Improvement of sea ice thermodynamics with variable sea ice salinity and melt pond parameterizations in an OGCM

Enhanced representation of sea ice processes in ocean modeling studies is required for advancing our under-standing and prediction of the climate variability. In the present study, we improved the sea ice thermodynamics in an OGCM by introducing processes for variable sea ice salinity and melt ponds on sea ice. The former affects the latent heat and conductivity of sea ice as well as the salt budget in the ocean-sea ice system. We formulated salinification and desalination processes based on previous observational studies. Melt ponds enhance sum-mertime surface melting by reducing the surface albedo. In our formulation of the melt pond thermodynamics, the shortwave radiation absorbed in the pond was partly released to the air and partly conducted to the un-derlying sea ice. By adopting these parameterizations, seasonal cycles of the sea ice salinity distributions in both hemispheres were generally reproduced. Simulated seasonal and interannual variations in the Arctic melt pond area and associated surface albedo reductions were basically consistent with observational estimates. The melt pond effect reduced the Arctic sea ice volume by about 1.5 x 103 km3 in the summer; half of this impact remained throughout the year. Our results suggest that the new parameterizations have the potential to improve the seasonal evolution of the sea ice thickness and concentration distributions, which would contribute to prediction studies.