1. Introduction
In the midtwentieth century Newton (1950) summarized extant ideas concerning the formation and maintenance of prefrontal squall lines1 in the warm sectors of cyclones east of the Rocky Mountains. Three theories were discussed. The oldest, described by Holzman (1936), Lichtblau (1936), Lloyd (1942), Showalter and Fulks (1943), Crawford (1950) and others, attributed the formation and maintenance of warm sector squall lines to a cold front aloft (CFA). The second theory, proposed by Tepper (1950), was that squall lines are triggered and maintained by gravitational waves that move along temperature inversions in warm sectors. The third theory, which Newton (1950) believed was most likely correct, was “that a mass of rain-cooled air sets up a shallow local cold front in the warm sector, which subsequently releases the thunderstorm energy ahead of the real cold front.” Two-dimensional and 3D cloud modeling simulations of squall lines (e.g., Rotunno et al. 1988; Weisman et al. 1988) show that when a squall line forms in this way it can be self-maintained and long-lived in some environments, without assistance from any externally supplied lifting mechanisms. It should be noted, however, that this theory pertains only to the maintenance of squall lines, not their initiation.
In a summary of current ideas concerning squall line formation in the warm sector of cyclonic storms east of the Rocky Mountains, Ahrens (2000) states that “pre-frontal squall-line thunderstorms of the middle latitudes represent the largest and most severe type of squall line,” and they “may extend for over 1000 km, with huge supercell storms causing severe weather.” However, Ahrens states “there is still debate as to exactly how pre-frontal squall lines form.”
The CFA theory for severe warm sector squall lines had generally fallen out of favor. However, beginning with Locatelli et al. (1989), recent research based on observations and mesoscale model simulations has led to a resurgence of interest in CFAs. Locatelli et al. (1989) proposed that when a cold front overtakes a trough in the lee of the Rocky Mountains, it can move aloft, forming a structure similar to a warm-type occlusion. They also proposed, and subsequent studies (Hobbs et al. 1990; Locatelli et al. 1995; Locatelli and Hobbs 1995; Locatelli et al. 1998; Stoelinga et al. 2000) confirmed, that with appropriate instability east of the lee trough, a CFA can initiate squall lines. In this case, the surface pressure trough, which the squall line precedes, is a lee trough, usually with dryline characteristics, and it is structurally akin to a surface warm occluded front. However, in these situations, the trough is typically analyzed by the National Weather Service (NWS) as a surface cold front. This leads to the incorrect implication that the squall line is a prefrontal or warm-sector squall line when, in fact, the squall line is intimately connected to the leading edge of a deep tropospheric CFA.
Other recent studies have found that CFAs are identifiable features that can spawn severe weather (Neiman et al. 1998; Neiman and Wakimoto 1999; Koch and Siedlarz 1999). Neiman and Wakimoto 1999 state “It is becoming increasingly apparent that these Pacific fronts often do not intersect the ground over the Great Plains but remain aloft associated with their parent upper-level short-wave troughs. Although these fronts are decoupled from the surface, it is important to note that they can still be associated with severe weather and heavy precipitation over the central United States.”
Two CFAs that have been extensively studied through both observations and modeling were associated primarily with what could be described as benign or marginally severe weather. The CFA rainband of 6 March 1986 (Locatelli et al. 1989) was stratiform in nature. The CFA squall line of 8–9 March 1992 (Locatelli et al. 1995; Locatelli et al. 1998; Stoelinga et al. 2000) produced an F0 tornado and ∼2.5 cm-sized hail, but neither high winds nor significant damage was reported along the majority of its length. These two cases were chosen more for the availability of data from field projects than for their severity. The motivation for the present study is to explore whether a CFA played a role in some large outbreaks of severe weather in the central United States.
The Palm Sunday outbreak (11–12 April 1965) and the Super Outbreak (3–4 April 1974) are widely regarded as two of the most extensive and violent outbreaks of tornadoes in the United States in the last half century. Of particular interest, specifically with regard to the role of a CFA in these outbreaks, is that in both the Palm Sunday and Super Outbreak the convection was not collocated with a surface frontal feature, but instead was organized into long (800–2000 km) squall lines. Furthermore, Hoxit and Chappell's (1975) study of the Super Outbreak, and the Fujita et al. (1970) study of the Palm Sunday outbreak, provide evidence of CFAs. In fact, Fujita refers to the role of a forward-tilted cold front (where drier air was found at the surface on the cool side of the front) in the development of the convection.2
Recently, the NCEP–NCAR reanalysis dataset (Kalnay et al. 1996) has made it possible to run accurate mesoscale model simulations of historical weather events that were previously ill-suited for model simulation due to lack of quality analyses for initial and boundary conditions. The NCEP–NCAR Reanalysis dataset provides such analyses by reassimilating all available past data from the period 1957–96 into a modern global model-based data assimilation system. Using these Reanalysis data, and the Pennsylvania State University–National Center for Atmospheric Research fifth-generation Mesoscale Model (MM5), simulations of the Palm Sunday and Super Outbreaks have been completed with considerable fidelity. The model simulations showed that both tornado outbreaks were associated with a CFA squall line. This paper will describe the model results for the Super Outbreak.
It will be shown through both conventional analysis (section 3) and model analysis (section 4) that of the three squall lines in the conventionally analyzed warm sector of the Super Outbreak, the first was associated with a structure resembling an undular bore. The second squall line, which produced the most numerous supercells and damaging tornadoes, was associated with a CFA. The third squall line was not captured with sufficient accuracy by the MM5 model simulations to arrive at any firm conclusions concerning its origin.
2. Overview of the Super Outbreak
The Super Outbreak is well known for the enormous number and intensity of tornadoes it spawned: 148 over a period of 20 h! The tornadoes occurred in thirteen states with a total tornado pathlength of 3241 km. When compared to two other famous tornado outbreaks, the Palm Sunday and Tristate outbreaks (total tornado pathlengths 1372 km and 703 km, respectively), it is clear that the Super Outbreak was truly a phenomenal event. While the Super Outbreak is best known for the tornadoes it spawned, other significant weather occurred during the lifetime of its associated cyclone, which traversed the central and eastern United States from 2 to 5 April 1974. On 2–3 April, eastern Colorado experienced strong winds and snow. At the same time the intensifying storm caused winds in excess of 29 m s−1, and widespread blowing dust, over the state of New Mexico. On 2–4 April, 5–30 cm of snow fell in Nebraska and 26 m s−1 winds caused snow drifts over 2 m in depth.
During the Super Outbreak, the tornadoes were unusually intense, large, and long lasting. This can be seen in a map of the tornado paths coded by intensity (Abbey and Fujita 1983) shown in Fig. 1. Given the large percentage of tornadoes rated F3 or greater on the Fujita damage scale, and the number of hook echoes (Forbes 1978) associated with the individual convective storms containing tornadoes, it is clear that the squall lines in the Super Outbreak contained a significant amount of embedded supercellular convection. Forbes (1975) concluded that during some of the period the three squall lines were really three broken lines of supercells, and that the three squall lines were made up of as many tornadic supercell storms as nontornadic.
Hoxit and Chappell (1975) analyzed extensively the synoptic conditions associated with the Super Outbreak. Synoptic analysis of the Super Outbreak was also done by Agee et al. (1975), who remark “an astonishing aspect of the surface weather … is that the warm sector of the extratropical cyclone was capable of supporting simultaneously three tornado-producing squall lines, which were responsible for nearly all of the tornadoes.” The radar summary (Radar Summary Charts obtained from National Climatic Data Center) and operationally analyzed fronts (obtained from microfilm copies of the North American Surface Charts prepared by the National Meteorological Center, now called NCEP) for 0000 UTC 4 April 1974, superimposed on our isobaric analysis (Fig. 2), confirms that at a time when the three squall lines were simultaneously active, they were located either wholly or partly within the warm sector of the cyclone as analyzed by the NWS. The NWS-analyzed frontal positions are shown since this was the analysis believed to be correct at the time of the storm. In section 3, it will be shown the extent to which this frontal analysis agrees with our analysis of the various surface fields. The middle squall line (labeled 2 in Fig. 2) was clearly the longest band of the three at this time. Tornadoes in progress at 0000 UTC 4 April (Fig. 2) also suggest that squall line 2 was the most severe. This is further confirmed by categorizing all the Super Outbreak tornadoes (not just those on the ground at 0000 UTC 4 April) according to the squall lines with which they were associated (Fig. 3), and noting that squall line 2 was associated with the largest number of tornadoes. Comparison of Figs. 1 and 3 also shows that all of the F5 tornadoes, and most of the F3 and F4 tornadoes, were associated with squall line 2.
