RRFS_A Change Log

'
Cycle Experiment(s) Description
2024/6/13 12Z RRFS_A
  • The UPP code was updated (6/12: #fc85241).
  • WCOSS2 production switch from Dogwood to Cactus.
    • 2024/6/9 00Z RRFS_A
  • Call the non-LSM stochastic perturbation generation (SPPINT only) more frequently for REFS members 1 and 2, every 72 s instead of every 180 s.
    • 2024/6/2 18Z RRFS_A
  • Call the non-LSM stochastic perturbation generation (SPPINT only) more frequently for REFS members 3-5, every 72 s instead of every 180 s.
    • 2024/5/31 18Z RRFS_A
  • The NSSL microphysics perturbed parameters were reduced by 30-40% for REFS members 3-5.
    • 2024/5/22 12Z RRFS_A
  • RRFS was updated to use obsproc v1.2.
    • 2024/5/20 18Z RRFS_A
  • The GSI code has been updated (5/7: #529b6ea).
  • GVF update at 03Z and 15Z for fixing missing value in surface file.
  • The UPP code has been updated (5/10: #1916cb2).
  • The ufs-weather-model code has been updated (5/7: #6431581).
  • Call the stochastic perturbation generation (LNDPINT and SPPINT) more frequently, every 180 s instead of every 360 s.
  • Change the RAVE data directory to /lfs/h2/ops/prod/dcom.
    • 2024/4/24 12Z RRFS_A
  • WCOSS2 production switch from Cactus to Dogwood.
    • 2024/4/18 17Z RRFS_A, RRFSFW
  • 15 minute output is now produced for the first 18 hours of the forecast.
  • The RRFS fire weather nest was added to the RRFS workflow.
  • Cycle smoke/dust at the 06/18Z spinup cycle if the 04/16Z smoke/dust cycle is skipped.
  • The GSI code was updated to use the release/rrfs.v1.0.0 branch (3/26: #b53740a).
  • The UPP code was updated (4/12: #8bc3fc9). This includes a bug fix for blowing snow effect on visibility, and the removal of various fields from the grib2 output.
  • A bug fix for smoke and dust is included.
  • The version number was updated to v0.9.1
    • 2024/3/28 12Z RRFS_A
  • Call the stochastic perturbation generation (LNDPINT and SPPINT) more frequently, every 360 s instead of every 720 s.
    • 2024/3/27 12Z RRFS_A
  • Implemented a fix for the streak issue. fv_sg_adj was changed to 7200, n_sponge was changed to 65, and sg_cutoff was changed to -1.
  • Added GSI analysis to the cold start step at 03Z and 15Z.
  • Use RRFS/RAP observation dumps instead of RRFS_e/RAP_e dumps.
  • Call the stochastic perturbation generation (LNDPINT and SPPINT) less frequently, every 720 s instead of every 360 s.
  • Added time averaged smoke/dust.
  • Zero cloud fraction issue for RRFS ensemble members 1 and 4 was fixed.
  • Turned on fix for RRFS write component hangs.
  • The ufs-weather-model code was updated (3/19: #b1d7667)
  • The GSI code was updated (3/24: #f7e93ab)
  • The version number was updated to v0.8.9.
  • WCOSS2 production switch from Dogwood to Cactus.
    • 2024/3/20 00Z RRFS_A
  • Call the stochastic perturbation generation more frequently, every 360 s instead of every 720 s. There was an ensemble forecast failure using 720 s.
    • 2024/3/16 00Z RRFS_A
  • Call the stochastic perturbation generation more frequently, every 720 s instead of every 3600 s. The change to call it less frequently in v0.8.5 resulted in several ensemble forecast failures.
    • 2024/3/12 21Z RRFS_A
  • Turn on GLM FED Assimilation with GSI hybrid EnVar.
  • Call the stochastic perturbation generation less frequently.
  • Add EnKF analysis to the blending cycle.
  • Use the version of FVCOM data with lake ice turned on.
  • Write restart files every 12 hours for 60-hour forecasts.
  • Update the smoke processing to use the operational feed of RAVE data. Note that smoke and dust are turned back on with this version update.
  • Add GTG, IFI, and ATM to the product generation task.
  • Increase the Metar cloud base impact radius from 15 to 20. 20 is also used by the HRRR.
  • The rrfs-workflow code was updated (3/8: #8e9aff2).
  • The ufs-weather-model code was updated (3/5: #d53f87a).
  • The UFS_UTILS code was updated (2/27: #33cc663).
  • The UPP code was updated (3/1: #10bfc06). Use the 10-m wind gust calculation from RAP/HRRR.
  • The version number was updated to v0.8.5.
    • 2024/2/28 12Z RRFS_A
  • Turn on GLM FED Assimilation with EnKF.
  • GTG output from the UPP was added.
  • Turn on the blending of the RRFS EnKF and GDAS EnKF perturbations.
  • The ufs-weather-model code was updated (2/23: #a3818c9).
    1. Local model change from GDIT to improve UFS I/O.
    2. Local model change for a smoke/dust bug fix to the plumerise. Note that smoke/dust are still turned off in this version update.
  • The GSI code was updated (2/23: #7c4a571).
  • The UPP code was updated (2/1: #2953e46).
  • The version number was updated to v0.8.3.
    • 2024/2/28 12Z RRFS_A, RRFSFW
  • WCOSS2 production switch from Cactus to Dogwood.
    • 2024/2/13 15Z RRFS_A
  • RRFS observation dumps are now used for the EnKF cycle.
  • Note that between the 2/11 15Z cycle and this one, the EnKF was not assimilating any observations.
    • 2024/2/11 15Z RRFS_A
  • RRFS observation dumps are now used for the deterministic cycle.
    • 2024/1/31 12Z RRFS_A, RRFSFW
  • WCOSS2 production switch from Dogwood to Cactus.
