EMC/HPC/MPC/NCO/CPC synergy meeting highlights 09/30/02
This meeting was led by Bill Bua and followed a generic agenda denoted by the subtitles below. Attendees included John Ward, Jim Hoke, Larry Burroughs, Stephen Jascourt, Hua-Lu Pan, Zoltan Toth, Ed Danaher, new MPC SOO Joe Sienkiewicz, and Dave Michaud. Julia Zhao, a new NCO employee working on the climate model implementation, was an invited guest.
1. IBM SP
John Ward reported that the new computer (now called the R6000, since it uses new “clustering technology”) has been installed in an IBM facility in Gathersburg, MD (both development and operational systems). The moratorium on model codes starts Tuesday, 8 October, other than the GFS T254L64. The network for communications with the new system was to be completed by Verizon today (9/30). Access to key people will begin during the week of October 7, 2002. Acceptance testing of the new system will begin November 1, 2002 (including tests of GEMPAK and HPC/MPC scripts) and is anticipated to last 30 days, after which the system would be generally available for all users. Names of the two systems have been temporarily set to “frost” and “snow”; permanent names for the operational and development systems will be determined later.
Code conversion to the new machine has had a few glitches, but nothing insurmountable. Best estimates for when it will become operational are in the late January to early February time frame, the same as reported at the August synergy meeting.
The old operational system will remain available until we begin to vacate the Bowie facility at the end of August 2003, and will be devoted to the regional reanalysis project. The old development system (bsp), however, will be removed in February 2003.
2. Notes from EMC
a. Global Modeling Group: Hua-Lu Pan reported that the T254L64 version of the Global Forecast System model is still expected to become operational on October 29. The current parallel with analysis changes only was successfully run; it will be run starting this week at the T254L64 resolution until implementation.
b. Mesoscale Modeling Group: No report.
c. Global Ensembles: The increase to 4 times per day ensembles is likely to take place in February 2003 when the new supercomputer is declared operational. It will also include higher resolution T126L28 out to 7.5 days (rather than the current 3.5 days), 6-hour breeding of perturbations (rather than 24 hours), and the same relocation of tropical vortices as found in the operational MRF from the GFS.
d. Short Range Ensembles: Nothing new to report.
e. Marine Modeling and Analysis Branch (MMAB) - Larry Burroughs reported that the wave model (WWIII) was now running four times per day for the Marine Prediction Center (MPC), and that the new MPC SOO was developing GEMPAK scripts to graphically display the data in NMAP. Vessel icing model is now running four times per day year-round, rather than just seasonally. After the moratorium, the MMAB will be testing an open ocean fog/visibility prediction system, which will require Eta-look-alike inputs from the GFS, which the MMAB hopes to have available by late winter to spring 2003. The current coastal fog forecast will be tested to 84-hours, with forecasts at 3-hour intervals, and run at four cycles per day.
3. Input to EMC and other information from Operational Centers
a. Hydrological Prediction Center (HPC) reported that the preparations for the 2nd Winter Weather Experiment (WWE) begin October 1, 2002. As was true last year, the purpose of the WWE is to explore the effectiveness of national/field office collaboration, and do a Short Range Ensemble Forecast (SREF) ‘shakedown’. What is different this year is that many more offices will be participating, including all of Eastern Region and eight offices in the Eastern Time Zone in Central Region (WFOs in Marquette, Gaylord, Detroit, and Grand Rapids MI; Jackson and Louisville KY; and Indianapolis and Northern IN). The effect of the winter weather coordination effort on field and HPC workload will also be examined. What hopefully will also be different is that there will be adequate significant winter weather to conduct a good experiment!
The HPC winter weather graphical products provided for the experiment will be changed a bit based on last year’s experiment. For example, they will graphically outline those counties expected to reach winter weather warning criteria. SREF-provided storm tracks will also be made available. Other “tweaks” will be made as warranted during the WWE. The two-times-per-day winter weather ‘chat room’, which seemed cumbersome to many of the users, will be changed to a 15-minute conference call lead by HPC on the SREF product indications for a specific winter weather event.
The formal experiment will be conducted from November 1, 2002 to April 1, 2003. HPC SOO Pete Manousos will be the point of contact.
b. Marine Prediction Center (MPC) reported that they have been very pleased with the performance of the GFS forecasts of tropical cyclone development, track, and strength, and even with the difficult model
problem of forecasting tropical-to-extratropical cyclone conversion.
4. Next Meeting Proposed Monday, October 28th, 2002 at noon in room 209.