3. Conventional analysis
The important characteristics of the operationally analyzed surface frontal features for the Super Outbreak can be seen by comparing them with subjective temperature (Fig. 4a), dewpoint (Fig. 4b), and pressure analyses at 0000 UTC 4 April 1974. An east–west stationary front is well supported by the temperature, dewpoint, pressure, and winds fields (Fig. 2). A primary cold front is analyzed from Illinois to Louisiana. It has a temperature gradient located to its west, but it is also an axis of maximum temperature. The moisture gradient along this cold front (Fig. 4b) suggests that the dryline characteristics of this feature are more robust than the cold frontal characteristics. The cold front analyzed by the NWS is situated along most of its length in the middle of the moisture gradient. In fact, until the time when the secondary cold front (located from Iowa–Texas at 0000 UTC 4 April) caught up with the analyzed primary cold front, the dewpoint temperature gradient across the primary cold front was larger than the temperature gradient across the primary cold front. The NWS-analyzed warm front is not well supported by the temperature, pressure, or moisture fields.
The upper air analyses taken from Hoxit and Chappell (1975) for 0000 UTC 4 April 1974 are shown in Figs. 5a–c. The analyses shown are unmodified, except for the addition of a line marking the leading edge of the baroclinic zone, which was embedded in the short-wave trough at 850, 700, and 500 hPa. These positions have been transposed to Fig. 6, where the positions of the three squall lines and the surface cold front, as analyzed by the NWS, are also shown. This superposition shows clearly that the baroclinic zone was ahead of the position of the surface cold front analyzed by the NWS at this time, at both 700 hPa (from Iowa to Louisiana) and at 500 hPa (from Iowa to Missouri). The southern portion of the 850-hPa baroclinic zone that curves westward through Texas is the upper extension of the secondary cold front, not the primary cold front. The southern half of squall line 2 is also aligned with the furthest forward extent (near 700 hPa) of the baroclinic zone aloft (i.e., the CFA).
Hoxit and Chappell (1975) analyzed vertical cross sections through the Super Outbreak at 0000 UTC 4 April 1974. One of their cross sections cuts through the baroclinic zone at all levels (Fig. 7). Again, the analysis is unmodified, except for the addition of a line marking the leading edge of the baroclinic zone. This leading edge, which represents the forward extent of the Pacific cold front (i.e., the CFA), has a definite tipped-forward structure from the surface position of the analyzed cold front up to around 700 hPa. The CFA is located over Nashville, Tennessee (BNA), and Fig. 6 shows that BNA is located in the center of squall line 2 at this time. To examine the connection between the CFA and squall line 2, and also the developments of the other two squall lines, a mesoscale model simulation was employed and analyzed, as described below.
4. Mesoscale model analysis
a. The numerical model
The mesoscale model MM5 (Grell et al. 1994), was used to simulate the Super Outbreak. The MM5 model is a nonhydrostatic, primitive equation model that uses the terrain-following sigma vertical coordinate, and a rectangular grid on a conformal map projection. Specific physical parameterizations include the Medium-Range Forecast (MRF) model planetary boundary layer parameterization, the Kain–Fritsch (1990) cumulus parameterization, and the explicit cloud microphysical scheme implemented by Reisner et al. (1993). In the present study, the MM5 was run on a 54-km coarse outer domain with an 18-km resolution inner nested domain. The 18-km inner grid covered an area from northern Lake Michigan to southern Louisiana, and from eastern Colorado to eastern Virginia. Both domains had a 30-hPa vertical grid spacing. The model was initialized at 1200 UTC 2 April 1974 with the NCEP–NCAR Reanalysis dataset used for initial and boundary conditions.
b. Comparison of the simulated and observed storm systems
The MM5 model was used to generate surface pressure, wind, and precipitation fields for model hour 36 valid at 0000 UTC 4 April 2000 (Fig. 8). The corresponding surface temperature and dewpoint fields (Fig. 9) were used to mark surface frontal features for this 36-h forecast. When the model-generated frontal features are compared to the frontal features analyzed by the NWS for the same time (Fig. 2), two frontal features are seen to be different. First, the model did not develop the weaker of the two warm fronts seen in the NWS analysis (i.e., the warm front in Indiana, Ohio, and The Virginias shown in Fig. 2). Second, the simulated N–S front along the thermal ridge has been labeled as an occluded front rather than a cold front. The reason for this will be apparent from the remaining model analysis.
A comparison of the analyzed fields and the model-generated wind, temperature, and dewpoint fields (Figs. 4a and 4b) shows that the model accurately captured the main features of the surface wind, temperature, and moisture fields associated with the surface fronts. The model captures the maximum in temperature along the occluded front (i.e., the primary cold front analyzed by the NWS), a temperature decrease behind the occluded and secondary cold front, and the temperature gradient north of the stationary front. The model replicates the dewpoint gradient along the occluded front, and the intrusion of moist air northward ahead of the occluded front. The model also reproduces the southerly winds ahead of the occluded front, the southwesterly winds between the occluded front and the secondary cold front, the westerly winds behind the secondary cold front, and the north-northeasterly winds north of the low pressure center.
There are some differences between the simulation and the observations. The simulated stationary front was ∼320 km further north than observed. The simulated low pressure center was 6 hPa lower than observed and was shifted NW by about 200 km. In the simulation the maximum dewpoint was about 3°C too low near the Gulf of Mexico and 2°C too high just south of the stationary front. The surface winds in the model are almost three times greater than those observed near the low pressure center. We attribute this increase to the stronger pressure gradient associated with the deeper simulated system. The simulation has a greater area west of the low pressure center that has subfreezing temperatures than do the observations.
Comparisons of the model precipitation field (Fig. 8) with the actual precipitation field (Fig. 2) shows that only traces of squall line 1 are in the model precipitation field. In contrast, squall line 2 is well defined in the model precipitation field. Moreover, the model reproduces the locations of squall lines 1 and 2 in the correct relationship to the surface fronts, as well as the correct speed of each of the squall lines. Squall lines 1 and 2 rotated with time from a N–S orientation to a more NE–SW orientation as a consequence of the higher eastward speed at their northern ends. In comparing the speeds of the MM5 model squall lines and the actual squall lines, the speeds perpendicular to, and at the midpoint of, each squall line were measured. Measured this way, the actual speed of squall line 1 was 19 m s−1 and the model speed was also 19 m s−1. The model speed for the northern end of squall line 1 is ∼6 m s−1 slower than the observations. Some of this difference is due to the difficulty in locating the northern end of the simulated squall line 1. Squall line 1 formed at ∼0535 UTC 3 April while in the simulation it formed at ∼0500 UTC. The actual speed of squall line 2 was 10 m s−1, while the model speed was 9 m s−1. Squall line 2 formed at ∼1235 UTC 3 April while in the simulation it formed at ∼1600 UTC. As a consequence of the later start time of squall line 2 in the simulation, the position of the simulated squall line 2 is about 300 km behind its observed position at 0000 UTC 3 April. The model does not reproduce the third squall line, although it does produce general precipitation in the same region, which is located east of the occluded front and west of squall line 2.
While the MM5 model simulations vary from the observations in some respects, these differences do not impact our conclusion that the basic structure of the model storm is similar to that of the observed storm. In particular, since the model captures the pattern of the temperature, wind, and moisture fields of the Super Outbreak, and it accurately reproduces the location and movement of squall lines 1 and 2, the model simulation can be used to study the origin and maintenance of squall lines 1 and 2.
c. Squall line 1
Shown in Fig. 10 are the model-simulated frontal positions and the location of squall line 1, valid at 0800, 1400, and 1800 UTC 3 April. Squall line 1 moves away from the occluded front in an arc, and by 1800 UTC (Fig. 10c) it has begun to break apart into scattered bands.
