A one-dimensional thermodynamic model of sea ice is presented that focuses on those features that are most relevant to interactions with the atmosphere, namely the surface albedo and leads. It includes a surface albedo parameterization that interacts strongly with the state of the surface, and explicitly includes meltwater ponds. The lead parameterization contains a minimum lead fraction, absorption of solar radiation in and below the leads, lateral accretion and ablation of the sea ice, and a prescribed sea ice divergence rate. The model performed well in predicting the current climatic sea ice conditions in the central Arctic when compared with observations and other theoretical calculations. Results of parameter sensitivity tests produced large equilibrium ice thicknesses for small values of ice divergence or large values of minimum lead fraction as a result of positive feedback mechanisms involving cooling of water in the leads. The ice thickness was also quite sensitive to the meltwater runoff fraction and moderately sensitive to the other parameters in the melt pond parameterization, a result of the strong dependence of the surface albedo, and hence the net flux, on the surface condition. To further investigate the physical interactions and internal feedback processes governing the sea-ice-lead system, sensitivity tests were also performed for each of the external forcing variables.

The model's equilibrium sea ice thickness was extremely sensitive to change in the downward longwave and shortwave fluxes and atmospheric temperature and humidly, moderately sensitive to the value of the ocean heat flux, and insensitive to values of wind speed, snowfall, and rainfall in the immediate vicinity of the baseline forcing, although significant changes in thickness occurred for larger variations in wind speed and snowfall. Four important positive feedback loops were identified and described: (1) the surface albedo feedback, (2) the conduction feedback, (3) the lead solar flux feedback, and (4) the lead fraction feedback. The destabilizing effects of these positive feedbacks were mitigated by two strong negative feedbacks: (1) the outgoing longwave flux feedback, and (2) the turbulent flux feedback. Considering the strong influence which sea ice has on global atmospheric and oceanic circulation patterns, it is essential that climate models be able to treat these feedback processes appropriately. ¿ American Geophysical Union 1993