Future NCEP Guidance Support for Surface Transportation

1. Future NCEP Guidance Support for Surface Transportation Stephen Lord Director, NCEP Environmental Modeling Center 26 July 2007

2. Overview • Weather for Roads, Air transportation, etc. • National picture • New ensemble products • Local picture • Downscaling • Real-time Mesoscale Analysis (RTMA) • Land Information System (LIS) • Dynamical – Statistical approach • Marine applications • Waves • Water levels • Data availability • What’s needed to move ahead

3. New Ensemble Products fromNCEP Storm Prediction Center • NCEP Short-Range Ensemble Forecast (SREF) System • National coverage ~ 30 km grid • Probabilistic guidance with extremes SREF Maximum (any member) 3h Accumulated Snowfall SREF Pr[Ptype = ZR] and Mean P03I (contours) SREF 6h Calibrated Probability of Snow/Ice Accum Accumulation based on MADIS road surface condition D. Bright NCEP/SPC

4. SREF Likely PTYPE and Mean P03I (contours) 24 h Fcst Precip Type, Amount 32 F Isotherm ZR Snow D. Bright NCEP/SPC IP Rain

5. Downscaling • Future computing requirements • National scale ~20 years to reach sufficient resolution • Dynamical-statistical approach • Real time Mesoscale Analysis (RTMA) • Land Information System (LIS) • Bias correction and statistical processing • Components under development

6. 5 km National (NGDG) grid (eventually 2.5 km) Hourly analysis Focus on “drawing to obs” (mesonet) Temperature, precipitation, surface wind, dew point Anisotropic (e.g. land-water contrast) Analysis uncertainty To include cloud cover Will cover CONUS, Alaska, Hawaii, Puerto Rico, Guam Real-Time Mesoscale Analysis (RTMA) M. Pondeca J. Purser G. DiMego NOAA/GSD - RUC RTMA Temperature Analysis (° F) (17Z 6/14/07) RTMA Temperature Analysis Uncertainty (° F) (17Z 6/14/07) RTMA 1-hour Precipitation Analysis (inches) (01z 6/14/07)

7. Land Information System (NASA/NOAA) S. Kumar Jim Geiger C. Peters-Lidard J. Meng K. Mitchell • Land states forced by • Observed precipitation • Model solar, long wave radiation, cloudiness • Noah Land Surface Model (LSM) defines skin temperature, soil moisture, etc. • Can be run at 1 km resolution (below) Surface (skin) Temperature 50 km area Washington DC NASA LSM GFS forcing 00 UTC 1 July – 21 UTC 1 July 00 UTC 7 PM 03 UTC 10 PM 06 UTC 1 AM 09 UTC 4 AM 12 UTC 7 AM 15 UTC 10 AM 18 UTC 1 PM 21 UTC 4 PM

8. Dynamical Statistical Approach • Bias correction of forecast fields with respect to model analysis (e.g. NAM) • “Downscaling Transformation” (DT) • Produces time-dependent differences between coarse forecast model (e.g. 12 km NAM) and RTMA (5 km) • Downscaled (local) fcst = NAM fcst + Bias correction + DT • On local grid • Probabilistic products • Created from ensemble systems (SREF, GENS) through Bayesian Model Averaging (BMA) approach • Applications for • Road transportation • Air transportation management (NEXTGEN) • Severe weather forecasting

9. Marine ApplicationsMulti-Grid Wave Modeling Higher coastal model resolution Deep ocean model resolution Highest model resolution in areas of special interest Multi-grid wave model tentative resolutions in minutes for the parallel implementation in FY2007-Q4. Hurricane nests moving with storm(s) like GFDL and HWRF Wave ensemble system application for ship routing

10. NCEP Real-Time Ocean Forecast System (RTOFS)Operational December 2005, upgraded June 2007 • RTOFS provides • Routine estimation of the ocean state [T, S, U, V, W, SSH] • Daily 1 week forecast • 5 km coastal resolution • Initial and boundary conditions for local model applications • Applications • Downscaling support for water levels for shipping • Water quality • Ecosystem and biogeochemical prediction • Improved hurricane forecasts • Improved estimation of the atmosphere state for global and regional forecasts • Collaboration with NOAA/NOS Chesapeake Bay

11. Product Availability • Three levels of information • Routinely delivered • Pointwise, single-valued, downscaled MLF* from all available guidance on NDGD grid • Description of forecast uncertainty through probability density function (mode & 10/90 %ile) • Accompanying post-processed fields • Meteorologically consistent • Closest to MLF* • “On-demand” (via publicly accessible server) • Individual ensemble member forecasts available • Prototype: NOMADS * MLF – Most Likely Forecast

12. What’s Needed? • Written requirements for surface transportation to NWS • Operational (and research) computing resources • Acceleration of current dynamical-statistical efforts • Outreach and coordination with local users

13. Concurrent execution of global and regional forecast models (Phase 2) Analysis Global/Regional Model Domain Model Region 1 • Real time boundary and initial conditions available hourly • “On-demand” downscaling to local applications • Similar to current hurricane runs but run either • Centrally at OR • Locally (B.C, I. C. retrieved from on-line data) • No boundary or initial conditions older than 1 hour • Flexibility for “over capacity” runs (e.g. Fire Wx, Hurricane) • Using climate fraction must be planned • No impact on remainder of services • For NEXTGEN: A consistent solution from global to local with a single forecast system and ensembles providing estimate of uncertainty Model Region 2 Local Solution