P1.2 MERGERS BETWEEN ISOLATED SUPERCELLS ANDQUASI-LINEAR CONVECTIVE SYSTEMS: A PRELIMINARY STUDY

Adam J. French∗ and Matthew D. ParkerNorth Carolina State University, Raleigh, North Carolina

1. INTRODUCTIONPast research within the convective storms commu-nity has often viewed supercell thunderstorms andlarger-scale quasi-linear convective systems (QLCSs,commonly referred to as �squall lines�) as distinctentities to be studied in isolation. While this haslead to immeasurable advances in the understand-ing of supercells and squall lines it overlooks the factthat these organizational modes can be found occur-ring simultaneously, in close proximity to each other(e.g. French and Parker 2008), often resulting inthe merger of multiple storms. While the basic con-cept of mergers between convective cells has beenstudied at great length (e.g. Simpson et al. 1980,Westcott 1994), a number of questions remain re-garding the processes at work when well-organizedconvective storms merge and interact. Of particu-lar interest is what this might mean in terms of theproduction of severe weather including large hail,damaging wind, and tornadoes. With this in mind,the present study provides a �rst step in an investi-gation of the dynamics that govern mergers betweensupercells and QLCSs, through a review of severalmerger cases.

2. BACKGROUNDPast work on interactions between supercells andother storms, including other supercells, single ormulti-cellular storms, and squall lines have lead to anongoing debate as to whether mergers favor or hindersupercellular organization and tornado production.Modeling studies by Bluestein and Weisman (2000),Finley et al. (2001), and Jewett and Wilhelmson(2006) have suggested that mergers can be detrimen-tal to long-lived supercell structures, with enhancedprecipitation causing strong cold pools and evolutiontowards linear modes. Non-favorable mergers have

∗Corresponding author address: Adam J. French, Depart-ment of Marine, Earth, & Atmospheric Sciences, North Car-olina State University, Campus Box 8208, Raleigh, NC 27695�8208. E-mail: ajfrench@ncsu.edu

been observed in actual cases as well. In the case dis-cussed by Knupp et al. (2003) two di�erent supercellstorms were observed to permanently lose supercel-lular characteristics following mergers with a left-moving split from an earlier storm. In a similar case,Lindsey and Bunkers (2005) examined a merger be-tween a left-moving supercell and a tornadic right-mover, �nding that the merger appeared to disruptthe circulation of the right-mover and temporarilysuspended tornado production. Furthermore, in astudy examining long-lived supercells, Bunkers et al.(2006) listed �the supercell merges or interacts withother thunderstorms which can either destroy its cir-culation or cause it to evolve into another convec-tive mode� as one mechanism for supercell demise,however they also pointed out that �not all stormmergers were destructive�. Indeed, Lee et al. (2006)observed that as many of 60 % of cell mergers ob-served in the 19 April 1996 case resulted in increasedcell rotation, and 54% of the tornadoes on this dayoccurred within 15 minutes of a merger. Thus theimpact of cell mergers on supercell longevity is farfrom certain.

This uncertainty carries over to cases of supercell-squall line mergers as well. A number of cases citedin the literature pertain to events wherein a mergerbetween a squall line (often exhibiting bow-echocharacteristics) and a supercell facilitates tornado-genesis (e.g Goodman and Knupp 1993; Saboneset al. 1996; Wolf et al. 1996, Wolf 1998, and LaPentaet al. 2005). In many of these cases the supercell isobserved to remain coherent following the merger,with the merged system taking on a �comma� or�S�-shaped re�ectivity pattern (Sabones et al. 1996,Wolf 1998). This is considered by some to be a sur-prising result, as the expectation would be that theout�ow from the squall line would undercut the su-percell's updraft and disrupt its mesocyclone (Wolf1998). Indeed, other cases appear to exemplify thismore �expected� behavior whereby bow-echo super-cell mergers lead to a rapid decrease in supercell or-ganization, (Calianese et al. 2002). This sometimeseven occurs within an event where favorable interac-

1

tions were also observed (Wolf et al. 1996). Addingcomplexity to this debate is the frequently observedincrease in both storm strength and rotation priorto storm mergers, suggesting that the two stormsmay be interacting in some way before they actuallymerge (Przybylinski 1995; Wolf et al. 1996; LaPentaet al. 2005).

