Special Paper: A Global Vegetation Model Based on the Climatological Approach of Budyko

Abstract

A global vegetation model based on the climato- logical approach of Budyko is developed. The major vege- tation zones of the world are predicted by a two- dimensional ordination of a Dryness Index and Potential Evaporation, which is derived from radiation balance. Mean temperature of the warmest month is also used to separate the Ice/Polar Desert, Tundra, and Taiga zones. Pre- dictions of vegetation distributions are made using a global climate database interpolated to a 0.5? by 0.5? terrestrial grid. The overall impression from examining the resulting global vegetation map is that the modified Budyko model predicts the location and distribution of the world's vegeta- tion fairly well. Comparison between model predictions and Olson's actual vegetation map were based on Kappa statistics and indicate good agreement for Ice/Polar Desert, Tundra, Taiga, and Desert (even though we predict too much Desert). Agreement with Olson's map was fair for predicting the specific location of Tropical Rain Forest and Tropical Savannas, and was good for predicting their general location at a larger scale. Agreement between Olson's map and model predictions were poor for Steppe, Temperate Forest, Tropical Seasonal Forest, and Xerophy- tic Shrubs, although the predictions for Temperate Forest and Tropical Seasonal Forest improved to fair at a larger scale for judging agreement. Agreement with the baseline map of Olson was poor for Steppe and Xerophytic Shrubs at all scales of comparison. Based on Kappa statistics, overall agreement between model predictions and Olson's map is between fair and good, depending on the scale of comparison. The model performed well in comparison to other global vegetation models. Apparently the calculation of radiation balance and the resulting Dryness Index and Potential Evaporation provides important information for predicting the distribution of the major vegetation zones of the world.