Abstract

Sustained warming is displacing montane species upslope, yet trees already at the upper limit of their mountains face a finite vertical refuge. This study assessed whether the world’s highest-elevation tree, Polylepis tarapacana, can remain within climatically suitable terrain as the climate warms across its South-Central Andean range (Peru, Bolivia, Chile, and Argentina), with field validation in Tacna, southern Peru. The aim was to quantify, under CMIP6 scenarios extending to 2100, how much suitable habitat the species retains and what fraction of any loss leaves no reachable higher-elevation refuge. A MaxEnt model, benchmarked against Random Forest, was calibrated on the full known range from GBIF occurrences and field plots using collinearity-filtered bioclimatic predictors and a sampling-bias-corrected background, projected as a multi-GCM ensemble under CMIP6 twenty-first-century scenarios, and validated externally against 148 field-mapped stand polygons (12,874 ha). A novel elevational-ceiling and “no-escape” framework quantified, for each scenario, the suitable area lost without any reachable higher-elevation refuge. The models discriminated suitable habitat well (external AUC =0.85, Boyce =0.98). Under the highest-emission pathway, suitable habitat in Tacna contracts from −13% (2030s) to −100% (2090s) and the highly suitable class collapses by 99.4%; across the partial-loss scenarios, the suitability centroid shifts upslope by up to +195 m. By 2081–2100 the no-escape fraction reaches 100% at every tested dispersal buffer—to date the first quantitative evidence of mountaintop-extinction risk for the world’s highest-elevation tree. These results argue for anticipatory ex situ conservation and Peru–Bolivia cross-border protection of the highest-elevation Altiplano refugia, before the vertical escape route closes.

Keywords