Event Summary      National Weather Service, Raleigh NC

May 14, 2006 Severe Weather Event Updated 2006/05/24 Event Headlines - ...The 'Mother's Day' severe Weather Event resulted in widespread large hail reports with some reports of wind damage that may have been associated with weak tornadoes... ...The National Weather Service in Raleigh issued 65 Severe Thunderstorm and Tornado Warnings on Sunday, May 14, 2006... ...An F1 tornado touchdown was confirmed just north of Broadway in Lee County by a National Weather Service survey team. Winds were estimated at 100 MPH with path damage of one half mile long and 100 yards wide... ...Widespread hail fog was reported after the storms... Event Overview - Widespread severe weather swept across North Carolina on May 14, 2006. Reports of hail from golf ball to tennis ball size were numerous, especially across the Piedmont and northern Coastal Plain. There was one confirmed tornado touchdown and numerous other reports of funnel clouds. A large, powerful, and persistent area of low pressure in the mid levels of the atmosphere had lingered over the Great Lakes region for several days before dropping southeastward into the Ohio Valley over the weekend. A very cold pool of air was also associated with the low pressure system. A strong jet of westerly winds greater than 110 mph in the upper levels of the atmosphere became a major player in this outbreak as it rounded the southern side of the low pressure system. Typically, vigorous large scale lift can be found in the left front portion of such upper jets, and this helped contribute to numerous strong thunderstorms as the moist air at the surface was forced upward to produce the widespread storms. A cold front extended down the Appalachian Mountains during the mid morning hours with a developing area of low pressure centered over the central NC mountains. A warm front which had initiated some thunderstorms earlier in the morning moved to near the Virginia / North Carolina border by late morning. Sunshine heating the low levels of the atmosphere combined with the cold pocket of air at mid levels resulted in an unstable atmosphere by early afternoon. East of the low pressure system across the Piedmont and Coastal Plain, south and southeast surface winds shifted and increased up through the low levels of the atmosphere, becoming southwesterly as a low level jet entered the Carolinas. The turning of the wind direction and the increasing wind speed with height combined with the unstable air mass allowed several storms to begin rotating and become supercells. The intense rotating updrafts in these storms helped keep ice particles aloft, tumbling within each storm, growing in size before finally falling to the ground as large hail. The powerful rotation inside these supercells produced numerous funnel clouds across the area. National Weather Service Doppler radar indicated two particularly dangerous tornadic storms, one over Lee county that had a confirmed touchdown, and another one that maintained a steady strong state from the beginnings of its lifetime over Alamance county and as it tracked eastward across Durham county, northern Wake county, southern Granville County, to Edgecombe county. This second storm also moved roughly eastward, whereas many other non-severe storms were moving more northeastward. This “right turning” nature is a common indicator of a supercell. Many reports of funnel clouds were received, however most of them did not touch down. This was likely due to the fact that these storms had somewhat high cloud bases. The marginal amounts of moisture in the very low levels of the atmosphere, a condition that is often necessary for tornadoes to “spin up” at the ground, meant that the majority of the funnels were not able to extend to the surface, or did so only briefly. Severe Weather Reports - Text of severe weather reports across central North Carolina

Satellite Visible satellite imagery from 1945Z (345 PM EDT) on Sunday, May 14, 2006 The visible satellite image below from 1945Z (345 PM EDT) shows scattered thunderstorms across portions of central North Carolina. These storms were the second in a series of waves of thunderstorms that moved across this portion of the state. Severe thunderstorms were reported around this time in Durham, Northern Wake, and Franklin counties. A Java Loop of Visible satellite imagery from 1102Z (702 AM EDT) through 2315Z (715 PM EDT) Sunday, May 14, 2006 is available.

