NCEP Synergy Meeting Highlights: January 28, 2013

This meeting was led by Wallace Hogsett (HPC) and attended by Dave Novak and Keith Brill (HPC); Bill Bua (UCAR/COMET); Chris Caruso-Magee (NCO); John Derber, Eric Rogers, Stephen Barry, Hendrik Tolman, and Mary Hart (EMC); Scott Scallion (MDL); Richard Pasch (NHC); Israel Jirak, (SPC); Bill Ward (PR); Carven Scott (AR); Andy Edman (WR); Dave Radell and Brian Miretzky (ER); Aiyun Zang (NOS); Roberto Garcia (Puerto Rico WFO).  


1. NOTES FROM NCO (Chris Caruso-Magee)

NCO is working with developers on the WCOSS (new supercomputer) transition, which is about 40% done. The GFS and NAM are going well. The global wave model work is ongoing, hurricane wave model work is beginning, the combined FNMOC wave ensemble work is ongoing, and the Great Lakes wave model is close to being done. The SREF and GEFS transitions are scheduled to begin as early as early February. Three weeks per major model transition is allotted from the Senior Production Analyst (SPA) perspective. Then, one week is allotted for the developer to validate the SPA's first test using canned data, another week for the SPA to set up the jobs using live data, and then another week for the developer to validate the SPA's test using live data. So, it could be up to three weeks after the initial SPA test until final developer validation.

For outside folks, we are not sure how they will access the data. We will use the MAG website to help with evaluation, and some data will be pushed to the NCEP FTP server. Regions and WFOs should stay tuned; customer validation is the last step. Also, MOS is completely in on WCOSS, and the last piece came in from NOS on Friday. Both are in great shape.

The HWRF system, not including upgrades, has been fully tested on Tide. The GFDL is not ready because their developers did not have access to Tide until very recently. Parallels may be made available to NHC. The full transition from CCS to WCOSS will occur in August, but that doesn’t mean that the work is done at that time. There is an issue with the load balancer, which distributes jobs on the various nodes. The load balancer issue will not be resolved until April, and then the SPAs will have to work to get everything on the final version of WCOSS.

Discussion about coarse GFS grids and MRF-lookalike grids: WR and AR need the 381 km data, but not necessarily the 190 km. Puerto Rico uses and needs the 00Z and 12Z runs at 190 km at least out to Day 7. NCO wants to get rid of MRF lookalikes (low resolution) before the transition to WCOSS. Most don’t care about those, and that’s a number one goal. See this link for more information: http://www.nws.noaa.gov/os/notification/pns13gfslegacy.htm

Why do you need the 190 km grids? Northern Hemisphere 381 km data are used to see Rossby waves moving in. Are fields produced at 12z and 00z? Puerto Rico WFO looks at 4 cycles.

2. NOTES FROM EMC

2a. Global Climate and Weather Modeling Branch (GCWMB) (John Derber and Yuejian Zhu)

The WCOSS transition and preparations for the next implementation after the transition are the main activities. The next GFS will be T1148 semi-Lagrangian, which will be an improvement in the forecast resolution. AWIPS should not change; the GRIB files will be the same resolution. The new GFS (~17km grid spacing) parallels will begin in October and will go operational in early 2014.  

The group discussed the possibility of higher resolution GFS data for users. Half-degree fields are produced now; some groups want even higher resolution. It is possible to produce higher resolution data, but getting the data to the field via AWIPS is the issue. It’s not useful to produce the high resolution data if it cannot be disseminated. There is a need for support to get global half-degree GFS data out to the field.

There is some push from above for a GDAS implementation before the WCOSS transition to include MetOp-B and NPP data. Nothing has been decided, and there are some problems getting data. This could occur before the transition in the May timeframe. It will only be done with approval as we don’t want to slow the transition.

Non-NCEP update: The Canadian Global Ensemble will be upgraded in the coming weeks to a higher-resolution (horizontal and vertical) and unified physics version. The initial results show a skill increase for days 1 – 7.

2b. Mesoscale Modeling Branch (MMB) (Eric Rogers)

MMB is working on the WCOSS transition and running tests for the NAM. Changes include RRTM radiation and convection changes, which are showing good impact on synoptic scales and some good impact on convective systems. Implementation of the new NAM will not occur for a while after the WCOSS transition.

A bug was discovered in the SREF BUFR sounding output for NMM-B members. Precipitation is supposed to be hourly, but for the NMM-B it’s hourly except for every three hours, when it’s three hourly. Jun Du has fix, which could be a one-liner [UPDATE: SREF BUFR fix implemented February 4]. The RTMA implementation didn’t make it onto the last CCS change. The next version of the RAP will be implemented late next year.

2c. Marine Modeling and Analysis Branch (MMAB) (Hendrik Tolman)

The ocean models are making good progress on WCOSS. There were issues with IBM in the past, but they’re doing well now. Tracer is almost done, and the other two are very close to finished. We have found a workaround, although not a solution, for the MPI libraries.

The Great lakes model is going in either by the old fashioned way or with a physics upgrade on the new machines. The physics upgrade is uncertain, as it could cause work for NCO. MMAB is ready to go, but we are not driving the boat. The global wave model will cover the Arctic in year or so, at least up to 88N. RTOFS model is waiting for global DA so that the hurricane implementations are linked in terms of manpower. Ice concentration and models were discussed a few months ago at the production review meeting.

3. NATIONAL OCEAN SERVICE (NOS, Aiyun Zang)

We sent the Great Lakes package to NCO, and they are testing it. So far, the Tide output looks reasonable.

4. FEEDBACK FROM MDL/OPERATIONAL CENTERS/REGIONS

4a. MDL (Scott Scallion)

Working on WCOSS transition. A major announcement is that the gridded LAMP data is now being disseminated across the SBN.

4b. NCEP Centers and NWS Regions

Hydrological Prediction Center (HPC, Dave Novak): No report.

Storm Prediction Center (SPC, Israel Jirak): No report.

National Hurricane Center (NHC, Richard Pasch):

What role will NHC play in the WCOSS transition? NCO: The new output will not be bit-by-bit reproducible, but no egregious errors are expected. We can’t disseminate the data via AWIPS, so we’ll rely on the MAG site and FTP to get the data to users.

Ocean Prediction Center (OPC): No representative.

Aviation Weather Center (AWC): No representative.

Pacific Region (PR, Bill Ward):

The group continued to discuss the field’s need for global GFS data. Pacific Region not only monitors the northern hemisphere, but also the southern hemisphere to track storms there and around the dateline. Western Region also tracks storms a long way around the globe, and farther out than most people understand. AWIPS work is the challenge; we would love half-degree global data. We need to get the data to AWIPS, as well as in AWIPS. It would be great to get some NWSHQ communications folks on these calls. This is an issue that needs to go to a higher level.

It may be possible to use NOMADS to slice and dice the data, or to regrid the data to fit it through the pipe.

Alaska Region (AR, Carven Scott): No report.

Eastern Region (ER, Dave Radell and Brian Miretzky): No report.

Western Region (WR, Andy Edman): No report.

Southern Region (SR, no representative): No report.

Central Region (CR, no representative): No report.


5. NESDIS (no representative): No report.

6. The next Synergy Meeting will be held at 2:30 pm EST on Monday, 4 Mar 2013 in NCWCP conference room 2155, with remote teleconferencing capability.