GFS/GDAS CHANGES SINCE 1991
History of recent modifications to the global forecast/analysis system
Some of the changes (TPB > 442) are documented in
Global Modeling Branch Technical Procedures Bulletins(TPB)
Another source of information is the
Parallel run home page
- 08/80: R30L12
- 10/83: R40L12
- 04/85: MRF85 R40L18 - GFDL physics
- 06/86: MRF86 R40L18 - convection to tropopause
- 08/87: MRF87 T80L18 - diurnal cycle, moisture all layers
- 01/88: MRF88 T80L18 - interactive clouds
- 03/91: T80L18 to T126L18, silhouette to mean orog, new SST,
marine stratus, reduced horizontal diffusion
- 06/91: SSI analysis
- 08/93: Arakawa-Schubert convection, 28 layers
- 01/95: SSM/I PW over sea; analysis changes; various physics changes (TPB 417 (not on Web))
- 10/95: Satellite radiances instead of T retrievals; ERS-1 winds; physics changes (TPB 428 (not on Web))
- 06/96: Adjustments to PBL and convection
- 02/97: SSM/I water vapor discontinued
- 11/97: Elimination of "valley snow" and other small changes
- 01/98: TOVS-1b radiances; vertical diffusion inc in free atmosphere
- 06/15/98: T170L42 to day 3.5; physics changes; 3D ozone; analysis changes
(See TPB 449)
- 07/22/98: First emergency implementation: fix to convection, horizontal diffus,
plant evap: see last section of TPB 449 above.
- 10/06/98: Second emergency implementation: Back to T126L28, better fits to data, more
iterations in analysis, physics changes
- 12/01/98: Snow resolution increased from 2.0 to 0.5 degrees; snow depth field added;
snow depth no longer estimated by model
- 01/08/99: Use of VAD winds eliminated due to problems with light wind speeds
- 03/08/99: Introduction of high-resolution data from the
The data are radiances from the AMSU-A and HIRS-3 intruments.
The NOAA-11 Satellite is no longer providing AMSU data and will
soon be unable to provide HIRS data.
- 01/24/2000: Resolution upgraded from T126L28 to T170L42, restoring the resolution
used from June 15 through Oct 5, 1998. No other changes in the
forecast/analysis system. The MRF is run at T170L42 through day 7,
then at T62L28 through day 16. The AVN is run at T170L42 out to 84
hours four times a day.
(See TPB 452)
- 05/17/2000: AVN available out to 126h at full (T170) resolution at 00Z and 12Z.
- 06/27/2000: Resolution of ensemble members increased from T62 to T126 for first 60 hours of forecast.
- 07/06/2000: Hurricanes and tropical storms in the model's guess field
are relocated to the official Tropical Prediction Center position in each
6-hour analysis cycle.
( TPB 472)
- 08/29/2000 18Z: Data cutoff time for the 06Z and 18Z final analysis (FNL)
extended from 4 hours to 6 hours.
- 10/01/2000 06Z: Package of minor changes:
Observations and analysis:
New obs error diagnosis,rawinsonde radiation correction,
effects of balloon drift in time and space included
Forecast and post processing:
Improved orography, reduced gaussian grid over polar regions, new surface albedo
climatology, single scattering albedo adjusted
- 11/01/2000 12Z: ERA wind data erratic; turned off.
- 02/13/2001 12Z: Satellite radiance and moisture analysis changes plus smoothed output MSLP.
- 05/15/01 12Z:
(See TPB 484).
- Inclusion of cloud condensate as a history variable
- Use of the cloud condensate in the calculation of radiative transfer
- Stronger quality control for AMSU radiances
- Momentum mixing included in deep convection
- Refinement of hurricane relocation algorithm
- SST anomaly damped toward climatology during forecast, with 90-day relaxation time
- 06/27/01 12Z: Minor increase in vegetation fraction.
- 07/24/01 12Z: The data source for the daily 1-degree sea surface
temperature analysis is changed from NOAA-14 to NOAA-16
due to instrument drift.
