Hail to the size of golf balls litters the ground near Belton, South 
Carolina, on Friday evening, 21 April 2006.  Image courtesy of David 
Rogers, Anderson Independent-Mail.  Used by permission.

1.  Introduction

Several severe thunderstorms developed across northeast Georgia and
the upstate of South Carolina on Friday evening, 21 April 2006.  The 
thunderstorm activity continued into the early morning hours of Saturday,
22 April.  In all, the National Weather Service (NWS) Weather Forecast 
Office (WFO) in Greer, South Carolina, located at the Greenville - 
Spartanburg Airport (GSP), issued 26 Severe Thunderstorm Warnings and
2 tornado warnings during the period from 8 am Friday to 8 am Saturday.  
[Note:  All times in this document from this point on are referred to
in Universal Time Coordinated (UTC), which is Eastern Daylight Time
plus four hours.]  Most of the severe weather produced by the storms 
Friday evening came in the form of large hail, while several reports 
of wind damage were received early Saturday morning (Fig. 1).  The most 
intense severe thunderstorm of the evening produced a swath of large hail 
up to the size of golf balls across the area from Belton, South Carolina, 
to the southwest corner of Greenville County near Fork Shoals.

Click here to see a list of all Local Storm Reports for this event.

Figure 1.  Severe weather reports for the 24-hour period ending 1207 UTC
22 April 2006.  The map may not reflect all the storm reports received at 
WFO GSP post-event.

The Belton hail storm was interesting from a radar perspective in that
it displayed many of the classic characteristics of a large-hail-producing
severe thunderstorm, as we shall see in later sections.

2.  Synoptic Pattern and Pre-Storm Environment

The upper air analysis from Friday morning indicated the environment 
would be favorable for deep convection later in the day across northeast 
Georgia and Upstate South Carolina.  The 500 mb objective analysis from 
the Storm Prediction Center (SPC) indicated a pool of slightly cooler 
temperatures aloft just upstream over the Mississippi Delta region along 
with a mid-level jet streak (Fig. 2).  A weak short wave seen as the 
inflection in the 576 dm height contour over Arkansas and Louisiana was 
expected to lift northeast across Tennessee later in the day with the 
upper ridge holding over the Carolinas.  The 300 mb analysis showed a 
region of upper diffluence moving over the western Carolinas.  Meanwhile, 
the 850 mb analysis showed a weak thermal ridge immediately upstream with 
moisture transport implied by a southerly flow from the Gulf of Mexico.  
The surface analysis at 1200 UTC from the Hydrometeorological Prediction 
Center (HPC) showed a cold front extending from low pressure over northern
Wisconsin, through the mid-Mississippi Valley, to the Ozarks (Fig. 3).

Figure 2.  SPC objective analysis of 500 mb geopotential height, 
temperature, and wind for 1200 UTC 21 April 2006.  Click on image to enlarge.

Figure 3.  HPC surface pressure and fronts analysis for 1200 UTC 21 April.  
Click on image to enlarge.

The upper air sounding observed at Peachtree City, Georgia (FFC), at 
1200 UTC on 21 April (Fig. 4) gave further support to the idea that 
severe thunderstorms were possible later in the day over northeast 
Georgia and the western part of South Carolina.  The sounding showed a 
steep lapse rate in the 700-500 mb layer and a surface-based Convective 
Available Potential Energy (CAPE) of greater than 2100 J/kg modified for 
expected conditions in the afternoon.  Wind shear was expected to be 
sufficient to support the development of a few supercell thunderstorms.  
However, the environment farther to the north and east was not expected
to be as favorable, as seen in the Greensboro, North Carolina, sounding.
The favorable environment for severe thunderstorms was reflected in the 
Day 1 Convective Outlook, updated at 2000 UTC, which included northeast 
Georgia and the western tip of South Carolina in a Slight Risk.

Figure 4.  Skew-T log P diagram of the upper air sounding (left) and 
hodograph (right) for FFC at 1200 UTC 21 April.  The table at the bottom 
shows many of the severe weather indices used by the SPC.  Click on image 
to enlarge.

