New GSI fix files: netcdf global_berror and l31 files

[RussTreadon-NOAA]

Type of static data

Fix

Target directory

Fix: gsi

Type of change

• Update existing files (Version update needed)
• Remove files (Version update needed)
• Add new files to existing directory

Any workflow changes needed?

No change needed

Please list all related issues. If this request requires a workflow update (including a fix version update), please open a separate global-workflow issue to track the change.

None

Pickup location

Hera: /scratch1/NCEPDEV/da/Russ.Treadon/git/gsi/20241127

Size

134.1 MB

Additional information

The files in /scratch1/NCEPDEV/da/Russ.Treadon/git/gsi/20241127 serve two purposes:

1. netcdf format global_berror files to be used for GSI -vs JEDI DA v18 DA validation. The netcdf global_berror files may also be implemented as part of GFS v17
2. Fix files to run ultra low-resolution GSI with 31 levels

None of the files in /scratch1/NCEPDEV/da/Russ.Treadon/git/gsi/20241127 are currently in /scratch1/NCEPDEV/global/glopara/fix/gsi/20241022 so the files in 20241127 could be added to /scratch1/NCEPDEV/global/glopara/fix/gsi/20241022 without changing the version number. Of course, this is EIB's call.

[KateFriedman-NOAA]

None of the files in /scratch1/NCEPDEV/da/Russ.Treadon/git/gsi/20241127 are currently in /scratch1/NCEPDEV/global/glopara/fix/gsi/20241022 so the files in 20241127 could be added to /scratch1/NCEPDEV/global/glopara/fix/gsi/20241022 without changing the version number. Of course, this is EIB's call.

@RussTreadon-NOAA Does this mean that the files that are in the current 20241022 set but not in the new 20241127 set are being retired and will not be used again? While combining would indeed spare us a new timestamp, I am inclined to do a new timestamp so we have a fresh set that doesn't contain files being retired (if they are being retired). Let me know, thanks!

[RussTreadon-NOAA]

@KateFriedman-NOAA , no GSI fix files are being retired. We need to retain the binary global_berror files for the foreseeable future. This request simply adds netcdf counterparts to some, not all, global_berrror files. 31 layer files are also added to support future ultra low resolution g-w CI.

[KateFriedman-NOAA]

@RussTreadon-NOAA Understood, thanks for explaining that the current files should remain. Given this, I will add the new files to the current timestamp. Stay tuned...

[KateFriedman-NOAA]

@RussTreadon-NOAA I have pulled the new GSI fix files into the 20241022 timestamp on Hera. Please confirm they are good in their new home and we will sync them to the other platforms. Thanks!

[role.glopara@hfe09 20241022]$ pwd
/scratch1/NCEPDEV/global/glopara/fix/gsi/20241022
[role.glopara@hfe09 20241022]$ rsync -azv /scratch1/NCEPDEV/da/Russ.Treadon/git/gsi/20241127/* .
sending incremental file list
global_anavinfo.l31.txt
global_berror.l127y194.nc
global_berror.l127y386.nc
global_berror.l127y770.nc
global_berror.l127y98.nc
global_berror.l31y26.nc
global_berror.l31y50.nc
global_hybens_info.l31.txt

sent 103,644,614 bytes  received 168 bytes  15,945,351.08 bytes/sec
total size is 132,374,049  speedup is 1.28

[RussTreadon-NOAA]

Thank you @KateFriedman-NOAA for installing the new fix files on Hera. The fix files are in the correct location. I built and ran the GSI global_4denvar ctest using a netcdf global_berror file copied from the g-w GSI fix directory. Everything looks good!

[KateFriedman-NOAA]

Thanks for checking @RussTreadon-NOAA! I have synced the new gsi fix files to WCOSS2, Jet, and AWS. Working on HPSS now.

@WalterKolczynski-NOAA please sync gsi/20241022 to MSU, thanks!

[KateFriedman-NOAA]

The gsi/20241022 fix is now also archived to HPSS.

[RussTreadon-NOAA]

Thank you @KateFriedman-NOAA . I built GSI on Dogwood using the updated fix and am now running ctests. I'll do the same on Orion and Hercules after the updated GSI fix is in place.

[WalterKolczynski-NOAA]

MSU synchronized

[KateFriedman-NOAA]

@RussTreadon-NOAA Once you've tested and confirmed the files are good on MSU, please respond with that and feel free to close this issue. Thanks!

[RussTreadon-NOAA]

Thank you @WalterKolczynski-NOAA . I see the files in the expected location. Hercules and Orion builds are underway.

[RussTreadon-NOAA]

GSI ctests on Orion, Hera, and WCOSS2 completed with acceptable results. Hercules jobs remain pending in the queue. Hercules and Orion use the same g-w GSI fix so the Orion success is sufficient.

[RussTreadon-NOAA]

Thank you @KateFriedman-NOAA and @WalterKolczynski-NOAA for installing the new files. I'm seeing acceptable behavior on several machines. Closing this issue.