On the use of δ18Oatm for ice core dating

Deep ice core chronologies have been improved over the past years through the addition of new age constraints. However, dating methods are still associated with large uncertainties for ice cores from the East Antarctic plateau where layer counting is not possible. Indeed, an uncertainty up to 6 ka is associated with AICC2012 chronology of EPICA Dome C (EDC) ice core, which mostly arises from uncertainty on the delay between changes recorded in δ18Oatm and in June 21st insolation variations at 65°N used for ice core orbital dating. Consequently, we need to enhance the knowledge of this delay to improve ice core chronologies. We present new high-resolution EDC δ18Oatm record (153–374 ka) and δO2/N2 measurements (163–332 ka) performed on well-stored ice to provide continuous records of δ18Oatm and δO2/N2 between 100 and 800 ka. The comparison of δ18Oatm with the δ18Ocalcite from East Asian speleothems shows that both signals present similar orbital and millennial variabilities, which may represent shifts in the InterTropical Convergence Zone position, themselves associated with Heinrich events. We thus propose to use the δ18Ocalcite as target for δ18Oatm orbital dating. Such a tuning method improves the ice core chronology of the last glacial inception compared to AICC2012 by reconciling NGRIP and mid-latitude climatic records. It is especially marked during Dansgaard-Oeschger 25 where the proposed chronology is 2.2 ka older than AICC2012. This δ18Oatm – δ18Ocalcite alignment method applied between 100 and 640 ka improves the EDC ice core chronology, especially over MIS 11, and leads to lower ice age uncertainties compared to AICC2012