A tornado touched down just north of Fair Play, South Carolina, on 8 April 2010, partially removing the roof of this home. Image by Tony Sturey, NWS.

Author's Note: The following report has not been subjected to the scientific peer review process.

1.  Introduction

During the afternoon and evening of Thursday, 8 April 2010, severe 
thunderstorms developed ahead of a cold front over north Georgia and 
moved across Upstate South Carolina and the western Piedmont of North 
Carolina.  The strongest of the storms were part of a short, quasi-
linear convective system (QLCS, or squall line).  The QLCS produced 
two tornadoes in Oconee County and Anderson County, South Carolina.  
The first tornado occurred near Fair Play, South Carolina, at about
501 pm EDT (Eastern Daylight Time), or 2101 UTC (Universal Time 
Coordinated).  Part of a roof was blown off a manufactured home and 
numerous trees were snapped or uprooted.  The damage was rated at EF0 
intensity on the Enhanced Fujita Scale.  At about 526 pm EDT (2126 UTC), 
another tornado touched down near Pendleton, South Carolina.  The 
second tornado was stronger than the first and produced a path of 
damage about two miles long.  Several barns and horse stables sustained 
roof damage, several out-buildings were destroyed, and numerous trees 
were snapped or uprooted.  The damage from the second tornado was rated 
at EF1 intensity.  The same system produced damaging wind gusts over 
northern Anderson County and west central Greenville County, South 
Carolina, before it weakened.  Other reports of damage were received 
from Cleveland, Lincoln, Iredell, and Davie counties in North Carolina.

[Note:  All times in this report from this point forward are expressed 
in Universal Time Coordinated (UTC), which is EDT plus four hours.  To
convert from UTC to EDT, subtract four hours from the UTC time.]

Click here to view a list of severe weather reports for 8 April 2010.

Figure 1.  Preliminary reports of tornadoes, wind damage, and hail 
received during the 24 hours ending at 1200 UTC on 8 April 2010.  Click 
on image to enlarge.

The Fair Play tornado appeared to be associated with a break in the
QLCS similar to what has been observed with tornadic QLCS events in
the western Carolinas, such as the Bessemer City, North Carolina,
tornado (January 2006) or the Liberty and Moore, South Carolina, 
tornadoes (January 2007).  The studies linked above dealt with QLCS 
tornadoes in a high shear - low CAPE (HSLC) environment in the cool 
season.  The pre-storm environment on 8 April was characterized by 
high shear and storm-relative helicity (SRH), low lifting condensation 
level (LCL), and enough convective available potential energy (CAPE, 
a measure of buoyancy of air parcels) to support the development of 
supercell thunderstorms.  What set the Fair Play storm apart was the 
relatively high CAPE compared to other QLCS tornadoes associated
with line-breaks.  Another interesting low reflectivity feature was
noted wrapping around the leading edge of the QLCS prior to the
development of the tornado.

2.  Synoptic Features

For several days leading up to the event, forecasters at the National
Weather Service Greenville-Spartanburg (GSP) office expected conditions 
to be favorable for at least isolated severe thunderstorms to the east 
of the Appalachians on 8 April.  The Day 2 Convective Outlook issued by 
the Storm Prediction Center (SPC) at 1730 UTC on 7 April featured a 
Slight Risk of severe thunderstorms over the Piedmont of the Carolinas.  
The Slight Risk was expanded to cover even more of the western Carolinas 
when the Day 1 Convective Outlook was issued at 0547 UTC on 8 April.

At 1200 UTC on 8 April, the 500 mb analysis featured a deep trough over 
the middle of the country with a low center over Wisconsin (Fig. 2).  A 
deep southwest flow aloft was spread over the Southeast and Carolinas 
between the upper ridge over the western Atlantic and the upper trough 
axis that extended from the upper low south along the Mississippi River 
to the northwest Gulf of Mexico.  A short wave trough was located over 
the Arklatex region with a belt of stronger winds in excess of 60 kt 
ahead of the short wave from east Texas to the Ohio Valley.  The eastward
movement of the short wave was expected to help spread the stronger
winds over the Carolinas later in the day.

