NCEP Synergy Meeting Highlights: August 19, 2013
This meeting was led by Wallace Hogsett (WPC) and attended by Keith Brill and David Novak (WPC); Chris Caruso Magee (NCO); John Derber, Eric Rogers, Geoff DiMego, Hendrik Tolman, Stephen Barry, David Dewitt, and Mary Hart (EMC); Joe Sienkiewicz (OPC); Kathy Gilbert (MDL); Steve Weiss, Andy Dean, and Israel Jirak (SPC); Richard Pasch (NHC); David Bright (AWC); Brian Miretzky (ER); Jeff Craven (CR); John Eise (CR SSD); Andy Edman (WR); Aijun Zhang (NOS)
1. NOTES FROM NCO (Chris Caruso Magee)
Things are going smoothly on WCOSS, and NCO is shifting the focus to new implementations. The first changes will be an update to the MAG page, upgrades to bufr and grib2 libraries, andn grib2 output from global RTOFS. Air quality for OCONUS was implemented on Aug. 20.
NCO is working to move NHC wind speed probability scripts into production, and the next implementations will be RTMA, Great Lakes wave, GFS MOS cool season wind, MOS extension into the Pacific and Mexico, San Francisco Bay domain for NOS, and the RAP. All of these are scheduled for Q1. (The RAP will likely be delayed into Q2 based on discussions at the RAP kick-off meeting held on 8/27).
As of August 20, GSI now includes more data from MetOp-B, CrIS, AMSU-A, AMSU-B, and SEVERI from Meteosat-10. This data will provide redundancy and resiliency to future satellite failures, and negligible forecast impact is expected. NCO is working with NESDIS to go from 3-hourly to 1-hourly satellite winds, and the hope is for an accelerated RFC within 3 weeks. NCO is meeting with centers who run in production to get things moved officially into the production job suite.
To streamline work and minimize mistakes, NCO is requiring subversion to be used in handing off codes to be implemented on the supercomputer. EMC (Tolman) noted that so far it’s very efficient and that using the vertical SVN structure makes working with NCO easier. For major implementations, the SPAs can check the newest version, set it in a production-like environment, get timing/resources set and any bugs worked out, then move the code into an official 30-day parallel. It may be slow at first with some overhead, but in the long run it will help the implementation process. So far, half of the production suite is versioned, but it doesn’t include some of the big systems yet, including the GFS, NAM, and GSI.
2. NOTES FROM EMC
2a. Global Climate and Weather Modeling Branch (GCWMB) (John Derber)
In addition to the GDAS data additions, WAFS flight-level 410 is next, and that will be combined with a fix to the maximum wind speed to ensure that we don’t have higher maxima at some levels than the max. Most of the work now is focused on next spring’s GFS/GEFS upgrade. It will be a huge resolution increase, and there are a lot of issues, such as reduced memory on WCOSS. We expect to have a prototype version finalized within a few months.
Question from NESDIS: Are you looking to divest any NESDIS datasets? Nothing has gone bad, so we won’t remove it unless it’s bad. We will move SMIS into operations next implementation.
Question from NHC: What impact do you expect from the new satellite data in the GFS? Minimal. The new system will be more robust and resilient to future satellite failures (some are beyond lifetime now).
2b. Mesoscale Modeling Branch (MMB) (Geoff DiMego)
We are working on the RTMA implementation first, then the RAPv2, which must go in before the HRRR. These will occur Q1/Q2. Next implementations will be the high-res window and the NAM. The hope was to go to 3km for both systems, but that has been pushed back to 2015, later in phase one of WCOSS. More time is needed to test, and more experience on WCOSS is necessary. The NAM includes significant upgrades, and the high-res window will be full CONUS twice daily, Alaska twice daily, codes will be upgraded, and the NMM will be replaced by NMM-B. Because of the decision not to go to 3-km resolution, the high-res bundle will not be as large as initially expected. The hope is that the high-res window would be preempted with four storms, not one or two. This is a benefit.
Question from SPC: Will the high-res ARW be improved? Yes, but not all the way down to 3-km resolution.
