• Adlerman, E. J., and Droegemeier K. K. , 2000: A numerical simulation of cyclic tornadogenesis. Preprints, 20th Conf. on Severe Local Storms, Orlando, FL, Amer. Meteor. Soc., 591–594.

  • Agee, E. M., Snow J. T. , and Clare P. R. , 1976: Multiple vortex features in the tornado cyclone and the occurrence of tornado families. Mon. Wea. Rev., 104 , 552563.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Biggerstaff, M. I., and Coauthors, 2005: The Shared Mobile Atmospheric Research and Teaching Radar: A collaboration to enhance research and teaching. Bull. Amer. Meteor. Soc., 86 , 12631274.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bluestein, H. B., and Pazmany A. L. , 2000: Observations of tornadoes and other convective phenomena with a mobile, 3-mm wavelength, Doppler radar: The spring 1999 field experiment. Bull. Amer. Meteor. Soc., 81 , 29392951.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bluestein, H. B., and Weisman M. L. , 2000: The interaction of numerically simulated supercells initiated along lines. Mon. Wea. Rev., 128 , 31283149.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bluestein, H. B., and Wakimoto R. M. , 2003: Mobile radar observations of severe convective storms. Radar and Atmospheric Science: A Collection of Essays in Honor of David Atlas, Meteor. Monogr., No. 52, Amer. Meteor. Soc., 105–136.

    • Search Google Scholar
    • Export Citation
  • Brady, R. H., and Szoke E. J. , 1989: A case of nonmesocyclone tornado development in northeast Colorado: Similarities to waterspout formation. Mon. Wea. Rev., 117 , 843856.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Burgess, D. W., Wood V. T. , and Brown R. A. , 1982: Mesocyclone evolution statistics. Preprints, 12th Conf. on Severe Local Storms, San Antonio, TX, Amer. Meteor. Soc., 422–424.

  • Byers, H. R., and Braham R. R. Jr., 1949: The Thunderstorm Project. U.S. Government Printing Office, 287 pp.

  • Changnon S. A. Jr., , 1976: Effects of urban areas and echo merging on radar echo behavior. J. Appl. Meteor., 15 , 561570.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chin, H-N. S., and Wilhelmson R. B. , 1998: Evolution and structure of tropical squall line elements within a moderate CAPE and strong low-level jet environment. J. Atmos. Sci., 55 , 30893113.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cunning, J. B., Holle R. L. , Gannon P. T. , and Watson A. I. , 1982: Convective evolution and merger in the FACE experimental area: Mesoscale convection and boundary layer interactions. J. Appl. Meteor., 21 , 953978.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Davies, J. M., Doswell C. A. III, Burgess D. W. , and Weaver J. F. , 1994: Some noteworthy aspects of the Hesston, Kansas, tornado family of 13 March 1990. Bull. Amer. Meteor. Soc., 75 , 10071017.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dowell, D. C., and Bluestein H. B. , 2002a: The 8 June 1995 McLean, Texas, storm. Part I: Observation of cyclic tornadogenesis. Mon. Wea. Rev., 130 , 26262648.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dowell, D. C., and Bluestein H. B. , 2002b: The 8 June 1995 McLean, Texas, storm. Part II: Cyclic tornado formation, maintenance, and dissipation. Mon. Wea. Rev., 130 , 26492670.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fierro, A. O., Gilmore M. S. , Wicker L. J. , Mansell E. R. , Straka J. M. , and Rasmussen E. N. , 2004: Electrification and lightning in an idealized boundary-crossing supercell simulation of 2 June 1995. Preprints, 22d Conf. on Severe Local Storms, Hyannis, MA, Amer. Meteor. Soc., CD-ROM, 16B.7.

  • Finley, C. A., Cotton W. R. , and Pielke R. A. Sr., 2001: Numerical simulation of tornadogenesis in a high precipitation supercell. Part I: Storm evolution and transition into a bow echo. J. Atmos. Sci., 58 , 15971629.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Finley, C. A., Cotton W. R. , and Pielke R. A. Sr., 2002: Tornadogenesis in a simulated HP supercell. Preprints, 21st Conf. on Severe Local Storms, San Antonio, TX, Amer. Meteor. Soc., 531–534.

  • Fujita, T. T., 1960: A detailed analysis of the Fargo tornadoes of June 20, 1957. U.S. Weather Bureau Research Paper 42, Chicago, IL, 67 pp.

  • Fujita, T. T., Bradbury D. L. , and Van Thullenar C. F. , 1970: Palm Sunday tornadoes of April 11, 1965. Mon. Wea. Rev., 98 , 2669.

  • Gilmore, M. S., and Wicker L. J. , 2002: Influences of the local environment on supercell cloud-to-ground lightning, radar characteristics, and severe weather on 2 June 1995. Mon. Wea. Rev., 130 , 23492372.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gilmore, M. S., Straka J. M. , and Rasmussen E. N. , 2004: Precipitation uncertainty due to variations in precipitation particle parameters within a simple microphysics scheme. Mon. Wea. Rev., 132 , 26102627.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Goodman, S. J., and Knupp K. R. , 1993: Tornadogenesis via squall line and supercell interactions: The November 15, 1989, Huntsville, Alabama, tornado. The Tornado: Its Structure, Dynamics, Prediction, and Hazards, Geophys. Monogr., Vol. 79, Amer. Geophys. Union, 183–199.

    • Crossref
    • Export Citation
  • Jewett, B. F., Wilhelmson R. B. , and Lee B. D. , 2002: Numerical simulation of cell interaction. Preprints, 21st Conf. on Severe Local Storms, San Antonio, TX, Amer. Meteor. Soc., 353–356.

  • Kogan, Y. L., and Shapiro A. , 1996: The simulation of a convective cloud in a 3D model with explicit microphysics. Part II: Dynamical and microphysical aspects of cloud merger. J. Atmos. Sci., 53 , 25252545.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lee, B. D., and Wilhelmson R. B. , 1997: The numerical simulation of non-supercell tornadogenesis. Part II: Evolution of a family of tornadoes along a weak outflow boundary. J. Atmos. Sci., 54 , 23872415.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lee, B. D., Jewett B. F. , and Wilhelmson R. B. , 2006: The 19 April 1996 Illinois tornado outbreak. Part I: Cell evolution and supercell isolation. Wea. Forecasting, 21 , 433448.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lemon, L. R., 1976: The flanking line, a severe thunderstorm intensification source. J. Atmos. Sci., 33 , 686694.

  • Malkus, J. S., 1954: Some results of a trade-cumulus cloud investigation. J. Meteor., 11 , 220237.

  • Markowski, P. M., Rasmussen E. N. , and Straka J. M. , 1998: The occurrence of tornadoes in supercells interacting with boundaries during VORTEX-95. Wea. Forecasting, 13 , 851859.

    • Search Google Scholar
    • Export Citation
  • Markowski, P. M., Rasmussen E. N. , and Straka J. M. , 2002: Direct surface thermodynamic observations within rear-flank downdrafts of nontornadic and tornadic supercells. Mon. Wea. Rev., 130 , 16921721.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • NCDC, 1996: Storm Data. Vol. 38, No. 4, 410 pp. [Available from National Climatic Data Center, 151 Patton Ave., Asheville, NC 28801.].

  • NSSL, 2000: WATADS (WSR-88D Algorithm Testing and Display System) 2000: Reference guide for version 10.2. NSSL, 200 pp. [Available from Storm Scale Applications Division, National Severe Storms Laboratory, 1313 Halley Circle, Norman, OK 73069.].

  • Purdom, J. F., 1976: Some uses of high-resolution GOES imagery in the mesoscale forecasting of convection and its behavior. Mon. Wea. Rev., 104 , 14741483.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rasmussen, E. N., Davies-Jones R. P. , Doswell C. A. III, Straka J. M. , Carr F. H. , Eilts M. D. , and MacGorman D. R. , 1994: The Verification of the Origins of Rotation in Tornadoes Experiment: VORTEX. Bull. Amer. Meteor. Soc., 75 , 9951005.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rasmussen, E. N., Richardson S. , Straka J. , Markowsiki P. , and Blanchard D. , 2000: The association of significant tornadoes with a baroclinic boundary on 2 June 1995. Mon. Wea. Rev., 128 , 174191.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Sabones, M. E., Agee E. M. , and Akridge M. , 1996: The Palaski County and West Lafayette, Indiana, tornadoes, 26–27 April, 1994: A case study of supercell (mesocyclone) and squall-line bow echo interaction. Preprints, 18th Conf. on Severe Local Storms, San Francisco, CA, Amer. Meteor. Soc., 746–750.

  • Simpson, J., 1980: Downdrafts as linkages in dynamic cumulus seeding effects. J. Appl. Meteor., 19 , 477487.

  • Simpson, J., and Woodley W. L. , 1971: Seeding cumulus in Florida: New 1970 results. Science, 172 , 117126.

  • Simpson, J., Westcott N. E. , Clerman R. J. , and Peilke R. A. , 1980: On cumulus mergers. Arch. Meteor. Geophys. Bioklimatol. Ser. A, 29 , 140.

  • Stumpf, G. J., Witt A. , Mitchell E. D. , Spencer P. L. , Johnson J. T. , Eilts M. D. , Thomas K. W. , and Burgess D. W. , 1998: The National Severe Storms Laboratory Mesocyclone Detection Algorithm for the WSR-88D. Wea. Forecasting, 13 , 304320.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tao, W-K., and Simpson J. , 1984: Cloud interactions and merging: Numerical simulations. J. Atmos. Sci., 41 , 29012917.

