Global model of vegetation migration: incorporation of climatic variability

Climate change induced by increasing concentrations of greenhouse gases in the atmosphere is expected to transform global distribution of terrestrial vegetation. The analysis of global vegetation distributions determined by earth's climates is usually based on biogeographical classifications. These classifications assume that flora in any location is predetermined by certain parameters of climate, and that change in these parameters will produce instant changes in plants composition. Application of the models employing biogeographical classifications to different scenarios of climate change is generally limited with the hypothesis of 'dynamic equilibrium': the rate of climate change should be comparable with the rate of vegetation response. The estimated rate of hypothesized global warming is now considered to be fast enough to fall outside these limits. Therefore, an alternative assumption that the forests will be extirpated, but will not migrate to the new territory within the time limits of CO, doubling, can be considered a better approximation of future vegetation dynamics. We propose a model to study the effect of native variability of input data on model prediction, portraying vegetation migration as a stochastic process. We included into analysis only variability of climate and assumed that the only process affected by changing climate variability is vegetation mortality (i.e. regional extirpation). Our simulations show that the probabilistic approach can significantly change the forecasts of global vegetation redistribution. The predicted by the model extirpation area for boreal forests is significantly less, than the area predicted by the equilibrium model without forest in-migration. (C) 2000 Elsevier Science B.V. All rights reserved.