| Summary: | American Geophysical Union Fall Meeting, 12-16 December 2016, San Francisco The Marine Isotope Stage (MIS) 13 interglacial is unusual in that warm Northern Hemisphere conditions were accompanied by relatively cool Southern Hemisphere conditions and because it was preceded by a mild glaciation (MIS 14) with less ice volume and higher CO2 levels than the two preceding glacial maxima. Here we investigate Late Pleistocene glacial cycles, and MIS 13 in particular, using two relaxation models from García-Olivares & Herrero [2013] that describe the relationships between global ice volume (V), atmospheric CO2 (C) and the extent of the Antarctic ice shelves (A). The two models differ in parameterizing deep ocean stratification as either a function of V and A (model 3τ) or as a function of C and A (model LS). Note that global ice volume, V, is most closely related to Northern hemisphere climate, whereas C and A are most closely related to Antarctic climate. Here we present the results of using a sea level stack [Spratt & Lisiecki, 2016] as the ice volume tuning target instead of benthic δ18O. We find that tuning to the sea level stack dramatically improves the simulation of MIS 13 in the 3τ model. With the sea level stack, 3τ correctly reproduces the weak amplitudes of MIS 13 and 14 and a double peak in CO2 during MIS 13, whereas the LS model does not reproduce these features using either tuning target. The first peak in CO2 follows a minor ice volume decrease at ~530 kyr but significantly precedes a second, larger sea level rise at ~500 kyr. The later sea level rise coincides with a second benthic δ18O decrease and likely triggered the second CO2 peak. This two-step transition to peak interglacial conditions might be caused by deep ocean stratification and Antarctic ice cover acting out of phase: weakened stratification produced an initial pulse of CO2 from the deep ocean, but because Antarctic warming was unusually weak, the Antarctic ice shelf remained relatively wide and less CO2 than usual was released from ...
|
|---|