What is the role of sea ice in the climatic response to Milankovitch forcing?

ANDRÉS ANTICO; OLIVIER MARCHAL; LAWRENCE A. MYSAK

Simposio; Climate change: from the geologic past to the uncertain future: A tribute to André Berger; 2008

Universidad Louvain-la-Neuve

Cycles at Milankovitch frequencies have been argued to be present in paleoclimate records from ocean sediment cores for epochs such as the Pliocene when little or no ice was present on land. We have investigated the time-dependent response of a land ice-free climate to Milankovitch forcing using a coupled atmosphere-ocean-sea ice model. The ocean model component is a zonally averaged model of the circulation in 5 basins (Arctic, Atlantic, Indian, Pacific and Southern Oceans). The sea-ice model component is purely thermodynamic and the atmosphere model component is a one-dimensional (latitudinal) energy balance model. The coupled model is used to conduct three different numerical experiments on the time-dependent response of climate to Milankovitch forcing for the time interval from 5 Myr to 3 Myr ago. The first experiment does not include the sea-ice model component. The second experiment includes the sea-ice model component but the effect of sea ice on ocean salinity is neglected. Finally, the third experiment includes the sea ice component and the effect of sea-ice cover on sea surface freshwater fluxes is incorporated. The first experiment (no sea-ice) confirms some results obtained in a previous study (Brickman et al., 1999): a) the annual mean surface air temperature linearly responds to changes in annual mean insolation at the obliquity and eccentricity frequencies, b) the response of deep ocean temperature to eccentricity is suppressed, and c) ocean convection rectifies the effect of precession on deep ocean temperature. However, the inclusion of sea ice modifies significantly the climate responses to Milankovitch forcing: a) in both polar regions, sea ice is highly sensitive to changes in local summer insolation at the precession, obliquity and eccentricity frequencies , b) sea ice variations rectify the effect of precessional forcing on the annual mean surface air temperature at high latitudes, and c) the response of deep ocean temperature drastically changes when the effect of sea ice on salinity is incorporated in the model. At precessional frequencies, the temperature variations of deep ocean water of northern and southern origin are approximately in (out of) phase when sea-ice volume changes does (not) modify ocean salinity; this is a consequence of the strong effect of surface salt fluxes on the convective heat fluxes in the Southern Ocean. These numerical experiments suggest that sea ice plays an important role in the time-dependent response of the climate system to Milankovitch forcing and should provide a useful reference for future investigations based on more complete models.