Realistic representation of vertical melt pond thermodynamics in the CICE sea ice model

Abstract:

Realistic parametrisation of melt ponds plays a significant role in the ability of sea ice models to accurately predict seasonal evolution of the ice cover in the Arctic by enhancing the albedo feedback mechanism during the melting season. In addition, the presence of melt ponds during winter refreezing has a potential to delay the basal growth of sea ice. Effects of melt pond formation and refreezing on sea ice predictability present an uncertainty in climate simulations, as these do not realistically account for storage and transfer of heat in the extra layer of melt water. We address this issue by modifying the vertical thermodynamics scheme in the sea ice model CICE, to include the small scale physics of melt pond evolution. It allows for concurrent phases to exist in each sea ice category (snow, ice lid on top of melt pond, melt pond, ice beneath the pond) and explicitly resolves the energy transfer between them. Early runs suggest this modification affects the summer and autumn sea ice thickness and concentration not only in regions with persistent melt ponds but at the scale of the entire Arctic basin. We investigate the sensitivity of the model to the new melt pond scheme, its impacts on Arctic seasonal variability and the mechanisms through which it affects the sea ice melt rate and albedo.