    • 2024/1/31 00Z RRFS_A
  • Scale and variable dependent localization (SDL/VDL) was turned on.
  • Smoke and dust have been turned off until model instability/crashes can be resolved.
  • The ufs-weather-model code was updated (1/26: #932a532).
    1. Aerosol interpolation bug fix
    2. Hailcast bug fix
    3. CLM lake bug fix
    4. MYNN-EDMF PBL
      1. Cosmetic fix to the cloud fractions in (relatively rare) conditions with unsaturated grid cells and positive mixing ratios
      2. Very slight increase in diffusion in stable conditions
      3. Momentum transport bug fix - missing rho
    5. Grell-Freitas convection: Adjust the scale-awareness in the first hour forecast to reduce precipitation bias
    6. Smoke and Dust: Fix the diurnal cycle of smoke, dust emission over frozen soil, dry deposition
  • The UPP code was updated (1/28: #6331f0b).
  • The GSI code was updated (12/7: #2353eaa).
  • The rrfs-workflow was updated to the most recent version.
  • The version number was updated to v0.7.9.
    • 2024/1/10 12Z RRFS_A, RRFSFW
  • WCOSS2 production switch from Cactus to Dogwood.
    • 2024/1/7 00Z RRFS_A
  • The ufs-weather-model code was updated to include two bug fixes related to hailcast and aerosol interpolation, which contributed to several RRFS crashes.
  • The non-variational cloud anaylsis was turned back on.
    • 2024/1/4 09Z RRFS_A
  • The non-variational cloud analysis was turned off; it was identified as the cause of several model crashes.
    • 2023/12/22 05Z RRFS_A
  • The ufs-weather-model code was updated (12/21: #991d652).
    1. RUC LSM
      1. A change in the diagnostic of the snow depth accumulation. The old diagnostic included snow melting, but the new version only includes the snowfall with variable density. This will reduce the cold bias.
      2. The old diagnostics that included snow melting in the computation of snowfallacc had a bug in the computation of snow melt. Snow melt and snow sublimation should not be multiplied by snowfrac.
      3. The change to uncomment snow_mosaic=0 for fresh snow changes the model performance. This will help with the deficient snow depth on the ground.
    2. A bug fix to GFS_diagnostics, which was outputting weasdi (water equivalent snow depth over ice) instead of snodi (snow depth over ice).
    3. CLM Lake
      1. A new flag is created to track new freezing lake ice. Let sfc_diff calculate stability variables over the new icy grids for GFS PBL in the second surface loop, which will otherwise remain as missing values.
      2. A bug fix to use near-surface density instead of near-surface pressure.
    • 2023/12/8 21Z RRFS_A
  • The RRFSv1 multiphysics ensemble configuration was turned on (was single physics to this date). CLM lake is turned off in members 2 and 5 owing to a compatiblity bug when lake ice is present.
  • Update recentering to use the deterministic control as the base.
  • Turn on building precipitation in the cloud analysis. Intended to address the low precip bias in the DA cycle.
  • The following physics updates:
    1. MYNN-EDMF PBL
      1. Fixed 8-plume model instead of variable number plume.
      2. Bug fix to correct grid mean vs in-cloud mixing ratios. Now requires namelist variable lcnorm (set to true).
      3. Adjustments to cloud pdf to better fit modified Thompson scheme.
    2. MYNN surface layer
      1. Tweak to Czil parameter to improve diurnal cycle of T and wspd.
    3. Thompson aerosol-aware microphysics
      1. Change to nucleation rates to fix marine drizzle problem.
      2. Minimum effective radii of 7 microns over water.
    4. RUC LSM
      1. Adding leaf area index, wilting point, and soil field capacity to the history files.
      2. Adding a placeholder for the fire heat flux provided by smoke-dust code.
      3. A change to the transpiration function to take into account the non-linear distribution of roots.
  • Bug fix for the convective reflectivity calculation.
  • The ufs-weather-model code was updated (12/6: #1f7dd77).
  • The version number was updated to v0.7.7.
    • 2023/12/6 12Z RRFS_A, RRFSFW
  • WCOSS2 production switch from Dogwood to Cactus.
    • 2023/12/5 18Z RRFS_A
  • snwdph, sncovr_ice, sncovr, and snodl were removed from the ensemble re-centering due to forecast failures.
    • 2023/11/30 20Z RRFS_A
  • snwdph, sncovr_ice, sncovr, and snodl were added to the ensemble re-centering.
    • 2023/11/26 00Z RRFS_A
  • The EnVar radar-reflectivity data assimilation was updated to reduce the high reflectivity bias. The GSI code was updated (Nov 7: #a3e13da)
  • Updated the twice per day ensemble re-initialization to use 7-h GDAS EnKF instead of 30h GEFS EnKF. Offset the ensemble re-initialization time by 1h for 7h GDAS availability; re-initialization now occurs at 0700 and 1900 UTC instead of 0600 and 1800 UTC.
  • The update_LBC step was added and runs after the data assimilation tasks complete.
  • The smoke and dust fix files were updated.
  • Added the 2024 co2 fix files.
  • A bug fix using abi_g18 to replace abi_g17 was implemented.
  • More grib precision was added to ASNOW and PRATE. The UPP code was updated (Nov 7: #78f369b)
  • The version number was updated to v0.7.5.
    • 2023/11/14 12Z RRFSFW
  • Implemented the RRFS 65-level vertical configuration.
    • 2023/11/9 09Z RRFS_A
  • The rrfs-workflow repository is now used for running the RRFS_A instead of the feature/RRFS_dev1 branch of the NOAA-GSL fork of the regional_workflow repository.
  • The UPP code was updated (Oct 18: #9a191de)
  • The version number was updated to v0.7.3.
