INFRA - Dom
- Work on upcoming Skylab v7 release - address bug fixes and finalize data ingest tasks
- Roll out of spack-stack-1.6.0 on all platforms; this release is not used for the Skylab v7 release (it uses 1.5.1), but we expect to switch over within the first quarter of 2024
- Ongoing work on data ingest suite (more observations, test running MPAS experiments in near real-time)
- More repositories were onboarded in the Terraform Github management tool; additional repositories (fv3-jedi and mpas-jedi related) are being discussed at management level and are currently on hold
ALGO - Christian
4D ensemble covariances now also available via SABER covariances
New localization setup for the 4DEnVar gfs skylab experiments (with Nate) with horizontal localization lengthscales varying by model levels.
Capability to plot lat-lon analyses and increments on model levels (as well as pressure levels) - work in progress
Setup for a 2DVar for snow DA (with Clara Draper, PSL/NOAA)
Big Atlas interface standardization change is now merged!
Removing last variable change from FGAT cost function
Modifying MPAS Skylab workflow to handle different resolution for En-Var minimization
UFS full cycling high-res 80 members
Score calculation progress
OBS - Hui
Obs team is actively working with CRTM to further improve the reflectance assimilation capability. Its prototype was implemented and tested in the past quarter; Obs team has also added QC flags entry into all UFO operators, which opens the pathway to limit the UFO simulations to the observations passing QC filters. The use of QC flags has been implemented to the CRTM operator at this point. Coordinated PRs among OOPS, UFO, UFO-data are under review; Meanwhile, OBS team is actively working on optimize observation models for both MW and RO, which allows adaptive errors to be used in different weather scenarios (e.g., cloudy areas, ducting situations); MetOffice visitoring program continues to support the OBS team to refactor the RO operators and allow the separation of filters and operators for this particular obs type.
CRTM - Cheng
CRTM v3.1.0 was released on Dec 21th, 2023, with new features such as active sensor and DDA cloud tables. We are working on finalizing the user guide for CRTM v3.1.0.
There were a lot of discussions following this:
Ricardo Todlling: still testing CRTM v2.4.0 and saw some unexpected results that need to be sorted out. Concerned about the release of CRTM v3.1.0 would discontinue the support of older versions of CRTM in skylab.
Jack Zhiquan Liu: NCAR saw similar unexpected results of brightness temperature with v2.4.0 versus v2.3.
Emily Liu: her term did not have any issue with CRTM v2.4.0 being operationally used.
Don: had a lot of discussions on different CRTM versions with Ben Johnson; restate that CRTM v2.4.1 was not meant for operational use, but a bridge between v2.x to v3.x.
Yannik/Hui: let's wait for BenR and BenJ to come back and have an extended discussion on how to support different versions of CRTM in JEDI/UFO.
SOCA - Travis
- skylab: 3D EXPLICIT_DIFFUSION implemented as an alternative to BUMP_NICAS. Should perform better with the 1/4 deg ocean grid we use (w/ other resolutions, BUMP_NICAS might perform better)
- skylab - cycling marine-ufs, is in progress, probably won't make it into v7, but should be working before quarterly review
- skylab-marine should be working now after some bug fixes
- but what is going on with Intel on orion/hercules??? 3dvar is horribly slow
COMPO - Jerome
EnVar geos aerosol will go in release. (Non-cycled)
Working on HTLM. Not this release.
TEMPO
GMAO -
No updates due to holidays
ROMS - Hernan
Got his ctests working with Atlas change
NRL - Sarah
Looking at Minimizers in OOPS
Looking at interpolators
Have seen stripack collinear error. (PR fix has now been merged as of this morning.)
EMC - Guillaume
No updates, but complaints
Found a bug initializing sea ice
MMM - Jake
Cheyenne gone. Using Derecho.
Mostly using gnu compilers on Derecho. Uses less memory. But Derecho better with Intel than Cheyenne was.
All-sky IR DA.
Met Office - David
Spectral waveband covariance
Issues when trying to combine their DA and OBS systems