Download as pdf or txt
Download as pdf or txt
You are on page 1of 22

4220 MONTHLY WEATHER REVIEW VOLUME 136

The Effect of Mesoscale Heterogeneity on the Genesis and Structure of Mesovortices


within Quasi-Linear Convective Systems

DUSTAN M. WHEATLEY* AND ROBERT J. TRAPP


Purdue University, West Lafayette, Indiana

(Manuscript received 26 June 2007, in final form 2 January 2008)

ABSTRACT

This study examines the structure and evolution of quasi-linear convective systems (QLCSs) within
complex mesoscale environments. Convective outflows and other mesoscale features appear to affect the
rotational characteristics and associated dynamics of these systems. Thus, real-data numerical simulations
of two QLCS events have been performed to (i) identify and characterize the various ambient mesoscale
features that modify the structure and evolution of simulated QLCSs; and then to (ii) determine the nature
of interaction of such features with the systems, with an emphasis on the genesis and evolution of low-level
mesovortices.
Significant low-level mesovortices develop in both simulated QLCSs as a consequence of mechanisms
internal to the system—consistent with idealized numerical simulations of mesovortex-bearing QLCSs—
and not as an effect of system interaction with external heterogeneity. However, meso-␥-scale (order of
10 km) heterogeneity in the form of a convective outflow boundary is sufficient to affect mesovortex
strength, as air parcels populating the vortex region encounter enhanced convergence at the point of
QLCS–boundary interaction. Moreover, meso-␤-scale (order of 100 km) heterogeneity in the form of
interacting air masses provides for along-line variations in the distributions of low- to midlevel vertical wind
shear and convective available potential energy. The subsequent impact on updraft strength/tilt has impli-
cations on the vortex stretching experienced by leading-edge mesovortices.

1. Introduction In the idealized numerical simulations of QLCSs


(e.g., Thorpe et al. 1982; Rotunno et al. 1988; Weisman
An important unresolved issue in severe weather re-
1992, 1993; Skamarock et al. 1994; Weisman and Davis
search is the relationship between quasi-linear convec-
1998; Weisman and Trapp 2003; Trapp and Weisman
tive systems (QLCSs) such as squall lines and bow ech-
2003) and other convective storms, the initial condi-
oes and the complex mesoscale environments within
tions are supplied by a one-dimensional sounding that
which they evolve. Some observational results (e.g.,
varies only with height; horizontal homogeneity along
Klimowski et al. 2000; Przybylinski et al. 2000;
model vertical levels is assumed. The resultant class of
Schmocker et al. 2000) suggest that mesoscale variabil-
solutions—while arising from an incomplete formula-
ity in the wind and thermodynamic fields due, for ex-
tion of deep moist convection (see Balaji and Clark
ample, to convective outflows can significantly affect
1988)—agree with observations of highly organized,
the rotational characteristics of severe QLCSs. Never-
self-sustaining QLCSs (e.g., Burgess and Smull 1990;
theless, idealized numerical simulations produce appar-
Jorgensen and Smull 1993). Many hypotheses put forth
ently severe QLCSs in the absence of environmental
to explain the morphology, dynamics, and severe
heterogeneity.
weather potential of QLCSs are derived from idealized
modeling, and some have been validated by subsequent
* Current affiliation: NOAA/National Severe Storms Labora- studies. For example, a growing body of observational
tory, Norman, Oklahoma. evidence (e.g., Atkins et al. 2005; Wheatley et al. 2006;
Wakimoto et al. 2006) has confirmed the role of low-
level mesovortices in the production of damaging sur-
Corresponding author address: Dustan M. Wheatley, NOAA/
National Severe Storms Laboratory, National Weather Center,
face winds within QLCSs, as described in Trapp and
120 David L. Boren Blvd., Norman, OK 73072. Weisman (2003).
E-mail: dustan.wheatley@noaa.gov But idealized numerical simulations that develop me-

DOI: 10.1175/2008MWR2294.1

© 2008 American Meteorological Society


Unauthenticated | Downloaded 03/13/24 12:05 PM UTC
NOVEMBER 2008 WHEATLEY AND TRAPP 4221

soconvective systems are unable to fully document the An issue especially in the latter observational studies
environmental conditions affecting the life cycles of is the difficulty in establishing the unambiguous effect
these systems, as they cannot reproduce the inherent of environmental heterogeneity, due to the lack of a
four-dimensionality of the environment. homogeneous control. Toward this end, Atkins et al.
(1999) used a cloud-resolving model to quantify the
a. Conceptual models of storm–boundary influence of preexisting boundaries on supercell evolu-
interactions that produce tornadoes tion. In their control “no-boundary” run, the model was
A long series of modeling and observational studies initialized in the standard way, using a single-sounding
have focused on the life cycles of convective storms in representation of a tornadic environment. Initial con-
the presence of environmental heterogeneity. Consider ditions for the second, “boundary” experiment were
the evidence that a kinematic boundary is at times a chosen to resemble the meteorological conditions near
necessary condition for the development of nonsuper- Garden City, Kansas, on 16 May 1995, which was char-
cell tornadoes (e.g., Carbone 1983; Brady and Szoke acterized by a preexisting trough line. In both sets of
1989; Wakimoto and Wilson 1989; Mueller and Car- experiments, warm thermal bubbles were used to initi-
bone 1987; Lee and Wilhelmson 1997a,b). In the con- ate convection.
ceptual model of nonsupercell tornadogenesis, small- The inception and evolution of low-level rotation
scale circulations form along a convergence or shear [i.e, significant vertical vorticity of O(10⫺2 s⫺1) within
boundary through the release of a horizontal shearing the primary updraft] within the storms simulated in
instability. The circulations strengthen to tornadic in- the boundary runs tended to be earlier, stronger, and
tensity through stretching in the updraft of overhead more persistent, when compared to the no-boundary
convection. runs. Moreover, vorticity analyses of the two storms
A preexisting1 thermal boundary has been suggested revealed that the origin of rotation in the boundary
as a necessary condition for the development of signif- runs differed in part from the usual conceptualization
icant tornadoes [i.e., F2 or greater on the Fujita damage of low-level mesocyclogenesis (e.g., Rotunno and
intensity scale (Fujita 1981)]. Maddox et al. (1980) de- Klemp 1985; Davies-Jones and Brook 1993), in which
veloped a physical model of boundary layer wind fields streamwise horizontal vorticity along the forward-
to explain the seeming natural tendency of thunder- (rear) flank gust front is vertically tilted by the primary
storms to become more severe and even tornadic, upon
updraft (rear-flank downdraft). Now, the low levels
interaction with a synoptic-scale front or external thun-
on the cool-air side of the boundary were a source re-
derstorm outflow boundary. In their conceptual model,
gion for parcels rich in horizontal baroclinic vorticity,
boundary layer vertical wind profiles owing to shallow
which then flowed into the updraft in a streamwise
baroclinic zones enhance moisture content, conver-
manner. This behavior was consistent in those simu-
gence, and vertical vorticity, providing for tornadic
lated storms that propagated along the boundary,
thunderstorms in an otherwise unsupportive environ-
sometimes at small angles toward the warm-air side of
ment.
the boundary.
Observational studies born of the Verification of
By employing an otherwise idealized model with me-
the Origins of Rotation in Tornadoes Experiment
soscale variability in the environmental conditions, the
(VORTEX; Rasmussen et al. 1994) have considered a
number of cases where tornadogenesis via thunder- experimental methodology of Atkins et al. (1999) and
storm interactions with shallow baroclinic zones ap- more recently Richardson et al. (2007) represent the
peared likely. In particular, Markowski et al. (1998) most robust treatments of numerically simulated con-
found that nearly 70% of the significant tornadoes that vective storms interacting with their heterogeneous en-
occurred during VORTEX were associated in some vironment. It should be noted, though, that their initial
way with external shallow baroclinic zones. Moreover, conditions are defined analytically (without enhance-
of this subset, most of the tornadoes occurred within ment by observations) to closely resemble desired en-
the baroclinic zones of the boundaries, and generally vironmental conditions, and significant uncertainties
within 30 km of the boundaries’ leading edges. exist within this input, as stated by the authors. Fur-
thermore, the interaction is also necessarily “one way,”
hence, the convective scale is not permitted to modify
1
Here, the qualifiers “preexisting” and “external” refer to a
the mesoscale, which then cannot feedback to the con-
thermal and/or kinematic boundary generated by a source outside
the convective cell/system affected. Synoptic-scale fronts (i.e., vective scale. For these reasons, it is uncertain whether
warm fronts and stationary fronts) and old thunderstorm outflow their idealized results can be generalized to the real
boundaries are both examples of preexisting thermal boundaries. atmosphere.