A series of cross sections in Fig. 11 along the line section AA′ (see Fig. 10 for location of the cross section) reveals the thermal and wind structure associated with the development of squall line 1. By 0800 UTC 3 April (Fig. 11a), the Pacific cold front has already caught up to the surface trough/dryline and formed an occluded front (notice the slight tipped-forward structure of the Pacific cold front as it interacts with the stable layer east of the trough). By 1400 UTC 3 April (Fig. 11b), squall line 1 is clearly associated with an eastward-moving region of rising air (marked by the arrows) that resembles a borelike structure. This feature is borelike for several reasons: 1) the steplike increase in the height of the potential temperature contours at the location of the feature, 2) the front-to-rear relative flow through the feature, and 3) the decrease in the relative front-to-rear flow behind the feature. These characteristics are typical of undular bores (e.g., Simpson 1987). In spite of the coarse model resolution, a suggestion of undulations can be seen behind the initial updraft in Fig. 11a. By 1800 UTC 3 April (Fig. 11c), the undular borelike structure is still apparent, though not as sharp as in its early stage (Fig. 11a). We postulate that this structure was caused by the forward advance of the Pacific cold front into the stable air mass to the east of the lee trough/dryline, and that squall line 1 was driven by the vertical motions associated with the undular borelike structure.
The question arises as to how much of the borelike structure is caused by the associated convection (squall line 1) and how much is attributable to the bore itself. While the convection did eventually develop along the entire length of the borelike structure in the simulation, at the time it first developed the bore did not have convection along its entire length. In places where the bore and convection were collocated (Fig. 11), it is difficult to separate out direct effects of convection from the externally forced lifting and temperature perturbations associated with the bore. Updrafts occurring at 300–500 hPa within the shaded column in Fig. 11b are likely a response to the diabatic heating produced by parameterized convection, as may the smaller-scale fluctuations in the potential temperature field at the leading edge of the borelike structure. However, while convection can produce localized vertical velocity and perturbations in potential temperature, the fact remains that the feature in question represents a narrow transition between a broad region of strong easterly relative flow and vertically lower isentropes to the east, and a broad region of weaker easterly relative flow and vertically higher isentropes to the west, with relative airflow moving through the feature. These features suggest the presence of a borelike structure on a spatial scale significantly larger than that of the squall line itself.
d. Squall line 2
A time series of surface maps depicting the movement of the model-simulated squall line 2 valid for 1600 and 2000 UTC 3 April, 0000 and 0400 UTC 4 April (Fig. 12) shows that squall line 2 develops several hours after the formation of the model-simulated squall line 1, which was also the case for the observed squall lines. Squall line 2 also moves away in an arc from the surface occluded front. It is noteworthy that the arc of squall line 2 moves uninterrupted across the stationary front (as seen in Figs. 12c and 12d). Further diagnosis of the 3D structure of squall line 2 and its environment at model hour 36 (valid at 0000 UTC 4 April) (Fig. 12c) is given below.
The model-simulated 500–950-hPa thickness valid for 0000 UTC 4 April superimposed on the frontal positions and the precipitation field (Fig. 13) shows a thickness ridge in advance of the surface occluded front, and the ridge follows the arc shape of squall line 2. The thickness pattern indicates that the cooling in the atmosphere is in advance of the surface trough (labeled as an occluded front). That is, the model simulations show the Pacific cold front overtaking the lee trough/dryline and moving aloft eastward of the surface position of the lee trough/dryline, much as a cold front would occlude with a warm front to form a warm occlusion.
Maps of potential temperature, the horizontal gradient of potential temperature (hereafter, thermal gradient), and negative horizontal advection of potential temperature (hereafter, cold advection) at various pressure levels, are of tremendous utility in locating the model-simulated position of the Pacific cold front relative to the surface occluded front. For the present study, such plots were generated every 100 hPa from 900 to 500 hPa. However, only the plots at 800 hPa at 0000 UTC 4 April are shown here to demonstrate how the Pacific cold front was located at a particular pressure level.
Several features can be seen in the potential temperature distribution and the thermal gradient (Fig. 14a), as well as the cold advection field (Fig. 14b). The arc of enhanced thermal gradient running from Kansas through Minnesota to the Great Lakes shows up as cold advection only in Nebraska and Kansas, where the short segment of a surface cold front that is an extension of the baroclinic zone of the stationary front is located. North of Nebraska the arc of the enhanced thermal gradient corresponds to the surface stationary front and is therefore not a region of cold advection at 800 hPa. Regions of enhanced thermal gradient and cold advection at 800 hPa associated with the surface secondary cold front are seen clearly from Missouri through Oklahoma. A third arc-shaped region of both enhanced cold advection and thermal gradient is marked with a heavy dashed line near the center of Fig. 14. This region is associated with the baroclinic zone of the Pacific cold front. The leading edge of the Pacific cold front was identified from similar maps at the 500-, 600-, 700-, and 900-hPa levels (not shown). The positions of the Pacific cold front at these pressure levels are superimposed in Fig. 15. In Louisiana, the Pacific cold front tips backward (westward) with height, similar to a conventional cold front. From Louisiana northward, the Pacific cold front above 900 hPa protrudes ahead of its position at 900 hPa and ahead of the surface occluded front. If the CFA is defined as the farthest forward extent of the Pacific cold front regardless of vertical level, it can be seen from Fig. 15 that the CFA is coincident with the position of squall line 2.
The relationship of the CFA to squall line 2 can be further diagnosed by examining selected fields along a vertical cross section through the line BB′ shown in Fig. 15. The baroclinic zone of the Pacific cold front is evident in the potential temperature and the thermal gradient (Fig. 16a). The leading edge of the Pacific cold front is marked by a heavy dashed line in Fig. 16a. As was deduced from the horizontal positions of the Pacific cold front at different pressure levels (Fig. 15), the Pacific cold front clearly tips forward (eastward) in the lower levels of the atmosphere. In a squall-line-relative (and CFA-relative) sense, the air below 800 hPa that approaches from the right of the cross section shown in Fig. 16b moves toward the Pacific cold front at a relative speed of as much as 20 m s−1. It then encounters strong convergence beneath the nose of the Pacific cold front and rises into the updraft associated with squall line 2. Some of the air near the surface continues westward until it meets the surface occluded front where convergence at the surface is maximized. This is consistent with the largest surface wind shift being at this location. However, in three dimensions, the strongest convergence is directly beneath the nose of the CFA, about 50 hPa above the surface. Associated with this convergence are the main updraft and squall line, which are coincident with the CFA. This relationship between the CFA, strong convergence, and the squall line updraft, exists along the length of the simulated squall line 2; this is why the CFA and squall line 2 are collocated in Fig. 15. Therefore, we propose that squall line 2, which produced the strongest and most numerous tornadoes of the three squall lines in the Super Outbreak of 2–5 April 1974, was directly connected to lifting associated with a CFA.
5. Sensitivity to surface heat fluxes and evaporative cooling
Because the structural difference between conventional backward-tilting cold fronts and nonconventional forward-tilting cold fronts is generally confined to the lowest few kilometers of the atmosphere, it is reasonable to hypothesize that physical processes that affect the thermal structure of the boundary layer, such as surface heat fluxes or the formation of a cold pool through evaporation of precipitation, might have a significant impact on the type of frontal tilt that develops as a Pacific cold front moves into the air mass east of the lee trough/dryline. To test this hypothesis in the case of the Super Outbreak, sensitivity tests were run using MM5. In these sensitivity tests either surface heat fluxes, or evaporative cooling, or both, were switched off in the model simulation starting at forecast hour 26 (valid at 1400 UTC 3 April 1974, or 0800 local standard time). This time was chosen since it is close to the “node” in the diurnal cycle between the typical nocturnal minimum and afternoon maximum in temperature. In the tests without evaporative cooling, the cooling was removed from both the bulk microphysical scheme and the cumulus parameterization. In the following discussion, the full-physics model simulation used in section 4 will be referred to as the control simulation.
Figure 17a shows the difference in potential temperature (averaged over the lowest 0.5 km) between the control simulation and the no-surface-flux sensitivity test at forecast hour 30 (valid for 1800 UTC 3 April 1974). The model-simulated cloud is also shown, based on the 0.0025 g (kg)−1 contour of vertically averaged cloud water. The surface heat fluxes tended to heat the boundary layer away from cloudy areas due to the larger insolation. On the other hand, it can be seen from Fig. 17b that evaporative cooling tended to cool the boundary layer beneath the cloudy areas, due to precipitation in these regions. The control run was colder than the sensitivity run only in cloud regions.
