On 11 November 1995, a strong cold front induced a narrow squall line of low-topped (< 25,000 ft) thunderstorms which swept rapidly across central and eastern North Carolina. A tornado watch was issued for portions of western North Carolina for the afternoon hours followed by a severe thunderstorm watch for eastern North Carolina for the evening hours. The National Weather Service Forecast Office at Raleigh (RAH) issued 20 severe thunderstorm warnings as the line passed through its county warning area. Nineteen of these warnings were verified by widespread wind damage reports while one verified with a brief F1 tornado touchdown (Fig. 1).

This event is noteworthy for two reasons. First, the tornado was associated with a classic bow, or line echo wave pattern, on Weather Surveillance Radar - 1988 Doppler (WSR-88D) reflectivity imagery and an extremely well pronounced convergence boundary on storm relative velocity imagery. Second, the widespread wind damage occurred primarily with non-severe criteria (< 50 kt) winds. The purpose of this paper is to provide a mesoscale summary of the event for future reference as well as a training reference for use in conjunction with archived WSR-88D data.

This was a highly dynamic event featuring a strong, rapidly moving surface cold front (Fig. 2) in phase with a deep, negatively-tilted trough aloft. In Fig. 2, note that there is a distinct post-frontal trough; there are two abrupt wind shift lines over southern sections of the front, from south to southwest along the front, and from southwest to west or northwest along the trough.

The major factor that inhibited convection was a deep south to southwest flow at all levels. This flow was producing warm air advection aloft which effectively capped strong convection during most of the squall line's life span. For example, the 500-mb temperature at Greensboro, North Carolina (GSO), had warmed from -15 C at 1200 UTC on 10 November to -10 C at 1800 UTC on 11 November.

Figure 3 contains both observed and modified hodographs from a special 1800 UTC rawinsonde sounding at GSO and a 0000 UTC sounding at Newport, North Carolina (MHX). The squall line passed GSO at approximately 2200 UTC and produced a tornado in Edgecombe County at around 0100 UTC. As such, these hodographs are representative of atmospheric conditions ahead of the front.

Storm relative helicity measures the potential rotation that can be realized by a storm moving through a vertically sheared environment (Davies-Jones et al. 1990). It is critically dependent on existent shear vorticity, storm motion, and the strength of the storm inflow. The Skew-T / Hodograph Analysis and Research Program (SHARP; Hart and Korotky 1991) calculates an initial storm motion 30 to the right of the 0-6 km mean wind using 75% of its magnitude. This simulates the movement that may be expected from a strong supercell (Leftwich 1990).

From Fig. 3, the observed 0-3 km storm relative helicities (using the SHARP-calculated storm motion of about 240 at 35 kt) ranged from 437 (m/s)² at GSO to 494 (m/s)² at MHX, indicative of strong and violent tornadoes, respectively (Davies-Jones et al. 1990). However, when the hodographs are modified with a storm motion derived from WSR-88D imagery (270 at 30 kt), the helicity values increase alarmingly to about 575 (m/s)² at both locations.

No amount of directional shear can compensate for weak inflow (Lazarus and Droegemeier 1990), generally defined as less than 20 kt. In this case, observed low-level inflow winds generally ranged from 30 to 40 kt, while low-level winds using modified storm motion ranged from 45 to nearly 60 kt. It is apparent that low-level shear was significant and any deep convection could potentially produce tornadic activity.

Figure 4 contains unmodified rawinsonde soundings for GSO (1800 UTC) and MHX (0000 UTC). As noted above, these soundings should be representative of conditions prior to frontal passage. A stability analysis reveals stable conditions over central North Carolina and moderately unstable conditions in eastern North Carolina. Some of the more prominent stability features calculated from these soundings are as follows:

1. The convective-available potential energy (CAPE) over central North Carolina was minimal, while modest CAPE (though incalculable from the MHX sounding since the rawinsonde did not reach the equilibrium level) was present in eastern North Carolina;
2. Convective cap strength decreased from a strongly suppressive 4.5 C in central North Carolina to a less suppressive 1.9 C in eastern North Carolina; and,
3. Lifted indices decreased from stable values in central North Carolina (+4) to moderately unstable values (-2) in eastern North Carolina. This is also shown in the lifted index product from the AFOS Data Analysis Program (Fig. 5).

During the first 2 hours that the squall line was detected by the WSR-88D at RAH, it remained nearly linear (Fig. 6a) with an eastward movement around 40 kt (see Klazura and Imy 1993, for a detailed description of WSR-88D products). The maximum tops were below 25,000 ft and the 50-dBZ reflectivity cores ranged from only 8,000 to 10,000 ft (not shown). At this point, RAH was issuing severe thunderstorm warnings for every county along the line due to the vast number of tree damage reports that were received.