Unauthenticated | Downloaded 03/13/24 12:05 PM UTC


4222 MONTHLY WEATHER REVIEW VOLUME 136

b. Conceptual models of a QLCS interacting with complex mesoscale environments and the subsequent
its environment implications for the rotational characteristics of these
systems. The inherent incompleteness of conventional
At present, only a few observational studies in the
and even experimental observations has precluded a
literature have considered how mesoscale heterogene-
satisfactory resolution of this scientific issue. To this
ities might affect the dynamics and severe weather po-
end, complete data are only available in the form of
tential of QLCSs. Collectively, Klimowski et al. (2000)
numerical model fields. There are examples of numeri-
and Przybylinski et al. (2000) have suggested that an
cal studies of QLCSs, such as Bernardet and Cotton
isolated convective cell(s) in advance of a linear MCS,
(1998) and Coniglio and Stensrud (2001), that have in-
upon merger with the main convective line, plays a role corporated mesoscale variability in their initial condi-
in subsequent bow-echo genesis. While not a prerequi- tions, albeit through different means, but none of these
site for these types of interactions, these cells often studies directly consider the subsequent impact of these
initiate in proximity to and are indicative of a preexist- features on the simulated system. Largely, modeling
ing surface boundary and associated region of en- studies of QLCSs are still limited by the use of envi-
hanced low-level convergence. Such mergers, albeit lo- ronments that are horizontally homogeneous, despite
cally, may intensify ongoing convection, which in turn growth in computing power and the development of
should force more intense (rainy) downdrafts. At last, a advanced numerical weather prediction models. The
deeper (i.e., stronger) cold pool and associated region latter numerical models, though, can be run with het-
of hydrostatically induced high pressure accelerates the erogeneous initial conditions interpolated from a large-
near-surface horizontal flow. scale analysis (e.g., the 40-km Eta analysis) for “real-
Analogous to the findings of Markowski et al. (1998) data” cases, as well as a land surface model.
on supercell–boundary interactions, Przybylinski et al. Thus, using real-data numerical simulations of severe
(2000) suggested an association between QLCS– QLCSs, the objectives herein are to
boundary interactions and tornadogenesis within
1) identify and characterize the various ambient meso-
QLCSs. Similar interaction between a QLCS and its
scale features that modify structure and evolution of
heterogeneous mesoscale environment is noted by
simulated QLCSs; and then
Schmocker et al. (2000). In this conceptual model,
2) determine the nature of interaction of such features
streamwise horizontal vorticity enhancement occurs
with systems, emphasizing the dynamics of genesis
along a preexisting, line-normal thermal boundary,
and evolution of low-level mesovortices.
where baroclinicity is appreciable. The production of
cyclonic vertical vorticity occurs as parcels flow into the In section 2, candidate events for simulation and
base of the updraft and undergo considerable vertical analysis are identified, followed by a discussion of the
vortex stretching. Indeed, Przybylinski et al. (2000) research methodology employed in this study. Section 3
noted that the development of severe weather (primar- provides an overview of the simulations. In section 4,
ily damaging “straight-line” winds) often followed the the mechanism(s) for low-level mesovortex genesis is
interactions between QLCSs and preexisting surface considered within a real-data framework, and then
boundaries. Moreover, they noted a near one-to-one compared to the fundamental explanation given from
correspondence between tornado damage (as inferred an idealized perspective. The effects of meso-␥- and
from damage surveys) and mesovortex tracks (as in- meso-␤-scale heterogeneities on the rotational charac-
ferred from single-Doppler radar), which argues for the teristics of the simulated QLCS are examined in sec-
enhanced potential for QLCS tornadogenesis. tions 5 and 6, respectively. Finally, in section 7, the
While intuitive, this mechanism for low-level rotation results of this study are summarized, their implications
fails to explain the development of near-surface rota- are discussed, and suggestions for future research on
tion. In fact, Davies-Jones (1982a,b) contended that this topic are offered.
such an “in, up, and out” type flow will only produce
significant vertical vorticity after parcels flowing into 2. Methodology
the updraft have ascended a few kilometers. Abrupt
upward turning of the streamlines, which are coincident The fully compressible, nonhydrostatic Advanced
with the vortex lines, would require uncharacteristic Research Weather Research and Forecasting (WRF-
horizontal variations in vertical velocities over very ARW; Skamarock 2005) version 2 is used to perform
short distances. real-data numerical simulations of the mature, exten-
Results of the studies summarized above indicate the sive bow echo on 6 July 2003 and the squall-line bow
need to investigate the interactions of QLCSs with their echo on 24 October 2001. Schemes used to parameter-

Unauthenticated | Downloaded 03/13/24 12:05 PM UTC


NOVEMBER 2008 WHEATLEY AND TRAPP 4223

FIG. 1. Horizontal grids used for real-data numerical simulations of quasi-linear convective systems on (a) 6 Jul
2003 and (b) 24 Oct 2001. In each case, d01, d02, and d03 utilize 9-, 3-, and 1-km horizontal grid spacings,
respectively.

ize physical processes include the six-class Purdue Lin are spaced less than 100 m near the surface to over 1 km
microphysical scheme (Chen and Sun 2002), the Noah at model top (i.e., the 50-hPa pressure surface).
land surface model (Chen and Dudhia 2001), and the Interpretation of the modeling results that follow is
Yonsei University (YSU) planetary boundary scheme guided by our philosophical approach to real-data nu-
(Noh et al. 2003), and, for atmospheric radiation, the merical simulations. A successful simulation must rep-
Rapid Radiative Transfer Model (longwave; Mlawer et licate the salient structural characteristics of the ob-
al. 1997) and the fifth-generation Pennsylvania State served system and its proximity environment. In this
University–National Center for Atmospheric Research regard, conventional observations are used to verify the
(PSU–NCAR) Mesoscale Model (MM5; Dudhia) representativeness of the results. The observed and
scheme (shortwave; Dudhia 1989). All simulations take simulated systems, though, are viewed most appropri-
account of earth’s rotation. ately as corresponding, but not identical, parts. Particu-
The initial conditions for the real-data simulations larly, at increasingly smaller spatial scales, the observa-
are supplied by the 40-km Eta model analysis, with tions required to confirm the dynamical characteristics
boundary conditions updated on a 3-h interval using the of (some) system–environment interactions produced
Eta model forecasts. Model computations are per- in these simulations are not presently available. Like
formed on a parent grid and two inner or nested grids earlier idealized modeling studies (later verified by ob-
that are fully two-way interactive. The parent grid uses servations), additional case studies and confirmatory
a horizontal grid point spacing of 9 km to allow for the observations will be needed to further generalize in-
influence of large-scale forcing mechanisms. The effects sights gleaned from this initial study.
of deep cumulus convection are parameterized with the
Kain–Fritsch scheme (Kain 2004) only on this grid. The 3. Overview of the simulations
intermediate and finest grids, with horizontal grid point
spacings of 3 and 1 km, respectively, effectively provide
a. The 6 July 2003 bow echo: A warm-season
for convective-system (Weisman et al. 1997) and cloud- example
resolving modeling. The placement of the intermediate The bow echo of 6 July 2003, observed during the
grid encompasses the region of convection initiation, Bow Echo and MCV Experiment (BAMEX; Davis et
while that of the finest grid reflects the movement of al. 2004), is one subject of this study. This bowing con-
the observed (and simulated) QLCS during it late for- vective system evolved from intense multicellular con-
mative and mature stages; the domain constructions for vection over northeastern Nebraska (see Fig. 2a), just
the specific cases are shown in Fig. 1. The parent grid at downwind of a shortwave upper-level trough and be-
every coarse-grid time step specifies nested lateral neath moderately strong flow (21 m s⫺1) at 500 hPa.
boundary conditions. The vertical grid has 31 levels that The atmosphere in proximity to the mature system was

