[Standard Update]: WMO filenames should have more flexibility

[MatthewPyle-NOAA]

What are the limitations of the current standard?

Filename format for files in the wmo sub-directory seems overly restrictive, and is tied to a legacy concept of grid numbers that really isn't a part of the GRIB2 standard.

What changes are being proposed to the current standard?

Generalize the concept of "dom" in these examples to allow for descriptive text (na, conus, ak, etc.) in addition to grid numbers. Only grid numbers are currently listed.

grib2.aqm.tHHz.08hr_o3.227
grib2.akrtma.tHHz.2dvaranl.198
grib2.sref.tHHz.spread.212

How do these changes improve the current standard?

It makes the standard more modern and forward looking, and adds flexibility for people submitting code.

What are potential impacts by changing this standard?

Different systems may have somewhat different wmo product file names (but this difference is subtle compared to the differences seen in current production)

How will this standard be enforced?

In the same way as the current standard, but just slightly more permissive on one aspect of it.

Additional context or notes (optional)

[MatthewPyle-NOAA]

@AndrewBenjamin-NOAA may have other comments about it

[AndrewBenjamin-NOAA]

If you take a look at ON388 Table B you'll notice a few things:

• There EMC had hundreds of grids, grids numbers can only go to 255 as a restriction of the grib2 format. Currently all of those numbers are taken.
• The NAMconusnest grid (used by REFS) isn't even listed here, I suspect for that reason.
• There is not be enough room to add in the North American RRFS grid.
• There may be more grids in operations not listed.

Retiring legacy models goes a long way in beginning the process to clean these grids up (and Table B). But retiring models doesn't equal removal of all its associated grids, because it is very likely stakeholders will ask for look-alike products on the same grid (FAA has done this for RRFS). There would also need to be some sort of legacy Table B where users would be able to look identify historical grids. All of that to say it is not a straight forward process to add grid numbers for RRFS, and therefore have grid numbers in the AWIPS filenames.

HiResW and HRRR and even a few RAP AWIPS filenames are examples of systems currently in operations that do not follow this standard, so there is some precedence for not having grid numbers in the filenames, especially given the restrictions above.

The RRFS and REFS are brand new systems, simplifying products and grids from legacy models. I think this a great opportunity to create a more flexible AWIPS filename convection as well, especially with momentum towards AWIPs in the cloud.

[JamesPolly-NCO]

We'll allow the use of letters in the domain descriptor.

Change to the standard will be some examples showing the use of letters as a domain description rather than numbers.

Does that work for @MatthewPyle-NOAA and @AndrewBenjamin-NOAA ?

[MatthewPyle-NOAA]

That works for me!

[AndrewBenjamin-NOAA]

That works for me as well.