1. Introduction
On 24 June 2003 a regional tornado outbreak occurred across parts of eastern South Dakota and central and southern Minnesota. During a period from 2214 UTC 24 June through 0412 UTC 25 June, 94 tornadoes were reported. Although the majority of the tornadoes were rated less than F2 on the Fujita scale (Fujita 1971), there were 10 significant tornadoes, including an F4 tornado at Manchester, South Dakota. This event does not meet the more robust tornado outbreak criteria proposed by Doswell et al. (2006) but does meet the definition set forth by Galway (1977), which is related to the number of tornadoes only. In addition to tornadoes, hail of up to 7.6 cm (3.0 in) in diameter and a measured wind gust to 50 m s−1 were recorded. The severe weather was responsible for over $38 million in damage to areas of southeast South Dakota and southern Minnesota.
Postevent analysis indicated there were three primary sectors of thunderstorm development (Fig. 1). The first area, defined as sector 1, occurred along a warm front and was responsible for the most significant tornadoes. Visual appearance [from one of the authors (JMB)] and radar characteristics indicated a classic supercell structure in this region. Sector 2 was located in the warm sector of the surface cyclone and produced the majority of the tornadoes. In this sector, storm morphology transitioned from mesocyclones with maximum rotational velocity (Vr) located at or above 5 km prior to 0030 UTC 25 June to below 3 km after 0030 UTC 25 June, during which time 32 tornadoes occurred. In addition, tornadoes in this region showed a general increase in damage rating with time, with the most damaging tornadoes occurring after 0100 UTC 25 June. Supercells also developed along the cold front in sector 3 and also were responsible for tornadoes. This would be the source region for a quasi-linear convective system (QLCS) that later produced tornadoes in sector 2.
Sectors for mode of tornadogenesis with surface boundaries and cyclone position at 2100 UTC 24 Jun 2003. See text for description of sectors.
Citation: Weather and Forecasting 26, 6; 10.1175/WAF-D-10-05021.1
Owing in part to the early summer occurrence of the tornado outbreak, mixed-layer (ML) convective available potential energy (CAPE) exceeded 3000 J kg−1 as analyzed by the Rapid Update Cycle (RUC; Benjamin et al. 2002). These values of CAPE extended across the warm sector to the cool side of the warm front by late afternoon. With the relatively uniform MLCAPE across all three sectors during the event, it is hypothesized that a variation in the vertical wind profile is responsible for the different mesocyclone evolution patterns over a relatively small geographical area. This study will examine how these variations impacted the vertical shear distribution and corresponded to differences in mesocyclone evolution through the evening.
Shear as an important element in severe weather forecasting long has been recognized (Fawbush and Miller 1954). Some early studies into severe weather forecasting (e.g., Darkow and Fowler 1971; Maddox 1976; Rasmussen and Wilhelmson 1983) used observed soundings to analyze stability and the vertical wind profiles for severe weather. Early numerical modeling studies also noted the importance of shear for severe storms. Klemp and Wilhelmson (1978), using a straight hodograph, modeled the splitting of convection into left- and right-moving cells with respect to the mean wind. Wilhelmson and Klemp (1978) also varied the low- and midlevel environmental winds and found that curvature of the hodograph can affect the evolution of right- or left-moving supercells. Further studies included varying the buoyancy and shear to model the spectrum of thunderstorm morphology (Weisman and Klemp 1982). The origin of midlevel rotation within thunderstorms was shown to be the tilting of horizontal vorticity into the vertical (Davies-Jones 1984; Rotunno and Klemp 1985; Weisman and Rotunno 2000). The ability of an updraft to rotate depends at least in part on the amount of low-level shear, or horizontal vorticity that can be tilted into the vertical. An assessment of this potential can be expressed by storm-relative helicity (SRH; Davies-Jones et al. 1990). Although SRH has been shown to be a valuable tool for supercell forecasting, Markowski et al. (1998b) identified how values of it can vary in space and time within a meso-β-scale or smaller domain. Bunkers (2002) suggested that a measure of deep-layer shear, both bulk and total, is a better indication of supercell potential due to the dependence of SRH on storm motion.
The exact ingredients and their relative levels of importance for the development of low-level rotation within supercells remain unknown. Earlier research highlighted the tilting upward of the thunderstorm-generated horizontal baroclinicity along the gust front (Rotunno and Klemp 1985). Brooks et al. (1994a) found that, in addition to the low-level shear, the development of the low-level mesocyclone is a function of the deep-layer environmental storm-relative winds for the purpose of distributing hydrometeors into the forward and rear flanks of the thunderstorm, and suggested that operational forecasters concentrate not only on the low-level wind profile, but also on the midlevel storm-relative winds near storms for tornado forecasting. The strength and ability of a low-level mesocyclone to produce a tornado also has been related to the presence of a preexisting boundary with which the thunderstorm interacts (Markowski et al. 1998a; Atkins et al. 1999; Rasmussen et al. 2000). Low-level moisture also has been shown to affect the low-level mesocyclone and tornadogenesis. Brooks et al. (1994b) found that low-level absolute humidity is important to the low-level mesocyclone. Based on a large sample of proximity soundings from the RUC, Thompson et al. (2003) found that the lifted condensation levels (LCLs) were significantly lower for tornadic supercells. Markowski et al. (2002) showed that tornadic storms were also associated with smaller dewpoint depressions in the inflow. Supercells with smaller dewpoint depressions were also observed to have a warmer rear-flank downdraft (RFD).
Recent research highlights the role of low-level vortex lines and their evolution within the RFD (Straka et al. 2007). Baroclinically produced vortex lines associated with horizontal vorticity within the RFD are initially tilted downward toward the surface by descending air, stretched horizontally, and then arched upward at the rear of the low-level updraft. This mechanism may explain the increased near-surface rotation, as horizontal vorticity is tilted into the vertical (Straka et al. 2007). The near-surface rotation may be aided by the presence of strong low-level environmental shear, which could strengthen the low-level circulation. This would increase the dynamic pressure perturbation and enhance the updraft, which increases the rate that vortex lines are arched upward within the updraft and in turn increases the near-surface rotation in supercells. Diagnosing the presence of vortex lines in the RFD or evaluating the direct impacts of low-level environmental shear on the dynamic pressure gradient are beyond the scope of this paper. Instead, we will document how changes in environmental low-level shear coincide with changes in the evolution of the low-level mesocyclone as resolved by Weather Surveillance Radar-1988 Doppler (WSR-88D).