- 08/15/2001 06Z: Package of minor changes:
Observations and analysis:
Higher resolution sea ice mask
Forecast and post processing:
Bug corrections in gravity wave drag, randomization of convective cloud tops,
land surface evaporation with trace snow cover; minor adjustments in
effective radius for ice crystals, autoconversion rate for ice, evaporation
of falling precip, and critical RH for condensation
- 10/09/2001: Snow depth is updated daily at 00Z from observations;
at 06Z, 12Z and 18Z the model guess is used. Formerly the same
00Z update was reinserted in each of the day's cycles.
- 01/15/2002 12Z: Quikscat surface winds included.
- 03/05/2002 : AVN runs four times a day out to 384 hours.
Resolution is T170L42 to 180h, thereafter T62L28.
- 04/23/2002 00Z: MRF is replaced by the 00Z AVN. Look-alike MRF
grids will still be available daily for several months on th
ftp server at
- Sept-Oct 2002: Name changes:
The AVN will be referred to as the Global Forecast System model (GFS).
- 10/29/02 12Z:(See draft TPB)
- Resolution change
old: T170L42 to 180h, T62L28 to 384h
new: T254L64 to 84h, T170L42 to 180h, T126L28 to 384h
- Analysis and observation changes: background error recomputed,
AMSU-A channels 12 and 13 from NOAA-15 and NOAA-16 and HIRS from NOAA-16 used,
METAR surface pressure observations used, divergence tendency constraint
in tropics turned off.
- 03/11/03 12Z: NOAA 17 1B radiances assimilated, NOAA-16 AMSU-A radiances
restored, QuikSCAT winds superobbed at 0.5 degrees.
- 08/28/03 12Z: RRTM longwave radiation from AER installed:
More trace gases (CH4, N2O, CFC's)
and better tropospheric water vapor absorption.
- 10/28/03 12Z: NOAA-17 AMSU-A radiances turned off
- 11/20/03 12Z: Package of minor analysis changes
- 12/09/03 12Z: Vertical diffusion added to ozone
- 02/24/04 12Z: Mountain blocking: Parameterization of the separation of
airflow in the vertical with passage over mountainous terrain.
- 03/09/04 12Z: Ensemble run 4 times daily. Resolution T126 0-180 h,
then T62 to 384 h.
- 05/25/04 12Z: Turn off NOAA-16 HIRS/3 observations
- 05/31/05 12Z:(See TPB
- includes changes made 06/14/05 and 07/07/05)
- Resolution change:
old: T254L64 to 84h, T170L42 to 180h, T126L28 to 384h
new: T382L64 to 180h, T190L64 to 384h
- Increase mountain blocking, decrease vertical diffusion and modify sea ice.
- Land Surface Model: Upgraded from OSU 2-L LSM to 4-L NOHA LSM.
- 06/14/05 12Z: Increase canopy resistance of vegetation
- 07/07/05 12Z: Correct error in temperature near top of model
- 01/31/05 18Z: Radiation code corrected to prevent blowups in ensemble forecasts
- 05/30/06 12Z: GFS Ensembles now have 14 members, 4 times daily. Each cycle now
has its own control member. Resolution is T126L28 through 384 h. New ensemble
transform method used to generate members.
- 08/22/06 12Z:(See Briefing Notes):
- New orography&land-sea mask, improved snow analysis, and new ozone physics.
- New glacial ice and fix error in downward LW radiation at earth's surface.
- Infrastructure changes (e.g. Full ESMF compliance).
- 05/01/07 12Z:(See Briefing Notes):
- Change vertical coordinate to hybrid sigma-pressure, reducing some upper air model errors.
- Unify the NCEP 3DVAR assimilation system under the GSI, improving some performance metrics without affecting others and preparing for future analysis improvements.
- Add new observing systems.
- Modernize the radiation package.
- Increase output particularly for hydrology.
- 09/25/07 12Z: GSI Upgrade
- Modify GSI code to permit assimilation of SBUV-8 and JMA wind data.
- Upgrade to latest version of GSI. Incremental improvement due to addition of new data types.