  • Tao, W-K., and Simpson J. , 1989: A further study of cumulus interactions and mergers: Three-dimensional simulations with trajectory analyses. J. Atmos. Sci., 46 , 29743004.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wakimoto, R. M., and Wilson J. W. , 1989: Non-supercell tornadoes. Mon. Wea. Rev., 117 , 11131140.

  • Weaver, J. F., and Nelson S. P. , 1982: Multiscale aspects of thunderstorm gust fronts and their effects on subsequent storm development. Mon. Wea. Rev., 110 , 707718.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Weaver, J. F., and Purdom J. F. W. , 1995: An interesting mesoscale storm-environment interaction observed just prior to changes in severe storm behavior. Wea. Forecasting, 10 , 449453.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Weaver, J. F., Purdom J. F. W. , and Szoke E. J. , 1994: Some mesoscale aspects of the 6 June 1990 Limon, Colorado, tornado case. Wea. Forecasting, 9 , 4561.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Westcott, N., 1984: A historical perspective on cloud mergers. Bull. Amer. Meteor. Soc., 65 , 219226.

  • Westcott, N., 1994: Merging of convective clouds: Cloud initiation, bridging, and subsequent growth. Mon. Wea. Rev., 122 , 780790.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Westcott, N., and Kennedy P. C. , 1989: Cell development and merger in an Illinois thunderstorm observed by Doppler radar. J. Atmos. Sci., 46 , 117131.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wicker, L. J., and Wilhelmson R. B. , 1995: Simulation and analysis of tornado development and decay within a three-dimensional supercell thunderstorm. J. Atmos. Sci., 52 , 26752703.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wiggert, V. G., Lockett J. , and Ostlund S. S. , 1981: Rainshower growth histories and variations with wind speed, echo motion, location and merger status. Mon. Wea. Rev., 109 , 14671494.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wilkins, E. M., Sasaki Y. K. , Gerber G. E. , and Chaplin W. H. Jr., 1976: Numerical simulation of the lateral interactions between buoyant clouds. J. Atmos. Sci., 33 , 13211329.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Witt, A., Eilts M. D. , Stumpf G. J. , Mitchell E. D. , Johnson J. T. , and Thomas K. W. , 1998: Evaluating the performance of WSR-88D severe storm detection algorithms. Wea. Forecasting, 13 , 513518.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wolf, R., Pryzbylinski R. , and Berg P. , 1996: Observations of a merging bow segment and supercell. Preprints, 18th Conf. on Severe Local Storms, San Francisco, CA, Amer. Meteor. Soc., 740–745.

  • View in gallery

    Cell maximum radar reflectivity centroid tracks at the 0.5° elevation angle for 1940–0200 UTC 19–20 Apr 1996. See legend for track-type designations. The letters D, W, and OC specify storms initiating near the dryline, warm front, or dryline–warm front occlusion boundaries, respectively. The M and I represent storm merger and nonmerger interactions. Cell designations with an S appended designate storms resulting from storm splits. Tracks with arrows specify storms that were still under way at 0200 UTC. See text for details on the convective initiation cell tracking window and on cell numbering.

  • View in gallery

    As in Fig. 1 except with only the cyclonic supercell tracks and those contributory cells that were involved directly or indirectly in mergers with supercells. The bold blue lines designate supercell tracks of the radar reflectivity centroid. The superposed green and red line segments indicate when the storm had a demonstrable mesocyclone or TVS, respectively, based largely on WATADS algorithms. Tornado tracks are indicated in pink and labeled in the order they occurred. The light gray cell tracks are for two large discrete cells that developed along the flanking lines of two preexisting supercells. The gray designations for D16 indicate either a cell merger or nonmerger interaction from unplotted cells.

  • View in gallery

    KILX radar reflectivity imagery (0.5°) at 2159, 2216, 2234, 2252, and 2309 UTC. The cyan dashed lines represent cell tracks. Specific cell mergers referenced in the text are labeled. Cells experiencing multiple mergers have the sequential merger number appended (e.g., D16M3 represents the third merger event for D16). Reflectivities less than 15 dBZ are not plotted.

  • View in gallery

    KILX radar reflectivity imagery (3.3°) featuring the type C merger D12FM (merger of D12 and D12F). Panels are at times 2353, 2359, and 0004 UTC. The projected maximum reflectivity centroids for D12 and D12F are indicated in the 0004 UTC panel for geographical reference.

  • View in gallery

    KILX radar reflectivity imagery (0.5°) featuring the type D merger of D16 and D18. Panel times are 2309, 2321, and 2332 UTC.

  • View in gallery

    Histogram of tornado incidence times relative to the time of the cell merger partitioned in 5-min intervals for the 15-min period before and after cell merger.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 202 171 10
PDF Downloads 110 93 8

The 19 April 1996 Illinois Tornado Outbreak. Part II: Cell Mergers and Associated Tornado Incidence

View More View Less
  • 1 WindLogics, Inc., Grand Rapids, Minnesota
  • | 2 Department of Atmospheric Sciences, University of Illinois at Urbana–Champaign, Urbana, Illinois
© Get Permissions
Full access

Abstract

In the 19 April 1996 Illinois tornado outbreak, cell mergers played a very important role in the convective evolution. With a large number of cells forming within a short time period, the early stages of cell organization were marked by cell merger interactions and cell attrition that led to a pattern of isolated tornadic supercells as described in Part I of this study. Twenty-six mergers were documented and analyzed. Storm-rotation-induced differential cell propagation accounted for 58% of these 26 cell mergers while differing cell speeds prompted 27% of the mergers. Cell merger characterizations were utilized to describe the cell reflectivity coalescence morphology including aspects of new cell development, development along the periphery of an existing cell, or an upward pulse in the cell intensity of a dominant cell. In cases where the merging cells were of similar intensity, a rapidly developing cellular pulse “bridging” the two echoes was often observed. When the relationship between short-term cell intensity changes and cell mergers was examined, it was found that the maximum reflectivity tendency showed a bias toward higher reflectivity for the product storm. Depending upon the radar elevation angle utilized, 27%–44% of mergers were associated with an increase in peak reflectivity while 40%–58% of the storms realized little or no increase. With respect to short-term cell rotation changes, the merger signal was marked. Depending upon the length of the evaluation window, in 44%–60% of the mergers, there was evidence of a merger-associated increase in cell rotation. When considering the association between tornado occurrence and cell mergers, a striking 54% of the tornadoes occurred within 15 min before or after a cell merger. This high percentage is strongly suggestive of a physical relationship between storm mergers and tornadogenesis. A discussion is presented of potential merger scenarios and favorable ambient environmental conditions that may have been conducive to tornadogenesis in this event. Suggestions are presented to raise the awareness level of forecasters to key aspects of cell evolution and interaction in nowcasting severe convection.

Corresponding author address: Dr. Bruce D. Lee, WindLogics, Inc., 201 NW 4th St., Grand Rapids, MN 55744. Email: blee@windlogics.com

Abstract

In the 19 April 1996 Illinois tornado outbreak, cell mergers played a very important role in the convective evolution. With a large number of cells forming within a short time period, the early stages of cell organization were marked by cell merger interactions and cell attrition that led to a pattern of isolated tornadic supercells as described in Part I of this study. Twenty-six mergers were documented and analyzed. Storm-rotation-induced differential cell propagation accounted for 58% of these 26 cell mergers while differing cell speeds prompted 27% of the mergers. Cell merger characterizations were utilized to describe the cell reflectivity coalescence morphology including aspects of new cell development, development along the periphery of an existing cell, or an upward pulse in the cell intensity of a dominant cell. In cases where the merging cells were of similar intensity, a rapidly developing cellular pulse “bridging” the two echoes was often observed. When the relationship between short-term cell intensity changes and cell mergers was examined, it was found that the maximum reflectivity tendency showed a bias toward higher reflectivity for the product storm. Depending upon the radar elevation angle utilized, 27%–44% of mergers were associated with an increase in peak reflectivity while 40%–58% of the storms realized little or no increase. With respect to short-term cell rotation changes, the merger signal was marked. Depending upon the length of the evaluation window, in 44%–60% of the mergers, there was evidence of a merger-associated increase in cell rotation. When considering the association between tornado occurrence and cell mergers, a striking 54% of the tornadoes occurred within 15 min before or after a cell merger. This high percentage is strongly suggestive of a physical relationship between storm mergers and tornadogenesis. A discussion is presented of potential merger scenarios and favorable ambient environmental conditions that may have been conducive to tornadogenesis in this event. Suggestions are presented to raise the awareness level of forecasters to key aspects of cell evolution and interaction in nowcasting severe convection.