Not all tornadoes, or even all significant tornadoes, are associated with persistent low-level mesocyclones. On 24 June, a significant tornado was produced in sector 2 from a QLCS mesovortex. The mesovortices associated with a QLCS have been shown to be produced by the tilting of horizontal, baroclinically produced, crosswise vorticity into the vertical (Trapp and Weisman 2003). Trapp and Weisman (2003) also indicated that this process occurs near the surface and develops vertically, at times reaching the midlevels of the QLCS. Much like supercell mesocyclones, the mesovortices were found to be stronger and longer lived as the low-level environmental shear increased (Weisman and Trapp 2003). Other mechanisms in regards to QLCS tornadoes have been proposed by Atkins and Laurent (2009) and Wakimoto et al. (2006).
On 24 June, important changes in both the bulk and cumulative shears were noted in space and time. This appears to have affected the evolution of the mesocyclones. Operationally, to a forecaster with warning responsibility, anticipating changes in environmental shear is very important, as storm characteristics supporting tornadogenesis can change rapidly. Using hodographs and time–height plots of rotational velocity, we will show how the vertical distribution of the cumulative shear corresponded to different mesocyclone evolution patterns over a relatively small area, and how this can help operational forecasters to better prepare for warning responsibility.
Section 2 is the methodology. Section 3 will present a brief synoptic overview of the conditions on 24 June. The evolution of tornadoes and mesocyclones will be discussed in section 4, followed by a discussion in section 5 on forecasting implications. Finally, section 6 will provide concluding remarks.
2. Methodology
a. Observational data
North American upper-air radiosonde data, surface observations, and 1-km visible satellite data were obtained from the Cooperative Program for Operational Meteorology, Education, and Training (COMET) data archives. Both surface and upper-air data were displayed using the General Meteorological Package (GEMPAK; desJardins et al. 1991). WSR-88D radar data from Sioux Falls, South Dakota (KFSD), were archived at the National Weather Service Weather Forecast Office (NWS WFO) in Sioux Falls from data transmitted to the Advanced Weather Interactive Processing System (AWIPS). Radar data were received every 5 min with 14 levels of beam tilt from 0.5° to 19.5°.1 The WSR-88D radial velocity data were available at ~0.5 m s−1 intervals.
Hodographs were used to calculate the shear distribution associated with different mesocyclones during the event. Because observed proximity soundings were not available, these hodographs were constructed using gridpoint soundings from the RUC analysis for the hour around the development of the mesocyclone. Thompson et al. (2003) showed that the RUC analysis provides a reasonable approximation of the supercell environment. They found that the model errors were generally smallest just after 0000 and 1200 UTC. Wind data were interpolated every 500 m from 500 m above ground level (AGL) to 8 km AGL. Surface wind data were taken from the observation closest to the gridpoint sounding location and also located in the inflow region of the supercell.
Both total shear and bulk shear were calculated from the RUC wind profile. In many operational forecast settings, the wind difference in a layer has been used in lieu of the actual bulk shear, which is the wind difference divided by depth. In this study, we provided both numbers; the shear will be listed first with units of s−1 with the more familiar wind difference shown in parentheses with units of m s−1. The latter is included to provide a common reference for operational forecasters. This will be done for both bulk shear, which is the magnitude of the vector difference from the top and bottom of the layer, and for total shear, which is the summation of vector differences for smaller segments within the fixed layer of the hodograph.
The use of total shear has received limited attention in the literature in regard to severe weather forecasting. Bunkers (2002) noted that for a curved hodograph, the difference between the bulk shear and total shear can be as much as a factor of 2. One concern is that, for a given hodograph, total shear can vary depending upon the depth over which each shear segment is calculated. To avoid including insignificant changes in wind speed and direction over depths of <500 m, the total shear is calculated using a summation of 0.5-km segments (Bunkers 2002). Using this methodology, Bunkers (2002) found that 0–6-km total shear of 0.0030–0.0040 s−1 (20–25 m s−1) is commonly associated with supercells. Total shear will be presented in a time versus height graph to illustrate how changes in the distribution of shear appear to affect mesocyclone development and evolution.
b. Rotational velocity
Similar to Atkins et al. (2005), rotational velocity (Vr) was calculated for mesocyclones from their initiation to their dissipation through the depth of the storm. A mesocyclone was identified when a Vr of 10 m s−1 was observed at two adjacent elevation angles and persisted for at least three volume scans (approximately 10 min). After identifying a mesocyclone, the Vr was calculated by finding the maximum inbound and outbound velocities within the mesocyclone. The maximum inbound and outbound velocities had to be located within 10 km of each other. The Vr is then the average of the absolute value of the maximum inbound and maximum outbound velocities.2
Bad radial velocity data at times existed from the KFSD WSR-88D, including instances of range-folded data near the circulation. In other instances, data were determined to be improperly dealiased if one of the following two criteria were met:
radial continuity—adjacent values along the radial differed by >25 m s−1 and in these cases data from the entire radial were not used, and
vertical continuity–radial velocity values change by >25 m s−1 for adjacent elevation angles. Only range bins that failed the test were not used. In cases where vertical continuity failed at multiple elevation angles (i.e., from 0.5° to 1.5° and from 3.4° to 4.3°), then all elevation angles in between also were not used.
3. Environmental discussion
The 250-hPa 0000 UTC 25 June analysis indicated a 45 m s−1 jet from the central Rocky Mountains to the northern plains (Fig. 2a). The region from eastern South Dakota into central and southern Minnesota was located in the exit region of the jet. Clark et al. (2009) indicated that this was a favorable region for low-level ascent and severe weather reports. The upper-level southwest flow allowed an elevated mixed layer (EML; Lanicci and Warner 1991) to spread across the region. The average environmental lapse rate from 700 to 500 hPa indicated that values >7.0°C km−1 extended from Colorado and New Mexico to the Great Lakes (Fig. 2b). Below the EML, 850-hPa dew points of 15°C or greater extended from the southern plains to Minnesota (not shown). The combination of the 850-hPa moisture and the lapse rates created conditional instability favorable for deep moist convection (DMC).
