1 / 52

The NCEP Operational Modeling Suite: FY11 Highlights

William. M. Lapenta Deputy Director Environmental Modeling Center NOAA/NWS/NCEP Contributions from many others. The NCEP Operational Modeling Suite: FY11 Highlights. NWS Seamless Suite of Forecast Products Spanning Weather and Climate. Outlook. Guidance. Threats Assessments.

anise
Download Presentation

The NCEP Operational Modeling Suite: FY11 Highlights

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. William. M. LapentaDeputy DirectorEnvironmental Modeling CenterNOAA/NWS/NCEPContributions from many others The NCEP Operational Modeling Suite: FY11 Highlights

  2. NWS Seamless Suite of ForecastProducts Spanning Weather and Climate Outlook Guidance Threats Assessments Forecast Lead Time Forecasts Watches Warnings & Alert Coordination Benefits NCEP Model Perspective Forecast Uncertainty Years Seasons Months • Climate Forecast System 2 Week • North American Ensemble Forecast System • Global Ensemble Forecast System 1 Week • Land Surface • Ocean • Waves • Tsunami • Space Weather • Tropical Cyclone • Global Forecast System • Short-Range Ensemble Forecast Days • North American Forecast Hours • Rapid Update Cycle for Aviation Minutes • Dispersion Models for DHS • GFDL • HWRF Health Aviation Recreation Ecosystem Agriculture Commerce Hydropower Environment Maritime Fire Weather Life & Property Energy Planning Reservoir Control Emergency Mgmt Space Operations

  3. Upgrades to Global System • Data Assimilation (17 December 2009) • Assimilate: • NOAA-19 AMSU-A/B, HIRS • RARS 1b data • NOAA-18 SBUV/2 and OMI • Tropical Storm SLP • Improved use of GPS RO observations • Refractivity forward operator • Better QC procedures Metop/GRAS, GRACE-A and CHAMP, COSMIC • Increased Horizontal Resolution and Physics Upgrade (28 July 2010 • Horizontal resolution • T382 (35km) to T574 (~28km) & 64L for 0 to 192hr forecast • Remains T190 (~70km) & 64L for 192-384 hr forecast • Upgrade GFS physics (Radiation, shallow/deep convection and PBL) • Modified SW and LW radiation calculations (aerosol effects, co2, cloud interaction) • Detrainment from all levels (deep convection) • PBL diffusion in inversion layers reduced (decrease erosion of marine stratus) • Includes gravity wave drag (effect of topography)

  4. Impact of GFS Physics Upgrade Implemented 28 July 2010 Updated GFS physics package eliminates grid-point precipitation “bombs” • Modify GFS shallow/deep convection and PBL • Detrainment from all levels (deep convection) • PBL diffusion in inversion layers reduced (decrease erosion of marine stratus) • GSI/GFS Resolution • T382 (~35km) to T574 (~28km) & 64L 24 h accumulated precip ending 12 UTC 14 July 2009 Operational GFS Observed Upgraded Physics GFS

  5. Physics Upgrade has a Positive Impacton GFS Quantitative Precipitation Precipitation Verification (CONUS) 29 Jun – 28 Jul 2010 • 24h, 48h & 72h forecasts statistics • GFS upgrade implemented 28 July 2010 • Precipitation patterns remain unchanged (ETS similar) • High bias for large precipitation thresholds significantly reduced • Impact on lower precipitation thresholds minimal Equitable Threat 2.6 0.8 Bias

  6. 500mb AC Days 0-9 NH for August 2010 6

  7. 500mb AC Days 0-9 NH for September 2010 6

  8. 500mb AC Days 0-9 SH for September 2010 6

  9. 500mb AC Day 5 for the Past 31 Days NH 20º- 80º 0.6 SH 20º- 80º 0.6

  10. GFS Upgrade EPAC Hurricane Track Error2008 Hurricane Season 32% Improvement over 2008 Prod 25% Improvement over 1Q10 Prod 2008 Operational – Blue December GFS Package – Red T574 with upgraded Physics - Green

