CICE error: ice: init_grid: ANGLE out of expected range

[ddai]

Hi all,

I am trying to do a basic control run with the following configuration:
Version - cesm2.1.3
Compset - 1850_CAM40_CLM40%SP_CICE_DOCN%SOM_MOSART_SGLC_SWAV
Grid - f19_g16

Nothing is changed after creating the case, apart from setting the XML variable DOCN_SOM_FILENAME = pop_frc.b.c40.B1850CN.f19_g16.100105.nc.

The model builds and submits without errors. However, the run fails on the ICE initialisation step with the following error

ice: init_grid: ANGLE out of expected range

I have seen a similar error in another thread here but it doesn't seem like they are the same. Is there something that I am missing? Any help on this is greatly appreciated!

 

[dbailey]

What is your initial condition for the sea ice? Sounds like you have missing values in your grid variables. Also, are you using g16? In this version of the code it should be g17.

 

[ddai]

Thank you for your swift reply!
The initial conditions are unchanged and set up automatically. I am indeed trying to use g16 here, and did try a run with g17 which succeeded. The g17 run yielded a snowball Earth which I thought was strange given that it is just a control run and I was expecting roughly a steady climate, similar to outputs I've gotten from other control runs with different grid resolutions. However given the boundary condition file in my original message I would have thought that g16 would be the right grid since it's in the name of the BC file? Is there any documentation explaining why we should use g17 with the configuration?

 

[dbailey]

Not really. The main change is that the Caspian Sea became land with the change to g17. This should not be an issue with using a g16 as an initial file as the data in the Caspian are ignored. I think the problem is missing values in grid variables. Now, I see you are using CAM4.0 ... that might be the issue with the snowball Earth. I would check in the CAM forum.

 

[ddai]

I see, thank you, it's good to have the confusion around the grid cleared up. We will be exploring some other configurations to see if the problem occurs again and will ask in the CAM forum if it does.