Corresponding author address: Dr. Bruce D. Lee, WindLogics, Inc., 201 NW 4th St., Grand Rapids, MN 55744. Email: blee@windlogics.com

1. Introduction

On 19 April 1996, an outbreak of tornadic supercell thunderstorms struck portions of Illinois and adjacent states (NCDC 1996). In Lee et al. (2006, hereafter Part I of this study), we described the magnitude of this event, which included 39 tornadoes in Illinois and an additional 20 tornadoes in the surrounding states of Iowa, Indiana, and Missouri. A fascinating aspect of this case involves the early evolution and interaction of cells that formed along a dryline just west of the Mississippi River in Missouri, a warm front oriented northwest–southeast through central Illinois, and a dryline–warm front occlusion boundary oriented northwest–southeast through southeastern Iowa. Radar analysis of this evolving convection presented in Part I revealed a complex pattern of storm splits and mergers as seen in Fig. 1 (reintroduced from Part I). Of particular interest, from more than a dozen cells forming along the dryline in northeastern and eastern Missouri at approximately 2100 UTC, only two large long-lived supercells remained by 2330 UTC that tracked across central Illinois, producing 13 tornadoes. In a broader context, supercells simultaneously formed along all three boundaries on this day. Building on Part I analyses that included convective initiation, cell evolution, and processes resulting in supercell isolation, a major objective of this paper is to examine storm interactions in the context of merger morphology and merged cell intensity changes and to qualify the types of cell mergers. Additionally, given the large number of cell mergers and tornadoes, this case provides an opportunity to investigate the statistical relationship between cell interactions and tornado occurrence.

Historically, cell mergers have been recognized as having a profound influence on quantities such as cell size, persistence, or precipitation production (Byers and Braham 1949; Malkus 1954; Simpson and Woodley 1971; Changnon 1976; Simpson et al. 1980; Wiggert et al. 1981). The physical processes ongoing in storm mergers are not well understood. Westcott’s (1984) review article on cloud mergers provides a comprehensive overview on the varied hypotheses of cloud merger processes and addresses the rather contentious topic of what constitutes a “cloud merger.” In this paper we use the term “cell merger” to mean the merger of reflectivity regions of two previously distinct cells resulting in only one remaining cell assessed at the lowest radar elevation angle (0.5°). More specifically, two cells are considered “merged” when two distinct reflectivity maxima, identified with the particular cells, are no longer present. Note that for the analysis of 19 April 1996 herein, “interaction” and “merger” are used synonymously unless specific reference is made to a nonmerger interaction. In most cases, due to radar distance limitations, we can only infer storm interaction from reflectivity evolution.

In two-dimensional numerical simulations Wilkins et al. (1976) found that for merging pairs of buoyant thermals, the merged pair behaves as a single thermal with twice the buoyancy of a solitary thermal. Kogan and Shapiro (1996) used idealized three-dimensional simulations of pairs of convective clouds in an environment with no vertical shear to investigate the physical mechanisms of cloud mergers. Whether or not convective clouds merged was dependent upon the distance between the cell-initiating thermal perturbations. As in the Wilkins et al. study, merged updrafts were stronger than the updraft of the single-cell control simulation. The modeling work of Tao and Simpson (1984, 1989) showed that merging cells are linked by new updraft growth in a convergent “bridge” region between interacting cells. This low-level convergence region was found to be produced near the leading edge of cell outflows and in outflow collision regions. These results were consistent with the cloud bridge observations for merging cells and the proposed mechanism for cell merger provided by Simpson (1980). In a numerical study of the evolving convective elements within a tropical squall line, Chin and Wilhelmson (1998) noted the merger of two storms through a new bridge cell. The concept of new cell growth bridging the reflectivity regions between two previously individual cells will be shown in section 2 to be the most frequent mode for storm mergers within the study area on 19 April 1996. Westcott and Kennedy (1989) and Westcott (1994), using case study data in specific convective environments, documented the reflectivity coalescence morphology in radar-observed cell mergers. In the former study, cell mergers resulted from differential storm motion and new cell growth between storms. In both of these studies, transport of hydrometeors into the region between cells was proposed as a means of developing a reflectivity bridge that connected the originally separate cells.

Of particular applicability to this study, the idea that mergers may influence supercell thunderstorm evolution has been a subject of interest for several decades. Lemon (1976) reported on a case where a prominent flanking line cell merged with the weak echo region of a supercell resulting in an apparent increase in cell rotation and updraft intensity. In a numerical study, Bluestein and Weisman (2000, hereafter BW2000) showed that the evolution of storms that developed along a finite section of a boundary in an environment generally supportive of supercells was highly dependent upon storm interactions in the internal part of the line. Cell interactions were principally dependent upon differential cell motion that was largely controlled by the characteristics of the vertical profile of the horizontal wind. Supercell evolution in BW2000 ultimately depended upon the direction of the vertical shear vector with respect to the orientation of the initiating boundary. Some facets of the convective evolution in this case study will be compared with the results of BW2000.

The influence of mergers on tornadogenesis is also poorly understood. The observational study by Sabones et al. (1996) documented tornado occurrence coincident with the interaction between a supercell and squall line. In a similar case of squall line–supercell interaction, Goodman and Knupp (1993) suggest a potential linkage between an increase in tornado intensity and storm interaction. Wolf et al. (1996) described a case where the merging of supercells appeared to be related to the formation of a strong tornado. In their case, outflow interactions occurring with the merger were cited as a potential key mechanism for the associated tornadogenesis. The modeling investigations of Finley et al. (2001, 2002) document a case of tornadogenesis coinciding with the merger of two supercells. The merging supercells produced an updraft significantly stronger than either of the individual cells. This result was consistent with the aforementioned idealized modeling results of Wilkins et al. (1976) and Kogan and Shapiro (1996). Finley et al. (2002) also documented a marked increase in low-level vertical vorticity stretching tendency during and just after the updraft merger, which resulted in the development of a strong low-level vortex.

The authors have noted on frequent occasions that storm merger events seem to be qualitatively related to tornado occurrence within a short time of the merger. Conversations with National Weather Service forecasters and other severe storms researchers have also revealed this same perceived association. Since the 19 April outbreak featured numerous cell mergers and other interactions, and also featured many tornadoes, it appeared to be an ideal case for attempting to quantify this link between cell merger and tornado occurrence. Families of tornadoes from isolated supercells were characteristic of this event (Fujita 1960; Fujita et al. 1970; Agee et al. 1976). Some of these families can be seen in the west-southwest to east-northeast pattern of tornado tracks in Fig. 1 of Part I. One prolific tornadic supercell in central Illinois produced 10 tornadoes between 2244 and 0208 UTC. Were cyclic tornadogenesis and changes in storm intensity in these supercells prompted by the merger events? While the observational dataset does not allow much latitude in directly identifying the storm-scale processes ongoing in a merger interaction that might lead to tornadogenesis, quantitative associations can be made.

The reader should reference Part I for an overview of the synoptic conditions associated with this outbreak as well as for details on the data sources and techniques utilized in the cell tracking analysis. Part II of this study is organized in the follow way. Section 2 focuses on cell merger characterizations and statistics along with an analysis of the association of storm mergers with storm intensity changes. Section 3 discusses the association between tornado occurrence and storm mergers. Finally, in section 4 we offer a summary of the results and discuss why cell mergers may be related to tornadogenesis. We finish this section with a discussion of the potential applicability of these findings to improved nowcasting

2. Cell mergers

Cell interactions in the form of mergers were an important factor influencing supercell development and evolution in this severe convective outbreak. In this study, 24 mergers documented in Part I and indicated in Fig. 1 are included in the analysis. Although we did not, in general, track flanking line cells or consider flanking line cells as separate entities from the host cell, we did add two additional cases to this dataset where strong and discrete cells initiating along a supercell’s flanking line merged with the parent supercell. In this section, we identify the scenarios leading to cell mergers and characterize the types of mergers observed. Additionally, the influences of mergers on cell rotation and intensity are documented. To facilitate the discussion of this section’s topics, Fig. 2 was created and includes cell tracks of the cyclonic supercells and those contributory cells that were involved directly or indirectly in mergers with supercells. Indirect involvement, in this sense, means that a cell merged with a cell that eventually merged with a supercell (e.g., W25 merged with W23, which later merged with W20). In this case study, the vast majority of all documented mergers involved cells in these supercell and contributory cell groups (e.g., the central Illinois supercells D12 and D16). The merger locations are indicated in Fig. 2 as well as the radar-derived rotational character of the storms. The cell rotation character was largely based on WSR-88D algorithm output from the Weather Surveillance Radar-1988 Doppler (WSR-88D) Algorithm Testing and Display System (WATADS 10.2; NSSL 2000) and is indicated on the tracks. Tornado tracks are also indicated in Fig. 2 based on the Storm Prediction Center historical tornado archive.

a. Scenarios leading to cell merger

An important requirement for the understanding of cell interactions on this day involves the identification of the scenarios that brought the convective elements together. Actual characterizations of how the radar echoes appeared to merge are documented in section 2b. Analysis of radar imagery from Lincoln, Illinois (KILX); St. Louis, Missouri (KLSX); and Davenport, Iowa (KDVN); revealed that 58% of the 26 mergers involved differential cell propagation apparently owing largely to the observed rotation. The various interaction scenarios that entailed one or more rotating cells included the following:

  1. a cyclonically rotating supercell moving to the right of the mean wind intersecting the track of a nonrotating cell moving in the direction of the mean wind (e.g., D16 and D17 in Fig. 2),

  2. an anticyclonically rotating cell moving to the left of the mean wind intersecting the track of a nonrotating or cyclonically rotating cell (e.g., D16S and D10 in Fig. 2), and

  3. a cyclonically rotating supercell moving to the right of the mean wind intersecting the track of another cyclonic supercell with less deviant motion (e.g., W8 and W12 in Fig. 2).

