1Institut Pierre-Simon Laplace/Laboratoire des Sciences du Climat et de l'Environnement, CEA-UMR INSU/CNRS 8212-UVSQ, 91191 Gif-sur-Yvette, France
2Laboratoire de Glaciologie et Géophysique de l'Environnement, CNRS-UJF, 38400 St Martin d'Hères, France
3Climate and Environmental Physics, Physics Institute, and Oeschger Centre for Climate Change Research, University of Bern, Sidlerstrasse 5, 3012 Bern, Switzerland
4Centre for Ice and Climate, Niels Bohr Institute, University of Copenhagen, Juliane Maries Vej 30, 2100, Copenhagen, Denmark
5Alfred Wegener Institute for Polar and Marine Research, P.O. Box 120161, 27515 Bremerhaven, Germany
6University of Trieste, Department of Geological, Environmental and Marine Sciences, Via E. Weiss 2, 34127 Trieste, Italy
Abstract. Since its discovery in Greenland ice cores, the millennial scale climatic variability of the last glacial period has been increasingly documented at all latitudes with studies focusing mainly on Marine Isotopic Stage 3 (MIS 3; 28–60 thousand of years before present, hereafter ka) and characterized by short Dansgaard-Oeschger (DO) events. Recent and new results obtained on the EPICA and NorthGRIP ice cores now precisely describe the rapid variations of Antarctic and Greenland temperature during MIS 5 (73.5–123 ka), a time period corresponding to relatively high sea level. The results display a succession of long DO events enabling us to highlight a sub-millennial scale climatic variability depicted by i) short-lived and abrupt warming events preceding some Greenland InterStadial (GIS) (precursor-type events) and ii) abrupt warming events at the end of some GIS (rebound-type events). The occurrence of these secondary events is suggested to be driven by the Northern Hemisphere summertime insolation at 65° N together with the internal forcing of ice sheets. Thanks to a recent NorthGRIP-EPICA Dronning Maud Land (EDML) common timescale over MIS 5, the bipolar sequence of climatic events can be established at millennial to sub-millennial timescale. This provides evidence that a linear relationship is not satisfactory in explaining the link between Antarctic warming amplitudes and the duration of their concurrent Greenland Stadial (GS) for the entire glacial period. The conceptual model for a thermal bipolar seesaw permits a reconstruction of the Antarctic response to the northern millennial and sub-millennial scale variability over MIS 5. However, we show that when ice sheets are extensive, Antarctica does not necessarily warm during the whole GS as the thermal bipolar seesaw model would predict.