Name:   NCEP Ensemble

The NCEP real-time forecasts consist of a 16-member ensemble run every day. The S2S archive contains all the NCEP real-time forecasts since 1st January 2015, and the associated re-forecasts.

1. ENSEMBLE VERSION

Ensemble identifier code:    CFSv2
Short Description:    Global ensemble forecast system for monthly and seasonal predictions
Research or operational: Operational
Data time of first forecast run:   15/03/2011. CFSv2 was operational at the end of March 2011. No forecasts before April 1, 2011 were given to the public.

 2. Configuration of the EPS

Is the model coupled to an ocean model?   Yes, from day 0
If yes, please describe ocean model briefly including frequency of coupling and any ensemble perturbation applied: Ocean model is GFDL MOM4 that has a spatial resolution in the zonal direction of 0.5° and in the meridional direction, 0.25° from 10°S to 10°N, progressively decreasing to 0.5° from 10° to 30°, and is fixed at 0.5° beyond 30° in both hemispheres. There are 40 levels in vertical.
Is the model coupled to a sea Ice model? Yes - Sea ice model is part of MOM4p0
If yes, please describe sea-ice model briefly including any ensemble perturbation applied: thermodynamic and dynamic sea ice model from the GFDL Sea Ice Simulator (Griffies, S. M.,M. J.Harrison, R. C. Pacanowski, and A. Rosati, 2004: Technical guide to MOM4.GFDLOcean Group Technical Rep. 5, 337 pp. [Available online at http://www.gfdl.noaa.gov/;fms.])
Is the model coupled to a wave model? No
If yes, please describe wave model briefly including any ensemble perturbation applied: ECMWF wave model.  N/A
Ocean model: MOM4p0
Horizontal resolution of the atmospheric model: T126 (about 100 km) Number of model levels: 64
Top of model: 0.02 hPa
Type of model levels: sigma-pressure hybrid coordinates
Forecast length: 45 days (1080 hours)
Run Frequency: 4 cycles/day
Is there an unperturbed control forecast included?: Yes
Number of perturbed ensemble members: Three perturbed members each 6-hour cycle
Integration time step: 20 minutes

 3. Initial conditions and perturbations

Data assimilation method for control analysis: Climate Forecast System Reanalysis (CFSR)
Resolution of model used to generate Control Analysis: T384/L64 for hindcast and T574/L64 for real-time forecasts after 2011
Ensemble initial perturbation strategy: Add a small perturbation into atmospheric, oceanic and land analysis at each cycle
Horizontal and vertical resolution of perturbations:  3D, all levels and variables
Perturbations in +/- pairs: Yes, there is one +/- pair

4. Model Uncertainties perturbations:

Is model physics perturbed? No
Do all ensemble members use exactly the same model version? Yes
Is model dynamics perturbed? No
Are the above model perturbations applied to the control forecast? N/A

 5. Surface Boundary perturbations

Perturbations to sea surface temperature? Yes
Perturbation to soil moisture? Yes
Perturbation to surface stress or roughness? Yes
Any other surface perturbation?  Everything is changed through Initial condition perturbations, not physics perturbations
Are the above surface perturbations applied to the Control forecast?  No
Additional comments

6. Other details of the models:

Description of model grid: Gaussian grid
List of model levels in appropriate coordinates: http://journals.ametsoc.org/doi/suppl/10.1175/2010BAMS3001.1/suppl_file/10.1175_2010bams3001.2.s1.pdf
What kind of large scale dynamics is used?  Spectral
What kind of boundary layer parameterization is used? http://journals.ametsoc.org/doi/pdf/10.1175/2010BAMS3001.1
What kind of convective parameterization is used? Simplified Arakawa–Schubert convection with momentum mixing
What kind of large-scale precipitation scheme is used? RH Criteria
What cloud scheme is used? Prognostic cloud condensate from which cloud  cover is diagnosed
What kind of land-surface scheme is used? NOAH Land model
How is radiation parametrized? See references below
Other relevant details? For further details on model configuration and physics, see citations below

 7. Re-forecast Configuration

Number of years covered: 1999-2010
Produced on the fly or fix re-forecasts? Fix
Frequency:   Everyday; 4 runs/day
Ensemble size: 1 member
Initial conditions: CFSR
Is the model physics and resolution  the same as for the real-time forecasts: Yes
If not, what are the differences: NA
Is the ensemble generation the same as for real-time forecasts? CFSR analysis for each cycle
If not, what are the differences: NA

 8. References:

Comprehensive description of the model physics:

Saha, S., and Coauthors, 2010: The NCEP Climate Forecast System Reanalysis. Bull. Amer. Meteor. Soc., 91, 1015.1057. doi: 10.1175/2010BAMS3001.1

Saha, S. and Coauthors, 2014: The NCEP Climate Forecast System Version 2 J. Climate, 27, 2185–2208. doi: http://dx.doi.org/10.1175/JCLI-D-12-00823.1

9. Configuration  in the S2S archiving

The NCEP re-forecasts dataset is a "fixed" dataset which means that the re-forecasts are produced once from a "frozen" version of the model and are used for a number of years to calibrate real-time forecast. The NCEP re-forecasts consist of a 4-member ensemble run every day from 1st January 1999 to 31 December 2010. Because of the large volume of this re-forecast dataset, the NCEP re-forecasts are archived on a daily basis in the S2S database. Each day,  the NCEP re-forecasts corresponding to this day from 1999 to 2010 are archived in  the S2S database. This means that the complete NCEP re-forecast database will only be  available at the end of 2015.

As for the other models, NCEP re-forecasts are archived in the S2S database with 2 date attributes:

• hdate which corresponds to the actual starting date of the re-forecast
• date which correspond tot he ModelVersionDate.Since the NCEP re-forecasts are "fixed" re-forecasts this ModelVersiondate is the same for all the re-forecasts and equal to 20110301. This variable will change when a new version of CFS will be implemented.

10  Issues in the S2S database

• Step 0 is not available for all surface  instantaneous fields, except for mean sea level pressure (msl)
• Step 0-24 (day 1) not available for all daily-mean fields except 2-metre dewpoint temperature, total cloud cover and Convective available Potential Energy (CAPE)
• For some re-forecast dates, only the 00Z runs could be recovered, which means that the control, ensemble members 1, 2 and 3 are identical and correspond tot he 00Z run. 
  The list of affected dates is:
  

  19990101,

  19991110 -- 19991012,

  20001110 -- 20001119,

  20011110 -- 20011119,

  19990210 -- 19990221,

  20000210 -- 20000221,

  20010210 -- 20010221.