- 12/04/07 12Z: Use of Unified Post Processor in GFS
- 12/04/07 12Z: GFS Ensemble (NAEFS/TIGGE) UPGRADE
- GFS deterministic model bias corrected (bc) output through F180
- GFS bc output used in GEFS processing
- Combine NCEP/GEFS (20 member) and CMC/GEFS (20 member) to produce:
Ensemble mean and anomaly, ensemble spread and mode along with 10%,50%,90%
probability forecast at 1x1 degree resolution.
- Statistical downscaling by using RTMA as a reference: a. At NDGD (5km) resolution
(CONUS), generate mean and mode along with 10%, 50% and 90% probability forecast
- 12/09/08 12Z: Post Change
- Formulation change in freezing level calculation
- Adding several new parameters to output files
- A fix to the calculation of precipitation type
- 02/24/09 12Z: GSI Upgrade
- Inclusion of METOP IASI data; Use of variational qc;
Addition of background error covariance input file;
Reduction of number of airs water vapor changels used
- Change in land/snow/ice skin temperature variance
- Flow dependent reweighting of background error variances
- Use of new version and coefficients for community radiative transfer model
- Modification of height assignment for height based wind observations;
Modification of surface land use file to remove a few
permanent (~12) glacial points to improve surface temperature forecasts over those points.
- 12/15/09 12Z: GFS, GSI and Post upgrades, see NWS TIN for the detail.
- GFS: Upgrading ESMF to Version 3.1.0rp2. Changing output definition of low cloud to combine
the previously separately defined boundary-layer cloud and low cloud. Establishing a baseline/housekeeping
version of the GFS for future implementations.
- GSI: new data sources and improved numerical techniques are
included in GSI. The analysis changes include:
- Assimilating tropical storm pseudo sea-level pressure observations,
NOAA-19 HIRS/4 and AMSU-a brightness temperature, RARS 1b data,
NOAA-18 SBUV/2 and AURA OMI ozone, EUMETSAT-9 atmospheric motion vectors.
- Using uniform thinning mesh for brightness temperature data.
- Improving assimilation of GPs radial occultation data. RE-tuned observation errors.
- Adding dry mass pressure constraint.
- Using goddard modeling and assimilation office/emc merged code including hooks for
4dvar and observation sensitivity.
- Updating background error covariance file.
- Post: the post processors used for GFS and GDAS are
unified and two minor changes are introduced into the GFS.
Unification of the post processing codes results in
slightly less smooth fields in the gdas.
A number of diagnostic variables are improved and additional
parameters aree added to the model output files.
- 07/28/2010 12Z: GFS Q3FY2010 Resolution and Physics Major Upgrade, see NWS TIN and
EMC Verification for more details.
- Resolution and ESMF
* Eulerian T574L64 for fcst1 (0-192hr) and T190L64 for fcst2 (192-384 hr) .
* ESMF 3.1.0rp2
- Radiation and cloud
* Changing SW routine from ncep0 to RRTM2
* Changing longwave computation frequency from three hours to one hour
* Adding stratospheric aerosol SW and LW and tropospheric aerosol LW
* Changing aerosol SW single scattering albedo from 0.90 in the operation to 0.99
* Changing SW aerosol asymmetry factor. Using new aerosol climatology.
* Changing SW cloud overlap from random to maximum-random overlap
* Using time varying global mean CO2 instead of constant CO2 in the operation
* Using the Yang et al. (2008 JAMC) scheme to treat the dependence of direct-beam surface albedo on solar zenith angle over snow-free land surface
- Gravity-Wave Drag Parameterization
* Using a modified GWD routine to automatically scale mountain block and GWD stress with resolution.
* Compared to the T382L64 GFS, the T574L64 GFS uses four times stronger mountain block and one half the strength of GWD.
- Removal of negative water vapor
* Using a positive-definite tracer transport scheme (Yang et al., 2009) in the vertical to replace the operational central-differencing scheme to eliminate computationally-induced negative tracers.
* Changing GSI factqmin and factqmax parameters to reduce negative water vapor and supersaturation points from analysis step.
* Modifying cloud physics to limit the borrowing of water vapor that is used to fill negative cloud water to the maximum amount of available water vapor so as to prevent the model from producing negative water vapor.
