Ocean/Sea Ice: Assimilation Setup: Sea Ice



The last sequence consists of the assimilation of sea ice concentration, from NSIDC or CMIP5. Sea-ice concentration is defined as the fraction of ocean area covered by sea ice and assumed to be the same variable as the sum of the five types of fractional sea ice in the CICE model where the total fractional ice coverage is:

  • aice=0 if there is no ice
  • aice=1 if there is no open water
  • 0< aice <1 if there is both ice and open water
Since the total concentration aice is a diagnostic variable in CICE, one would need to project the observations onto the five types of ice concentrations. An implementation in which the observed ice fraction was decomposed into the five prognostic types of ice fraction using the same ratios as the background was found to be numerically unstable most of the time. Instead, T and S increments are computed using the covariances between aice, T and S from the static ensemble. Again, to account for the small size of the ensemble, and the resulting errors in the estimation of far away correlations, a localization procedure is used. The parameters of the localization are given in line 8 of Table 1. The observation error is prescribed as 0.05.