In short, a survey of past research reveals caseswhere squall line-supercell interactions lead to an en-hancement or at least persistence of supercell struc-tures, as well as those that appear to destroy pre-existing supercell structures. In the cases wherethe supercell is able to persist, several cases suggestthat tornadogenesis is instigated by the merger (e.g.Goodman and Knupp 1993; Sabones et al. 1996).While this dichotomy of behaviors has been well doc-umented in the literature, it remains unclear as towhy some mergers favor sustained supercell struc-tures, while other destroy them. The present studylooks address this by delving deeper into the dy-namic processes at work during merger events. Thispaper outlines the �rst step in achieving this goal,focusing on identifying key characteristics of thesemerger events, both via an examination of past liter-ature, as well as through analysis of new cases. Thiswill set the stage for numerical simulations, whichwill provide a means to examine the dynamic pro-cesses at work during these mergers, aiming to pro-vide better insight as to what may cause a supercell-squall line merger to be favorable vs unfavorable forthe supercell's circulation.

3. METHODSThis work utilized operationally available datafrom across the continental United States. Caseswere identi�ed using a combination of the StormPrediction Center's Online Severe ThunderstormEvents Archive (http://w1.spc.woc.noaa.gov/exper/archive/events/) and the NCAR MMM OnlineImage Archive (http://www.mmm.ucar.edu/imagearchive/) to search for events that featured iso-lated supercells and squall lines between January2000 and June 2009. Initially, the search domainfocused solely on Oklahoma, in the interest of be-ing able to exploit a larger number of observationaldatasets. However, in the interest of increasing thenumber of cases, other events from around the coun-try were included as well. These latter cases wereidenti�ed from past investigations of multi-mode(squall line and supercell) cases done by the �rstauthor, rather than through a systematic search re-lated to the present work. An initial list of 63 cases,was narrowed to 16 �most promising� cases for which

SBCAPE: 1250 J/kg

SBCIN: -4 J/kg

MLCAPE: 853 J/kg

MLCIN: -1 J/kg

0-6 km shear: 30 m/s

Shear direction: 74 deg.

0-1 km SRH: 287 m2/s2

0-3 km SRH: 320 m2/s2

SBCAPE: 2771 J/kg

SBCIN: -3 J/kg

MLCAPE: 2174 J/kg

MLCIN: -29 J/kg

0-6 km shear: 35 m/s

Shear direction: 65 deg.

0-1 km SRH: 176 m2/s2

0-3 km SRH: 325 m2/s2

a)

b)

Figure 1: Skew-T log-P diagrams and hodographs represent-ing the background environments for the four cases presented:a) 10 November 2002 1800 UTC sounding from Wilming-ton, OH, b) 24 May 2008 0000 UTC sounding from DodgeCity, KS, c) 11 February 2009 0000 UTC sounding from FortWorth, TX, and d) 19 April 2009 0000 UTC sounding fromLamont, OK. Hodographs for the tropospheric wind pro�leare included in the upper right corner of each sounding, alongwith a wind pro�le, both of which are in units of m/s (halfbarb = 5 ms, full barb = 10 ms, �ag = 25 ms. Calculatedsurface based (SB) and 100 mb mixed layer (ML) CAPE andCIN are provide for each sounding, along with 0-6 km bulkshear and 0-1 and 0-3 km SRH.