    • 2023/11/9 06Z RRFS Ensprod
  • The RRFS ensemble products now include the HRRR in addition to the 6 single-physics RRFS members for each cycle; 14 total members when including a time-lagged version of each.
    • 2023/11/8 12Z RRFS_A, RRFSFW
  • WCOSS2 production switch from Cactus to Dogwood.
    • 2023/10/20 00Z RRFSFW
  • The ufs-weather-model code was updated (Oct 3: #41811b8)
  • The UFS_UTILS code was updated (Oct 4: #751bb18)
    • 2023/10/19 21Z RRFS_A
  • The ufs-weather-model code was updated (Oct 3: #41811b8)
  • The UFS_UTILS code was updated (Oct 4: #751bb18)
  • The UPP code was updated (Oct 6: #fae617b)
  • The GSI code was updated (Oct 6: #978b7e7)
  • Hourly forecasts were extended from 12 to 18 hours for all non 00/06/12/18Z cycles.
  • Compression of history files and parallel netcdf were enabled.
  • CLM lakes were turned on.
  • GOES-18 clear sky radiance data was added.
  • N21 CrIS-FSR/ATMS data are now assimilated.
  • The version number was updated to v0.7.1.
    • 2023/10/11 12Z RRFS_A, RRFSFW
  • WCOSS2 production switch from Dogwood to Cactus.
    • 2023/9/13 12Z RRFS_A, RRFSFW
  • WCOSS2 production switch from Cactus to Dogwood.
    • 2023/9/11 12Z RRFSFW
  • The 06/18Z cycles were added.
  • Fractional vegetation and soil information was added to the fix files.
    • 2023/9/7 12Z RRFSFW
  • The ufs-weather-model code was updated (Jul 10: #f1f0180)
  • The UFS_UTILS code was updated (Jun 26: #ae483bd)
  • Grell-Freitas (GF) deep convection was turned on.
    • 2023/9/1 15Z RRFS_A
  • The UPP flat file (postxconfig) was updated.
  • The UFS_UTILS code was updated (Aug 21: #8a479da)
  • The version number was updated to v0.6.7.
    • 2023/8/30 18Z RRFS_A
  • SPPT was turned off in the RRFS_A ensemble.
    • 2023/8/22 06Z RRFS_A
  • SPPT was changed from 0.5 to 0.25.
    • 2023/8/19 15Z RRFS_A
  • Grell-Freitas (GF) deep convection was turned on.
    • 2023/8/19 00Z RRFS_A
  • SPPT was activated in the RRFS_A ensemble. It was inadvertently turned off since the ensemble members began running over the North America domain on 7/24.
    • 2023/8/15 18Z RRFS_A
  • 32-bit physics is now used instead of 64-bit physics.
  • The version number was updated to v0.6.5.
    • 2023/7/24 12Z RRFS_A
  • RRFS single-physics 5-member ensemble forecasts are now run over the 3-km North America domain.
  • hord_tr was changed from 10 to 8.
    • 2023/7/17 12Z RRFS_A
  • RRFS deterministic forecasts are now run over the 3-km North America domain.
  • The non-hydrostatic SIM solver with "hydrostatic assumption" was added automatically by updating to a newer version of the dycore than April 28th.
  • The version number was updated to v0.6.1.
    • 2023/6/30 12Z RRFS_A (CONUS)
  • The RRFS_A (CONUS) real-time parallel was turned off in order to transition to the North America domain.
    • 2023/5/30 12Z RRFS_A (CONUS)
  • Bug fix for snow mixing in MYNN.
  • Use dt_atmos = 36, k_split = 2, n_split = 5, dt_inner = 36 to replace dt_atmos = 60, k_split = 6, n_split = 3, dt_inner = 20.
  • Turned on SPPT and sppt=0.7 in single and multi-physics ensemble forecasts.
  • Turned on smoke-direct feedback functions (deterministic only).
  • The version number was updated to v0.5.2.
    • 2023/5/8 18Z RRFS_A (CONUS)
  • The version number was updated to v0.5.0.
    • ~2023/4/12 12Z RRFS_A (CONUS)
  • Switched to using the develop branch of ufs-community/ufs-weather-model (#cbf1567).
  • Switched to using the develop branch of NOAA-EMC/GSI.
  • Updated CRTM from v2.3 to v2.4.
  • The version number was updated to v0.4.2.
    • 2023/4/3 00Z RRFS_A (CONUS)
  • The single and multi-physics 9-member ensembles were turned on.
  • Turned on process_bufr_enkf and process_radarref_enkf for the EnKF cycle.
  • The version number was updated to v0.4.0.
    • 2023/2/20 00Z RRFS_A (CONUS)
  • The version number was updated to v0.3.8.
    • 2023/2/6 12Z LAMDAX
  • The LAMDAX parallel was turned off on WCOSS2.
    • 2023/2/5 00Z RRFS_A (CONUS)
  • Variable SLR/SDEN was added.
  • Smoke and dust components were added.
  • Turned on FVCOM processing.
  • Turned on hourly EnKF.
  • The version number was updated to v0.3.6.
    • 2023/1/18 00Z RRFS_A (NA)
  • The UPP code was updated (Dec 12: #324a195)
  • The new parallelized product generation script is utilized.
  • The version number was updated to v0.3.3.
    • 2022/12/14 00Z RRFS_A (CONUS)
  • The bufrsnd job was added.
    • 2022/11/20 00Z RRFS_A (CONUS)
  • The ufs-weather-model code was updated (Nov 3: #e639efc)
  • Direct radar reflectivity assimilation in GSI Hybrid EnVar and EnKF was added.
  • Updates to MYNN PBL, RUC LSM, and Thompson MP were added.
  • UPP updates were added, including adding snow variables, smoke and dust post-processing, GOES-16 simulated satellite brightness temperatures, and IFI functions.
