Model tuning and its impact on modelled climate change response: Hudson Bay sea ice, a case study

The impact of climate model parameterization is examined using Hudson Bay sea ice as a case study. The implicit diffusion of an upstream scheme used in a three dimensional ocean general circulation model is locally determined by using a one dimension proxy model of Hudson Bay with atmospheric, cryospheric and oceanic components. Using this proxy model, distinct pairs of values for the thermal conductivity of sea ice and thermal diffusivity of sea water are found that reproduce the current climatology of sea ice thickness in Hudson Bay. These pairs of values are subsequently used in a 3°C warming scenario. The resulting ice thicknesses are shown to be highly dependent on the choice of these pairs of values as well as the seasonal distribution of the warming.