Figure 2.  Objective analysis of 500 mb geopotential height (black 
contours), temperature (red dashed contours), and wind barbs at 1200 UTC 
on 8 April 2010.  Click on image to enlarge.

The stronger winds were also apparent on the 700 mb analysis (Fig. 3) 
from central Mississippi to middle Tennessee and southwest Ohio.  At
850 mb (Fig. 4), a cold front aloft was suggested from middle Tennessee 
to southern Louisiana by the wind at Birmingham, Alabama, which was 
backed relative to the southwest flow elsewhere across the Southeast.
This backing of the wind at 850 mb was expected to increase the potential
for rotating updrafts over the Piedmont as the cold front aloft 
approached from the west later in the day.

Figure 3.  Objective analysis of 700 mb geopotential height (black 
contours), temperature (red and blue dashed contours), dewpoint (green 
contours), and wind barbs at 1200 UTC on 8 April 2010.  Click on image 
to enlarge.

Figure 4.  Objective analysis of 850 mb geopotential height (black 
contours), temperature (red and blue dashed contours), dewpoint (green 
contours), and wind barbs at 1200 UTC on 28 March 2010.  Click on image 
to enlarge.

At the surface, a cold front stretched from middle Tennessee to southeast
Louisiana (Fig. 5).  The convective mode was expected to be linear later 
in the day because of the mainly unidirectional winds aloft and the deep 
layer shear vector nearly parallel to the cold front, as outlined in the 
Day 1 Convective Outlook issued by the SPC at 1257 UTC.  A broken line of 
convection across eastern Tennessee and northwest Georgia (Fig. 6), with 
a trailing region of stratiform rain ahead of the cold front, was expected 
to gradually weaken and dissipate during the morning hours.  In fact, the 
convection weakened through the early afternoon as expected, and left
considerable cloudiness across the western Carolinas that slowed the 
destabilization of the air mass east of the mountains.  However, severe 
thunderstorms, remained a possibility as the low level jet moved across 
the Carolinas during the time of peak heating.  The Slight Risk was 
unchanged on the Day 1 Convective Outlook updated at 1750 UTC.

Figure 5.  HPC Surface fronts and pressure analysis at 1200 UTC on 
8 April 2010.  Click on image to enlarge.

Figure 6.  Composite reflectivity mosaic centered on GSP at 1300 UTC on  
8 April 2010.  Click on image to enlarge.

At 1800 UTC, the cold front reached from east Tennessee and across the 
northwest tip of Georgia to eastern Alabama (Fig. 7).  Water vapor 
satellite imagery showed the leading edge of dry air at mid levels moving 
across Alabama (Fig. 8), indicative of stronger winds and subsidence
behind the cold front.  Between 1800 UTC and 1900 UTC, a new line of 
convection developed near Atlanta, Georgia, which continued to strengthen
as it moved across north Georgia (Fig. 9).

Figure 7.  HPC surface fronts and pressure analysis central east sector 
at 1800 UTC on 8 April 2010.  Click on image to enlarge.

Figure 8.  GOES-12 water vapor satellite imagery at 1745 UTC on 8 April 
2010.  Brightness temperatures are given by the color scale on the
bottom of the image, with red and black corresponding to drier air at
mid levels.  Click on image to enlarge.

Figure 9.  As in Fig. 6, except at 1957 UTC on 8 April 2010.  Click on 
image to enlarge.

Click here to view a 17 frame Java loop of GOES-12 water vapor channel 
satellite imagery from 1145 UTC on 8 April to 0345 UTC on 9 April.

Click here to view a 10 frame Java loop of a regional reflectivity mosaic 
from 1158 UTC to 2100 UTC on 8 April.

