Yannick opened the meeting with no announcements and moved on to the general updates from the CSDA teams

At the end of the meeting Yannick made an announcement that the JCSDA Science Workshop will take place next week (May 16 - May 18) in Boulder, and because of this we will not have the weekly (special topics) meeting next Thursday (May 18).

INFRA

Dom gave the following updates:

  • Logging has been added to R2D2
    • INFO level, which is on by default, gives concise messages such as status
    • DEBUG level, which is off by default, gives verbose debug messages
  • We performed some github cleanup
    • Archived old repos
    • Removed obsolete/unused repos
    • Changed access permissions so that only the organization level admins (Tom, Yannick, Dom) can create new repos
      • We need to be more deliberate about creating new repos moving forward
      • Please put in requests for new repo creation to the organization level admins
  • Spack-stack
    • Just completed a massive sync of the spack fork with the authoritative spack repo
      • Picked up many bug fixes, improvements, etc.
      • Facilitates with our submitting PRs to the authoritative spack repo
    • New spack-stack builds are available on Neptune, Hera and Gaea C5
    • Our spack-stack and fork of spack repos are moving from the NOAA-EMC organization to the JCSDA public organization tomorrow (May 12, 2023)
  • The IODA Fortan API is ready for testing
    • Patrick and Steve H will help
  • Good progress is being made with the IODA reader improvements
    • We are coordinating this development with EMC and the Met Office
  • Good progress is being made with cloud services (running EWOK experiments) and the CI pipeline
  • Orion has been horrendously slow this week
    • MSU IT is working on it, but we don't have a target date when these issues will be addressed

ALGO

  • Anna gave the following update
    • General bug fixes and clean up have been issued on behalf of running EWOK experiments
    • Good progress with the low resolution experiments
      • Using Feb, 2023 dates
      • For demo and tutorial purposes
      • Finding a balance between lower resolution to run fast with high enough resolution that the results show similar patterns compared with the high resolution experiments
    • Hailing is making good progress on a c96 resolution experiment for a tutorial scheduled for this June
    • Yannick has an FGAT PR in review that makes some helpful adjustment (such as the temporal location of the forecast)
    • VADAR is making good progress
      • Cleanup, bug fixes
      • Next step is to work on interface updates/improvements
    • Recent BUMP/SABER updates
    • Francois H has PRs for ATLAS and ECKIT (in the ECMWF repos) that are under review
      • Once these get merged and tagged, we will need to update the spack-stack versions to these new tag versions
    • Investigating using ATLAS unstructured grids for the interpolator when using a point cloud

OBS

  • Francois V gave the following update
    • CAPS(?) team is making good progress with ground based radar obs operator development
    • NRL is working on the ingest of single RO WMO bufr file data
    • Good progress with the NCEPLIBS-bufr based converter
      • Emily add ATMS updates
      • Recent PR got merged that properly enforces the C++14 standard
      • SSMI converter is work in progress

CRTM

  • No update from the team, but Dom mentioned that we are having issues with GDEX downloads
    • For the short term we will start using an S4 ftp site
    • For the long term we will switch over to using an AWS S3 bucket

COMPO

  • Finished the upgrades on IODA format for trace gas, all repos a and databases are updated
  • Reminded the group that ozone retrievals that uses amtvertinterplay could be upgraded to use columnretrieval
  • Cleaned up aero test in fv3-jedi, removed 1/2 of the tests
  • Finished all PR and ingest for skylab small aro and trace gas, they now need to be merged or ingested
  • Still working on cycling GEOS in ewok for a 3DVar
  • Helping GMAO to better linearize their AOD operator and have a log transform in the non-linear operator
  • Currently working on a generic interface to produce H(x) using atlas geometries

SOCA

  • Working with Dom and Anna to get SOCA using the proper version of UFS
    • Made a lot of progress but still have remaining issue with an adjoint related feature
  • No labels