To see what difference the lack of surface heating in the cloud-free areas, and the lack of evaporative cooling in the cloudy areas, does to the shape of the Pacific cold front in the vertical, a cross section along CC′ in Figs. 17a and 17b was constructed from the sensitivity model run with no surface fluxes and no evaporative cooling (Fig. 18a). This cross section shows potential temperature and thermal gradient, wind in the cross section relative to the speed of squall line 2, and upward vertical velocity fields. Cross section CC′ is the same length, parallel to, and about 100 km north of the position of cross section BB′ (Fig. 16). The position of cross section CC′ was moved northward since, in the “no-surface-flux, no-evaporative-cooling” model sensitivity run, squall line 2 does not extend quite as far south as it does in the control run. Comparison of Figs. 16a,b and Fig. 18a shows that without surface heating behind the Pacific cold front (in the cloud-free region), and without evaporative cooling ahead of it (in the cloudy region), a more conventional cold front is formed that is much less tipped forward. The difference in potential temperature between the control run and the no-surface-flux, no-evaporative-cooling model run for cross section CC′ is shown in Fig. 18b, where the Pacific cold front is marked by the heavy dashed line. The control run was about 2–3 K warmer than the sensitivity run in the boundary layer behind the Pacific cold front, and about 1–2 K colder than the sensitivity run in the boundary layer ahead of it. The combination of the heating in the cloud-free air behind the Pacific cold front, and the evaporative cooling in the cloudy region ahead of the Pacific cold front, tended to lessen the thermal gradient across the Pacific cold front as it approached from the west. Thus, to some extent, the occluded-like structure was directly caused by the diabatic effects of surface heating and evaporative cooling on that day.
Comparison of Figs. 18a and 16a shows that, in the boundary layer and in the cloud-free region behind the Pacific cold front, the lapse rate is still close to neutrally stable, even after 10 h without surface heating. To test the robustness of this result, the surface fluxes were turned off at model hour 2, a full 34 h before 0000 UTC 4 April (model hour 36). The resulting cold air mass behind the Pacific cold front at model hour 36 (not shown) was still nearly neutrally stable. This suggests that the lower stability of the air behind the cold front was a cumulative result of several days of enhanced surface heating, as the air in the cloud-free region behind the Pacific cold front traversed the Rocky Mountains.
6. Discussion
a. Squall line 1 (the undular bore)
Additional evidence that the first squall line in the Super Outbreak was caused by a wavelike phenomenon can be found in the study of Miller and Sanders (1980). They analyzed data from more than 200 surface stations during the Super Outbreak, and concluded that small-scale pressure fluctuations, which were superimposed on synoptic-scale pressure falls, were highly suggestive of gravity wave activity. They identified 10 wave packets of considerable spatial and temporal continuity, which were led by a wave (which they called A) that had the strongest pressure signal and was the only wave packet that consistently moved with a region of organized convection (which they called alpha).
Figure 19, derived from two figures of Miller and Sanders (1980), shows the positions of their convective region alpha (which corresponds to our squall line 1), the location of their wave packet A, and the position of lines of echoes from the NWS radar summary maps. It can be seen that for over 12 h, wave packet A was coincident with or slightly ahead of lines of echoes within the convective region alpha (squall line 1). In fact, Miller and Sanders conclude that “no other packet had such a simple and long-lasting relationship to a convective line.” Pressure traces taken from Miller and Sanders for the three stations shown in Fig. 20 (Fort Wayne, IN; Louisville, KY; and Birmingham, AL) clearly show the pressure signal associated with wave packet A. Miller and Sanders marked the mesoscale low pressure signals associated with wave packet A on the pressure traces. The most prominent pressure feature of wave packet A is the pressure jump of ∼2–3 hPa that follows the start of the wave packet. Miller and Sanders state that ∼69% of precipitation episodes associated with this wave packet occurred on rising pressure or near its mesoscale pressure ridge. From Fig. 11 it can be seen that there is an abrupt rise in the potential temperature lines as the model squall line 1 moves over a fixed surface location, which would produce a corresponding abrupt increase in the surface pressure as shown for the wave packet A in Fig. 20.
There is considerable evidence in the literature that wavelike phenomena can initiate convection. Uccellini (1975) described a case of severe convection that was initiated north of a stationary front in Iowa and Wisconsin by synoptic-scale gravity waves. Shreffler and Binkowski (1981) reported on strong thunderstorm activity over Iowa that was the source of “pressure jump lines” that moved SSE causing numerous rain showers and thunder. Stobie et al. (1983) described convection north of a Midwest stationary front in which convective cells moved with the triggering gravity wave for a significant portion of the wave track. Convective storms near an Oklahoma dryline, which were apparently triggered by a gravity wavelike disturbance, were studied by Koch and McCarthy (1982). Koch et al. (1988) described an intense mesoscale convective complex that resulted in part from a high degree of organization and periodic renewal of deep convection provided by a gravity wave. Ferretti et al. (1988) described convective systems that were closely linked to gravity waves, although not in a consistent manner. Schneider (1990) determined that large-scale gravity waves, which propagated northward of a low pressure center in a region of general snowfall, were associated with winds in excess of 30 m s−1, cloud-to-ground lightning, and an increase in snowfall rate. Koch and Siedlarz (1999) also found evidence of squall lines associated with gravity waves. Ralph et al. (1999) suggested that gravity waves triggered a short-lived line of deep convection ahead of a cold front.
Abdullah (1955) described a solitary wave (Christie et al. 1978) traveling north of a stationary front that produced some scattered showers. He speculated that such a wave might have produced severe weather if conditions had been more unstable. Ramamurthy et al. (1990) reported on very-large-amplitude solitary waves that propagated over 1000 km before dissipating; distinct, narrow precipitation bands accompanied the waves over some of this distance. Rottman et al. (1992) speculated that a solitary wave, generated by the explosive convective development of a thunderstorm over the eastern slopes of the Rocky Mountains, triggered a second thunderstorm further east over the high plains.
Carbone et al. (1990) reported on a gravity current outflow from a mesoscale convective system in Nebraska that, after reaching a dryline, initiated an internal undular bore east of the dryline. This undular bore continued into Kansas where it triggered a ∼300 km long squall line. Karyampudi et al. (1995) determined that a Pacific cold front was the source of the density current that initiated an internal undular bore that subsequently moved eastward where it intercepted a warm front and dryline to produce severe weather over Nebraska and Kansas.
The squall line analyzed by Tepper (1950) was the type referred to by Ahrens (2000) as a prefrontal or warm-sector squall line, which are typically 1000 km or greater in length. Other than Tepper's case, none of the studies previously discussed attribute the formation of such an extensive squall line to any of the three pressure-jump-forming mechanisms (gravity waves, solitary waves, or undular bores). Therefore, squall line 1 in the Super Outbreak may be the first documented case of an extensive warm-sector squall line produced by a pressure jump, which in this case was most likely an internal undular bore.
Neither the model resolution in the present study, nor the hourly resolution of the plots of the station pressures shown by Miller and Sanders, are fine enough to resolve individual undulations of the borelike structure associated with squall line 2. However Clarke (1989) was able to simulate an undular bore with a 20-km horizontal model grid spacing with results he called “surprisingly good.” This gives credence to our assertion that our model simulation was able to capture the gross features of an undular bore. A combination of the two independent studies (our modeling results and Miller and Sander's work), and the evidence from other research that wavelike phenomenon can initiate convection, supports the hypothesis that squall line 1 was generated by an internal undular bore.
b. Squall line 2
Squall line 2 contained the majority of tornadoes, and also the majority of violent tornadoes (rated F4 or F5), in the Super Outbreak. It has been shown that the second squall line in the Super Outbreak was organized by a CFA. Thus, a CFA is not only important for organizing convection in general, but it can initiate and organize extremely destructive and extensive (1000 km or longer) squall lines that may contain violent tornado-producing supercells.
The frontal lifting associated with the Pacific cold front occurs most strongly at the height and horizontal position of the CFA. Quasigeostrophic (QG) diagnostics, as embodied in jet-streak dynamics, Q-vector diagnosis, or the “omega equation,” are frequently used to understand the lifting associated with baroclinic processes in midlatitudes. However, the spatial scale at which QG dynamics is valid (∼1000 km) is much larger than the spatial scale of interest here. A CFA is the result of a Pacific cold frontal surface that is forward-tilted in the lower troposphere, with a leading edge aloft that extends ahead of its associated surface trough by a few 10s to a few 100 km. This small region of forward-tilt of the cold frontal surface (as compared to the cyclone and cold front as a whole) is scarcely discernible in a properly filtered QG analysis of a CFA (as seen, for example, in Stoelinga et al. 2000). In the QG framework there is little difference between a cold front that has a CFA and a standard cold front—the same rules of QG theory still apply. If the large-scale flow is frontogenetical, which it generally is when a cold air mass displaces a warm air mass, a thermally direct circulation must occur. Stated in another way, the lifting associated with the CFA is not a mystery that needs to be unlocked with a detailed QG analysis, just as the lifting associated with an active conventional cold front is not mysterious. The important point is that some Pacific cold fronts have a unique structure (i.e., a forward-tilt in the lower troposphere and an associated CFA) that leads to the main frontal lifting occurring ahead of the apparent surface location of the front, and convective bands that form from such lifting are essentially frontal bands, not warm-sector bands.