2

Level II WSR-88D single site radar data were ob-tained. An initial analysis of the Level II data fur-ther narrowed the list to 8 archetypal cases thatcontained well-organized supercell and QLCS modesthat merged during their lifetimes. Other factorsincluding availability of radar data throughout theevent, and proximity to the radar were also takeninto account in �nalizing case selection.

For the initial analysis presented in this work,base re�ectivity and base velocity WSR-88D datawere examined for each case to provide an overviewof storm evolution before, during and after themerger(s) between the squall line and super-cell(s) of interest. In addition, a general pic-ture of the background environment was ascer-tained using operationally available surface andupper air observations and archived SPC Meso-analysis graphics (for events after 2006, Bothwellet al. 2002). Finally, data from the NCDC StormData archive (http://www4.ncdc.noaa.gov/cgi-win/wwcgi.dll?wwEvent~Storms) were also used to pro-vide details on instances of severe weather (namelytornadoes) that occurred in conjunction with thestorms involved in the merger.

4. CASESIn the interest of providing more detail, a sub-set of4 of the more representative cases will be presentedin this paper. These cases were selected as they il-lustrate several key commonalities seen between all8 cases, while also each also serving to illustrate aunique characteristic of interest. A short summaryof each case will be provided below, followed by adiscussion of some of the common themes identi�edfrom the initial analysis of all 8 cases in the followingsection.

4.1 10 November 200210 November 2002 was witness to a widespread out-break of severe weather and tornadoes across theeastern half of the United States, from the GreatLakes to the Gulf Coast. The present analysis willfocus on the early part of this event, particularlythe interaction between a supercell and squall line inwestern Ohio. The synoptic environment was char-acterized by a strong, negatively tilted upper-leveltrough over the central United States, with an at-tendant surface low pressure system over northernMichigan and a surface cold front extending from theGreat Lakes to the Gulf coast. An 18 UTC soundingfrom Wilmington, Ohio (Fig. 1a) contained 30 ms−1of 0-6 km shear and 1250 J kg−1 of surface-based

SBCAPE: 1844 J/kg

SBCIN: 0.0 J/kg

MLCAPE: 1396 J/kg

MLCIN: 0 J/kg

0-6 km shear: 34 m/s

Shear direction: 61 deg.

0-1 km SRH: 267 m2/s2

0-3 km SRH: 240 m2/s2

c)

d)

SBCAPE: 966 J/kg

SBCIN: -13 J/kg

MLCAPE: 746 J/kg

MLCIN: -2 J/kg

0-6 km shear: 28 m/s

Shear direction: 30 deg.

0-1 km SRH: 106 m2/s2

0-3 km SRH: 165 m2/s2

Figure 1: (Continued)

3

KIWX

10 Nov. 2002

2000 UTC

KIWX

10 Nov. 2002

2025 UTC

KIWX

10 Nov. 2002

2045 UTC

KIWX

10 Nov. 2002

2104 UTC

supercell

supercell

supercell

remnants

of

supercell

KIWX KIWX

KIWX KIWX

a) b)

c) d)

Figure 2: Time series of radar re�ectivity images from theNorth Webster, IN (KIWX) WSR-88D for 10 November 2002.Radar re�ectivity is shaded per the color bar on the left ofeach image, and times are included in the lower-right corner.

convective available potential energy (SBCAPE), fa-voring supercells (Rasmussen and Blanchard 1998).A 0-1 km storm-relative helicity (SRH) of 287 m2s−2further favored storm rotation and indicated thattornadoes were possible (Rasmussen and Blanchard1998).

