Event Summary
     National Weather Service, Raleigh NC

March 28, 2007 Severe Weather Event
Updated 2009/03/09





Event Headlines -
...The National Weather Service in Raleigh issued 11 Severe Thunderstorm Warnings on Wednesday, March 28, 2007 with over 30 reports of severe weather...
...With little shear but sufficient instability and cold air aloft the primary threat of severe weather was with hail...
...Forecasters used elevated reflectivity cores as an effective warning methodology with this event...
...Some flash flooding was reported but radar precipitation estimates were exaggerated because of enhanced reflectivities associated with hail inside the storm...


Event Overview -
Thunderstorms developed rapidly between 200 PM and 300 PM on March 28, 2007 across the northern portions of the northeast Piedmont and the northern Coastal Plain of North Carolina. The convection developed in a narrow band of instability ahead of a backdoor cold front that was moving slowly southwest across the state. At 2000Z or 400 PM the convection was thriving along a thermal boundary and in the vicinity of a moisture discontinuity stretching from Roxboro in Person county, to Oxford in Granville county, to Enfield in Halifax county. RUC soundings valid during the afternoon in the northeast Coastal Plain showed surface based CAPE values around 2000 J/Kg, precipitable water values of 1.50" (abnormally moist for late March), a freezing level around 11,000 ft. and a wet bulb zero height of approximately 10,500 ft. RUC data also showed steep 1000-500mb lapse rates of 7 to 7.5 C/km during the afternoon. Wind shear was relatively weak with maximum winds of 30 kts through 300 mb. The combination of steep lapse rates and relatively low freezing and wet bulb zero levels indicated hail would likely be the primary severe weather threat. Aloft, a 500 mb ridge was located over the Mid-Atlantic with the ridge axis along and just west of the Appalachian mountains.

With a west to northwest flow aloft (SPC 20Z 700MB Analysis | SPC 20Z 500MB Analysis) the convective cells moved slowly southeast along the instability axis. In general, the thunderstorms over the RAH CWA peaked in intensity shortly after 20Z or 400 PM and then gradually diminished across northern portions of the northeast Piedmont and the northern Coastal Plain through around 630 PM. By 00Z or 800 PM the backdoor cold front had moved southwestward into central North Carolina. An outflow boundary which was first apparent on radar shortly after 19Z (330 PM) had progressed over 80 miles southwest to near Asheboro, Sanford, and Clinton by 0030Z (830 PM).


Severe Weather Reports -
Text of severe weather reports across central North Carolina







Satellite

Visible satellite imagery from 1945Z (345 PM EDT) on Wednesday, March 28, 2007

The visible satellite image below from 1945Z (345 PM EDT) shows the individual convective cells across the northern Coastal Plain and the northern Piedmont of North Carolina. A radar image from the krax WSR-88D also at 1945Z shows the convective cells oriented northwest to southeast across the area. At approximately the same time, large hail up to 1.0 inch in diameter was reported near Aventon in Nash County at 342 PM with 1.0 inch hail also reported in Oxford at 347 PM.

A Java Loop of visible satellite imagery between 1445Z (1045 AM EDT) through 2255Z (655 PM EDT) on Wednesday March 28, 2007 is available here. The loop shows the initial line of cumulus clouds that slide south from far southern Virginia into northern North Carolina from around 1500Z to 1800Z. The showers and thunderstorms that would eventually produce the severe thunderstorms rapidly developed between 1855Z (255 PM EDT)and 1955Z (355 PM EDT).





Surface Analysis

Analyzed surface map from 1800Z (200 PM EDT) on Wednesday, March 28, 2007

The surface analysis from 18Z (200 PM EDT) on Wednesday, March 28, 2007 depicts a "back-door" cold front stretching from northwest to southeast across northern and eastern North Carolina. A weak area of low pressure was located near the front across the western Piedmont of North Carolina. The airmass behind the front was notably cooler with temperatures in the upper 60s to mid 70s across Virginia and the Chesapeake bay region with northeast to easterly winds.

Surface analysis from 1800Z (200 PM EDT) Wednesday, March 28, 2007



Mesoscale Data

Analyzed mean sea level pressure (black) and surface wind barbs from SPC at 20Z on Wednesday, March 28, 2007 (4 PM EDT).
The cold front is apparent with the light convergent wind flow across central and eastern North Carolina.

SPC Analysis at 20Z on Wednesday, March 28, 2007.



Analyzed surface temperatures (red), dewpoints (brown/green), and wind barbs from SPC at 20Z on Wednesday, March 28, 2007 (4 PM EDT).
The frontal zone is clearly seen with the tight temperature and dewpoint gradient stretching across central across North Carolina. Note the more unstable airmass with temperatures well into the 80s and dewpoints in the lower 60s across much of central North Carolina.

SPC Analysis at 20Z on Wednesday, March 28, 2007.



Analyzed low level lapse rates in the 0-3 km layer (blue, green, orange, and red) from SPC at 20Z on Wednesday, March 28, 2007 (4 PM EDT).
Note the large area with a conditionally unstable atmosphere that has low level lapse rates in a layer that is close to becoming "absolutely unstable."

SPC Analysis at 20Z on Wednesday, March 28, 2007.



