The differences among the CO₂ concentrations projected with the various SRES scenarios considered are larger than the differences caused by inclusion or omission of climate-mediated feedbacks. The range of uptake rates projected by process-based models for any one scenario is, however, considerable, due to uncertainties about (especially) terrestrial ecosystem responses to high CO₂ concentrations, which have not yet been resolved experimentally, and uncertainties about the response of global NPP to changes in climate (Cramer et al., 1999). A smaller feedback would be implied if, as some models indicate, global NPP increases with warming throughout the relevant range of climates and no forest die back occurs. Larger positive feedbacks would be implied if regional drying caused partial die back of tropical forests, as some of the DGVMs in Cramer et al. (2001), and one coupled climate-carbon model study of Cox et al. (2000), suggest; however, another coupled climate-carbon model study (Friedlingstein et al., 2001) suggests a smaller feedback. Uncertainty also arises due to differences in the climate responses of ocean models, especially as regards the extent and effects (biological as well as physical) of increased stratification in a warmer climate (Joos et al., 1999b).

In conclusion, anthropogenic CO₂ emissions are virtually certain to be the dominant factor determining CO₂ concentrations throughout the 21st century. The importance of anthropogenic emissions is underlined by the expectation that the proportion of emissions taken up by both ocean and land will decline at high atmospheric CO₂ concentrations (even if absolute uptake by the ocean continues to rise). There is considerable uncertainty in projections of future CO₂ concentration, because of uncertainty about the effects of climate change on the processes determining ocean and land uptake of CO₂. These uncertainties do not negate the main finding that anthropogenic emissions will be the main control.

Large-scale manipulations of terrestrial ecosystems have been proposed as a means of slowing the increase of atmospheric CO₂ during the 21st century in support of the aims of the Kyoto Protocol (Tans and Wallace, 1999; IPCC, 2000a). Based on current understanding of land use in the carbon cycle, the impacts of future land use on terrestrial biosphere-atmosphere exchanges have the potential to modify atmospheric CO₂ concentrations on this time-scale. Direct effects of land-use changes are thought to represent about 10 to 30% of total anthropogenic CO₂ emissions (Table 3.1), so there is scope for either intended or unintended changes in land use to reduce or increase total anthropogenic emissions. But the possibilities for enhancing natural sinks have to be placed in perspective: a rough upper bound for the reduction in CO₂ concentration that could be achieved by enhancing terrestrial carbon uptake through land-use change over the coming century is 40 to 70 ppm (Section 3.2.2.2), to be considered against a two to four times larger potential for increasing CO₂ concentraion by deforestation, and a >400 ppm range among the SRES scenarios (Figure 3.12).

IPCC Homepage