Meeting notes courtesy of Francois V.
OBS Team
Decoder:
Total Electron Content PR is ready to go. Has been tested on Navy Machines and elsewhere. Uses public CDAAC input files. There is a next step where a consensus is needed on the new variables required. The observation itself is relatively straightforward; the Earth Centered Earth Fixed coordinates for the location may need a discussion plan to entail NRL, SWPC and anyone from NASA and UKMO who is interested. https://github.com/JCSDA-internal/ioda-converters/pull/1505
WSF-M Weather Sat Follow-on, Microwave (Imager) mission. This does have some fully polarized channels, the converter has been tested and is fully working for the proxy data provided. https://github.com/JCSDA-internal/ioda-converters/pull/1512
Working with Youlong from NOAA on the IMS snow product, this is a file that has the snow cover product but also an entry for a converted snow depth. Further there is some, standardization and convention updates. We are working to ensure all the land converters are up to date: https://github.com/JCSDA-internal/ioda-convreak. Some CI coefficient files need also to be updated. https://github.com/JCSDA-internal/ufo/pull/3015erters/pull/1510
UFO:
EUMETSAT SAF project with Neill Bowler and Hailing that is abstracting GNSS-RO QC checks. One of the last PRs deals with an update to the Meteo France super refraction check. https://github.com/JCSDA-internal/ufo/pull/3239
QC flag propagation from OOPS, is ongoing work. Note that any new operator should include these QCFlags and handling of them. New developments to propagate these QC flags: https://github.com/JCSDA-internal/ufo/pull/3157
GOES global lightning operator: PR was submitted last quarter by the OU CAPS team, recently we’ve been able to do our thorough review and new updates are included. This is ready for a closer look by all, please take a look if you’re interested. https://github.com/JCSDA-internal/ufo/pull/3308
VarBC: is out of draft mode and has the required approvals so could be merged. Thanks to Cory, Andrew and Mike Cooke for comments, we would like to have a review from NASA. We will merge next Wednesday.
ALGO
- Working with the compo team, we are testing the HTLM using a geos ensemble with a 15 min temporal resolution and PSEUDO model. The initial results look good with the HTLM reducing linearization error for atmospheric chemistry variables. We have also run the HTLM at high resolution (c384) using a UFS ensemble but only at an hourly temporal resolution. If UFS forecasts in JEDI can be repaired we will likely scale up our experimental setup with GEOS to c384 once we can produce at least a 15 min temporal resolution ensemble.
- We have also added the ability to run a forecast from the analysis directly in the variational application as a first step toward the development of continuous DA capabilities in JEDI.
- Working on tuning background error covariances for snow 2DVar with Clara Draper (PSL).
- Preparation for model naming convention: will hold sprint last week of May.
- Nearly finished an update to the b-matrix training suite that makes yaml configuration more robust and allows for more user-friendly adjustment of training settings.
- Working on the GEOS JEDI interface and making good progress (with Dom and Matt Thompson's help), still working on passing information from JEDI back to GEOS with other people at GMAO.
- Working on continued improvements of the interpolators: work to reduce memory usage and improve robustness on different atlas grids.
INFRA
R2D2 team in conjunction with GMAO continues to develop and finalize the initial client/server architecture. End user testing should begin soon.
Two main peices to mention here.
Evan had to create a OpenAPI 3.0 parser from scratch to replace a dependency. This was a non-trivial technical task.
Tariq has created a client API generator that automatically creates the r2d2 client from the web service Open API specification.
Ashley Griffin + OBS
Add ingest capability for NASA GMI (ready to merge)
Ashley
Create an infrastructure knowledge base and are in the process of populating it with information. This is using UCAR's Confluence wiki and will be living document with various INFRA procedures, support, and troubleshooting. https://wiki.ucar.edu/display/JEDI/Infrastructure+Knowledge+Base
Dom
JEDI-UFS integration and JEDI-GEOS integration: consolidating FMS branches and move to a newer version of the FV3 dynamical core in JEDI’s FV3-dry-dycore-only setup.
Move Github self-hosted CI runners to new EC2 instance (affects r2d2, simobs, skylab, ufs-bundle, spack-stack) - will complete this week. Ping INFRA if you notice errors.
Evan
We are doing a trial test of the addition to oasim and rttov to our CI builds, this is available now in the jedi bundle via a build flag (thanks Anna)
We are seeing sporadic new failures in CI integration tests due to several ufo and saber tests affecting the saber and ufo repositories, these will be investigated
Steve
IODA: Phil UnderwoodUKMO and Steve completed a prototype for the ObsSpace improvements, it’s now in a feature branch
SOCA
- I am continuing my SOCA deleting and refactoring spree
- I removed about 2000 lines of soca code as states / increments are stored primarily in atlas fieldsets now ( this may be of interest to FV3 and others)
- I should be opening the generic diffusion operator draft PRs in saber/oops/soca/fv3 today or early next week.
- Hamideh is getting the NOBM (ocean biogeochemistry) backgrounds into R2D2, needed for ocean color (OASIM) work
- Kriti is starting to look at HTLM for soca
EMC
GFS-17:
Working on real time marine observations.
Working out B matrix (BUMP) issues for aerosols and adding to workflow.
Workflow adjustments for snow DA. IODA converters for land.
Re-starting WCDA experiments (Ocean - Sea Ice - Atm[GSI])
Preparing JEDI for running on the operational machine.
Regional (RRFS V2)
Adding SABER component for multi-grid Beta filter background model and running with Vader.
PR adding Vader to MPAS-JEDI.Atmospheric DA
LETKF: investigating issue where reading h(x) has a massive memory footprint (18gb for single ATMS).
Observer verification continues. Investigating GeoVaLs differences for GNSSRO. Working on aircraft filters.
Looking into failures when observations missing from a cycle.
Prototyping a database modernisation of EMC obs tanks.
Interacting with UFS using cubed sphere increments.
Interfacing to JCB repo for building JEDI executables.