You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 2 Next »

  • Infra
    • Spack-stack 1.9 (1.9.1) is ready for site installs and, should be tagged soon. 1.9.1 includes fixes for fckit and ectrans builds, and updated wgrib2
      • Updated discover spack-stack - note discover's 1.9.0 install includes all the 1.9.1 fixes  (Ashley)
      • Presubmit CI is now on 1.9  (Evan)
      • AWS Developer Instances  (Steve G)
      • We are skylab-CI before EOW (Evan)
      • Track platform installs on this spack stack issue: https://github.com/JCSDA/spack-stack/issues/1485.
    • New obs in the ingest suite: SSMIS, TEMPEST, and COWVR (Ashley + Obs. team)
    • R2D2 Finishing up optimizations of R2D2's new production server. (Eric)
      • Finishing up new testing framework (Steve G)
      • Added localhost Docker container for new R2D2 localhost server (Steve G)
      • Distributing instructions for test users to try it out on Hercules, Discover, and localhost today. (Eric)
      • We will switch over to the new system next Wednesday early in the morning. (Eric)

Interfaces and Optimization –

- added channel-selection to ObsSpace OSDF backend [SteveV and SteveH]
- benchmarking the ObsGroup vs OSDF in IODA after implementing a crude/temporary means to populated the OSDF from the ObsGroup. Benchmark is using mpas-jedi H(x) application, focus on cris and iasi instruments with many channels. [SteveH]
- enabling surface correction operator to handle geometry and geopotential heights as inputs [Raji]
- working on VADER LVCs to make more flexible (see PRs, including saber and fv3jedi adaptions) [SteveV]
- working on porting ectrans transforms to GPU [Liam]
- building JEDI repos on Nvidia compilers [FrancoisH]
- ongoing testing with Ricardo for improvements to the interpolations used with GSI-B in SABER [FrancoisH]OBS -


OBS –

The 2nd ROMEX meeting took place in Darmstadt at EUMETSAT. Presenters from all partners (NOAA, NASA, NRL & JCSDA participated in the meeting. JCSDA (Hailing) presented results with the JEDI-based system. 



UFO:


Mike Cooke’s PR #3610 has added the first leg of few PRs to complete varBC by record.  We need this reviewed by the impacted partners NOAA and NASA.


Nghi Do for from Scripps UCSD had a fix with an ROPP Airborne Radio Occultation ctest is not routinely exercised and had fallen behind.


JianJun has a series of PRs for all-sky implementation. The 1st one was merged to get potential temperature at selected levels. There are additional PRs, one in fv3-jedi, one on IODA conventions, and a third one adding a ctests of the function related to the cloud fraction calculations. Please review those PRs if you’re interested.


Praveen from NOAA has a PR to bring in Eumetsat Cloud and Aerosol Detection code in JEDI. That code was originally written by Andrew Collard.


Ricardo has a PR to provide options to set CO2 for passing into CRTM.  A constant which can be defined via yaml, or to enable a global Constant based on an equation derived from data at Mauna Loa. We are reviewing this and requesting the activation of the Mauna Loa feature and the addition of a ctest.

CRTM –

  • The CRTM team is participating in the VIS/UV DA workshop held in Italy this week.
  • Pending PR on unit tests regarding surface emissivity currently under review.
  • Bug/issue regarding surface emissivity jacobians (V3.0+, reported by Jack Liu)
  • CRTM k-matrix unit test failures with intel compiler (reported by Evan)


Compo

Jerome


NASA presentation to the Earth Science Directorate about our TEMPO success
ISSI presentation about our advances in Visible L1b data assimilation

R2D2 discover switch to CSS

first set of TEMPO DA runs for publication finished

Help on aerosol dependency with SOCA OASIM development



SOCA

Thank for taking the notes today. Here are the bulletpoints:

  1. Travis has been working on general SOCA code optimization and AI based covariances 
  2. Kriti has been working on scientific testing of HTLM and having some issues which Christian is helping with

Discussion with Hernan: Travis plans on adding vertical impicit diffusion next AOP.

  1. Hamideh has been working on wrapping her OASIM PRs

Algo –

Nate

  • merged PR in MPAS-JEDI and Skylab for using the CCPP model variable names with MPAS (a big thanks to BJ for his help with this)
  • fixed the MPAS 120km static B Skylab experiment (it needed a new background generated from the version of MPAS in the jedi-bundle). The new background will be ingested into r2d2 soon.
  • ProcessPerts for multi-scale localization is now in EWOK & Skylab.
  • Working on a fix in saber for letting process perts work better when increment includes surface fields.

Christian

  • Adding adaptive regularization and SVD truncation options to the HTLM. Testing with SOCA and GEOS
  • Testing Skylab additions for continuous DA
  • Slowly working to get single stream MPAS in Skylab
  • Continuing on continuous DA code changes in OOPS.


Clémentine

  • Doing some testing of our skylab experiments with spack 1.9.1
  • Narrowing down (?) on convergence issues with GEOS variational, debug mode failures from Dan, ... and other fv3jedi subtleties
  • Testing gsi 4denvar with Ricardo
  • No labels