  11. Atlantic Hurricane Track Error (NM)2010 Hurricane Season (through 16 September) NCEP GFS ECMWF GFS similar to EC

  12. Planned Global Data AssimilationSystem Upgrade (FY11) • Joint Development with JCSDA, NASA, ESRL • New Observations • ASCAT winds • Radar and Lightning data IO and cloud analysis for RR(GSD, DTC) • NOAA-19 SBUV/2 • NOAA-19 MHS – remove NOAA-15 AMSU-BQC and data handling • QC and Data Handling • Upgrade quality control of MHS/AMSU-B water vapor channels • Improve location of Buoys in vertical (move from 20 to 10m) • Improved OMI QC • Removal of redundant SBUV/2 total ozone • Retune SBUV/2 ozone ob errors • Relax AMSU-A Channel 5 QC • Remove down weighting of collocated radiances

  13. The NCEP Climate Forecast System Reanalysis Suranjana Saha, Shrinivas Moorthi, Hua-Lu Pan, Xingren Wu, Jiande Wang, Sudhir Nadiga, Patrick Tripp, Robert Kistler, John Woollen, David Behringer, Haixia Liu, Diane Stokes, Robert Grumbine, George Gayno, Jun Wang, Yu-Tai Hou, Hui-ya Chuang, Hann-Ming H. Juang, Joe Sela, Mark Iredell, Russ Treadon, Daryl Kleist, Paul Van Delst, Dennis Keyser, John Derber, Michael Ek, Jesse Meng, Helin Wei, Rongqian Yang, Stephen Lord, Huug van den Dool, Arun Kumar, Wanqiu Wang, Craig Long, Muthuvel Chelliah, Yan Xue, Boyin Huang, Jae-Kyung Schemm, Wesley Ebisuzaki, Roger Lin, Pingping Xie, Mingyue Chen, Shuntai Zhou, Wayne Higgins, Cheng-Zhi Zou, Quanhua Liu, Yong Chen, Yong Han, Lidia Cucurull, Richard W. Reynolds, Glenn Rutledge, Mitch Goldberg Bulletin of the American Meteorological Society Volume 91, Issue 8, pp 1015-1057. doi: 10.1175/2010BAMS3001.1

  14. Improved Precipitation Signal in CFSR Correlation of intraseasonal precipitation with CMORPH R1 R2 CFSR Contours are shaded starting at 0.3 with 0.1 interval.

  15. Advancement of Climate Forecast System (CFS) Planned for Q2FY11

  16. 0 UTC 6 UTC 12 UTC 18 UTC Operational Configuration for the Climate Forecast System V2 (FY11) A total of 16 CFS runs every day • 4 runs @ 9 months length • 3 runs @ 1 season • 9 runs @ 45 days 9 Month Seasonal 45 Day

  17. Preliminary Analysis Shows Improved Western Hemisphere MJO Signal in CFSV2 WH-MJO Index 09 Feb to 13 Mar WH-MJO Index 09 Feb to 13 Mar CFS Operational CFS V2 Skill Skill Days Days WH-MJO Index 09 Nov to 13 Dec WH-MJO Index 09 Nov to 13 Dec CFS Operational CFS V2 Skill Skill Days Days

  18. Hybrid Ensemble-Variational Data Assimilation Development A partnership to develop and implement a hybrid 4D-VAR system • Joint venture between ESRL, NASA/GMAO, Univ of Oklahoma and NCEP • Hybrid uses ensemble-based information to improve representation of flow-dependent background errors • Must serve global ensemble system as well as analysis • Global, regional and hurricane applications

  19. Hybrid Cycling Experiments • Full resolution cycling experiment • T382L64 deterministic analysis and forecasts • Same period as EnKF runs (2008 Hurricane Season) • Re-ran control run utilizing latest versions of GSI/GFS • “One-way coupled” • T190L64 perturbations from offline EnKF run used for B estimate [no feedback to ensemble system] • Localization scales for alpha set similar to EnKF run • Not exact since current localization implementation for hybrid differs from EnKF • (b1)-1=0.5; TLNMC utilized • conservative parameter settings • Provide baseline, test mechanics, and find potential issues