Unauthenticated | Downloaded 03/13/24 12:05 PM UTC


4224 MONTHLY WEATHER REVIEW VOLUME 136

FIG. 2. Comparison of radar reflectivity factor at 0.5° elevation and ⬃2-km AGL simulated radar reflectivity
factor at (a) ⬃0300 UTC (from the KOAX WSR-88D)/0130 UTC 6 Jul 2003, (b) ⬃0445 UTC (from the KOAX
WSR-88D)/0300 UTC 6 Jul 2003, and (c) ⬃2200 UTC (from the KIWX WSR-88D)/2300 UTC 24 Oct 2001. Note
the discrepancies in timing between companion panels in (a), (b), and (c), which are because of the evolution of
the simulated systems.

quite unstable and moist, with ML CAPE of 3079 J Such high instability apparently compensated for the
kg⫺1 in the 0000 UTC sounding from Omaha, Ne- relatively weak 0–2.5-km shear of 10 m s⫺1 (Figs. 3a,b),
braska, and surface dewpoints in excess of 18°C over thus, allowing for the formation of the strong system
eastern Nebraska and extreme western Iowa (Fig. 3a). shown in Fig. 2b. While short-lived, the bow echo of

Unauthenticated | Downloaded 03/13/24 12:05 PM UTC


NOVEMBER 2008 WHEATLEY AND TRAPP 4225

FIG. 3. Proximity sounding and hodograph from (a), (b) OAX (Omaha, NE) observations
at 0000 UTC 6 Jul 2003, and from (c), (d) DTX (Detroit, MI) and (e), (f) ILN (Wilmington,
OH) observations at 0000 UTC 25 Oct 2001. In (b), (d), and (f), the first four nodes along the
hodograph trace correspond to 925, 850 (⬃1400 m), 700 (⬃3000 m), and 500 hPa (⬃5700 m).

Unauthenticated | Downloaded 03/13/24 12:05 PM UTC


4226 MONTHLY WEATHER REVIEW VOLUME 136

FIG. 4. Images of radar reflectivity factor and ground-relative radial velocity at (a) 0.5° elevation, from the
WSR-88D KOAX at 0554:13 and 0554:33 UTC 6 Jul 2003, respectively; and (b) 0.5° elevation, from the WSR-88D
KIWX at 2050:19 and 2050:39 UTC 24 Oct 2001. Range rings are displayed at 60-km intervals. Inset in (a) shows
storm-relative motion within the dashed box.

6 July 2003 exhibited the salient features of more orga- is initialized at 0000 UTC 5 July 2003, approximately
nized systems—including a well-defined rear-inflow jet 21 h before the occurrence of initial convective cells.
(RIJ), line-end vortices, and a diversity of low-level Such a period of time is necessary for the simulated
mesovortices (e.g., see Fig. 4a)—and produced tens of environment to evolve from an otherwise cold start,
high wind reports [(National Climatic Data Center) which possesses no cloud water and only that hetero-
NCDC 2003], with widespread F0-intensity (and more geneity resolved in the initial conditions.
localized F1-intensity) wind damage occurring across
eastern Nebraska and western Iowa.
To simulate this event, the WRF model is applied to
1) SYSTEM-SCALE STRUCTURE

a domain that covers the Great Plains, upper Midwest, Comparable to the observed severe bow echo (see
and the lower Ohio River Valley (Fig. 1a). The model Figs. 2a,b), the simulated bow echo has a structure and

Unauthenticated | Downloaded 03/13/24 12:05 PM UTC


NOVEMBER 2008 WHEATLEY AND TRAPP 4227

evolution described as follows: At approximately 2100 4 km MSL. The structure and evolution of MV2, as well
UTC 5 July, several individual convective cells initiate as other low-altitude mesovortices within this simulated
over the southwest quadrant of South Dakota, ahead of bow system (for the strengths of other mesovortices,
a synoptic-scale shortwave trough embedded in north- see ahead Fig. 17), are consistent with QLCS observa-
westerly flow (not shown). These convective cells move tions (e.g., Atkins et al. 2004, 2005; Przybylinski et al.
to the east-southeast and merge with ongoing convec- 2000) and the idealized numerical simulations of (e.g.,
tion over northeastern Nebraska (Fig. 2a). By 0130 Weisman and Trapp 2003).
UTC, the interaction of these cells leads to a mesoscale
convective system greater than 100 km in length, as
b. The 24 October 2001 squall-line bow echo: A
evidenced by a continuous line of simulated radar re-
cool-season example
flectivity in excess of 35 dBZ (Fig. 2a). Over the next
few hours, this system grows upscale into a mature, The squall-line bow echo on 24 October 2001 is the
extensive bow echo approximately 200 km in length, other subject of this study. Occurring in October, this
which moves through the Omaha metropolitan area event exemplifies a severe bow echo forced at the syn-
and into western Iowa (Fig. 2b). optic scale by a dynamic pattern more typical of cool-
Although occurring 1–2 h early compared to the ob- season months. Ahead of a strong, migrating low pres-
servations, the salient features of the observed system sure system over the upper Midwest, meteorological
(rear-inflow jet, line-end vortices, etc.) are reproduced conditions across northern Indiana, southern Michigan,
well by the WRF model at high resolution. To demon- and surrounding areas were marginally unstable, with
strate this point, the system-scale structure of the simu- ML CAPE of 500–1000 J kg⫺1 in the 0000 UTC sound-
lated bow system during its mature stage, at 0300 UTC, ings from Detroit, Michigan, and Wilmington, Ohio
is examined (Fig. 5a). At approximately 2.00 km AGL, (Figs. 3c,e). Large vertical wind shear was concentrated
a narrow zone of nearly continuous updraft occurs at low levels in these same soundings, with (south)
along the system’s leading edge. A concentrated rear- southwesterly winds increasing to greater than 25 m s⫺1
inflow jet extends tens of kilometers rearward of the at 1–2 km AGL (Figs. 3c–f). Within this low-instability/
active leading-line convection. Furthermore, a domi- high-shear parameter space, an extensive, south–north-
nant cyclonic vortex is evidenced on the northern end oriented squall line develops across the eastern half of
of the convective system by a general cyclonic turning the United States. This mesovortex-bearing system
of the velocity vectors. (e.g., see Fig. 4b) was host to 13 tornadoes across north-
eastern Indiana, and numerous incidents of high wind
were also reported (NCDC 2001). Results of a real-data
2) SUBSYSTEM-SCALE STRUCTURE
numerical simulation of this bow-echo event are sum-
Here the subsystem-scale structure of the simulated marized below.
bow system during its mature stage, at approximately The WRF model is applied to a domain that covers
0300 UTC, is examined (Fig. 5b). At approximately the eastern one-half of the United States (Fig. 1b). We
400 m AGL, several small-amplitude waves occur along note that the placement of the finest grid is chosen such
the leading edge of the cold pool (i.e., the 304-K po- that it covers the area of embedded bowing segments.
tential temperature contour), and are collocated with The model is initialized at 0000 UTC 24 October 2001,
low-level mesovortices. For reference, a “significant” approximately 18 h before convective storms.
mesovortex is one with a diameter ⱖ4 km (four hori-
zontal grid intervals on d03), maximum vertical vortic-
ity ⱖ0.01 s⫺1, appreciable vertical depth (⬎1 km), and
1) SYSTEM-SCALE STRUCTURE