    • 2022/10/16 00Z RRFS_A (NA)
  • Extended forecast length to 60 hours at 06Z and 18Z.
    • 2022/10/10 00Z RRFS_A (NA)
  • Hourly forecast length was increased from 3 to 18 hours.
    • 2022/9/27 02Z RRFS_A (NA)
  • Updated product generation script to generate "ak", "hi", and "pr" grib2 files.
    • 2022/9/19 00Z RRFS_A (NA)
  • Turned on hybrid analysis using GDAS ensemble.
    • 2022/9/18 00Z RRFS_A (NA)
  • Updated the product generation script to generate "namerica" and "conus" grib2 files.
    • 2022/9/16 00Z RRFS_A (CONUS)
  • Regional ensemble members from EnKF are used in the analysis at 19Z, 20Z, 21Z, 22Z, 23Z, and 00Z.
    • 2022/9/6 00Z RRFS_A (CONUS)
  • Hourly forecast length was increased from 12 to 18 hours.
    • 2022/8/22 12Z LAMDAX
  • The ufs-weather-model code was updated (Aug 18: #4640a44)
    • 2022/8/20 00Z RRFS_A(CONUS), RRFS_A(NA)
  • Updated CRTM from version 2.3.0 to version 2.4.0.
    • 2022/8/11 00Z RRFS_A(CONUS)
  • Updated Thompson microphysics scheme to reduce reflectivity bias.
  • Hourly forecast length was increased from 3 to 12 hours to support the AWT.
    • ~2022/8/1 00Z RRFS_A(NA)
  • Official start of the RRFS_A North America parallel.
    • 2022/7/21 00Z LAMDAX
  • The LAMDAX parallel was restarted on WCOSS2.
  • The ufs-weather-model code was updated (Jul 15: #d3d2565)
    • 2022/7/10 00Z RRFS_A(CONUS)
  • The RRFS_A CONUS parallel was restarted on WCOSS2.
    • 2022/7/05 00Z LAMDA, LAMDAX, RRFS_A
  • The real-time parallels were shut off on WCOSS1.
    • 2022/6/16 12Z LAMDA
  • The ufs-weather-model code was updated (Jun 14: #830b507)
    • 2022/6/16 00Z LAMDAX
  • The ufs-weather-model code was updated (Jun 14: #830b507)
    • 2022/6/3 00Z LAMDA, LAMDAX
  • GSI analysis script changed to use the early RAP data dump ("rap_e") instead of the late RAP data dump ("rap") for analyses valid at 00Z and 12Z.
    • 2022/5/10 00Z LAMDA
  • GSL's update_bc code will be used in place of the three EMC lamda_utils codes (create_expanded_restart_files_for_DA, move_DA_update_dat, and prep_for_regional_DA) to put the GSI analysis into the 00-h boundary condition file for warm started forecasts during the DA cycle.
    • 2022/5/8 00Z LAMDA-X
  • GSL's update_bc code will be used in place of the three EMC lamda_utils codes (create_expanded_restart_files_for_DA, move_DA_update_dat, and prep_for_regional_DA) to put the GSI analysis into the 00-h boundary condition file for warm started forecasts during the DA cycle.
    • 2022/5/2 12Z LAMDA, LAMDA-X
  • The resolution of the analysis grid was increased from ~9 km to ~6 km to match RRFS_A (grid_ratio_fv3_regional=2.0)
    • 2022/4/29 00Z RRFS_A
  • The Global 127-Level B-matrix (770, 127) was added, and the value of hszcl was updated to match LAMDA
  • Use a single domain for the restart file instead of 4 subdomains
    • 2022/4/25 12Z RRFS_A
  • The bufr sounding job was added
  • Snow and ice update at 00Z in the product cycle
  • Add MERRA2 aerosol data
  • hord was changed from 5 to 6
  • vtdm4 was changed from 0.03 to 0.02
  • Call radiation every 15 mins instead of every 20 mins
  • dz_min was changed from 6 to 2. If delz is less than dz_min, it is set to dz_min
    • 2022/4/11 00Z RRFS_A
  • Turn on the non-var cloud analysis
  • Use the C3359 grid to start producing output on the HRRR / NAM nest grid
  • Use 2-m T/Q from the FV3LAM model as the background for the surface observation operator in the GSI
  • Add soil moisture and temperature fields adjustment
  • Modify the io_layout to 1,4
  • Products for the HWT are now being created
    • 2022/4/2 00Z LAMDA
  • The C3359 fix files were updated to use pazi=0.0 (SW corner point is 1,1 for all fixed fields).
    • 2022/3/31 00Z LAMDA, LAMDAX
  • The diag table was updated to use unified output names and remove legacy products.
    • 2022/3/30 12Z RRFS_A, LAM
  • The RRFS_A parallel was added, and the LAM parallel was turned off.
    • 2022/3/29 12Z LAMDAX
  • The ESG compute grid for the North American grid was expanded, as well as the downstream North American output grid.
  • The ufs-weather-model code was updated (Mar 24: #a46c7ed)
    • 2022/3/28 12Z LAMDA
  • The ufs-weather-model code was updated (Mar 24: #a46c7ed)
    • 2022/3/18 12Z LAMDAX
  • The io_layout was modified from 1,1 to 1,15 to speed up the model initialization. This results in the grid_specs and restart files being written in 15 pieces, and a mppnccombine step is added to recombine the file pieces.
    • 2022/3/18 00Z LAMDA
  • The io_layout was modified from 1,1 to 1,14 to speed up the model initialization. This results in the grid_specs and restart files being written in 14 pieces, and a mppnccombine step is added to recombine the file pieces.
    • 2022/3/10 00Z LAM, LAMDA, LAMDAX
  • The ufs-weather-model was updated (Feb 25: #34a5895).
