Figure 3.2: Variations in atmospheric CO2 concentration on different time-scales. (a) Direct measurements of atmospheric CO2 concentration (Keeling and Whorf, 2000), and O2 from 1990 onwards (Battle et al., 2000). O2 concentration is expressed as the change from an arbitrary standard. (b) CO2 concentration in Antarctic ice cores for the past millenium (Siegenthaler et al., 1988; Neftel et al., 1994; Barnola et al., 1995; Etheridge et al., 1996). Recent atmospheric measurements at Mauna Loa (Keeling and Whorf, 2000) are shown for comparison. (c) CO2 concentration in the Taylor Dome Antarctic ice core (Indermühle et al., 1999). (d) CO2 concentration in the Vostok Antarctic ice core (Petit et al., 1999; Fischer et al., 1999). (e) Geochemically inferred CO2 concentrations, from Pagani et al. (1999a) and Pearson and Palmer (2000). (f) Geochemically inferred CO2 concentrations: coloured bars represent different published studies cited by Berner (1997). The data from Pearson and Palmer (2000) are shown by a black line. (BP = before present.)
Despite the importance of biological processes for the ocean's natural carbon cycle, current thinking maintains that the oceanic uptake of anthropogenic CO2 is primarily a physically and chemically controlled process superimposed on a biologically driven carbon cycle that is close to steady state. This differs from the situation on land because of the different factors which control marine and terrestrial primary productivity. On land, experiments have repeatedly shown that current CO2 concentrations are limiting to plant growth (Section 18.104.22.168). In the ocean, experimental evidence is against control of productivity by CO2 concentrations, except for certain species at lower than contemporary CO2 concentrations (Riebesell et al., 1993; Falkowski, 1994). Further, deep ocean concentrations of major nutrients and DIC are tightly correlated, with the existing ratios closely (but not exactly, see Section 22.214.171.124) matching the nutritional requirements of marine organisms (the "Redfield ratios": Redfield et al., 1963). This implies that as long as nutrients that are mixed into the ocean surface layer are largely removed by organic carbon production and export, then there is little potential to drive a net air-sea carbon transfer simply through alteration of the global rate of production. Terrestrial ecosystems show greater variability in this respect because land plants have multiple ways to acquire nutrients, and have greater plasticity in their chemical composition (Melillo and Gosz, 1983). There are, however, extensive regions of the ocean surface where major nutrients are not fully depleted, and changes in these regions may play a significant role in altering atmosphere-ocean carbon partitioning (see Section 126.96.36.199).
The increase of atmospheric pCO2 over pre-industrial levels has tended to increase uptake into natural CO2 sink regions and decreased release from natural outgassing regions. Contemporary net air-sea fluxes comprise spatially-varying mixtures of natural and anthropogenic CO2 flux components and cannot be equated with anthropogenic CO2 uptake, except on a global scale. Uptake of anthropogenic CO2 is strongest in regions where "old" waters, which have spent many years in the ocean interior since their last contact with the atmosphere, are re-exposed at the sea surface to a contemporary atmosphere which now contains anthropogenic CO2 (e.g., Sarmiento et al., 1992; Doney, 1999). In an upwelling region, for example, the natural component of the air-sea flux may be to outgas CO2 to the atmosphere. The higher atmospheric pCO2 of the contemporary atmosphere acts to reduce this outgassing relative to the natural state, implying that more carbon remains in the ocean. This represents uptake of anthropogenic CO2 by a region which is a source of CO2 to the atmosphere. The additional carbon in the ocean resulting from such uptake is then transported by the surface ocean circulation, and eventually stored as surface waters sink, or are mixed, into the deep ocean interior. Whereas upwelling into the surface layer is quantitatively balanced on a global scale by sinking, the locations where deep waters rise and sink can be separated by large horizontal distances.
Air-sea gas transfer allows older waters to approach a new steady state with higher atmospheric CO2 levels after about a year at the sea surface. This is fast relative to the rate of ocean mixing, implying that anthropogenic CO2 uptake is limited by the rate at which "older" waters are mixed towards the air-sea interface. The rate of exposure of older, deeper waters is therefore a critical factor limiting the uptake of anthropogenic CO2. In principle, there is sufficient uptake capacity (see Box 3.3) in the ocean to incorporate 70 to 80% of anthropogenic CO2 emissions to the atmosphere, even when total emissions of up to 4,500 PgC are considered (Archer et al., 1997). The finite rate of ocean mixing, however, means that it takes several hundred years to access this capacity (Maier-Reimer and Hasselmann, 1987; Enting et al., 1994; Archer et al., 1997). Chemical neutralisation of added CO2 through reaction with CaCO3 contained in deep ocean sediments could potentially absorb a further 9 to 15% of the total emitted amount, reducing the airborne fraction of cumulative emissions by about a factor of 2; however the response time of deep ocean sediments is in the order of 5,000 years (Archer et al., 1997).
Using time-series and global survey data, the increasing oceanic carbon content has been directly observed, although the signal is small compared to natural variability and requires extremely accurate measurements (Sabine et al., 1997). A long-term increase of surface water CO2 levels tracking the mean atmospheric CO2 increase has been observed in the ocean's subtropical gyres (Bates et al., 1996; Winn et al., 1998) and the equatorial Pacific (Feely et al., 1999b). However, very few such time-series exist and the response of other important oceanic regions to the atmospheric pCO2 increase cannot yet be assessed. Inter-decadal increases in DIC concentrations at depth have been resolved from direct measurements (Wallace, 1995; Peng et al., 1998; Ono et al., 1998; Sabine et al., 1999). The total amounts of anthropogenic CO2 accumulated in the ocean since the pre-industrial era can also be estimated from measurements using recent refinements (Gruber et al., 1996) of long-standing methods for separating the natural and anthropogenic components of oceanic DIC. A comparison of such analyses with ocean model results is presented in Section 3.6.3.
Other reports in this collection