  20. GFS EnKF HYB GFS EnKF HYB GFS EnKF HYB Preliminary Results Fits of 06h forecasts to radiosondes (O-F, wind) for 20080815-20080915 Northern Hemisphere Tropics Southern Hemisphere Hybrid fits are better than control 3DVAR and EnKFEnKF fits are for high-resolution deterministic forecast (not ensemble mean)

  21. EnKF Testing During2010 Hurricane Season • Data assimilation: 80-member T254L64 using 2010 GFS • Ensemble forecasts: 20 members to 5 days lead • T574L64 and T254L64 control from ens mean also • Bug fixes, mid-August 2010: • Assimilate more hyperspectral data (AIRS and IASI) ; last year’s we mistakenly used only lowest ~1000 channels. • Fixed scan-angle bias correction bug for satellite radiances; wasn’t updating bias. 21

  22. Global Hurricane Track Error (NM)2010 Hurricane Season (through 16 September) 22

  23. Supercomputing at NCEP IBM Power6 p575 • 69.7 Teraflops Linpack • 156 Power6 32-way Nodes • 4,992 processors @ 4.7GHz • 19,712 GB memory • 320 TB of disk space per system • 17 PB tape archive Fairmont, West Virginia Cirrus— (backup) Gaithersburg, Maryland Stratus— (primary)

  24. Production Suite on IBM P6September 2010 Number of Nodes Glob HYCOM 12 06 06 12 00 18 00 Time (UTC)

  25. Production Suite on IBM P6Projected: September 2011 Number of Nodes 06 12 00 18 00 25 Time (UTC)

  26. Thanks for your attention

  27. GFS Upgrade Atlantic Hurricane Track Error (2008 Hurricane Season) 2008 Operational – Blue December GFS Package – Red T574 with upgraded Physics - Green

  28. Regional Ensemble: system upgrade (SREF, 2011) • Current SREF System: • 4 models (WRF-NMM, WRF-ARW, ETA, RSM), 32km, 3.5 days, 21 members, Breeding IC perturbations • Planned Changes: • Eliminate 6 Eta and 5 RSM members (11 members) • Add 7 NEMS-NMMB, 2 WRF-ARW and 2 WRF-NMM members (11 members) • Update WRF-NMM and WRF-ARW model versions • Increase horizontal resolution from 32km to 22 km • Downscaled ET perturbations from global ensemble • Expected Benefits: • Reduced maintenance costs without legacy models (Eta & RSM) • Inclusion of state-of-the-art model – improved forecasts • More detailed and accurate forecasts especially for precipitation • More consistent with global ensemble

  29. Dyn Dyn Dyn Phy Phy Phy Wrt Wrt Wrt Chm Chm Chm NEMS V.3 Component Structure Start time Run length MAIN Ensemble Component NEMS Coupler Component EARTH(1:N) Choice of physical domain and component Ice Ocean Atmos Component execution HiRAM? NMM GFS FIM Domains(1:M) Individual nests Physics Dynamics Choices And I/O 1 Below the dashed line the cores can be organized as best suits their own needs.

  30. FIM design – vertical coordinate • Hybrid (sigma/ isentropic) vertical coordinate • Improved fidelity in vertical structure using quasi-material surfaces, reduced vertical dispersion. • Grid points that follow vertical motion (“Lagrangian” grid) can prevent numerical dispersion during wave-induced vertical transport. (q coordinate ) • Improved transport by reducing numerical dispersion from vertical cross-coordinate transport, improved stratospheric/tropospheric exchange. • Used in NCEP Rapid Update Cycle (RUC) model • Used in HYCOM ocean model • Builds on work by • Rainer Bleck, • RUC group (Benj et al. 2004-Mon.Wea.Rev.), • Don Johnson et al. (U. Wisconsin),

  31. GFS FIM (GFS physics) FIM and GFS March – April 2010 NH – Day 6 Dec 2009 – Feb 2010 NH SH FIM – GFS difference

  32. GFS NMM-B NMM-B NMM-B and GFS (spectral) 20 June 2009 – 16 June 2010 500 hPA Anomaly Correlation

  33. NAM NEMS based NMM Bgrid replaces Egrid Parent remains at 12 km Multiple Nests Run to ~48hr ~4 km CONUS nest ~6 km Alaska nest ~3 km HI & PR nests ~1.5-2km DHS/FireWeather/IMET possible Rapid Refresh WRF-based ARW Use of GSI analysis Expanded 13 km Domain to include Alaska Experimental 3 km HRRR NCEP Mesoscale Modeling for CONUS: Planned Q3-Q4 FY11 WRF-Rapid Refresh domain – 2010 RUC-13 CONUS domain Original CONUS domain Experimental 3 km HRRR

