the northern Guangdong limestone region under the SSP1-2.6, SSP2-4.5, and SSP5-8.5 scenarios for 2061-2080. We then integrated these projections into a bivariate spatial framework that couples current species richness with net richness change (ΔS) to delineate climate trajectory zones. Models performed well (AUC 0.907-0.984; TSS 0.749-0.933). Annual precipitation (BIO12), precipitation of the driest month (BIO14), and mean temperature of the warmest quarter (BIO10) were the most pervasive drivers across the assemblage. Total suitable area changed by less than 10% across all scenarios. Highly suitable habitat, however, often contracted even where total area held steady, identifying habitat-quality erosion rather than extent loss as the principal threat. Under SSP1-2.6 it shrank for 11 of 12 species, most in Toona sinensis (-88%) and Celtis sinensis (-85%), while Sinosideroxylon wightianum alone gained (up to +208%). Centroids of highly suitable habitat migrated 3.3-90.9 km in varying directions, most erratically under SSP1-2.6. Counterintuitively, the SSP1-2.6 scenario produced the most severe degradation. Its high-risk degradation zone covered 85.2% of the region, against 51.0% under SSP2-4.5 and 54.0% under SSP5-8.5. Critically, areas of highest current species richness coincided spatially with the largest climate-driven degradation risk. Within this high-richness core, the climatically stable zones of the limestone mountains of Yingde City and Yangshan County emerged as the most resilient afforestation targets. These findings provide a spatially explicit, climate-informed framework for tree species selection and afforestation site prioritization in the karst regions of southern China, directly supporting regional ecological restoration planning and long-term forest management under climate change.