The δ18O of atmospheric O2 from air inclusions in the Vostok Ice Core: Timing of CO2 and ice volume changes during the penultimate deglaciation

One important question about the CO2-climate connection is how increasing atmospheric pCO2 levels affect climate and thereby the mass balance of continental ice sheets. A record of atmospheric CO2 variations over the last 160,000 years has recently been constructed by analyzing the trapped gas in the Vostok ice core [Barnola et al., 1987]. The relationship between changes in atmospheric CO2 and the size of the continental ice sheets has been difficult to ascertain because the CO2 record is obtained from ice cores while the ice volume record has been constructed from the stable isotopic composition of biogenic CaCO3 in deep-sea sediment cores. In order to compare these two records in a more precise manner, we present a record of the isotopic composition (δ18O) of atmospheric O2 trapped in the Vostok ice core, and propose that it may be considered a proxy for the δ18O of seawater and hence ice volume. Having a record of atmospheric CO2 along with a continental ice volume proxy in trapped air in the same ice core allows us to compare the timing of changes in these two parameters with little uncertainty in the relative ages of important events. Our results suggest that, during the penultimate glacial termination, atmospheric CO2 began to increase at least 3 kyr before the initial introduction of meltwater to the oceans. Possible errors in the relative age-depth curve of the Vostok ice core and uncertainties in the influence of biological and hydrologic processes on the δ18O of atmospheric O2 introduce some uncertainty into our conclusions. However, our results are in general agreement with the observed phase relationship between atmospheric CO2 and ice volume [Imbrie et al., 1984; Imbrie et al., 1989] (inferred from changes in the δ13C difference between contemporaneous planktonic and benthic forams deposited in deep-sea cores and the SPECMAP ice volume record).