Fix reading of mixed layer forcing file. #578
Conversation
There was a problem hiding this comment.
In ice_forcing.F90, the currents might still be too strong, but you can try it like this.
Instead of moving init_forcing_ocn later in the subroutine, would it work to move init_forcing_atmo up? sst is needed in init_state.
Doesn't ocn_data_dir need to be generic in set_nml.ml, as for other (atmo) forcing?
|
Very good points. I'm actually now thinking I could move the call to allocate the arrays to init_forcing_ocn instead and leave the calls in the current location in CICE_InitMod.F90. You are right about ocn_data_dir as well. |
|
Sorry for the repeated commits/pushes. I had troubles with conflicts from the new time manager stuff. |
|
Looks like I still have some issues. |
|
@dabail10 since you've gotten some runs completed, have the issues mentioned above been fixed? |
|
I think the code is fixed, but we don't have the forcing file on Zenodo yet. |
|
I'm confused, why was this closed and not merged? |
|
Not sure. I think it needs the forcing file? |
|
@dabail10, it looks like you closed this. we can reopen it. I thought this had to be merged regardless of whether there is an available file on zenodo. |
|
I can reopen. Not sure what happened. |
|
I see what happened. I thought it was merged and I deleted the branch. My bad. |
| ocn_data_type = 'ncar' | ||
| ocn_data_format = 'nc' | ||
| ocn_data_dir = 'ICE_MACHINE_INPUTDATA/CICE_data/forcing/gx1' | ||
| oceanmixed_file = 'pop_frc.gx1.nc' |
There was a problem hiding this comment.
Need have a standard name for the oceanmixed_file and then commit to Zenodo.
There was a problem hiding this comment.
How about ocean_forcing_gx1.nc?
There was a problem hiding this comment.
It should be named similarly to the atmospheric forcing, but that is not standardized. It could follow the JRA tx1 example,
CESM_ocean_monthly_forcing_gx1.nc -- probably overkill since its location in the forcing directory structure would identify it: CICE_data/forcing/gx1/CESM/MONTHLY/
So ocean_forcing_gx1.nc works.
There was a problem hiding this comment.
@eclare108213 mentioned having 2D in the name as well. Then if we put it under the forcing/gx1 area, we shouldn't need the "gx1" label.
There was a problem hiding this comment.
I think we want a filename that doesn't rely on it's directory location to describe it. We also don't want the same filename in multiple directories. So we want to avoid having multiple "ocean_forcing.nc" files in multiple directories. When the file is separated from it's directory, it should be unique (relative to other input data files) and still somewhat descriptive. I am completely fine with something like "CESM_ocean_monthly_climatology_forcing_gx1_210324.nc". But it need not be that descriptive. It's always a balance. When/if we update this file in the future, we'll need to give it a new name as well. We need to make sure we never have multiple versions of the same filename, ever.
There was a problem hiding this comment.
I've added the Kay et al. 2015 reference to the CESM-LE in the metadata. I think the metadata can be very descriptive as well as the Zenodo entry. I believe we should keep the filename simple.
|
I would go with more descriptive rather than less.
|
|
Yes, the metadata will have essential info, so we don’t need that kind of distinguishing label on the file name. I still prefer including gx1, and 2D is helpful. So maybe ocean_forcing_clim_2D_gx1.nc?
|
|
I like that. I will fix the PR. |
|
Updated set_nml.ml option. The forcing file has been added to Zenodo at 10.5281/zenodo.4633898. |
|
The new forcing file and location needs to be added to the input data wiki page. Otherwise this is ready. Thank you! |
|
The forcing file information was added to the input data page. |
For detailed information about submitting Pull Requests (PRs) to the CICE-Consortium,
please refer to: https://github.com/CICE-Consortium/About-Us/wiki/Resource-Index#information-for-developers
PR checklist
This fixes a bug where the mixed-layer forcing was not initialized correctly as the array was not allocated.
dabail10 (D. Bailey)
https://github.com/CICE-Consortium/Test-Results/wiki/cice_by_mach_forks#cheyenne
I have added a new test, set_nml.ml that turns on oceanmixed_ice with a SOM forcing file from a CESM run. The main change is to move the call to init_forcing_ocn in CICE_InitMod.F90. Also, the default has changed to use a 2D forcing file with everything at the gridcell centers. I've also added some helpful extra diagnostic information in the netCDF read routines.