A supercell thunderstorm moves through Abbeville, South Carolina, on Saturday 15 March 2008. Numerous reports of hail and wind damage were received from Abbeville County and a tornado touched down near Due West. Image courtesy of GwdToday.com, photographer unknown.

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

1. Introduction

On 15 March 2008, a series of supercell thunderstorms produced a significant outbreak of severe weather across Upstate South Carolina and northeast Georgia (Fig. 1). Numerous reports of large hail were received, including one report of softball size hail in Anderson County, South Carolina. In addition, three tornadoes were confirmed, the strongest being the EF2 tornado that tracked over southeast Elbert County, Georgia. The other two tornadoes affected Franklin and Hart counties in Georgia, and northeast Abbeville County in South Carolina, and were rated at EF0 intensity. This was an unusual event for the region, as the atmospheric ingredients necessary for supercell thunderstorms rarely coexist across northeast Georgia and the western Carolinas. Nevertheless, the severe weather episode was well-forecast several days in advance by the National Weather Service (NWS) Weather Forecast Office (WFO) at Greenville-Spartanburg (GSP) and by the NWS Storm Prediction Center (SPC).

Click here for a list of storm reports from this event

Figure 1. Preliminary SPC storm reports for 15 March 2008. Tornado reports are shown in red, large hail reports are shown in green, and damaging wind reports are shown in blue. Black squares represent reports of wind gusts of 65 knots or greater, while black triangles are reports of hail 2 inches in diameter or larger.

Note that all times in this report are referenced to Coordinated Univeral Time (UTC), which is four hours ahead of Eastern Daylight Time (EDT). To convert to EDT, subtract four hours from the UTC time.

2. Background

The possibility of a significant outbreak of severe weather occurring on 15 March 2008 within the GSP County Warning and Forecast Area (CWFA) was first identified by GSP forecasters on 11 March. The Area Forecast Discussion issued by GSP at 1800 UTC mentioned the resemblance of forecast soundings from the National Centers for Environmental Prediction?s (NCEP) Global Forecasting System (GFS) model to the "loaded gun" sounding (Fig. 2) often associated with Great Plains supercell outbreaks.

Figure 2. An example of a "loaded gun" sounding.

One of the important characteristics of this sounding that contributes to its strong potential instability is the layer of steep, nearly dry-adiabatic lapse rates that exists above the moist boundary layer. This layer is referred to as the "elevated mixed layer." Originating from the warm and dry high-elevation regions of northern Mexico, strong southwest flow in the 5,000 to 10,000 feet layer often advects this airmass north and east over the southern Great Plains and lower Mississippi Valley during Spring. As this warm, dry air overspreads relatively cool, moist maritime air originating from the Gulf of Mexico, a strong inversion becomes established at the top of the marine layer. This inversion inhibits or "caps" convective development. Due to the high degree of potential instability in the sounding, removal of the cap often results in explosive convective development, hence the term "loaded gun." It is quite rare for the elevated mixed layer to advect much further east than the lower Mississippi Valley (Lanicci and Warner, 1991). This is due to the fact that convection often significantly modifies (moistens and warms) this layer before it reaches Georgia and the Carolinas. This results in significant weakening of mid-level lapse rates, and therefore weaker potential instability.

Subsequent model runs of the GFS continued to indicate the potential for an outbreak of severe weather on 15 March, increasing forecasters? confidence that a significant event would occur. The Hazardous Weather Outlook (HWO) issued by GSP at 0925 UTC on 12 March specified that the Piedmont and Foothills of the western Carolinas and northeast Georgia were the most likely areas to receive severe weather on Saturday, 15 March.

The SPC Day 3 Convective Outlook, issued at 0722 UTC on 13 March (Fig. 3) featured a slight risk for severe weather across much of the Southeast. Meanwhile, GSP continued to highlight the potential for a severe weather outbreak in the HWO issued at 0936 UTC on 13 March and specifically mentioned areas south of Interstate 85 as being the highest threat area. With confidence increasing that a significant event was likely, GSP began making plans to augment staffing for 15 March.

Figure 3. SPC graphical Day 3 Convective Outlook issued 0722 UTC 13 March 2008.

