UC Santa Barbara

Benthic δ 18O is often used as a stratigraphic tool to place marine records on a common age model and as a proxy for the timing of ice volume/sea level change. However, Skinner and Shackleton (2005) found that the timing of benthic δ 18O change at the last termination differed by 3900 years between one Atlantic site and one Pacific site. These results suggest that benthic δ 18O change may not always accurately record the timing of deglaciation. We compare benthic δ 18O records from 20 Atlantic sites and 14 Pacific sites to evaluate systematic differences in the timing of terminations in benthic δ 18O. Analysis of sedimentation rates derived from the alignment of benthic δ 18O suggests a statistically significant Atlantic lead over Pacific benthic δ 18O change during the last six terminations. We estimate an average Pacific benthic δ 18O lag of 1600 years for Terminations 1-5, slightly larger than the delay expected from ocean mixing rates given that most glacial meltwater probably enters the North Atlantic. We additionally find evidence of ∼4000-year Pacific δ 18O lags at approximately 128 ka and 330 ka, suggesting that stratigraphic correlation of δ 18O has the potential to generate age model errors of several thousand years during terminations. A simple model demonstrates that these lags can be generated by diachronous temperature changes and do not require slower circulation rates. Most importantly, diachronous benthic δ 18O responses must be taken into account when comparing Atlantic and Pacific benthic δ 18O records or when using benthic δ 18O records as a proxy for the timing of ice volume change. Copyright 2009 by the American Geophysical Union.