Accounting for 27% of the mergers, the next most common scenario involved differential speeds for cells generally moving in the same direction. Smaller yet, differential directional propagation for reasons that could not be determined (e.g., when neither storm appeared to have rotation) composed 8% of the mergers. The final 8% of interactions included the two flanking line cell–supercell mergers. In this case, the storm tracks did not cross as in previous scenarios. Instead, the cells interacted by way of new cell growth between the original entities, which, as is documented next, is a common morphological mode of cell merger.

b. Merger characterization

Before proceeding further, it is useful to revisit the question: What constitutes a cell merger? In the broadest sense, since the weather radar is our primary observational tool in this study, we consider a cell merger to entail the joining of two radar echoes. More specifically, two cells are considered “merged” when two distinct reflectivity maxima, identified with the particular cells, are no longer present at the lowest scanning angle (0.5°). This criterion established the formal “time of merger” for this study. Although this definition may seem straightforward in a most simplistic sense, the morphological changes when the two interacting cells come into close proximity can be quite complex and are not thoroughly understood. Assuming a storm interaction is close enough to a fixed-location WSR-88D to resolve distinct cellular changes, the main difficulty in observationally studying storm mergers lies in the temporal resolution limitations of the 5–6-min volume scans. The rapidity of storm structural changes in the time period of a single volume scan can be very dramatic as will be shown in the following characterizations. The deployment of mobile Doppler radars [e.g., the C-band Shared Mobile Atmospheric Research and Teaching Radars (SMART-Rs); Bluestein and Wakimoto (2003)] could play a role in gathering essential high temporal and spatial resolution data on storm mergers in future field projects. Where possible, multiple elevation scans from KILX, KLSX, and KDVN were utilized to examine the merging process because the time offset between radars gave an effective higher temporal resolution. In contrast to most observational radar studies, numerical modeling investigations of storm interactions have the advantage of high temporal resolution and often spatial resolution relative to the phenomena being modeled (e.g., Kogan and Shapiro 1996; BW2000).

To characterize the cell mergers of 19 April 1996, the interactions were subdivided into four descriptive categories and one category for undetermined cell unions. Given the numerous cell interactions (Fig. 2) and range of merger types occurring with the array of west-central Illinois cells from which the isolated supercells D12 and D16 emerged, Fig. 3 was designed to provide a KILX cell configuration reference to accompany the merger descriptions below. Where appropriate, figures utilizing a higher time sampling are employed to demonstrate specific type C and D mergers.

Type A mergers are characterized as those where a larger cell incorporates a historically smaller cell with no recognizable new cell development, reflectivity area expansion, or intensification of the primary cell. As innocuous as these mergers appear, structural changes related to rotational variations may occur just after the cell union. Type A mergers accounted for 23% of all mergers. The D12M2 merger (i.e., the second merger experienced by the dominant cell D12) indicated in Fig. 3 was a type A merger.

Type B mergers involve the interaction between a dominant and subordinate cell whereby the reflectivity of the dominant cell pulses upward and/or reflectivity development occurs on the periphery of the dominant cell at or near the point of merger. In the latter case, the peripheral cell reflectivity expansion does not appear as a discrete new cell between the colliding cells, in contrast to the type C mergers described below. Perhaps with better time or space resolution of the radar imagery, discrete cellular growth may have been seen occurring between the colliding cells in some of these cases. Type B mergers accounted for 23% of all mergers. Type B mergers included D12M1, D16M2, and D16M3 indicated in Fig. 3.

Type C mergers involve the development of a distinct reflectivity maximum between the approaching cells when they come in close proximity. This new cellular growth “bridged” the two original reflectivity maxima. This mode of merger through new bridge development was favored for interactions where the difference in cell intensity was small. Of particular applicability to this merger type, the numerical modeling research of BW2000 has shown a new updraft pulse between colliding cell updrafts and above the colliding outflow boundaries of the merging cells. The new updraft produced a precipitation area that created a short-lived proxy-reflectivity bridge. The new cell quickly replaced the merging cells. Past two-dimensional (Tao and Simpson 1984) and three-dimension (Tao and Simpson 1989; Chin and Wilhelmson 1998) simulations of tropical convection have shown bridge cell growth between merging cells. Observational studies have also supported the idea of cell mergers through the development of bridge cell growth between the original cells (Simpson 1980; Cunning et al. 1982; Westcott and Kennedy 1989). Consistent with BW2000, in cases where cells were on a collision course, we found that when the cell intensities were not greatly different, the precipitation (reflectivity) regions merged aloft first, followed by a coalesced reflectivity structure through the lowest scanning levels. The type C merger was the most frequent type with 31% of all mergers. The first merger D16 undergoes (i.e., D16M1 referenced in Fig. 3), was a type C merger as was the first merger experienced by D10 (i.e., D10M). Another marked example of a type C merger may be seen in the large flanking line–supercell merger (D12FM) featured in the KILX reflectivity imagery in Fig. 4. As the cell spacing gradually narrows, a new coalesced reflectivity maximum develops very rapidly between the cells. The unified reflectivity maximum in Fig. 4 at 0004 UTC is nearly exactly at the midpoint location between the projected centroid positions of the former separate cells. Multiple elevation scans support the idea of a new cellular pulse between the two storms. This new cellular development links the former reflectivity maxima in the remarkably short period of one volume scan.

The morphological changes related to cell rotation evolution are also striking. Already at 2353 UTC there exists some indication of a cyclonic and anticyclonic reflectivity hook couplet. In just 6 min (2359 UTC), this rotational couplet inferred in the reflectivity hooks becomes very well defined with corresponding cyclonic and anticyclonic circulation centers evident in the KILX radial velocity imagery (not shown). As is evident at 0004 UTC, the cyclonic circulation became dominant with a reflectivity structure indicative of a thin precipitation streamer making up the neck of the hook echo. This very narrow reflectivity hook neck is reminiscent (although significantly wider due to the resolution of the WSR-88D) of the finescale hook structures observed from portable Doppler radar by Bluestein and Pazmany (2000). Tornado 16 shown in Fig. 2 was first reported 6 min prior to the appearance of a unified reflectivity pattern (i.e., at a time very close to the middle panel in Fig. 4). This tornado was one of the stronger (F3 intensity) and longer-lived (21 min) tornadoes on this day. In section 3, evidence for an association between tornado incidence and cell merger is presented.

Type D merger prompts distinct cellular growth at or very near the point of coalescence. The new cellular growth does not occur between the approaching cells as in type C, but occurs as a distinct new cell just as a subordinate cell loses identity upon merging with a dominant cell. While this could be considered a merger by-product, we have defined this as a distinct merger type since we are interested in the very short-term cell morphology inclusive of new cellular growth that appears to be prompted by the colliding cells. This merger characterization is the least frequent of the merger types documented on this day (4%). However, the single incidence of the type D merger involved a most interesting cell evolution as documented in Fig. 5. In this figure, a very small but intense cell resembling a “minisupercell” develops by 2332 UTC coincident with the merger of D18 and the supercell D16. The mesocyclone associated with this new cell developed very quickly after initiation. Tornadogenesis also rapidly commenced in this cell, with the first two authors observing a tornado by 2330 UTC (tornado 15 in Fig. 2). As an additional refinement to the convective initiation analysis, it appears that as D18 interacted with the flanking line of supercell D16, convective initiation ensued. The 2321 UTC panel in Fig. 5 shows a reflectivity arc that is likely the flanking line of cell D16. The incipient cell may have been triggered by the collision between outflow from D18 and the rear-flank downdraft boundary from supercell D16 (that was locationally coincident with the flanking line).

Type O mergers (“other”) consisted of cases very difficult to typecast due to radar spatial and temporal gaps and due to complex multiple cell interactions. This category comprised 19% of all mergers.

c. Relationship between cell merger and changes in storm intensity

How mergers influence the short-term intensity of storms is an important question for the storm research, forecasting, and warning communities. We confine our focus to the short term (≤18 min) given the difficulty involved in linking a specific merger event to a change in cell intensity farther downstream, especially when a cell may experience multiple storm interactions (e.g., D12 and D16 in Fig. 2). This analysis is, in large part, an assessment of the influence of mergers on supercells since cells designated as supercells are involved in 81% of all the mergers documented here; however, since we are also interested in cell intensity changes in nonsupercell storms after a merger, we include all mergers where possible. Two gauges are used to address changes in cell intensity: reflectivity and rotation.

The first measure includes an analysis of cell peak reflectivity tendency tracked from one volume scan before the time of merger through three volume scans after the merger. In this first measure, we make the assumption that a storm with higher peak reflectivity represents a “more intense” storm, presumably with a stronger updraft. We note the possibility that in some cases of storm merger, reflectivity could also temporarily increase simply due to the number of particles increasing per unit volume as the precipitation areas merge (Fierro et al. 2004). However, there is ample evidence in the literature supporting the relationship between increasing reflectivity (or analogous precipitation field) and updraft strength in cases of cell merger where the product storm showed intensification (e.g., Kogan and Shapiro 1996; Finley et al. 2001, 2002). Thus, the idea that observed reflectivity increases are likely related to updraft intensification for a product storm from a cell merger has relevance as an indicator of storm intensity, despite its deficiencies.

