hotfix to assign all missing QualityMarker values to 15

[guoqing-noaa]

This PR is to address the real-time MPAS model crash issue due to unreasonably large increments.
It is caused by observations with missing QualityMarker values, which can not be handled by our current rrfs-workflow MPASJEDI settings.

We will try to make a permanent solution, but it will take a while.

For now, we will put all missing QualityMarker values to 15 so they can be filtered out by specific obs filters.

The added Python script is based on offline_add_var_to_ioda.py developed by Donnie.

Thanks @chunhuazhou, @spanNOAA and @hu5970 for quickly identifying the cause.

[delippi]

I see now, the missing QMs are causing those unreasonably large increments. Nice job finding the problem.

I also preferred the word "modify" over "tweak" in the name but not such a big deal.

Can someone send me the path to a run which has the missing QualityMarks for the data? I would like to double check the yaml and also try some modifications that I have in mind for a more robust fix. @guoqing-noaa @chunhuazhou @HuiLiu-NOAA

[HuiLiu-NOAA]

@guoqing-noaa : This can be fixed easily in prepbufr2ioda.yaml by setting the QualityMarker to 15 as default value.

[guoqing-noaa]

@guoqing-noaa : This can be fixed easily in prepbufr2ioda.yaml by setting the QualityMarker to 15 as default value.

@HuiLiu-NOAA Could you show which line we should change?

rrfs-workflow/parm/prepbufr_aircft.yaml

Line 6 in c31d81e

observations:

[chunhuazhou]

@delippi You can find the ioda_aircar.nc with missing QC on jet at /lfs6/BMC/wrfruc/Chunhua.Zhou/data/saved_jedivar_cases/MPAS_conus12km/2025022701/jedivar_01.3DVar.newQC/data/obs