production/AQM.v7.0.8

AQMv7.0.8 aims to support the upgrade of RAVE v2 along with bug fixes for AQM model, bias-correction bull's eye issue and WMO files for hourly average O3.

production/AQM.v7.0.7

This release/tag brings in the latest changes made to the production code delivered to NCO in support of the AQMv7.0 implementation.

production/AQM.v7.0.1

Production AQM v7.0.1

ufs-srw-v2.2.0

The Earth Prediction Innovation Center (EPIC) and the Unified Forecast System (UFS) community are proud to announce the public release of the Short-Range Weather Application (SRW App) version 2.2.0. This release is an update to the version 2.1.0 release from November 2022 and reflects a number of changes currently available in the SRW App develop branch. The Application is designed for short-range (up to two days) regional forecasts located anywhere on the globe. The new release includes the following updates:

• Bug fixes since the v2.1.0 release
• Addition of the supported FV3_RAP physics suite (PR #811) and support for the RRFS_NA_13km predefined grid
• Addition of FV3_GFS_v17_p8 physics suite (PR #574)
• Updates to CCPP that target the top of the main branch (which is ahead of CCPP v6.0.0). View a detailed summary of CCPP updates that came in ahead of the SRW App v2.2.0 release.
• Expansion of Level 1 platforms to include Derecho, Hercules, and Gaea C5 (PRs #894, #898, #911)
• Transition to spack-stack modulefiles for most supported platforms to align with the UFS WM shift to spack-stack (PRs #913 and #941)
• Overhaul of WE2E testing suite (see, e.g., PRs #686, #732, #864, #871)
• Improvements to the CI/CD automated testing pipeline (see, e.g., PRs #707 and #847)
• Incorporation of additional METplus verification capabilities (PRs #552, #614, #757, #853)
• Integration of the Unified Workflow's templater tool (PR #793)
• Ability to create a user-defined custom workflow (PR #676)
• Option to use a custom vertical coordinate file with different distribution of vertical layers (PR #813) and documentation on how to use this feature (PR #888)
• Incorporation of plotting tasks into the workflow (PR #482); addition of ability to plot on both CONUS and smaller regional grid (PR #560)
• Addition of a sample verification case (PR #500) with documentation
• A new tutorial chapter in the documentation (PR #584)
• Incorporation of UFS Case Studies within the WE2E framework (PRs #736 and #822)
• Air Quality Modeling (AQM) capabilities (unsupported but available; see PR #613)
• Miscellaneous documentation updates to reflect the changes above

The v2.2.0 User's Guide has been updated to reflect these improvements to the Application, and numerous test cases with all associated data files are also available to the public in the SRW App data bucket). Interested users can get support by submitting a question through the GitHub Discussions Q & A page.

Known Issues

• The v2.2.0 container does not run on Derecho.
• The path to the container on Derecho should be /glade/work/epicufsrt/contrib/containers

production/AQM.v7.0.0

• Production AQM v7.0.0

ufs-srw-v2.1.0

The Earth Prediction Innovation Center (EPIC) and the Unified Forecast System (UFS) community are proud to announce the public release of the UFS Short-Range Weather (UFS SRW) Application version 2.1.0. This release is an update to the version 2.0.0 release from June 2022 and reflects a number of changes currently available at the head of the UFS SRW development branch. The Application is designed for short-range (up to two days) regional forecasts located anywhere on the globe. The new release includes the following updates:

• Conversion to a Python workflow (from the former shell workflow)
• Improved container support, including the option to run across compute nodes using Rocoto
• Updates to physical parameterizations within the Common Community Physics Package (CCPP) (see detailed description here)
• Support for the Unified Post Processor (UPP) to be run inline
• Addition of a multi-purpose code clean-up script (devclean.sh)
• Improved support for the METplus verification suite

The User's Guide, available here, has been updated to reflect these improvements to the Application, and numerous test cases with all associated data files are also available to the public here. Interested users can get support by submitting a question through the Q&A Category on the ufs-srweather-app GitHub Discussions page.