* Changing the minimum value of water vapor mass mixing ratio in radiation from 1.0e-5 in the operation to 1.0e-20. Otherwise, the model artificially injects water vapor in the upper atmosphere where water vapor mixing ratio is often below 1.0e-5.
- Hurricane relocation
* Running hurricane relocation at the 1760x880 forecast grid instead of the 1152x576 analysis grid
* Posting GDAS pgb files first on Guassian grid (1760x880), then convert to 0.5-deg for hurricane relocation.
- Post processing and Utility
* Posting GFS forecast master pgb files on Gaussian grid, then copygb to 1-deg for postprocessing and archive.
* Using a 20-bit and faster copygb instead of the operational 16-bit copygb
* Using a new chgres which has double precision and has a fix in dry air mass (pdryini2=0)
- Snow analysis
*Using T574 compatible high-resolution snow analysis
- Upgraded Boundary Layer Scheme
* Include stratocumulus-top driven turbulence mixing.
* Enhance stratocumulus top driven diffusion when condition for cloud top entrainment instability is met.
* Use local diffusion for the nighttime stable PBL.
* Background diffusion in inversion layers below 2.5km over ocean is reduced by 70% to decrease the erosion of stratocumulus along the costal area.
- New mass flux shallow convection scheme
* Detrain cloud water from every updraft layer
* Convection starting level is defined as the level of maximum moist static energy within PBL
* Cloud top is limited to 700 hPa.
* Entrainment rate is given to be inversely proportional to height and detrainment rate is set to be a constant as entrainment rate at the cloud base.
* Mass flux at cloud base is given to be a function of convective boundary layer velocity scale.
- Updated deep convection scheme
* Eliminate Random cloud top, and cloud water is detrained from every cloud layer of the single cloud.
* Finite entrainment and detrainment rates for heat, moisture, and momentum are specified.
* Similar to shallow convection scheme, entrainment rate is given to be inversely proportional to height in sub-cloud layers and detrainment rate is set to be a constant as entrainment rate at the cloud base.
* Above cloud base, an organized entrainment is added, which is a function of environmental relative humidity
* increased momentum background diffusivity for winds only
* convective overshooting: increased cloud water detrainment in upper cloud layers
- 05/09/2011 12Z: GSI/GFS Q2FY2011 Minor Upgrade.
- GSI Changes
* Improved OMI QC
* Removal of redundant SBUV/2 total ozone
* Retune SBUV/2 ozone ob errors
* Relax AMSU-A Channel 5 QC
* New version of CRTM 2.0.2
* Inclusion of Field of View Size/Shape/Power for Radiative transfer
* Inclusion of GPS RO data from SAC-C, C/NOFS and TerraSAR-X satellites
* Remove down weighting of collocated radiances
* Limit moisture >= 1.e-10 in each outer iteration and at end of analysis
* Inclusion of uniform (higher resolution) thinning for satellite radiances
* Improve location of Buoys in vertical (move from 20 to 10m)
* Improved GSI code with optimization and additional options
* Recomputed background errors
* Inclusion of SBUV and MHS from NOAA-19 – removal of AMSU-A NOAA-15
* Ambiguous vector quality control for ASCAT (type 290) data
- GFS and Post Changes
* New Thermal Roughness Length -- Reduced land surface skin temperature cold bias
and low level summer warm bias over arid land areas.
* Set minimum moisture Value in Stratosphere to 1.0E-7
* Reduce background diffusion in the Stratosphere
* Correct error in the 192 hr, 12-hr precipitation bucket
* add a new membrane sea level pressure (SLP) field to eliminate smoothing of MSLP over water to prevent misalignment of pressure and wind fields.
- 05/22/2012 12Z: GSI Q3FY2012 Major Upgrade.
- A hybrid variational/ensemble assimilation system is employed. In this system,
the background error used to project the information in the observations into the analysis
is created by a combination of a static background error (as in the prior system) and a
new background error produced from a lower resolution (T254) Ensemble Kalman Filter.
see NWS TIN
and EMC Parallels .
