Abstract: By using model combination method, this paper simulated the changes of response variable (tree species distribution area at landscape level under climate change) under three scenarios of environmental spatial heterogeneous level, analyzed the differentiation of simulated results under different scenarios, and discussed the effects of environmental spatial heterogeneity on the larger spatial extrapolation of the tree species responses to climate change observed in sampling plots. For most tree species, spatial heterogeneity had little effects on the extrapolation from plot scale to class scale; for the tree species insensitive to climate warming and the azonal species,spatial heterogeneity also had little effects on the extrapolation from plot-scale to zonal scale. By contrast, for the tree species sensitive to climate warming,spatial heterogeneity had effects on the extrapolation from plot scale to zonal scale, and the effects could be varied under different scenarios.

Key words: climate change, forest landscape prediction, spatial heterogeneity, spatial extrapolation, scale, LANDIS, sampling plot data