(a) The 250-hPa heights (solid, every 60 m), and wind isotachs (shaded every 5 m s−1). Station plots include height (dam), temperature and dewpoint (°C), and wind (full barb, 10 m s−1; half barb, 5 m s−1). (b) The 700-hPa analysis of height (solid, every 30 m), temperature (dashed, every 3°C), and the 700–500-hPa lapse rates >6.5°C km−1 are shaded.
Citation: Weather and Forecasting 26, 6; 10.1175/WAF-D-10-05021.1
The mesoscale environment on 24 June also was favorable for DMC. The surface analysis at 1200 UTC 24 June showed a cold front from northern Minnesota to western Nebraska (Fig. 3a). An outflow boundary from previous convection extended from southern Minnesota to northern Nebraska. During the next 9 h, the outflow boundary became stationary before developing into a warm front and moved north extending from southeast South Dakota into southern Minnesota by 2100 UTC (Fig. 3c). The warm front continued to move north through the late afternoon as the cold front over westernNebraska accelerated east (Fig. 3d). Figure 4 provides snapshots of how the outflow boundary evolved during the day. At 1500 UTC, the outflow boundary was located across northeast Nebraska and northwest Iowa. By 1800 UTC, the boundary was still visible from north-central Nebraska into southwest Minnesota as it evolved into a warm front. Finally, at 2100 UTC, the northern edge of the widespread cumulus field extending from southeast South Dakota to southern Minnesota denoted the location of the front.
Surface analysis from (a) 1200 UTC 24 Jun, (b) 1800 UTC 24 Jun, (c) 2100 UTC 24 Jun, and (d) 0000 UTC 25 Jun 2003. Short dotted lines denote outflow boundaries, long dashed lines indicate surface trough position, and other frontal analysis is conventional.
Citation: Weather and Forecasting 26, 6; 10.1175/WAF-D-10-05021.1
Visible satellite pictures from (a) 1500 UTC 24 Jun, (b) 1800 UTC 24 Jun, and (c) 2100 UTC 24 Jun 2003.
Citation: Weather and Forecasting 26, 6; 10.1175/WAF-D-10-05021.1
The northward surge of the warm front allowed for a very warm and moist boundary layer to move into southeast South Dakota and southwest Minnesota. Conditional instability was observed from upper-air soundings at Omaha, Nebraska (KOAX), and Aberdeen, South Dakota (KABR) (Fig. 5). KOAX, located in the warm sector, exhibited a moistening and deepening boundary layer from 1800 to 0000 UTC. Using the virtual temperature correction (Doswell and Rasmussen 1994), the KOAX sounding indicated <−10 J kg−1 of 100-hPa ML convective inhibition (CIN) by 0000 UTC. The sounding also indicated 3880 J kg−1 of MLCAPE. The warm sector remained south of KABR through 0000 UTC, which led to a frontal inversion on both soundings. Although most unstable (MU) CAPE was 2068 J kg−1 due to steep midlevel tropospheric lapse rates, MLCAPE was 688 J kg−1 with MLCIN of −288 J kg−1. The KABR sounding also exhibited strong directional shear below 700 hPa, similar to RUC hodographs in sector 1 (shown later).
Skew T–logp diagrams of radiosondes from (a) Omaha (KOAX) and (b) Aberdeen (KABR) at 1800 UTC 24 Jun 2003 (dashed) and 0000 UTC 25 Jun 2003 (solid). For both (a) and (b), wind barbs on the left are from 1800 UTC and on the right are from 0000 UTC. Wind barbs are as in Fig. 2.
Citation: Weather and Forecasting 26, 6; 10.1175/WAF-D-10-05021.1
In addition to the large amount of CAPE, a favorable wind profile for supercells was present for both locations at 1800 UTC, with wind speeds increasing and veering with height. By 0000 UTC, the 0–6-km bulk shear at KOAX decreased from 0.0037 s−1 (22 m s−1) to 0.0025 s−1 (15 m s−1), while the bulk shear remained at 0.0058 s−1 (35 m s−1) at KABR. This acted to increase the gradient in bulk shear across the central and northern plains (Fig. 6). The greatest values of bulk shear were located in the cool sector of the cyclone, with deep-layer shear decreasing in the warm sector over southeast South Dakota and eastern Nebraska.
RUC analysis of bulk shear from the mean 0–500-m AGL wind to 6 km AGL contoured every 10 × 10−4 s−1 at 0000 UTC 25 June.
Citation: Weather and Forecasting 26, 6; 10.1175/WAF-D-10-05021.1
4. Evolution of tornadoes and mesocyclones
a. Temporal distribution of tornadoes
The movement of the warm front was important in the evolution of convection across the northern plains. The first supercells developed along the warm front over southeast South Dakota between 2030 and 2200 UTC (Fig. 4c). The warm front continued to focus convective activity across sector 1 through the evening (e.g., Fig. 7). Other thunderstorms developed in the weakly capped warm sector south of this warm front over northeast Nebraska around 2200 UTC. These storms spread across sector 2 during the evening.
KFSD 0.5° beam tilt reflectivity in dBZ at 0058 UTC 25 Jun 2003 with surface features overlaid: sector (top) 1 and (bottom) 2.
Citation: Weather and Forecasting 26, 6; 10.1175/WAF-D-10-05021.1
Although thunderstorms developed in sectors 1 and 2 prior to 2200 UTC, the evolution of tornado reports through the event in each sector was very different (Fig. 8). In sector 1, thunderstorms spawned tornadoes soon after developing. Reports of tornadoes, some F2 or greater, persisted through 0230 UTC 25 June before the convective mode changed. Despite the MLCAPE > 3000 J kg−1 and favorable deep-layer shear in sector 2, initial thunderstorms only produced five F0 or F1 tornadoes. The tornado frequency increased after 0000 UTC. More importantly, the intensity of the tornadoes also increased, with the first of two F2 tornadoes occurring after 0100 UTC. Was there a difference in mesocyclone evolution between the two sectors? To answer this question, time–height graphs of rotational velocity will be examined similar to Atkins et al. (2005).
