NCEP Synergy Meeting Highlights: October 31, 2011
This meeting was led by Bill Bua (UCAR-COMET), and attended by Rebecca Cosgrove (NCO), Jordan Alpert (EMC), Yuejian Zhu (EMC), Sura Saha (EMC), Geoff DiMego (EMC), Vera Gerald (EMC), Mary Hart (EMC), Steve Silberberg (AWC), Michael Brennan (NHC), Steve Weiss (SPC), Keith Brill (HPC), Carven Scott (AR), Andy Edman (WR), Chris Smallcomb (WR), Brian Miretzky (ER), Pete Browning (CR), Bernard Meisner (SR), Melissa Kreller (SR), Kathy Gilbert (MDL) and Daryl Kleist (EMC)
1. NCO (Rebecca Cosgrove)
Implementations in October 2011
The NAM-B was implemented at 12 UTC 18 October 2011, as was the HYCOM global RTOFS. NAM-B high resolution nests have replaced the NAM downscaled grids for the 60-hour duration of these forecasts.
Next up in the implementation queue:
· 1 November 2011: NCEP/Fleet Numerical Meteorological and Oceanographical Center (FNMOC) combined wave ensemble, using the NCEP and FNMOC WaveWatch3 models and perturbed wind forcing.
· 8 November 2011: Hi-Res Window output will be made available on the Satellite Broadcast Network (SBN). All grids should work except for Guam, and the Puerto Rico grid will need a software change in AWIPS before that data can be properly displayed.
· Global Ensemble Forecast System: Parallel should be ready to run by mid- to late-November for a planned mid- to late-February implementation
· Rapid Refresh (RR, Rapid Update Cycle [RUC] replacement): The Technical Information Notice (TIN) and implementation schedule is currently being prepared
For the latest schedule updates see: http://www.nco.ncep.noaa.gov/pmb/changes/
2. NOTES FROM EMC
2a. Global Climate and Weather Modeling Branch (GCWMB) (Jordan Alpert, Yuejian Zhu)
Data Assimilation/Hybrid 3-D VAR Ensemble Kalman Filter (3D-VAR/EnKF)
A parallel forecast has begun which was started at 1 June 2011 and has been run through late July 2011. It will be caught up to real time, the code delivered to NCO, and then be run as a parallel for evaluation and implementation in Spring 2012. Daryl Kleist’s presentation at the end of the meeting will cover the 3D_VAR/EnKF; discussion of the presentation is found at the end of the notes.
2b. Global Ensemble Prediction System (GEFS)
See notes under NCO for implementation schedule, and previous notes for planned configuration.
2c. Mesoscale Modeling Branch (MMB) (Geoff DiMego)
The NMM-B was implemented as noted by NCO.
Rapid Refresh (RR) Replacement of RUC
Rapid Refresh (RR): If all were to go perfectly, implementation of the new RR would take place 12 January 2012. Additionally, the first North American Rapid Refresh Ensemble (NARRE) based on time-lagged forecasts of the RR will be provided. While it will not be available on AWIPS, there will be ftp access to the data.
More information and results can be found here (http://www2.emc.ncep.noaa.gov/impDoc.html) under "Implementation Briefing" in last column of the NAM 3.0 row; and under the EMC home page (http://www.emc.ncep.noaa.gov/) under "NAM Upgrade" on the right, click "Description of Changes".
Both the NAM NMM-B and RR will have an additional upgrade late in the year, but no specific changes have been identified; stay tuned.
Finally, Geoff DiMego will send out an e-mail about where to obtain NMM-B Fire Weather inner nest graphics (and to find out what domain was run in the nest) to the field.
A upgrade hopefully will be able to be implemented in Spring 2012. Below repeats the details in previous synergy meeting notes:
• Expanded domain to cover the Canadian portion of the NWRFC area
• ~3 km resolution for mainland Alaska and 1.5 km for Juneau area
• GSI/2D Var improvements for CONUS domain (to update to the same method as for Guam & Hawaii)
• Improved winds over lakes and oceans.
The real-time parallel for the next implementation of the SREF is nearing completion. It was thought that because of computer resource limitations, a choice has had to be made to balance higher resolution with physics diversity. With the help of IBM staff, however, it appears that it will be possible to implement a SREF with horizontal grid spacing of 16-17 km, essentially half the grid spacing used now. Other changes will include
• Elimination of Eta and RSM
• Addition of the NMMB
• More diverse physical parameterizations
• An initial condition perturbation strategy using a hybrid Breeding and Ensemble Transform method, and analyses from GDAS, NDAS, and Rapid Refresh
• Post-processing enhancements
2d. Marine Modeling and Analysis Branch (MMAB) (Vera Gerald)
Global Hybrid Coordinate Ocean Model (HYCOM/RTOFS)
The Global RTOFS was implemented into operation on October 25, 2011.
RTGSST high resolution
Mention was made of the failure of the AMSR-E; this failure will affect the 0.25°x0.25° Optimal Interpolation (OI) SST analysis, so that only visible and IR data are available. In areas where cloudiness is an issue, like over the Gulf Stream and Great Lakes during the cold season, there will be spells without observations. This will affect the quality of the OI SST analysis. Since AMSR-E is likely “deceased” there will be no such data for the foreseeable future.
The global NCEP/FNMOC wave ensemble model is scheduled for implementation on
November 1, 2011. This combines the Navy and NCEP Wave Watch 3 ensemble forecast systems.
Spectral output resolution increase for the multi-scale wave model is still on schedule for FY12Q1. A parallel is expected to begin on 22 November 2011. RFC’s have been submitted
To set up “dev” parallel runs; making it easier for operational implementation.
In Q2FY12, implementations will be performed that will make the wave model physics among all versions run at NCEP consistent with each other.
3. NATIONAL OCEAN SERVICE
4. FEEDBACK FROM MDL/OPERATIONAL CENTERS/REGIONS
4b. NCEP Centers and NWS Regions
HPC, NHC, SPC, AWC – Present, nothing to report.
OPC – No report.
Central Region – Has a cool moist near-surface GFS problem that they have report to EMC pending.
Eastern Region, Alaska Region, Southern Region, Western Region – Present, nothing to report.
Pacific Region – No report.
5. NOTE: The annual NCEP model review dates have been set for 6-8 December 2011. Formal review will be on 6-7 December, with the morning of 8 December set aside for more informal discussions.
6. Daryl Kleist, EMC/Data Assimilation Team presented on the current configuration and testing of the Hybrid 3DVAR-Ensemble Kalman Filter data assimilation system (EnKF). EnKF differs from standard 3DVAR in that an ensemble of analyses and forecasts are used to create flow dependent first guess error covariance (i.e, shape and size of error “pattern”). This is rather than the standard fixed error covariance determined from all previous first guess errors, regardless of flow regime. What makes the data assimilation “hybrid” is the use of relative weighting to linearly combine the EnKF and 3DVAR covariances. This is necessary to reduce the sampling error from the EnKF (80 members out of a huge number of degrees of freedom!) and to increase the effective resolution of the error structure (EnKF is at coarse resolution, while 3DVAR is at the resolution of the operational GFS).
Daryl then showed examples indicating improvement in the GFS at all lead times to 7 days and gave information on how to follow progress with development of the real-time parallel. This included a verification URL. Forecast graphics will be made available when the real-time parallel is running if not before.
Daryl’s presentation will be made available in a separate e-mail to the Synergy group.
7. Given the upcoming NCEP Model Suite Review December 6-8, the next scheduled Synergy Meeting will be deferred until Monday, January 30, 2012.