IFPS Science Steering Team Forum

1. IFPS Science Steering Team Forum November 8th, 2005

2. ISST Membership • Greg Mann – WFO Detroit (CR) – Lead • Jim Nelson – WFO Anchorage (AR) – Backup Lead • Brad Colman – WFO Seattle (WR) • Tom Salem – WFO Glasgow (WR) • Bill Ward – Pacific Region Headquarters • Steve Keighton – WFO Blacksburg (ER) • Dan St. Jean – WFO Grey (ER) • Jeffrey Medlin – WFO Mobile (SR) • Ken Falk – WFO Shreveport (SR) • Lee Anderson – NWSHQ / OS&T – Facilitator

3. Overview • ISST Roadmap • RTMA / AoR • Gridded MOS • Day 4-7 assessment • Digital Forecast Process • Open Discussion

4. ISST Roadmap

5. Roadmap • In an effort to maintain continuity and productivity, the team has devised a roadmap of team activities that have been categorized in the following manner: • Immediate • Influencing • Monitoring • Visionary • Supporting • Ongoing

6. Immediate • Days 4-7 assessment • Given the wide variety of sources available for use in the generation of extended forecast grids, a need for a comprehensive assessment of guidance performance is required to support decision makers. The ISST will encourage and support a formal WFO/HPC/EMC Days 4-7 grid assessment through the development of a suite of objective metrics for the various sources of guidance/grid initializations • RTMA field assessment • There are several overarching objectives to this assessment: To expose forecasters to real-time objective NDFD-matching gridded analyses. To facilitate early development of WFO tools and approaches that will utilize the mature RTMA/AOR in the preparation and critique of forecasts. Provide subjective field input to EMC to identify weaknesses and biases that are undoubtedly present in the early baseline versions of the RTMA. • Digital Forecast Process • Provide both broad ideas and specific input on issues related to improving the scientific validity of methods related to the entire digital forecast process. This item is a larger umbrella under which many other ISST roadmap items fall, and is at the heart of our charter. The more immediate goal of this item is to incorporate field input and ISST opinion on key questions aimed at better defining the digital forecast process the NWS should be aiming for, and to produce a white paper summarizing these ideas and offering suggestions on how to get there. The longer term goal would be to push for and support specific activities that would help the NWS achieve the desired DFP state, and periodically re-visit what this should look like as resources and needs change.

7. Immediate • Gridded MOS field assessment • To assess whether gridded MOS fields are beneficial to forecast operations. The assessment would also look at how the forecasters used the gridded MOS fields. • GFE system enhancements (real-time feedback, integrity, QC) • Continue to support ESRL/GSD (FSL) efforts to develop the GFE infrastructure to create enhancements that ultimately ease the forecaster workload and create better scientifically sound products. • Centralized bias correction and smart initialization of grids • Influence incorporation of an effective Bias Correction scheme and to enact smart initialization at a central location. This will reduce the overhead related to ifpInit in the field offices and ensure a consistent application of smart initialization across the entire domain.

8. Influencing • OCONUS AoR • Monitor CONUS AoR actions and develop an OCONUS requirements White Paper to implement an AoR in the OCONUS. • Short Term Forecast Methods/Tools • Develop a loose framework of methods and tools resulting in more efficient and accurate short term gridded forecasts

9. Monitoring • NWS Concept of Operations changes (impacts on DFP) • Facilitate the discussion of evolving the gridded forecast paradigm in the proposed restructured NWS Concept of Operations

10. Visionary • Probabilistic/uncertainty (including tropical fields) • To transition the NDFD and digital services from a highly deterministic set of forecasts to one that is has a better balance of probabilistic and uncertainty information with the highly detailed geoclimatic information. • Future elements and required support guidance • To recommend and encourage the creation of new grids (either forecast or guidance) that will make the forecast process more efficient, accurate, and scientifically correct. Also, to increase the amount of information produced in the gridded database to be useful for all partners and the public. • LDFD / RDFD / NDFD • Explore the advantages of a multi-tiered digital forecast database within the frame work of a scientifically sound digital forecast process, and also considering what makes most sense for both the users, from the national to the local level. Explore answers to the following kinds of questions: 1) What would be the differences between a local, regional, and national database in terms of resolution, elements, forecast projections, update frequency, products generated from them, etc? 2) What entities within the NWS would be responsible for contributing to each of them? 3) How would they be interconnected? When answering some of these questions, we have to consider how these multiple tiers can most efficiently be produced within the current structure of the NWS, as well as potential future structure (therefore, this is tied in heavily with “CONOPS Changes”, “DFP”, and several other ISST roadmap issues). • Role of local models • Identify and explore the utility of local models in the augmentation and enhancement of the gridded forecast process.

