|
| 1 | +GFS v15.2.0 RELEASE NOTES |
| 2 | + |
| 3 | + |
| 4 | +PRELUDE (taken from GFS v15.2.0 SCN) |
| 5 | + |
| 6 | + GFS version 15.1 was implemented into operation at the 12Z cycle on June 12, 2019. It was the first |
| 7 | + GFS implementation with the finite volume cubed-sphere (FV3) dynamical core as the Weather Service’s |
| 8 | + Next Generation Global Prediction System (NGGPS). |
| 9 | + |
| 10 | + GFS version 15.2 is a minor upgrade. The major change to the system is to ingest new and replacement |
| 11 | + satellite observations for data assimilation. It also contains a few other minor upgrades and bug fixes. |
| 12 | + |
| 13 | + 1) Assimilate new satellite observations |
| 14 | + * GOES-17 AMVs |
| 15 | + * GOES-17 has already replaced GOES-15 as the operational GOES-West satellite. Distribution of |
| 16 | + GOES-15 products is expected to cease around December 2019 and active assimilation of GOES-17 |
| 17 | + AMVs is required to prevent a gap in data coverage. |
| 18 | + * Metop-C AMSU and MHS |
| 19 | + * Metop-C is now the primary satellite in the 9:30 AM orbit. Assimilation of these data provide |
| 20 | + additional impact as well as adding robustness to the system. |
| 21 | + * KOMPSAT-5 (GPS-RO) |
| 22 | + * Provides additional robustness to the system. |
| 23 | + * Addition changes are made to VIIRS AMV ingest code to allow continued use after an expected change |
| 24 | + to the BUFR format. |
| 25 | + |
| 26 | + 2) Assimilate buoyb sea-surface temperature (SST) data |
| 27 | + TAC2BUFR changes in the buoy network resulted in a reduction of available SST measurements from buoys |
| 28 | + to 10% or less of expected levels. Obsproc and GSI changes were required to restore data counts to |
| 29 | + previous levels. |
| 30 | + |
| 31 | + 3) New product: Graphical Turbulence Guidance (GTG) |
| 32 | + Global Turbulence product generated using NCAR GTG algorithm will start being disseminated on NOMADS. |
| 33 | + The product will be available 3 hourly from F06 to F36 and horizontal resolution will be a quarter |
| 34 | + degree. gfs.t{CC}Z.gtg.0p25.grb2f{HHH} |
| 35 | + |
| 36 | + 4) Update the Unified Post Processor(UPP) to address a mask issue of several land surface fields over |
| 37 | + water bodies in grib2 products. |
| 38 | + |
| 39 | + This update will make GFS.v15.2 p-grid products to be more consistent with GFS.v14 products, and |
| 40 | + remove spurious soil moisture along coastlines. These land surface fields include Soil Moisture, |
| 41 | + Soil Temperature, Liquid Volumetric Soil Moisture, WEASD, Snow Depth, Water Runoff, GFLUX Ground |
| 42 | + Heat Flux, WILT Wilting Point, and FLDCP Field Capacity. |
| 43 | + |
| 44 | + Affected product files are: |
| 45 | + gfs.t{CC}Z.pgrb2.0p25.F{HHH} |
| 46 | + gfs.t{CC}Z.pgrb2b.0p25.g{HHH} |
| 47 | + gfs.t{CC}Z.pgrb2.0p50.F{HHH} |
| 48 | + gfs.t{CC}Z.pgrb2b.0p50.g{HHH} |
| 49 | + gfs.t{CC}Z.pgrb2.1p00.F{HHH} |
| 50 | + gfs.t{CC}Z.pgrb2b.1p00.g{HHH} |
| 51 | + gfs.t{CC}Z.sfluxgrbf{HHH}.grib2 |
| 52 | + Where CC is cycle for 00, 06, 12, 18 UTC, and HHH is forecast hour. |
| 53 | + |