The analysis uses WSR-88D 0.5° elevation reflectivity data from the closest available radar site. A cell is assumed to have a positive reflectivity trend if its maximum reflectivity increases for at least two consecutive volume scans and the maximum reflectivity is larger in the final scan than in the first. This reflectivity trend analysis is further refined by considering those cases having changes of less than 5 dBZ to be representative of cells exhibiting little or no trend. Since in most mergers there is a stronger cell involved, this reflectivity tendency analysis is largely an examination of the reflectivity tendency of the stronger cell as influenced by the cell interaction. As shown in Table 1, in a majority of the merger cases, there was little or no change in the storm’s reflectivity. There was a modest bias toward constructive merger effects as shown by the larger number of cells with positive rather than negative reflectivity tendencies. Since the storm’s reflectivity regions were found to merge aloft first and a higher scan elevation might provide an earlier indication of hydrometeor growth resulting from an updraft increase, cell maximum reflectivity statistics were also compiled for the 2.4° scanning elevation. As can be seen in Table 1, the reflectivity tendency exhibited a similar character to that found for the lowest scanning angle; however, there was a stronger propensity for storms to realize an increase in reflectivity with 44% of the cases showing cell strengthening. A similar reflectivity tendency analysis using a narrowed evaluation window that included just the three volume scan periods after the merger time revealed very similar results.

To ensure that these results were not just sampling a longer-term trend of increasing reflectivity, and since most of the mergers involved a supercell, we looked at the peak reflectivity trend over a 1.5-h period (starting at about the 1-h point after the cell first met the tracking criteria) for 10 long-lived supercells. Mean peak reflectivity changes between the beginning and midpoint, and between the beginning and endpoint of this 1.5-h period were less than 5 dBZ. Thus, a substantial mean increase in peak reflectivity for the group of cells was not present. In another comparative test using five longer-lived cells not involved in a merger, a peak reflectivity variability analysis for the same 1.5-h period noted previously indicated no substantial peak reflectivity trend bias. In summary, there were a significant percentage of mergers (∼27%–44%) that were coincident with a substantial increase in short-term storm reflectivity and a much smaller percentage that weakened.

A large number of merger cases exhibited no substantive trend in reflectivity tendency (40%–58%). In comparison, the modeling studies of BW2000 and Jewett et al. (2002) found that most of the cell interactions resulted in generally weaker storms; however, a few cell interactions resulted in at least brief increases in rotation. The more narrow application of what constitutes an “interaction” in this study may account for some of the differences with these modeling papers. It is noteworthy in the context of severe storm environments that model simulations and observations support scenarios where cell interactions resulted in an increase in cell intensity. A robust signal for constructive mergers was found by Kogan and Shapiro (1996) in idealized cloud merger simulations that showed when convective clouds merged, the resultant updraft was stronger than a comparative single-cell control simulation. A number of observational studies have documented cell intensity increases with storm mergers including Simpson and Woodley (1971), Changnon (1976), and Lemon (1976). In an eastern Illinois radar study of convective evolutions for two summer season case days featuring many cell mergers, Westcott (1994) found postmerger cell growth (in terms of maximum reflectivity, cell-top height, or echo area) was associated with the premerger growth tendency and age of the cores involved. Westcott also states that “a synergistic interaction between merging cores was not obvious.” Given the variability in the merger responses as documented in this research and the cited studies, there is much to be learned about the guiding environmental parameters that influence the storm intensity tendency for cell interactions.

For the second assessment of storm intensity changes associated with cell mergers, a similar analysis was conducted for cell rotation using the WSR-88D cell rotation algorithms. The assumption was made that the quality of cell rotation increases as the progressive algorithm thresholds for shear (S), mesocyclone (M), and tornado vortex signature (TVS) are met. There are a numbers of potential problems with this assumption, not the least of which is the influence distance from the radar has on the minimum detectable signature. This problem is mitigated to some degree by radar algorithms that compensate for the beam-spreading problem by range normalizing the shear signatures (Stumpf et al. 1998). Nonetheless, this system for assessing rotation quality trends could be systematically applied to the mergers. A change in rotational status was not limited to an assessment at a particular location (e.g., a cell’s persistent circulation center), but reflected the change in the peak rotation for colliding cells. For instance, a cell could exhibit mesocyclone-categorized rotation in two consecutive volume scans at a particular storm-relative position, but in these same two scans, at a location closer to the collision point, a younger mesocyclone-categorized circulation that strengthened to TVS status would constitute an overall increase in rotation quality for a storm. The baseline period used for this analysis included the three volume scans before and after the scan time centered on merger completion (test A). We extended the time of interest before the reflectivity merger since outflow from the merging storms could interact and influence the rotational character of the storm well before the reflectivity cores came together. Of the 26 mergers documented, the D10M merger was neglected since the merger product cell quickly merged with another storm, compromising the postmerger analysis window. The character of the cyclonic rotation trends for the mergers is documented in Table 2. Four categories have been employed to describe these rotational trends. A rotation increase (decrease) includes any jump (drop) in rotational state (e.g., M to TVS) from one volume scan to the next. As seen in Table 2, there appears to be a significant bias toward an increase in cell rotation for cell mergers, with 20% of mergers showing an explicit positive rotation trend (type I). Surprisingly, only 4% of the cell mergers indicated a clear reduced rotation trend. In another 52% of the mergers, there was an increase in cell rotation (type II) in the evaluation period that also featured a rotation decrease.

A plausible explanation for the rotation tendency exhibiting both increases and decreases in the evaluation period involves the pulselike character of the rotation intensification frequently observed with mergers. The rapid rotation increases may be prompted by cell interaction dynamics (Lemon 1976; Wilkins et al. 1976; Kogan and Shapiro 1996; Finley et al. 2001, 2002) followed by a systematic rotation decline when the interaction dynamics no longer exist. This may be similar to the pulselike vorticity increase noted by BW2000 for one of their colliding cell cases. The possibility also exists that, in some cases, a perceived rotation decrease following an increase might be attributed to a contraction of an intensifying vortex to an unresolvable scale. If the merger had a negative influence on storm rotation, one might expect either a decrease in cell rotation with no subsequent increase (as shown in Table 2) or a temporary decrease followed by an increase in rotation as the cell reestablished its premerger level of rotation. Eliminating the three instances of the latter scenario from the type II total results in 40% of the cases that were likely constructive type II merger events for cell rotation. Combining this figure with the 20% of cases displaying explicit increases yields a striking 60% of the mergers associated with cell rotation intensification.

A supplementary analysis, designated test B in Table 2, using a reduced time window for evaluation (eliminating the earliest and latest volume scans), produced results similar to test A, although with a slightly smaller percentage of explicit constructive mergers. Thus, even with a 10–12-min reduction in the evaluation period, 44% of the mergers featured in test B were type I or type II constructive mergers (applying the type II discriminating criteria used for test A). Perhaps the most important measure of a merger-related increase in cell rotational intensity lies in the association of tornado occurrence with cell mergers. This topic is addressed in the next section.

3. Association of tornado incidence with storm mergers

The motivation for analyzing the association between cell mergers and tornado occurrence resulted, in part, from a cursory examination of tornado path and radar data that appeared to indicate a qualitative connection. Given the large number of tornadoes and cell mergers for this event, an examination of this tornado–cell merger association seemed potentially fruitful. Figure 2 provides a reference for the tornado paths in the tracking domain. Note that since the tornadic region of most storms was typically southwest of the reflectivity centroid, given the cell motions, the tornado paths were generally south of the cell track.

Before the data on the association between tornado occurrence and cell merger are presented, it is useful to examine the storm types associated with the tornadoes. In a qualitative sense, a brief inspection of Fig. 2 would indicate that most of the tornadoes were associated with supercell thunderstorms. In fact, the analysis reveals that, for the plotted tornadoes, 81% were associated with plotted supercells. This percentage rises to 97% when unplotted supercells are included in the analysis. The cells that evolved into these unplotted supercells formed after the 2230 UTC endpoint used for the tracking analysis. The southern periphery tornadoes 33, 35, and 38 shown in Fig. 2 are examples of tornadoes associated with unplotted supercells. Although the vast majority of the tornadoes on this day were likely associated with the parent supercell mesocyclone, radar data limitations due to scanning distance prevent a more definitive statement. Some tornadoes could have developed in a manner more consistent with nonsupercell tornadoes (Brady and Szoke 1989; Wakimoto and Wilson 1989; Lee and Wilhelmson 1997) along the flanking line of the supercell.

To assess the association between tornado occurrence and cell mergers, Storm Data incidence data for the tornadoes plotted in Fig. 2 were compared to the estimated time of cell merger (based on the same criteria noted in section 2b). Since Storm Data reports beginning and ending times for tornadoes, or just a single time in the case of brief tornadoes, the tornado incidence times used here correspond to the earliest time reported. A 15-min period before and after the estimated merger time was used as an evaluation window. Given the complexity of storm processes and external influences on the parent supercell (e.g., other outflow interactions, subsequent mergers, etc.), this narrow ±15 min span was selected to improve the probability that tornadogenesis–merger interactions would be adequately sampled in our analysis. A primary reason for including the 15-min period prior to the time of reflectivity merger involves the hypothesis that, in some cases, outflow interactions between the merging cells could induce tornadogenesis. The outflow interaction accompanying cell mergers often occurs before low-level reflectivity merger (or the analogous merger of liquid water mixing ratio fields in modeling studies; e.g., see BW2000, their Fig. 12). New updraft or an updraft pulse from the cell merger, a potential factor in triggering tornadogenesis, may lead the reflectivity field merger at the lowest scanning angle. There are a number of caveats to the perceived association between the timing of tornadogenesis and cell merger. Storm Data tornado occurrence times in some instances may be imprecise (Witt et al. 1998) and the estimated merger times, while based on prescribed criteria stated in section 2b, are approximate. Tornadogenesis processes may be influenced by prior cell interactions occurring much earlier than the tornado event. In this regard, much needs to be learned about the tornadogenesis process and its variants, and the dynamics of storm interactions and their potential relationship with tornadogenesis.

