Interaction between the thermohaline circulation and sea ice in Dansgaard-Oeschger cycles.

Abstract

During the last glacial period large fluctuations on a millennial timescale are observed in the Northern Hemisphere. Several mechanisms have been proposed to explain these Dansgaard- Oeschger (DO) cycles. The most commonly accepted mechanism is related to oscillations in the Atlantic meridional overturning circulation (AMOC), brought about by periodic changes in the freshwater flux into the North Atlantic. In a more recent theory it is proposed that the dominant process is subsurface warming in the Nordic Seas induced by increasing sea ice cover, which eventually leads to instability of the water column. Both of these theories correspond well to part of the observations, but neither are unable to explain the full range of observational data. Here we combine both mechanisms in a conceptual model study. In addition we include a land ice component to capture the effect of Heinrich events. It is shown that the model results correspond well to observational data and that the combination of these mechanisms is able to explain more of the features of the DO-cycle than each of the individual mechanisms. The full conceptual model provides an explanation for the structure of the events and accounts for the duration of stadials and interstadials and the variability therein. It shows why Heinrich events are triggered during stadials and how they can delay the occurrence of a subsequent DO-event. As a consequence it also explains one of the most striking features of the DO-cycle; the double peak in the probability distribution of the waiting time between consecutive events.