The main focus of the project will be on improving the forecast initialization. The reason is that the forecast quality of the neXtSIM-F forecasting platform (which is running operationally as of October 2015) is based on the quality of the three core components. These components are:

1. Model: The sea-ice model itself, which is named neXtSIM

2. Forcing: Atmospheric and oceanic forcing data taken from ECMWF and TOPAZ

3. Initialization: Estimation of the sea-ice conditions at the beginning of the forecast

Components 1 and 2 are difficult to improve substantially, because both the model and the forcing are the best currently available. In contrast, the initial state estimation can be improved substantially. As the thickness forecast is especially sensitive to the initial estimation, the thickness forecast would benefit greatly from an improved initialization. This is also true for drift and deformation forecast, which would benefit from improved damage initialization.

We will improve the initialization by introducing three additional data sources to the forecast platform in addition to the AMSR2 sea ice concentration and SMOS thin sea ice thickness dataset we are currently working on for the current phase of the project. These will be the real- time OSISAF ice concentration provided by MET Norway (http://osisaf.met.no/p/ice/), the ice thickness derived from CryoSat provided in near-real time by UCL-CPOM (http://www.epom.ucl.ac.ulc/csopr/seaice.html , and the ice damage derived from satellite monitored ice deformation.