| 54 | + 5) The vessel icing program uses OISST as input. OISST will not be ported from WCOSS Phase 1 to Phase 3 |
| 55 | + after Phase 1 is decommissioned in 2020. A decision was made to move the vessel icing program |
| 56 | + within the Unified Post Processor(UPP) and use GFS forecast skin temperature as input. Current vessel |
| 57 | + icing product in operation (sice.tCCz.siceg) has a 1-deg resolution and is disseminated to the public at |
| 58 | + |
| 59 | + ftp://ftp.ncep.noaa.gov/pub/data/nccf/com/omb/prod/sice.yyyymmdd/ |
| 60 | + |
| 61 | + New vessel icing product will be included as a variable (ICEG) in GFS p-grid products gfs.tCCz.pgrb2.xpxx.fxxx |
| 62 | + and gfs.tCCz.pgrb2b.xpxx.fxxx at 0.25, 0.5, and 1.0-deg resolutions, and be disseminated to the public at |
| 63 | + |
| 64 | + ftp://ftp.ncep.noaa.gov/pub/data/nccf/com/gfs/prod/gfs.yyyymmdd/ |
| 65 | + |
| 66 | + 6) Added three stations to station time series bufr soundings: |
| 67 | + |
| 68 | + 006011 62.02N 6.76W TOR 00 Thorshvan, Denmark 54 Usr Rqst 4-19 |
| 69 | + 999913 15.51S 128.15E WYN 00 Wyndham aerodrome Australia 4 Usr Rqst 1-19 |
| 70 | + 999914 57.48N 7.36W EGPL 00 Benbecula, Scotland, UK 6 Usr Rqst 1-19 |
| 71 | + |
| 72 | + The affected output files are: |
| 73 | + gfs_yyyymmddhh.sfc |
| 74 | + gfs_yyyymmddhh.snd |
| 75 | + gfs.tCCz.bufrsnd.tar.gz |
| 76 | + |
| 77 | + Three additional files for the stations: |
| 78 | + bufr.006011.yyyymmddhh |
| 79 | + bufr.999913.yyyymmddhh |
| 80 | + bufr.999914.yyyymmddhh |
| 81 | + |
| 82 | + 7) Reduction of water temperature biases in small lakes. |
| 83 | + For small lakes adequate observations do not always exit to support the analysis of lake surface |
| 84 | + temperature, often leading to significant departures from both the climatology and real-time observation. |
| 85 | + Two changes were introduced to ensure that lake temperatures do not deviate from the climatology when |
| 86 | + observations are not available. The first change is to replace a surface mask file at 0.5-degree |
| 87 | + resolution with the one on the T1534 Gaussian grid (~13km) to prevent unrealistic SST climatology |
| 88 | + from being used for updating the background of the near sea-surface temperature analysis over small |
| 89 | + water bodies, such as those in the Great Salt Lake. The second change is to reduce the relaxation |
| 90 | + time scale of the SST to climatology in GDAS forecast step from 90 days to 10 days. |
| 91 | + |
| 92 | + 8) Changes to NOAAPORT/SBN |
| 93 | + Product Removals |
| 94 | + * GADS FAX product which contains tropical surface analysis in TIF format with G4 compression. |
| 95 | + |
| 96 | + |
| 97 | + |
| 98 | +IMPLEMENTATION INSTRUCTIONS |
| 99 | + |
| 100 | + * NOAA Vlab GIT is used to manage GFS.v15.2 code. The SPA(s) handling the GFS.v15.2 implementation need to have |
| 101 | + permission to clone Vlab gerrit repositories. So far Wojciech Cencek has been given access to all GFS.v15.2 |
| 102 | + related git repositories. Please contact [email protected] or [email protected] if there is any VLAB |