11. Supporting • Input on Digital Services Directives Transition (10-23xx) • Ensure scientific integrity for each weather element in 10-23xx Directive series • Gridded MOS • Support the efforts of MDL and ESRL/GSD in getting gridded MOS into GFE. • Gridded Verification • To encourage the creation of gridded verification to allow forecasters to assess their skill at forecasting in the gridded world. Also, to encourage the forecasters to look at there verification scores on a grid and improve on their forecasts. • CONUS AoR • Support development and execution of RTMA/AoR, including strong advocacy position. • Smart Tool/Init Team • To support the role of the ST/SIT through direct correspondence and quarterly updates from the Team Chair. • ESRL/GSD & MDL projects (as requested) • Act as sounding board/testbed/advisory committee to ESRL/GSD & MDL on an as-needed, as-requested basis • Climatology • Maintain and Improve Current Climatology Grids for GFE

12. Ongoing • Input on directives • Periodic review of roadmap (Quarterly) • Training

13. Delivery Vehicles • Forums • Briefings (Corp Board S&T Sub-Committee, DS-PAC, Regions, MICs, SOOs, etc) • Input to DS-PAC • Field surveys • White papers • One-pagers or memos • Form new teams • Ex-officio membership on other teams • Dialogue with OS&T Director, regions, SOOs, etc • Input on training requirements

14. RTMA / AOR

15. NWS AOR IWT Lee Anderson (co-chair), OST Brad Colman (co-chair) Fred Branski, OCIO Geoff DiMego, NCEP EMC Brian Gockel, OST MDL Dave Kitzmiller, OHD Chuck Kluepfel, OCWWS Art Thomas, OCWWS Bill Ward, NWS PR Al Wissman, OOS NCEP/EMC Geoff DiMego Ying Lin Manuel Pondeca John Horel, University of Utah Bob Aune, NESDIS Other partners and supporters: NCDC, NESDIS OAR, FSL, OST SEC AMS The RTMA/AOR Team

16. Mesoscale Analysis Committee (MAC) • Robert Aune, NOAA/NESDIS University of Wisconsin Space Sciences and Engineering Center • Stanley Benjamin, Forecast Systems Laboratory • Craig Bishop, Naval Research Laboratory • Keith A. Brewster, Center for Analysis and Prediction of Storms The University of Oklahoma • Brad Colman (Committee Co-chair), NOAA/National Weather Service -- Seattle • Christopher Daly, Spatial Climate Analysis Climate Service Oregon State University • Geoff DiMego, NOAA/ National Weather Service National Centers for Environmental Prediction • Joshua P. Hacker, National Center for Atmospheric Research • John Horel (Committee Co-chair), Department of Meteorology, University of Utah • Dongsoo Kim, National Climatic Data Center • Steven Koch, Forecast Systems Laboratory • Steven Lazarus, Florida Institute of Technology • Jennifer Mahoney, Aviation Division Forecast Systems Laboratory • Tim Owen, National Climatic Data Center • John Roads, Scripps Institution of Oceanography • David Sharp, NOAA/National Weather Service -- Melbourne Ex Officio: • Andy Edman, Science & Technology Committee representative • LeRoy Spayd, Meteorological Services Division representative • Gary Carter, Office of Hydrology representative • Kenneth Crawford, COOP/ISOS representative

17. Goal: A comprehensive set of the best possible analyses of the atmosphere at high spatial and temporal resolution with particular attention placed on weather and climate conditions near the surface

18. Brief Background and Motivation • WR SOO/DOH IFPS White Paper recommendations: • Develop a national real-time, gridded verification system • Produce objective, bias-corrected model grids for WFO use • 2003: S&T Committee endorsed concept of NDFD-grid matching analyses of forecast parameters • June 2004: Community summit to assess requirements and capabilities • August 2004: OST Director established MAC (Mesoscale Analysis Committee) • October 2004: MAC recommended NOAA develop and implement suite of consistent sensible weather analysis products using current and future technologies: • Develop a strategy for prototype AOR or proof-of-concept (the RTMA -- Real Time Mesoscale Analysis) • Provide mesoscale analyses hourly and within 30 minutes of valid time • Pursue an archive-quality analysis (and complete historical re-analysis)