The SPC Day 2 convective outlook issued at 0541 UTC on 14 March (Fig. 4a) continued to indicate a slight chance for severe weather across much of the Southeast. The probabilistic Day 2 outlook (Fig. 4b) featured an area of 30% or greater probability of severe weather that covered the southern half of the GSP CWFA. The updated probabilistic Day 2 outlook, issued at 1729 UTC on 14 March (Fig. 4d), featured an area of 10% or greater chance of significant severe weather across much of northern Georgia and the lower Piedmont and northern Midlands of South Carolina.

Figure 4. Categorical Day 2 Convective Outlook (a.) and Probabilistic Outlook (b.) issued by SPC at 0541 UTC 14 March 2008. Categorical Outlook (c.) and Probabilistic Outlook (d.) issued at 1729 UTC on 14 March. The light blue hatched area in (d.) indicates the greater than 10% probability of significant severe weather.

Finally, the Day 1 Convective Outlook issued by SPC at 0559 UTC on 15 March (Fig. 5) indicated a moderate risk for severe weather for much of north Georgia, and the Upstate and Midlands of South Carolina. The probabilistic outlook issued at this time featured a large area of 30% or greater probability of large hail and damaging wind, along with a 15% or greater probability of tornadoes across much of South Carolina, as well as central and north Georgia. The threat for severe weather was considered significant enough for the SPC to issue a Public Severe Weather Outlook at 1004 UTC.

Figure 5.(a.) Categorical Day 1 Convective Outlook issued by SPC at 0559 UTC 15 March 2008. Probabilistic forecasts are shown for (b.) large hail, (c.) tornadoes, and (d.) damaging wind gusts.

3. Synoptic Characteristics

On the 500 hPa analysis at 1200 UTC on 15 March, a low amplitude, fast, quasi-zonal flow was observed across much of the southern half of the country (Fig. 6). A weak short wave trough was moving off the East Coast, while another weak short wave trough was moving out of the Great Plains and into the Mississippi Valley. Short wave ridging was noted between these two features over the Appalachian Mountains.

Figure 6. SPC 500 hPa objective analysis of geopotential height, temperature, and wind barbs at 1200 UTC 15 March.

The 300 mb analysis (Fig. 7) at 1200 UTC on 15 March indicated the upper-level reflection of the short wave trough over the Mississippi Valley, with a strong jet streak (125 - 150 knots) upstream of the trough over the southern Plains. This resulted in strong upper divergence over much of the Mississippi Valley and western Ohio Valley. The rear entrance region of a weaker jet associated with the short wave trough over the east coast produced a second area of upper divergence over Georgia and the Carolinas. This area of forcing was responsible for isolated, marginally severe thunderstorms over the lower Piedmont of South Carolina during the late morning hours.

Figure 7. SPC 300 hPa objective analysis of isotachs, streamlines, and divergence at 1200 UTC 15 March.

The 1200 UTC surface analysis from the Hydrometeorological Prediction Center (HPC)(Fig. 8) indicated an area of low pressure across southern Arkansas, with a quasi-stationary frontal boundary extending from the surface low across Mississippi and Alabama and into north Georgia and the Midlands of South Carolina. The frontal boundary was clearly seen in the observations shown on the regional surface plot at the same time (Fig. 9). An airmass characterized by temperatures and dewpoint temperatures in the 60s was observed along the Coastal Plain. Meanwhile, temperatures and dewpoints were in the 40s across the Piedmont and Foothills of the Carolinas.

Figure 8. HPC analysis of surface fronts and pressure at 1200 UTC 15 March.

Figure 9. Regional surface plot at 1243 UTC 15 March, including surface observations taken at 1150 UTC. Data are plotted according to the traditional station model.

The similarity between the observed sounding from Birmingham, Alabama (BMX), at 1200 UTC on 15 March (Fig. 10) and the sounding in Figure 2 was notable. A very moist layer of air was indicated from the surface to around 850 hPa. This layer was capped by a strong temperature inversion. Very steep lapse rates were observed above the inversion. Although this sounding featured very little instability (106 J/kg), the potential instability in the sounding was quite high, owing to the steep mid-level lapse and the moist boundary layer. Modifying this sounding for a surface temperature of around 70 degrees F, and a dewpoint of around 65 degrees F, produced convective available potential energy (CAPE) in excess of 2000 J/kg, a relatively large CAPE value for the Southeast in March.

Figure 10. Observed sounding from Birmingham, Alabama (BMX), at 1200 UTC 15 March. The red line is the temperature sounding and the green line is the dewpoint sounding. Wind barbs are shown in the column on the right. A hodograph is indicated in the upper left corner. Image created using RAOB v.5.8 for Windows.

