| description abstract | A coupled atmosphere?mixed layer ocean GCM (GENESIS2) is forced with altered orbital boundary conditions for paleoclimates warmer than modern (6 kyr BP) and colder than modern (115 kyr BP) in the high-latitude Northern Hemisphere. A pair of experiments is run for each paleoclimate, one with sea-ice dynamics and one without, to determine the climatic effect of ice motion and to estimate the climatic changes at these times. At 6 kyr BP the central Arctic ice pack thins by about 0.5 m and the atmosphere warms by 0.7 K in the experiment with dynamic ice. At 115 kyr BP the central Arctic sea ice in the dynamical version thickens by 2?3 m, accompanied by a 2 K cooling. The magnitude of these mean-annual simulated changes is smaller than that implied by paleoenvironmental evidence, suggesting that changes in other earth system components are needed to produce realistic simulations. Contrary to previous simulations without atmospheric feedbacks, the sign of the dynamic sea-ice feedback is complicated and depends on the region, the climatic variable, and the sign of the forcing perturbation. Within the central Arctic, sea-ice motion significantly reduces the amount of ice thickening at 115 kyr BP and thinning at 6 kyr BP, thus serving as a strong negative feedback in both pairs of simulations. Ice motion causes the near-surface air to cool in both sets of experiments, however, thus representing a positive feedback at 115 kyr BP and a negative feedback at 6 kyr BP. The excess cooling with ice motion at 115 kyr BP is caused by the enhanced, advective spreading of the ice pack into the North Atlantic dominating over the warming tendency from the thinner central Arctic sea ice. The reduced atmospheric warming due to ice dynamics at 6 kyr BP is caused by sea-ice ridging, a thickening process that partially counteracts the orbitally induced atmospheric warming perturbation. | |
