General announcements
- Next meeting in two weeks (no special topic for next week's time slot)
- Send in any special topics you are interested in presenting and discussing
Infra - Evan
- Ashley added a new ingest suite for ERA5 Analysis
- Evan and others are looking into doing EWOK tasks in a Singularity container
- Good progress with test and installation automation for R2D2 (Eric, SteveG)
Interfaces - Liam
- Raji submitted two PRs
- C++ implementation of surface pressure operator using the UKMO and GSL schemes
- Updates to filters for better traceability using the oops::Log::trace() channel
- SteveH and SteveV are working on the new IODA reader refactoring and integration of the new ObsSpace DataFrame
- New reader steps: Load (file I/O), Filter, Distribute
- Filter step implementation is done, Load (SteveH) and Distribute (SteveV) steps are work in progress
- Met Office contributed a IODA parallel ODB reader (in develop)
- NOAA-EMC contributed a IODA parallel BUFR reader (PR under review)
- Many thanks to our partners for these valuable contributions
- Liam is continuing work on GPU offloading in Atlas, Ectrans for various steps including adjoint, inverse, etc.
Algo - Clementine
- Christian is back from PTO and is working on the continuous DA capability
- Nate is adding a feature for writing out lat, lon files from any model which will be used to compute scores
- Clementine
- Adding capability to compute scores using WeatherBench
- Assisted with work to run EWOK tasks in a Singularity container
- Cleaning up the EWOK repo, including removing obsolete code
COMPO - Jerome
- PRs submitted for ingest of new aerosol obs types
Marine - Kriti
- Work in progress with 4DVar, HTLM
OBS - BenR
- Github discussion (https://github.com/orgs/JCSDA-internal/discussions/171) for extending the ground based radar operators
- Ocean surface wind converter (GNSSR, COWVR)
- Ashley helped out with the COWVR radiance ingest suite
- This provides a nice guide for further ingest suite development
- SSMIS unit test updates
- Ncdiag converter updates including a fix for an issue with azimuth angle metadata
- Contributions from partners:
- Mean gaussian filter which can be used for actions such as super-obb'ing
- Work in progress to reduce spatial density of GeoVaLs
- Various UFO bug fixes from Jianjun
CRTM - Cheng
- Final check for instrument coeff. in netcdf format
- Once this is complete, we will be ready to assemble the 3.2 release
- New coeff. gen for PACE - visible channels
- COMPO will be one of the first users to try this out
- Various bug fixes
- Azimuthal Jacobian for visible channels
- Surface emissivity Jacobian
- Various ctest updates
NOAA-EMC - Dan
- Focus is on testing and qualifying flows including ones that use JEDI
- Realtime runs
- Reproducibility checks
- Good results in general
- Some obs types are showing poor results and are being investigated
- H(x) differences of 10's of Kelvin, has some dependency on the number of MPI tasks being used
- Plan moving forward is to dump out quantities in single precision and call it good when differences are within roundoff errors
- Qualification with entire workflows in underway
- Cleaning up the build processes including eliminating forked repos (eg, MOM6)
- Moved to spack-stack-1.9.1
- Ron's parallel BUFR reader PR has been waiting for spack-stack-1.9.1 and is now ready for review
GMAO - Ricado
- Checking GSIBEC results and will keep EMC updated
- Finishing up work with Francois H. to implement interpolator changes needed by GSIBEC
- Wrapping up fixes for CO2 UFO operator issues
- Put knobs into GMAO workflow for cycling DA
- Seeing good results with SWELL
- Starting to look into ingesting SABER(?) obs into JEDI
- Measurements from mesosphere, upper stratosphere
- Have a converter to BUFR, working on path to ingest in JEDI
MMM - ChrisS
- Developing/updating reference configs for workflows
- Prep for MPAS-JEDI tutorial scheduled for this June in Scotland
- Updated obs2ioda converter to directly output IODA V3 format
- Working on hourly cycling in DA flows
Met Office - MikeC
- We don't have any updates from the UKMO this week which were not part of the quarterly review.
- We have had bank holidays over the easter period.
- In terms of general work:
We continue to implement JEDI operationally for the observation processing and the trialing looks good
This is planned for operational implementation at the end of 2025
We are currently testing working on HTLM, hybrid-B, multi resolution testing in preparation for our next generation HPC benchmarking exercise
Due later in 2025
OAR - no update
ROMS - HernanA
- Qualifying more realistic 4DVar flows
- Error analysis
- Looking at innovation, increments, background variance, etc.
- Working out tuning and calibration for the flows
NRL - no update