The southeast regional radar composite image at 1200 UTC on 15 March indicated two areas of elevated convection north of the frontal boundary over the region (Fig. 11). One was located over Upstate South Carolina, but a larger area was over northern Mississippi. Meanwhile, the capping inversion appeared to be inhibiting convective development over the warm sector. Based on this information, only minor changes were made to the Day 1 Convective Outlook when it was updated at 1227 UTC.

Figure 11. Southeast regional radar image at 1200 UTC on 15 March 2008. Image obtained from the Iowa Environmental Mesonet.

During the middle part of the morning, all signs pointed to the rapid development of severe thunderstorms in the favorable environment ahead of convection moving from north Alabama into north Georgia. As a result, the SPC issued a Tornado Watch for much of north Georgia and the Upstate and Midlands of South Carolina at 1500 UTC (Fig. 12). The moderate risk area was expanded on the updated Day 1 Convective Outlook issued at 1614 UTC and as of late morning, a severe weather outbreak appeared to be likely.

Figure 12. Outline of Tornado Watch number 119 issued by the SPC at 1500 UTC on 15 March 2008.

The absence of convection in the warm sector during the morning of 15 March and resultant lack of overturning allowed the plume of steep mid-level lapse rates sampled by the 1200 UTC BMX sounding to advect over the western Carolinas and northeast Georgia during the afternoon. Increasing south to southeast flow responding to the surface low caused the frontal boundary to move slightly to the north during the afternoon. Meanwhile, dewpoint temperatures in the 60s were being drawn northward to the boundary (Fig. 13). As the upper jet streak continued to dig into the Southeast, increasing upper divergence within the left exit region lifted and cooled the capping inversion, allowing for explosive release of the potential instability over much of South Carolina and north and central Georgia.

Figure 13. As in Fig. 9, except at 1643 UTC.

The 1800 UTC sounding from BMX (Fig. 14) revealed the evolution of the atmosphere during the morning and early afternoon. An inversion remained, but had been lifted to around 700 hPa. Despite the presence of this inversion, the atmosphere was no longer capped, as heating and moistening of the boundary layer had increased instability sufficiently so that an air parcel lifted from the surface experienced positive buoyancy from the lifted condensation level (LCL) through the depth of the troposphere. The sounding yielded almost 1500 J/kg of CAPE. Storm Relative Helicity (SRH) in the sounding was 375 m2/s2 in the 0-3 km layer, and 309 m2/s2 in the 0-1 km layer. These values of CAPE and SRH were well within the range of values found by Lane (2008) to be characteristic of significant tornado environments over the western Carolinas and northeast Georgia.

Figure 14. As in Fig. 10, except at 1800 UTC 15 March.

4. Convective Evolution

By 1700 UTC, scattered supercell thunderstorms developed along the stationary front across northern Alabama and north Georgia. These storms produced widespread severe weather across these areas as they moved along the stationary front toward the Carolinas. GSP forecasters, recognizing the threat that these storms posed to the GSP CWFA, issued a Special Weather Statement (SPS) at 1706 UTC to alert customers and partners of the increasing potential for dangerous thunderstorms. The growing threat was also noted by the SPC at 1722 UTC. The first warning issued by WFO GSP during the afternoon of 15 March was a Tornado Warning at 1818 UTC for portions of Elbert, Hart, and Franklin Counties in Georgia, as a well-organized supercell moved across northeast Georgia. The composite reflectivity from the Greer (KGSP) Weather Service Radar - 1988 Doppler (WSR-88D) at 1840 UTC and 1848 UTC (Fig. 15) showed the initial supercell moving into Franklin County. A sequence of storm relative velocity (SRV) images from KGSP (Fig. 16) indicated a deep, but relatively broad mesocyclone at 1840 UTC. However, by 1848 UTC, a smaller, more intense circulation was observed near the center of the mesocyclonic circulation on the Franklin and Hart county border (Fig. 17). A gate-to-gate velocity couplet of around 80 knots (35 knots outbound + 45 knots inbound) was observed at 0.5 degrees. This velocity differential qualified this circulation as a Tornado Vortex Signature (TVS). However, the radar algorithm did not identify this feature as a TVS until the next volume scan. Incidentally, an EF0 tornado was in progress in association with this rotational signature across southeast Franklin and extreme southwest Hart County (Fig. 18). However, the tornado was near the end of its life cycle when the gate-to-gate shear became evident in radar data. By the time the radar algorithm identified the circulation as a TVS, the tornado had dissipated.

