A new conceptual model for interpolar climate coupling during the Ice Ages

The thermal bipolar seesaw model by Stocker and Johnsen is widely used to explain the connection between abrupt Dansgaard-Oeschger events in Greenland and their climate response in Antarctica. We now have better-dated records from Greenland and Antarctica with higher resolutions which provide critical information into furthering our understanding of interpolar climate coupling.

Records of water isotopic composition from Greenland and Antarctic ice cores were used as a proxy for past temperatures. The NGRIP core is dated via the Greenland Ice Core Chronology 2005. The Antarctic ice cores come from five different locations: WAIS Divide, EDML, EDC, Dome Fuji, and Talos Dome, and were synchronized using volcanic activity. MatLab was the primary tool used for data analysis, combining statistical correlation and modeling on the updated data sets.

First, we use the updated data sets to replicate the bipolar seesaw. WAIS Divide, Talos Dome, and EDC produced the best depiction of the seesaw, whereas EDML and Dome Fuji had lower correlations even though their proximity is much closer to the South Atlantic.

Next, we suggest a simple conceptual model that can replicate both millennial and orbital-scale Antarctic climate during the last ice age using greenhouse gas forcing, surface albedo, and the AMOC. Motivated by recent modeling results, we propose that Antarctic climate simply reflects the mean ocean temperature; in this view it is the global ocean interior, rather than the Southern ocean, that acts as the heat reservoir in the bipolar seesaw. Interstadial periods represent periods of oceanic heat loss via the North Atlantic, thereby cooling Antarctica. By using three AMOC states (strong, weak, and off), we can simulate the D/O cycle, confirming the seesaw is part of the machinery of glacial cycles. The success of our simple model approach, combined with results from recent GCM simulations, suggests a revised view of the seesaw concept may be warranted.