general time and space coherency. Comparable to the observed severe squall-line bow
A west–east cross section through MV2 (see Fig. 5b) echo (Fig. 2c), the simulated squall-line bow echo has a
at 0300 UTC provides further insight into the structural structure and evolution described as follows: by 1800
characteristics of a significant mesovortex (Fig. 6a). UTC, convective cells associated with a strong, prefron-
MV2 is located within a zone of potential temperature tal convergence line organize into a linear convective
gradient behind the advancing edge of the cold pool. system over central Illinois and southeastern Missouri
The maximum vertical vorticity associated with MV2, (not shown). Around 2300 UTC, the main convective
0.016 s⫺1 or approximately 1.6 times mesocyclone-scale line extends from southern Michigan southwestward
vorticity, occurs at low levels. An earlier west–east over northeastern Arkansas, as evidenced by a continu-
cross section through MV2 confirms that it develops ous line of simulated radar reflectivity in excess of 35
first near the surface (not shown) and builds upward to dBZ. This main convective line develops pronounced

Unauthenticated | Downloaded 03/13/24 12:05 PM UTC


4228 MONTHLY WEATHER REVIEW VOLUME 136

FIG. 5. Horizontal cross sections at ⬃2.0 and ⬃0.4 km AGL at (a), (b) 0300 UTC 6 Jul 2003 and (c), (d) 2300
UTC 24 Oct 2001. Positive and negative vertical velocities are shown, in (a) and (c), every 5 m s⫺1 as solid and
dashed black lines, respectively, with no zero contour. Positive vertical vorticity values are shown, in (b) and (d),
every 500 ⫻ 10⫺5 s⫺1 as solid black lines. In (a)–(d), rainwater mixing ratios greater than 1, 3, and 5 g kg⫺1 are
shaded light, medium, and dark gray, respectively, and every tenth ground-relative, horizontal wind vector is
plotted. Full and half barbs denote speeds of 10 and 5 m s⫺1, respectively. Low-level mesovortices are denoted with
an MV# on the plots at ⬃0.4 km AGL. Tick marks are displayed every 20 km.

bowing segments over the northern half of Indiana and 2) SUBSYSTEM-SCALE STRUCTURE
southern Michigan, each with their own rear-inflow jets
(Fig. 5c). It should be noted that the timing of the simu- At approximately 400 m AGL, a number of low-level
lated squall line, though, lags the observed system by mesovortices can be found along-line during the simu-
about 1 h. lated system’s time on the finest grid (e.g., Fig. 5d).

Unauthenticated | Downloaded 03/13/24 12:05 PM UTC


NOVEMBER 2008 WHEATLEY AND TRAPP 4229

FIG. 6. West–east vertical cross sections of vertical vorticity, vertical velocity, and potential temperature (see
color bar) through (a) MV2 in Fig. 5b at 0300 UTC 6 Jul 2003 and (b) MV3 in Fig. 5d at 2300 UTC 24 Oct 2001.
Vertical vorticity (black lines) is contoured every 400 ⫻ 10⫺5 s⫺1 with no zero contour and dashed negative
contours. Vertical velocity (gray lines) is contoured very 5 m s⫺1, with no zero contour and dashed negative
contours.

Maximum vertical vorticities associated with these the vertical velocity, VH is the horizontal velocity vec-
mesovortices, at times, exceed 3 times mesocyclone- tor, and F␨ is the mixing term. The first and second
scale vorticity. A west–east cross section through MV3 terms on the RHS of the above equation represent the
at 2300 UTC shows its maximum vertical vorticity, like tilting of horizontal vorticity in the vertical and the
that of MV2 from the 6 July 2003 bow-echo event, oc- stretching of absolute vertical vorticity, respectively. As
curring at low levels (Fig. 6b). in Trapp and Weisman (2003), the time-integrated con-
tributions from the tilting and stretching processes in
Eq. (1) are calculated along trajectories as follows:
4. Analysis of low-level mesovortex genesis
a. The 6 July 2003 bow echo
The development of a significant low-level mesovor-
␨a共x, y, z, t兲 ⫽ ␨a共x0, y0, z0, t0兲 ⫹ 冕t0
t
␻H ⭈ ⵱Hw dt

tex on the finest grid is now examined. At 0300 UTC,


MV2 is one of several significant mesovortices along
the system’s leading edge (see again Fig. 5b). The de-
⫺ 冕t0
t
␨a⵱H ⭈ VH dt. 共2兲

velopmental history of MV2 can be traced back to a


cyclonic–anticyclonic vortex couplet some 30 min ear- Here, model data predicted every 1 min are used to
lier (Fig. 7). At this earlier time, a nearly symmetrical diagnose these forcing terms as well as ␨a. Trajectories
(particularly in regards to vortex intensity) couplet of all parcels in the vicinity of significant low-level
straddles a low-level downdraft and local maximum in mesovortices are considered. The total time integration
the rainwater mixing ratio field. This implication is that period is t ⱕ 10 min.
this and other mesovortices are formed as horizontal Backward trajectories are calculated for parcels
vorticity associated with the cold pool is tilted into the spaced every 1 km within a cubic region encompassing
vertical by convective downdrafts. MV2. All of the parcels originate from either the inflow
To confirm the deductions from Fig. 7, consider the or the rear of the system. Figure 7 shows a representa-
following form of absolute vertical vorticity equation: tive trajectory projected into the plane. Times series of
parcel height show that the positive vertical vorticity
d␨a associated with MV2 is generated in a downward cur-
⫽ ␻H ⭈ ⵱Hw ⫺ ␨a⵱H ⭈ VH ⫹ F␨, 共1兲 rent of air, as baroclinic horizontal vorticity is tilted
dt
vertically (Fig. 8). Beyond 0236 UTC, the terminus for
where ␨a ⫽ ␨ ⫹ f(␪) is the absolute vertical vorticity, ␪ the above trajectories, MV2 is maintained along the
is the latitude, ␻H is the horizontal vorticity vector, w is gust front for greater than 1 h.

Unauthenticated | Downloaded 03/13/24 12:05 PM UTC


4230 MONTHLY WEATHER REVIEW VOLUME 136

FIG. 7. Approximately 400-m AGL horizontal cross section of potential temperature contours (every 4 K),
vertical velocity (see color bar), and vertical vorticity contours (every 400 ⫻ 10⫺5 s⫺1, with no zero contour and
dashed negative contours), at 0236 UTC 6 Jul 2003. Tick marks are plotted every 1 km.

b. The 24 October 2001 squall-line bow echo (Fig. 10a). (The southern end of the squall line enters
this domain with well-developed vortices.) This spacing
Horizontal components of vorticity are quite large in is consistent with the linear theory of Miles and How-
the prefrontal environment—typically between 0.03 ard (1964), which predicts that the wavelength of the
and 0.04 s⫺1—owing to an intense southerly low-level instabilities is ⬃7.5 times the vortex sheet width, and
jet (20–25 m s⫺1 at ⬃500 m AGL). Upward tilting of with the idealized numerical simulations of Lee and
this eastward-directed horizontal vorticity generates Wilhelmson (1997a,b), which suggest that this wave-
vertical vorticity values of 0.001–0.003 s⫺1 at ⬃250– length can be roughly half this theoretical value. Fur-
750 m AGL (Fig. 9). With subsequent stretching, the ther evidence of horizontal shearing instability as a
rate of vorticity production shown in Fig. 9 readily ex- mechanism for the generation of low-level mesovorti-
plains the narrow band of positive vertical vorticity ces is the general lack of anticyclonic–cyclonic vortex
along the east flank of the system-scale updraft— couplets at all stages of vortex development (see Fig.
especially below 1 km AGL—where ␨ ⬃0.004 s⫺1. 10b).
The emergence of vorticity maxima and ultimately In summary, the basic mechanism for low-level
low-level mesovortices from this 3- (⫾1) km-wide vor- mesovortex genesis within the real-data numerical
tex sheet appears to be associated with the release of a simulation of the warm-season QLCS event of 6 July
horizontal shearing instability (see, e.g., Carbone 1983; 2003 is dynamically equivalent to the process described
Mueller and Carbone 1987). At 2100 UTC, a number of within the idealized modeling framework of Trapp and
small-scale vortices separated by ⬃10–20-km distance Weisman (2003). As just shown, mesovortices occurring
can be found along the northern end of the squall line within the cool-season QLCS event of 24 October 2001