  • This commit includes a fix for a transient increase in noise caused by incorrect processing of winds along the lateral boundary on the final acoustic time step before fresh boundary information is ingested into the model.
    • 2022/3/9 00Z LAMDA
  • Replace the NAM B-Matrix (192, 42) with the Global 127-Level B-Matrix (770, 127), and use the tuned hzscl value and anavinfo file.
    • 2022/3/3 00Z LAMDAX
  • Replace the NAM B-Matrix (192, 42) with the Global 127-Level B-Matrix (770, 127), and use the tuned hzscl value and anavinfo file.
    • 2022/2/16 12Z LAM, LAMDA, LAMDAX
  • The UPP code for generating the GOES brightness temperature output was updated in all the LAM parallels (Feb 11: #583f76b)
    • 2022/1/19 00Z LAMDA, LAMDAX
  • The create_expanded_restart_files_for_DA code, which runs before the model in the LAMDA to create larger "placeholder" fv_core and fv_tracer restart files w/ larger halos (used later in the workflow to put the analysis into the 00-h LBC files), was changed so that these files were created in netcdf-4 classic format, and so that the chunk sizes were set correctly to allow the parallel I/O in the GSI to work efficiently (reducing run time by 10-20%).
    • 2022/1/11 00Z LAMDA, LAMDAX
  • The GSI code was updated in LAMDA/LAMDA-X (Dec 21: #7226690)
  • Includes an additional change from the feature/parallel_ioV1 branch (Jan 5: #f7d8619) - adds the nf90_mpiio parameter in the nf90_open arguments for parallel IO for the FV3LAM
    • 2022/1/10 12Z LAM, LAMDA, LAMDAX
  • The ufs-weather-model code was updated (Dec 27: #3d995d7) which includes the following updates to the MYNN PBL scheme:
    1. Improved numerical stability
    2. Removes all negative qv printouts
    3. Bug fix for the interpolation of mass-flux clouds from interface levels to mid-levels which cause excessive LWP in shallow-cumulus fields
    4. Fixes for smoke/chemical mixing (more changes expected in the future)
    5. Improvements to the excessive cloud bias in polar regions
    • 2021/11/30 12Z LAM, LAMDA, LAMDAX
  • The following changes were made to the real-time parallels:
    1. The ufs-weather-model code was updated (Nov 18: #805421d)
    2. All grib2 output now uses 3 digits for the forecast hours
    3. nord was changed from 2 to 3
    4. d4_bg was changed from 0.15 to 0.12
  • The new nord/d4_bg values yield improvements in the precip/reflectivity stats.
    • 2021/11/29 00Z LAMDA, LAMDAX
  • After the emergency production switch on November 5, the halo4 mosaic fix file was inadvertently being used with the LAMDA and LAMDAX forecast jobs. This was fixed with the 11/29 00Z cycle.
    • 2021/10/20 00Z LAMDA, LAMDAX
  • The halo3 mosaic fix file is now used for the LAMDA and LAMDAX forecast jobs. The halo4 mosaic fix file is still used for the pre-processing jobs.
    • 2021/10/19 12Z LAM, LAMDA, LAMDAX
  • The halo3 mosaic fix file is now used for the LAM forecast job. The halo4 mosaic fix file is still used for the pre-processing jobs.
  • The mean absolute pressure tendency per timestep is now available in all parallels.
    • 2021/10/5 00Z LAM, LAMDA
  • The ufs-weather-model code was re-compiled with upp/10.0.10 to resolve an issue with the 1000-mb height field.
    • 2021/10/1 12Z LAMDAX
  • The ufs-weather-model code was re-compiled with upp/10.0.10 to resolve an issue with the 1000-mb height field.
    • 2021/9/23 12Z LAM, LAMDA, LAMDAX
  • vtdm4 was changed from 0.075 to 0.02 in all the parallels.
    • 2021/9/15 12Z LAM, LAMDA, LAMDAX
  • The ufs-weather-model code was re-compiled with upp/10.0.9 and the pratemax field was added to the output.
  • The chgres_cube code was updated (#e67789f), and the combined GFSv16 grib2 files are now used for LBC generation.
    • 2021/8/25 12Z LAM, LAMX
  • The LAMX parallel has been renamed to LAM.
    • 2021/8/13 12Z LAMX, LAMDAX
  • The Puerto Rico output grid was expanded.
    • 2021/8/9 12Z LAM, LAMX, LAMDA, LAMDAX
  • The following changes were made to the real-time parallels:
    1. The ufs-weather-model code has been updated in all parallels (Jul 30: #ddcd809)
    2. Add the GSL drag suite to all parallels.
    3. Use new fix files which contain fixes to topography filtering and unifying the Earth radius between the model and UFS_UTILS codes.
    4. Add psm_bc = 1 to the namelist to improve model stability.
    5. Vertical interpolation in LAMX is now done by chgres_cube instead of by the model to match what is done in LAMDA and LAMDAX.
    6. The LAM CONUS parallel has been turned off. The LAMX will be renamed to LAM soon.
    • 2021/7/26 12Z LAM, LAMX
  • The LAM and LAMX parallels have been updated to use the non-combined GFSv16 grib2 files for LBC generation.
  • The current version of chgres_cube can't handle the combined grib2 files of GFSv16, particularly for specific humidity, because it has fewer isobaric levels than UGRD and other prognostic variables.
    • 2021/7/25 00Z LAMDA, LAMDAX
  • The LAMDA parallels have been updated to use the non-combined GFSv16 grib2 files for LBC generation.
  • The current version of chgres_cube can't handle the combined grib2 files of GFSv16, particularly for specific humidity, because it has fewer isobaric levels than UGRD and other prognostic variables.
