National Convective Weather Forecast Product

1
National Convective Weather Forecast Product

• Currently demonstrated at the Aviation Weather
  Center, with airlines, and on the Internet
• Provides graphical information regarding the
  current and forecast locations of thunderstorms
• Produces 60-min extrapolation forecasts of
  thunderstorm activity on a national scale
• Uses a tracker to follow and extrapolate a
  combined lightning and radar reflectivity field
• Future refinements will include forecast of
  thunderstorm growth and decay

• On the graphic above, the current locations of
  thunderstorm hazards as determined from radar and
  lightning data are depicted based on a 6-level
  intensity scale
• Levels 3 and above are considered convectively
  active regions where thunderstorm hazards are
  present
• The forecast is represented as a cyan colored
  polygon that shows the forecast position of level
  3 and greater storms in one hour, and a vector
  indicating the direction of storm movement
• The forecast is annotated with the speed of the
  storm (in knots) and the top heights (in 100s of
  feet)

2
Convective Hazard Detection
Radar
Lightning Rate

• Radar
• Unisys - VIL and Echo Tops National Mosaics
• Echo Tops allows removal of AP and ground
  clutter
• Lightning
• Global Atmospherics
• Lightning provides
• More timely product
• Radar is 10 to 25 min old
• Lightning is
• Highlights updraft regions
• Fills in regions of radar data voids or latency

Detection Field and Forecast
3
Extrapolation Forecast

• Eliminate Stratiform
• Stratiform-Convective Partitioner
• Steiner and Houze (1995)
• Remove perishable scales
• Wilson 1966
• Elliptical filter -
• Wolfson 1999 (15B.1 Thursday 400 pm
  The FAA Terminal Convective Weather Forecast
  Product Scale Separation Filter Optimization)
• Determine motion vectors
• Dual-threshold cell tracker
• Dixon and Weiner (1993)

4
NCWF Summer 1999 Evaluations

• Tech Center usability study at Comair and Delta
• Comair
• Baseline - June 1-3
• Training at NCAR - June 7
• Training at Comair - June 14-17
• Delta
• Training - Late June
• Usability study July and Aug
• AGFS - RTVS
• June through August
• Currently sending data to FSL in real-time
• Internal validation

5
(No Transcript)
6
(No Transcript)
7
(No Transcript)
8
(No Transcript)
9
Notes on verification statistics

• Different scores measure different things
• CSI H / (M H F)
• Measures relative accuracy
• TSS PODy PODn -1
• Measures discrimination between Yes and No
  observations
• PODy
• Measures proportion of convective area that is
  correctly forecast to have convection
• PODn
• Measures proportion of non-convective area that
  is correctly forecast to not have convection
• FAR
• Measures proportion of forecast convective area
  that is incorrect
• Bias
• Measures the extent of over- or under-
  forecasting

• H Hits
• M Misses
• F False Alarms
• PODy H / (H M)
• PODn proportion of No area that was correctly
  forecast to be No
• FAR F / (H F)
• Bias (F H) / (M H)

10
NCWF Verification notes

• Independent verification provided by Real-Time
  Verification System (RTVS), FSL/NOAA and NCAR
  verification groups, summer of 1999
• http//www-ad.fsl.noaa.gov/afra/rtvs/convective/ma
  in_convective.html
• Observations based on national convective
  detection product (NCDP) over 10 minutes prior to
  forecast valid time
• 20-km grid used as basis for verification
• Observations on a 4-km grid
• Filter At least 12 (50) 4-km boxes within a
  20-km box must have a convective observation in
  order for the 20-km box to count as a Yes
  observation
• Verification approach is demanding - it
  strictly requires overlap of forecasts and
  observations (i.e., no slop allowed for
  getting close!)
• Results are consistent with previous studies of
  these types of evaluations

11
Example 4 June1999, 00 UTC
12
4 June 1999Forecasts valid at 0000 UTC
Area efficiency (PODy / Area) x 100