Climate Change 2001:
Working Group I: The Scientific Basis
Other reports in this collection

3.3.3 Variations in Atmospheric CO2 during the Past 11,000 Years

Natural variations in CO2 during the past 11,000 years (Figure 3.2c) have been small (about 20 ppm) according to the best available measurements, which are from the Taylor Dome ice core (Smith et al., 1999; Indermühle et al., 1999). These measurements show a short-lived maximum around 11 kyr BP, followed by a slight fall, which may have been caused by increasing carbon storage in the terrestrial biosphere. Atmospheric CO2 concentration was about 260 ppm at its Holocene minimum around 8 kyr BP and increased towards about 280 ppm in the pre-industrial period. The same pattern and the same CO2 concentration levels over the past 8 kyr have also been shown in another ice core, BH7 near Vostok (Peybernès et al., 2000). The causes of these changes are not known. Preliminary 13C measurements (see Box 3.6) suggest that this increase may have been due to a gradual reduction in terrestrial carbon storage (Indermühle et al., 1999; Smith et al., 1999) but others have considered an oceanic explanation more likely.

Atmospheric CO2 concentrations have also been reconstructed indirectly, from stomatal index measurements on sub-fossil leaves (Van de Water et al., 1994; Beerling et al., 1995; Rundgren and Beerling, 1999; Wagner et al., 1999). Stomatal density and stomatal index of many species respond to atmospheric CO2 (Woodward, 1987; Woodward and Bazzaz, 1988) but are influenced by other environmental variables as well (Poole et al., 1996). One recent stomatal index record, interpreted as implying high (up to 350 ppm) and rapidly fluctuating CO2 concentrations in the early Holocene (Wagner et al., 1999), is clearly incompatible with the ice core record of Indermühle et al. (1999), whereas a continuous stomatal index record from 9 kyr BP onwards (Rundgren and Beerling, 1999) has shown concentration trends consistent with the ice-core records.

Figure 3.2b shows the excellent agreement among different high-resolution Antarctic ice cores covering the past 1,000 years. Atmospheric CO2 concentration fell by about 8 to 10 ppm during the Little Ice Age (from 1280 to 1860, see Chapter 2) (Figure. 3.2b, c; Barnola et al., 1995; Etheridge et al., 1996; Indermühle et al., 1999; Rundgren and Beerling, 1999). A slight contemporaneous increase in 13C of atmospheric CO2 has led to the suggestion that this effect was caused by enhanced carbon storage on land (Francey et al., 1999b; Trudinger et al., 1999).

Figure 3.4: Partitioning of fossil fuel CO2 uptake using O2 measurements (Keeling and Shertz, 1992; Keeling et al., 1993; Battle et al., 1996, 2000; Bender et al., 1996; Keeling et al., 1996b; Manning, 2001). The graph shows the relationship between changes in CO2 (horizontal axis) and O2 (vertical axis). Observations of annual mean concentrations of O2, centred on January 1, are shown from the average of the Alert and La Jolla monitoring stations (Keeling et al., 1996b; Manning, 2001; solid circles) and from the average of the Cape Grim and Point Barrow monitoring stations (Battle et al., 2000; solid triangles). The records from the two laboratories, which use different reference standards, have been shifted to optimally match during the mutually overlapping period. The CO2 observations represent global averages compiled from the stations of the NOAA network (Conway et al., 1994) with the methods of Tans et al. (1989). The arrow labelled “fossil fuel burning” denotes the effect of the combustion of fossil fuels (Marland et al., 2000; British Petroleum, 2000) based on the relatively well known O2:CO2 stoichiometric relation of the different fuel types (Keeling, 1988). Uptake by land and ocean is constrained by the known O2:CO2 stoichiometric ratio of these processes, defining the slopes of the respective arrows. A small correction is made for differential outgassing of O2 and N2 with the increased temperature of the ocean as estimated by Levitus et al. (2000).

3.3.4 Implications

The Vostok record of atmospheric CO2 and Antarctic climate is consistent with a view of the climate system in which CO2 concentration changes amplify orbitally-induced climate changes on glacial/inter-glacial time-scales (Shackleton, 2000). Changes during the present inter-glacial (until the start of the anthropogenic CO2 rise) have been small by comparison. Although complete explanations for these changes in the past are lacking, high-resolution ice core records establish that the human-induced increase of atmospheric CO2 over the past century is at least an order of magnitude faster than has occurred during the preceeding 20,000 years.

Box 3.6: Stable carbon isotopes in atmospheric CO2.

13C, a measure of the relative abundance of the two stable carbon isotopes, 13C and 12C, in atmospheric CO2 gives in principle similar possibilities to O2 for the partitioning of atmospheric CO2 uptake (Keeling et al., 1979, 1980; Mook et al., 1983; Keeling et al., 1989; Francey et al., 1995; Keeling et al., 1995). The principle of using d13C to separate land and ocean components of the carbon budget relies on the fractionation during photosynthesis by C3 plants, which discriminates against 13C. This fractionation leads to biospheric carbon being depleted in 13C by about 18‰ relative to the atmosphere. In contrast, exchanges with the ocean involve relatively small fractionation effects. Changes in the 13C/12C ratio of atmospheric CO2 thus indicate the extent to which concurrent CO2 variations can be ascribed to variations in biospheric uptake. The calculation also requires specification of the turnover times of carbon in the ocean and on land, because fossil fuel burning implies a continuous release of isotopically light carbon to the atmosphere. This leads to a lowering of the atmospheric 13C/12C isotope ratio, which takes years to centuries to work its way through the carbon cycle (Keeling et al., 1980; Tans et al., 1993; Ciais et al., 1995a,b).

There are some complications. C3 plants discriminate against 13C more strongly than C4 plants (Lloyd and Farquhar, 1994), thus the distributions of C3 and C4 photosynthesis need to be known. The oceanic disequilibrium can in principle be estimated observationally (Tans et al., 1993; Heimannn and Maier-Reimer, 1996; Bacastow et al., 1996; Gruber et al., 1999), while the terrestrial disequilibrium has to be estimated by means of models (e.g., Ciais et al., 1999). Langenfelds et al. (1999) and Battle et al. (2000) have shown that recently estimated values for the disequilibrium terms lead to consistency between the partitioning of CO2 uptake into land and ocean uptake based on O2 and on d13C measurements.

Other reports in this collection