Notes from JEDI meeting 5/25/2023


Yannick Opened with two Announcements:

  • After discussion at JCSDA Science Workshop, partners will now give updates at this JEDI meeting, in addition to the core team leads.
  • Code changes that have impact across repos will now be announced in a new way, via the "Discussions" tab on the "JCSDA-internal" organization page:


INFRA-Dom

  • Making progress on EWOK-CI
  • Spack stack 1.4.0 almost rolled out.
  • There's a new version of HDF5. Can’t use mpich older than 4.0.3
  • Theres a new Jedi-infra-support slack channel. If someone doesn't have access they can also use the new Discussions page (INFRA Category) on JCSDA-internal.


ALGO-Anna

  • Atlas & eckit changes are in ECMWF repos but require c++ 17, so need to wait for that rollout to JEDI repos.
  • Work in VADER will soon introduce a new model Traits class for ModelData. Watch for more information.
  • JEDI EDU was used and went really well. Positive comments from Peter Jan van Leeuwen.


COMPO-Jerome

  • Running hofx. Working! Unexpected.
  • Vertical levels are flipped. Need to work on that.


OBS-Hui

  • During testing found some filters need to be fixed so they can be used without bias correction
  • Merging some PRs. 
  • CRTM 2.4.1 now in develop.
  • IODA converters. Work with EMC and NCEP decoder.
  • Setting up Skylab experiments


CRTM-Ben J.

  • Working on CRTM v3 test in UFO. Getting close to rolling that out.


Marine-Kriti

  • No updates due to lots of team PTO


EMC-Cory


GMAO-Ron Gelaro

  • Work on Obs Configuration. Down to Marine Surface Wind. That finishes conventional obs.
  • Still struggling with 4D issues. Might be an incompatibility with the JEDI 4D cost function?
  • Working on Spack-stack compatibility
  • Getting Dom NCCS account.


Met Office-David S.

  • Very close to surface obs done on UFO
  • Marek going to work with Steve H. to improve IODA input performance
  • Marek working on hybrid covariance
  • Looking at problem with increment being off by a factor of 2. Anna investigated and thinks there's no issue in FV3. (Apparent "problem" there is because test was between two grid points.) Anna thinks issue is in Lfric-lite, so Met Office will keep investigating.


NRL-Sarah

  • Working on new obs. GNSSRO
  • Spinning up new HPC’s.
  • 3dvar fgat cycling test


NCAR-MMM-Jake

  • Working on All-sky ATMS assimilation.
  • Looking at bad q (moisture) analysis in stratosphere
  • Looking at CRTM problem with Jacobian magnitude


ROMS-Hernan

  • Recently presented on JEDI in Rome.
  • Dealing with ifort compiler issue with BUMP in their environment.
  • No labels