Yannick: Big update this week is the release of JEDI-Skylab v2 last week: https://www.jcsda.org/jediskylab

ALGO

  • The UnstructuredInterpolator accuracy fix is expected to be merged on Friday October 21. See details in https://github.com/orgs/JCSDA-internal/teams/jedi-models/discussions/85 . Expect changes in references for all model interfaces that use UnstructuredInterpolator.
  • At the Monday Algorithms meeting there were two discussions: on obs perturbations for EDA, and time interpolation in 4DEnVar.
  • PR in oops (in review) updating how obs perturbations are done: changes in yamls + additional capability to perturb before filters.
  • Linear variable change capability in VADER is now available and used via fv3-jedi. VADER code sprint coming up the week of November 7th to populate VADER recipes and start using them in different model interfaces.
  • SABER blocks refactoring PR merged last week. Multiple updates to spectral covariance blocks, GSI covariance blocks, BUMP blocks (faster variance filtering).
  • Started discussions on GeoVals refactoring requirements.

INFRA

  • Improved ioda reader: issues will be available soon including descriptions of the approaches (Steve Herbener).
  • ioda-converters updates; PR to make converters installable via pip
  • Work started on mysql implementation of r2d2 (changes coming later in the quarter) (Eric Lingerfelt)
  • Work on containers; looking into issues with singularity containers and graphical tools for jedi-edu. (Maryam Abdi-Oskouei)
  • Solving issues with building spack stack on narwhal.
  • Progress on using ufs-bundle to run ufs within jedi, currently updating the code to work with the newest stack and code.

CRTM

  • CRTM transitioning to v3 repo, fully public and will start shallow (without 2.x history, should be much faster to clone). v3 will be mostly backwards compatible to v2.4, in particular in interfaces. First release of v3.0 will have limited capabilities in terms of polarization. Aiming for December release.
  • Merging radar simulator for space-based simulator (Ben Johnson and Isaac Moradi). A few weeks before merge, but part of v3.0.
  • Working on melting layer addition (Ben J).

OBS

  • Looking into consolidating ioda converters.
  • Observations in Skylab. Skylab v2 additional obs: synop, metar, iasi, cris, ascat. Skylab v3: new study period 15 Feb - 15 Mar 2022. New data planned for v3: mhs, goes, tempest, cowvr, tropics, ship, buoy, windborne data.
  • Plan to look in more detail into VarBC for clear sky and start looking into VarBC for all sky. Also looking in more detail into correlated obs error.
  • Talked to Clara Draper about IMS snow product (snow cover) – will have better snow obs.
  • Hui Shao is working on backgrounds and ensembles for the next testing period

SOCA

  • Working on tuning letkf, towards hybrid letkf and envar.
  • PR open for OASIM. EMC will take a lead on taking care of technical issues. We’ll have an obs operator for ocean color!

COMPO

  • Gfs aerosols ensembles ready for R2D2.
  • TROPESS CO converter almost ready.
  • Started looking into 3DVar runs with ewok for gfs aerosols for the next release. Next step is looking into EDA runs with gfs aerosols within ewok.


EMC (Cory Martin)

  • Code sprint at EMC last week, focused on reproducing GSI H(x) and QC with UFO. Made some progress, but still more work to do.

GMAO (Ron Gelaro)

  • Similar development at GMAO. Reached out to EMC on the next step of handling duplicate obs and obs error inflation, plan to meet next week to discuss.
  • No labels