Figure 15. Composite reflectivity from KGSP radar at (a.) 1840 UTC and (b.) 1848 UTC.

Figure 16. Storm relative velocity images from KGSP at 1840 UTC at (a.) 0.5 degrees, (b.) 1.3 degrees, (c.) 2.4 degrees, and (d.) 3.1 degrees. The white arrows in (b.) indicate the motion associated with a couplet of inbound and outbound targets that identify a low level mesocyclone. The speed is given by the color table at the lower right, where negative values represent motion toward the radar and positive values represent motion away from the radar, by convention.

Figure 17. As in Fig. 16, except at 1848 UTC. The white arrow shows the strongest gate-to-gate shear on the 0.5 degree scan.

Figure 18. Track of tornado across parts of Franklin County and Hart County, Georgia, on 15 March 2008. The tornado track is delineated by the heavy black line below Franklin Springs and Royston. The background map is from Delorme Street Atlas USA.

The environment across northeast Georgia remained highly favorable for additional supercell thunderstorms and tornadoes during the early afternoon. In fact, the SPC upgraded the Day 1 Convective Outlook at 1940 UTC to a high risk of severe weather across part of northeast Georgia and the lower Piedmont of South Carolina.

A Tornado Warning was issued for Elbert County, Georgia, at 1954 UTC for another supercell thunderstorm moving out of north central Georgia. A series of reflectivity images from KGSP (Fig. 19) showed the low-level structure of this storm, including an impressive hook echo, as it was producing an EF2 tornado across extreme southern Elbert County. The three-dimensional cross section from KGSP at 2016 UTC depicted in Figure 20 indicated classic supercell structure, including a strongly tilted updraft, bounded weak echo region (BWER), and a high reflectivity core suspended aloft. These features were indicative of a very strong, rotating updraft.

Figure 19. Radar reflectivity from KGSP at 0.5 degrees at (a.) 2016 UTC, (b.) 2021 UTC, (c.) 2025 UTC, and (d.) 2029 UTC on 15 March. The white arrow in (b.) indicates the location of a well-defined hook echo. The reflectivity scale is shown by the color bar on the left of each image.

Figure 20. Radar volumetric cross-section of the Elbert County storm at 2016 UTC. The white arrows indicate the location of a BWER, tilted updraft, and an elevated hail core. The color scale is the same as in Figure 19. Image made with GR2Analyst software.

A sequence of SRV images from KGSP corresponding to the reflectivity images in Figure 19 indicated a rapidly strengthening low-level circulation (Fig. 21). Note that the 2012 UTC SRV image is shown in place of the 2016 UTC image due to velocity dealiasing errors. A gate- to-gate rotational couplet of around 85 knots was evident at 2012 UTC. By 2020 UTC, this couplet strengthened to greater than 100 knots. The EF2 tornado developed at 2015 UTC and dissipated over the extreme southeast corner of the county at 2030 UTC (Fig. 22). Although an operator-defined TVS was identified prior to tornado development, the radar algorithm did not identify this feature as a TVS until near the end of this tornado's life cycle.

Figure 21. Images of 0.5 degree storm relative velocity from KGSP at (a.) 2012 UTC, (b.) 2021 UTC, (c.) 2025 UTC, and (d.) 2029 UTC on 15 March. The white arrow in (b.) denotes the gate-to-gate shear of greater than 100 knots. The speed values are given by the color table in the lower right corner of the figure. By convention, negative values represent motion toward from the radar while positive values represent motion away from the radar.

Figure 22. Track of tornado across the southern part of Elbert County, Georgia, on 15 March 2008. The tornado track is delineated by the heavy black line. The background map is from Delorme Street Atlas USA.

As the tornadic supercell thunderstorms moved out of the eastern end of the original Tornado Watch and into the eastern Piedmont of South Carolina, a new Tornado Watch was issued by the SPC at 2035 UTC. The Elbert supercell thunderstorm continued to produce periodic tornadoes as it moved over the South Carolina Midlands during the late afternoon and early evening. In addition, very large hail from this storm caused millions of dollars in property damage as the core of the storm moved over southern Greenwood County (Fig. 23).

