Aim Species in the tropics respond to global warming by altitudinal distribution shifts. Consequences for biodiversity may be severe, resulting in lowland attrition, range-shift gaps, range contractions and extinction risks. We aim to identify plant groups (growth forms, families, endemic status) with higher than average risks.
Location South Ethiopian highlands.
Methods Based on observational data from mainly unexplored and remote mountain regions, we applied a published model to project the consequences of an upward shift of thermal site conditions on the altitudinal distribution of 475 plant species. Annual average temperature increases of up to 5 °C were evaluated. Differences between groups of species were analysed by a permutation procedure and Generalized Linear Models.
Results Because of a limited regional species pool, even mild warming is projected to create strong potential risks concerning lowland attrition, i.e. the net loss of species richness because of upward range shifts in the absence of new species arriving. Likewise, many species are expected to face range-shift gaps, i.e. the absence of an overlap between future and current altitudinal ranges already under mild warming scenarios. Altitudinal contractions and mountain-top extinctions will potentially become important when warming exceeds 3.5 °C. Mean area per species is projected to decline by 55% for the A2 emissions scenario (+4.2 °C until 2100) because of the physical shape of the mountains. Higher than average vulnerability is expected for endemic species as well as for herbs and ferns. Plant families that are especially threatened are identified.
Main conclusions Lowland biotic attrition and range-shift gaps as predicted by a simple model driven by shifts of isotherms will result in novel challenges for preserving mountain biodiversity in the inner tropics. Whereas contractions of occupied area are expected to threaten endemic and already endangered species in particular, we suggest that conservation priorities can be identified based on simple prognostic models even without precise regional warming scenarios.