Graph of the number of tornadoes reported each hour within sectors (a) 1 and (b) 2.
Citation: Weather and Forecasting 26, 6; 10.1175/WAF-D-10-05021.1
b. Warm frontal supercells: Sector 1
1) Mount Vernon supercell: 2112–2232 UTC 24 June
A thunderstorm developed south of Mount Vernon, South Dakota, around 2053 UTC 24 June. While this storm was associated with two mesocyclones during its lifetime, only the first was examined (Fig. 9). The storm was ~120 km from the radar. It was initially located south of the surface warm front but moved north of it around 2152 UTC. A hodograph constructed from the RUC at 2100 UTC showed a favorable profile for cyclonic mesocyclones, as shown by Rasmussen and Wilhelmson (1983; see Fig. 9b). The 0–6-km AGL3 total shear was 0.0055 s−1 (33 m s−1) and the bulk shear was 0.0045 s−1 (27 m s−1). In the 0–1-km layer, the shear was 0.0100 s−1 (10 m s−1), while in the 3–6-km layer, the total shear was 0.0040 s−1 (12 m s−1) and the bulk shear was 0.0037 s−1 (11 m s−1). Over the next 20 min, the storm organized, and by 2112 UTC, it had developed weak rotation near 6 km that extended to a depth of over 5 km by 2117 UTC (Fig. 9c). This mesocyclone persisted for 90 min, with Vr exceeding 20 m s−1 for approximately an hour.
(a) Track of the centroid of the Mount Vernon mesocyclone from 2117 to 2237 UTC 24 Jun 2003. County names are boldfaced in gray and city names are set in black. (b) RUC hodograph at 2100 UTC 24 Jun 2003 near Mount Vernon. (c) Time–height analysis of the maximum rotational velocity of the mesocyclone. Thick black lines denote tornado life spans.
Citation: Weather and Forecasting 26, 6; 10.1175/WAF-D-10-05021.1
Meanwhile, the rotation below 3 km remained much weaker. Once the storm crossed the warm front, Vr below 3 km rapidly increased. As noted by Markowski et al. (1998a), the presence of a warm front can provide additional horizontal and vertical vorticity to aid in the development of low-level mesocyclones. Fifteen minutes after crossing the warm front, Vr increased to 17 m s−1 and the midlevel mesocyclone intensified. At 2215 UTC, an F0 tornado developed south of Mount Vernon. It dissipated after 2 min, and an F2 tornado formed near Mount Vernon and lasted for about 15 min. The mesocyclone rapidly dissipated after the tornado ended, and a nontornadic second mesocyclone formed (not shown).
2) Manchester supercell: 2358 UTC 24 June–0058 UTC 25 June
The most damaging tornado to develop on 24 June, rated F4, occurred near Manchester, South Dakota, between 0033 and 0057 UTC 25 June. This storm developed north of the surface warm front after the merger of two supercells southwest of Iroquois, South Dakota, at 2353 UTC (Fig. 10a). A RUC hodograph from 0000 UTC near Manchester showed the 0–6-km total shear was 0.0042 s−1 (25 m s−1), while the bulk shear was 0.0035 s−1 (21 m s−1); the 0–1-km shear was 0.0110 s−1 (11 m s−1) (Fig. 11b). The total shear in the 3–6-km layer decreased but was 0.0030 s−1 (9 m s−1), with bulk shear of 0.0023 s−1 (7 m s−1). This was also a typical hodograph for a classic supercell.
As in Fig. 9, but for the mesocyclone associated with the Manchester F4 tornado from 2358 UTC 24 Jun to 0058 UTC 25 Jun.
Citation: Weather and Forecasting 26, 6; 10.1175/WAF-D-10-05021.1
As in Fig. 9, but for mesocyclone 1 southwest of Vermillion from 2147 to 2232 UTC 24 Jun 2003.
Citation: Weather and Forecasting 26, 6; 10.1175/WAF-D-10-05021.1
The midlevel mesocyclone rapidly developed after the storm merger at 2358 UTC (Fig. 10c). By 0003 UTC, Vr exceeded 20 m s−1 above 9 km and remained >20 m s−1 for almost 45 min. The low-level mesocyclone developed more quickly below 3 km than in the previous case. In this case, the low-level mesocyclone was observed as a secondary maximum in Vr and somewhat separated from the midlevel mesocyclone. This was similar to the mesocyclone evolution discussed in Rasmussen et al. (2000). As near Mount Vernon, a tornado developed and persisted for over 20 min. After the tornado and mesocyclone dissipated, another mesocyclone developed that also produced an F2 tornado (not shown).
c. Warm-sector supercells: Sector 2
1) Vermillion supercell mesocyclone 1: 2152–2232 UTC 24 June
At 2142 UTC, a supercell developed southwest of Vermillion, South Dakota, in northeastern Nebraska. Several mesocyclones developed with this storm as it moved into southeast South Dakota. The first mesocyclone developed at 2152 UTC near Sholes, Nebraska, and persisted for 40 min (Fig. 11a). The hodograph from the 2200 RUC showed that the 0–6-km total shear was 0.0052 s−1 (31 m s−1), while the bulk shear was 0.0028 s−1 (17 m s−1) (Fig. 11b). The total shear was stronger than along the warm front, but the deep-layer bulk shear was marginal for supercell development (Rasmussen and Blanchard 1998; Houston et al. 2008) and lower than in sector 1. The shear was distributed throughout the hodograph though, with the 3–6-km total shear 0.0040 s−1 (12 m s−1) and the bulk shear 0.0037 s−1 (11 m s−1). The 0–1-km bulk shear was similar to locations in sector 1, with values of 0.0100 s−1 (10 m s−1).
Like the warm-frontal supercells, the initial mesocyclones began at the height of 7 km at 2152 UTC and 10 min later had a depth of over 5 km. By 2217 UTC, the maximum Vr exceeded 20 m s−1. Unlike the warm-frontal mesocyclones, a strong circulation was unable to develop below 3 km. No tornado was observed, and the mesocyclone dissipated after 2232 UTC.
