Using history files as ice initial condition

[Camille013]

Hi all,

I am trying to use a time-averaged .nc history file as an initial condition for a fully coupled (B1850 compset) CESM1 run. I set the namelist variable ice_ic to my file, but the model crashes during initialization without any error in the ice log file. In the ice log it just ends on the line: "Using restart dump= <my file>" Any thoughts on how to correctly initialize from my own nc file?

Camille

 

[dbailey]

Just talked to Keith Lindsay, and we don't have an easy way to initialize the CICE model from history files. The issue is enthalpy / energy. This was likely not saved in the history files and is a 4D field. I think you are better off using some other sea ice restart file from an existing B1850 run, or potentially just setting ice_ic to 'default' which is a simple internal initialization of the sea ice. Check the cesm log for other errors related to the sea ice. If this is a thermodynamic error, then something about your restart is not compatible.

 

[Camille013]

Got it, thanks. Just so I understand correctly, doing the internal sea ice initialization (i.e. setting ice_ic='default') would essentially initialize ice based on the SST initial condition? Trying to understand this because I believe I did such an initialization with my very warm (8xCO2) ocean initial condition, but it still initialized with a lot of sea ice.

Thank you again!

 

[dbailey]

You might need a code fix. The issue is the SST is not computed from the ocean before the sea ice is initialized. However, if you use Tair instead, this is better.