A squall line began developing along the cold frontin central/eastern Indiana during the late morninghours on the 10th and progressed eastward. By1930 UTC, as the line approached the Ohio/Indianaborder, a supercell developed approximately 40 kmahead of the line in eastern IN and raced northeast-ward, parallel to the line (not shown). A complexinteraction between the supercell and line began byapproximately 20 UTC (Fig. 2a). North and westof the supercell, the squall line began to weaken andlag in its eastward motion. Meanwhile, extendingsouthwestward from the supercell, the line appearedto surge eastward and intensify (Fig. 2b-c). Thisoverall pattern continued for approximately an hour,with the supercell fully merging with the northernend of the line by 21 UTC (Fig. 2d). During thisperiod, the line surged east of the supercell to itssouth, however a clear separation between the tworemained evident in both the WSR-88D re�ectiv-ity (Fig. 2c) and velocity data (not shown). Duringthis same period, the supercell produced a long-tracktornado with damage rated at F-4 on the Fujita tor-

nado intensity scale1. Following the merger, whichappeared to result from squall line out�ow overtak-ing the supercell, supercell characteristics in bothre�ectivity and velocity data are quickly lost as it isabsorbed by the line (Fig. 2d).

4.2 23 May 2008A widespread severe weather and tornado outbreakoccurred across southwestern Kansas on 23 May2008, consisting of several intense, long-lived super-cells two of which eventually merged with a squallline. The background environment strongly favoreda supercellular mode, with 35 ms−1 of 0-6 km shear,325 m2s−2 0-3 km SRH (notably 0-1 km SRH wasapproximately half that, 176 m2s−2), and 2771 Jkg−1 of SBCAPE (Fig. 1b). An intense upper levellow to the west, along with a surface dryline and coldfront in the region promoted widespread convectiveinitiation. The initial convective mode across west-ern Kansas consisted of multiple isolated supercells,one of which produced an EF-4 tornado in west cen-tral Kansas. By 0000 UTC, a squall line began to de-velop in northwestern Kansas and subsequently grewsouthward along a north/south boundary. Ahead ofthis line, isolated supercells continued to be favored,with 3 supercells becoming dominant just east of thedeveloping squall line. Of these, one dissipated as anisolated storm, while the other two merged with theline.

The �rst of these mergers facilitated a complexstorm evolution as it occurred. As the supercell(S1 in Fig. 3a-d) approached the squall line, theline broke northwest of the supercell. S1 mergedwith the segment of the line south of this break(Fig. 3a-b), producing an EF-2 tornado at the ap-proximate merger time. Following the merger, S1remained de�nable in both re�ectivity (Fig. 3b-d)and velocity �elds for at least an hour, appearing toenhance straight-line winds within the line to thesouth. Meanwhile, the portion of the squall linenorth of the break, very rapidly evolved into a su-percell (S3 in Fig. 3b-c), developing strong rotation,and producing an EF-2 tornado of its own beforeeventually dissipating.

The other pre-line supercell (S2 in Fig. 3a-f) con-tinued to persist ahead of the line while the �rstmerger took place (Fig. 3a-d). It moved north andparallel to the line, and produced an EF-3 tornadothat was on the ground for 35 km, which was the

1Depending on the case both the original Fujita scale (Fu-jita 1971) as well as the newer, �enhanced� Fujita (EF) scale(McDonald and Mehta 2006) are used to describe tornadodamage, based on which was used during the damage survey.

4

longest track surveyed during the event. EventuallyS2 merged with the remnants of S1 at the north endof the line, maintaining its supercellular character-istics for about an hour and enhancing straight-linewinds to its south as well (Fig. 3e-f). Notably, thethe termination of the tornado damage associatedwith S2 appears to roughly coincide with the merger,and no further tornado damage was reported withthis particular storm following the merger.

KDDC

24 May 2008

0303 UTC

KDDC

24 May 2008

0321 UTC

KDDC

24 May 2008

0345 UTC

KDDC

24 May 2008

0408 UTC

KDDC

24 May 2008

0431 UTC

KDDC

24 May 2008

0459 UTC

S1 S2

S3

S1

S1

S1

S2

S2

S2

S2S2

S3

S3

KDDC KDDC

KDDCKDDC

KDDCKDDC

a) b)

c) d)

e) f )

Figure 3: As in Fig. 2, but from the Dodge City, KS (KDDC)radar for the 23 May 2008 case.