What transpired during the day was a general lack of deep convection 
across most of the western Carolinas.  Extensive debris cloudiness left 
over from widespread convection in the morning over the Great Valley of 
east Tennessee effectively cut down on solar heating across western North 
Carolina and most of the Upstate of South Carolina.  A severe thunderstorm 
produced half-dollar sized hail near Marble in Cherokee County, North 
Carolina, around 1500 UTC, while a similar storm produced marble to penny 
sized hail over Rabun County, Georgia, around 1530 UTC.  However, both 
storms quickly diminished as they moved northeast into more stable air.  
Middle and high level clouds remained across the western Carolinas through 
at least 1900 UTC, as seen in Aqua MODIS imagery.  The SPC objective 
analysis of CAPE and convective inhibition at 1900 UTC was indicative of 
marginal instability across most of the western Carolinas.  It was not
until after 2100 UTC that some measure of destabilization occurred over
the upper Savannah River Valley.  Surface observations around 2200 UTC 
showed dewpoints pooling in the lower 60s in the area from Anderson to 
Athens and Gainesville, Georgia, with winds backing to the southeast at
Anderson and Greenwood (Fig. 5).  The higher dewpoints combined with 
surface heating allowed by thinning cloud cover resulted in a CAPE above
1000 J/kg and eliminated the remaining convective inhibition (Fig. 6).

Figure 5.  Surface observations plot at 2200 UTC 21 April.  Click on 
image to enlarge.

Figure 6.  SPC objective analysis of surface based CAPE and CINH at 
2200 UTC 21 April 2006.  Click on image to enlarge.

In response to the increasing instability, thunderstorms developed 
quickly over northeast Georgia by 2200 UTC.

3.  Radar Observations of the Belton Storm

The convective cell that would become the severe thunderstorm over 
Anderson County had its genesis over northeast Georgia around 2130 UTC.  
The shower moved northeast over Elbert and Hart counties to a position 
near Hartwell Dam through about 2215 UTC, at which time a new updraft 
formed on its southern flank to the southeast of Hartwell, Georgia.  Over 
the next ten minutes, the updraft quickly developed into a thunderstorm 
over the Hart and Anderson county line below Hartwell Dam, which is also 
located about 45 miles to the southwest of the Weather Surveillance Radar- 
88 Doppler (WSR-88D) at GSP.  At 2236 UTC, the radar reflectivity in the 
core of the thunderstorm, now located about 7 miles west of Iva, South
Carolina, exceeded 65 dBz above the freezing level (Fig. 7).  Radar 
reflectivity values this large are usually associated with large hail 
stones, which are generally more reflective than a similar number of large 
rain drops in a radar sampling volume.  A Severe Thunderstorm Warning was 
issued for Anderson County at 2237 UTC, with the expectation that hail of 
at least the size of a penny size would fall in the area from Starr to 
Anderson and Belton.

Figure 7.  Radar reflectivity at 2.9 degree scan from KGSP radar at 
2236 UTC.  Note the core of greater than 65 dBz west of Iva.  The KGSP 
radar is located off the upper right corner of the image.  Click on 
image to enlarge.

Click here to view a 27 frame java loop of KGSP 0.5 degree base reflectivity.

The severe thunderstorm continued to move northeast at 25 mph across 
the middle of Anderson county over the next 25 minutes.  A new and more 
vigorous updraft developed on the southeast flank of the storm about 
6 miles south of Anderson at 2305 UTC, with a tremendous upward surge 
in radar echo top and reflectivity values in the storm core.  On the 
2309 UTC volume scan, reflectivity surpassed 70 dBz and a three-body 
scatter spike (Zrnic 1987, Lemon 1998) was observed on radar at about 
17,000 feet above ground level, both of which are indicative of very 
large hail (Fig. 8).