As the squall line encountered more moderately unstable air in eastern North Carolina, the central portion of the line accelerated and a classic bow, or line echo wave pattern, emerged (Fig. 6b). Inspection of the 4-panel base reflectivity in Fig. 7, indicates a movement of 40 kt in the northern and southern portions of the line, while the central section had accelerated to 48-52 kt.

As the bow accelerated, a break in the convection occurred (Fig. 7, 0042 UTC) and immediately redeveloped (Fig. 7, 0048 UTC). This redevelopment took place close to the intersection point of the two distinct boundaries indicated on the storm relative motion imagery in Fig. 8. The tornado occurred shortly before 0100 UTC at or very near this point (marked with an X on Figs. 7 and 8). The velocity signature at this point is primarily convergent, although it could be argued that cyclonic rotation is also occurring. The 50-dBZ reflectivity cores were stronger, but still unimpressive, ranging from 10,000 to 14,000 feet (not shown).

All available manual and Automated Surface Observing System (ASOS) wind observations were collected after this event (nine sites) and revealed a surprising fact; with the exception of one documented 59-kt gust recorded at Pope Air Force Base (northern Cumberland County), peak wind gusts from all sources ranged from only 35 to 41 kt, well below the severe criteria threshold of 50 kt.

Damage reports during this event were very widespread with numerous residences and vehicles damaged by falling trees. There were also numerous power outages caused by trees falling onto power lines. Winds in this range are not uncommon and a very similar event occurred on 24 January 1996, during which a squall line produced numerous wind gusts in excess of 30 kt, but no damage. Upon closer scrutiny, it was found that the vast majority of tree damage during this event was from uprooted trees, instead of from trees which had been snapped off. A survey of local emergency officials confirmed that most of the trees which had fallen were shallow-rooted pines.

The cause of the damage stems primarily from soil moisture conditions leading up to the event. A comprehensive survey of rainfall reports from official and ASOS stations as well as cooperative weather observers found that rainfall had averaged about 2.50 inches for the 5- day period, and about 4.25 inches for the 10-day period prior to the event. Closer examination of 1-min winds from ASOS stations showed an abrupt (< 2 min) wind shift from south with gusts to 25 kt to southwest or west with gusts from 35 to 40 kt. The nearly saturated soils could not support the shallow-rooted pines when the stress on the root systems abruptly shifted in direction and increased in magnitude.

As the squall line moved into the RAH county warning area, the office began receiving wind damage reports almost immediately. Pinpointing counties for which to issue severe thunderstorm warnings involved timing the arrival of the line. The widespread nature of the damage was surprising at the time, but post storm analysis showed that shallow rooted pine trees anchored in nearly saturated soil caused the vast majority of the damage. Soil moisture must be considered when strong (albeit non-severe) winds are anticipated, particularly when accompanied by an abrupt wind shift.

Limited instability and a strong convective cap over central North Carolina suppressed deep convection which might have tapped into the existent storm vorticity to induce tornadogenesis. Intersecting convergence boundaries were noted on relative velocity imagery and had been monitored for about 3 h with no activity noted. As the line encountered unstable air and a weaker convective cap over eastern North Carolina, the central portion of the line accelerated, producing a break in the convection directly above the intersection point of the convergence boundaries. A tornado was spawned as the convection regenerated.

This event provides a textbook example of non-supercell tornadogenesis associated with a boundary layer vortex, or misocyclone. Burgess et al. (1993) provides an excellent discussion of misocyclone tornadogenesis including a comprehensive list of additional references. In short, misocyclones tend to develop along or at the intersection of mesoscale surface boundaries (in this case, intersecting convergence boundaries). Developing updrafts produce non-supercell tornadoes by stretching and strengthening these pre-existing misocyclones. Due to the limited horizontal and vertical extent of these vortices (generally confined to the boundary layer), they are unlikely to be directly detected by radar beyond a range of about 25 n mi. The surface boundaries along which they develop, however, are detectable to a range of 50 n mi or more.

This event illustrates several of the precursor conditions that a radar operator must consider prior to a potential severe weather event: 1) current mesoscale environmental conditions and how they vary across the county warning area, 2) conceptual models of the weather pattern that is unfolding and how the mesoscale features are likely to affect this pattern, and 3) knowledge of the antecedent hydrologic conditions leading up to the event.

Thanks are extended to Rodney Gonski, NWSFO RAH, for technical assistance; Dr. Allen Riordan, North Carolina State University Meteorology Dept., for mesoscale surface analysis; and, Chris Vandersip, North Carolina State University Graduate Program, for warning verification.