Variability of global biome patterns as a function of initial and boundary conditions in a climate model

The use of one-way coupling of an equilibrium-response vegetation, or biome, model with atmospheric circulation models is critically assessed. Global biome patterns from various, equally likely numerical realisations of present-day climate are compared. It has been found that the changes in global biome patterns to be expected from interdecadal variability in the atmosphere affect 9-12% of the continental surface (Antarctica excluded). There is no unique difference pattern, although changes are mainly induced by the variability of annual moisture availability and, to a lesser extent, by winter temperatures. This variability of biome patterns reflects the uncertainty in the estimate of equilibrium vegetation patterns from finite time interval climatologies. Changes in biome distributions between present-day climate and anomaly climate, the latter induced by an increase in sea-surface temperatures and atmospheric CO2, are larger than and different in kind from the changes due to inderdecadal variability. Roughly 30% of the land surface is affected by these changes. It appears that the strongest and most significant signal is seen for boreal biomes which can be attributed to an increase in near surface temperatures.