The heterogeneous global structure of Dansgaard-Oeschger events in paleoclimate archives and climate simulations

Abstract:

Dansgaard-Oeschger (DO) events are among the most iconic examples of past abrupt climate changes. They are prominently seen in Greenland ice cores during Marine Isotope Stage 3, when a warming of more than 10K within a few decades occurred multiple times. While the imprint of DO events in the North Atlantic and its interplay with Antarctica is well-established, its global fingerprint is less studied. However, to understand the potential environmental and societal impacts of any future tipping points, a more complete understanding of the spatial patterns of past abrupt events is important.

We analyze the spatial structure of temperature and precipitation during DO events using a large database of lake sediment, speleothem, and marine sediment records. The chronological uncertainties and limited temporal resolution of most proxy records hamper the identification of the exact onset and temporal shape of DO events on a global scale. Instead, we construct regional composites that constrain the spatial temperature and precipitation difference patterns between interstadials and preceding stadials. This approach additionally enables the classification of DO events according to their spatial similarity.

Secondly, we show that large differences exist in the spatial responses of temperature and precipitation in simulations of DO-like events and freshwater-induced AMOC changes with three climate models of varying complexity. Comparison against the proxy-based estimates indicates that, in particular, precipitation responses are more realistically simulated in models of higher complexity.