19. Why does the ISST see this as a critical effort? • WR SOO/DOH IFPS White Paper recommendations: • Develop a national real-time, gridded verification system • Provide full-resolution NCEP model grids • Produce objective, bias-corrected model grids for WFO use • Implement methods to objectively downscale forecast grids • Incorporate climatology grids into the GFE process • Deliver short and medium-range ensemble grids • Produce NDFD-matching gridded MOS • Modify the GFE software to ingest real-time data • Optimize ways to tap forecaster expertise • Real-time, and delayed mode, analyses will be used to: • Verify NWS gridded forecasts • Support model post processing for bias removal • Support forecast guidance development efforts (gridded MOS, etc.) • Enhance forecasters’ situational awareness • Serve as a logical starting point for short-term gridded forecasts • Develop a more robust, climatological database for various studies and applications

20. Program will be executed in three phases • Phase I – Real-time Mesoscale Analysis (RTMA) • Hourly within 30 minutes • Prototype, or proof-of-concept, for AOR • NCEP, FSL, and NESDIS volunteer to build first phase • Matures into quality real-time analysis component • Phase II – Analysis of Record • State-of-the-science analysis (best possible) • Delayed for late arriving data assets • Methodology to be determined (likely community effort) • Accepted ‘truth’ for use in studies and verification • Phase III – Reanalysis • Apply mature AOR methodology retrospectively • 30 year time history of AORs

21. RTMA logistics and timeline • Experimental grids are now complete • Hourly, 5-km NDFD grid, GRIB2 • EMC objective evaluation and comparison • Field assessment early CY2006 • Operational at NCEP Q3 FY2006 • Distribution of analyses and estimate of analysis error/uncertainty via AWIPS SBN as part of OB7 upgrade – end of FY2006 • Archived at NCDC

22. Initial RTMA products • Products transmitted hourly via SBN to AWIPS for field offices • Temperature (2 m) • Dew point temperature (2 m) • Wind (10 m, direction and speed) • Quantitative precipitation (Stage II) • Sky cover • Also includes estimates of analysis error • Spatially and temporally varying • Reflects observation density, observation quality and background quality • Also reflects representativeness of observations • Additional products (e.g., max. temp.) developed and provided later • RTMA information archived at NCDC

23. RTMA Methodology • Temperature, dew point, and wind elements • RUC forecast/analysis (13 km) is downscaled by FSL to 5 km NDFD grid • Downscaled RUC then used as first-guess in NCEP’s 2DVar analysis of ALL surface observations • Estimate of analysis error/uncertainty • Precipitation – NCEP Stage II analysis • Sky cover – NESDIS GOES sounder effective cloud amount

24. Downscaled 2 m Temperature Original 13 km Downscaled 5 km

25. NCEP obtains full complement of observations ALL Surface Obs = 89126 total

26. NCEP GSI-2DVAR Anisotropic covariance functions

27. Wind-following auto-correlation function for moisture for a test point located at (x=140,y=180)

28. Very first examples!

29. NCEP RTMA Precipitation Analysis • NCEP Stage II (real-time) and Stage IV (delayed) precipitation analyses are produced on the 4-km Hydrologic Rainfall Analysis Project grid • The existing multi-sensor (gauge and radar) Stage II precipitation analysis available 35 minutes past the hour • RTMA is mapped to the 5 km NDFD grid and converted to GRIB2 ORIGINAL NDFD GRIB2

30. Hourly Gages Available for Stage II Precipitation Analysis

31. Sky Cover: Effective Cloud Amount • Effective Cloud Amount (ECA, %) • Derived from GOES sounder • Mapped onto 5-km NDFD grid • Converted to GRIB2 for NDGD • Robert Aune, NESDIS (Madison, WI) GOES-12 IR image (11um) ECA from GRIB2 file – 5km grid Derived ECA from GOES-12

32. Evaluation plans (Nov-Jun) • NCEP will establish webpage displaying RTMA • NCEP will attempt to acquire gridded forms of: • ADAS from University of Utah / WR • STMAS from FSL Steve Koch • MatchOb from AWIPS • NCEP will conduct cross validation analysis • Steve Lazarus (ADAS expert) will help with inter-comparison of 2DVar with other analyses • NCEP and ISST will conduct field assessment

33. RTMA Field Assessment • ISST will work with EMC and Kirby Cook (WR) to distribute initial fields to test offices • Grids will be displayable in D2D and ingested into IFPS/GFE • Offices will provide feedback on pluses/minuses – i.e., help define the gap between this proof-of-concept and needed operational quality • Several offices from each CONUS Region will participate • Up and running around the first of year • Results will feedback into EMC development • Will provide test grids for developing gridded verification interfaces (BoiVer and FSL)