Unauthenticated | Downloaded 03/13/24 12:05 PM UTC


NOVEMBER 2008 WHEATLEY AND TRAPP 4231

FIG. 8. Time series of parcel height (thin solid line), absolute vertical vorticity (small dashed line), and the
time-integrated contributions to vertical vorticity (thick solid line) from the tilting (medium dashed line) and
absolute vorticity-stretching terms (large dashed line) for a representative parcel originating in the inflow. Time
“0” corresponds to the release time of 0236 UTC 6 Jul 2003.

appear due to the release of a horizontal shearing in- potential temperature (␪e) (Fig. 11). In the case of the 6
stability. In each case, the genesis mechanism is in di- July 2003 bow-echo event, one can identify two distinct
rect response to the intrinsic dynamics of the system— scales of heterogeneity, excluding the horizontal vari-
as suggested in the aforementioned idealized experi- ability in temperature associated with the system-
ments—and not from some system interaction with generated cold pool (Fig. 11a). The meso-␤-scale het-
external (from the system) mesoscale heterogeneity. erogeneity is related to an airmass boundary across cen-
This finding is significant given that such variability can tral Nebraska and west-central Iowa, and is reserved
arise in real-data numerical simulations without further for discussion in section 6. The meso-␥-scale heteroge-
refinement, and previous observational studies of neity is a nearly circular outflow boundary displaced
mesovortex-bearing QLCSs have described this type of just to the northwest of the northern end of the main
role for system–environment interactions. Subsequent convective line. This outflow is associated with a cluster
sections, though, will demonstrate how preexisting het- of convective cells that initiated 1–2 h earlier over ex-
erogeneities can affect the structure and evolution of a treme southeastern South Dakota, and have since
preexisting vortex. moved east-southeast over west-central Iowa. The
southern periphery of this outflow possesses a cross-
5. An effect of meso-␥-scale heterogeneity boundary ␪e gradient of the order of 10 K km⫺1, and is
oriented normal to the system’s leading edge.
When present, heterogeneity in the mesoscale envi- We now consider the effect of this mesoscale hetero-
ronments of the QLCSs is normally revealed well in geneity on the developmental history of MV1, another
maps of the magnitude of the gradient of equivalent leading-edge mesovortex that forms during the mature

Unauthenticated | Downloaded 03/13/24 12:05 PM UTC


4232 MONTHLY WEATHER REVIEW VOLUME 136

FIG. 9. A forward trajectory: Time series of parcel height (thin solid line), absolute vertical vorticity (small
dashed line), and the time-integrated contributions to vertical vorticity (thick solid line) from the tilting (medium
dashed line) and absolute vorticity-stretching terms (large dashed line) for a representative parcel originating in
the inflow. Time “0” corresponds to the release time of 2130 UTC 24 Oct 2001.

stage of the simulated bow echo (see Fig. 5b). At 0253 tensification continues through 0303 UTC (Fig. 12c),
UTC, MV1 is one of two mesovortices embedded after which mesovortex strength increases by nearly
within an elongated region of positive vertical vorticity 200%, to nearly 0.03 s⫺1.
along the northern end of the system’s leading edge To confirm the nature of this QLCS–boundary inter-
(Fig. 12a). The vorticity maximum of approximately action, backward trajectories were calculated for par-
0.01 s⫺1 associated with MV1 is located just on the cels within a cubic region encompassing MV1 (see Fig.
warm-air side of the line-normal boundary. Like MV2, 13 for their projections into the horizontal plane). The
this mesovortex is generated via the tilting of baroclinic upper bound for this cubic region is the model sigma
horizontal vorticity within a downdraft; its intensifica- level at approximately 1 km AGL. Similar to the analy-
tion appears to be linked to a focused region of en- sis of low mesovortex genesis in section 4, parcels
hanced low-level convergence, produced by the inter- within the vortex region at 0303 UTC—the time of
section of the system-generated cold pool and bound- peak mesovortex intensity—originate from the inflow
ary (Fig. 12d). and behind the gust front (Fig. 13a). In this case,
Over the next few minutes, in a storm-relative frame- though, nearly 50% of the considered parcels originate
work, the boundary moves toward the south-southeast along the boundary (Fig. 13a). Horizontal cross sections
(Fig. 12b). By 0256 UTC, the local maximum in the of horizontal vorticity vectors show appreciable stream-
convergence field associated with the boundary spa- wise horizontal baroclinic vorticity associated with the
tially corresponds with MV1 (Fig. 12e), presumably boundary (Fig. 13b). However, as shown in Fig. 13b and
promoting the stretching and subsequent amplification confirmed in Fig. 14, tilting of such vorticity is relatively
of its vertical vorticity. The process of mesovortex in- weak for parcels flowing along the boundary. For the

Unauthenticated | Downloaded 03/13/24 12:05 PM UTC


NOVEMBER 2008 WHEATLEY AND TRAPP 4233

tical vorticity born in the boundary contribute to MV1,


which as demonstrated is amplified by significant
stretching due to the QLCS–boundary interaction. It
should be noted that parcels originating farther behind
the boundary begin with small amounts of negative vor-
ticity, which tends to zero with stretching.
While significant, the QLCS–boundary interaction
described above is a relatively short-lived process. Be-
yond 0303 UTC, the convergence maximum has moved
beyond the center of circulation, which is increasingly
dislocated from the leading edge of the gust front (not
shown). MV1 begins to weaken in the same period of
time.
To summarize, these simulation results show that the
effect of the boundary is limited to the intensification of
MV1.

6. An effect of meso-␤-scale heterogeneity


The meso-␤-scale heterogeneity in Fig. 11a is related
to a low-level airmass boundary that had been quasi-
stationary over east-central Nebraska and west-central
Iowa for 1–2 days prior to the bow-echo event. Across
this transition zone, the low-level flow is south-
southwesterly over southern parts of eastern Nebraska
and western Iowa, and becomes light and variable (to,
at times, easterly) north-northwest of this area (shown
for the simulation in Fig. 15a), producing significant
convergence across central Nebraska and west-central
Iowa. This east–west-oriented boundary is distin-
guished from the surrounding area by a relatively subtle
meridional temperature gradient. Significant low-level
moisture convergence occurs along and north of the
boundary, as revealed by the sharper gradient in
equivalent potential temperature (shown for the simu-
lation in Fig. 11a).
A number of these environmental conditions play a
significant role in both the formation and mature stages
of the simulated system. For example, the pooling of
FIG. 10. Horizontal cross section, on the lowest model level, of
moisture near the boundary and the associated desta-
(a) vertical vorticity (every 250 ⫻ 10⫺5 s⫺1, with no zero contour bilization helps to control the east-southeasterly orien-
and dashed negative contours), at the two times indicated. (b) The tation and motion of the initial convection storms (Figs.
inner box and “x” shows the geographic area and the center of a 2a,b). Additionally, the vertical juxtaposition of weak
mesovortex at 2130 UTC, respectively. Also in (b), vertical vor- easterly flow at the surface and westerly flow at 700 hPa
ticity contours (every 250 ⫻ 10⫺5 s⫺1, with no zero contour and
dashed negative contours) and potential temperature contours
produces a belt of enhanced 0–3-km shear along and
(every 4 K). Tick marks are plotted every 20 km in (a) and every north of the boundary (Fig. 15c). The shear distribution
1 km in (b). at midlevels is qualitatively similar to the low-level dis-
tribution, with greater values occurring in proximity to
the boundary (Fig. 15d). This horizontal variation in the
parcel examined, the tilting process contributes no low- and midlevel shear fields produces significant
more than 0.002 s⫺1 (i.e., ⬃10% or less) to mesovortex along-line variability during the mature stage of the
strength at 0303 UTC, which is representative of the lot simulated system: the northern half of the simulated
of parcels considered. However, small amounts of ver- system propagates through stronger shear along and