    • 2021/6/14 12Z LAM, LAMX, LAMDA, LAMDAX
  • The following changes were made to the real-time parallels:
    1. The ufs-weather-model code has been updated in all parallels (Jun 3: #2c882f0)
    2. Implemented the 65-level vertical configuration in LAM/LAM-X/LAMDA
    3. Added MERRA2 aerosol data to all parallels
    4. The GSI code has been updated in LAMDA/LAMDA-X (May 21: #6cec754)
    5. The LAMDA-X parallel will now run over the LAM-X (RRFS/North American) domain
    • 2021/5/21 00Z LAMDAX
  • Removed the Valid Time Shifting method and implemented the 65-level vertical configuration.
    • 2021/5/13 00Z LAMDAX
  • Turn off trajectory correcton of the global ensemble used in the LAMDA-X GSI analysis while it is running Valid Time Shifting.
    • 2021/5/6 12Z LAMDAX
  • Removed the change related to the treatment of total mass based pressure and introduced the Valid Time Shifting (VTS, Huang and Wang 2018) method into the hybrid DA.
  • With VTS, the ensemble size is increased by including ensemble members valid at different lead times but initialized at the same previous DA cycle analysis. This effectively increases the ensemble size from 80 to 160 members.
    • 2021/4/26 12Z LAM, LAMDA, LAMDAX
  • The size of the ESG grid for the LAM, LAMDA, and LAMDA-X real-time parallels has been increased.
  • This modification allows us to create output on the HRRR / NAM nest CONUS output grid.
    • 2021/4/15 00Z LAM, LAMX, LAMDA, LAMDAX
  • sg_cutoff = 10000.0 was added to the namelists of all the real-time parallels.
  • sg_cutoff overrides n_sponge, allowing more control for the depth over which the 2dz damping occurs at upper levels.
    • 2021/4/13 00Z LAM, LAMX, LAMDA, LAMDAX
  • The following aerosol/radiation changes were made to the real-time parallels:
    1. iaer was changed from 111 to 5111
    2. icliq_sw was changed from 1 to 2
    3. iovr was changed from 1 to 3
    • 2021/4/9 00Z LAMDA, LAMDAX
  • The following changes were made to the real-time DA parallels:
    1. LAMDA: Start assimilating direct broadcast (DB) satellite radiance data
    2. LAMDAX: GSI interface for FV3-LAM changed to correctly treat total mass based pressure (moisture air plus hydrometeors) from warm restart files
    • 2021/3/29 12Z LAM, LAMX, LAMDA, LAMDAX
  • The following changes were made to the real-time parallels:
    1. The inline post capability is utilized, eliminating the need to run a separate UPP job in the workflow.
    2. The MYNN surface layer scheme replaces the GFS surface layer scheme.
    3. LAMDAX only: The assimilation of surface observations was turned back on, and the GSI will start using direct broadcast (DB) satellite radiance data.
    • 2021/2/23 00Z LAMDA, LAMDAX
  • Corrected an error in the GSI script which led to no aircraft data being assimilated.
    • 2021/2/18 12Z LAMDA, LAMDAX
  • The following changes have been made to the DA parallels:
  • LAMDA: DA cycle is coldstarted directly from a GDAS 6-h forecast (previous LAMDAX configuration)
  • LAMDAX: Assimilation of all surface wind, temperature, and moisture observations is turned off
    • 2021/2/11 12Z LAMX
  • The LAMX parallel will use an n_split of 5 instead of 6.
  • The model failure that prompted the n_split increase to 6 was successfully re-run with n_split = 5.
    • 2021/1/29 00Z LAMX
  • The LAMX parallel has switched from 3-hourly LBC updates to hourly LBC updates.
  • A bug was discovered in the LAMX namelist which was causing it to incorrectly use 3-hourly LBC updates.
    • 2021/1/27 00Z LAMDA, LAMDAX
  • Xiaoyan Zhang fixed a small bug in the move_BC_update_data routine.
  • The delz values from the GSI analysis at the lateral boundary were being overwritten by extrapolated (or dummy) values of delz created in a later part of the code.
    • 2021/1/8 00Z LAM, LAMX, LAMDA, LAMDAX
  • The LAM parallels have been updated to use GFS grib2 files for LBC generation instead of GFS nemsio files.
    • 2020/12/21 12Z LAM, LAMX, LAMDA, LAMDAX
  • The LAM parallels have been updated to use more recent ufs-weather-model code (Dec 2: #63591b6).
  • The newer code includes a fix to allow for quilting with non-factors for the layout, as well as changing dz_min from 2m to 6m. If delz is less than dz_min, it is set to dz_min.
    • 2020/12/19 00Z LAMDAX
  • Official start of the FV3LAM-DA-X parallel.
  • The FV3LAM-DA-X parallel is the same as the FV3LAM-DA, but with the DA cycle coldstarted directly from a GDAS 6-h forecast, instead of from a 6-h LAM forecast as is done in the LAM-DA control run.
    • 2020/12/03 00Z LAM, LAMX, LAMDA
  • effr_in has been changed from false to true.
  • effr_in is a flag whch determines whether or not to use effective radii of cloud species in radiation.
    • 2020/11/24 12Z LAM, LAMDA
  • The LAM and LAMDA parallels will use the extended Schmidt gnomonic (ESG) grid.
    • 2020/11/12 12Z LAMX
  • The LAMX will use an n_split of 6 instead of 5.
  • Increasing n_split from 5 to 6 resolved a model failure.
    • 2020/11/05 12Z LAM, LAMX, LAMDA
  • The LAM, LAMX, and LAMDA have switched from 3-hourly LBC updates to hourly LBC updates.
    • 2020/10/23 12Z LAMDA
  • LAMDA is again using RAPv5 radar winds. Shun Liu fixed a bug in the GSI routine that reads this data.
    • 2020/10/22 12Z LAM, LAMX, LAMDA
  • hord_dp has been changed from -6 to +6. This change was recommended by GFDL. Tests were conducted and minimal impacts are expected.
