Docstoc

CWSU Oberlin Workstation ETA

Document Sample
CWSU Oberlin Workstation ETA Powered By Docstoc
					Using a Mesoscale Model to
Identify Convective Initiation
   in an ARTCC/CWSU
         Environment
                 Warren R. Snyder
           NOAA/NWS Weather Forecast Office
                  Albany, New York

                Mark R. McKinley
        NOAA/NWS Center Weather Service Unit
                  Oberlin, Ohio

                      Allison R. Vegh
     Department of Earth and Atmospheric Sciences
   University at Brockport State University of New York
                   Brockport, New York
Why does convective initiation matter to
 the National Airspace System (NAS)
     Causes of air traffic Delays
      •   76 % of involve weather
      •   24% involve thunderstorms
      •   17% Ceilings
      •   14% each for visibilities and wind

     Result…Ground Stops & Delays
      • Significant costs from $3K to 300K/flight
      • Safety Issues
      • Your sitting in the airport delayed for hours or
        days !!!
         Purpose of the Study
   Improve convection initiation forecasts of
    CWSU Oberlin using a mesoscale model
   Can better forecasts reduce ground stops,
    improving convection avoidance, and
    earlier or later re-routings
   Proof of concept
   How much $$$ can the industry save?
   To run the full modeling system in a
    CWSU costs $49 a month for a T1 and
    $4000 for the PC
   How much does jet fuel cost?
      WSETA project at CWSU Oberlin,
        Ohio and WFO Albany, NY
   Training Provided to CWSU Staff April 2004
   Model data posted on CSTAR server twice a day via gempak
    graphics on the internet.
   SUNY Student compared six fields identified by MIC/SOO as most
    likely to be indicative of convective initiation using data from
    Summer 2004 to early June 2005.
   Parameters with best performance
    • Hourly Convective Precipitation
    • 700 HPA Omega
   Modeling system - Dell Pentium 4 , 2400 mHz Linux PC
   Model run for 24 hours. 06 UTC and 18 UTC runs with output
    posting 0800 UTC/2000 UTC
   Kain-Fritcsh Convective Parameterization, Nested - Outer nest
    resolution is 15 km, Inner Nest 7.5 km resolution, Diffusion is 0.30
    versus 1.0 in operational models. Most parameters configured at
    SOO/STRC baseline.
Study Area – Oberlin Service Area
          Squared off
                           Results Part I
Performance   850 hPa   Hrly          700 hPa   850 hPa Jet   250 hPa      BL
              Theta-E   Convective    Omega                   Divergence   Convergence
                        Precipitation
Good              6          5            9          4             6            6

Acceptable        8         19           18          5            14           12

Poor              15         5            2          17            9           11



Good             21%        17%          31%        14%          21%          21%

Acceptable       28%        66%          62%        17%          48%          41%

Poor             52%        17%          7%         59%          31%          38%

   Definitions - % of convection
    in area forecasted by model
    parameter                            HCP - Acceptable or Good 83%
     • Good – 75% or more                700 hPa Omega- Acceptable or
     • Acceptable – 25% to 75%           Good 93%
     • Poor – Less than 25%
Hourly Convective Precipitation
  (HCP) vs. 700 hPa Omega
     Part II – Comparing 700 hPa
     Omega and HCP forecasts to
               NLDN Data
   Used all convective events from June to
    September 2005 in CWSU Oberlin area
   NLDN data plots every lightning stroke at
    exact lat/lon points
   Assess accuracy of model parameters in
    time and space
   UAlbany students extracted the data
   Software developed by Vasil Koleci to plot
    lighting data hourly over the area
   87 events identified, 3 dropped as only
    SHRA and no lightning occurred
      Comparison of Model 700 hPa
        Omega/HCP Convective
      Configuration with NLDN data
   Excellent (4)–Model data match in
    location, and structure/orientation.
   Good – (3) If structure/orientation are
    very similar but location is off by 160 km,
    or location is within 160 km and
    structure/orientation are different
   Fair – (2) If they are both in the same ¼
    of a state or states, or overlap each other
    25% or less
   Poor – (1) No match
        Results for Configuration

   700 hPa Omega had an average of
    3.08 configuration. 74 events where
    good or excellent, only 10 fair and
    poor

   HCP – Had an average 2.81
    configuration. 63 Events good or
    excellent, 21 Fair or Poor
        Results for Timeliness
 700 hPa Omega and HCP forecasted
  the time of initiation at the same
  time in all but 9 cases
 Average model time error for
  convective initiation was 37.5
  minutes
    • 54 events were forecast within an hour
    • 19 events between one and two hours
    • 11 events between two and three hours
               Distance Error
   Distance from the model convection to the
    actual lighting at initiation
   Average for 700 hPA Omega 33 km
    • All but one event within 240 km
    • 240 km is the distance a jet travels in 15
      minutes
   Average for HCP 86 km.
    • All but 5 events within 240 km
   Difference likely the result of the much
    more specific locations of HCP versus the
    general broader areas of 700 hPa Omega
           Future Directions
 Model converted over to the WRF
 WRF data being provided to CWSUs at
  Nashua and Oberlin
 Will undertake a similar study during the

  2007 convective season using the WRF
 New products available such as model
  Composite Reflectivity
            References
 Based on ER Tech Attachment 2006-01
  “Using a Mesoscale Model to Identify
  Convective Initiation in an
  ARTCC/CWSU Environment” 2006:
  Snyder W.R, McKinley M.R. and Vegh
  A.R.
 Original Paper can be found online at

 http://www.werh.noaa.gov/SSD/erps/t
  a/ta2006-01.pdf

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:7
posted:7/5/2011
language:English
pages:14