4.3 10 February 2009The 10 February 2009 event was a cold-season severeweather outbreak across central Oklahoma, consist-ing of several isolated supercells early on, and a su-percell that merged with a squall line later in theevent. The large scale environment featured strongupper-level dynamics and favorable conditions forsupercells (34 ms−1 0-6 km shear, 240 m2s−20-3 kmSRH, 1844 J kg−1 SBCAPE, Fig. 1c), along with asurface cold front that forced the squall line. Early

in the event, the dominant convective mode was iso-lated supercells, several of which produced EF1/2tornadoes in the Oklahoma City metro area (notshown). Continued convective initiation and cellmergers along a surface cold front eventually lead tothe development of a squall line across western Ok-lahoma. As the squall line matured, a single isolatedsupercell developed near the Oklahoma/Texas bor-der and moved north into south-central Oklahoma(Fig. 4a). The supercell rapidly intensi�ed and wenton to produce EF-2 and EF-4 tornado damage alonga 33 mile track prior to its merger with the line (Fig.4a-c). The line weakened overall prior to the merger,however there was no distinct break as was observedsome other cases (Fig. 4c). The supercell main-tained its identity following the merger, remainingevident as a region of locally intense re�ectivity (Fig.4c-d) with attendant mesoscale circulation noted inthe velocity data (not shown). Additionally, a kinkin the the line developed at the approximate loca-tion of the merger, with the portion of the line southof the kink being displaced further east after themerger took place (Fig. 4d). There was no furthertornado activity following the merger.

4.4 18 April 2009The �nal case, 18 April, 2009 di�ers from the oth-ers in that it was not associated with a widespreadsevere weather outbreak. Rather, it consisted of alocalized event in north central Oklahoma consistingof a supercell and squall line that developed nearlyconcurrently and merged shortly thereafter. Thesynoptic environment was characterized by a broadupper level trough over the central plains with a sur-face low-pressure center located in north-central Ok-lahoma and cold front extending to the southwest.The 0000 UTC sounding from Lamont, Oklahoma(Fig. 1d) suggests an environment supportive of su-percells with 28 ms−1 of 0-6 km shear and 966 Jkg−1 of CAPE. Additionally, 0-1 and 0-3 SRH werecomparatively weaker than in the other cases (106and 165 m2s−2respectively) which may explain thelack of tornado production associated with this case.

An isolated supercell developed just ahead of acluster of cells that rapidly began evolving into asmall line (Fig. 5a). As both features continued toorganize they began to merge, with the line ostensi-bly overtaking the supercell from behind (Fig. 5b).The merger occurred just as the supercell's meso-cyclone appeared to be strengthening and the cir-culation contracting, however it weakened consid-erably as the merger took place, only to re-emergeas a strong, but broader circulation shortly there-

5

after (not shown). This circulation, located at thenorth end of the developing line, appears qualita-tively similar to a book-end vortex (e.g. Weisman1993, Weisman and Davis 1998, Atkins et al. 2004),featuring both a broad, strong rotation as well asa spiral pattern in the re�ectivity �eld (Fig. 5c).Straight-line winds increased to the south of thiscirculation (based on Doppler velocities) and bow-ing of the convective line was evident within the re-�ectivity �eld in this region as well (Fig. 5c). Thisfeature persisted for over an hour as the line intensi-�ed and progressed across central OK (Fig. 5d). Ofparticular interest in this case is the relative youthof both systems at the time of the merger, as theydeveloped concurrently and then merged within ap-proximately 1 hour of initial development. This iscompared to some of the other cases wherein the su-percell and/or squall line evolved individually for along period prior to merging.

5. COMMONALITIES BETWEEN CASESThe primary motivation for this initial study wasto examine a number of squall line-supercell mergercases to determine useful areas of focus for future,more detailed analysis and study. With this in mind,several common features that stood out between the8 case examined are summarized below.

