review comments

[anton-seaice]

No description provided.

[DeniseWorthen]

For the UFS side, I think the original placement of the code block in ice_comp_nuopc inside the CESMCOUPLED ifdef would also work, at least for how we intend to use these pio options at this point (via namelist). I think it does no harm (other than generating a warning message) where it is now, but I'm also fine if you revert to the original placement.

One question...the code includes the comment

! The only reason to set these is to detect in ice_history_write if the chunk/deflate settings are ok.

To me, this implies that the same type of checking might not happen if the pio options are set via namelist?

[anton-seaice]

For the UFS side, I think the original placement of the code block in ice_comp_nuopc inside the CESMCOUPLED ifdef would also work, at least for how we intend to use these pio options at this point (via namelist). I think it does no harm (other than generating a warning message) where it is now, but I'm also fine if you revert to the original placement.

Ah - I didn't think this through. It is in the right place without this PR.

• If CESMCOUPLED is set, then the iotype comes through NUOPC & CESM shared routines and we need to set the namelist _format options in the driver.
• If CESMCOUPLED is not set, then format comes through the namelist and is handled in ice_init in the same way as the standalone configuration.

One question...the code includes the comment

! The only reason to set these is to detect in ice_history_write if the chunk/deflate settings are ok.

To me, this implies that the same type of checking might not happen if the pio options are set via namelist?

Its about lines like this:

if (restart_format=='hdf5' .and. restart_deflate/=0) then
         status = pio_def_var_deflate(File, vardesc, shuffle=0, deflate=1, deflate_level=restart_deflate)

in ice_history_write.F90 and ice_restart.F90. If the format is set in the namelist then that if statement works fine too. I will try and reword. Maybe something like 'CICE history and restart handling needs these set to enable chunking / deflate'

[apcraig]

@anton-seaice, please close or revise this PR as needed. Let me know when it's ready.

[anton-seaice]

@DeniseWorthen , I assume the nuopc driver checks are ok without this change?

[DeniseWorthen]

@apcraig Yes, thanks. I've been testing w/o the change in ice_comp_nuopc.F90