Unauthenticated | Downloaded 03/13/24 12:05 PM UTC


4234 MONTHLY WEATHER REVIEW VOLUME 136

FIG. 11. Horizontal cross sections of the magnitude of the ␪e gradient on the lowest model
level (see color bar; K m⫺1) at (a) 0300 UTC 6 Jul 2003 and (b) 2300 UTC 24 Oct 2001. In (a)
and (b), the 1 g kg⫺1 rainwater mixing ratio contour is contoured (dashed line). In (a), black
dots denote the positions of 40-km line-averaged cross sections shown in Fig. 16. Tick marks
are displayed every 20 km.

Unauthenticated | Downloaded 03/13/24 12:05 PM UTC


NOVEMBER 2008 WHEATLEY AND TRAPP 4235

FIG. 12. Approximately 250-m AGL horizontal cross sections of equivalent potential temperature (every 4 K)
at (a) 0253, (b) 0256, and (c) 0303 UTC; and of divergence/convergence contours (dashed/thin solid lines; every
400 ⫻ 10⫺5 s⫺1, with no zero contour), at (d) 0253, (e) 0256, and (f) 0303 UTC. In (a)–(c), rainwater mixing ratios
between 1 and 3 g kg⫺1 are lightly shaded, with values greater than 3 g kg⫺1 darkly shaded. In (d)–(f), convergence
values greater than 1200 ⫻ 10⫺5 s⫺1 are shaded. In (a)–(f), positive vertical vorticity contours are plotted every
0.005 s⫺1 (thick solid lines). Tick marks are displayed every 1 km.

north of the boundary, while its southern half propa- level mesovortices along and north of the boundary
gates though the area of weaker shear south of the tend be deeper, stronger, and longer lived than their
boundary (Figs. 15c,d). Consistent with the theoretical counterparts on the southern end of the line (Fig. 17a).
discussion of Rotunno et al. (1988), this configuration is Vertical vorticity associated with northern-end meso-
supportive of deeper, stronger, and more vertically vortices exceeds 0.03 s⫺1 compared to 0.01 s⫺1 for
erect updrafts along the leading edge of the northern mesovortices on the southern end of the line. This find-
half of the simulated system (see Figs. 16a,b). ing is consistent with Weisman and Trapp (2003), who
The thermodynamic conditions of the storm proxim- noted the dependence of low-level mesovortices within
ity environment also may contribute to the variations in squall lines and bow echoes on environmental vertical
updraft strength/tilt shown in Fig. 16. Here, the distri- wind shear.
bution of most unstable (MU) CAPE is qualitatively As a reference, it is interesting to compare mesovor-
similar to the distributions of low- and midlevel shear, tex evolution in the 6 July 2003 QLCS to that in the 24
with large amounts of MU CAPE in proximity to the October 2001 QLCS. The numerous vortices on 24 Oc-
boundary (Fig. 15a). Also, convective inhibition (CIN) tober 2001 have strengths comparable to those occur-
is smallest in proximity to the boundary (Fig. 15b). The ring on 6 July 2003, yet their tracks are nearly continu-
environmental sensitivity experiments of Weisman ous (Fig. 17b). This reflects the relatively homogeneous
(1993) found that greater system-scale organization environment encountered by the 24 October 2001
could be promoted by simply increasing environmental QLCS, as illustrated by the nonexistence of a ␪e gradi-
CAPE (above at least 2000 J kg⫺1), even for relatively ent in excess of 1 K km⫺1 outside of the system-
weak low- to midlevel shear values. generated cold pool (Fig. 11b). Hence, we can conclude
These effects of the meso-␤-scale heterogeneity on that the effect of mesoscale heterogeneity appears suf-
the system-scale structure also have consequences on ficient but not necessary for the later-stage develop-
the evolution of low-level mesovortices along the lead- ment of low-level mesovortices.
ing edge of the system. The strong, erect updrafts on
the northern end of the line provide greater opportu- 7. Summary and discussion
nity for deep lifting and vortex stretching than do those Real-data simulations of two severe QLCS events
on the southern end of the line. Not surprisingly, low- have been presented, with an emphasis on the relation-

Unauthenticated | Downloaded 03/13/24 12:05 PM UTC


4236 MONTHLY WEATHER REVIEW VOLUME 136

FIG. 13. Approximately 250-m AGL horizontal cross sections, at 0303 UTC of (a) positive
vertical vorticity (every 500 ⫻ 10⫺5 s⫺1) and equivalent potential temperature (every 4 K);
and of (b) vertical velocity contours (see color bar; areas of downdraft are enclosed by dashed
contours) and horizontal vorticity vectors. A vector with a magnitude of 0.1 s⫺1 exactly
reaches the tail of the next adjacent vector. In (a), the projections of 3D trajectories released
at 0303 UTC (thin gray lines) are displayed. In (b), the projection of a 3D trajectory origi-
nating from the boundary is displayed. In (a) and (b), tick marks are plotted every 1 km.

Unauthenticated | Downloaded 03/13/24 12:05 PM UTC


NOVEMBER 2008 WHEATLEY AND TRAPP 4237

FIG. 14. Time series of (a) parcel height (thin solid line), vertical vorticity (small dashed line), and the time-
integrated contributions to vertical vorticity (thick solid line) from the tilting (medium dashed line) and absolute
vorticity-stretching terms (large dashed line), for a representative parcel originating along the boundary; and of (b)
the magnitude of the horizontal vorticity vector. Time “0” corresponds to the release time of 0303 UTC 6 Jul 2003.

ship between these systems and the complex environ- studies of mesovortex-bearing QLCSs. The simulated
ments within which they evolve. The 6 July 2003 event environment on 6 July 2003 possessed meso-␥- (order
exemplifies a warm-season bow echo, and is character- of 10 km) and meso-␤-scale (order of 100 km) hetero-
ized by considerable heterogeneity on multiple scales. geneities in the form of convective outflow and an air-
The 24 October 2001 event is a cool-season, squall-line mass boundary, respectively, and mesovortices occur
bow echo that formed in the relative absence of meso- before any significant interaction with these environ-
scale heterogeneity, but was strongly forced on the syn- mental features. With the 24 October 2001 QLCS
optic scale. event, mesovortices readily form within a relatively ho-
The simulations replicate the characteristics of highly mogeneous mesoscale environment, as gauged by the
organized, severe QLCSs, and thus capture the salient magnitude of the ␪e gradient, and their strengths were
features of the observed systems. For the 6 July 2003 comparable to those of its warm-season counterpart.
QLCS event, the model reproduces a mature, extensive This behavior, whereby mesovortices form as a conse-
bow echo with a horizontal scale well in excess of 100 quence of the intrinsic dynamics of the system, is con-
km. On the system-scale, this simulated bow system sistent with previous idealized numerical simulations of
possesses a dominant cyclonic vortex and weaker anti- highly organized QLCSs (see Trapp and Weisman
cyclonic vortex on its northern and southern flanks, 2003).
respectively, as well as a concentrated RIJ tens of kilo- The genesis mechanisms of the low-level mesovorti-
meters rearward of the leading-edge convection. For ces are dissimilar between the two cases. For the warm-
the 24 October 2001 QLCS event, the model repro- season case, mesovortices are formed by the tilting of
duces a more dynamically forced squall line with a hori- crosswise horizontal baroclinic vorticity in downdrafts,
zontal scale of several hundred kilometers. Comparable consistent with the idealized modeling results of Trapp
with the observed system, a number of embedded bow- and Weisman (2003). For the cool-season case, tilting of
ing segments occur over the northern half of Indiana the strong environment horizontal vorticity produces a
and extreme southern Michigan. vortex sheet, within which mesovortices appear to form
In both cases on the subsystem-scale, several low- through the release of a horizontal shearing instability.
level mesovortices with horizontal scales of 5–10 km In each case, vortex intensification is a consequence of
form on the simulated systems’ leading edges. The posi- the stretching.
tive vertical vorticity associated with these circulations Mesoscale heterogeneity can still have an effect on
is maximized near the surface, and extends into the the dynamics of QLCSs, however. The nature of such
midlevels of the troposphere. system–environment interaction appears dependent
External heterogeneity is shown to play no detect- upon the scale of the heterogeneity. In the case of the
able role in mesovortex genesis, which runs counter to 6 July 2003 simulated bow system, a leading-edge meso-
what has been suggested by previous observational vortex undergoes rapid intensification after interacting