3.  Pre-Storm Environment

The environment across northeast Georgia and western South Carolina 
ahead of the convection moving across north Georgia at 2000 UTC was 
characterized by high shear and a modest amount of buoyancy.  The 
shear in the surface to 1 km layer was at least 40 kt across most of 
Upstate South Carolina (Fig. 10a).  Shear values of 40 kt or greater 
in the layer from the surface to 2.5 km are thought to be sufficient 
to the production of low-level mesovortices in a QLCS as suggested by 
the modeling studies of Weisman and Trapp (2003).  The lowest 100 mb 
mean lifting condensation level (LCL) was on the order of 500 to 
750 m (Fig. 10b).  The combination of a relatively high 0-1 km shear 
of 40 kt and relatively low LCL of 750 m suggested a relatively high 
probability that a tornado would occur (Brooks and Craven, 2002).  
The effective storm relative helicity (SRH) was on the order of 
200–300 m2 s-2 (Fig. 10c), which was indicative of a high potential for 
tornadic supercells.  Values of CAPE for the most unstable air parcels 
were on the order of 500–1000 J kg-1 (Fig. 10d).  Although the CAPE was 
modest, the high values of effective SRH and effective bulk wind shear 
(Fig. 10e) yielded a supercell composite parameter value between 2 and 
4 (Fig. 10f), which was indicative of a high potential for right-moving 
supercells.  

Figure 10.  SPC objective mesoanalysis at 2000 UTC on 8 April 2010 
showing (a) surface - 1 km shear (kt; blue contours) and shear vector
(kt; barbs), (b) 100 mb mean parcel LCL height (m AGL; green and brown
contours, color fill above 2000 m), (c) effective SRH (m2s-2; blue 
contours), effective inflow base (m AGL; color fill), and storm motion
(kt; barbs), (d) most unstable CAPE (J kg-1; brown contours) and lifted
parcel level (m AGL, black contours and color fill), (e) effective bulk
shear (kt; blue contours and color fill), and (f) supercell composite
parameter for the effective layer (blue contours) and Bunkers storm 
motion (kt; barbs).  The area of interest, Oconee County (O), Pickens
County (P), and Anderson County (A), South Carolina, is shaded light
blue.  Click on each image to enlarge.

The less-than-optimal thermodynamic environment was thought to be a 
limiting factor, but the favorable shear would support isolated 
tornadoes and wind damage, as stated by the SPC in a mesoscale 
discussion issued at 2000 UTC.  The continued development of a line 
of thunderstorms toward more favorable buoyancy motivated the SPC to 
issue a Tornado Watch (#0066) for part of northeast Georgia and all 
of Upstate South Carolina at 2031 UTC.

4.  Radar Observations

Between 2000 UTC and 2030 UTC, the convection moving across north 
Georgia had consolidated into a short QLCS from western Oconee County, 
South Carolina, to Barrow County, Georgia.  A bowing segment of the 
line was located over Franklin County, Georgia.  A comparison of 0.5 
degree scans from the KGSP radar between 1957 UTC and 2031 UTC showed 
evidence of a rear inflow jet impinging upon the upshear side of the 
QLCS immediately behind the bowing segment (Fig. 11).  A relative 
desiccation of reflectivity was noted behind the bowing segment, which 
could have been the result of drier air descending from mid-levels 
behind the QLCS.  The base velocity data showed an area of stronger 
inbound echoes immediately behind the bow in western Franklin County.

Figure 11. KGSP radar 0.5 degree (a) base reflectivity and (b) base 
velocity at 1958 UTC, and (c) base reflectivity and (d) base velocity 
at 2031 UTC.  The reflectivity color table is shown on the left bar and 
the velocity color table is shown on the right color bar.  Note the hole 
of no reflectivity over Homer in (c).  The diamond shape in (c) and 
(d) denotes the location of a radar-detected hail core.  Note the region 
of higher velocity in the darker green color to the west of the diamond 
in (d) corresponding to the concave reflectivity gradient in (c).  Click 
on image to enlarge.

An interesting feature developed on the southeast flank of the bowing 
segment of the QLCS between 2023 UTC and 2036 UTC (Fig. 12).  Within a 
larger area of stratiform-like precipitation preceding the QLCS, a patch 
of low reflectivity appeared over southern Franklin County by 2027 UTC.  
There was some evidence of the low reflectivity area first on the 
1.8 degree scan at 2019 UTC, which suggested downward development of 
this feature.  The low reflectivity area elongated north to south and 
moved into the western part of Hart County, Georgia, by 2036 UTC, and 
then wrapped around the eastern flank of the QLCS near Fair Play by 
2044 UTC.  Meanwhile, an inflection point developed along the QLCS 
between 2031 UTC and 2044 UTC, and was located between Lavonia and Lake 
Hartwell at 2044 UTC.

