Verification of the 88D Hail Detection Algorithm at WFO Cheyenne

1
Verification of the 88D Hail Detection Algorithm
at WFO Cheyenne

• Mike Weiland
• WFO Cheyenne

2
(No Transcript)
3
Outline of Presentation

• Reason for Study
• Hail Climatology
• HDA
• Past Verification Studies
• Results
• Future Plans and Ideas

4
(No Transcript)
5
Reason for Study

• Hail is a primary severe weather threat in the
  WFO Cheyenne CWA.
• Billions of dollars in crop and property damage
  is caused each year by hail.
• Hail Detection Algorithm is one of the forecast
  tools available.

6
Frequency of Severe Hail
7
Frequency of Any Sized Hail
8
The Hail Detection Algorithm

• Developed in 1982 was based upon structural
  characteristics of typical southern plains severe
  hailstorms.
• Latest HDA is reflectivity based and utilizes the
  RADAP II VIL algorithm.
• Uses the latest Storm Cell Identification and
  Tracking (SCIT) algorithm.
• From the SCIT, the height and maximum
  reflectivity of each storm component is used to
  create a vertical reflectivity for each cell.
• HDA also uses the height of 45 dBz return above
  the melting level.

9
The Hail Detection Algorithm(Severe Hail Index)

• Develops a reflectivity to hail relation.
• Filters out most of the lower reflectivities as
  they tend to be associated with liquid water.
• Uses a temperature weighted vertical integration
  (since hail growth only occurs at temperatures
  lt0C and most severe hail growth occurs with
  temperatures near -20C or colder.
• Is used to create the Maximum Expected Hail Size
  (MEHS)

10
The Hail Detection Algorithm

• Works best on the plains
• Witts study in Colorado (1998) showed that the
  HDA correctly determined the cell that was
  producing hail 92 of the time.
• The POD from Witts study is 87 for 1 hail and
  96 for 2 diameter hail.

11
Past Area HDA Verification Studies

• Vasiloff (UT) mean absolute error was .22 and
  POD of HDA 86.
• Maddox (AZ) variations in terrain have a large
  impact on POSH. POSH has a large overforecasting
  bias.
• POSH of small storms seems to be too high.

12
Methodology

• 185 severe hail reports within the WFO Cheyenne
  CWA from May-August 2005 and 2006 were compared
  with the output from the HDA.
• Using archived data, the HDA value and VIL for
  each severe report was found.

13
Concerns

• The maximum hail size may not have been captured
  in some storms.
• The study only looked at severe hail.

14
HDA Hail Size Compared to Hail Size (.75
orlarger) 185 events in the WFO Cheyenne CWA in
2005 and 2006

• Hail Size Events HDA vs. Hail Size
  (inches)
• gt1.00 35 .0493
• .88 to 1.00 86 .4502
• .75 62 .5141
• Total 185 .3306
• Hail Size Events HDA Larger HDA
  Smaller
• gt1.00 35 18 times
  14 times
• .88 to 1.00 86 62 times
  15 times
• .75 62 55 times 4
  times
• Total 185 135 times
  33 times

15
HDA Hail Size Compared to Hail Size (.75
orlarger) Based on Distance from the KCYS
88D185 events in the WFO Cheyenne CWA in 2005
and 2006

• Greater than 125 nmi
• Hail Size Events HDA vs. Hail Size
  (inches)
• gt1.00 14 -.1429
• .88 to 1.00 28 .1807
• .75 14 .8570
• Hail Size Events HDA Larger HDA
  Smaller
• gt1.00 14 4 times
  9 times
• .88 to 1.00 28 15 times
  9 times
• .75 14 8 times
  4 times

16
HDA Hail Size Compared to Hail Size (.75
orlarger) Based on Distance from the KCYS
88D185 events in the WFO Cheyenne CWA in 2005
and 2006

• 76 to 125 nmi
• Hail Size Events HDA vs. Hail Size
  (inches)
• gt1.00 11 .0091
• .88 to 1.00 28 .3146
• .75 28 .4792
• Hail Size Events HDA Larger HDA
  Smaller
• gt1.00 11 9 times
  2 times
• .88 to 1.00 28 22 times
  2 times
• .75 28 25 times
  0 times

17
HDA Hail Size Compared to Hail Size (.75
orlarger) Based on Distance from the KCYS
88D185 events in the WFO Cheyenne CWA in 2005
and 2006

• 26 to 75 nmi
• Hail Size Events HDA vs. Hail Size
  (inches)
• gt1.00 3 .0133
• .88 to 1.00 20 .4105
• .75 13 .3077
• Hail Size Events HDA Larger HDA
  Smaller
• gt1.00 3 1 time
  1 time
• .88 to 1.00 20 16 times
  1 time
• .75 13 11 times
  0 times

18
HDA Hail Size Compared to Hail Size (.75
orlarger) Based on Distance from the KCYS
88D185 events in the WFO Cheyenne CWA in 2005
and 2006

• 26 nmi or less
• Hail Size Events HDA vs. Hail Size
  (inches)
• gt1.00 4 .4375
• .88 to 1.00 11 .4173
• .75 11 .3863
• Hail Size Events HDA Larger HDA
  Smaller
• gt1.00 4 3 times
  0 times
• .88 to 1.00 11 10 times
  0 times
• .75 11 10 times
  0 times

19
HDA Hail Size Compared to Hail Size (.75
orlarger) in Relation to VIL (gt55 dBz) 108
events in the WFO Cheyenne CWA in 2005 and 2006

• Hail Size Events HDA vs. Hail Size
• gt1.00 10 .5250
• .88 to 1.00 13 .8915
• .75 17 .8529
• Hail Size Events HDA Larger HDA
  Smaller
• gt1.00 10 8 times
  1 time
• .88 to 1.00 13 13 times
  0 times
• .75 17 17 times
  0 times

20
HDA Hail Size Compared to Hail Size (.75
orlarger) in Relation to VIL (lt55 dBz) 108
events in the WFO Cheyenne CWA in 2005 and 2006

• Hail Size Events HDA vs. Hail Size
• gt1.00 8 -.4375
• .88 to 1.00 41 .2351
• .75 19 .3553
• Hail Size Events HDA Larger HDA
  Smaller
• gt1.00 8 3 times
  5 times
• .88 to 1.00 41 27 times
  8 times
• .75 19 15 times 3
  times

21
Conclusion

• This study, using one radar over a 2 year period,
  would suggest that in most cases, the HDA does
  indeed overestimate the actual hail size by
  between one-quarter and three-quarters of an
  inch.
• This result is in agreement with past studies of
  the HDA. Though the overestimation seems to be
  slightly greater in the CYS CWA.

22
Ways to possibly improve the HDA

• Use storm top divergence to get the severe hail
  size. AT WFO Cheyenne, we have found that
  50 to 75 kts or more works best.
• Look at storm structure (BWERs, shear etc.) to
  help with hail detection.
• Look at the Mesocyclone Intensity

23
Changes Planned in the future for the HDA

• Within 3-5 years dual polarization techniques
  will be added to the HDA.

24

• Questions?
• Thank you.