Including the tornadoes shown in Fig. 2, with the exception of number 37 (due to inadequate radar coverage), 30% of the tornadoes were associated with a plotted cell merger. If we include the two large flanking line mergers alluded to previously, this figure rises to 35%. However, this value does not reveal the extent of the association given the additional mergers resulting from storms initiating after 2230 UTC that were not plotted in Figs. 1 or 2. The analysis was extended to examine tornado-merger incidence including these additional interactions. This expanded analysis revealed that a striking 54% of the tornadoes (20 of 37) occurred within 15 min of a cell merger. To understand the temporal distribution of tornado incidence with respect to the time of merger, the histogram in Fig. 6 has been created. The incidence data in Fig. 6 show an approximately symmetric distribution of the tornado–merger time differences. Although there is only a small numerical bias for tornado incidence following the merger time, a more noteworthy aspect of the distribution is that 55% of the tornado incidence times are within ±5 min of the time of merger (or about one radar volume scan on either side of the estimated merger time). This compares to only 15% of the tornadoes occurring in the evaluation period between ±10–15 min of the merger time. The results from the analysis of tornado incidence and cell merger times lend credence to the postulation that cell merger and tornado incidence were connected.

Another approach employed to gauge the association between cell mergers and tornadoes involved determining the percentage of mergers that were accompanied by tornadoes. As used in the prior analysis, a 15-min period before and after the estimated merger time was utilized as an evaluation window. Considering all the mergers shown in Fig. 1, 46% were associated with tornadoes. If this analysis is refined further by just considering the mergers involving supercells (see Fig. 2), a marked 57% of the mergers were associated with tornadoes. Included in this analysis are two tornadoes located just east of the plotting domain, that occurred in the evaluation window of the last merger involving supercell W20.

Due to the magnitude of the task and convolution of the track mapping, in general, we did not plot in Figs. 1 or 2 cell mergers or other interactions between cells initiating before 2230 UTC and those initiating after 2230 UTC unless the late initiating cells resulted from a storm split or were one of the two flanking line cells plotted (Fig. 2). As an example of the complexity of cell morphology in this event, additional merger and nonmerger interactions are indicated in Fig. 2 for supercell D16 from cells not plotted in Fig. 1. The nonmerger interactions noted here were cases where an approaching cell from the southwest appeared to interact with the flanking line of D16 just southwest of the main storm core but did not merge with the main body of the storm. For this prolific tornado-producing storm (10 tornadoes, 9 plotted), six merger and two nonmerger interactions were documented. The frequent cyclic tornadogenesis episodes (Burgess et al. 1982; Davies et al. 1994; Wicker and Wilhelmson 1995; Adlerman and Droegemeier 2000; Dowell and Bluestein 2002a) associated with D16 that occurred approximately coincident with merger and nonmerger interactions are, at least superficially, suggestive of a relationship. Note that of the nine plotted tornadoes for D16, six of these occurred within 15 min of a merger while a seventh tornado occurred with a marked cell interaction that did not qualify for merger designation. The recent papers by Dowell and Bluestein (2002a, b) noted a case of tornadogenesis (McLean storm tornado 1) coincident with the parent supercell’s merger with the precipitation region of an apparently dying cell. Assuming that an important element in cell mergers involves the relevant outflow boundary interactions, just how these interactions and related storm morphological changes compare to the broader scenario of storm-boundary interactions is unknown. Might some common mechanisms exist that are conducive to tornadogenesis? Observational papers by Purdom (1976), Weaver and Nelson (1982), Weaver et al. (1994), Weaver and Purdom (1995), and Wolf et al. (1996) have implied that storm-boundary interactions were associated with tornadic storms. In a study of supercells interacting with low-level boundaries during The Verification of the Origins of Rotation in Tornadoes Experiment (VORTEX; Rasmussen et al. 1994), Markowski et al. (1998) found that 70% of significant tornadoes occurred near a boundary. Similarly, both Rasmussen et al. (2000) and Gilmore and Wicker (2002) found increased tornado probability and stronger low-level mesocyclones for storms crossing an outflow boundary during the 2 June 1995 west Texas Panhandle VORTEX intercept that featured a localized outbreak of supercells and tornadoes. High-resolution numerical simulations and comprehensive observational datasets of storm merger interactions coincident with tornado occurrence are required to understand the relevant dynamics that may be conducive to tornadogenesis.

4. Summary and discussion

In the 19 April 1996 severe thunderstorm and tornado outbreak, cell mergers played a very important role in the evolution of convection. Twenty-six mergers were documented and analyzed. In addressing why cells merged, we found that 58% of the mergers involved rotationally induced differential cell propagation that led to storm interaction. Three typical scenarios for cell interactions involving one or more rotating cells included 1) a cyclonically rotating supercell moving to the right of the mean wind intersecting the track of a nonrotating cell moving in the direction of the mean wind, 2) an anticyclonically rotating cell moving to the left of the mean wind intersecting the track of a nonrotating or cyclonically rotating cell, and 3) a cyclonically rotating supercell moving to the right of the mean wind intersecting the track of another cyclonic supercell with less deviant motion. Clearly, the differing rotational properties of the various cells enhanced the probability for numerous mergers while fostering a scenario where, after a few hours, the supercells became increasingly isolated (Figs. 1 and 2). The other substantial mechanism for mergers involved cells moving in generally the same direction but at differing speeds. In this latter case, likely differences in cloud depth and implied cloud-bearing wind may be responsible for the differing cell speeds.

Four characterization types were created to describe the morphological changes, if any, from the mergers. With the exception of type A mergers, types B, C, and D all appeared to be associated with new reflectivity development or an increasing reflectivity pulse of the dominant cell in a mismatched intensity merger. If the cell reflectivity difference of the merging entities was not large, there was the tendency for type C mergers that involved new cellular growth between the merging cells analogous to that found in the modeling work of BW2000. In some of these type C merger instances, the analysis of the morphological evolution was challenging given the rapidity of the cell merger. With volume scans of 5–6 min, in initial examinations it appeared as though two reflectivity regions were merely merging, when in fact, there was a cellular pulse “bridging” the merging reflectivity areas as they moved in very close proximity (Westcott and Kennedy 1989). In some cases the speed of the morphological evolution was striking as in the type C case featured in Fig. 4 and the type D case featured in Fig. 5. In both of these instances, storm rotation increased markedly and was accompanied by tornadogenesis episodes.

The assessment of merger-associated short-term changes in storm intensity through an analysis of maximum reflectivity showed a bias toward higher reflectivity for the product storm. Depending upon the radar elevation angle used in the analysis (0.5° versus 2.4°), there was a significant percentage of mergers (∼27%–44%) that demonstrated an increase in short-term peak reflectivity intensity and a much smaller percentage that exhibited lower maximum reflectivity after merging. A large number of merger cases exhibited no substantive trend in reflectivity tendency (40%–58%). In a second assessment of merger-associated changes in cell intensity, short-term cell rotation tendency analysis revealed a marked interaction signal. Depending upon the length of the evaluation window (two versus three volume scans before and after the merger time), in 44%–60% of the mergers there was either an explicit type I or constructive type II increase in cell rotation. A much smaller percentage of cases exhibited reduced rotation with the storm merger. A direct approach to addressing the relationship between storm mergers and changes in low-level storm rotation lies in analyzing the association between tornado occurrence and merger time. A striking 54% of the tornadoes (20 of 37) occurred within ±15 min of a cell merger. Of this group, 55% of the tornadoes occurred within ±5 min of the estimated time of merger. In an alternate analysis, tornadoes were associated with 57% of the mergers that involved supercells shown in Fig. 2. These statistics are strongly suggestive of a physical relationship between storm mergers and tornadogenesis, at least in this outbreak. Supercell D16 [the Interstate Highway 72 (I-72) supercell] is a compelling example for this plausible relationship. Of the nine plotted tornadoes from this storm, six occurred within ±15 min of a merger and one other tornado occurred with a marked nonmerger interaction. In general, it appeared that multiple storm interactions incurred by mature supercells (such as the D16 storm) were related to cyclic tornadogenesis episodes.

