. The impact of uncertainty in Arctic Ocean precipitation on estimates of snow depth and density over sea ice

Abstract
Precipitation is a fundamental input for snow accumulation-melt models developed to estimate snow thickness and density for use in laser and radar altimeter retrievals of sea ice thickness. Of the current crop of reanalyses, only four product will extend through the planned operations of ICESat2 and CryoSat2: NASA‘s MERRA2, NOAA’s CFSRv2, ECMWF’s ERA5, and the Japanese JRA55. No one reanalysis is better than the others: MERRA2 and CFSRv2 have positive biases of about 50 mm based on pan-Arctic mean precipitation; JRA55 has a negative bias of 50 mm. Biases vary spatially: MERRA2 and CFSRv2 tend to have a wetter eastern Arctic than JRA55. ERA5 is still in production. A full run from 1979 to present is slated to be available by the end of 2018. Such large spread in precipitation estimates introduces uncertainty in estimates of snow cover over sea ice. In this presentation, we evaluate spread of precipitation estimates, along with other forcings, from the four covering the ICESat2 and CryoSat2 periods, and the associated uncertainty in estimates of snow depth and density for North Pole drifting stations, and SHEBA. We also explore the impact on snow cover uncertainty of bias correcting precipitation.