Figure 8.  Radar reflectivity at 5.0 degree scan from KGSP radar at 
2309 UTC.  Note the very high reflectivity (greater than 70 dBz) between 
Anderson and Homeland Park and the three-body scatter spike which appears
as weaker reflectivity extending along the radial to the southwest of 
Homeland Park.  The KGSP radar is located off the upper right corner of 
the image.  Click on image to enlarge.

A three-body scatter spike (TBSS) occurs when hydrometeors inside the 
thunderstorm cell, typically large water-coated hail stones when viewed 
with a 10-cm wavelength radar such as the WSR-88D, reflect some of the 
energy within the radar beam down to the ground.  A portion of this 
energy scatters back upward into the thunderstorm, and then scatters a 
third time back toward the radar antenna.  The radar detects the main 
energy scattered back from the hail stones, plus a smaller amount of 
energy which the radar perceives to be at a greater distance because 
of the triple reflection (hail stone-ground-hail stone) that adds 
extra time between the initial radar pulse and the subsequent return 
of energy.  Thus, the TBSS appears as a “flare echo” (Wilson and Reum 
1986, 1988) of weaker reflectivity at mid-levels of the storm on the 
back side of the reflectivity core when viewed along a radial (Fig. 9).  
A TBSS is typically seen down range from a high reflectivity core (65 dBz 
and greater) in the presence of large hail stones (2.5 cm diameter or 
greater) between 10 and 30 minutes before the hail reaches the ground 
(Lemon 1998).

Figure 9.  Schematic diagram showing the path of the radar beam during 
three-body scattering by large hail stones in a thunderstorm core.  
Adapted from Wilson and Reum (1988).

The thunderstorm continued to intensify as it moved across the area 
between Anderson and Belton through 2325 UTC, perhaps as the result
of enhanced low level convergence and moisture as suggested by the 
Mesoscale Discussion issued by the Storm Prediction Center.  At mid-
levels of the storm, a mesocyclone was evident, which is typical for 
a severe thunderstorm with a very strong updraft.  With the original 
Severe Thunderstorm Warning set to expire at 2330 UTC, the warning was 
re-issued and upgraded to a Tornado Warning for Anderson County at 
2327 UTC, due to the presence of rotation developing on the lowest 
scan of the KGSP radar.  A few minutes later, at 2330 UTC, a Severe 
Thunderstorm Warning was issued for southern Greenville County based 
on the movement of the storm to the northeast.

Figure 10.  Base velocity at 0.4 degree scan from KGSP radar at 
2331 UTC.  Note the couplet of inbound targets (green) and outbound 
targets (red) near Belton, indicative of cyclonic rotation in the
storm cloud.  Click on image to enlarge.

Large hail probably began falling at the ground a few miles west of Belton 
between 2327 and 2331 UTC, judging by the high reflectivity (65 dBz) on 
the lowest elevation scan from the KGSP radar (Fig. 11).  Between 2331 
and 2335 UTC, the severe thunderstorm reached its peak intensity as it 
moved over the north side of Belton.  Radar scans at mid levels of the 
storm showed a pronounced TBSS, weak echo region, and mesocyclone.  
Numerous reports of large hail were received from the Belton area as the 
storm passed, with sizes ranging from penny to golf ball (Fig. 12).

Click here to view a 14 frame java loop of reflectivity from the 2331 UTC 
volume scan from the KGSP WSR-88D radar.

Figure 11.  Radar reflectivity at 0.4 degree scan from KGSP radar at 
2331 UTC.  Note the high reflectivity (greater than 65 dBz) on the west 
side of Belton.  Large hail is most likely reaching the ground at this 
time.  Click on image to enlarge.

Figure 12.  Large hail typically falls in a range of sizes.  Note the 
large stone near the bottom of the railing which appears to be approximately 
1.5 inches wide, compared to the brick step.  When reporting hail to the 
National Weather Service, always remember to include an estimate of the 
largest stones.  Image courtesy of Cynthia Ownbey.

