Oxygen-Isotope Analyses and Pleistocene Ice Volumes

The oxygen-isotope record from fossil foraminifera in deep-sea sediments is commonly used as a proxy for global ice volume. The linkage between δ18O and ice volume, however, is probably nonlinear. We have developed a simple numerical model of the isotopic response of the oceans to ice-volume change. The major features it simulates are (1) the changing mean isotopic composition of snow as a function of ice volume (colder snow temperatures forced by climate change and higher-elevation accumulation areas imply more negative mean δ18O); (2) the nonequilibrium isotopic composition of ice sheets (the past history of an ice sheet is integrated into its mean isotopic composition, which introduces a lag of isotopic “ice volume,” i.e., the measured δ18O record, scaled to ice-volume units, behind true ice volume); (3) selective preservation of isotopically more negative (colder, higher-latitude) ice (this geographic effect can selectively amplify or dampen the isotopic response to the ice-volume signal). We illustrate the response of our model to simple hypothetical ice-volume transitions of ice growth and ice decay. Sensitivity tests are illustrated for all model parameters. The results suggest that oxygen-isotope records reproduce the general patterns of ice-volume change fairly accurately. The foraminiferal isotope record, however, may misrepresent the true amplitude of the ice-volume signal and lag true ice volume by 1000 to 3000 yr.