The results of this study raise some questions about the merger interactions of this event. Why do the postmerger peak reflectivities generally show either unchanged or greater values, and why do the collisions appear to foster at least a short-term increase in cell rotation when simulations of cell interactions have demonstrated a generally negative influence on storm intensity (BW2000; Jewett et al. 2002)? It should be noted that in the two modeling references, there were a small number of cases where constructive effects were documented from select mergers. To the degree to which storm outflow collisions influence the product storm in the merger cases, we believe the actual storm outflow may have had only small temperature deficits due to the near-storm environment having a deep layer of high relative humidity above the surface (Fig. 5 in Part I) and a low lifting condensation level (LCL). A check on five central Illinois surface stations that fortuitously experienced a direct outflow passage from one of three different supercells (D12, D16, W11) confirms that the outflows only had small temperature deficits. In four of the cases, the supercell’s reflectivity region of ≥50 dBZ passed over the station. In the remaining case, the station was on the southern edge of a region with reflectivities ≥50 dBZ. The station observations reveal outflow temperature deficits <3°C at all sites with the average deficit of 1.9°C. In only one of these five cases did the dewpoint drop by more than 1°C. In none of these cases did the prestorm surface observations appear notably influenced by evaporatively cooled outflow from other cells. From the perspective of reduced buoyancy in the storm inflow, these mild outflows likely only had a modest influence on the merged storm and appear to have, in some cases, prompted a pulse in storm intensity. A constructive outflow interaction could result in increased low-level convergence, a strengthened low–midlevel updraft, and a resultant increase in vertical vorticity through stretching that may have been essential for enhancing storm rotation and promoting tornadogenesis on this day. The comparison and interpretation of these results with those of numerical simulations would be very sensitive to the outflow intensity of the model storms, which are dependent, in part, on the sounding utilized and assumptions in the microphysical parameterizations employed (e.g., Gilmore et al. 2004). Historically, small outflow temperature deficits have usually not been produced in supercell thunderstorm simulations. It is qualitatively worth noting that the air in the rear-flank downdraft (RFD) during the Jacksonville tornado (number 11 in Fig. 2) felt unusually mild for outflow to the University of Illinois chase teams with similar responses from interviews with Urbana residents apparently within the RFD during the Urbana tornado (number 32 in Fig. 2). These admittedly qualitative observations when coupled with the small quantitative outflow temperature–dewpoint deficits noted previously may be consistent with the thermodynamic signal noted by Markowski et al. (2002) for supercell RFDs associated with tornadoes. At the time of these observations, the observers were not predisposed to be looking for “warm/mild” RFD signals given that these observations were taken well before the dissemination of Markowski et al.’s findings on RFD thermodynamics.

Although this is a single-outbreak case study where the findings are applicable to similar environments, there are conclusions we can draw that are of interest to forecasters when nowcasting aspects of severe convection. Forecasters should pay special attention to storm mergers as an indicator of a heightened tornado threat, especially when the background storm environment features high relative humidity and low LCLs that sometimes lead to outflows with small temperature deficits. To this end, the identification of developing weaker cells whose anticipated paths could intersect the projected position of a preexisting supercell may provide essential lead time. There exists at least circumstantial evidence that subsequent cell mergers with a supercell may prompt cyclic tornadogenesis.

Given the potential benefits to severe storm nowcasting derived from a better understanding of cell merger processes, further research is required. In a preliminary modeling study to identify the dependence of storm interactions on the product storm intensity (Jewett et al. 2002) the authors found that a narrow range of collision orientations positively influenced merged cell intensity. Further numerical studies investigating the parameters space of merger dependence on cell initiation and storm environment are needed. Additional observational case studies and/or field experiments are required to understand the frequency with which tornadoes are associated with cell mergers and other interactions. Finally, future comprehensive high space and time resolution multi-Doppler radar studies are essential for gaining a better understanding of the cloud-scale processes involved in various modes of cell merger. The deployment of mobile C-band radars (e.g., the SMART-Rs; Biggerstaff et al. 2005) in a field project setting with their favorable attenuation characteristics (for mobile platforms) could be quite effective in gathering the requisite cell merger datasets.

Acknowledgments

This study was funded by NSF Grants ATM-9986672, ATM-0102579, ATM-0432408, and ATM-0449753. We thank Matt Gilmore and Cathy Finley for providing many helpful suggestions that improved this manuscript. We thank the three anonymous reviewers for their thoughtful comments and suggestions. We appreciate the use of John Hart’s SeverePlot software for plotting the historical tornado reports from the Storm Prediction Center archive. The Department of Atmospheric Sciences at the University of Illinois provided essential archived data and software support. The authors recognize the field support contributions of Tim Shy and Tom Grzelak and the nowcasts provided by Edward Kieser.

REFERENCES

  • Adlerman, E. J., and Droegemeier K. K. , 2000: A numerical simulation of cyclic tornadogenesis. Preprints, 20th Conf. on Severe Local Storms, Orlando, FL, Amer. Meteor. Soc., 591–594.

  • Agee, E. M., Snow J. T. , and Clare P. R. , 1976: Multiple vortex features in the tornado cyclone and the occurrence of tornado families. Mon. Wea. Rev., 104 , 552563.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Biggerstaff, M. I., and Coauthors, 2005: The Shared Mobile Atmospheric Research and Teaching Radar: A collaboration to enhance research and teaching. Bull. Amer. Meteor. Soc., 86 , 12631274.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bluestein, H. B., and Pazmany A. L. , 2000: Observations of tornadoes and other convective phenomena with a mobile, 3-mm wavelength, Doppler radar: The spring 1999 field experiment. Bull. Amer. Meteor. Soc., 81 , 29392951.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bluestein, H. B., and Weisman M. L. , 2000: The interaction of numerically simulated supercells initiated along lines. Mon. Wea. Rev., 128 , 31283149.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bluestein, H. B., and Wakimoto R. M. , 2003: Mobile radar observations of severe convective storms. Radar and Atmospheric Science: A Collection of Essays in Honor of David Atlas, Meteor. Monogr., No. 52, Amer. Meteor. Soc., 105–136.

    • Search Google Scholar
    • Export Citation
  • Brady, R. H., and Szoke E. J. , 1989: A case of nonmesocyclone tornado development in northeast Colorado: Similarities to waterspout formation. Mon. Wea. Rev., 117 , 843856.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Burgess, D. W., Wood V. T. , and Brown R. A. , 1982: Mesocyclone evolution statistics. Preprints, 12th Conf. on Severe Local Storms, San Antonio, TX, Amer. Meteor. Soc., 422–424.

  • Byers, H. R., and Braham R. R. Jr., 1949: The Thunderstorm Project. U.S. Government Printing Office, 287 pp.

  • Changnon S. A. Jr., , 1976: Effects of urban areas and echo merging on radar echo behavior. J. Appl. Meteor., 15 , 561570.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chin, H-N. S., and Wilhelmson R. B. , 1998: Evolution and structure of tropical squall line elements within a moderate CAPE and strong low-level jet environment. J. Atmos. Sci., 55 , 30893113.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cunning, J. B., Holle R. L. , Gannon P. T. , and Watson A. I. , 1982: Convective evolution and merger in the FACE experimental area: Mesoscale convection and boundary layer interactions. J. Appl. Meteor., 21 , 953978.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Davies, J. M., Doswell C. A. III, Burgess D. W. , and Weaver J. F. , 1994: Some noteworthy aspects of the Hesston, Kansas, tornado family of 13 March 1990. Bull. Amer. Meteor. Soc., 75 , 10071017.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dowell, D. C., and Bluestein H. B. , 2002a: The 8 June 1995 McLean, Texas, storm. Part I: Observation of cyclic tornadogenesis. Mon. Wea. Rev., 130 , 26262648.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dowell, D. C., and Bluestein H. B. , 2002b: The 8 June 1995 McLean, Texas, storm. Part II: Cyclic tornado formation, maintenance, and dissipation. Mon. Wea. Rev., 130 , 26492670.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fierro, A. O., Gilmore M. S. , Wicker L. J. , Mansell E. R. , Straka J. M. , and Rasmussen E. N. , 2004: Electrification and lightning in an idealized boundary-crossing supercell simulation of 2 June 1995. Preprints, 22d Conf. on Severe Local Storms, Hyannis, MA, Amer. Meteor. Soc., CD-ROM, 16B.7.

  • Finley, C. A., Cotton W. R. , and Pielke R. A. Sr., 2001: Numerical simulation of tornadogenesis in a high precipitation supercell. Part I: Storm evolution and transition into a bow echo. J. Atmos. Sci., 58 , 15971629.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Finley, C. A., Cotton W. R. , and Pielke R. A. Sr., 2002: Tornadogenesis in a simulated HP supercell. Preprints, 21st Conf. on Severe Local Storms, San Antonio, TX, Amer. Meteor. Soc., 531–534.

  • Fujita, T. T., 1960: A detailed analysis of the Fargo tornadoes of June 20, 1957. U.S. Weather Bureau Research Paper 42, Chicago, IL, 67 pp.

  • Fujita, T. T., Bradbury D. L. , and Van Thullenar C. F. , 1970: Palm Sunday tornadoes of April 11, 1965. Mon. Wea. Rev., 98 , 2669.

  • Gilmore, M. S., and Wicker L. J. , 2002: Influences of the local environment on supercell cloud-to-ground lightning, radar characteristics, and severe weather on 2 June 1995. Mon. Wea. Rev., 130 , 23492372.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gilmore, M. S., Straka J. M. , and Rasmussen E. N. , 2004: Precipitation uncertainty due to variations in precipitation particle parameters within a simple microphysics scheme. Mon. Wea. Rev., 132 , 26102627.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Goodman, S. J., and Knupp K. R. , 1993: Tornadogenesis via squall line and supercell interactions: The November 15, 1989, Huntsville, Alabama, tornado. The Tornado: Its Structure, Dynamics, Prediction, and Hazards, Geophys. Monogr., Vol. 79, Amer. Geophys. Union, 183–199.

    • Crossref
    • Export Citation
  • Jewett, B. F., Wilhelmson R. B. , and Lee B. D. , 2002: Numerical simulation of cell interaction. Preprints, 21st Conf. on Severe Local Storms, San Antonio, TX, Amer. Meteor. Soc., 353–356.