- Other Analysis Changes:
* Use GPS RO bending angle rather than refractivity
* Include compressibility factors for atmosphere
* Retune SBUV ob errors, fix bug at top
* Update radiance usage flags
* Prepare for monitoring NPP and Metop-B satellite data
* Add NPP ATMS satellite data, GOES-13/15 radiance data, and SEVERI CSBT radiance product
* Include satellite monitoring statistics code in operations
* Add new satellite wind data and quality control
- 09/05/2012 12Z: GFS Minor Change.
- A look-up table used in the land surface scheme to control Minimum Canopy Resistance and Root Depth Number
was updated to reduce excessive evaporation. This update is aimed to mitigate GFS cold and moist biases
found in the late afternoon over the central United States when drought conditions existed in summer of 2012.
see NWS TIN
and EMC Parallels .
- 01/14/2015 12Z: T1534 Semi-Lagrangian GFS Major Upgrade
( NWS TIN )
- Model Changes
* Upgrade from current operational T574 Eulerian (~23km) to T1534 Semi-Lagrangian (~13km)
* Use high resolution daily RGT SST instead of weekly OI SST, and use daily sea ice analysis
* Extend high resolution forecast from 8 days to 10 days.
* Use McICA radiation approximation
* Reduced drag coefficient at high wind speeds
* Hybrid EDMF PBL scheme and TKE dissipative heating
* Retuned ice and water cloud conversion rates, background diffusion of momentum and heat
* Retuned orographic gravity-wave forcing and mountain block etc
* Change from Lagrangian to Hermite interpolation in the vertical to reduce stratospheric temperature cold bias
* Restructured physics and dynamics restart fields and updated sigio library
* Consistent diagnosis of snow accumulation in post and model
* Compute and output frozen precipitation fraction
* Divergence damping in the stratosphere to reduce noise
* Added a tracer fixer for maintaining global column ozone mass
* Stationary convective gravity wave drag
* New blended snow analysis to reduce reliance on AFWA snow
* Changes to treatment of lake ice to remove unfrozen lake in winter
* Modified initialization to reduce a sharp decrease in cloud water in the first model time step
* Correct a bug in the condensation calculation after the digital filter is applied
* Replace Bucket soil moisture climatology by CFS/GLDAS
* Add the vegetation dependence to the ratio of the thermal and momentum roughness
* Fixed a momentum roughness issue
* Accumulation bucket changed from 12 hour to 6 hour between day 8 and day 10
- GSI Changes
* convert GFS GSI to vertical structure
* increase horizontal resolution of ensemble from T254 to T574
* reduce number of second outer loop iterations from 150 to 100.
* changes in radiance assimilation: upgrade to CRTM v2.1.3
* move to enhanced radiance bias correction scheme
* correct bug in AMSU-A cloud liquid water bias correction term
* assimilate new radiances: F17 an F18 SSMIS, MetOp-B IASI
* turn off known bad channels: AQUA AIRS channels 321, NOAA-19 AMSUA channel 7, NOAA-19 MHS channel 3
* increase ATMS observation errors: increase channels 6 - 10 from 0.3 K to 0.4 K, increase channels 11 - 12 from 0.4 K to 0.45 K
* turn on cloud detection channels for monitored instruments: NOAA-17, -19 HIRS, GOES-13 and -14 sounders
* changes in assimilation of atmospheric motion vectors (AMV): assimilate NESDIS GOES hourly AMVs, improve AMV quality control
* improve GPS RO quality control
- 05/11/2016 12Z: Data Assimilation and Model Upgrade
( NWS TIN )
- Data Assimilation Upgrade
* Upgrade the 3D Hybrid Ensemble-Variational to 4D Hybrid Ensemble-Variational Data Assimilation System
* Multivariate Ozone update
* Assimilate all-sky (clear and cloudy) radiances
* Bias correct aircraft data
* Modify relocation and storm tracking to allow hourly tropical cyclone relocation
* other upgrades (e.g. CRTM, Data selection/thinning, AMV winds, etc.)
- Model Upgrade
* Corrections to land surface to reduce summertime warm, dry bias over Great Plains
* Hourly output fields through 120-hr forecasts
* Improved icing probability products and new icing severity product
* add five more levels from 10 hPa to 1 hPa in post-processed pgb files