Mike Bell & Jeff Ridley, Met Office

We propose to extend the ocean data assimilation scheme, used operationally on a daily basis at the Met Office, to include assimilation of sea-ice data. The data which will be given priority for assimilation are sea-ice concentration from AMSR and SSMI, and sea-ice velocity vectors from scatterometer instruments and SSMI. Sea-ice depth data will be a priority for assimilation when they have been validated. The value of surface temperature and snow depth data from AMSR for atmospheric models will be assessed. The processing of sea-ice concentration data and the representation of melt ponds in the sea-ice model need to be improved. At least one of these issues will be explored. An initial assimilation scheme building on the ocean assimilation code will be written. Problems with its performance and the impact of each data type will be assessed in a global model with 1/2o grid spacing and an Arctic model with a 1/6o grid. We will attempt to understand and solve the most serious problems with the initial assimilation scheme before performing a 12-month long demonstration assimilation The resulting sea-ice concentrations will be evaluated against in situ data, and by their impact on the assimilation of ATOVS data. The results from the demonstration will also be used to estimate statistics on the model error covariances and observation error variances.