Restart file regridding for GCHP

[yuanjianz]

Your name

Yuanjian Zhang

Your affiliation

WashU

What happened? What did you expect to happen?

In newer version of gcpy (1.7.1 as this issue is written), the restart file regridding changes the requirement for template file. Now it requires a file with the target resolution, which is inconvenient in my use case.

E.g. If we want to run a simulation in C40, we need a template file already in C40 resolution. But it is hard to generate one for regular users before the regridding actually happens.

Another bug is found for stretched grid restart file. To let MAPL recognize the restart file, some global attributes for netcdf is required:
E.g,

:STRETCH_FACTOR = 2.1f ;
:TARGET_LAT = 44.f ;
:TARGET_LON = 263.f ;

However, if using the default restart files in GEOSCHEM_RESTART as template files, these flags exist and not cleaned by the regridding:

:stretch_factor = 1. ;
:target_longitude = 170. ;
:target_latitude = -90. ;

which causes MAPL fails to recognize the restart file grid.

What are the steps to reproduce the bug?

Regrid restart files using gcpy

Please attach any relevant configuration and log files.

No response

What GCPy version were you using?

1.7.1

What environment were you running GCPy on?

Local cluster

What python environment were you using?

geoschem-gcpy in conda-forge

Will you be addressing this bug yourself?

Yes

Additional information

No response

[tkoskela]

I think I'm also encountering the MAPL bug this issue mentions. Is there any known work around for it?

[yantosca]

Tagging @1Dandan whom I believe made that change for the stretched-grid regridding.

[yuanjianz]

@tkoskela if you are running into the bug with stretched grid restart file, a simple workaround is to manually rewrite the global attributes of that file.

[yantosca]

@tkoskela @yuanjianz @lizziel @1Dandan, these tags:

stretch_factor = 1. ;
target_longitude = 170. ;
target_latitude = -90. ;

were updated in PR #381, I think for a new MAPL update if i remember properly.