Modeling the evolution of snow, snow ice and ice in the Baltic Sea

A numerical 1-dimensional fine grid sea ice thermodynamic model is constructed accountingspecially for: (1) slush formation via flooding and percolation of rain- and snow meltwater,(2) the consequent snow ice formation via slush freezing, and (3) the effects of snow compactionon heat diffusion in snow cover. The model simulations from ice winter period 1979–90 areviewed against corresponding observations at the Kemi fast ice station (65°39.8’N, 24°31.4’E).The 11-year averaged model results show good overall consistency with corresponding totalice thickness observations. The model slightly overestimates the snow ice thickness and underestimatesthe snow thickness in February and March, which is mainly addressed to the modelassumption of isostatic balance (i.e., slush formation via flooding), which was probably notfully satisfied at the coastal Kemi fast ice station. Supposing that this assumption is neverthelessgenerally valid away from the very coastal fast ice zone, an estimate for sea ice sensitivity tochanges in winter precipitation rate is produced. Increased precipitation leads to an increaseonly in snow ice thickness with little change in total ice thickness, while a reduction in precipitationof more than — 50% causes a significant increase in total ice thickness. The difference inmodeled total ice thickness for the case of artificially neglecting snow ice physics is about 25%, which indicates the importance of including snow ice physics in a sea ice model dealing withthe seasonal sea ice zone.