The severe thunderstorm crossed the Saluda River into southern Greenville 
County between 2340 and 2345 UTC in a somewhat weakened state.  Hail up
to the size of quarters continued to fall across the area from Belton to
Fork Shoals.  Although strong mid-level rotation continued, low-level 
rotation remained weak.  An apparent funnel cloud was sighted along 
Dunklin Bridge Road in southern Greenville County (Fig. 13), but no 
tornadoes were actually sighted nor were any reports of wind damage 
received from either Anderson or Greenville counties.

Figure 13.  Funnel cloud sighted off Dunklin Bridge Road in southern
Greenville County, South Carolina.  Quarter-sized hail stopped falling 
moments before.  Image captured by Mike & Christy Countryman.

Across the north side of Belton, hail accumulated to a depth of several 
inches in some spots.  In the aftermath of the severe thunderstorm, 
large hail remained on the ground across the Belton area for an hour or 
more as it gradually melted.  The hail resulted in dangerous driving 
conditions along Highway 20 north of Belton (Fig. 14).  Driving on a
hail covered road is akin to driving on small ball bearings.  Hail fog 
formed when the accumulated hail lowered the temperature near the ground 
to the point where the air became saturated (Fig. 15).  The hail fog 
locally reduced the visibility, making driving conditions even more 
dangerous.

Figure 14.  A significant accumulation of smaller sized hail stones resembles 
a snowfall and can be just as difficult to drive through.  This was the 
case along State Highway 20 between Belton and Cheddar after the hail ended.  
Images courtesy of David Rogers of the Anderson Independent-Mail.

Figure 15.  Significant amounts of hail will also locally affect the 
temperature and visibility.  Large hail remains on the ground an hour 
after it fell on the north side of Belton at intersection of Highway 20 
and Calhoun Road.  Note the poor visibility along Highway 20.  Image 
courtesy of David Rogers of the Anderson Independent-Mail.

Additional severe thunderstorms developed Friday night and early Saturday
morning across the northern part of the Upstate and the Piedmont of North 
Carolina, but none were as intense as the Belton storm.

4.  Summary 

A severe thunderstorm produced a swath of penny to golf ball sized hail 
from just south of Anderson, across Belton, to the southern part of 
Greenville County near Fork Shoals during the early evening of 21 April 
2006.  The initial occurrence of large hail in the Belton area was 
preceded by a Severe Thunderstorm Warning issued by the National Weather 
Service Office in Greer with 50 minutes of lead time.  Radar observations
of the storm showed a distinct three-body scatter spike approximately
16 minutes prior to the occurrence of golf ball sized hail on the north
side of Belton.  Although strong mid-level rotation was also observed, 
a tornado did not occur.

Acknowledgements

Thanks to John Cessarich at WYFF-TV for providing the images captured by 
Cynthia Ownbey and Mike and Christy Countryman.  The Anderson Independent-
Mail provided the contact to David Rogers, who captured the images of the 
hail on the north side of Belton.  The upper air and mesoscale analysis 
graphics were obtained from the Storm Prediction Center.  The surface 
analysis was obtained from the Hydrometeorological Prediction Center.  
Radar images were created using the Java NEXRAD Viewer and Data Exporter, 
obtained from the National Climatic Data Center.  The Aqua MODIS imagery
was obtained from the Space Science and Engineering Center at the 
University of Wisconsin at Madison.

References

Lemon, L. R., 1998: The radar “three-body scatter spike”: An operational 
     large-hail signature. Wea. Forecasting., 13, 327–340.

Wilson, J. W., and D. Reum, 1986: “The hail spike”: Reflectivity and velocity 
     signature. Preprints, 23d Conf. on Radar Meteorology, Snowmass, CO, Amer. 
     Meteor. Soc., 62–65. 

Wilson, J. W., and D. Reum, 1988: The flare echo: Reflectivity and velocity 
     signature. J. Atmos. Oceanic Technol., 5, 197–205. 

Zrnic, D. S., 1987: Three-body scattering produces precipitation signature of 
     special diagnostic value. Radio Sci., 22, 76–86.