Sensitivity of General Circulation Models to Changes in Sea-Ice Cover

Abstract

The last ten years have seen a number of experiments with atmospheric general circulation models (GCMs) to investigate the effect on the modelled atmosphere of changing the prescribed sea-ice cover. Tes responses obtained were often zonally asymmetric, demonstrating the limitations implicit in conducting zonal mean studies. For example, Newson (1973) found that by completely removing ice cover in the northern hemisphere in winter, in addition to the expected great warming at low levels in the Arctic, there was about a 1K cooling in the zonal mean at 30–50° N (Fig. 1). However, this was a manifestation of much larger cooling locally over the United States (8K), E Siberia (6K) and N Europe (2K) (Fig. 2). It was found that the Atlantic low moved SW when the ice was removed, reducing the warm low level westerlies over W Europe. Warshaw and Rapp (1972) however, in similar experiments with a 2-level model, found a zonal mean warming extending further south (Fig. 3), which they attributed to the outbreaks of polar air being less severe. It was not stated how much of this warming could be because of local heating, resulting from the removal of ice in Hudson Bay, for example. Flechter et al. (1973), using the same model as Warshaw and Rapp, found a smaller meridional temperature gradient tending to reduce baroclinic instability and the zonal mean westerles, but the low level heating and slight high level cooling greatly reduced static stability which has the effect of increasing baroclinic instability.