2) Vermillion supercell mesocyclone 3: 2227–2333 UTC 24 June
A second mesocyclone formed in the Vermillion supercell at 2212 UTC and persisted for only 20 min (not shown). Fifteen minutes later, a third mesocyclone formed to the south of the first circulation. This circulation formed near Coleridge, Nebraska, and moved toward Vermillion, South Dakota (Fig. 12a). Total shear decreased in the warm sector as the short-wave trough approached. The 3–6-km total shear decreased from 0.0040 (12) to 0.0037 s−1 (11 m s−1). The 3–6-km bulk shear decreased from 0.0037 (11) to 0.0027 s−1 (8 m s−1) (Fig. 12b). At the same time, the 0–1-km bulk shear increased to 0.0120 s−1 (12 m s−1). This was the result of a larger increase in wind speed around 1 km as the low-level jet (LLJ) developed across southeast South Dakota and eastern Nebraska. Wind speeds at 1 km increased from 16 to 17.5 m s−1 in 1 h.
As in Fig. 9, but for mesocyclone 3 near Vermillion from 2227 to 2332 UTC 24 Jun 2003.
Citation: Weather and Forecasting 26, 6; 10.1175/WAF-D-10-05021.1
Like the first circulation with this supercell, the third mesocyclone rapidly developed above 6 km. By 2247 UTC, the Vr was over 25 m s−1 and reached a peak of 38.7 m s−1 at 2304 UTC. With the increased low-level shear, a circulation existed below 3 km and exceeded 10 m s−1 for 40 min. During this time, two brief F0 tornadoes were reported in northeast Nebraska and were the first tornadoes in the warm sector. Unlike the Manchester mesocyclone, a secondary maximum in Vr did not form below 3 km. After 2322 UTC, the circulation below 3 km was <10 m s−1, and the mesocyclone dissipated after 2333 UTC.
A fourth mesocyclone formed at 2328 UTC and also produced a brief F0 tornado west of Vermillion (not shown). Like the previous mesocyclones, the strongest circulation was above 6 km.
3) Centerville supercell mesocyclones 2 and 3: 0018–0118 UTC 25 June
At 2318 UTC, a second supercell formed along the Missouri River west of Vermillion. The first mesocyclone formed at 2338 UTC and lasted until 0038 UTC (not shown). This mesocyclone evolved like those in northeastern Nebraska, with the strongest circulation remaining above 6 km and a weaker circulation extending below 3 km. No tornado was reported with this mesocyclone.
Two mesocyclones, separated by 10 km, formed at 0018 UTC and eventually merged with the primary cell during their life cycle. Mesocyclone 2, which merged first with the supercell, formed south of Wakonda, South Dakota, and moved northeast (Fig. 13a). The 0–6-km total shear increased to near 0.0065 s−1 (39 m s−1) with the bulk shear near 0.0023 s−1 (14 m s−1) (Fig. 13b). Although, little shear was located in the 3–6-km layer, the 0–1-km shear increased to 0.0018 s−1 (17.5 m s−1) due to the strengthening LLJ. The mesocyclone developed above 3 km early in its life cycle, with Vr > 20 m s−1 observed by 0033 UTC. The circulation above 6 km reached its peak at 0040 UTC and slowly weakened over the next 20 min. Meanwhile, a secondary maximum in the rotational velocity rapidly developed below 3 km after 0028 UTC, with a brief F0 tornado observed. Unlike the Vermillion mesocyclones, the low-level circulation persisted and strengthened. After merging with the primary cell between 0043 and 0048 UTC, the Vr at the 0.5° elevation angle increased to 30.5 m s−1. At the same time, the mesocyclone began to move to the northwest and an F2 tornado developed. The circulation then weakened rapidly. Unlike the Vermillion supercell, the remnant circulation was most evident below 6 km.
As in Fig. 9, but for mesocyclone 2 near Centerville from 0018 to 0058 UTC 25 Jun 2003.
Citation: Weather and Forecasting 26, 6; 10.1175/WAF-D-10-05021.1
Centerville mesocyclone 3 followed a similar track with Centerville mesocyclone 2 (Fig. 14a), with a mesocyclone that exhibited two distinct stages (Fig. 14c). The first stage lasted for 20 min (0018–0038 UTC) and was similar to the evolution of Vermillion mesocyclone 1. A circulation rapidly developed around 6 km, with Vr exceeding 20 m s−1 within a few minutes of formation. The rotation did extend below 3 km, but no secondary maximum was observed on radar. The midlevel rotation briefly weakened around 0038 UTC, with a new Vr maximum developing around 3 km at 0043 UTC. The mesocyclone intensified, with Vr exceeding 30 m s−1 by 0058 UTC. As the mesocyclone merged with the primary supercell, the rotation at the 0.5° elevation angle increased to 31.1 m s−1 and moved to the northwest, with an F1 tornado forming at 0111 UTC. Both the tornado and mesocyclone dissipated after 0118 UTC.
As in Fig. 9, but for mesocyclone 3 near Centerville from 0018 to 0123 UTC 25 Jun 2003. The constructed hodograph is from Centerville at 0100 UTC 25 Jun 2003.
Citation: Weather and Forecasting 26, 6; 10.1175/WAF-D-10-05021.1
4) Centerville supercell mesocyclone 4: 0113–0148 UTC 25 June
The next mesocyclone observed with the Centerville supercell formed around 0113 UTC north of Centerville (Fig. 15a). The circulation formed near the weakening Centerville mesocyclone 3. By 0200 UTC, the 0–6-km total shear was 0.0070 s−1 (42 m s−1), with the bulk shear at 0.0022 s−1 (13 m s−1). A majority of this shear was below 3 km (Fig. 15b). The 0–1-km shear increased to 0.0190 s−1 (19 m s−1) by 0200 UTC.
As in Fig. 9, but for mesocyclone 4 near Centerville from 0113 to 0148 UTC 25 Jun 2003.