Although none of the previous analyses of the Super Outbreak attributed the formation of what we have referred to as squall line 2 to a CFA, Danielsen (1975a,b) ascribed the convection associated with the “now famous storm of 3 April 1974” (the Super Outbreak) to a “negative sloping cold front.” Danielsen's concept of a negative sloping cold front over the Great Plains can be seen in the reproduction of a figure from his work (Fig. 21). Danielsen states that squall lines can form behind the leading edge of a negative sloping cold front, which can be 150–250 km in advance of the surface front, which, because of strong contrasts in moisture at the surface front, is often described as a dryline. This is an apt description of the CFA described in our analysis of the Super Outbreak.3
c. Thickness patterns and the CFA
Hobbs et al. (1990) proposed that, in addition to the two types of squall lines (prefrontal and ordinary) classified by Cotton and Anthes (1989), there is a third type of squall line, namely, the CFA squall line. The findings presented in this paper, and also by Locatelli et al. (1998) and Stoelinga et al. (2000), confirm this suggestion. However, although the existence of such squall lines is now well documented, it is generally difficult to locate a CFA in the conventional data that are available in real time. This is because the standard rawinsonde network is generally too coarse to identify the location of the CFA. Furthermore, upper-level analyses of observations and forecast model outputs are typically made available only at standard pressure levels (850, 700, 500 hPa, etc.), leaving significant gaps in information on the vertical thermal structure of the lower troposphere. For example, in the Super Outbreak storm the “nose” of the CFA was located (in the model simulations) at ∼800 hPa, which would not be discernible from analyses on standard pressure levels. A definitive analysis of the horizontal and vertical location of a CFA requires examination of lower-tropospheric horizontal temperature data at a minimum of 100-hPa vertical increments, as well as vertical cross sections specifically chosen to cut through the Pacific cold frontal surface (such as the horizontal and vertical cross sections shown in Figs. 14–16).
There is one field, however, that is traditionally included in standard model outputs that has proven useful in locating the position of a CFA. This is the 1000–500-hPa thickness field. Lumb (1950) recommended using the thickness field to ascertain the position of upper fronts when they are ahead of surface fronts. Hobbs et al. (1990) suggested that thickness and sea level pressure charts would show a thickness “front” ahead of a surface pressure trough at the location of the CFA. This is the case for the Super Outbreak, in which the model-simulated thickness field (for the 950–500-hPa layer in this case) show a thickness front in advance of the surface trough (Fig. 13). This thickness front, while not in the exact location of squall line 2, has the same arc shape, and is within 100 km of the position of squall line 2. We conclude that the thickness field can provide a quick way of assessing whether a storm east of the Rocky Mountains is likely to contain a CFA.
d. Surface heat flux and evaporative cooling
There can be either increased solar heating ahead of or behind a cold front, depending on which side of the front has greater cloud cover. Also, there can be evaporational cooling that is either stronger ahead of or behind a cold front depending on the distribution of precipitation across the front. All of these processes have been shown to have an effect on the horizontal temperature gradient across a cold front.
Koch (1984) noted that a low overcast cloud deck on the cold side of a cold front, and scattered clouds and clear skies on the warm side of a cold front, enhanced a deep frontal updraft and led to a severe squall line. Dorian et al. (1988) concluded that differential heating across a cold front can be a significant factor in strengthening the front and in the subsequent formation of convection along the front. Businger et al. (1991) found enhanced heating on the warm side of a cold front, which strengthened frontogenesis, as did Sanders (1999). Numerical calculations by Segal et al. (1993) suggest that in summer shading on the cold side of a cold front may enhance the average thermal contrast across the front by as much as 5°C. Numerical studies by Koch et al. (1995, 1997) support these general conclusions.
Differential evaporative cooling across a cold front has been shown to increase the temperature gradient across the front. Oliver and Holzworth (1953) found that warm-season precipitation falling into dry air immediately behind a cold front can cause sudden accelerations of the cold front. Thorpe and Clough (1991) showed that the temperature deficit behind cold fronts is enhanced by diabatic processes, such as the evaporation of snow. Gallus and Segal (1999) found that cloud shading and precipitation evaporation played comparable roles in increasing the horizontal temperature gradient across the front by maintaining colder air in the cold sector behind the cold front. Huang and Emanuel (1991) found through numerical simulations that evaporation of rain as it falls into the drier air on the colder side of a frontal zone can significantly increase the rate of frontogenesis.
While studies such as those described above have shown that differential diabatic processes (either surface heating or evaporative cooling) across a cold front can increase the strength of the front, other studies have focused on situations similar to that of the present study, namely, those in which diabatic processes acted to decrease the horizontal temperature gradient across a cold front. Sanders (1955) made two important observations that are relevant to the present paper in connection with a cold front in the Midwest. The first was that the air behind the cold front was heated by the surface of the earth at a rate of a few °C per day. The second was that, as a consequence of this heating, the vertical lapse rate was close to dry adiabatic on the cold side of the front. In contrast, the warm side of the front (i.e., the warm sector) was not heated as rapidly, resulting in a much more stable lapse rate. Danielsen (1975a) broadened these observations to include an explanation for severe dust storms in the U.S. Southwest and the high plains. He proposed that the arid soil of the high plains in the southwestern United States was strongly heated by solar radiation. This heating, which can occur in the clear air behind a Pacific cold front as it traverses the Rocky Mountains, can produce a deep mixed layer extending from just above the surface to 10 000–17 000 ft. (3–5 km) MSL. This well-mixed layer allows a downward transport of momentum causing strong surface winds and blowing dust. In section 1 of the present paper it was noted that the intensifying Super Outbreak caused winds in excess of 29 m s−1 and blowing dust over the entire state of New Mexico. This blowing dust occurred in dry, well-mixed air behind the Pacific cold front (Hoxit and Chappell 1975). The sounding at Midland, Texas, for 0000 UTC 3 April 1974 (when Midland was located ∼100 km west of the Pacific cold front) indicates a deep mixed layer up to 480 hPa. In section 5 it was shown that model simulations with surface heat fluxes withheld suggested that the lower-tropospheric stability of the air mass behind the Pacific cold front was a cumulative result of several days of enhanced surface heating in the cloud-free air behind the Pacific cold front as it traversed the Rocky Mountains. Even with one day of surface heating withheld, the air behind the Pacific cold front remained close to well mixed and still maintained a slightly forward tilt in the lower troposphere.
Other researchers have commented on the importance of evaporative cooling ahead of a cold front in decreasing the temperature gradient across the front. Katzfey and Ryan (1997) used a mesoscale model to show that the movement of a cold front was retarded through the creation of a prefrontal cold dome produced by the evaporation of precipitation into the air ahead of the front. Neiman et al. (1998) found that a rain-cooled layer ∼500 m deep acted to decouple a Pacific cold front from the surface. As a result, the passage of the Pacific cold front over Arkansas was not detected in the surface temperature and moisture fields, but was evident in the pressure field. We found a similar result for the Super Outbreak, in which a pool of air cooled by evaporation retarded the forward movement of the Pacific cold front in the lower troposphere. Even though the precipitation was generated at the frontal boundary, the evaporatively cooled downdrafts did not become incorporated into the Pacific cold air mass; rather, they generated a separate mesoscale air mass that remained distinct from, and retarded the progression of, the Pacific cold air mass.
The previous discussion has focused on the effects of diabatic processes on frontal structures. It is also of interest to examine the effect of diabatic processes on the development of convection. Surface heating is generally thought to aid the development of convection by destabilizing the lower troposphere. For example, Hoxit and Chappell (1975) attributed the record temperatures over portions of eastern Kentucky and western West Virginia, and the resulting destabilization of the lapse rate, to strong surface heating in the warm sector during the Super Outbreak. Evaporative cooling is generally thought to be important for producing cold pools that provide the continuous lifting necessary to trigger new convective cells and thereby maintain long-lived convective systems (Rotunno et al. 1988). However, in a study of a CFA-squall line in the U.S. Midwest, Locatelli et al. (1998) found that the formation and propagation of the squall line was not dependent on a cold pool. Instead they showed that the squall line was maintained by lifting associated with a CFA. The results of the sensitivity studies performed in this paper support the view that the lifting associated with a CFA is sufficient to develop and maintain organized convection, even with surface heating and evaporative cooling withheld from the model simulations.
