Predicting generalized ecosystem groups with the NCAR CCM: first steps towards an interactive biosphere

Canopy-plus-soil "big-leaf" models of the land surface now exist and are being incorporated in global climate models. These big-leaf models could he the basis for the incorporation of an interactive land biosphere into global models. However, their use depends upon satisfactory specification of the distribution of plants (and soils), and while such distributions can be obtained for the present-day, they must be manufactured for all other climatic scenarios. Thus, one prerequisite for the incorporation of an interactive land biosphere into global climate models is the successful prediction of the natural vegetation (or "climax" vegetation) for climatic scenarios such as the doubling of atmospheric carbon dioxide. (It is also necessary to predict agricultural land-use classes, but that is outside the scope of this paper.) A highly generalized (nine classes) grouping of Holdridge life zones has beer, used here as a means of investigating three-step "coupling" of a land-surface scheme into a global climate model. The investigation includes the derivation of nine "natural" ecosystems for the present day climate using temperature and precipitation derived from three experiments undertaken with the NCAR community climate model. These predictions differ from one another and both differ significantly from the prescribed classification groupings of ecosystem complexes used with one big-leaf, land-surface scheme--the Biosphere Atmosphere Transfer Scheme (BATS). On the other hand, these highly generalized groupings show relatively little sensitivity to the temperature changes induced by doubling atmospheric CO2 and even the inclusion of the precipitation disturbances in the doubled-CO2 scenario causes changes in the ecosystem "predictions" that are only of similar degree, or smaller, than the differences between sets of life-zone classes generated from present-day or doubled-CO2 climates. This result implies that the combined global field of annual biotemperature and precipitation" is not vet consistently predicted by GCMs, not that vegetation is likely to be insensitive to climate change.