4.6. Data Assimilation

Data assimilation (DA) is the scientific process of combining external data with numerical model forecasts. There are several ways this can be done including by adjusting the model initial conditions (internally or externally) or adjusting the model solution as it evolves in time. Various data assimilation options are being introduced in CICE and are described below.

4.6.1. Data Assimilation on restart

The namelist variable, restart_mod, specifies the restart DA mode. By default, this namelist value is set to none which disables the feature. The current active options are adjust_aice and adjust_aice_test.

With adjust_aice and adjust_aice_test, the category averaged aice value is modified at restart to specified values using the method implemented in cicecore/cicedyn/infrastructure/ice_restart_driver.F90 subroutine direct_adjust_aice. This method adjusts aice, vice, vsno, qice, and sice in all categories to be consistent with the category average aice specified. It also adjusts several thermodynamic variables such as temperature and salinity (see [46]). adjust_aice reads in a sea ice concentration field from an external file. The field is currently hardwired to ‘sic’ and the file is currently hardwired to ‘sic.nc’. The field must be on the model grid. adjust_aice_test modifies the aice field read on restart internally. The current implementation rounds the aice values read at restart to the nearest 1/100th. This mode exists primarily to test the feature.