7. Conclusions
Based on observational analyses, a control model simulation, and model sensitivity studies of the Super Outbreak of 2–5 April 1974, we have arrived at the following main conclusions:
There is strong evidence that the first squall line (referred to as squall line 1 in this paper) in the Super Outbreak was driven by lifting at the leading edge of an undular bore. The borelike feature in the model simulation resulted from the forward advance of a Pacific cold front into the stable air mass to the east of a dry line/lee trough.
The second squall line within the Super Outbreak (referred to as squall line 2 in this paper), which produced the most numerous and violent tornadoes, was directly connected to the lifting associated with the furthest forward protrusion of a Pacific cold front (i.e., a CFA) as it occluded with a lee trough/dryline.
The diabatic effects of evaporative cooling at the Pacific cold front, and the daytime surface heating behind the Pacific cold front, were important factors in the formation of the occluded structure. These effects combined to lessen the horizontal temperature gradient across the lower portions of the Pacific cold front.
The model simulations produced and maintained squall line 2 without the benefit of surface heating and evaporational cooling.
Acknowledgments
This research was supported by grants ATM-9632580 and ATM-9908446 from the National Science Foundation (NSF), Division of Atmospheric Sciences, Mesoscale Dynamic Meteorology Program. The MM5 model code and computer resources were provided by NCAR, which is sponsored by the NSF.
REFERENCES
Abbey, R. F., and T. T. Fujita, 1983: Tornadoes: The tornado outbreak of 3–4 April 1974. The Thunderstorm in Human Affairs, E. Kessler, Ed., University of Oklahoma Press, 37–66.
Abdullah, A. J., 1955: The atmospheric solitary wave. Bull. Amer. Meteor. Soc., 36 , 511–518.
Agee, E., C. Church, C. Morris, and J. Snow, 1975: Some synoptic aspects and dynamic features of vortices associated with the tornado outbreak of 3 April 1974. Mon. Wea. Rev., 103 , 318–333.
Ahrens, C. D., 2000: Meteorology Today. 6th ed. Brooks/Cole, 528 pp.
Businger, S., W. H. Bauman, and G. F. Watson, 1991: The development of the Piedmont front and associated outbreak of severe weather on 13 March 1986. Mon. Wea. Rev., 119 , 2224–2251.
Carbone, R. E., J. W. Conway, N. A. Crook, and M. W. Moncrieff, 1990: The generation and propagation of a nocturnal squall line. Part I: Observations and implications for mesoscale predictability. Mon. Wea. Rev., 118 , 26–49.
Christie, D. R., K. J. Muirhead, and A. L. Hales, 1978: On solitary waves in the atmosphere. J. Atmos. Sci., 35 , 805–825.
Clarke, R. H., 1989: Mortality of morning glories. Aust. Meteor. Mag., 37 , 167–172.
Cotton, W., and R. Anthes, 1989: Cloud and Storm Dynamics. Academic Press, 630–657.
Crawford, M. E., 1950: A synoptic study of squall lines. Bull. Amer. Meteor. Soc., 31 , 351–357.
Danielsen, E. F., 1975a: The relationship between severe weather, major dust storms and rapid large-scale cyclogenesis, Parts I and II. Subsynoptic Extratropical Weather Systems: Observations, Analysis, Modeling, and Prediction, Notes From a Colloquium, Summer 1974. National Center for Atmospheric Research, 215–241.
Danielsen, E. F., . 1975b: Generation and triggering of severe convective storms by large-scale motions. Open SESAME: Proc. of the Opening Meeting, Boulder, CO, NOAA Environmental Research Laboratories, 165–185.
Dorian, P. B., S. E. Koch, and W. C. Skillman, 1988: The relationship between satellite-inferred frontogenesis and squall line formation. Wea. Forecasting, 3 , 319–342.
Ferretti, R., F. Einaudi, and L. W. Uccellini, 1988: Wave disturbances associated with the Red River valley severe weather outbreak of 10–11 April 1979. Meteor. Atmos. Phys., 39 , 132–168.
Forbes, G. S., 1975: Relationship between tornadoes and hook echoes on April 3, 1974. Preprints, Ninth Conf. on Severe Local Storms, Norman, OK, Amer. Meteor. Soc., 280–285.
Forbes, G. S., . 1978: Three scales of motion associated with tornadoes. U.S. Nuclear Regulatory Commission Tech. Rep. NUREG/CR 0363, 359 pp.
Fujita, T. T., D. L. Bradbury, and C. F. van Thullenar, 1970: Palm Sunday tornadoes of April 11, 1965. Mon. Wea. Rev., 98 , 29–69.
Gallus, W. A., and M. Segal, 1999: Diabatic effects on late-winter cold front evolution: Conceptual and numerical model evaluations. Mon. Wea. Rev., 127 , 1518–1537.
Glickman, T. S., Ed.,. 2000: Glossary of Meteorology. 2d ed. Amer. Meteor. Soc., 855 pp.
Grell, G. A., J. Dudhia, and D. R. Stauffer, 1994: A description of the Fifth Generation Penn State/NCAR Mesoscale Model (MM5). NCAR Tech. Note NCAR/TN-398+STR, 121 pp. [Available from National Center for Atmospheric Research, Boulder, CO 80301-3000.].
Hobbs, P. V., J. D. Locatelli, and J. E. Martin, 1990: Cold fronts aloft and the forecasting of precipitation and severe weather east of the Rocky Mountains. Wea. Forecasting, 5 , 613–626.
Holzman, B., 1936: Synoptic determination and forecasting significance of cold fronts. Mon. Wea. Rev., 64 , 400–413.
Hoxit, L. R., and C. F. Chappell, 1975: Tornado outbreak of April 3–4, 1974: Synoptic analysis. NOAA Tech. Rep. ERL 338-APCL 37, 48 pp.
Huang, H. C., and K. A. Emanuel, 1991: The effects of evaporation on frontal circulations. J. Atmos. Sci., 48 , 619–628.
Huschke, R. E., Ed.,. 1959: Glossary of Meteorology. Amer. Meteor. Soc., 638 pp.
Kain, J. S., and J. M. Fritsch, 1990: A one-dimensional entrainment/detrainment plume model and its application in convective parameterization. J. Atmos. Sci., 47 , 2784–2802.
Kalnay, E., and Coauthors. 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77 , 437–471.
Karyampudi, V. M., S. E. Koch, C. Chen, J. W. Rottman, and M. L. Kaplan, 1995: The Influence of the Rocky Mountains on the 13–14 April 1986 Severe Weather Outbreak. Part II: Evolution of a prefrontal bore and its role in triggering a squall line. Mon. Wea. Rev., 123 , 1423–1446.
Katzfey, J. J., and B. F. Ryan, 1997: Modification of the thermodynamic structure of the lower troposphere by the evaporation of precipitation: A GEWEX cloud system study. Mon. Wea. Rev., 125 , 1431–1446.
Koch, S. E., 1984: The role of an apparent mesoscale frontogenetic circulation in squall line initiation. Mon. Wea. Rev., 112 , 2090–2111.
Koch, S. E., and J. McCarthy, 1982: The evolution of an Oklahoma dryline. Part II: Boundary-layer forcing of mesoscale systems. J. Atmos. Sci., 39 , 237–257.
Koch, S. E., and L. M. Siedlarz, 1999: Mesoscale gravity waves and their environment in the central United States during STORM-FEST. Mon. Wea. Rev., 127 , 2854–2879.
Koch, S. E., R. E. Golus, and P. B. Dorian, 1988: A mesoscale gravity wave event observed during CCOPE. Part II: Interactions between mesoscale convective systems and the antecedent waves. Mon. Wea. Rev., 116 , 2545–2569.
Koch, S. E., J. T. McQueen, and V. M. Karyampudi, 1995: A numerical study of the effects of differential cloud cover on cold frontal structure and dynamics. J. Atmos. Sci., 52 , 937–964.
Koch, S. E., A. Aksakal, and J. T. McQueen, 1997: The influence of mesoscale humidity and evapotranspiration fields on a model forecast of a cold-frontal squall line. Mon. Wea. Rev., 125 , 384–409.
Lichtblau, S., 1936: Upper cold fronts in North America. Mon. Wea. Rev., 64 , 414–425.