Citation: Weather and Forecasting 26, 6; 10.1175/WAF-D-10-05021.1
This mesocyclone moved to the northwest, but its overall evolution was much different from that observed with earlier mesocyclones in sector 2. It developed rapidly throughout the lowest 9 km of the atmosphere (Fig. 15c), with the largest Vr located near 3 km through the life span of the mesocyclone. At the 0.5°-elevation angle, the Vr exceeded 25 m s−1 within 10 min of forming, and an F2 tornado was observed. From 0133 to 0143 UTC, the Vr below 3 km decreased from 21 to <10 m s−1, with no circulation observed at the 0.5° elevation angle by 0148 UTC. The evolution of this mesocyclone was similar to that observed with QLCS, as discussed by Funk et al. (1999) and Atkins et al. (2005).
5) Hartford mesovortex: 0313–0338 UTC 25 June
A QLCS was located west of Sioux Falls at 0313 UTC. A mesovortex within the QLCS developed approximately 20 km west of Sioux Falls, producing an F1 tornado near Hartford, South Dakota (Fig. 16a). The 0–1-km shear was 0.0190 s−1 (19 m s−1) with a profile that was similar to the tornadic QLCS hodograph shown by Funk et al. (1999).
As in Fig. 9, but for mesocyclone near Hartford from 0253 to 0328 UTC 25 Jun 2003.
Citation: Weather and Forecasting 26, 6; 10.1175/WAF-D-10-05021.1
Like many QLCS circulation patterns, the Hartford mesovortex had a short life cycle that lasted only 25 min (Weisman and Trapp 2003). Within 5 min of being detected, the circulation between 1 and 5 km was greater than 35 m s−1 and an F1 tornado was observed. While the proximity of the mesovortex to the radar prevented sampling above 6 km, the decrease in Vr above 5 km indicated that the strongest circulation may have been near the surface through the entire evolution. After 25 min, the circulation no longer was detectable by radar. This is similar to what was documented by Funk et al. (1999) and Atkins et al. (2005) for tornadic QLCS mesovortices.
5. Discussion
Rasmussen et al. (2000), drawing also on prior work by Markowski et al. (1998a), offer a hypothesis for future testing that “significant tornadoes require augmentation of storm-relative helicity … beyond what is usually thought to be associated with environments conducive to tornado outbreaks … In other words, it is plausible that the large scales rarely provides sufficient vorticity and CAPE for significant tornadoes. There is no clear evidence in previous case studies of the occurrence of significant tornadoes without boundaries being present” (their italics). While our dataset does not provide similar meso-β-scale resolution for observations, there is no evidence of a discernible boundary in sector 2 during the event. Yet 40 tornadoes formed in this area, 2 of which produced F2 damage. Given the data available for this study, it was not possible to determine definitively the process of tornadogenesis. However, as Rasmussen et al. (2000) noted, a low-level mesocyclone is necessary for the development of significant supercell tornadoes. The results of this study can be used to examine how the low-level mesocyclones evolved and what factors appeared to influence their evolution.
Figure 17 shows the magnitude of the bulk vector difference in 500-m layers from the surface to 8 km. In sector 1, there was a maximum in shear between 0.5 and 1 km, with a second maximum between 2 and 5 km through 0000 UTC. This environment favored a mesocyclone evolution with Vr first increasing in the midlevels of the atmosphere. As supercells interacted with the warm front, the Vr increased below 3 km, followed by tornadoes developing similar to what was described by Rasmussen et al. (2000) and Trapp et al. (1999). In sector 2, the event began with sufficient deep-layer shear for supercells, but prior to 0000 UTC, the magnitude of the low-level shear was less than what was observed in sector 1. After 0000 UTC, shear decreased in the midtroposphere at the same time the 0–1-km shear increased. This structure in the shear distribution remained the same through 0400 UTC 25 June. After 0030 UTC, the mesocyclone evolution changed and became similar to that described by Atkins et al. (2005).
Shear distribution as a function of height and time from 2100 UTC 24 Jun to 0300 UTC 25 Jun in (a) sector 1 based upon the RUC analysis wind field near Manchester and (b) sector 2 based upon the RUC analysis wind field near Centerville.
Citation: Weather and Forecasting 26, 6; 10.1175/WAF-D-10-05021.1
The change in the distribution of shear in sector 2 can be attributed to the development of an LLJ from Kansas into Minnesota. The LLJ developed as the 250-hPa jet streak moved into the northern plains. The development of the LLJ can be seen in the Neligh, Nebraska (NGLN1), wind profiler (Fig. 18a) after 2200 UTC as wind speeds increased below 3 km. Farther north, the velocity azimuthal display (VAD) wind profiler (VWP) from KFSD (Fig. 18b) also showed an increase in the low-level wind speeds after 2200 UTC that remained over the area through 0400 UTC. Based upon the RUC hodograph near Centerville at 0100 UTC 25 June, the 0–1-km shear was 0.0150 s−1 and is similar to that found in the vicinity of boundaries where low-level mesocyclones were observed to develop below 3 km (Rasmussen et al. 2000; Fig. 19). Similar to Kis and Straka (2010), this suggests that one mechanism by which the LLJ may be important in low-level mesocyclogenesis is through increasing horizontal vorticity. At the same time, the boundary layer remained uncapped with the 0–3-km CAPE > 125 J kg−1, indicating convection was likely surface based. The redistribution of shear to the lowest 3 km of the atmosphere coincided with an increase in Vr below 3 km and an increase in the number and rating of tornadoes in sector 2.
Time–height display of wind from 1800 UTC 24 Jun to 0400 UTC 25 Jun for (a) the wind profiler in Neligh and (b) the VWP from KFSD.
Citation: Weather and Forecasting 26, 6; 10.1175/WAF-D-10-05021.1
Plot of surface relative vorticity every 1 × 10−5 s−1 at 0000 UTC 25 Jun 2003 from the unmodified RUC analysis. Shaded are values >2 × 10−5 s−1, and dashed lines are <0.
Citation: Weather and Forecasting 26, 6; 10.1175/WAF-D-10-05021.1
A scenario in which the LLJ exists in the presence of an uncapped boundary layer appears uncommon. In most cases, the LLJ does not develop until the free atmosphere has decoupled from the boundary layer (Mitchell et al. 1995). The stabilization of the boundary layer means that surface-based convection is less likely to occur. Conversely, the MLCAPE necessary for convective updrafts is more common on summer afternoons when LLJs are uncommon (Mitchell et al. 1995). A situation when both sufficient surface-based CAPE (SBCAPE) and shear exist in the absence of a discernible boundary is when there is a dynamically induced LLJ (Uccellini and Johnson 1979). On 24 June 2003, the presence of both SBCAPE and an LLJ appears to have influenced the development of low-level mesocyclones.