    • 2020/10/15 12Z LAMX
  • Official start of the FV3LAM-X parallel. The FV3LAM-X is identical to the FV3LAM but runs over a 3-km North American domain.
    • 2020/10/14 LAMDA
  • Starting with the 00Z cycle, update the geometric layer depth (delz) in the GSI analysis.
  • Starting with the 12Z cycle, force the GSI jobs to use the NEXRAD level 2 radar wind obs dumped for the NAM instead of those for parallel RAPv5. Since LAMDA moved from the Cray to Dell P3.5, the GSI puts an error message saying it can not read the RAPv5 radar winds. The error does not occur when using NAM-based radar wind obs created on the Dell.
    • 2020/10/05 12Z LAM, LAMDA
  • The gravity wave drag parameters for the LAM and LAMDA parallels were modified.
  • cdmbgwd was changed from 3.5, 0.01 to 0.88, 0.04.
    		•
    2020/09/02 12Z LAM
  • The LAM parallel (formerly SAR-X) has been moved from the WCOSS Cray to the WCOSS Dell 3.5.
  • The SAR-X parallel has been shut off, and a LAM-X parallel will be coming soon.
    • 2020/08/31 12Z LAMDA
  • The LAM-DA parallel (formerly SAR-DA) has been moved from the WCOSS Cray to the WCOSS Dell 3.5.
    		•
    2020/07/21 12Z SAR, SAR-X, SARDA
  • The ufs-weather-model code has been updated to use the latest version from the develop branch.
  • Refer to the following links for more information about what was changed (recommend opening links in a new tab):
    1. Update GFDL dynamic core to the GFDL 201912 release (GFDL tech memo)
    2. Update GFDL dynamic core to the GFDL 201912 release (GitHub PR)
    3. Tom's dycore regional DA fix (GitHub PR)
    4. Update develop from gsd/develop 20200630 (GitHub PR)
    • 2020/07/01 00Z SAR
  • Parameterized convection has been turned off in the FV3SAR.
  • A bug was discovered where even if convection is turned off in the namelist, it must also be commented out in the CCPP suite definition file.
    • 2020/06/12 12Z SAR, SAR-X
  • halo0 versions of the static fix files (maximum_snow_albedo, snowfree_albedo, slope_type, soil_type, vegetation_type, vegetation_greenness, substrate_temperature, facsf) are now used by the forecast job instead of the halo4 versions.
  • We originally switched to use the halo0 files on 5/22. The change was backed out when there was a production switch on 5/28.
    • 2020/06/03 00Z SAR, SAR-X, SARDA
  • fhcyc was changed from 24 to 0.
  • If fhcyc is set to anything other than 0 in the input.nml, climatology fields will be applied at initialization and at every fhcyc interval (in hours) afterward during the forecast.
    • 2020/05/08 SARDA
    1. Wind rotation fix (obs were rotated to grid-relative, background winds were on earth-relative for FV3SAR option)
    2. GOES-17 radiances turned on in monitor mode
    2020/05/04 SARDA
    1. Changed to use Thompson microphysics and MYNN PBL physics (same as SARX)
    2. The SARDA parallel now producing output directly onto the lambert conformal grid.
    3. Using observations from the parallel RAPv5 dumps.
    2020/04/22 SAR, SAR-X
  • The SAR and SAR-X parallels now produce output directly onto the lambert conformal grid.
    • 2020/04/18 SARDA
  • Start of SAR parallel initialized from a 6-h data assimiltion cycle wih hourly analysis updates on the WCOSS Cray.
  • Uses same model settings as SAR parallel
    1. FV3GFS v15 CCPP suite with GFDL microphysics
    2. Vertical remapping (kord=9) and timestepping (dt_atmos=36, n_split=5)
    3. Call radiation every 15 mins instead of every 20 mins.
    4. Output data in compressed netcdf format.
    5. n_sponge=9 (number of levels above 100 mb)
    6. Produce bufr sounding output.
    7. Runs with DA boundary cnditions changes (boundary blending code with 10 blending rows, insert GSI analysis is boundary tendency file).
    • 2020/04/01 SAR, SAR-X
  • The SAR and SAR-X parallels have been moved from WCOSS phase 2 to the WCOSS Cray.
  • The vertical remapping (kord=9) and timestepping (dt_atmos=36, n_split=5) changes in SAR-X have been added to the SAR.
  • The SAR-X parallel is now running with the Thompson microphysics and MYNN PBL schemes.
  • Both parallels now call radiation every 15 mins instead of every 20 mins.
  • Both parallels now output data in compressed netcdf format.
  • Both parallels now produce bufr sounding output.
  • Both parallels have switched from the IPD to the CCPP physics driver.
  • Both parallels have switched from using global_chgres to chgres_cube.
  • Both parallels now utilize the boundary blending code with 10 blending rows.
  • Both parallels now use 9 sponge layers instead of 24.
    • 2020/01/24 SAR-X
  • The values of dt_atmos and n_split were changed to 36 and 5 in the SAR-X parallel. These settings resolved the model failures.
    • 2020/01/16 SAR-X
  • The values of dt_atmos and n_split were reverted back to 30 and 6 in the SAR-X parallel due to several model failures.
    • 2020/01/14 SAR, SAR-X
  • A bug fix was applied to the post code used by the SAR and SAR-X parallels.
  • The winds coming out of the FV3 model are Earth-relative, and the UPP had been incorrectly assuming the winds were grid-relative.
    • 2019/12/13 SAR, SAR-X
  • Both sets of difference plots were updated to use pcolormesh instead of contourf.
  • It was discovered that contourf does not accurately portray the difference fields.
    • 2019/11/22 SAR, SAR-X
  • The FV3SAR parallel now uses the NAM vertical level structure.
  • Radiation is now called every 20 mins instead of every 60 mins in both parallels.