Figure 12.  KGSP 0.5 degree base reflectivity at (a) 2023 UTC, 
(b) 2027 UTC, (c) 2031 UTC, (d) 2036 UTC, (e) 2040 UTC, and (f) 2044 UTC.  
The reflectivity values are given by the color table in the upper left 
corner of each image.  The yellow arrow in (c) shows the location of the 
weak echo channel and the length of the arrow is 10 miles.  Click on 
image to enlarge.

Between 2040 UTC and 2052 UTC, the bowing segment of the QLCS surged 
forward at approximately 60 kt from northeast Franklin County, across 
extreme southern Oconee County, to Townville, South Carolina (Fig. 13 
and 14).  A break in the line occurred on both the 1.3 degree and 
1.8 degree scans by 2052 UTC, which corresponded to a range of elevation 
from 7,000 ft to 13,000 ft AGL in the storm.  No break was discernable 
on the lowest two elevation cuts, although a forward bulge was evident 
to the northeast of Fair Play.

Figure 13.  KGSP base reflectivity at 2040 UTC on the (a) 0.5 degree, 
(b) 0.9 degree, (c) 1.3 degree, and (d) 1.8 degree scans.  The reflectivity 
color table is given by the table in the upper left corner of each image.  
Click on image to enlarge.

Figure 14.  As in Fig. 13, except for 2052 UTC.  Click on image to enlarge.

Although broad rotation was associated with the bowing segment of the 
QLCS, maximum rotational velocity and shear were considered insignificant 
on the lowest three elevation scans from the KGSP radar through 2048 UTC 
(Fig. 15), as both measures of mesocyclone strength remained in the "weak" 
or "minimal" categories.  A small jump was noted in rotational velocity 
at all three levels at 2052 UTC, but this only brought the rotation into 
the "minimal mesocyclone" range, where it remained during the 2057 UTC 
volume scan.  The shear remained minimal but with an increasing trend 
through 2057 UTC as the distance between regions of inbound and outbound 
storm relative motion decreased.  On the 2101 UTC scan, the maximum 
rotational velocity showed another small increase, but the maximum 
rotational shear almost doubled to 0.0201 s-1, which was at the threshold 
between the "tornado possible" and the "tornado probable" regions of the 
Rotational Shear Nomogram.  The tornado touched down 0.5 miles north-
northwest of Fair Play at this time.  Prior to the tornado, there was no 
conclusive evidence to suggest a mesocyclone developing downward from 
mid-levels.  The rotational shear dropped back to pre-tornado values at 
2105 UTC but rotational velocity on the 0.5 degree scan maximized at 31 kt, 
which placed the rotation at the threshold of the minimal and moderate 
categories on the mesocyclone nomogram.  By this time, the tornado had 
already lifted.

Figure 15.  Maximum rotational velocity (a) and rotational shear (b) observed 
on the lowest three elevation scans from the KGSP radar from 2031 UTC to 
2151 UTC.  The velocity and shear data are organized according to the Fair 
Play storm (solid lines on the left) or the Pendleton storm (dashed lines 
on the right).  The pink colored vertical bars represent the time of the 
Fair Play tornado (2101 UTC) and the Pendleton tornado (2126 UTC to 
2129 UTC).  Click on image to enlarge.

A large coherent area of outbound storm relative motion was present by 
2052 UTC over southwest Oconee County, but the inbound portion of the 
circulation was not as well organized (Fig. 16).  Most of the inbound 
targets were along and ahead of the southeast flank of the QLCS with a 
smaller area east of Lavonia, Georgia.  By 2057 UTC, the circulation 
had the appearance of tightening because of a patch of inbound targets 
drawing closer to the outbound region.  At 2101 UTC, when the tornado 
was on the ground, a very small inbound area (two range gates) was 
adjacent to the larger outbound portion of the rotational signature.  
The maximum outbound storm relative motion increased 66% between 2057 UTC 
and 2105 UTC (from 26 kt to 43 kt) over southern Oconee County on the 
0.5 degree scan.  Reflectivity images gave only subtle clues prior to 
the tornado, mainly limited to a weak echo channel oriented along 
Interstate 85 in southern Oconee County punching into the rear flank of 
the QLCS (Fig. 17).  The break in the QLCS was not apparent on the 
0.5 degree elevation scan until 2105 UTC, by which time the tornado had 
already lifted.

