Simulations of potential vegetation distribution, natural fire frequency, carbon pools, and fluxes are presented for two DGVMs (Dynamic Global Vegetation Models) from the second phase of the Vegetation/Ecosystem Modeling and Analysis Project. Results link vegetation dynamics to biogeochemical cycling for the conterminous United States. Two climate change scenarios were used: a moderately warm scenario from the Hadley Climate Centre and a warmer scenario from the Canadian Climate Center. Both include sulfate aerosols and assume a gradual CO2 increase. Both DGVMs simulate a reduction of southwestern desert areas, a westward expansion of eastern deciduous forests, and the expansion of forests in the western part of the Pacific Northwest and in north-central California. Both DGVMs predict an increase in total biomass burnt in the next century, with a more pronounced increase under the Canadian scenario. Under the Hadley scenario, both DGVMs simulate increases in total carbon stocks. Under the Canadian scenario, both DGVMs simulate a decrease in live vegetation carbon. We identify similarities in model behavior due to the climate forcing and explain differences by the different structure of the models and their different sensitivity to CO2. We compare model output with data to enhance our confidence in their ability to simulate potential vegetation distribution and ecosystem processes. We compare changes in the area of drought-induced decreases in vegetation density with a spatial index derived from the Palmer Drought Severity Index to illustrate the ability of the vegetation to cope with water limitations in the future and the role of the CO2 fertilization effect.