New GFS/FV3 model background files (C12) containing all hydrometeors and corresponding AMSU-A (n19) observation file

[gthompsnJCSDA]

Description

To provide for more thorough testing of code that calculates effective radii, we need GFS/FV3 background files with all hydrometeors (rain, snow, graupel, cloud water, and cloud ice) found in the "tracer" file. Also, we can use a AMSU-A (NOAA-19) observation file that has been stripped down to channels 1-5, 15 which are the most sensitive to the presence of cloud and precipitation. The new AMSU-A file contains only those 6 channels and was subsetted in Location dimension by striding every 35 points out of its original 81000 points, making the final vector length 2315.

Issue(s) addressed

See the related one in UFO
UFO 2631

Acceptance Criteria (Definition of Done)

Positive reviews

Dependencies

Coordinate together with fv3-jedi PR 764

Impact

None expected.

[climbfuji]

@cmgas I think those might files overlap with what I gave you. @gthompsnJCSDA are these warm start files (it looks like to me), and if yes, what is the cold start date and what are the physics options used?

[danholdaway]

Good idea to reflect what is in these files in practice.

[gthompsnJCSDA]

@cmgas I think those might files overlap with what I gave you. @gthompsnJCSDA are these warm start files (it looks like to me), and if yes, what is the cold start date and what are the physics options used?

@cmgas @climbfuji Yes I copied something from r2d2 archive on Orion and I am attempting to reduce it data volume from 81000(Location) x 15(Channel) down to 2315x6 so it occupies less space. I am having massive struggles with the decimated obs file working in my CI-tests for the corresponding PR in fv3-jedi due to issues with _FillValue. Hopefully I can fix that problem quickly.

[cmgas]

@gthompsnJCSDA how did you produce these gfs files?
Some of it overlaps yes, @climbfuji but these wouldn't be sufficient to run UFS.

[gthompsnJCSDA]

@gthompsnJCSDA how did you produce these gfs files? Some of it overlaps yes, @climbfuji but these wouldn't be sufficient to run UFS.

They were in either your directory or the main r2d2 archive directory (on Orion) that Fabio sent to me. If you want, I can run FV3-GFS (prototype v16) simulation at C768L127 for you. I have run it multiple times myself.

[cmgas]

@gthompsnJCSDA we will likely change all the backgrounds later to have one date with all existing variables (will help future development).
In the meantime though we are making an effort to get rid of all colons in names, can you name these files so they are consistent with other fv3-jedi-data background files?