Surface Analysis Analyzed surface map from 2100Z (500 PM EDT) Sunday, May 14, 2006 The surface analysis from 21Z (500 PM EDT) Sunday, May 14, 2006 depicts a cold front stretching across the far western Piedmont of the Carolinas. Several areas of low pressure are located along the cold front with a warm front stretching near the Virginia/North Carolina border. A Java Loop of surface analysis imagery from 00Z Sunday, May 14, 2006 through 06Z Monday, May 15, 2006 shows the approach of the cold front, the passage of the surface low and then the cold front moving offshore.

Mesoscale Data Analyzed mean sea level pressure (black) and surface wind barbs from SPC at 23Z on Sunday, May 14, 2006 (7 PM EDT). A cold front can be seen stretching southwest from south-central Virginia across the western Piedmont of North and South Carolina and into northeast Georgia. An area of low pressure developed along the front and is located just northeast of Charlotte at 23Z 300 MB wind barbs (brown), 300 MB isotachs (red) and analyzed 300 MB divergence (purple) from SPC at 23Z on Sunday, May 14, 2006 (7 PM EDT). Note the 110 knot jet max located over southern Tennessee and northern Alabama with a large area of divergence and implied upward vertical motion over central and southern North Carolina. Lifting Condensation Level (red, blue, and green) from SPC at 23Z on Sunday, May 14, 2006 (7 PM EDT). The LCL (Lifting Condensation Level) is the level at which a parcel becomes saturated. It is a reasonable estimate of cloud base height when parcels experience forced ascent. Note that much of central NC has LCL values between 800 and 1200 meters. Analyzed Storm Relative Helicity (SRH) in the 0-3 km layer (light blue, blue, and dark blue) from SPC at 23Z on Sunday, May 14, 2006 (7 PM EDT). Note the axis of greatest SRH across the eastern and southeastern Piedmont and southern coastal plain. The SRH is a measure of the potential for cyclonic updraft rotation in right-moving supercells, and it is calculated for the lowest 3-km layers above ground level. Analyzed Bulk Richardson Number Shear (green), and CAPE (red) from SPC at 23Z on Sunday, May 14, 2006 (7 PM EDT). The Bulk Richardson Number Shear (BRN) is the greatest across the southern Coastal Plain where significant shear is present along with elevated values of CAPE. The greatest CAPE values range between 500 and 1000 J/kg across the Coastal Plain. BRN values of 35-40 or greater have been associated with supercells. NWS Composite Reflectivity Imagery from 2313Z on Sunday, May 14, 2006 (713 PM EDT). The composite reflectivity imagery is from the approximate time in which the analysis imagery above is valid.

Broadway Tornado An F1 tornado briefly touched down 6 miles east of Sanford, North Carolina, or about 2 miles north of Broadway, at approximately 758 PM EDT, May 14, 2006. Structural damage was minor, however there was significant tree damage from uprooted and snapped trees beginning near the intersection of Salem Church Road and Dalrymple Farm Road, continuing for approximately one half mile east through a heavily wooded area just southwest of Copeland Road. Winds were estimated at 100 MPH with path damage of one half mile long and 100 yards wide. The initial damage began at approximately 750 PM EDT in the city limits of Sanford where straight line winds produced minor roof damage to an old warehouse at the intersection of third and Courtland Drive. Shortly thereafter, county highway patrolmen spotted a funnel cloud moving east northeast along Highway 42. Residents along Highway 42 and Avents Ferry confirmed the funnel cloud report, noting a distinct howling sound and very strong winds. The strong straight line winds were estimated at 60 to 75 MPH, snapping small trees on adjacent nearby property. At 758 PM EDT, near the intersection of Salem Church Road and Dalrymple Farm Road, significant damage began with numerous large trees snapped, minor damage to a home, and a destroyed barn. The damage continued east northeast through a heavily wooded forest for approximately one half mile. It is believed that during this one half mile path, the tornado made several brief touchdowns, where there were obvious signs of destruction with downed trees rested in various directions. The damage climaxed in the wooded forest where winds where estimated at 100 MPH. Broadway Tornado Track

