Ensure consistent use of orography files across the workflow.

[ClaraDraper-NOAA]

Describe the issue.

The fix files used by global_workflow are in the process of being updated. This makes this a good time update the orography files that we're using to the latest ones, and to make sure that we're using the same files everywhere, and that we have checks built in to catch inconsistencies in the files.

Currently, we're using these files: /scratch1/NCEPDEV/stmp2/Michael.Barlage/models/vector/v2t_data/tile_files/C96.mx100_frac/oro_C96.mx100.tile*

(note the non-standard file name, this is currently hard coded in vector2tile)

The orog files need to be specified in:

• tile2vector namelist (read in by tile2vector)
• jedi yaml files (read in by fv3-jedi)
• the namelist for the ufs driver also specifies the vector length.
• In https://github.com/barlage/fv3_mapping to calculate the mapping indexes for IMS (done offline from workflow)

Checks in the code: vector2tile checks that the input vector length is the same as the orog files for the vector2tile direction only. This will catch inconsistencies between the specified orog file and that used in the ufs model driver (assuming number of land tiles is an appropriate measure of differences). For completeness we should add a check for the tile2vector direction. JEDI will only catch an error in the resolution of the specified fix file ( I think), and we are using the land mask and elevation info in JEDI.

Required actions: -confirm that we can update the ufs model driver to use the latest orography files from the new fix directory. -update other components to use the latest orography files -add check to vector2tile on the vector length for the tile2vector case, and change code to read in standard file name -clean up use of orography info in fv3_mapping. I think it can be removed so that the code is only reading in the lat/lon info for the grid.

Test output

Will addressing this issue change the DA_IMS_test output? If not, what are the expected changes?

[ClaraDraper-NOAA]

@barlage I had a look through the workflow to see how we're handling the fix/orography files. If you can update the UFS model driver to the latest orography files (or confirm that we're already using them), I'll take care of the rest of what needs to be done. I will also change the workflow to add a link to the official fix files, and then specify this location for all times we need the orography files.

[ClaraDraper-NOAA]

@barlage Jeff's replay uses the mx100 files (for coupled ocean/ice), but in other instances we'll want the non-coupled grid. This means we're going to have to maintain flexibility to choose the orog files at run time. I'm planning to change the code so that: -the cycling script stages the orog files with a generic name at run time -each program uses the staged files -we maintain mapping indexes for all options, and chose the appropriate on at run time.

[ClaraDraper-NOAA]

Closing this issue. I've updated the code so that the orography directory and filename are now specified in the scripts, and are then written into the namelists (for vector2tile and JEDI). Note that the model does not directly use the orography files, and takes the number of land grids as an input. There are currently no checks to ensure that this is consistent.