EMC/HPC/MPC/NCO Synergy Meeting Highlights 6/05/00
Rapid improvements and modification to operational models and the platforms the models are run upon require maximum synergy and frequent interaction between the modeling community and the operational community within the National Weather Service (NWS). Optimal model performance and accelerated model development consistent with operational needs are important outcomes of such close working relationships.
To meet these goals, monthly meetings are attended by the Environmental Modeling Center (EMC) and the co-located operational centers including the Hydrometeorological Prediction Center (HPC), the Marine Prediction Center (MPC), and NCEP Central Operations (NCO) of the World Weather Building (WWB) in Camp Springs, Maryland.
THE FOLLOWING ARE HIGHLIGHTS FROM THE MOST RECENT MEETING. INFORMATION FORWARDED IN THESE HIGHLIGHTS ARE TO BE REGARDED AS TENTATIVE AND FOR PLANNING PURPOSES ONLY. THIS INFORMATION IS SUBJECT TO CHANGE AND SHOULD NOT BE RELIED UPON OR REFERENCED AS FIRM DEADLINES OR ESTABLISHED POLICY.
This meeting was led by Peter Manousos (SOO - HPC) and followed a generic agenda denoted by the subtitles below. Attendees included Carl Staton (Director - NCO), Dave Reynolds (Forecast Operations Branch Chief - HPC), John Ward (NCO), James Hoke (Director - HPC & MPC), Stephen Jascourt (COMET Post Doc. at HPC), Hua-Lu Pan (EMC - Global Modeling Group), Geoff DiMego (EMC - Mesoscale Modeling Group), Bill Bua, Zoltan Toth, and Steve Tracton.
III. Meeting Highlights
A. IBM SP
1. Current Status: John Ward stated that the IBM SP is stable in its current configuration (logically divided into a "Production Side" and a "Development Side"). There had been some problems with the GFDL Hurricane Model (GHM) running to completion, but that has been solved. Implementation of the GDAS to initialize the GHM and running the GHM out to 126 hours will occur this month as well as the first 60 hours of the global ensembles being run at T126 resolution. This month also includes testing of both GEMPAK metafile production on the SP as well as non-monolithic GEMPAK grid files. The non-monolithic GEMPAK grid files differ from the current paradigm of GEMPAK grid file production in that the grids will be able to be accessed as they are produced off the IBM SP rather than waiting for the entire grid file to be completed in order to be viewed.
2. Planning of PHASE II IBM SP: Carl Staton stated that IBM will upgrade the IBM SP using 4 CPUs per node and doubling the current disk space. This translates to a CPU-to-Memory ratio 4 times that what was originally planned at no extra cost. This upgrade is planned to implemented by the end of September with a December operational time line. This implies that recompiling of some codes will be necessary to facilitate the upgrade. Physically swapping hardware from phase I to phase II will remove development capacity during parts of August/September. Impacts include no new model implementations will be possible from approximately mid-August to maybe November (considering 30-45 days of operational testing for new implementations after the computer configuration change).
B. Notes from EMC
1. Mesoscale Modeling Group: Geoff DiMego discussed preparations for the planned August implementation of the 22-km horizontal resolution/50-layer vertical resolution Eta model. Code to run the EDAS have been handed over to NCO to be implemented on the SPs "production side". The model will run in open slots, which may not exist if for instance the hurricane model is running. The EDAS will be required to run a minimum of one forecast a day in order to maintain assimilation cycling, and will need to continue to do this for enough successive days to provide statistics for CAFTI review. The August bundle still includes changes to handling of small-scale and cloud physics, as well as incorporation of radiances into the 3D Var. Precipitation will NOT be incorporated into the direct assimilation as was previously planned.
2. Global Modeling Group: Hua-Lu Pan reviewed the decision to pull back the previously planned implementation of the MRFX due to the problems with the convective scheme. What still remains in the MRFX are changes to the analysis scheme (which includes incorporation of radiosonde drift, GOES 10 sounder information, and changes to handling of ozone) and the relocation of initialized tropical circulations. The earliest possible implementation of the "new" MRFX would be in August; however it is hoped that if offline experiments go well with the GHM, the relocation scheme will be implemented later this month.
C. Input to EMC from Operational Centers
Peter Manousos supplied a few comments from HPC forecasters mainly centered around a perceived dry bias with convection of the Eta. Geoff DiMego explained that part of the rational for the changes intended for the August bundle of the Eta is targeted to address these very problems. Nonetheless, Geoff requested that these comments be forwarded to him for review. Another comment was also forwarded to EMC from HPC forecasters concerning the AVN/MRF placing precipitation too far north. This too was noted by the Global Modeling Group.
Zoltan Toth, Steve Tracton, Jim Hoke, and Dave Reynolds discussed the utilization of ensembles in HPC operations. Specifically, how to at a minimum return to normal production of ensemble products lost due to the CRAY fire in fall of 1999; and how to then optimize the use of ensemble products in HPC operations. Finally, it was also decided to add CPC to the mailing list for both meeting attendance and reception of highlights.
IV. Next Meeting Proposed Monday July 24, 2000 at noon in room 209.