Lloyd, J. R., 1942: The development and trajectories of tornadoes. Mon. Wea. Rev., 70 , 65–75.
Locatelli, J. D., and P. V. Hobbs, 1995: A world record rainfall rate at Holt, Missouri: Was it due to cold frontogenesis aloft? Wea. Forecasting, 10 , 779–785.
Locatelli, J. D., J. M. Sienkiewicz, and P. V. Hobbs, 1989: Organization and structure of clouds and precipitation on the mid-Atlantic coast of the United States. Part I: Synoptic evolution of a frontal system from the Rockies to the Atlantic coast. J. Atmos. Sci., 46 , 1327–1348.
Locatelli, J. D., J. E. Martin, J. A. Castle, and P. V. Hobbs, 1995: Structure and evolution of winter cyclones in the central United States and their effects on the distribution of precipitation. Part III: The development of a squall line associated with weak cold frontogenesis aloft. Mon. Wea. Rev., 123 , 2641–2662.
Locatelli, J. D., M. T. Stoelinga, and P. V. Hobbs, 1998: Structure and evolution of winter cyclones in the central United States and their effects on the distribution of precipitation. Part V: Thermodynamic and dual-Doppler radar analysis of a squall line associated with a cold front aloft. Mon. Wea. Rev., 126 , 860–875.
Lumb, F. E., 1950: Upper frontal analysis in the Mediterranean. Meteor. Mag., 79 , 191–198.
Miller, D. A., and F. Sanders, 1980: Mesoscale conditions for the severe convection of 3 April 1974 in the east-central United States. J. Atmos. Sci., 37 , 1041–1055.
Neiman, P. J., and R. M. Wakimoto, 1999: The interaction of a Pacific cold front with shallow masses east of the Rocky Mountains. Mon. Wea. Rev., 127 , 2102–2127.
Neiman, P. J., F. M. Ralph, M. A. Shapiro, B. F. Smull, and D. Johnson, 1998: An observational study of fronts and frontal mergers over the continental United States. Mon. Wea. Rev., 126 , 2521–2554.
Newton, C. W., 1950: Structure and mechanisms of the prefrontal squall line. J. Meteor., 7 , 210–222.
Oliver, V. J., and G. C. Holzworth, 1953: Some effects of the evaporation of widespread precipitation on the production of fronts and on changes in frontal slopes and motions. Mon. Wea. Rev., 81 , 141–151.
Ralph, F. M., P. J. Neiman, and T. L. Keller, 1999: Deep-tropospheric gravity waves created by leeside cold fronts. J. Atmos. Sci., 56 , 2986–3009.
Ramamurthy, M. K., B. P. Collins, R. M. Rauber, and P. C. Kennedy, 1990: Evidence of very-large-amplitude solitary waves in the atmosphere. Nature, 348 , 314–317.
Reisner, J., R. T. Bruintjes, and R. M. Rasmussen, 1993: Preliminary comparisons between MM5 NCAR/Penn State model generated icing forecasts and observations. Preprints, Fifth Int. Conf. on Aviation Weather Systems, Vienna, VA, Amer. Meteor. Soc., 65–59.
Rottman, J. W., F. Einaudi, S. E. Koch, and W. L. Clark, 1992: A case study of penetrative convection and gravity waves over the PROFS mesonetwork on 23 July 1983. Meteor. Atmos. Phys., 47 , 205–227.
Rotunno, R., J. B. Klemp, and M. L. Weisman, 1988: A theory for strong, long-lived squall lines. J. Atmos. Sci., 45 , 463–485.
Sanders, F., 1955: An investigation of the structure and dynamics of an intense surface frontal zone. J. Meteor., 12 , 542–552.
Sanders, F., . 1999: A short-lived cold front in the southwestern United States. Mon. Wea. Rev., 127 , 2395–2403.
Schneider, R. S., 1990: Large-amplitude mesoscale wave disturbances within the intense Midwest extratropical cyclone of 15 December 1987. Wea. Forecasting, 5 , 533–558.
Segal, M., W. L. Physick, J. E. Heim, and R. W. Arritt, 1993: The enhancement of cold-front temperature contrast by differential cloud cover. Mon. Wea. Rev., 121 , 867–873.
Showalter, A. K., and J. R. Fulks, 1943: Preliminary report on tornadoes. U.S. Department of Commerce, Weather Bureau, War Advisory Council on Meteorology Restricted Rep. 1110, 162 pp.
Shreffler, J. H., and F. S. Binkowski, 1981: Observations of pressure jump lines in the Midwest, 10–12 August 1976. Mon. Wea. Rev., 109 , 1713–1725.
Simpson, J. E., 1987: Gravity Currents in the Environment and the Laboratory. Ellis Horwood Limited, 244 pp.
Stobie, J. G., F. Einaudi, and L. W. Uccellini, 1983: A case study of gravity waves–convective storms interaction: 9 May 1979. J. Atmos. Sci., 40 , 2804–2830.
Stoelinga, M. T., J. D. Locatelli, and P. V. Hobbs, 2000: Structure and evolution of winter cyclones in the central United States and their effects on the distribution of precipitation. Part VI: A multiscale numerical modeling study of the initiation of convective rainbands. Mon. Wea. Rev., 128 , 3481–3500.
Tepper, M., 1950: A proposed mechanism of squall lines: The pressure jump line. J. Meteor., 7 , 21–29.
Thorpe, A. J., and S. A. Clough, 1991: Mesoscale dynamics of cold fronts: Structures described by dropsondes in FRONTS 87. Quart. J. Roy. Meteor. Soc., 117 , 903–941.
Uccellini, L. W., 1975: A case study of apparent gravity wave initiation of severe convective storms. Mon. Wea. Rev., 103 , 497–513.
Weisman, M. L., J. B. Klemp, and R. Rotunno, 1988: Structure and evolution of numerically simulated squall lines. J. Atmos. Sci., 45 , 1990–2013.
Tracks of the tornadoes that occurred during the Super Outbreak of 2–5 April 1974. From Abbey and Fujita (1983). The F0–F5 are damage ratings based on the Fujita scale
Citation: Monthly Weather Review 130, 6; 10.1175/1520-0493(2002)130<1633:ANLATS>2.0.CO;2
Analysis of sea level pressure (solid lines in hPa), tornado locations (open circles), and frontal positions (from operational analysis), surface winds (a full barb is 5 m s−1) and cloud cover (standard symbols) for 0000 UTC 4 Apr 1974. The precipitation regions associated with the three tornado-producing squall lines (numbered 1–3) and the region of nonsquall-line precipitation (labeled A) are indicated by the different shadings
Citation: Monthly Weather Review 130, 6; 10.1175/1520-0493(2002)130<1633:ANLATS>2.0.CO;2
As in Fig. 1. but the tornado tracks are labeled by the squall line with which they were associated
Citation: Monthly Weather Review 130, 6; 10.1175/1520-0493(2002)130<1633:ANLATS>2.0.CO;2
As in Fig. 2. but (a) contours of temperatures (in °C), and (b) contours of dewpoint (in °C)
Citation: Monthly Weather Review 130, 6; 10.1175/1520-0493(2002)130<1633:ANLATS>2.0.CO;2
Upper-air maps of temperature (dashed lines in °C) and geopotential height (solid lines in m) for 0000 UTC 4 Apr 1974 at (a) 850, (b) 700, and (c) 500 hPa. From Hoxit and Chappell (1975). The heavy dashed and dotted lines are the leading edge of the baroclinic zone at the respective pressure levels
Citation: Monthly Weather Review 130, 6; 10.1175/1520-0493(2002)130<1633:ANLATS>2.0.CO;2
Precipitation regions, and the positions of the Pacific cold front at the surface (solid line), at 850 hPa (dashed line), 700 hPa (dashed–dotted line), and 500 hPa (dashed–double-dotted line) for 0000 UTC 4 Apr 1974. The dotted line connecting the solid circles shows the location of the cross section in Fig. 7
Citation: Monthly Weather Review 130, 6; 10.1175/1520-0493(2002)130<1633:ANLATS>2.0.CO;2
Cross section (along the dotted line shown in Fig. 6) from Omaha, NE (OMA) to Waycross, GA, (AYS) of potential temperature (in K) for 0000 UTC 4 Apr 1974. From Hoxit and Chappell (1975). The heavy dashed line indicates the position of the Pacific cold front
Citation: Monthly Weather Review 130, 6; 10.1175/1520-0493(2002)130<1633:ANLATS>2.0.CO;2
Model results valid for 0000 UTC 4 Apr 1974 (36-h forecast) showing sea level pressure (solid lines in hPa), winds (see figure legend), 1-h totals of the explicit and convective parameterized precipitation (see figure legend), and frontal positions (standard symbols). See text for an explanation of the squall line locations
Citation: Monthly Weather Review 130, 6; 10.1175/1520-0493(2002)130<1633:ANLATS>2.0.CO;2
As in Fig. 8 but showing (a) contours of temperature (in °C) and (b) contours of dewpoint (in °C)
Citation: Monthly Weather Review 130, 6; 10.1175/1520-0493(2002)130<1633:ANLATS>2.0.CO;2
