Modelling glacial climatic millennial-scale variability related to changes in the Atlantic meridional overturning circulation: a review

Since the discovery of glacial abrupt variability in Greenland and the North Atlantic about two decades ago, its modelling remains a conceptual and technical challenge. While the global climatic signature of Heinrich events can reasonably be ascribed to the fresh water discharge related to the massive iceberg armadas released by the Laurentide ice-sheet in the North Atlantic, the existence of such a triggering mechanism is elusive as far as Dansgaard-Oeschger type variability is concerned. Even if other mechanisms have been proposed, the majority of modelling attempts to better understand glacial climate variability has assumed it to be linked to abrupt variations in the Atlantic Meridional Overturning Circulation (AMOC). In this review we introduce the rationale behind this approach, i.e. the possible existence of several deep ocean circulation equilibria associated with distinct climate signatures. We then review the published fresh water hosing experiments and the mechanisms at work to explain the climate differences related to the different AMOC states. Most of these works have been performed under pre-industrial boundary conditions. The few hosing experiments run under glacial conditions show that the sensitivity of the climate system to fresh water hosing is very dependent on the background climate state. In order to rigourously compare several types of experiments (equilibrium runs vs transient runs with pre-industrial vs glacial boundary conditions) we present experiments especially run for this review with two climate models of intermediate complexity: UVic and LOVECLIM. Although there are similarities between the responses simulated by these models, there are also many differences which would benefit being compared to reconstructions such as those presented in this special issue. However, a strict comparison will require to consider boundary conditions as realistic as possible for Marine Isotopic Stage 3. Modelling glacial variability to compare model results with paleoclimatic data remains a big challenge. (C) 2010 Elsevier Ltd. All rights reserved.