Predicting Responses to Climatic Change of the Potential Vegetation Distribution in China.

Abstract

The potential vegetation distribution shift in China caused by global climatic change was predicted by the direct transfer function approach. Two types of statistical model, the discriminant analysis model and the multinomial logit model, were applied to explain the relationship between vegetation classification and climatic conditions. The logit model resulted in a more successful result than the discriminant analysis model. Thus, the effects of global climatic change were estimated using the logit model under two climatic change scenarios proposed by Robock et al. (1993): Scenario A, a 2-degree increase of mean annual temperature and a 20% increase of annual precipitation; Scenario B, a 4-degree increase of mean annual temperature and a 20% increase of annual precipitation. It was predicted that (1) the conifer forest distributed in northeast China at the present time may be deprived of its suitable habitat within the boundary of China, (2) the broadleaf forest distributed in east China may shift northward by around 3 degrees of latitude for Scenario A and 5 degrees for Scenario B, and (3) the desert region in west China may expand and steppe and savanna may decrease, especially in the case of Scenario B.