Unauthenticated | Downloaded 03/13/24 12:05 PM UTC


4238 MONTHLY WEATHER REVIEW VOLUME 136

FIG. 15. Horizontal cross sections, at 0000 UTC 6 July 2003, of (a) CAPE for the parcel in each column with maximum equivalent
potential temperature below 3000 m AGL (i.e., most unstable CAPE), (b) CIN for these same parcels, (c) every ninth 0–3-km shear
vector and corresponding contours, and (d) every ninth 0–6-km shear vector and corresponding contours. In (a) and (b), every ninth
ground-relative, horizontal wind vector from the lowest model level has been plotted. In (a)–(d), full and half barbs denote speeds of
5 and 2.5 m s⫺1 respectively; flags indicate speeds of 25 m s⫺1.

with meso-␥-scale heterogeneity in the form of convec- Meso-␤-scale heterogeneity in the form of interacting
tive outflow. Trajectory analysis shows the boundary to air masses provides for along-line variations in the
be the predominant source region for air parcels popu- strength of low- to midlevel vertical wind shear and the
lating the vortex region at peak intensity, many of amount of thermodynamic instability as the simulated
which possess horizontal vorticity of the order of 0.05 bow system moves across eastern Nebraska and west-
s⫺1. However, vertical vorticity equation analysis shows ern Iowa. North of the airmass boundary, stronger low-
that vortex intensification owes to the stretching of ver- to midlevel vertical wind shear, as well as larger
tical vorticity brought about by enhanced convergence amounts of CAPE, help to maintain lift along the gust
at the point of QLCS–boundary interaction. The time- front during the simulated system’s mature stage, de-
integrated contribution to vertical vorticity from the spite a strengthening cold pool. This system-scale be-
tilting term is negligible, in contrast with current con- havior feeds back onto the subsystem scale. Mesovor-
ceptual models of QLCS–boundary interactions. tices north of the boundary experience deeper lifting

Unauthenticated | Downloaded 03/13/24 12:05 PM UTC


NOVEMBER 2008 WHEATLEY AND TRAPP 4239

FIG. 16. The 40-km line-averaged vertical cross sections of potential temperature and circulation vectors through the (a) northern and
(b) southern ends of the system at 0300 UTC 6 Jul 2003. The horizontal components of circulation vectors are scaled such that a
magnitude of 25 m s⫺1 exactly reaches the tail of the next adjacent vector. The maximum vertical vectors in (a) and (b) are 10.8 and
6.6 m s⫺1 (see bottom-right corner of each plot), respectively.

FIG. 17. Mesovortex tracks for the periods (a) 0000–0400 UTC 6 Jul 2003 and (b) 2100 UTC 24 Oct–0000 UTC 25 Oct 2001. Track
plots represent the sum of 10-min positive vertical vorticity fields, for values greater than 10⫺2 s⫺1, over the periods of interest. In (a),
the shear-no shear line is derived from Fig. 15d.

Unauthenticated | Downloaded 03/13/24 12:05 PM UTC


4240 MONTHLY WEATHER REVIEW VOLUME 136

and achieve greater maximum vertical vorticities, while surface/hydrology model with the Penn State/NCAR MM5
those on the southern end of the line (within the nomi- modeling system. Part I: Model description and implementa-
tion. Mon. Wea. Rev., 129, 569–585.
nal shear region) are weaker and shorter lived.
Chen, S.-H., and W.-Y. Sun, 2002: A one-dimensional time de-
The present study has evidenced the effect of meso- pendent cloud model. J. Meteor. Soc. Japan, 80, 99–118.
scale heterogeneity on the rotational dynamics of Coniglio, M. C., and D. J. Stensrud, 2001: Simulation of a progres-
QLCSs, but the nature as well as sheer number of such sive derecho using composite initial conditions. Mon. Wea.
storm–environment interactions cannot possibly be de- Rev., 129, 1593–1616.
Davies-Jones, R. P., 1982a: A new look at the vorticity equation
scribed through analysis of two simulated events. To-
with application to tornadogenesis. Preprints, 12th Conf. on
ward a better understanding of these processes, data Severe Local Storms, San Antonio, TX, Amer. Meteor. Soc.,
assimilation experiments are now being performed with 249–252.
an ensemble Kalman filter (EnKF) and involve the use ——, 1982b: Observational and theoretical aspect of tornadogen-
of routine surface and upper-air observations. This part esis. Intense Atmospheric Vortices, L. Bengtsson and J. Light-
of the study will further consider the effect of mesoscale hill, Eds., Springer-Verlag, 175–189.
——, and H. E. Brook, 1993: Mesocyclogenesis from a theoretical
heterogeneity from the viewpoint of system initiation, perspective. The Tornado: Its Structure, Dynamics, Predic-
and thus will concentrate on seemingly unpredictable tion, and Hazards, Geophys. Monogr., Vol. 79, Amer. Geo-
QLCS events in which significant mesoscale variability phys. Union, 105–114.
appears to have played a role in system initiation and/or Davis, C., and Coauthors, 2004: The Bow Echo and MCV Experi-
evolution. ment: Observations and opportunities. Bull. Amer. Meteor.
Soc., 85, 1075–1093.
Dudhia, J., 1989: Numerical study of convection observed during
Acknowledgments. The authors wish to thank the the Winter Monsoon Experiment using a mesoscale two-
three anonymous reviewers for their constructive com- dimensional model. J. Atmos. Sci., 46, 3077–3107.
ments. Resources made available through NCAR’s Sci- Fujita, T. T., 1981: Tornadoes and downbursts in the context of
entific Computing Division and Purdue University’s generalized planetary scales. J. Atmos. Sci., 38, 1511–1534.
Rosen Center for Advanced Computing were used to Jorgensen, D. P., and B. F. Smull, 1993: Mesovortex circulations
seen by airborne Doppler radar within a bow-echo mesoscale
generate the WRF model simulations. This research convective system. Bull. Amer. Meteor. Soc., 74, 2146–2157.
was supported in part by the National Science Founda- Kain, J. S., 2004: The Kain–Fritsch convective parameterization:
tion, under Grant ATM-023344 (DMW and RJT), and An update. J. Appl. Meteor., 43, 170–181.
by a Bilsland Dissertation Fellowship (DMW). Klimowski, B. A., R. Przybylinski, G. Schmocker, and M. R.
Hjelmfelt, 2000: Observations of the formation and early evo-
lution of bow echoes. Preprints, 20th Conf. on Severe Local
REFERENCES
Storms, Orlando, FL, Amer. Meteor. Soc., 44–47.
Atkins, N. T., M. L. Weisman, and L. J. Wicker, 1999: The influ- Lee, B. D., and R. B. Wilhelmson, 1997a: The numerical simula-
ence of preexisting boundaries on supercell evolution. Mon. tion of nonsupercell tornadogenesis. Part I: Initiation and
Wea. Rev., 127, 2910–2927. evolution of pretornadic misocyclone circulations along a dry
——, J. M. Arnott, R. W. Przybylinski, R. A. Wolf, and B. D. Ket- outflow boundary. J. Atmos. Sci., 54, 32–60.
cham, 2004: Vortex structure and evolution within bow ech- ——, and ——, 1997b: The numerical simulation of nonsupercell
oes. Part I: Single-Doppler and damage analysis of the 29 tornadogenesis. Part II: Evolution of a family of tornadoes
June 1998 derecho. Mon. Wea. Rev., 132, 2224–2242. along a weak outflow boundary. J. Atmos. Sci., 54, 2387–2415.
——, C. S. Bouchard, R. W. Przybylinski, R. J. Trapp, and G. Maddox, R. A., L. R. Hoxit, and C. F. Chappell, 1980: A study of
Schmocker, 2005: Damaging surface wind mechanisms within tornadic thunderstorm interactions with thermal boundaries.
the 10 June 2003 Saint Louis bow echo during BAMEX. Mon. Wea. Rev., 108, 322–336.
Mon. Wea. Rev., 133, 2275–2296. Markowski, P. M., E. N. Rasmussen, and J. M. Straka, 1998: The
Balaji, V., and T. L. Clark, 1988: Scale selection in locally forced occurrence of tornadoes in supercells interacting with bound-
convective fields and the initiation of deep cumulus. J. At- aries during VORTEX-95. Wea. Forecasting, 13, 852–859.
mos. Sci., 45, 3188–3211. Miles, J. W., and L. N. Howard, 1964: Note on heterogeneous
Bernardet, L. R., and W. R. Cotton, 1998: Multiscale evolution of shear flow. J. Fluid Mech., 20, 331–336.
a derecho-producing mesoscale convective system. Mon. Mlawer, E. J., S. J. Taubman, P. D. Brown, M. J. Iacono, and S. A.
Wea. Rev., 126, 2991–3015. Clough, 1997: Radiative transfer for inhomogeneous atmo-
Brady, R. H., and E. J. Szoke, 1989: A case study of nonmesocy- sphere: RRTM, a validated correlated-k model for the long-
clone tornado development in northeast Colorado: Similari- wave. J. Geophys. Res., 102 (D14), 16 663–16 682.
ties to waterspout formation. Mon. Wea. Rev., 117, 843–856. Mueller, C. K., and R. E. Carbone, 1987: Dynamics of a thunder-
Burgess, D. W., and B. F. Smull, 1990: Doppler radar observations storm outflow. J. Atmos. Sci., 44, 1879–1898.
of a bow echo associated with a long-track severe windstorm. NCDC, 2001: Storm Data. Vol. 43, No. 10, 172 pp. [Available from
Preprints, 16th Conf. on Severe Local Storms, Kananaskis National Climatic Data Center, 151 Patton Ave., Asheville,
Park, AB, Canada, Amer. Meteor. Soc., 203–208. NC 28801-5001.]
Carbone, R. E., 1983: A severe frontal rainband. Part II: Tornado ——, 2003: Storm Data. Vol. 45, No. 7, 478 pp. [Available from
parent vortex circulation. J. Atmos. Sci., 40, 2639–2654. National Climatic Data Center, 151 Patton Ave., Asheville,
Chen, F., and J. Dudhia, 2001: Coupling an advanced land- NC 28801-5001.]