Known Issues

• When building a PDF of the v2.1.0 documentation locally, an error in the conf.py file causes the subtitle to read "Release v1.0." However, the contents of the documentation are correct for v2.1.0. If users want their PDF to reflect the correct version number, they can change line #30 of ufs-srweather-app/docs/UsersGuide/source/conf.py from release = 'v1.0' to release = 'v2.1.0' and rerun the command to generate the document.

ufs-srw-v2.0.0

24 June 2022: Short-Range Weather (SRW) Application v2.0.0

Release Notes

The UFS Community is proud to announce the public release (version 2.0.0) of the UFS Short-Range Weather Application. This newly updated weather application is designed for short-range (up to two days) regional forecasts located anywhere on the globe. It contains the Common Community Physics Package (CCPP) v6.0.0, version 10.1.0 of the Unified Post Processor (UPP) and version 3.0.0 of the UFS Weather Model. Documentation has been significantly enhanced and numerous test cases with all associated data files are now available to the public. The application is highly portable and detailed instructions describe how to build and run on a wide variety of platforms, including Ubuntu Linux and MacOS.

Documentation: User's Guide v2.0.0

Known Issues

• On MacOS, users must run brew install bash before using brew to upgrade bash. MacOS users also need to run brew install gsed.

• Bash/dash issue: On some systems, .sh points to bin/dash instead of bin/bash. On these systems, users can run sed -i -e 's/bin\/sh/bin\/bash/g' *.sh to resolve the issue.

• On Gaea: The staged data path in regional_workflow/ush/machine/gaea.sh must be updated from ncep to ncep_shared if users do not have access to the ncep directory. To do this, change staged_data_dir to "/lustre/f2/pdata/ncep_shared/UFS_SRW_App/v2p0/".

• When users build the HPC-Stack for use with the SRW App v2.0.0, they should clone the HPC-Stack srw-public-v2 branch as follows:

  git clone -b release/srw-public-v2 https://github.com/NOAA-EMC/hpc-stack.git

• Known WE2E Test Failures: Some failed tests will pass if run a second time individually. The following tests consistently fail:

  Platform	Test	Failure Stage	Error Message	What was done?
  Cheyenne (gnu)	specify_DOT_OR_USCORE	run_fcst	FATAL from PE4: set_group_update: mpp_domains_stack overflow, call mpp_domains_set_stack_size( 3013795) from all PEs.	Increased domains_stack_size in input.nml and the test still crashes
  Jet (Intel)	pregen_grid_orog_sfc_climo	run_fcst	FATAL from PE4: NetCDF: Start+count exceeds dimension bound: netcdf_read_data_2d: file:INPUT/C403_grid.tile7.halo3.nc- variable:x

ufs-v1.0.1

ufs-v1.0.1 release with updated versions of UPP, regional workflow, updated Users Guide.

16 September 2021: Short-Range Weather (SRW) Application v1.0.1

Tag ufs-v1.0.1

Release Notes

A bug fix release for the UFS Short-Range Weather (SRW) Application v1.0.1 was publicly released on 9/16/2021. Bug fixes related to:

• addressing a runtime error that occurs due to an lmod update on Hera (PR #542),
• resolving a filter_topo failure by removing "refine_ratio = ${refine_ratio}" in scripts/exregional_make_orog.sh (PR #547), and
• adding an array index for 'cycle_status' to the launch script (launch_FV3LAM_wflow.sh; PR #513)

are included in this release.

Updated documentation is available and now includes information on running the SRW App release using a Docker container (new Chapter 11).

The authoritative UFS SRW App umbrella repository is located at: https://github.com/ufs-community/ufs-srweather-app.

ufs-v1.0.0

ufs-v1.0.1 release for the Short-Range Weather Application

4 March 2021: Short-Range Weather (SRW) Application v1.0.0

Tag ufs-v1.0.0

Release Notes

A release description is provided here.

Known Bugs

• Input source string

  • The model code incorrectly ignores certain pre-processing fields available in the gfs_data.nc file unless the input source string written out by chgres_cube is set to be an FV3GFS Gaussian nemsio file.
  • Temporary fix:
    • This string can be manually set to “FV3GFS GAUSSIAN NEMSIO FILE” for all input sources (while unsatisfactory, this fix was deemed to be the best option to resolve this issue for the release). These changes can be made in the chgres_cube "write_data.F90" routine near line 1317.