• In just about every case, the supercell re-mains identi�able for at least an hour follow-ing the merger, and appears to have a signi�-cant impact on the evolution of the squall line.This would tend to suggest that the super-cell, rather than the squall line �dominates�the merger. As noted by Wolf (1998) this is asomewhat surprising result as the squall line'sout�ow appears to have little impact in termsof disrupting the mesocyclonic circulation ofthe supercell.

• In some cases the merger is preceded by a splitor break in the line, which may play a role inpromoting the longevity of the supercell fol-lowing the merger. In what is perhaps an ex-treme case, on 23 May 2008 this appeared tofacilitate the development of a new supercellwithin what had been the northern portion ofthe line.

• Post-merger evolution appears to vary basedon where along the squall line the merger oc-curs. For mergers that occurred approximatelyhalfway along the squall line, a break in theline sometimes preceded the merger, however

the line maintained its structure and intensitynorth and south of the merger. The char-acteristics of the supercell, including a cy-clonic circulation, were maintained, and gen-erally became embedded within the line fol-lowing the merger. When the merger occurredcloser to the north end of the line, a weak-ening of the line north of the merger was of-ten observed, and typically the remnants ofthe supercell subsequently became the north-ern end of the intense convective line. In thisscenario, the supercell mesocyclone often fur-ther evolved into a broader cyclonic circulationthat appears at least qualitatively similar tothe �bookend vortices� described by Weisman(1993), Weisman and Davis (1998), and Atkinset al. (2004).

KTLX

11 Feb. 2009

0028 UTC

KTLX

11 Feb. 2009

0126 UTC

KTLX

11 Feb. 2009

0149 UTC

KTLX

11 Feb. 2009

0231 UTC

supercellsupercell

supercell

kink in

line

KTLX KTLX

KTLX KTLX

a) b)

c) d)

Figure 4: As in Fig. 2, but from the Oklahoma City, OK(KTLX) radar for the 10 February 2009 case.

• Given the apparent persistence of supercell cir-culations following the merger, it is of interestto examine their vorticity source. Does the cir-culation continue to ingest and tilt horizontalvorticity from the background environment, oris it able to feed o� baroclinically generatedvorticity from the squall line's cold pool? Inthis regard, it would be useful to explore thesimilarities between the circulations that re-sult from a supercell/squall line merger, andthe mesovorticies that are sometimes observedwithin squall lines absent such a merger (e.g.

6

Weisman and Trapp 2003; Trapp andWeisman2003; Wakimoto et al. 2006).

• In most of the cases where tornadoes werepresent, the longest-track (and in some casesmost intense) tornadoes of the event were pro-duced by supercells that evolved for a pro-longed period ahead of, but in close prox-imity to the squall line. For the most partthese path lengths were at least twice thoseproduced by isolated storms. It should benoted that the large scale environment wasfavorable for supercells/tornadoes (i.e. largeshear/SRH), and other tornadoes were presentin most cases, however the longest track stormswere commonly associated with a pre-line su-percell. This, in addition to past observationsof storm intensi�cation as supercells and squalllines converge (e.g. Przybylinski 1995; Wolfet al. 1996; LaPenta et al. 2005), suggest thatthe storms are having an in�uence on one an-other prior to merging. One possible hypothe-sis to this end is that the squall line is alteringthe pre-line environment in a manner that fa-vors supercell intensi�cation/long-lived torna-does.

• These cases contained instances where amerger appeared to instigate tornadogenesis orat least maintain an ongoing tornado as wellas instances where the merger appeared to ei-ther terminate an ongoing tornado, or failed toinstigate tornadogenesis. This would suggest,as seen in the past literature, that mergers donot clearly favor or hinder tornadogenesis, butrather the details of a given merger event maydetermine if a tornado will form, the speci�csof which are not clearly understood.

