The Collaborative Effort Between Stony Brook University and

The Collaborative Effort Between Stony Brook University and the National Weather Service Part 3 – Integration of Mesoscale Models in Operational Weather Forecasting Jeffrey S. Tongue NOAA/NWS Upton, NY Matt Jones, Brian Colle and Michael Charles Stony Brook University

2003 What NWP Models are available to the Operational Forecaster ? ? ?

• NCEP: – NGM (*) – Eta (12 km Operational (*), 8 km HIRES (NMM)) – GFS (*) – 8 km WRF Experimental – RUC (*) 40/20 km and 10 km – SREF (15 members >> 20) • • Eta RSM Eta KF RUC (Grell) – GFS Ensemble (*) (11 Members)

Others: • Environment Canada – GEM – GSM – Ensemble GSM (16 Member) • NOGAPS • ECMWF(*) – Ensembles (50 Members) • UKMET (*) • Various Workstation Eta (*) • Numerous MM 5 (*)

What’s a Forecaster to do? ? ?

Maintain Situational Awareness ! • Formulate Synoptic Scale Conceptual Model • Identify Potential Mesoscale Processes • Utilizes Climatology • Maintain knowledge of Model Details/ Parameterizations

MOST IMPORTANTLY !! • OBSERVATIONS – – Surface Satellite Radar Others

Case 1 – October 27 th • Heavy Rain Over the Northeast

Valid: 12 UTC 27 OCT 03

Surface: Temps/Isobars

850 h. Pa: Heights/RH

24 HR Model QPF

00 UTC 27 Oct 03

AREA FORECAST DISCUSSION. . . UPDATED NATIONAL WEATHER SERVICE UPTON NY 737 PM EST SUN OCT 26 2003. DISCUSSION. . . WITH THE LOW MOVING NORTHWARD INTO CENTRAL NEW YORK AND THE RAIN SLOWER TO PROGRESS EASTWARD WILL UPDATE THE ZONES TO REDUCE THE CHANCE OF RAIN TO THE CHANCE CATEGORY WEST AND NORTH OF NEW YORK CITY BEFORE MIDNIGHT AND SLIGHT CHANCE EAST. THEN INCREASE PROBABILITY TO LIKELY AFTER MIDNIGHT AND OCCASIONAL SHOWERS TOWARD DAYBREAK. ALSO IN THE WEATHER GRIDS. . . WITH AREAS OF FOG ADDED THIS TO ALL THE GRIDS TONIGHT. ADJUSTED SKY COVER GRIDS TO GO WITH THE 100 PERCENT COVERAGE. &&

11: 15 PM 6. 7 um

How does NWS Improve its Coordination and Collaboration ? ? ? ? ?

12 Hr POP Ending: 12 UTC 27 Nov 03

Sky Cover Valid: 12 UTC 27 Nov 03

Dew Point Valid: 12 UTC 27 Nov 03

Temp/Wind Valid: 12 UTC 27 Nov 03

6 Hr QPF Ending: 12 UTC 27 Oct 03

6 Hr QPF Ending: 18 UTC 27 Oct 03

6 Hr QPF Ending: 00 UTC 28 Oct 03

Sky Cover (36 Hr Fcst) Valid: 12 UTC 28 Nov 03

• How do we SOLVE This Problem?

High Resolution Deterministic Models Result in … “Highly Detailed INACCURATE forecasts”

Can Ensembles Provide a better starting point? ?

Oct th 29 Event More Heavy Rain !!! MM 5 Ensemble Members Provide Different Results

48 Hour Fcst Valid: 00 UTC 30 Nov 03 24 Hour QPF

Something Needs to Change!

Summary • Move away from Deterministic Models to Populate the GFE… • • The Model of the Day! Higher Resolution – Better Model! The XYZ Model “looks” more realistic The XYZ Model “appears” better initialized • “Populated the Grids” with the XYZ model NO MORE !

Summary (cont) • Move towards Ensemble populated forecasts. – Means/deviations/category/extremes – Will it capture the atmosphere? • QPF • Wind • Temperature • Discover the benefits and weaknesses of ensembles. • Continue to “Add Value” where possible.

Plans… • • • Improve NWS bandwidth Issues. Integrate Ensemble Data in AWIPS. Integrate Ensemble Data into GFE. Develop Smart Tools. Expand Initial Conditions Ensemble (Canadian/No. GAPS) • Training Staff – PBL, Parameterizations, etc • Learn • Refine