For the SREF, the “bug fix” is scheduled as the #6 priority for Q1; the full SREF upgrade is scheduled for Q4.
2c. Marine Modeling and Analysis Branch (MMAB) (Hendrik Tolman)
The next upgrades will be to Great Lakes wave model (October), then the wave ensemble, which will go down to a half degree for all 20 members and may include newer physics. This has been coordinated with Fleet Numerical and Environment Canada. In mid-2014 all systems will be updated. We are updating the hurricane wave model with newer bathymetry and updated code. Early 2015 we plan to upgrade Great Lakes wave with higher resolution for NDFD. For the global wave, we plan to implement curvilinear grids throughout the Arctic. The Atlantic model will be upgraded with the same grid, but more integrated with global RTOFS with dynamic boundary conditions. Early in Q2, we hoped to end the Fukishima tracer model, but with linger leakage problems, we may continue. There is uncertainty here, as we may kill it, or it may be a major implementation with a larger grid.
We also plan to upgrade SSTs if MetOp data is available, but no huge impact is expected. We will also update the sea ice drift model to use the ensemble mean, and we’ll upgrade ice concentration.
3. NATIONAL OCEAN SERVICE (NOS, Aijun Zhang):
4. FEEDBACK FROM MDL/OPERATIONAL CENTERS/REGIONS
4a. MDL (Kathy Gilbert):
We made a fix to GFS winds associated with a surface roughness length change. For Pacific MOS, we’re adding sky and ceiling guidance for fifteen stations in the Pacific islands. We issued a TIN, and we’ll amend as needed. The 00Z and 12Z messages will be identical.
All implementations will be vertically structured, and we’re trying to use the Virtual Lab in Subversion for the MOS code. In Q2, we will add probabilistic grids over CONUS, including ptype. Also, a 20-km LAMP probabilistic convection product for lightning and convection will be included, as well as probabilistic storm surge.
Question: When will the field see these new things in AWIPS? We need to map implementations to SREC priorities. Bob Rood is a contact.
4b. NCEP Centers and NWS Regions
Weather Prediction Center (WPC, Wallace Hogsett): WPC just completed the Flash Flood and Intense Rainfall (FFaIR) Experiment, which aimed to improve our ability to forecast flash floods in the 0 – 12 h timeframe. We are in the early planning stages for the annual Winter Weather Experiment, which we hope to build on the success of prior experiments.
Storm Prediction Center (SPC, Israel Jirak): No report.
National Hurricane Center (NHC, Richard Pasch): No to report.
Space Weather Prediction Center (SWPC, no representative):
Ocean Prediction Center (OPC, Joe Sienkiewicz): Nothing to report
Aviation Weather Center (AWC, David Bright): Nothing to report.
Pacific Region (PR, no representative):
Alaska Region (AR, no representative):
Eastern Region (ER, Brian Miretzky):
The group discussed the Virtual Lab, which is open to anyone in NOAA.
Western Region (WR, Andy Edman): Nothing to report.
Southern Region (SR, no representative):
Central Region (CR, Jeff Craven):
FYI, we have noticed a 10-20+ minute improvement in the arrival of NAM and other data since the WCOSS transition. Some in CR want to change the way we process models, but can we count on the early arrival?
EMC: we tried to slow it down. Take advantage of it, but don’t expect the data to continue arriving early. Leave your processes where you have them, because we’ll use that window for other parts of the production suite.
NCO: The goal was to not change delivery times. With the NAM, we couldn’t slow it down. There is a difference in nodes (110 GB memory on CCS vs. 25-30 GB memory on WCOSS), even though there are more nodes on WCOSS. Don’t assume early delivery will continue. Lots of upgrades are coming.
John Eise, SSD CR: Does MDL provide QPF over CONUS? Yes, High-Res MOS QPF. http://www.nws.noaa.gov/mdl/hrqpf/
5. NESDIS (no representative):
The next Synergy Meeting will be held at 2:30 pm EDT on Monday, 30 September 2013 in NCWCP conference room 2890, with remote teleconferencing capability.