  34. 4 km NMM-B CONUS Nest – 36 h Fcst3 April to 27 September 2010 RMS Temperature Error Temperature Bias RMS Vector Wind Error

  35. Planned Global Ensemble Upgrade (Q4 FY11) • Current: GEFS and GEFS/NAEFS post processing • T190L28 for all 384 hours lead-time • 20+1 members per cycle, 4 cycles per day • Planned Upgrades: • 0 to 192h ---- T254L42 (was T190L28) • 192 to 384hr ---- T190L42 • Improving ensemble initialization • Improving STTP (stochastic total tendency perturbation) method

  36. Plans for Wave Modeling • Increase spectral resolution from (24 directions x 29 frequencies) to (36 directions x 50 frequencies) • Data assimilation • Adding curvilinear grids (NRL) • Develop a tri-polar grid for the Arctic Ocean • Singularity in spectral direction to be resolved and to be implemented with two way nesting • Unstructured grids (IFREMER, Aron Roland) • Currently being transitioned to the development version of the model along with the curvilinear grid model version • Intended for coastal use and direct coupling with unstructured grid storm surge models • Intended to be part of grid mosaic with two way nesting

  37. Plans for Wave Modeling • Physics upgrade • Wave generation (dissipation and non-linear interaction) in wind – wave models • First phase of this upgrade to be put in operations in FY 2011 • Coupling hurricane wave models to storm surge, including currents • A 3-way coupled (hurricane – wave – circulation) system for improved hurricane prediction. • A relocatable wave model • Support local wave modeling capability for WFOs and partners as established by pilot projects using SWAN • Supported for some WFOs by USACE • Basic system to be WAVEWATCH III centric, but will support SWAN • A quasi-stationary extension to the wave model being developed to increase computational efficiency

  38. FY11 : National AQ Forecasting Plans • Upgrade to NAM-B CMAQ V4.7 System • Updated Emissions from 2008 NEI • NAM-B loose grid coupling • Improved deposition, PBL physics • GEOS-Chem LBCs • Improved gas/aerosol mechanisms (CMAQ V4.7.1, CB05) • Development/Testing • 4 km NAM-B nest coupling • Smoke sources from NESDIS, dust • GOCART PM LBC coupling • Common NAM-B tight grid coupling • Surface PM2.5 data assimilation • Shallow/deep convective mixing • Ozone bias correction CB05 O3 Bias Developmental PM2.5 Forecast (ug/m3)

  39. Global aerosol forecast and analysis systemSarah lu, Ho-Chun Huang, C. Perez, Z. Janjic Goal: Improving weather and air quality forecasts by incorporating prognostic aerosols in GFS/NMMB and assimilating global aerosol information in GSI via NCEP-NASA/GSFC-Howard University collaborations MODIS fire emissions Regional AQF Emissions Global forecast and analysis system Dynamic LBCs GOCART Modeling Atmos. Correction SST Analysis Data Assimilation Algorithm Color key Validation NASA obs and tech ROSE project Various datasets AERONET, OMI, CALIPSO Satellite data NCEP DSSs

  40. Goddard Chemistry Aerosol Radiation and Transport (GOCART) 1x1° model driven by the NCEP GFS AERONET MODIS,OMI,MISR CALIPSO Dongchul Kim (NASA GSFC)

  41. Developing a Boundary Layer Height Analysis for Dispersion Applications • GOAL: Produce a National PBL analysis to drive dust, wildfire smoke models • Task 1: Extension of Real Time Mesoscale Analysis • Unified RTMA assimilation codes; added generalized tracers • Aircraft derived PBL assimilated into RTMA • Task 2: Utilizing NASA instrument technologies • Evaluated PBL from COSMIC GPS/RO using cloud clearing techniques • Derived PBL from MPLNET & used for evaluation • Coordinated a PBL field program to evaluate analysis PBL Hgt analysis : 10/18/09 w Aircraft & Sondes • FY11 planned activities • Complete calibration CALIPSO PBL heights • Ingest additional measurements into RTMA • Impact studies with dispersion models MPLNET/GSFC: Sept. 14, 2009