Unauthenticated | Downloaded 03/13/24 12:05 PM UTC


NOVEMBER 2008 WHEATLEY AND TRAPP 4241

Noh, Y., W. G. Cheon, S.-Y. Hong, and S. Raasch, 2003: Improve- search WRF Version 2. NCAR Tech. Note TN-468⫹STR,
ment of the K-profile model for the planetary boundary layer 88 pp.
based on large eddy simulation data. Bound.-Layer Meteor., Thorpe, A. J., M. J. Miller, and M. W. Moncrieff, 1982: Two-
107, 401–427. dimensional convection in non-constant shear: A model of
Przybylinski, R. W., G. K. Schmocker, and Y.-J. Lin, 2000: A mid-latitude squall lines. Quart. J. Roy. Meteor. Soc., 108,
study of storm and vortex morphology during the “intensify- 739–762.
ing stage” of severe wind mesoscale convective systems. Pre- Trapp, R. J., and M. L. Weisman, 2003: Low-level mesovortices
prints, 20th Conf. on Severe Local Storms, Orlando, FL, within squall lines and bow echoes. Part II: Their genesis and
Amer. Meteor. Soc., 173–176. implications. Mon. Wea. Rev., 131, 2804–2823.
Rasmussen, E. N., J. M. Straka, R. P. Davies-Jones, C. A. Doswell Wakimoto, R. M., and J. W. Wilson, 1989: Non-supercell torna-
III, F. H. Carr, M. D. Eilts, and D. R. MacGorman, 1994: The does. Mon. Wea. Rev., 117, 1113–1140.
Verifications of the Origins of Rotation in Tornadoes Experi- ——, H. V. Murphey, C. A. Davis, and N. T. Atkins, 2006: High
ment: VORTEX. Bull. Amer. Meteor. Soc., 75, 997–1006.
winds generated by bow echoes. Part II: The relationship
Richardson, Y. P., K. K. Droegemeier, and R. P. Davies-Jones, between the mesovortices and damaging straight-line winds.
2007: The influence of horizontal environmental variability
Mon. Wea. Rev., 134, 2813–2829.
on numerically simulated convective storms. Part I: Varia-
Weisman, M. L., 1992: The role of convectively generated rear-
tions in vertical shear. Mon. Wea. Rev., 135, 3429–3455.
inflow jets in the evolution of long-lived mesoconvective sys-
Rotunno, R., and J. B. Klemp, 1985: On the rotation and propa-
tems. J. Atmos. Sci., 49, 1826–1847.
gation of simulated supercell thunderstorms. J. Atmos. Sci.,
42, 271–292. ——, 1993: The genesis of severe, long-lived bow echoes. J. At-
mos. Sci., 50, 645–670.
——, ——, and M. L. Weisman, 1988: A theory for strong, long-
lived squall lines. J. Atmos. Sci., 45, 463–485. ——, and C. Davis, 1998: Mechanisms for the generation of me-
Schmocker, G. K., R. W. Przybylinski, and E. N. Rasmussen, soscale vortices within quasi-linear convective systems. J. At-
2000: The severe bow echo event of 14 June 1998 over the mos. Sci., 55, 2603–2622.
mid-Mississippi Valley region: A case of vortex development ——, and R. J. Trapp, 2003: Low-level mesovortices within squall
near the intersection of a preexisting boundary and a convec- lines and bow echoes. Part I: Overview and dependence on
tive line. Preprints, 20th Conf. on Severe Local Storms, Or- environmental shear. Mon. Wea. Rev., 131, 2779–2803.
lando, FL, Amer. Meteor. Soc., 169–172. ——, W. C. Skamarock, and J. B. Klemp, 1997: The resolution
Skamarock, W. C., M. L. Weisman, and J. B. Klemp, 1994: Three- dependence of explicitly modeled convective systems. Mon.
dimensional evolution of simulated long-lived squall lines. J. Wea. Rev., 125, 527–548.
Atmos. Sci., 51, 2563–2584. Wheatley, D. M., R. J. Trapp, and N. T. Atkins, 2006: Radar and
——, J. B. Klemp, J. Dudhia, D. O. Gill, D. M. Barker, W. Wang, damage analysis of severe bow echoes observed during
and J. G. Powers, 2005: A description of the Advanced Re- BAMEX. Mon. Wea. Rev., 134, 791–806.

Unauthenticated | Downloaded 03/13/24 12:05 PM UTC

You might also like