Bibliographie complète

Abstract The last deglaciation (20.0–10.0 kyr B.P.) was punctuated by two major cooling events affecting the Northern Hemisphere: the Oldest Dryas (OD; 18.0–14.7 kyr B.P.) and the Younger Dryas (YD; 12.8–11.5 kyr B.P.). Greenland ice core δ 18 O temperature reconstructions suggest that the YD was as cold as the OD, despite a 50 ppmv increase in atmospheric CO 2 , while modeling studies suggest that the YD was approximately 4–5°C warmer than the OD. This discrepancy has been surmised to result from changes in the origin of the water vapor delivered to Greenland; however, this hypothesis has not been hitherto tested. Here we use an atmospheric circulation model with an embedded moisture‐tracing module to investigate atmospheric processes that may have been responsible for the similar δ 18 O values during the OD and YD. Our results show that the summer‐to‐winter precipitation ratio over central Greenland in the OD is twice as high as in the YD experiment, which shifts the δ 18 O signal toward warmer (summer) temperatures (enriched δ 18 O values and it accounts for ~45% of the expected YD‐OD δ 18 O difference). A change in the inversion (cloud) temperature relationship between the two climate states further contributes (~20%) to altering the δ 18 O‐temperature‐relation model. Our experiments also show a 7% decrease of δ 18 O‐depleted precipitation from distant regions (e.g., the Pacific Ocean) in the OD, hence further contributing (15–20%) in masking the actual temperature difference. All together, these changes provide a physical explanation for the ostensible similarity in the ice core δ 18 O temperature reconstructions in Greenland during OD and YD. , Key Points Precipitation seasonality and inversion temperature changes behind YD‐OD δ 18 O enigma Local processes changes accounting up to 65% of the expected YD‐OD δ 18 O difference Moisture transport changes from the Pacific accounting only up to 20% of the expected YD‐OD δ 18 O difference