• Reading ice mixing ratio from wrfprs HRRR input files

  • chgres_cube does not read in ice mixing ratio (ice_wat) from wrfprs HRRR files because it does not search for the correct grib2 variable code
  • Temporary fix:
    • One option is to use HRRR wrfnat files instead. Ice mixing ratio is properly read from these files.
    • Code can be modified by changing the string "_1_84" to "_1_82" on lines 2609 and 2619 of UFS_UTILS/sorc/chgres_cube.fd/input_data.F90.

• Radiation wrapper

  • A CCPP-physics submodule issue (#565)
  • A unit conversion bug associated with the calculation of density impacts subsequent calculations of effective radius in Thompson MP that is used in RRTMG, impacting shortwave radiative fluxes.
  • A bug fix has been tested in gfsv16_ccpp global experiments and RRFS experiments. The results show significant impacts on the SW radiative fluxes and subsequently temperature when Thompson MP is used.

• Script to generate workflow (generate_FV3LAM_wflow.sh)

  • The ability of this script to catch errors is limited. If errors go unnoticed, it is possible that subsequent tasks (e.g. make_grid) fail.
    • Example: if the make_grid step fails with an error like
      arcx, arcy 0.0000 0.0000
In bestesg_geo; a nonpositive domain parameter, garcx or garcy
STOP Failure flag returned from get_bestesg
    • Inspect generate_FV3LAM_wflow.log for errors such as
      .../constants.sh: line 8: bc: command not found

• Removal of cron job on Cheyenne

  • The automatic removal of the cron job (used to relaunch the workflow) fails on Cheyenne and it will need to be removed by hand when the cycle has completed.

• On some MacOS and Linux platforms, stack limits may result in errors

  • Users must remove stack limits when running on MacOS by using the command ulimit -S -s unlimited, otherwise there may be runtime errors.
  • These errors can manifest in multiple ways; one example is during the execution of the “orog” program:
    ..
nx = 428
IM= 0 /= grid file nx/2= 214
Set IM = 214
JM= 0 /= grid file ny/2= 129
Set JM = 129
Command terminated abnormally.
0.52 real 0.00 user 0.00 sys
ERROR:
From script: "exregional_make_orog.sh"
...
  • Similarly, on Linux, users should set their stacksize to a large enough value (e.g. 65536, which corresponds to the macOS ulimit -S -s unlimited setting) or to “unlimited”:
    • Bash: ulimit -s unlimited
    • Tcsh: stacksize unlimited

• Loading the Python conda environment on Hera and Jet

  • C-shell users may encounter the following error when loading the Python conda environment on Hera and Jet:
    fi: Command not found
  • Note that this does not impact the conda environment and Python plotting still succeeds

• MYNN diagnostic output

  • A reported bug related to setting the bl_mynn_tkebudget control variable in GFS_typedefs.F90. In particular, the bl_mynn_tkebudget variable is not read in via namelist nor set. This should only affect diagnostic output of the MYNN PBL scheme. See this issue for more details.

Limitations / Disclaimers

• Noah-MP is an experimental land surface model that is still in relatively early stages of testing with the RRFS.
• When changing the MPI decomposition (parameter layout = x,y in the model input namelist input.nml), results are bit-for-bit identical for the RRFS_v1alpha suite at all levels of optimization (production/reproducibility/debug flags). For the GFS v15.2 suite, results are bit-for-bit identical only when using reproducibility or debug flags.
• Users may run into instabilities/warning messages during model integration, including near-CFL “Warn_K” messages, segmentation faults, and/or model failures with some model configurations. These limitations are representative of the developmental state of the application. Potential fixes include Increasing the “n_split” namelist parameter to 5 or 6 or decreasing the “dt_atmos” mamelist parameter.
• The SRW App cannot be run at resolutions lower than 3 km because the required topographic fix files for those resolutions are not available.