Polynya ice production of the Laptev Sea area for the winters 2007/08 and 2008/09 calculated from simulations with COSMO numerical weather prediction (NWP) model

Global Change and its predicted key impact on the Arctic bring the Laptev Sea to the centre of climate-related polar research. This Shelf Sea is known as being a highly productive area for the formation of new ice throughout the winter season. A main part of the ice production occurs in flaw polynyas which appear recurrently at the edge of the fast ice surrounding the coastal zones during wintertime. This work attempts to provide a method to reliably estimate the ice production in the Laptev Sea polynyas from model result s of the numerical weather prediction model COSMO. Our modeling approach contains the use of COSMO with 15 and 5 km horizontal resolution nested in global GME/ERA-Interim data to calculate ice production in polynyas. To account for realistic polynya representation polynya area is prescribed to the COSMO model by daily AMSR-E satellite data. The potential volume ice production is calculated from atmospheric net radiation fluxes. In contrast to preceding studies our new COSMO-based method takes into account the effect of polynyas on the atmosphere. Over open water, warmer 2m temperatures (COSMO in comparison to NCEP) lead to lower ice production. Over thin ice, surface temperature depends on air temperature and reduced air surface temperature gradients cause lower heat fluxes and less ice production than over open water. As warm-biased NCEP values are balancing the effects of our improvements the comparison of ice production retrieval based on NCEP data with our results show minor total differences only. Both methods are leading to results in same order of magnitude if the polynya is assumed to be covered with 10cm of thin-ice. This supports the thesis that either of them leads to feasible ice production values if thin ice within the polynya is accounted for in the calculation. In case of an open water polynya, however, our study underlines the impact of the atmospheric data on the ice production. Thus we conclude that it is of major importance to choose a validated ice thickness parameterization for the model