EMC Synergy Meeting Highlights February 28, 2005
This meeting was led by Peter Manousos and attended by Bill Bua, Eric Rogers, John Ward, Mary Hart, Daniel Pawlak, Stephen Lord, and Joe Sienkiewicz. Additionally, Stephen Jascourt, AWC’s Steven Silberberg, and SPC’s Steve Weiss attended by teleconference.
John Ward reported the new CCS went operational January 25th 2005 with no major problems. The second member of the Climate Forecast System (CFS) went operational as did the hybrid sigma-pressure version of the GFS (currently still only run in sigma mode). The extratropical cyclone tracker is running as well. Codes for the T382 version of the GFS have been received and will be implemented in parallel the first week of March. The last “Eta” bundle parallel codes will also be running the first week of March. Notification to evaluation participants have either been sent or are forthcoming. Additionally, the AQI model domain has been increased threefold (covers the eastern domain). A Great Lakes Wave parallel will also begin running in parallel toward the end of March. Also, the 13-km RUC will soon run in parallel with evaluation expected during April.
2. Notes from EMC
a. Global Modeling Group: Stephen Lord reported the T382 parallel is catching up to real time (a bug related to quality control of data around Antarctica - not rejecting data with unrepresentative surface pressures- has been rectified). A list of changes included in the GFS parallel is found in last meeting’s notes (January 31, 2005). Mountain blocking/gravity wave drag changes still remain in the parallel. GFS parallel output will be made ready in GEMPAK format for evaluation by NCEP Service centers. The intended operational implementation date for the parallel is early to mid May. The Global GSI parallel is expected thereafter.
b. Mesoscale Modeling Group: Eric Rogers reported the NAM parallel will run in real time. The parallel web site will point to the output when it becomes available. The output will also be made ready in GEMPAK format for evaluation by NCEP Service centers. Evaluations are due in by April 15. A crisis change had been implemented to fix discontinuous snow accumulation fields. Operational implementation is intended early to mid May. Additionally, a fully cycling WRF run including the GSI analysis will be ready within a few weeks.
c. Global Ensemble Prediction System: Zoltan Toth emailed the global ensemble testing for a spring implementation is slow due to IBM issues (lack of disk space, etc). Changes considered for implementation include a 6-hr cycling (instead of current 24-hr) and has exhibited very positive results. The hurricane perturbation relocation algorithm, the T126 out to 16 days, and 20 members per cycle changes are all being worked on still with no results yetThe Ensemble Transform Kalman Filter for initial perturbations will most likely will have to wait until fall. Finally, the Winter Winter Storm Recon program is going well.
d. Short Range Ensemble Prediction System: No report
e. Marine Modeling and Analysis Branch (MMAB): No report
3. Input from NCEP Centers
a. SPC – Requested for RUC 13 output in hourly frequency and reiterated request to EMC for access to GEMPAK grids for the upcoming evaluations.
b. AWC – reiterated SPCs comments.
c. HPC – inquired on a question posed by the field regarding the order of the operational suite (can the GFS be run prior to the NAM in order to provide the NAM a set of lateral boundary conditions based on the current forecast instead of using the forecast from the previous run). Stephen Lord stated the issue of the operational suite order is not new. Cliff Mass has brought this up to EMC a number of times. EMC can indeed run the GFS prior to the meso model. However, the tradeoff is an earlier data cutoff time for initial data ingest. That means the GFS would begin running before a complete set of initial obs arrived, especially POES data, degrading the GFS run and negating some of the advantage of using a later GFS run for boundary conditions. He also said improved versions of the models won’t fit on the NCEP computer to run simultaneously, so the NAM would have to be greatly delayed or EMC would have to continue to run old versions of the models. Technology and money (NPOESS and major upgrade to NCEP computer) could solve both of these problems, but that would be on the order of a decade from now, not soon.
4. The next meeting will be held March 28th, 2005 at noon in room 209 with remote conference capability.