Human activities, particularly burning fossil fuel, have increased atmospheric carbon dioxide (CO2) concentrations. Because CO2 traps heat, continued emissions are expected to change global climate. The extent of this change will depend not only on the rate of emissions, but also on carbon uptake by the oceans and the land.
According to some models, land ecosystems can sequester carbon fast enough to help to counteract CO2 emissions. Models featured in the Third Assessment Report of the Intergovernmental Panel on Climate Change (IPCC) suggest that increasing atmospheric CO2 alone could cause 350 to 890 Pg of carbon (1 Pg = 1015 g) to accumulate in the terrestrial biosphere by 2100. These amounts are equivalent to 22 to 57% of expected anthropogenic CO2 emissions in an intermediate emissions scenario (1, 2). The models suggest that atmospheric CO2 and climate change together could cause 260 to 530 Pg of carbon to accumulate, or 16 to 34% of emissions (1, 2).
These models probably exaggerate the terrestrial biosphere's potential to slow atmospheric CO2 rise. Ecosystem carbon accumulation may be constrained by nutrients, particularly nitrogen (3, 4), through mechanisms that are not well developed in or absent from the models.
How much nitrogen do the model projections require? The models distribute the future terrestrial carbon sink roughly equally between trees and soils. With no change in the carbon:nitrogen (C:N) ratios of trees (200) and soils (15), the CO2-only projections require 7.7 to 37.5 Pg of nitrogen; the CO2-climate projections require 2.3 to 16.9 Pg of nitrogen (see the figure) (5).
(Left) Nitrogen required to support terrestrial carbon uptake (1), compared to likely limits of nitrogen supply (green). For each model (2), values are shown for CO2-only (blue) and CO2-climate (red) projections. The upper nitrogen requirement assumes a fixed tree C:N of 200; the lower value assumes that all new tree carbon is allocated to wood. (Right) Discrepancy between nitrogen required for projected carbon uptake and likely nitrogen availability for CO2-only (blue) and CO2-and-climate-change (red) scenarios. Upper value: maximum calculated nitrogen required minus low nitrogen supply limit. Lower value: minimum nitrogen required minus high nitrogen supply limit.
Can increasing ecosystem C:N ratios reduce the nitrogen required? Tree C:N increases with atmospheric CO2 concentration (6, 7). But even allowing all the simulated increase in tree carbon to occur as wood (C:N = 500) only slightly reduces the amount of additional nitrogen required (see the figure). Soil C:N could also increase with rising atmospheric CO2 concentration, allowing soil carbon accumulation without additional nitrogen. This mechanism could allow some nitrogen transfer from soil to trees (6, 7), lowering the nitrogen demand associated with increased tree carbon. However, experimental studies show that when CO2 enrichment increases soil C:N, decomposing microorganisms require more nitrogen. This effect can reduce nitrogen mineralization, the main source of nitrogen for plants (8, 9). It is thus unlikely that increases in soil C:N could yield large increases in ecosystem carbon stocks.
With little contribution from increasing C:N, the carbon-uptake projections (1, 2) almost certainly require nitrogen accumulation. Nitrogen enters the terrestrial biosphere through atmospheric deposition and biological fixation, and is mainly lost through leaching and gaseous fluxes. We have estimated high and low nitrogen fluxes for each of these mechanisms (10).
To estimate future anthropogenic nitrogen deposition based on population-growth projections (11), we assume that per capita nitrogen deposition remains constant (low) or increases linearly to that of North America today (high) (12). We assume that 5% (low) to 10% (high) of that deposited nitrogen supports increased carbon storage (9). We estimate biological nitrogen fixation (12) to increase linearly by 10% (low) or 45% (high) with CO2 doubling (9). We further assume that nitrogen leaching losses are currently 36 Tg of nitrogen per year (13), and that nitrogen leaching would decline linearly with CO2 doubling by 0 (low) to 20% (high) (9).
Combining our high estimates, 6.1 Pg of nitrogen could accumulate by 2100 (see the figure). This amount is less than is required by all CO2-only simulations and by four of the six CO2-climate simulations (1, 2) (see the figure). Our low estimates of nitrogen accumulation yield only 1.2 Pg of nitrogen, insufficient for all simulations.
We have focused on nitrogen, but the situation may be worse for other nutrients, such as potassium and phosphorus, which are less subject to human or biological control than is nitrogen fixation. Models that incorporate nutrient cycling predict much less CO2 carbon uptake than models lacking these feedbacks (14). The next IPCC assessment must include models taking into account these feedbacks.
