A wide range of forests distributed across steep environmental gradients are found in Yunnan, southwest China. Climate change could profoundly change these forests by affecting species ranges. We produce predictions about species suitable habitat shifts and use these to (1) evaluate species range size change, loss and turn-over under no- and full-dispersal and nine climate change scenarios and (2) identify environmental variables responsible for current species richness and future local species losses.

　　Yunnan Province, Southwest China.

　　Using MaxEnt, we modelled current distributions of 2319 woody plant species, corrected for collecting bias and found that 1996 had significant spatial association with environmental factors. Using three General Circulation Models (GCMs: CGCM, CSIRO and HADCM3) for the years 2070–2099 (2080s), based on three emission scenarios for each GCM (A1b, A2a and B2a), we predicted the future geographic position of suitable habitat for each species.

　　Although most species were predicted to persist within Yunnan, with a maximum extinction rate of c. 6% under the most extreme climate change scenario, up to 1400 species (of the 1996 tested) are expected to lose more than 30% of their current range under the most extreme climate change scenario. Assuming no- or unlimited dispersal minimally affected these outcomes. Species losses were associated with increasing temperature variability and declining precipitation during the dry season.

　　To conserve Yunnan's woody flora, management efforts should focus on providing elevational migration routes at local scales, with priority for those areas located within previously identified conservation hotspots. As almost all species show range contractions, storage of genetic diversity in seed banks and botanical gardens would be sensible. A change in Yunnan's conservation policy will be needed to counter the predicted negative impacts of climate change on its flora.