Figure 16.  KGSP storm relative motion on the 0.5 degree scan at 
(a) 2052 UTC, (b) 2057 UTC, (c) 2101 UTC, and (d) 2105 UTC.  The color 
table in the upper left corner of each image shows the magnitude of the 
component of motion toward or away from the radar along the radial.  In 
general, green shades represent motion toward the radar and red shades 
show motion away from the radar.  The dashed purple line is the border 
between South Carolina and Georgia.  The blue line is Interstate 85.  
County boundaries are shown in grey.  The location labeled “Home” 
denotes the approximate location  where the Fair Play tornado touched 
down at 2105 UTC.  Click on image to enlarge.

Figure 17.  KGSP base reflectivity on the 0.5 degree scan at (a) 2052 UTC, 
(b) 2057 UTC, (c) 2101 UTC, and (d) 2105 UTC.  The color table in the 
upper left corner of each image gives the reflectivity value.  The dashed 
purple line is the border between South Carolina and Georgia.  The blue 
line is Interstate 85.  County boundaries are shown in grey.  The location 
marked “Home” is the approximate location where the Fair Play tornado 
touched down at 2101 UTC.  Click on image to enlarge.

Reflectivity data gave no meaningful clues as to the development of the 
second tornado.  The apparent break in the QLCS filled in by 2114 UTC.  
Broad rotation continued across the QLCS in the vicinity of Townville, 
South Carolina, at 2109 UTC and 2113 UTC (Fig. 18).  The rotation became 
better organized between 2113 UTC and 2122 UTC as the couplet moved 
across the area between Pendleton and Sandy Springs.  Maximum rotational 
velocity increased steadily and the distance between maximum inbound and 
outbound decreased (Fig. 15).  Rotational shear peaked at 0.037 s-1 (in 
the "tornado likely" category) on the 0.9 degree scan at 2126 UTC as the 
gate-to-gate difference between inbound and outbound storm relative motion 
peaked.  The Pendleton tornado touched down at this time.  The tornado 
remained on the ground through approximately 2129 UTC, after which 
rotational velocity and shear diminished.

Figure 18.  KGSP storm relative motion on the 0.5 degree scan at 
(a) 2109 UTC, (b) 2113 UTC, (c) 2118 UTC, (d) 2122 UTC, (e) 2126 UTC, 
and (f) 2130 UTC.  The location marked "Home" is the location where the 
Pendleton tornado touched down at approximately 2126 UTC.  The polygon 
boundary for Tornado Warning #0015 is shown as the broken light blue line.  
County boundaries are shown in light gray.  Click on image to enlarge.

Click here to view a 25 frame Java loop of KGSP 0.5 degree base reflectivity 
from 2002 UTC to 2143 UTC on 8 April.

Click here to view a 25 frame Java loop of KGSP 0.5 degree storm relative 
motion from 2002 UTC to 2143 UTC on 8 April.

5.  Summary 

Clues that a tornado was imminent were subtle up until the radar volume 
scan during which the Fair Play tornado touched down.  Evidence suggested 
the Fair Play tornado was preceded by a rear inflow jet which caused the 
QLCS to bow forward and a weak echo channel that wrapped around an 
inflection point on the convective line.  The tornado was accompanied by 
a break in the QLCS whereby the tornado formed at the southern end of the 
trailing line segment.  Similar radar signatures have been associated with 
QLCS tornadoes in HSLC environments (Lee and Jones 1998; McAvoy et al. 2000; 
Lane and Moore 2006).  However, the break in the line was not evident on 
the 0.5 degree elevation scan until the volume scan after the tornado had 
already lifted.  The maximum rotational shear did not reach the threshold 
for tornado warning issuance (greater than 0.020 s-1) until the tornado was 
already on the ground.

