Yannick opened the meeting. No general announcements

INTERFACE - Francois H

  • UFO
    - Integrate code from NCAR into obs2ioda converter to compress file sizes.
    - Raji working to remove QC flag from Linear simulate AD and TL functions.
    IODA
    - Steve H and V working on a Writer for OSDF ObsSpaces, which is a big fraction of the remaining work toward functional parity with ObsGroup-based ObsSpaces. Should be done in weeks ahead.
    - the ODB reader for OSDF-based ObsSpaces, from the Met Office, has been merged in.
    - Anna (with Steve H and V) has implemented a "redistribute" function for ObsSpaces using the (beta) OSDF container, which allows the MPI distribution to change in-memory.

INFRA - Evan

  • We have a release branch for spack-stack-2.1 ready for installation and testing
    • 2.1 includes incremental changes from the 2.0 release
      • Includes, but not limited to, upgrades for pnetcdf, gsibec, pyh5, etc.
    • The NEPTUNE environments have been successfully deployed with 2.1

ALGO - Yannick

  • Work in progress with the adoption of the new version of gsibec

OBS - Francois V

  • ioda converters

    • Tidy up time windowing with Madrigal TEC converter, and add sequenceNumber to collect vTEC and sTEC (PR1742 and PR1740).

  • UFO
    • 3 PRs related to RTTOV: 
      • Fix for unphysical cloud top pressure passed to RTTOV #4008
      • Add T and Q Jacobian in rttovcpp_setYdiag (in RTTOV 14 CPP interface) #3997
      • Reduce verbosity of RTTOV messages #3996
    • New ctest for Hubernorm functionality #4005. Need review and also the associated jedi-docs PR #984.
  • Technical Debt
    • Removing QC_flag in Simobs TL/AD functions #4006
    • Fixing the yaml files failing Yaml validation #3995
  • Space Weather

    • Vertical integration operator refactor PR merged UFO#3961


  • ObsProc Meeting - 3 presentations
    • RO operator comparison at EMC

    • upcoming bug fixes in the NOAA/CU radar operator

    • AtmVertInterpLay operator phase-out discussion

Ricardo asked about the impact of changes related to an operator for Ozone. A discussion ensued, and it was decided to continue this discussion off-line.

CRTM - Cheng

  • Working systematically on the upcoming CRTM v3.2 release, including scientific updates to the TL/AD code, as well as structural code improvements, including making NetCDF the default data format.
  • A feature branch will be created soon for testing purposes. We will coordinate closely with the OBS team and external partners before merging any changes into develop that could potentially introduce breaking changes.
  • Instrument coefficient testing is ongoing for TIO and PolSIR.

EMC - Dan (No update)

GMAO - Ricardo

  • Running comparison tests between GSI and JEDI
    • Working with EMC
    • New gisbec version
    • Updates to GSI
    • QC updates in UFO (eg, satwind)
    • Discovered a grid discrepancy issue with some BUFR converters
  • Testing GEOS LETKF and seeing some unexpected results. Trying single obs tests for now.
  • Bryan K added that he is working with BenR on a ioda converter for MTG IRS (hopefully I've got this right)

NRL - Sarah, Dom

  • NRL is performing in annual program reviews
  • Testing high altitude 4DVar
  • Investigating unexpected results with some satellite instruments
  • JEDI components required for NEPTUNE are now in spack-stack-2.1.0
    • oops, ioda, ioda-converters, ropp-ufo, ufo.
  • The crtm-fix package needs updating: the flattening of the CRTM coefficient files, introduced by EMC, needs to be reverted so that it works with JEDI. We will temporarily create symbolic links to mimic the flattened directory structure, but these will be deprecated after some time (end of 2026). Partners who rely on the flat directory structure need to update and should start now.

MMM - BJ

  • Junmei Ban has shared her findings on simulating active sensors (via CRTM) with Isaac Moradi, Ben Ruston, Ben Johnson. Waiting for their response.
  • Zhuming Ying is evaluating the impact of assimilating IASI. See some degradation from the temperature sounding channels. Looking at channel selection.
  • There is an effort to build the MPAS-Model as a Cmake ExternalProject (in progress). This is supposed to reduce the burden of maintaining the CMake build mechanism (from the MPAS-Model). 

UKMO - Mike C

  • Work is ongoing for UFO output cleanup
  • RTTOV 14 is close
  • Testing 4DVar configurations

ROMS - Hernan

  • Working on 4DVar application
  • Made some great progress with file I/O compression
    • Using PIO
    • Using netcdf built-in compression (zlib)
    • Seeing up to %60 reduction

Francois H commented that a group is investigating hdf5 plugins that enable newer compression schemes (LZ4, BLOSC) compared to the built-in scheme (zlib). The new schemes offer good compression with much less CPU runtime (applying the compression). Decided to continue this discussion off-line.

Maryam announced that her group just finished a composition oriented JEDI training, and she is willing to share the tutorials if anyone is interested. Tom offered to follow up with Maryam off-line.

  • No labels