The carbon balance of the terrestrial biosphere: Ecosystem models and atmospheric observations

Precise measurements in air are helping to clarify the fate of CO₂ released by human activities. Oxygen-to-nitrogen ratios in firn (the transition state from snow to ice) and archived air samples indicate that the terrestrial biosphere was approximately carbon-neutral on average during the 1980s. CO₂ release by forest clearance during this period must have been compensated for by CO₂ sinks elsewhere on land. Direct atmospheric O₂:N₂ measurements became available during the 1990s. These measurements indicate net terrestrial CO₂ uptake of ~2 Pg C/yr. From the north-south O₂:N₂ gradient, it has been inferred that about this amount was taken up by terrestrial ecosystems in the northern nontropics while additional CO₂ released by tropical-forest clearance must have been compensated for by additional, tropical, terrestrial CO₂ sinks. These and other atmospheric observations provide independent tests of carbon-cycle reconstructions made with process-based terrestrial ecosystem models. Such models can account for major features of the atmospheric-CO₂ record, including the amplitude and phase of the seasonal cycle of atmospheric-CO₂ concentration at different latitudes, and much of the interannual variability in the rate of increase of atmospheric CO₂. Models also predict direct effects of rising atmospheric-CO₂ concentration on primary production, modified by feedbacks at the plant and ecosystem levels. These effects translate into a global carbon sink the right order of magnitude to compensate for forest clearance during the 1980s. The modeled sink depends on continuously increasing CO₂ to maintain disequilibrium between primary production and carbon storage. There are still substantial differences among the carbon-balance estimates made by different models, reflecting limitations in current understanding of ecosystem-level responses to atmospheric-CO₂ concentration, especially with regard to the interactions of C and N cycling and interactions with land-use change. Scenario calculations nevertheless agree that if atmospheric CO₂ continues its rise unchecked then the terrestrial sink will start to decline by the middle of the next century, for reasons including saturation of the direct CO₂ effect on photosynthesis.