  • Kogan, Y. L., and Shapiro A. , 1996: The simulation of a convective cloud in a 3D model with explicit microphysics. Part II: Dynamical and microphysical aspects of cloud merger. J. Atmos. Sci., 53 , 25252545.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lee, B. D., and Wilhelmson R. B. , 1997: The numerical simulation of non-supercell tornadogenesis. Part II: Evolution of a family of tornadoes along a weak outflow boundary. J. Atmos. Sci., 54 , 23872415.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lee, B. D., Jewett B. F. , and Wilhelmson R. B. , 2006: The 19 April 1996 Illinois tornado outbreak. Part I: Cell evolution and supercell isolation. Wea. Forecasting, 21 , 433448.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lemon, L. R., 1976: The flanking line, a severe thunderstorm intensification source. J. Atmos. Sci., 33 , 686694.

  • Malkus, J. S., 1954: Some results of a trade-cumulus cloud investigation. J. Meteor., 11 , 220237.

  • Markowski, P. M., Rasmussen E. N. , and Straka J. M. , 1998: The occurrence of tornadoes in supercells interacting with boundaries during VORTEX-95. Wea. Forecasting, 13 , 851859.

    • Search Google Scholar
    • Export Citation
  • Markowski, P. M., Rasmussen E. N. , and Straka J. M. , 2002: Direct surface thermodynamic observations within rear-flank downdrafts of nontornadic and tornadic supercells. Mon. Wea. Rev., 130 , 16921721.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • NCDC, 1996: Storm Data. Vol. 38, No. 4, 410 pp. [Available from National Climatic Data Center, 151 Patton Ave., Asheville, NC 28801.].

  • NSSL, 2000: WATADS (WSR-88D Algorithm Testing and Display System) 2000: Reference guide for version 10.2. NSSL, 200 pp. [Available from Storm Scale Applications Division, National Severe Storms Laboratory, 1313 Halley Circle, Norman, OK 73069.].

  • Purdom, J. F., 1976: Some uses of high-resolution GOES imagery in the mesoscale forecasting of convection and its behavior. Mon. Wea. Rev., 104 , 14741483.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rasmussen, E. N., Davies-Jones R. P. , Doswell C. A. III, Straka J. M. , Carr F. H. , Eilts M. D. , and MacGorman D. R. , 1994: The Verification of the Origins of Rotation in Tornadoes Experiment: VORTEX. Bull. Amer. Meteor. Soc., 75 , 9951005.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Rasmussen, E. N., Richardson S. , Straka J. , Markowsiki P. , and Blanchard D. , 2000: The association of significant tornadoes with a baroclinic boundary on 2 June 1995. Mon. Wea. Rev., 128 , 174191.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Sabones, M. E., Agee E. M. , and Akridge M. , 1996: The Palaski County and West Lafayette, Indiana, tornadoes, 26–27 April, 1994: A case study of supercell (mesocyclone) and squall-line bow echo interaction. Preprints, 18th Conf. on Severe Local Storms, San Francisco, CA, Amer. Meteor. Soc., 746–750.

  • Simpson, J., 1980: Downdrafts as linkages in dynamic cumulus seeding effects. J. Appl. Meteor., 19 , 477487.

  • Simpson, J., and Woodley W. L. , 1971: Seeding cumulus in Florida: New 1970 results. Science, 172 , 117126.

  • Simpson, J., Westcott N. E. , Clerman R. J. , and Peilke R. A. , 1980: On cumulus mergers. Arch. Meteor. Geophys. Bioklimatol. Ser. A, 29 , 140.

  • Stumpf, G. J., Witt A. , Mitchell E. D. , Spencer P. L. , Johnson J. T. , Eilts M. D. , Thomas K. W. , and Burgess D. W. , 1998: The National Severe Storms Laboratory Mesocyclone Detection Algorithm for the WSR-88D. Wea. Forecasting, 13 , 304320.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Tao, W-K., and Simpson J. , 1984: Cloud interactions and merging: Numerical simulations. J. Atmos. Sci., 41 , 29012917.

  • Tao, W-K., and Simpson J. , 1989: A further study of cumulus interactions and mergers: Three-dimensional simulations with trajectory analyses. J. Atmos. Sci., 46 , 29743004.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wakimoto, R. M., and Wilson J. W. , 1989: Non-supercell tornadoes. Mon. Wea. Rev., 117 , 11131140.

  • Weaver, J. F., and Nelson S. P. , 1982: Multiscale aspects of thunderstorm gust fronts and their effects on subsequent storm development. Mon. Wea. Rev., 110 , 707718.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Weaver, J. F., and Purdom J. F. W. , 1995: An interesting mesoscale storm-environment interaction observed just prior to changes in severe storm behavior. Wea. Forecasting, 10 , 449453.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Weaver, J. F., Purdom J. F. W. , and Szoke E. J. , 1994: Some mesoscale aspects of the 6 June 1990 Limon, Colorado, tornado case. Wea. Forecasting, 9 , 4561.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Westcott, N., 1984: A historical perspective on cloud mergers. Bull. Amer. Meteor. Soc., 65 , 219226.

  • Westcott, N., 1994: Merging of convective clouds: Cloud initiation, bridging, and subsequent growth. Mon. Wea. Rev., 122 , 780790.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Westcott, N., and Kennedy P. C. , 1989: Cell development and merger in an Illinois thunderstorm observed by Doppler radar. J. Atmos. Sci., 46 , 117131.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wicker, L. J., and Wilhelmson R. B. , 1995: Simulation and analysis of tornado development and decay within a three-dimensional supercell thunderstorm. J. Atmos. Sci., 52 , 26752703.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wiggert, V. G., Lockett J. , and Ostlund S. S. , 1981: Rainshower growth histories and variations with wind speed, echo motion, location and merger status. Mon. Wea. Rev., 109 , 14671494.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wilkins, E. M., Sasaki Y. K. , Gerber G. E. , and Chaplin W. H. Jr., 1976: Numerical simulation of the lateral interactions between buoyant clouds. J. Atmos. Sci., 33 , 13211329.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Witt, A., Eilts M. D. , Stumpf G. J. , Mitchell E. D. , Johnson J. T. , and Thomas K. W. , 1998: Evaluating the performance of WSR-88D severe storm detection algorithms. Wea. Forecasting, 13 , 513518.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wolf, R., Pryzbylinski R. , and Berg P. , 1996: Observations of a merging bow segment and supercell. Preprints, 18th Conf. on Severe Local Storms, San Francisco, CA, Amer. Meteor. Soc., 740–745.

Fig. 1.
Fig. 1.

Cell maximum radar reflectivity centroid tracks at the 0.5° elevation angle for 1940–0200 UTC 19–20 Apr 1996. See legend for track-type designations. The letters D, W, and OC specify storms initiating near the dryline, warm front, or dryline–warm front occlusion boundaries, respectively. The M and I represent storm merger and nonmerger interactions. Cell designations with an S appended designate storms resulting from storm splits. Tracks with arrows specify storms that were still under way at 0200 UTC. See text for details on the convective initiation cell tracking window and on cell numbering.

Citation: Weather and Forecasting 21, 4; 10.1175/WAF943.1

Fig. 2.
Fig. 2.

As in Fig. 1 except with only the cyclonic supercell tracks and those contributory cells that were involved directly or indirectly in mergers with supercells. The bold blue lines designate supercell tracks of the radar reflectivity centroid. The superposed green and red line segments indicate when the storm had a demonstrable mesocyclone or TVS, respectively, based largely on WATADS algorithms. Tornado tracks are indicated in pink and labeled in the order they occurred. The light gray cell tracks are for two large discrete cells that developed along the flanking lines of two preexisting supercells. The gray designations for D16 indicate either a cell merger or nonmerger interaction from unplotted cells.

Citation: Weather and Forecasting 21, 4; 10.1175/WAF943.1

Fig. 3.
Fig. 3.

KILX radar reflectivity imagery (0.5°) at 2159, 2216, 2234, 2252, and 2309 UTC. The cyan dashed lines represent cell tracks. Specific cell mergers referenced in the text are labeled. Cells experiencing multiple mergers have the sequential merger number appended (e.g., D16M3 represents the third merger event for D16). Reflectivities less than 15 dBZ are not plotted.

Citation: Weather and Forecasting 21, 4; 10.1175/WAF943.1

Fig. 4.
Fig. 4.

KILX radar reflectivity imagery (3.3°) featuring the type C merger D12FM (merger of D12 and D12F). Panels are at times 2353, 2359, and 0004 UTC. The projected maximum reflectivity centroids for D12 and D12F are indicated in the 0004 UTC panel for geographical reference.

Citation: Weather and Forecasting 21, 4; 10.1175/WAF943.1

Fig. 5.
Fig. 5.

KILX radar reflectivity imagery (0.5°) featuring the type D merger of D16 and D18. Panel times are 2309, 2321, and 2332 UTC.

Citation: Weather and Forecasting 21, 4; 10.1175/WAF943.1

Fig. 6.
Fig. 6.

Histogram of tornado incidence times relative to the time of the cell merger partitioned in 5-min intervals for the 15-min period before and after cell merger.

Citation: Weather and Forecasting 21, 4; 10.1175/WAF943.1

Table 1.

Postmerger cell maximum reflectivity changes.

Table 1.
Table 2.

Cell rotation changes associated with mergers.

Table 2.
Save