34. Phase II: Analysis of Record • RTMA (improved version) will remain in place to support near-real-time needs of forecaster • Considerable research, development, and operational infrastructure (computing) required for complete suite of analysis products of sufficient accuracy to verify NDFD • Best possible analysis will require 3 or 4-DVar • Model needed to move info through space & time – WRF • Analysis capable of using ALL observations – GSI • Later data cut-off to acquire all obs – 18-48 hours • Should use obs precip to drive evolution of land-states lower boundary condition • Combination of 3D-Var and full-physics model requires substantial additional computing resources • Initial steps started to address NWS AOR funding for FY 2008 and beyond, emphasizing NCEP (OSIP, PPBES, etc.) • Additional funding sources need to be identified for broader community efforts to develop next-generation AOR

35. Gridded MOS

36. Gridded MOS • Why do we need a gridded MOS? • When will it be available? • What will gridded MOS include? • How will we know if it is worth the effort? • How do I learn more about gridded MOS?

37. Gridded MOS • ISST Whitepaper (2003) • To provide model bias correction of grids • High-resolution guidance • Improve centralized guidance

39. Gridded MOS • Produced at 12Z and 00Z • 5 km grid (CONUS) ** Initial Development • West of 110 degrees. • Max T, Min T • 3 hr Temp, Dew Point, & RH • 12 hr PoP, (6 hr PoP) • WRH will create IFPS SmartInits

40. Future Gridded MOS • More Variables • Wind direction, Wind speed, 6hr Prob. Thunder, 12 hr Prob. Thunder, Snowfall, Sky • Anticipated full deployment of grids Fall 2006 • Full CONUS (Summer 2006) • Alaska (Fall 2007) • Hawaii (Fall 2008)

41. Gridded MOS Assessment • ISST will conduct assessment for Gridded MOS • Will start with Medford and then Western Region. • Finally with the entire CONUS (OCONUS). Information on Gridded MOShttp://www.nws.noaa.gov/mdl/synop/gmos.html

42. Day 4-7 Assessment

43. Day 4-7 Guidance Assessment • The increasing number of sources (including derived sources and ensembles) of days 4-7 guidance necessitates that an objective evaluation of these sources be conducted in order to: • Foster effective decision making surrounding a proven starting point for the extended portion of the gridded forecast (which could be fluid) • Improve downscaling techniques (e.g. MatchGuidance, Regime PRISM) • Improve HPC gridded guidance (e.g., alternate first guess) • Facilitate bias corrected output • Facilitate assessment of new methods or models before implementation

44. Day 4-7 Guidance Assessment • Suggested procedure • Start soon (software, tools, and training requirements can be developed/defined) • Alternative verification schemes should be used to effectively account for geographical timeliness constraints and varying field practice (i.e. when 4-7 day grids prepared versus available guidance)

45. Day 4-7 Guidance Assessment • Suggested Procedure (continued) • Guidance to be evaluated initially: • HPC Points • HPC Grids • GFSX Grids • GFSX MOS • DGEX Grids • Duration - 1 year with quarterly reports (possibly ongoing) • Verification of grid sources will be completed at non-traditional verification points to capture skill at non-MOS forecast locations • Conducted with grids when RTMA is available

46. Day 4-7 Guidance Assessment • Assessment could be expanded to include other sources such as: Gridded MOS, ensemble MOS and other modeling center grids (e.g., CMC, EMCWF, UKMET, etc) • Sources used by HPC for extended guidance production

47. Digital Forecast Process

48. Digital Forecast Process Field Evaluation • Well over a year ago, the ISST set out to solicit field input on key issues related to the DFP • We drafted a one-page document highlighting what we felt were the key issues • Established an on-line forum for discussion (helpful, but limited input) • Developed three overarching questions based on original document and on-line forum input • ~1 year ago, created regional teams of SOOs and forecasters (initiated by Dave Sharp - SR ISST member at the time) • Either one team per question, or in some cases one team addressed all three questions

49. Digital Forecast Process Field Evaluation • The three questions: • “Within the limits of predictability, what are the optimal spatial and temporal resolutions needed to provide a useful and versatile digital service while maintaining scientific validity?” • “What is the best way to minimize discrepancies and produce a near-seamless NDFD without sacrificing accuracy and consistency?” • "How should each NCEP center support the WFOs contribution to the digital forecast process?"

50. Digital Forecast Process Field Evaluation • Input from the regional teams has now been compiled and summarized, and highlights are presented here • Over the next several months, the ISST will use the regional team input, add our input to the questions, and develop a series of documents (one addressing each question) with greater definition to the forecast process within the gridded paradigm • Findings from the CONOPS Tiger Team will certainly have an influence • These documents will hopefully be an important resource as various D.S. teams forge ahead