Atmospheric CO2 Concentrations over the Last Glacial Termination Eric Monnin, Andreas Indermühle, André Dällenbach, Jacqueline Flückiger, Bernhard Stauffer, Thomas F. Stocker, Dominique Raynaud, and Jean-Marc Barnola

Supplementary Material

For CO₂ measurements, each sample (of size 2.5 cm by 2.5 cm by 1.5 cm) is cracked in an evacuated and cooled (-30°C) needle cracker. The air from opened bubbles expands into a laser absorption cell. The CO₂ concentration of the gas in the absorption cell is measured by tuning an infrared laser six times over the absorption line of a vibration-rotation transition of the CO₂ molecule. Reference gas from the Scripps Institution of Oceanography at 251.7 ppmv is used for the calibration of each measurement. Control measurements are routinely (after about five measurements) performed by using other reference gases from the Scripps Institution of Oceanography at 321.06 and 342.03 ppmv to check the linearity of the system. The linearity of the measurement technique for CO₂ concentrations lower than the calibration reference gas can be checked indirectly by generating control measurements at lower pressure, generating similar absorption line depths than for measurements with lower CO₂ concentrations. Measurements on bubble-free single-crystal ice samples, to which reference gas is added and which are cracked afterward, yield a reproducibility of ±1.5 ppmv.

For the methane measurements, an ice sample (~40 g) is melted under vacuum and then refrozen from the bottom with an alcohol bath at -25°C. The methane concentration on the gas is measured with a gas chromatograph. The external precision from replicate and blank measurements is 10 ppmv (1s). For additional information see (1).

Possible Artifacts Produced in the Ice

Most chemical impurities in the ice show large variations over short depth intervals corresponding to one or several annual layers [annual layer thickness at Dome Concordia (Dome C) is ~10 to 20 mm in the considered depth interval]. If there is a production of CO₂ in the ice matrix, we expect a scatter of the samples in a depth interval that cannot be explained by the analytical uncertainty (±1.5 ppmv). A c² test yields agreement of the standard deviations with the analytical uncertainty. A detailed analysis of the Ca²⁺ record [continuous flow analysis with a resolution of 10 mm (2)], which is a qualitative indicator of the carbonate concentration, shows variations between 10 to 30 parts per billion per weight (ppbw). These variations correspond to a CO₂ production potential of 60 to 180 ppmv (3). Our CO₂ measurements performed with a similar depth resolution do not show variations that are correlated significantly with the Ca²⁺ variations. On the other hand, the concentration of H₂O₂, a likely oxidant, is below the detection limit (<1 ppbw) in the Dome C ice core, and therefore, no oxidation of organic matter by H₂O₂ is expected. Results from the Vostok (4) and Taylor Dome (5) ice cores are in agreement with this record within the analytical uncertainties (the time resolution of these records allows only to confirm the gross features). CO₂ measurements from the Byrd ice core (6-8) are about 5 to 10 ppmv higher, especially on samples older than 18 ka. The reason for these deviations is still unclear. Apart from these systematic deviations, the general trend of the Byrd values is in agreement with the new measurements from the Dome C ice core.

Assumptions Made on the Link Between Deuterium Content in the Ice and Southern Ocean Temperature

Strong climatic connections could exist between Antarctica and the Southern Ocean (SO), as the SO could export latent and sensible heat to Antarctica. Similarities between temperature reconstructions in Vostok and reconstructions of the SO sea surface temperature (SST) by deep sea sediment core records (9, 10) and fossil Antarctic diatoms (11) were found at least for the long-term features. The Dome C temperature reconstruction based on stable isotopes measurements is very similar to the record of Vostok, so it can be expected that the Dome C temperature is an indicator for the SO SST. It is not clear without ambiguity if structures on shorter time scales such as the Antarctic Cold Reversal (ACR) were also a SO SST signal. Some SST temperature reconstructions on deep sea sediment core records (9) and fossil Antarctic diatoms (11) show a distinct ACR, but on other cores (10, 11), the ACR cannot be recognized. Short-lived peaks in Antarctic isotope records could, in principle, be due to temporary coolings of the evaporation sites of the deposits associated with cold conditions in the Northern Hemisphere (12).

Procedure for Determining the Correlation Coefficient Between the CO₂ and dD Records

In order to compare the CO₂ and temperature records, we carried out the following procedure. We calculated the correlation coefficient of the CO₂ and dD records using the mean values of 11 dD data points (corresponding to a time interval of 200 to 500 years) to every corresponding CO₂ data point. This procedure was chosen in order to take into account the fact that the enclosure process of air into ice causes a natural averaging of the CO₂ record corresponding to ~10% of the gas-ice age difference and, therefore, a single CO₂ value represents a time interval of 200 to 500 years. In order to estimate possible leads or lags, we shifted the CO₂ record in steps of 10 years and calculated the correlation coefficient between CO₂ and dD after each step.

References

1. J. Chappellaz et al., J. Geophys. Res. 102, 15987 (1997).
2. R. Röthlisberger, M. A. Hutterli, S. Sommer, E. W. Wolff, R. Mulvaney, J. Geophys. Res. 105, 20565 (2000).
3. J. Tschumi, B. Stauffer, J. Glaciol. 46, 45 (2000).
4. J. M. Barnola, P. Pimienta, D. Raynaud, Y. S. Korotkevich, Tellus 43B, 83 (1991).
5. H. Fischer, M. Wahlen, J. Smith, D. Mastroianni, B. Deck, Science 283, 1712 (1999).
6. A. Neftel, H. Oeschger, T. Staffelbach, B. Stauffer, Nature 331, 609 (1988).
7. T. Staffelbach, B. Stauffer, A. Sigg, H. Oeschger, Tellus 43B, 91 (1991).
8. O. Marchal et al., Clim. Dyn. 15, 341 (1999).
9. C. D. Charles, J. Lynch-Stieglitz, U. S. Ninnemann, R. G. Fairbanks, Earth Planet. Sci. Lett. 142, 19 (1996).
10. C. Walbroeck et al., Clim. Dyn. 12, 113 (1995).
11. J.-J. Pichon et al., Paleoceanography 7, 289 (1992).
12. A. Mazaud, F. Vimeux, J. Jouzel, Earth Planet. Sci. Lett. 177, 219 (2000).