100-mb Mixed Layer CAPE (red contours) and 100-mb Mixed Layer CIN (shaded in blue) from SPC at 20Z on Wednesday, March 28, 2007 (4 PM EDT).
Note the narrow axis of instability in the 1000-1500 J/kg range extending from southwest Virginia to north-central North Carolina to eastern North Carolina along with the limited CIN.

SPC Analysis at 20Z on Wednesday, March 28, 2007.



300 MB wind barbs (brown), 300 MB isotachs (blue) and analyzed 300 MB divergence (purple) from SPC at 20Z on Wednesday, March 28, 2007 (4 PM EDT).
Note the area of divergence and implied upward vertical motion over eastern North Carolina in response to and central South Carolina.

SPC Analysis at 20Z on Wednesday, March 28, 2007.



NWS Composite Reflectivity Imagery from 2024Z on Wednesday, March 28, 2007 (424 PM EDT).
The composite reflectivity imagery is from the approximate time in which the analysis imagery above is valid.

Composite Reflectivity Imagery from 2024Z on Wednesday, March 28, 2007 .




Radar Loops

A Java Loop of krax base reflectivity imagery from 1800Z (200 PM EDT) through 2301Z (701 PM EDT) on Wednesday March 28, 2007 is available. This loop includes imagery at 15 minute intervals with a total of 22 frames.

A Java Loop of krax base reflectivity imagery from every scan between 1859Z (259 PM EDT) through 2330Z (730 PM EDT) on Wednesday March 28, 2007 is available. This loop includes imagery at 4 minute intervals with a total of 64 frames.


KRAX base reflectivity image from 2022Z on Wednesday, March 28, 2007 (422 PM EDT)





Warning Decision Process


KRAX 4 panel imagery from 1920Z (320 PM EDT) on Wednesday, March 28, 2007.
The images are of Composite Reflectivity in the upper left, Echo Top in the upper right, 0.5 degree reflectivity in the lower left, and VIL in the lower right.

click to enlarge


Once convection initiated, warning forecasters made use of a variety of products to assist in the warning process. Volume scan products such as Composite Reflectivity, VIL and Echo Tops were used to help prioritize which storms needed further investigation. The All-Tilts radar product was especially useful in storm interrogation because it allows forecasters to see every radar elevation angle as it comes in. All-tilts allows forecasters to easily monitor volume-scan to volume-scan convective trends throughout the depth of the storm, such as the strength and vertical motion of reflectivity cores, mid-level convergence and storm top divergence.

A recent study (Gerard 1998) of 39 convective events (64 storms total) in WFO's Cleveland, OH and Jackson, MS showed that convective cells have a high probability of being severe (96%) when 65+ dBZ reflectivity echoes are located at or above the freezing level. This signature worked quite well with storms during the afternoon of March 28, 2006. For example, at 1922Z (322 PM) a severe thunderstorm warning was issued for Granville County when a convective cell near Oxford attained a 63 dBZ core above the freezing level (at roughly 12,000 ft). The signature was seen on the 1.8 degree elevation slice via All-Tilts at 1920Z (320 PM). Several minutes later, at 1924Z (324 PM), the convective cell attained a 65 dBZ core at 24,000 ft, over 10,000 ft above the freezing level! This and subsequent signatures led to very high confidence that the storm in question was severe. Reports of penny and quarter sized hail accumulating on the ground were received at 1947Z (347 PM) in Oxford with this storm. As the afternoon progressed, warnings for subsequent storms were based on signatures similar to the storm in Granville County where very high (~65+ dBZ) reflectivity echoes were being sampled at or above the freezing level.



Flash Flooding

Thunderstorms trained along the nearly stationary boundary, and radar estimated rainfall rates exceeded 3 inches per hour. The storms were lined up parallel to Fishing and Stony Creeks, and flash flood warnings were issued for southern Halifax and northern Nash counties at 330 PM. These were followed by another flash flood warning for Granville county at 400 pm as storms continued to develop almost directly over the city of Oxford. The warnings for Nash and Granville counties were verified by road flooding, and numerous streets were reportedly closed by up to 2 feet of water in Oxford.

A Java Loop of krax storm total precipitation imagery from 1825Z (225 PM EDT) through 2321Z (721 PM EDT) on Wednesday March 28, 2007 is available at this link. This loop includes imagery at 15 minute intervals and shows the evolution of the radars precipitation estimates. Also shown in the loop are the flash flood warnings which appear as green polygons in the animation.

The rainfall estimated by the KRAX radar below were exaggerated by the much stronger reflectivity returns from the large hail cores within the storms. The pixels in blue represent radar rainfall estimates of 4 to 6 inches. Reliable rainfall reports of 2 to 4 inches were received the following morning.

click to enlarge


The two to four inches of rain caused very rapid rises on Fishing Creek and Swift Creek. Fortunately, the storms began to dissipate with the loss of diurnal heating and the creeks crested 2 to 3 feet below bridges across the streams. The height of the river gauge at Fishing Creek at Enfield and at Swift Creek at Hilliardston is shown in feet in the two images below. The images are courtesy of the National Water Information System from the USGS.


Fishing Creek at Enfield




Swift Creek at Hilliardston





Case study team -
Brandon Vincent
Michael Moneypenny
Brandon Dunstan
Jonathan Blaes

For questions regarding the web site, please contact Jonathan Blaes.


  • NWS Disclaimer.