6. FUTURE WORKThe case studies presented herein represent the �rststep in a much more in-depth study of mergers be-tween squall lines and supercells. Subsequent workwill focus on delving deeper into these cases to ex-amine the underlying dynamics at work within thesetypes of events. To accomplish this, we plan toexploit available observations to the fullest extentpossible, while also using numerical simulations, asthese may provide the best means for addressingsome of the hypotheses outlined in the previous sec-tion. This is especially true with regards to isolatingsome of the dynamic processes at work. We intendto utilize a combination of idealized simulations forspeci�c, controlled hypothesis testing, while also em-

KVNX

18 Apr 2009

2114 UTC

KVNX

18 Apr 2009

2147 UTC

KVNX

18 Apr 2009

2212 UTC

KVNX

18 Apr 2009

2321 UTC

supercell

supercell

reflectivity

swirl

reflectivity

swirl

KVNX KVNX

KVNXKVNX

a) b)

c) d)

Figure 5: As in Fig. 2, but from the Vance Air Force Base,OK (KVNX) radar for the 18 April 2009 case.

ploying case-study simulations to provide a meansof further analyzing speci�c cases and bridging thegap between observations and the idealized simu-lations. Given their inherent complexity (i.e. thepresence of multiple convective modes), advancedtechniques such as the use of data assimilation, non-homogeneous idealized base-state conditions, andidealized representation of persistent linear forcingmechanisms are likely to be necessary to accuratelysimulate these types of events.Acknowledgments

The authors would like to thank the members ofthe Convective Storms Group at NC State Univer-sity for thoughtful discussion and case suggestionsin the course of this work. This work is funded byNSF grants ATM-0552154 and ATM-0758509.

ReferencesAtkins, N. T., J. M. Arnott, R. W. Przybylinski,R. A. Wolf, and B. D. Ketcham, 2004: Vortexstructure and evolution within bow echoes. part I:Single-doppler and damage analysis of the 29 june1998 derecho. Mon. Wea. Rev., 132, 2224�2242.

Bluestein, H. B. and M. L. Weisman, 2000: The in-teraction of numerically simulated supercells initi-ated along lines. Mon. Wea. Rev., 128, 3128�3149.

Bothwell, P. D., J. A. Hart, and R. L. Thomp-son, 2002, An integrated three-dimensional objec-

7

tive analysis scheme in use at the Storm PredictionCenter. Preprints, 21st Conference on Severe LocalStorms, San Antonio, TX, Amer. Meteor. Soc.

Bunkers, M. J., M. R. Hjelmfelt, and P. L. Smith,2006: An observational examination of long-livedsupercells. Part I: Characteristics, evolution, anddemise. Wea. Forecasting, 21, 673�688.

Calianese, E. J., J. K. Jordan, E. B. Curran, A. R.Moller, and G. Woodall, 2002, The Mayfest high-precipitation supercell of 5 May 1995-a case study.Preprints, 21st Conference on Severe Local Storms,San Antonio, TX, Amer. Meteor. Soc.

Finley, C. A., W. R. Cotton, and R. A. Pielke,2001: Numerical simulation of tornadogenesis ina high-precipitation supercell. Part I: Storm evolu-tion and transition into a bow echo. J. Atmos. Sci.,58, 1597�1629.

French, A. J. and M. D. Parker, 2008: The initia-tion and evolution of multiple modes of convectionwithin a meso-alpha scale region. Wea. Forecasting,23, 1221�1252.

Fujita, T. T., 1971: Proposed characterization oftornadoes and hurricanes by area and intensity.Satellite and Mesometeorology Research ProjectReport 91, University of Chicago, 42 pp.

Goodman, S. J. and K. R. Knupp, 1993, Tornado-genesis via squall line and supercell interaction:The November 15, 1989, Huntsville, Alabama, tor-nado. The Tornado: Its Structure, Dynamics, Pre-diction, and Hazards. No. 79, Amer. Geophys.Union.