Figure 23. This manufactured home was damaged by wind driven hail near Callison, South Carolina, on 15 March 2008. Image courtesy of Greenwood County Emergency Management.

The final supercell of the day that affected the GSP CWA was not tornadic, but produced extremely large hail (tennis ball to soft ball size) across southern Anderson, northern Abbeville, and northern Greenwood counties. A series of reflectivity scans from KGSP at 2109 UTC is presented in Figure 24. This was around the time that softball size hail was reported by a trained spotter located 7 miles northeast of Iva in southern Anderson County. The images indicated an approximately 10,000 foot layer (between 10,000 feet and 20,000 feet AGL) in which a three-body scatter spike (TBSS) was evident extending southwest from the high reflectivity core. This was an indication of very large hail. The three-dimensional cross section in Figure 25 from 2109 UTC indicated a large area of greater than 60 dBz reflectivity from the low levels of the storm to about 20,000 feet. There was an area of greater than 70 dBz reflectivity at around 16,000 feet. This was well above the freezing level of about 10,000 feet as depicted on regional soundings at 1200 UTC and 1800 UTC. The 50 dBz reflectivity core extended to around 30,000 feet, which was around 10,000 feet higher than the -20 degrees C level as indicated on regional soundings. These factors all point to a thunderstorm that was producing very large hail.

Figure 24. Reflectivity images from KGSP at 2109 UTC at (a.) 1.8 degrees, (b.) 2.4 degrees, (c.) 3.1 degrees, and (d.) 4.0 degrees. The white arrow in (a.) indicates a three-body scatter spike indicative of large hail in the storm core.

Figure 25. As in Figure 20, except at 2109 UTC 15 March.

Strong to severe thunderstorms continued across the western part of South Carolina and northeast Georgia through the late afternoon hours. The passage of a cold front during the early part of the evening pushed the severe thunderstorms to the east of the GSP CWFA ending the threat for severe weather.

5. Summary

A rare outbreak of classic supercell thunderstorms affected portions of northeast Georgia and the Upstate of South Carolina on 15 March 2008. The outbreak was the result of the coexistence of strong vertical wind shear and moderate levels of instability, which was an unusual situation in this part of the country. The moderate levels of instability were largely the result of an atypical eastward advection of the elevated mixed layer (steep mid-level lapse rates). The superimposition of the left exit region of a sub-tropical jet streak and the right entrance region of a maximum in the polar jet resulted in sufficient synoptic- scale lift to remove the cap, allowing for explosive convective development on the warm sector side of the quasi-stationary front by early afternoon. Scattered supercell thunderstorms affected the GSP CWA from about 1830 UTC to 2130 UTC before moving into the South Carolina Midlands. Although large hail, and in some cases destructive hail, were the main impacts in the GSP CWA, two EF0 tornadoes and an EF2 tornado were also reported.

More images of large hail damage near Callison, South Carolina, on 15 March 2008. Images courtesy of Greenwood County Emergency Management.

References

Lane, Justin D., 2008:  A Comprehensive Climatology of Significant 
     Tornadoes in the Greenville-Spartanburg, South Carolina County 
     Warning Area (1880 - 2006).  Eastern Region Tech. Attachment
     2008-01, 35 pp. 

Lanicci, John M., and T. T. Warner, 1991: A Synoptic Climatology of 
     the Elevated Mixed-Layer Inversion over the Southern Great Plains 
     in Spring. Part I: Structure, Dynamics, and Seasonal Evolution.  
     Wea. Forecasting, 6, 198 - 213.

Acknowledgements

Pat Moore and Neil Dixon was responsible for conversion of the manuscript to html code. The convective outlooks, Tornado Watch products, and upper air analyses were obtained from the Storm Prediction Center. The surface fronts analyses were obtained from the Hydrometeorological Prediction Center. The surface observation plots, satellite imagery, and radar mosaics were obtained from the University Corporation for Atmospheric Research. The composite reflectivity mosaic was obtained from the Iowa Environmental Mesonet. Tornado track maps were created using Delorme Street Atlas USA 2006. Radar reflectivity cross sections were created using GR2Analyst version 1.22, by Gibson Ridge Software. The upper air sounding images were made using RAOB version 5.8 for Windows, by Environmental Research Services, LLC. Other radar images were created using the Java NEXRAD viewer from the National Climatic Data Center. Greenwood County (SC) Emergency Management provided most of the damage pictures.