  42. 2 Meter Temperature Trend in New Reanalysis(1979 – 2010) 1.02K GHCH_CAMS 0.94K CFSR 0.66K R1 • CFSR has less bias than R1 relative to GHCN_CAMS • Upward trend in CFSR larger than in R1 more like GHCN_CAMS

  43. SST and Precipitation Relationshipin New ReanalysisPrecipitation-SST lag correlation in tropical Western Pacific • Response of Precipitation to SST increase: • Warming too quick in R1 and R2 • simultaneous positive correlation in R1 and R2 Courtesy: Jiande Wang, EMC

  44. Hindcast Configuration for next CFS 9-month hindcasts will be initiated from every 5th day and will be run from all 4 cycles of that day, beginning from Jan 1 of each year, over a 28 year period from 1982-2009 This is required to calibrate the operational CPC longer-term seasonal predictions (ENSO, etc) There will also be a single 1 season (123-day) hindcast run, initiated from every 0 UTC cycle between these five days, over the 12 year period from 1999-2010. This is required to calibrate the operational CPC first season predictions for hydrological forecasts (precip, evaporation, runoff, streamflow, etc) In addition, there will be three 45-day (1-month) hindcast runs from every 6, 12 and 18 UTC cycles, over the 12-year period from 1999-2010. This is required for the operational CPC week3-week6 predictions of tropical circulations (MJO, PNA, etc) Total number of years of integration = 9447 years !!!!! Jan 2 0 6 12 18 Jan 1 0 6 12 18 Jan 3 0 6 12 18 Jan 4 0 6 12 18 Jan 5 0 6 12 18 Jan 6 0 6 12 18 1 season run 45 day run 9 month run Courtesy: Suru Saha, EMC

  45. Operational Configuration for next CFS There will be 4 control runs per day from the 0, 6, 12 and 18 UTC cycles of the CFS real-time data assimilation system, out to 9 months. In addition to the control run of 9 months at the 0 UTC cycle, there will be 3 additional runs, out to one season. These 3 runs per cycle will be initialized as in current operations. In addition to the control run of 9 months at the 6, 12 and 18 UTC cycles, there will be 3 additional runs, out to 45 days. These 3 runs per cycle will be initialized as in current operations. There will be a total of 16 CFS runs every day, of which 4 runs will go out to 9 months, 3 runs will go out to 1 season and 9 runs will go out to 45 days. 0 UTC 6 UTC 12 UTC 18 UTC 1 season run (3) 45 day run (9) 9 month run (4) Courtesy: Suru Saha, EMC

  46. CFSv1 versus CFSv2 Anomaly Correlation: All Leads (1-8), All Months (10) Green is good Red is not good Courtesy: Huug van den Dool, CPC

  47. Anomaly Correlation for other Regions (collaboration with EUROSIP and India) All Leads (1-8), All Months (10) Green is good Red is not good Courtesy: Huug van den Dool, CPC

  48. Preparing for Global Ocean Modeling Capability in FY11 • Motivation for global system: • NOS, OPC, TPC and WFO need for real time high resolution eddy-resolving ocean products for customers • Ocean component of future planned CFS upgrades • Coupled regional hurricane modeling • Centerpiece of integrated ocean modeling system • Based on a 1/12 degree HYCOM (HYbrid Coordinate Ocean Model) with a Pan-Am Global Grid (4500 x 3928) • 32 vertical hybrid layers (isopycnal in the deep, isolevel in the mixed layer and sigma in shallow waters) • Initialization based on a MVOI scheme developed by the US Navy that assimilates daily observations (T,S, U,V and sea surface height) in a sequential incremental update cycle to produce analysis • Daily global ocean forecasts (for up to a week) will be forced with the GFS surface fluxes, radiation, precipitation and momentum

  49. Gulf Stream & Kuroshio Current

More Related