  • The FV3SAR-X parallel now uses dt_atmos = 36, k_split = 2, n_split = 4.
  • kord has been changed from 11 to 9 in the FV3SAR-X parallel.
    • 2019/09/09 SAR-X
  • The FV3SAR-X parallel now uses the NAM vertical level structure. There are 60 vertical levels, with the lowest level at ~20 m AGL and the highest level at 2 mb.
    • 2019/08/29 SAR, SAR-X
  • Hord had inadvertently been set to 5 since the real-time runs were moved from the Dell to phase 2. It was switched back to 6.
    • 2019/07/30 SAR, SAR-X
  • FV3SAR forecasts have been moved from the Dell to phase 2.
  • Official start of the 12Z cycle of the FV3SAR run.
  • dt_atmos was changed from 90 to 30 secs and k_split was changed from 6 to 2.
    • 2019/07/22 Nest
  • The FV3NEST parallel has been turned off.
    • 2019/06/19 SAR, SAR-X
  • Official start of the FV3SAR-X parallel. The FV3SAR-X is an experimental version of the SAR, where the FV3SAR is the control run.
    • 2019/06/02 SAR, SAR-DA
  • The FV3SAR-DA parallel was turned off.
    • 2019/05/16 Nest, SAR, SAR-DA
  • nstf_name is changed to 2,0,0,0,0 to match what is used in the FV3GFS.
  • Changing the second flag from 1 to 0 means the NSST data will be read in from the sfc data files instead of cold started.
    • 2019/04/22 Nest, SAR, SAR-DA
  • NCO's official FV3GFS parallel has picked up the radiation and snow changes, so these will now be incorporated into the input files for chgres.
    • 2019/04/13 Nest, SAR, SAR-DA
  • hybedmf is changed from false to true. This turns on the hybrid EDMF PBL scheme which was implemented in the operational GFS in 2015.
  • Previously, the 2011 version of the GFS PBL scheme was being used.
    • 2019/04/06 Nest, SAR
  • Difference plots comparing the 3-km FV3SAR to the 3-km NAM CONUS nest have been added.
  • The Nest & SAR plots were modified such that the difference fields are SAR - Nest instead of Nest - SAR.
    • 2019/03/27 Nest, SAR, SAR-DA
  • The regional domains were refined.
  • 6 new domains were added: Boston-NYC, Central, Colorado, Seattle-Portland, SF Bay Area, LA-Vegas
    • 2019/03/22 Nest, SAR, SAR-DA
  • Plots of max/min hourly diagnostics have been added.
    • 2019/03/15 Nest, SAR, SAR-DA
  • Updated versions of the high resolution tiled surface data were inserted into the regional FV3 parallels.
  • These updated files were generated using 1-km data and contain more detail.
    • 2019/03/12 Nest
  • The orography for the FV3NEST (tile 7) will be the same as that used for the FV3SAR.
  • This will eliminate differences in how the terrain is handled for both configurations.
    • 2019/03/07 Nest, SAR, SAR-DA
  • High resolution (~3-km) tiled surface data was inserted into the regional FV3 parallels.
  • Previously, Gaussian files on a ~13-km grid were being used as fix files for the 3-km model.
    • 2019/03/05 Nest, SAR, SAR-DA
  • Radiation and snow changes that were added to the FV3GFS master repository were added to the regional FV3 parallels.
    • 2019/02/25 Nest, SAR, SAR-DA
  • Forecasts are now initialized off of NCO's official FV3GFS parallel.
    • 2019/02/19 Nest, SAR, SAR-DA
  • The cold low-level temperature bias currently observed in the global FV3GFS is also apparent in the regional FV3 parallels.
    • 2019/01/31 Nest, SAR, SAR-DA
  • A Mid-Atlantic domain has been added.
  • The Northeast domain has been modified.
    • 2019/01/30 Nest
  • The FV3NEST contains a persistent warm/moist bias that is most prevalent over the Pacific northwest.
  • Preliminary investigations reveal the FV3SAR is closer to reality than the FV3NEST.
  • Differences in how terrain is handled in the FV3NEST vs the FV3SAR may be a contributing factor.
    • 2018/12/14 Nest, SAR
  • Individual plots for the 3-km FV3NEST runs have been removed.
  • Difference plots comparing the 3-km FV3SAR to the 3-km FV3NEST have been added.
    • 2018/12/10 SAR-DA
  • FV3SAR-DA runs contain significantly more noise in the first 6 hours of the forecast when compared to the FV3SAR runs.
  • The DA increments are likely introducing a model imbalance.
    • 2018/11/30 SAR, SAR-DA
  • Individual plots for the 3-km FV3SAR runs have been removed.
  • Difference plots comparing the 3-km FV3SAR coldstarted off the FV3GFS to the 3-km FV3SAR-DA with a 6-h data assimilation cycle have been added.
    • 2018/11/08 Nest, SAR
  • 3-km FV3SAR and 3-km FV3NEST forecasts have been moved from phase 2 to the Dell.
    • 2018/10/24 Nest, SAR
  • The central latitude of the domain was changed from 35.5N to 38.5N. This change will allow the 3 km HRRR output grid to have no undefined grid points outside the FV3 integration domain.
    • 2018/07/03 SAR
  • Official start of the 3-km FV3 stand-alone regional (FV3SAR) run. The forecast is coldstarted off of the FV3GFS using 3-hourly boundary conditions.
    • 2018/06/16 Nest
  • Forecasts are now initialized off of EMC's official FV3GFS parallel.
    • 2018/05/09 Nest
  • Reflectivity is now output directly from the GFDL microphysics scheme instead of using the GFDL-modified Stoelinga algorithm.
    • 2018/05/01 Nest
  • Official start of the 3-km FV3 CONUS nest (FV3NEST) run within the FV3GFS global domain.