ON MODELING THE SEASONAL THERMODYNAMIC CYCLE OF SEA ICE IN STUDIES OF CLIMATIC-CHANGE

Recent work in modeling climatic changes due to increased atmospheric CO2 has shown the maximum change to occur in the polar regions as a result of seasonal reductions in sea ice coverage. Typically, sea ice thermodynamics is modeled in a very simple way, whereby the storage of both sensible and latent heat within the ice is ignored, and the effects of snow cover on conductivity and on surface albedo and of oceanic heat flux on bottom ablation may also be neglected. Whether omission of these processes is justified within the context of quantitatively determining regional climatic changes is considered. A related question, whether omission of ice dynamics can be justified, is not considered. Relatively complete 1-dimensional models of sea-ice thermodynamics were previously developed and tested for a variety of environmental conditions by Maykut and Untersteiner (1969, 1971) and by Semtner (1976). A simpler model which neglects the storage of sensible and latent heat is described in the appendix to Semtner (1976). In that model, the errors in annual-mean ice thickness which would arise from neglect of heat storage can be compensated by increases in albedo and in conductivity. The seasonal cycle of ice thickness predicted by such a model is examined and significant errors in phase (1 mo. lead) and in amplitude (.apprxeq. 50% overestimate) are found. The amplitude errors are enhanced as snowfall and oceanic heat flux diminish (or are neglected). This suggests that substantial errors may occur in climate simulations which use very simple formulations of sea ice thermodynamics, whereby early and excessive melting exaggerates the seasonal disappearance of sea ice. To illustrate the above point, 2 models are configured to examine the local response of Arctic sea ice to a quadrupling of atmospheric CO2. The 1st model neglects a number of physical processes and mimics the behavior of sea ice found in Manabe and Stouffer (1980), both for present and enhanced levels of CO2. The more complete second model gives a better simulation of Arctic ice for the present level of CO2 and shows a reduced response to CO2 quadrupling relative to that in Manabe and Stouffer (1980). In particular, the change in surface temperature is cut by a factor of 2. In view of this result, a more complete treatment of sea ice thermodynamics would seem warranted in further studies of climate change. Only a minor computational increase is required.