Insights into the Tornado Warning Decision Making Process Introduction - The decision to issue a tornado warning is often difficult. Most tornadoes are spawned by supercell thunderstorms. By definition, supercells are rotating thunderstorms. On Doppler radar, supercells are associated with features in the reflectivity data that are indicative of rotation. These features can be pronounced as for example when there is a “hook” or “pendant” extending from the core of the supercell or quite subtle as is the case when there is a “kidney bean” shape to the supercell. The principal point is that the presence of a supercell thunderstorm on radar and the indication of rotation as seen in the reflectivity are not necessarily indicative of a tornado. Indeed, while nearly all supercell thunderstorms produce severe weather (i.e., large hail and/or damaging winds), only 30% percent of all supercells produce tornadoes. 18 Tornado Warnings, 3 Confirmed Tornadoes May 14th - The Mothers Day outbreak of supercells was remarkable given the extent of severe weather in central North Carolina. There were 56 warnings issued that were verified with large hail and/or damaging winds and tornadoes while 65 total warnings were issued. Nearly 90 percent of all severe weather events were preceded by warnings. The accuracy of the warnings for this outbreak exceeds the state of the science and represents excellent success; however, it is reasonable to question the warning decision process that resulted in 18 tornado warnings and only 3 confirmed tornadoes. The Tornado Warning Process, May 14th - When National Weather Service forecasters issue tornado warnings they do so after considering a number of factors that include doppler radar data, satellite data, meteorological features analyzed in the upper air and at the surface, trained spotter and/or public reports of possible tornadoes. Tornado warnings are not routinely issued for supercell thunderstorms. Recall that only 30 percent of all supercells produce tornadoes. Nor are tornado warnings necessarily issued for reports of tornadoes by the public since there are all too often public reports that mistaken other cloud features for tornadoes. The presence of supercells on radar and frequent reports of tornadoes and funnel clouds were not the sole reasons for the issuance of 18 tornado warnings on the 14th of May. Key Question – Will the Supercell Spawn a Tornado ? Recall that supercell thunderstorms rotate. The question is will the rotating thunderstorm spawn a smaller scale but more violent and powerful rotation – a tornado? “Wall clouds” often precede the development of funnel clouds which precede the development of tornadoes. Yet not all “funnel clouds” reach the ground to become tornadoes. The science behind the mechanisms that allow a funnel cloud to reach the ground is not yet fully understood; however much is known. We do know that the higher the cloud ceilings and the lower the near ground dew points (a measure of moisture), the more difficult it is for a funnel cloud to reach the ground. On Mother’s Day May 14th, the cloud bases were relatively high and the surface dew points less than favorably moist. National Weather Service forecasters knew that the overall storm environment was capably of producing tornado touchdowns but factors limiting their occurrence were present. So Why Were 18 Tornado Warnings Issued on 14 May ? Consider the picture below taken by Dave Carroll of a possible funnel cloud/tornado touchdown near the Durham/Granville county line on May 14, 2006. Note how difficult it is to see the developing tornado. While there is the presence of a funnel extending from the cloud base, the circulation about the tornado on the ground is not well indicated. The combination of relatively high ceilings and limited low level moisture resulted in little condensation, making it extremely difficult to see the brief tornado touch downs. Realizing these circumstances, the verification of the Lee County tornado early in the severe weather outbreak and persistent sightings of funnels and possible tornadoes prompted National Weather Service forecasters to issue tornado warnings. Indeed, given the brief touchdowns and the lack of low level condensation to make a tornado more easily detectable, arguably 3 confirmed tornado touchdowns may well under represent the number of tornados that occurred. The decision to issue tornado warnings under the circumstance that were present on the 14th of May was not straight forward but did require a considerable understanding of the science behind tornado genesis.