6. Concluding remarks
The tornado outbreak of 24 June 2003 was associated with two modes of low-level mesocyclone development. By late afternoon, the 0–6-km bulk shear was >0.0043 s−1 (25 m s−1) while MLCAPE exceeded 3000 J kg−1. A surface warm front provided a mesoscale source for lift, and strong southeast flow during the preceding 48 h allowed ample moisture to be in place. All these ingredients were favorable for the development of rotating updrafts and supercells.
With the storms in sector 1, where both deep-layer shear and a surface boundary existed, storms rapidly evolved into classic supercells. Time–height sections of Vr showed that rotation developed around 6 km. Approximately 30 min after forming, the rotation increased below 3 km, and tornadoes occurred shortly thereafter. The evolution of the storms was similar to that discussed by Rasmussen et al. (2000) when supercells interacted with surface boundaries. In sector 2, there was also sufficient deep-layer shear prior to 0000 UTC, but no surface boundary was present, which limited the 0–1-km shear. Correspondingly, rotation was weak at 3 km compared to 6 km. Therefore, only five F0 or F1 tornadoes were observed, and no significant tornadoes occurred. As the evening progressed, the shear above 3 km decreased while the 0–1-km shear increased. The evolution of the mesocyclones changed by developing, or re-forming, closer to 3 km, with rapid increases in rotational velocities near the surface, and two stronger and longer-lived tornadoes occurred. By 0300 UTC, a mesovortex formed rapidly through the entire depth of the storm, with tornadoes observed within 10 min of the mesovortex development. The later mesovortex followed a pattern of evolution observed with QLCS tornadoes (Funk et al. 1999; Atkins et al. 2005).
The change in the mesocyclone evolution pattern in sector 2 coincided with the development of the LLJ. After 0000 UTC, the 0–1-km shear values increased to >0.0150 s−1. The LLJ, by increasing the horizontal vorticity below 1 km, may have influenced the development of the low-level mesocyclone. As the dynamically induced LLJ developed near 1 km shortly after peak heating, it was observed that the evolution of the mesocyclones became similar to that of the QLCS mesovortices, with the strongest Vr below 3 km (Atkins et al. 2005). We caution that this is only one event; researchers need to examine additional events where significant tornadoes appeared to develop without discernable surface boundaries when SBCAPE exists in a weakly capped boundary layer and values of 0–1-km shear approach values similar to those observed with boundaries.
Forecasters need to maintain awareness, not only of the presence of boundaries, but also of the development of dynamically induced LLJs that can change the distribution of the shear near the surface. This can be done by examining observational datasets within the warm sector such as velocity wind profiles (VWP) and/or wind profilers, as well as short-term models when surface-based convection is developing. Because LLJs can develop within a couple of hours, the evolution of mesocyclones and their associated weather can change rapidly in the midst of ongoing convection.
Acknowledgments
We thank Ray Wolf, Barbara Mayes, and Jeffrey Manion for their helpful suggestions that improved the manuscript. We also thank Roger Edwards and an anonymous reviewer for their constructive comments. We would also like to thank Jeffrey Chapman for help in the identification of separate mesocyclones and Greg Harmon and James Meyer for providing us the time and resources to complete this research.
REFERENCES
Atkins, N. T., and Laurent M. S. , 2009: Bow echo mesovorticies. Part II: Their genesis. Mon. Wea. Rev., 137, 1514–1532.
Atkins, N. T., Weisman M. L. , and Wicker L. J. , 1999: The influence of preexisting boundaries on supercell evolution. Mon. Wea. Rev., 127, 2910–2927.
Atkins, N. T., Bouchard C. S. , Przybylinski R. W. , Trapp R. J. , and Schmocker G. , 2005: Damaging surface wind mechanisms within the 10 June 2003 Saint Louis bow echo during BAMEX. Mon. Wea. Rev., 133, 2275–2296.
Benjamin, S. G., and Coauthors, 2002: RUC20—The 20-km version of the Rapid Update Cycle. NOAA Tech. Memo. OAR FSL 28, Forecast Systems Laboratory, 9 pp.
Brooks, H. E., Doswell C. A. III, and Cooper J. , 1994a: On the environments of tornadic and nontornadic mesocyclones. Wea. Forecasting, 9, 606–618.
Brooks, H. E., Doswell C. A. III, and Wilhelmson R. B. , 1994b: The role of midtropospheric winds in the evolution and maintenance of low-level mesocyclones. Mon. Wea. Rev., 122, 126–136.
Bunkers, M. J., 2002: Vertical wind shear associated with left-moving supercells. Wea. Forecasting, 17, 845–855.
Clark, A. J., Schaffer C. J. , Gallus W. A. Jr., and Johnson-O’Mara K. , 2009: Climatology of storm reports relative to upper-level jet streaks. Wea. Forecasting, 24, 1032–1051.
Darkow, G. L., and Fowler M. G. , 1971: Tornado proximity soundings wind analysis. Preprints, Seventh Conf. on Severe Local Storms, Kansas City, MO, Amer. Meteor. Soc., 148–151.
Davies-Jones, R., 1984: Streamwise vorticity: The origin of updraft rotation in supercell storms. J. Atmos. Sci., 41, 2991–3006.
Davies-Jones, R., Burgess D. , and Foster M. , 1990: Test of helicity as a tornado forecast parameter. Preprints, 16th Conf. on Severe Local Storms, Kananaskis Park, AB, Canada, Amer. Meteor. Soc., 558–592.
desJardins, M. L., Brill K. F. , and Schotz S. S. , 1991: GEMPAK5 user’s guide. NASA Tech. Memo. 4260, 232 pp. [Available from NASA, Code NTT-4, Washington, DC 20546-0001.]
Doswell, C. A., III, and Rasmussen E. N. , 1994: The effect of neglecting the virtual temperature correction on CAPE calculations. Wea. Forecasting, 9, 625–629.