Jewett, B. F. and R. B. Wilhelmson, 2006: Therole of forcing in cell morphology and evolutionwithin midlatitude squall lines. Mon. Wea. Rev.,134, 3714�2734.

Knupp, K. R., S. Paech, and S. Goodman, 2003:Variations in cloud-to-ground lightning characteris-tics among three adjacent tornadic supercell stormsover the tennessee valley region. Mon. Wea. Rev.,131, 172�188.

LaPenta, K. D., L. F. Bosart, T. J. Galarneau, andM. J. Dickinson, 2005: A multiscale examinationof the 31 May 1998 Mechanicsville, New York, tor-nado. Wea. Forecasting, 20, 494�516.

Lee, B. D., B. F. Jewett, and R. B. Wilhelmson,2006: The 19 april 1996 Illinois tornado outbreak.Part II: Cell mergers and associated tornado inci-dence. Wea. Forecasting, 21, 449�464.

Lindsey, D. T. and M. J. Bunkers, 2005: Observa-tions of a severe, left-moving supercell on 4 May2003. Wea. Forecasting, 20, 15�22.

McDonald, J. R. and K. C. Mehta, 2006: A Recom-mendation for an Enhanced Fujita Scale, Submittedto the NWS and other interested users. [AvailableOnline at http://www.spc.noaa.gov/efscale/].

Przybylinski, R. H., 1995: The bow echo: Obser-vations, numerical simulations and severe weatherdetection methods. Wea. Forecasting, 10, 203�218.

Rasmussen, E. N. and D. O. Blanchard, 1998: Abaseline climatology of sounding�derived supercelland tornado forecast parameters. Wea. Forecasting,13, 1148�1164.

Sabones, M., E. M. Agee, and M. Akridge, 1996,The Pulaski county and West Lafayette, Indianatornadoes, 26-27 April 1994: A case of supercell(mesocyclone) and squall line bow-echo interaction.Preprints, 18th Conf. Severe Local Storms, SanFrancisco, CA, Amer. Meteor. Soc.

Simpson, J., N. E. Westcott, R. J. Clerman, andR. A. Pielke, 1980: On cumulus mergers. Arch. Me-teor. Geoph. Biokl. Ser. A., 29, 1�40.

Trapp, R. J. and M. L. Weisman, 2003: Low-levelmesovortices within squall lines and bow echoes.part II: Their genesis and implications. Mon. Wea.Rev., 131, 2804�2823.

Wakimoto, R. M., H. V. Murphey, C. A. Davis,and N. T. Atkins, 2006: High winds generatedby bow echoes. part II: The relationship betweenthe mesovortices and damaging straight-line winds.Mon. Wea. Rev., 134, 2813�2829.

Weisman, M. L., 1993: The genesis of severe, long�lived bow echoes. J. Atmos. Sci., 50, 645�670.

Weisman, M. L. and C. A. Davis, 1998: Mech-anisms for the generation of mesoscale vorticeswithin quasi-linear convective systems. J. Atmos.Sci., 55, 2603�2622.

Weisman, M. L. and R. J. Trapp, 2003: Low-levelmesovortices within squall lines and bow echoes.part I: Overview and dependence on environmentalshear. Mon. Wea. Rev., 131, 2779�2803.

Westcott, N. E., 1994: Merging of convectiveclouds: cloud initiation, bridging and subsequentgrowth. Mon. Wea. Rev., 122, 780�790.

8

Wolf, P. L., 1998: WSR-88D radar depiction ofsupercell-bow echo interaction: Unexpected evolu-tion of a large, tornadic "comma-shaped" supercellover eastern Oklahoma. Wea. Forecasting, 13, 492�504.

Wolf, R., R. Przybylinski, and P. Berg, 1996, Ob-servations of a merging bowing segment and super-cell. Preprints, 18th Conf. Severe Local Storms, SanFrancisco, CA, Amer. Meteor. Soc.

9