| 103 | + access issue and/or the individual code managers listed under item #6) below. Please follow the following |
| 104 | + steps to install the package on WCOSS DELL |
| 105 | + |
| 106 | + 1) cd $NWROOTp3 |
| 107 | + 2) mkdir gfs.v15.2.0 |
| 108 | + 3) cd gfs.v15.2.0 |
| 109 | + 4) git clone --recursive gerrit:global-workflow . |
| 110 | + 5) git checkout feature/dev-v15.2 |
| 111 | + 6) cd sorc |
| 112 | + 7) ./checkout.sh |
| 113 | + This script extracts the following GFS components from gerrit |
| 114 | + MODEL -- tag nemsfv3_gfsv15.2.1 [email protected] |
| 115 | + GSI -- tag fv3da_gfs.v15.2.0 [email protected] |
| 116 | + UPP -- tag ncep_post_gtg.v1.1.4 [email protected] |
| 117 | + WAFS -- tag gfs_wafs.v5.0.9 [email protected] |
| 118 | + |
| 119 | + |
| 120 | + 8) ./build_all.sh |
| 121 | + *This script compiles all GFS components. Runtime output from the build for each package is written |
| 122 | + to log files in directory logs. To build an individual program, for instance, gsi, use build_gsi.sh. |
| 123 | + |
| 124 | + 9) ./link_fv3gfs.sh nco dell |
| 125 | + |
| 126 | + * Note: 1) ecflow suite definition and scripts are saved in gfs.v15.2.0/ecflow/ecf |
| 127 | + 2) ncep_post_gtg.v1.1.4 contains restricted GTG (Graphic Turbulence Guidance) code provided by |
| 128 | + NCAR. Please do not post the GTG code in any public domain. |
| 129 | + |
| 130 | + |
| 131 | + |
| 132 | + |
| 133 | +JOB CHANGES |
| 134 | + |
| 135 | + * See docs/Release_Notes.gfs_downstream.v15.2.0.txt |
| 136 | + |
| 137 | + |
| 138 | +SORC CHANGES |
| 139 | + |
| 140 | + * sorc/ |
| 141 | + * checkout.sh - update to check out the following tags |
| 142 | + * NEMSfv3gfs nemsfv3_gfsv15.2.1 |
| 143 | + * ProdGSI fv3da_gfsv15.2.0 |
| 144 | + * EMC_post_gtg ncep_post_gtg.v1.1.4 |
| 145 | + * EMC_gfs_wafs gfs_wafs.v5.0.9 |
| 146 | + * sorc/global_chgres.fd/ |
| 147 | + * sfcsub.F - bug fix. set logical variable (lmask) to be false for a land surface variable |
| 148 | + * sorc/global_cycle.fd/ |
| 149 | + * sfcsub.F - bug fix. set logical variable (lmask) to be false for a land surface variable |
| 150 | + |
| 151 | + |
| 152 | +SCRIPT CHANGES |
| 153 | + |
| 154 | + * scripts/ |
| 155 | + * run_gfsmos_master.sh.cray - remove reference to COMROOTp1 |
| 156 | + * run_gfsmos_master.sh.dell - remove reference to COMROOTp1 |
| 157 | + * additional script changes documented in docs/Release_Notes.gfs_downstream.v15.2.0.txt |
| 158 | + |
| 159 | + |
| 160 | +PARM/CONFIG CHANGES |
| 161 | + |
| 162 | + * parm/ |
| 163 | + * Two files are modified to set a |
| 164 | + * parm/config/config.base.emc.dyn - set 10 day relaxaion time scale to SST climatology in GDAS forecast |
| 165 | + * parm/config/config.base.nco.static - set 10 day relaxaion time scale to SST climatology in GDAS forecast |
| 166 | + |
| 167 | + * Two files were modified for adding three bufr stations: |
| 168 | + * parm/product/bufr_stalist.meteo.gfs |
| 169 | + * parm/product/bufr_stalist.meteo.gfs3 |
| 170 | + |
| 171 | + |
| 172 | +FIX CHANGES |
| 173 | + |
| 174 | + * Files in fix/fix_gsi altered by GFS DA v15.2. See GFS DA v15.2 release notes |
