Remove Unneeded RRFSv1 output fields

[EricJames-NOAA]

In collaboration with EMC, GSL would like to get the RRFSv1 output fields organized a bit better into the different types of GRIB2 files.

If it is possible, we would like to distribute a 2-D GRIB2 file to downstream users, which contains ONLY 2-D fields. If this is allowed, we may be able to remove all 2-D fields from the prslev and natlev files.

If it is not possible to distribute a 2-D GRIB2 file, we propose moving all 2-D fields into the prslev files, and removing them all from the natlev files. This way, downstream users who are only interested in 2-D fields will know which file type to grab.

For reference, we distribute wrfsfc files from HRRR which contain only 2-D fields. The HRRR wrfnat and wrfprs files both have all the 2-D fields appended to them.

Input is welcome from the UPP team and EMC!

[BenjaminBlake-NOAA]

@EricJames-NOAA Thanks for opening this issue! Let's definitely get this sorted out before code delivery to NCO.

As for the possibility of distributing a 2-D GRIB2 file, I will defer to @MatthewPyle-NOAA on whether that is possible at this point. If not, I agree that we should move all 2-D fields into the prslev files and remove them from the natlev files.

Also, I am pretty sure there are some 2-D fields which are duplicated in both the prslev and natlev files, so we should remove those from the natlev files.

[MatthewPyle-NOAA]

I feel like the NATLEV files are largely used for initializing other (MPAS) model runs - do they also utilize any 2D fields? I'm reluctant to make any changes to the 1.0 system at this point, but we could consider it for a v1.1ish upgrade. How many fields are duplicated between the PRSLEV and NATLEV files?

[BenjaminBlake-NOAA]

I'm not sure exactly what systems are currently using the natlev files for initialization, but the natlev files do contain a large number of 2-D fields, some of which are not present in the prslev files. The prslev files also have some 2-D fields that are not available in the natlev files.

By my rough estimate there are about 306 duplicate 2-D fields that are present in both the prslev and natlev files. So removing them all from the natlev files would be a fairly significant change. I agree with @MatthewPyle-NOAA that I am a bit reluctant to do such a large change for RRFSv1 this late in the process, but this could certainly be done as part of the RRFSv1.1 upgrade. @EricJames-NOAA what are your thoughts?

I did find one problematic field in the prslev output - VVEL at hybrid level 1. VVEL has been replaced by DZDT so this one should be removed.

[EricJames-NOAA]

@MatthewPyle-NOAA good point on initializing downstream model runs. So I guess we will need to keep at least some 2-D fields in the natlev files. Maybe an alternate approach would be to have all 2-D fields in the natlev files, and none in the prslev files.

As far as distributing an additional 2-D GRIB2 file, is that a possibility for RRFSv1.1? Or not possible at all? I think it would be very helpful for downstream users who have bandwidth challenges.

[BenjaminBlake-NOAA]

@EricJames-NOAA I met with Jason Levit (EMC's VPPPG branch chief) yesterday, and he noted that it is too late in the process to make any product changes for RRFSv1. So we will have to leave the prslev and natlev files as is for the initial implementation - unless we discover any fields that shouldn't be in there, like hybrid level 1 VVEL.

I think distributing an additional 2-D GRIB2 file could be considered for RRFSv1.1, and it would be helpful for downstream users. Perhaps at that time we could also consolidate all of the 2-D fields into prslev or natlev (such as putting them all in natlev like you suggested). That would significantly decrease the size of the prslev files. Will defer to @MatthewPyle-NOAA for a final decision on this.

[MatthewPyle-NOAA]

I did add this item on to the ever growing v1.x request list.

[BenjaminBlake-NOAA]

@EricJames-NOAA Since this product update request will be part of RRFSv1.1, @WenMeng-NOAA and I suggest that you add this information to issue #1162 and close this issue.

[EricJames-NOAA]

@BenjaminBlake-NOAA would you be able to handle removing the VVEL on hybrid level 1?

[BenjaminBlake-NOAA]

@EricJames-NOAA Sure I can definitely do that. Within the next few weeks (before code delivery to NCO) I will do a final check of all the RRFS GRIB2 output to see if there are any other fields that should be removed, and I can remove VVEL then.

Let's leave this UPP issue open so I don't forget about this, and rename it to "Remove Unneeded RRFSv1 Output Fields".