CORRELATIONS BETWEEN OXYGEN ISOTOPE RECORDS FROM CREVICE CAVE, MISSOURI AND GREENLAND DURING THE LAST GLACIAL PERIOD

Oxygen isotope variations within stalagmites from Crevice Cave in eastern Missouri have been resolved at the level of 30-200 years and can be compared to the GISP2 and GRIP ice-core records. The speleothem record extends from 65,000 to 25,000 years ago (ka) and is well-dated by over 30 high-precision U-Th dates. The correlation presented here is based on numerous similarities in both large-scale and detailed features of the oxygen isotope records. Both the inland location of Crevice Cave and the detailed mimicry of events at resolutions as high as decades indicate the role of the atmosphere in linking the Mid-continent and the North Atlantic, and imply synchroneity among events.

Assuming these correlations between the Missouri and Greenland records are correct and that similar events were contemporaneous, offsets in timing among the Crevice Cave, GRIP, and GISP2 records follow a pattern that is consistent with the methods by which the ice core time scales were derived. Between 25 and 45 ka, the Crevice Cave time scale is older than GRIP (flow modeled) and is closer in age but younger than GISP2 (layer counted back to approximately 45 ka). This might plausibly reflect a tendency in the GISP2 method to occasionally count extra yearly events. Especially large time scale deviations occur between 45 and 48 ka and between 55 and 59 ka. During the 45-48 ka interval, the ice core time scales are compressed relative to the Crevice Cave time scale (GISP2 more so than GRIP), possibly because the snow accumulation rates assumed during this interval are too high. This interval corresponds to Heinrich Event 5. In contrast, the ice core time scales during the 55-59 ka interval are expanded relative to the speleothem time scale (again, GISP2 more so than GRIP), possibly because the snow accumulation rates assumed during this interval are too low. This interval is one of high oxygen isotope values (both for the cave and ice core records), and of presumably warmer temperatures. Therefore, these observations support the GRIP tenet of scaling the snow accumulation rate to ice oxygen isotope values. The correlation presented here suggests, however, that the magnitude of this effect has been underestimated.