Doswell, C. A., III, Edwards R. , Thompson, R. L. , Hart J. A. , and Crosbie K. C. , 2006: A simple and flexible method for ranking severe weather events. Wea. Forecasting, 21, 939–951.
Fawbush, E. J., and Miller R. C. , 1954: The types of air masses in which North American tornadoes form. Bull. Amer. Meteor. Soc., 35, 154–165.
Fujita, T. T., 1971: Proposed characterization of tornadoes and hurricanes by area and intensity. SMRP Research Paper 91, University of Chicago, 42 pp. [Available from Wind Engineering Research Center, Box 41023, Lubbock, TX 79409.]
Funk, T. W., Darmofal K. E. , Kirkpatrick J. D. , DeWald V. L. , Przybylinski R. W. , Schmocker G. K. , and Lin Y.-J. , 1999: Storm reflectivity and mesocyclone evolution associated with the 15 April 1994 squall line over Kentucky and southern Indiana. Wea. Forecasting, 14, 976–993.
Galway, J. G., 1977: Some climatological aspects of tornado outbreaks. Mon. Wea. Rev., 105, 477–484.
Houston, A. L., Thompson R. L. , and Edwards R. , 2008: The optimal bulk wind differential depth and the utility of the upper-tropospheric storm-relative flow for forecasting supercells. Wea. Forecasting, 23, 825–837.
Kis, A. K., and Straka J. M. , 2010: Nocturnal tornado climatology. Wea. Forecasting, 25, 545–561.
Klemp, J. B., and Wilhelmson R. B. , 1978: Simulations of right- and left- moving storms produced through storm splitting. J. Atmos. Sci., 35, 1097–1110.
Lanicci, J. M., and Warner T. T. , 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, 181–197.
Maddox, R. A., 1976: An evaluation of tornado proximity wind and stability data. Mon. Wea. Rev., 104, 133–143.
Markowski, P. M., Rasmussen E. N. , and Straka J. M. , 1998a: The occurrence of tornadoes in supercells interacting with boundaries during VORTEX-95. Wea. Forecasting, 13, 852–859.
Markowski, P. M., Straka J. M. , Rasmussen E. N. , and Blanchard D. O. , 1998b: Variability of storm-relative helicity during VORTEX. Mon. Wea. Rev., 126, 2959–2971.
Markowski, P. M., Straka J. M. , and Rasmussen E. N. , 2002: Direct surface thermodynamic observations within rear-flank downdrafts of nontornadic and tornadic supercells. Mon. Wea. Rev., 130, 1692–1721.
Mitchell, M. J., Arritt R. W. , and Labas K. , 1995: A climatology of the warm season Great Plains low-level jet using wind profiler observations. Wea. Forecasting, 10, 576–591.
Rasmussen, E. N., and Wilhelmson R. B. , 1983: Relationships between storm characteristics and 1200 GMT hodographs, low level shear, and stability. Preprints, 13th Conf. on Severe Local Storms, Tulsa, OK, Amer. Meteor. Soc., J5–J8.
Rasmussen, E. N., and Blanchard D. O. , 1998: A baseline climatology of sounding-derived supercell and tornado forecast parameters. Wea. Forecasting, 13, 1148–1164.
Rasmussen, E. N., Richardson S. , Straka J. M. , Markowski P. M. , and Blanchard D. O. , 2000: The association of significant tornadoes with a baroclinic boundary on 2 June 1995. Mon. Wea. Rev., 128, 174–191.
Rotunno, R., and Klemp J. B. , 1985: On the rotation and propagation of simulated supercell thunderstorms. J. Atmos. Sci., 42, 271–292.
Straka, J. M., Rasmussen E. N. , Davies-Jones R. P. , and Markowski P. M. , 2007: An observational and idealized numerical examination of low-level counter-rotating vortices in the rear flank of supercells. Electron. J. Severe Storms Meteor., 2 (8). [Available online at http://www.ejssm.org/ojs/index.php/ejssm/article/viewFile/32/33.]
Thompson, R. L., Edwards R. , Hart J. A. , Elmore K. L. , and Markowski P. , 2003: Close proximity soundings within supercell environments obtained from the Rapid Update Cycle. Wea. Forecasting, 15, 682–699.
Trapp, R. J., and Weisman M. L. , 2003: Low-level mesovortices within squall lines and bow echoes. Part II: Their genesis and implications. Mon. Wea. Rev., 131, 2804–2823.
Trapp, R. J., Mitchell E. D. , Tipton G. A. , Effertz D. W. , Watson A. I. , Andra D. L. Jr., and Magsig M. A. , 1999: Descending and nondescending tornadic vortex signatures detected by WSR-88Ds. Wea. Forecasting, 14, 625–639.
Uccellini, L. W., and Johnson D. R. , 1979: The coupling of upper and lower tropospheric jet streaks and implications for the development of severe convective storms. Mon. Wea. Rev., 107, 682–703.
Wakimoto, R. M., Murphey H. V. , Davis C. A. , and Atkins N. T. , 2006: High winds generated by bow echoes. Part II: The relationship between the mesovortices and damaging straight-line winds. Mon. Wea. Rev., 134, 2813–2829.
Weisman, M. L., and Klemp J. B. , 1982: The dependence of numerically simulated convective storms on vertical wind shear and buoyancy. Mon. Wea. Rev., 110, 504–520.
Weisman, M. L., and Rotunno R. , 2000: The use of vertical wind shear versus helicity in interpreting supercell dynamics. J. Atmos. Sci., 57, 1452–1472.
Weisman, M. L., and Trapp R. J. , 2003: Low-level mesovortices within squall lines and blow echoes. Part I: Overview and dependence on environmental shear. Mon. Wea. Rev., 131, 2779–2803.
Wilhelmson, R. B., and Klemp J. B. , 1978: A numerical study of storm splitting that leads to long-lived storms. J. Atmos. Sci., 35, 1974–1986.
Data from 7.5° and 16.7° were unavailable from the KFSD radar archive and National Climatic Data Center archive.
Where significant divergence or convergence existed, Vr may be overestimated, since there was no objective technique for removing the divergent portion of the wind when calculating rotation.
All subsequent vertical layers are based on above ground level hereafter.
