Sync HAFS submodules and bring in HAFSv1 baseline configurations #166
Conversation
…rog files in HAFS workflow (Bin Liu and Xu Li).
…nto feature/hafs_sync_202210
…S FV3ATM-HYCOM coupling.
feature/hafs_sync_202210, which has FV3ATM-HYCOM coupling momentum flux related bug fixed in HYCOM-src. *And submodules hafs_forecast.fd and hafs_utils.fd were synced with their authoritative branches as of 10/22/2022.
static/moving nesting configuration in HAFS application/workflow.
Majority of them are default values, but could be changed/tested in future experiments. Meanwhile, change to use upoff of 2 (instead of 1) to make the forecast model more stable.
Note: The sync of GSI develop was done through a series discussion/working session among Ting Lei, Xu Lu (OU), Li Bi, Yonghui Weng, Biju Thomas, Bin Liu.
FV3dycore internal tracker output phtcf file.
…_gsi' into feature/hafs_sync_202210
which were synced to their latest authoritative branches as of 11/13/2022.
domain center, when config.domlon/lat are not set. storm, parent domain centered at storm center; auto, parent domain center determined based on storm center and 72-h forecasted storm location (method used in HWRF). *Enable automatically setting the moving nest domain location when i/jstart_nest, i/jend_nest are set as -999. *Updated and improved hafs_workflow.xml.in and site entity files. Note: Need python3 version with numpy and xarray. Todo: Update to use proper python3 version on RDHPCS.
…_gsi' into feature/hafs_sync_202210
…2 complier(Orion)
*Update NEXRAD related gsi_parm namelist options (according to Jason Sippel). *Fix a typo in ush/hafs/launcher.py related auto detection of moving nest domain location. *Added hafsv1a baseline conf file and cronjob driver.
|
@panll, @mrinalbiswas, @JunghoonShin-NOAA, @BijuThomas-NOAA, all the submodules are up to date now, please go ahead to start the application/workflow level regression tests when you get a chance. Thanks! |
|
RT passed on WCOSS2 |
|
Regression tests passed on Hera except fails at Analysis (e.g. rt_regional_mvnest_vida) with error message /apps/lmod/8.5.2/libexec/lmod bash load gsi_hera.intel |
|
Hello Bin, Regression test is completed successfully on Jet. |
@panll, Thanks for testing! This is somewhat strange. A few of us are actually running the HAFSv1 baseline configurations on Hera (with DA turned on) and we did not encounter this gsi_hera.intel or mkl module load issue. @BijuThomas-NOAA or @lbi2018, could you please help to take a look at this. Thanks! |
|
|
The "module purge" fixes the problem. Thanks! @BinLiu-NOAA @yonghuiweng |
|
Thanks to @BijuThomas-NOAA, @JunghoonShin-NOAA, @panll and @mrinalbiswas for conducting the application level regression tests promptly for this PR! And @panll, for the gsi analysis job module purge issue on Hera, we will ignore it for now, since others did not have this issue. If it keeps coming next time, we will try to address this next time. Thanks! |
Description of changes
Sync HAFS submodules and bring in HAFSv1a/a baseline configurations back to HAFS develop branch.
Issues addressed
Contributors
Lots of UFS HAFS application team members contributed to this PR. Details can be found in the Git commit log messages.
Tests conducted
Application-level regression test status
Running the HAFS application-level regression tests is currently performed by code reviewers after the developer creates the initial PR. As regression tests are conducted, the testers should use the checklist below to indicate successful regression tests. You may add other tests as needed. If a test fails, do not check the box. Instead, describe the failure in the PR comments, noting the platform where the test failed.
As usual, please help to conduct the HAFS application/workflow level regression tests for this PR.
@panll (Hera)
@mrinalbiswas (Orion)
@JunghoonShin-NOAA (Jet)
@BijuThomas-NOAA (wcoss2)