The environment on 8 April contained more CAPE and more SRH than what was 
present in the "broken-S" QLCS tornado events documented in the studies
above.  In fact, the environment on 8 April appeared more supportive of a 
QLCS with embedded supercellular elements and perhaps discrete right-moving 
supercells. One of the few tools the warning forecaster has for the issuance 
of a successful tornado warning in a "broken-S" situation is an awareness 
of the HSLC environment that usually accompanies such events.  Situational 
awareness among the warning team of the possibility of a "broken-S" signature 
was relatively low because the forecast area was not in the typical HSLC 
environment.  A QLCS with embedded bows and supercells was the expected mode 
of convection.

An interesting reflectivity feature developed on the southeast flank of 
the QLCS approximately 35 minutes before the Fair Play tornado.  A patch 
of lower reflectivity appeared to descend and elongate in the direction 
of the wind in the stratiform rain area ahead of the QLCS.  The lower 
reflectivity patch proceeded to wrap around the eastern flank of the QLCS 
about 10 – 15 minutes before tornadogenesis.  This feature had not been 
noticed in similar QLCS tornado events in the past.  The QLCS on 8 April 
had characteristics similar to the "leading stratiform" mode of mesoscale 
convective system outlined by Parker and Johnson (2000).  It would be 
interesting to re-investigate previous "broken-S" events in light of 
Parker and Johnson (2000) to determine the organizational mode of the 
QLCS and see if any were of the leading stratiform variety and if a 
similar feature was present.

Pictures of the damage from the Pendleton Tornado

Pictures of the damage observed from the Pendleton tornado on 8 April 2010.  
Click on image to enlarge.

References

Brooks, H. E., and J. P. Craven, 2002:  A database of proximity soundings 
     for significant severe weather.  Preprints, 21st Conf. on Severe 
     Local Storms, San Antonio, TX, Amer. Meteor. Soc.


Lane, J. D., and P. D. Moore, 2006:  Observations of a non-supercell 
     tornadic thunderstorm from a Terminal Doppler Weather Radar.  
     Preprints, 23rd Conf. on Severe Local Storms, St. Louis, MO, Amer. 
     Meteor. Soc.


Lee, L. G., and W. A. Jones, 1998:  Characteristics of WSR-88D velocity 
     and reflectivity patterns associated with a cold season non-supercell 
     tornado in upstate South Carolina.  Preprints, 19th Conf. on Severe 
     Local Storms, Minneapolis, MN, Amer. Meteor. Soc., 151-154.


McAvoy, B. P., W. A. Jones, and P. D. Moore, 2000:  Investigation of an 
     unusual storm structure associated with weak to occasionally strong 
     tornadoes over the Eastern United States.  Preprints, 20th Conf. on 
     Severe Local Storms, Orlando, FL, Amer. Meteor. Soc., 182-185. 


Parker, M. D, and R. H. Johnson, 2000:  Organizational modes of midlatitude 
     mesoscale convective systems.  Mon. Wea. Rev., 128, 3413-3436.


Weisman, M. L., and R. J. Trapp, 2003:  Low-level mesovortices with squall 
     lines and bow echoes.  Part I:  Overview and dependence on environmental 
     shear.  Mon. Wea. Rev., 131, 2779-2803.

Acknowledgements

The damage survey for the Fair Play and Pendleton tornadoes was
conducted by Tony Sturey and Justin Lane (NWS).  The local storm 
report map and upper air analyses were obtained from the Storm 
Prediction Center.  The mesoscale objective analysis was prepared 
by the SPC and obtained from the archive at the NWS office in Omaha, 
Nebraska.  The surface fronts and pressure analyses were obtained 
from the Hydrometeorological Prediction Center.  The satellite 
imagery and radar mosaic imagery was obtained from the University 
Corporation for Atmospheric Research.  Some of the radar graphics 
were created using the GRLevel2 Analyst software.  The plots of
maximum rotational velocity and shear were created using Microsoft 
Excel.  The tornado track map was made using Delorme Street Atlas 
USA 2009 Plus.