Radar Loops Java Loop of KRAX radar base reflectivity imagery and warnings during the first round of severe weather from 1134Z (734 AM EDT) through 1503Z (103 AM EDT) Sunday, May 14, 2006. Note - this loop includes 51 frames. Java Loop of KRAX radar base reflectivity imagery and warnings during the second round of severe weather from 1701Z (101 PM EDT) through 2101Z (501 PM EDT) Sunday, May 14, 2006. Note - this loop includes 58 frames. Java Loop of KRAX radar base reflectivity imagery and warnings during the third round of severe weather from 2101Z (501PM EDT) through 0031Z (831 PM EDT) Sunday, May 14, 2006. Note - this loop includes 51 frames. Java Loop of every half hour KRAX radar base reflectivity imagery and warnings from 1134Z (734 AM EDT) through 0358Z (1158 PM EDT) Sunday, May 14, 2006. Note - this loop includes imagery from every half hour (includes 34 frames !) Java Loop of all KRAX radar base reflectivity imagery and warnings from 1134Z (734 AM EDT) through 0358Z (1158 PM EDT) Sunday, May 14, 2006. Note - this loop includes imagery from every radar scan (includes 234 frames !) KRAX base reflectivity image from 2259Z on Sunday, May 14, 2006 (659 PM EDT)

Hail Fog Photos courtesy of Phil Owen and Wayne Hill. Hail is produced by thunderstorms, and thunderstorms in the spring and summer months are associated with relatively warm, moist air at the surface. When copious amounts of hail fall to the ground in a thunderstorm, enough to begin accumulating on the ground, a phenomena known as 'hail fog' can result. Hail fog occurs when hail that has accumulated on the ground cools the air above it. Weather patterns that typically support thunderstorm development often accompany a very warm and moist environment near the surface. If the hail on the ground can cool the overlying air to it's dewpoint, then water vapor will condense out of the air, and a cloud will form. Since this occurs at ground level, the end result is fog. The photos below are a couple examples of the widespread hail fog that was reported after the hail storms. (click on the image to enlarge.)

Up-close Brush with Lightning Photos courtesy of Brandon Vincent. Frequent cloud to ground lightning was observed with many of the severe thunderstorms on Sunday, May 14. The following pictures are of a tree that was struck by lightning between 720-725 PM EDT on Sunday, May 14th in Holly Springs. This lightning strike was witnessed by a National Weather Service employee. The lightning strike caused the bark to literally explode off the tree, leaving a few branches and a pile of bark on the ground next to it. The photos below were taken the following day after the weather had cleared.

Video from the Severe Weather Event Video courtesy of Brian Shrader. The video below was filmed by Brian Shrader at around 2:45 PM EDT on Sunday, May 14 2006. The location is about 1 mile north of downtown Fuquay-Varina in Wake County. Large hailstones between 1.0 and 1.25 inches in diameter (quarter to half-dollar size) can be seen bouncing off the ground and other objects. Leaves and small branches are being shredded off the trees. The All Hazards - NOAA Weather Radio alarm and warning broadcast can be heard in the background. (click on the image to view video.) Note - the video is 41 seconds in length and approximately 4.1 MB in size. The 4 panel radar image below is from the National Weather Service (NWS) Doppler Radar in Newport, NC at 1843Z or 243 PM EDT, very close to the time in which the video was shot. The NWS Raleigh radar was too close to the storm to get a complete scan of the storm. The 4-Panel image contains Composite Reflectivity (CR) imagery in the upper left, "8-bit" Reflectivity imagery “(0.5 slice)” in the upper right, Vertically Integrated Liquid (VIL) in the lower left, and Echo Tops (ET) in the lower right. (click on the image to enlarge.)

Selected Photographs of the Severe Weather Event Photos courtesy of Dave Carroll, Wayne Hill, Steve Marks, Rick Barber, John Hamilton, John Shea, Brandon Dunstan, Brandon Locklear, Michael Finnegan, Phil Owen, and Davis Troxler. (click on the image to enlarge.)

Case study team - Gail Hartfield Brandon Dunstan Brandon Locklear Brandon Vincent Kermit Keeter Jeff Orrock Jonathan Blaes For questions regarding the web site, please contact Jonathan Blaes. NWS Disclaimer.