Does the bipolar seesaw extend to the terrestrial southern mid-latitudes?

High precision comparison of Greenland and Antarctic ice cores, suggesting a pervasive antiphased temperature relationship between the polar hemispheres during the last glaciation, lends strong support to the bipolar seesaw model (EPICA, 2006). The extent to which reorganisation of ocean-heat transport during abrupt climate change events affected the southern mid-latitudes remains unclear, however, owing to a paucity of well-dated records with robust climate proxies, variability between some records, and varying interpretations of their significance. Here we present temperature reconstructions for three key pollen records recognised by the NZ-INTIMATE (NZ-I) group which, along with the preliminary NZ-I climate event stratigraphy (Alloway et al., 2007) and published marine records, are compared with polar ice core records for the interval 30–10 cal. ka. We focus on key events within the context of Dansgard Oeschger cycles 4–1 and The Antarctic Cold Reversal/Younger Dryas intervals. The New Zealand records are broadly consistent with an extended bipolar seesaw whereby the oceanic southern mid-latitudes are warmed at times of MOC weakening or cessation in the North Atlantic, and vice versa. Variability between records indicate that other factors must be involved, however, and nor do these records refute alternative models that predict an antiphased inter-hemispheric pattern. Nevertheless an extended bipolar model may explain an early onset of LGM conditions in New Zealand and elsewhere in the Southern Hemisphere at a time when interstadials GI3 and GI4 kept Greenland warm. Similar inter-hemispheric dynamics have been invoked to explain an earlier termination of the LGM in Antarctica than in Greenland (Wolff et al., 2009) which is also evident in the New Zealand records. A prominent mid-LGM interstadial complex observed in several New Zealand records, connected by tephrochronology may represent another antiphased event although stronger chronological control is needed to support this assertion. By comparison, the variability evident between New Zealand Lateglacial records cannot be explained simply by chronological imprecision and there may be latitudinal controls on the extent to which the Antarctic pattern is registered, as has been suggested from other records from the southern mid-high latitudes. Consolidation of these patterns is needed from more precise climate reconstructions at sites spread across latitudinal gradients in both the marine and terrestria