| 175 | + (sorc/gsi.fd/doc/Release_Notes.gfs_da.v15.2.0.txt) for details |
| 176 | + |
| 177 | + |
| 178 | +PRODUCT CHANGES |
| 179 | + |
| 180 | + * see SCN |
| 181 | + |
| 182 | + |
| 183 | +RESOURCE INFORMATION |
| 184 | + |
| 185 | + * Frequency of run |
| 186 | + * No change from GFS v15.1 |
| 187 | + |
| 188 | + * Commonly used libraries, compiler, and modules are defined in gfs.v15.2.0/modulefiles. For nemsfv3gfs, gsi, upp, wafs |
| 189 | + they maintain their own module files under gfs.v15.2.0/sorc/(fv3gfs gsi gfs_post global_wafs).fd/modulefiles |
| 190 | + * GSI updated to use bufr/11.3.0 |
| 191 | + |
| 192 | + * Data retention time under $COMROOTp3 for GFS.v15.2 should be the same as GFS.v15.1. |
| 193 | + |
| 194 | + * Disk space: |
| 195 | + * About 4 Gb more per gdas cycle due to additional observation data in gdas and enkf diagnostic files |
| 196 | + |
| 197 | + * Computational resources and run times: |
| 198 | + * Due to processing additional observation data the runtime for the following jobs increases with |
| 199 | + respect to GFS v15.1 as noted below |
| 200 | + * gfs_analysis : about 30 seconds longer (27.4 minutes for GFS v15.1 -vs- 27.9 minutes for GFS v15.2) |
| 201 | + * gdas_enkf_select_obs : about 1 minute longer (3.7 for GFS v15.1 -vs- 4.7 for GFS v15.2) |
| 202 | + * gdas_enkf_innovate_obs_grp*: about 30 seconds longer (14.8 for GFS v15.1 -vs - 15.3 for GFS v15.2) |
| 203 | + * gdas_enkf_update : about 20 seconds longer (6.4 for GFS v15.1 -vs- 6.7 for GFS v15.2) |
| 204 | + |
| 205 | + |
| 206 | + |
| 207 | +PRE-IMPLEMENTATION TESTING REQUIREMENTS |
| 208 | + |
| 209 | + * Which production jobs should be tested as part of this implementation? |
| 210 | + * The entire GFS v15.2 package needs to be installed and tested. EMC can run the same date |
| 211 | + and compare NCO and EMC output to confirm the EMC and NCO tests reproduce each other |
| 212 | + |
| 213 | + * Does this change require a 30-day evaluation? |
| 214 | + * No. |
| 215 | + |
| 216 | + |
| 217 | + * Suggested evaluators |
| 218 | + * Please contact the following EMC staff for the indicated components |
| 219 | + |
| 220 | + |
| 221 | + |
| 222 | + |
| 223 | + [email protected] - downstream products |
| 224 | + |
| 225 | + |
| 226 | +DISSEMINATION INFORMATION |
| 227 | + |
| 228 | + * Where should this output be sent? |
| 229 | + * No change from GFS v15.1 |
| 230 | + |
| 231 | + * Who are the users? |
| 232 | + * No change from GFS v15.1 |
| 233 | + |
| 234 | + * Which output files should be transferred from PROD WCOSS to DEV WCOSS? |
| 235 | + * No change from GFS v15.1 |
| 236 | + |
| 237 | + * Directory changes |
| 238 | + * No change from GFS v15.1 |
| 239 | + |
| 240 | + * File changes. |
| 241 | + * See SCN |
| 242 | + |
| 243 | + |
| 244 | +HPSS ARCHIVE |
| 245 | + |
| 246 | + No change from GFS v15.1 |
| 247 | + |
| 248 | + |
| 249 | + |
| 250 | +JOB DEPENDENCIES & FLOW DIAGRAM |
| 251 | + * No change from GFS v15.1 |
| 252 | + |
| 253 | + |
| 254 | +=========== |
| 255 | +Prepared by |
| 256 | +Fanglin.Yang@noaa |
| 257 | + |
| 258 | + |
| 259 | + |
| 260 | + |
| 261 | + |
0 commit comments