Model results showing squall line 1 (heavy dashed line), hourly totals of explicit and convective parameterized precipitation (see legend), and frontal positions (standard symbols) valid for (a) 0800, (b) 1400, and (c) 1800 UTC 3 Apr 1974
Citation: Monthly Weather Review 130, 6; 10.1175/1520-0493(2002)130<1633:ANLATS>2.0.CO;2
Model cross section along the line AA′ in Fig. 10 showing contours of potential temperature (solid lines in K), winds in the plane of the cross section relative to the motion of squall line 1 (see figure legend), position of the Pacific cold front, and the location in the cross section of squall line 1 (vertical shaded rectangle) valid for (a) 0800, (b) 1400, and (c) 1800 UTC 3 Apr 1974. The arrows point to the leading edge of the undular bore
Citation: Monthly Weather Review 130, 6; 10.1175/1520-0493(2002)130<1633:ANLATS>2.0.CO;2
Model results showing squall line 2 (heavy dashed line), hourly totals of explicit and convective parameterized precipitation (see figure legend), and frontal positions (standard symbols) valid for (a) 1600, and (b) 2000 UTC 3 Apr 1974; and (c) 0000, and (d) 0400 UTC 4 Apr 1974
Citation: Monthly Weather Review 130, 6; 10.1175/1520-0493(2002)130<1633:ANLATS>2.0.CO;2
Model results valid for 0000 UTC 4 Apr 1974 showing surface frontal positions (standard symbols), 500–950-hPa thickness (solid lines in 10s of meters), and hourly totals of explicit and convective parameterized precipitation (see figure legend)
Citation: Monthly Weather Review 130, 6; 10.1175/1520-0493(2002)130<1633:ANLATS>2.0.CO;2
Model results valid for 0000 UTC 4 Apr 1974 showing (a) potential temperature (solid lines in K) and horizontal potential temperature gradient (shading, see figure legend) at 800 hPa, and (b) potential temperature (solid lines in K) and negative horizontal potential temperature advection (shading, see legend). The heavy dashed line marks the location of the Pacific cold front. The lighter-shaded fronts are surface frontal positions taken from Fig. 13
Citation: Monthly Weather Review 130, 6; 10.1175/1520-0493(2002)130<1633:ANLATS>2.0.CO;2
Model positions of the Pacific cold front valid for 0000 UTC Apr 4 1974 at 900 hPa (solid line), 800 hPa (dotted line), 700 hPa (dashed–double-dotted line), 600 hPa (dashed line), 500 hPa (dashed–dotted line), and hourly totals of explicit and convective parameterized precipitation (see legend)
Citation: Monthly Weather Review 130, 6; 10.1175/1520-0493(2002)130<1633:ANLATS>2.0.CO;2
Model cross section along the line BB′ in Fig. 15 valid for 0000 UTC 4 Apr 1974 showing (a) potential temperature (solid lines in K) and horizontal potential temperature gradient (shading, see legend), and (b) horizontal potential temperature gradient (shading, see legend), winds in the cross section relative to the motion of squall line 2 (see legend), squall line 2 convergence (solid lines contoured every 1 × 10−5 s−1), and squall line 2 vertical velocity (dotted lines contoured every 2 cm s−1). The position of the Pacific cold front is shown by the heavy dashed line
Citation: Monthly Weather Review 130, 6; 10.1175/1520-0493(2002)130<1633:ANLATS>2.0.CO;2
Difference in potential temperature (averaged over the lowest 0.5 km, see legend) between the control model simulation and the model sensitivity test at forecast hour 30 valid for 1900 UTC 3 Apr 1974 for (a) the no-surface flux run, and (b) the no-evaporative cooling run. Also shown is an outline (heavy solid line) of simulated cloud, based on the 0.0025 g kg−1 contour of vertically averaged cloud water
Citation: Monthly Weather Review 130, 6; 10.1175/1520-0493(2002)130<1633:ANLATS>2.0.CO;2
Model cross section along the line CC′ in Fig. 17 at forecast hour 36 valid for 0000 UTC 4 Apr 1974 showing (a) the sensitivity model simulation with no surface energy fluxes and no latent cooling where the solid lines are contours of potential temperature (in K), the horizontal gradient of potential temperature (shading, see legend), winds in the cross section relative to the motion of squall line 2 (see legend), and vertical velocity associated with the Pacific cold front (contoured every 4 cm s−1), and (b) the sensitivity model simulation where the solid line are contours of potential temperature (in K), and the difference in potential temperature between the control run and sensitivity test (shading, see legend). The heavy dashed line is the location of the Pacific cold front
Citation: Monthly Weather Review 130, 6; 10.1175/1520-0493(2002)130<1633:ANLATS>2.0.CO;2
Superposition of analyzed wave packet A (heavy dashed line), convective region alpha (shaded area) and lines of echo from NWS radar summary maps (heavy dotted lines) for indicated UTC times. Adapted from Fig. 10 of Miller and Sanders (1980). The black circles indicate the locations of the stations whose altimeter settings are plotted in Fig. 20
Citation: Monthly Weather Review 130, 6; 10.1175/1520-0493(2002)130<1633:ANLATS>2.0.CO;2
Altimeter settings of the three stations shown in Fig. 19: Fort Wayne, IN (FWA), Louisville, KY (SDF), and Birmingham, AL (BHM). Adapted from Fig. 10 of Miller and Sanders (1980)
Citation: Monthly Weather Review 130, 6; 10.1175/1520-0493(2002)130<1633:ANLATS>2.0.CO;2
Danielsen's diagram of a negative sloping cold front–squall line type. (From an unpublished manuscript)
Citation: Monthly Weather Review 130, 6; 10.1175/1520-0493(2002)130<1633:ANLATS>2.0.CO;2
We use the term squall line here in the most general sense, consistent with the latest edition of the Glossary of Meteorology (Glickman 2000). The Glossary defines a squall line as “a line of active thunderstorms, either continuous or with breaks, including contiguous precipitation areas resulting from the existence of the thunderstorms.” The Glossary does not make any reference to the type of convective cells that comprise the line (ordinary cells or rotating supercells). This is particularly relevant to the squall lines in the super outbreak, which were well documented to have contained many supercells. Furthermore, the Glossary does not confine use of the term squall line to nonfrontal lines of convection [as did the original Glossary (Huschke 1959)]. Prior to publication of the new Glossary's definition, Stoelinga et al. (2000) suggested using the term convective rainband (rather than squall line) for a linear convective precipitation feature associated with a CFA (or any front in general). However, with the more general definition of a squall line given above, we revert to referring to a line of convection associated with a CFA as a squall line.
Fujita (1970) hypothesized that the development of a forward-tilted cold front would lead to superadiabatic lapse rates and thus an explosively unstable situation. However, in our studies we have not found superadiabatic lapse rates to form in conjunction with a CFA. In fact, with frontal surfaces represented as first-order discontinuities in temperature, a forward-tilted frontal surface does not require a superadiabatic lapse rate, as does a forward-tilted zero-order discontinuity. There can be weak or even neutral stability behind the forward-tilted portion of the Pacific cold frontal surface in such cases, but the convective development does not emanate from this steep lapse-rate air, which is typically quite dry. Rather, it is the lifting at or slightly ahead of the CFA that initiates convection in air with high CAPE ahead of the CFA.
The results from Danielsen's work discussed here, and shown in Fig. 21, were extracted from a manuscript by Danielsen entitled “A conceptual theory of tornadogenesis: Part I. Large-scale generation of severe storm potentials.” To the authors' knowledge, this manuscript was never published in its entirety. However, significant portions of the manuscript were published in the two cited references (Danielsen 1975a,b), although Fig. 21 and our quotes from Danielsen were not. A copy of Danielsen's unpublished manuscript can be obtained from the corresponding author of the present paper.