Storm Interactions: A Historical Perspective

Merger. Nudger. Collision. Interaction. All are words used in the Severe Local Storms community to describe the process when convective cells come into contact with one-another, influencing their behavior and evolution. Such changes in behavior may include increases in precipitation production and perhaps even the production of tornadoes (e.g., Finley et al. 2001; Lee et al. 2006; Nixon et al. 2024)

Though the phrase storm interactions has become widely used in the last several years following increasing interest in interactions in published literature (e.g. Flournoy et al. 2022; Fischer and Dalh 2023; Broyles et al. 2023; Nixon et al. 2024; Werkema et al. 2025; etc.), storm interactions have been studied for decades and observed for much longer. This is a history of storm interactions.

---

Why write this

As a student of Severe Local Storms and an armchair historian about any various topic that interests me, I hold a deep passion for the background of our science. My master's thesis (and PhD work) entails examining the influence of storm interactions on storm behavior and evolution. As such, I conducted an extensive review of every piece of literature I could find on the subject of storm interaction for my thesis' introduction. It fascinates me so much that I wanted to share it with you.

This will be a simple digestible history of storm interaction research. A discussion on what storm interactions are and their impact on storm behavior will be the topic of a future article.

---

The early years

1940s-1950s: The realization that reconnaissance radars may have application as tools for weather observation impacted the field of Meteorology in countless ways. One such application, which in this writer's view is perhaps among the most foundational studies of convection our field has known, was The Thunderstorm Project of the middle-late 1940s (Byers and Braham 1949). This joint government project (between the US Army Air Force, Navy, Weather Bureau and NASA's precursor, NACA) studied thunderstorms, primarily in Florida and Ohio, through armored aircraft, radiosondes, mobile radars, timelapse photography and more (Figure 1). While the primary results of the Thunderstorm Project included documentation of turbulence structure around thunderstorms, proof that radar was a useful tool for pilots, and documentation of thunderstorm cell lifecycles (a model largely still in use today), there is a particular sentence buried in that 331 page document that is relevant to the topic at hand, it reads:

"in a few cases, two adjacent lines having different orientations tended to merge, showing an apparent 'transfer' of echoes from one line to the other"

The Thunderstorm Project, as best as I can tell, holds the first reference to storm interactions in our field's literature. Being that the Thunderstorm Project was the first large-scale study of thunderstorms, its quite remarkable that we were observing storm interactions and considering their importance from the outset.

On May 28th, 1954, a severe weather event in central Illinois was being documented by the early APS-15 x-band radar at Champaign, Illinois. During the evening, around 5:00 PM local time, a tornado was produced east of Peoria, Illinois following an apparent "merger of echoes" (Figure 2). This event and its data were documented the following year by Stout and Hiser (1955). A fascinating excerpt from their study reads:

"An experienced radar operator on duty at the time, perceiving that a tornado might be present, notified other members of the staff. The tornado formed at about the time and place the edges of the echoes first touched."

This is the first instance of a documented temporal "relationship" between storm interactions and the genesis of tornadoes -- on the ex-Naval patrol aircraft x-band radar in 1954.

Almost exactly one year later, in northern Oklahoma and southern Kansas, a number of communities would be devastated by powerful tornadoes -- including the towns of Blackwell, Oklahoma and Udall, Kansas. The event in its entirety would tally 80 fatalities and nearly 300 injuries. Staats and Turrentine (1956) would conduct a study of this event, tracking the movement of radar echoes over the event's duration (Figure 3). Their analysis of echo trajectories yielded the curious observation:

"It is interesting to note that, in general, the trajectories seem to converge in the area east of Udall"

1960s-1970s: Following these events, a US Weather Bureau Eastern Region technical training memorandum titled A Review of Use of Radar in Detection of Tornadoes and Hail (Hamilton 1969) included a section that encouraged radar operators to watch for interacting echoes as the two aforementioned studies documented tornado development shortly following an interaction:

“Many investigators have detected tornado development when separate echoes merge into a single larger echo mass”

Small sidebar here: isn't it fascinating how little information our early pioneers could use to generate ideas and identify patterns? So many foundational theories and ideas formed early in the history of our science weren't found through intense interrogation of high-resolution simulations, but rather crudely drawn interpretations of radar data or simply a set of carefully-observant eyes. Fujita, Browning, Brandes, Fawbush, Miller, and Doswell are just a few names that come to mind who possessed this remarkable skill.

A study of an Oklahoma supercell thunderstorm on June 25th, 1969 by Lemon (1976) again revealed another instance of a storm interaction leading to changes in storm behavior. Lemon documented what he defined as a "flanking line cell merger", in which a "neighbor" cell merged into the rear-flank of a parent supercell (resembling what we might call a "rear-flank storm interaction" or "nudger" today). This interaction...

...“led to the intensification of the adjoining supercell through a merger process [which resulted in an] increase in buoyancy, vertical acceleration, and possibly rotation.”

“...in the area where [an auxiliary cell] merged with the supercell, the hook echo forms and is maintained”

That second line is the first instance in the literature of storm interactions being associated with changes in Rear-Flank Downdraft (RFD) behavior -- an important piece to the storm interaction puzzle today. The same year, a study out of the University of Chicago's Journal of Satellite and Mesometeorology found remarkably similar behavior while analyzing data from a localized tornado outbreak in southwest Missouri on March 26th, 1976 (Pearl 1976). The data Pearl analyzed, including radar, satellite, and Learjet photographs, revealed that tornado development seemed to follow the merging of "feeder echoes" and "feeder clouds" into the southern portion of the parent echo. Particularly relevant to my Master's work, Pearl speculated on what he observed in this line (my favorite historical quote) that reads:

“it is possible that in some cases the feeder cloud and the parent thunderstorm work in unison to produce a tornado, and in particular types of storms a hook echo may be the manifestation of their interaction”

While we understand today that this speculation isn't quite the case, it still shows that severe storm scientists are observing a relationship between storm interactions and "hook echo behavior" (i.e., RFD behavior) which is of particular interest today (e.g. Broyles et al. 2023; Werkema et al. 2025; Gillett et al. 2026).

1980s-1990s: Dr. Theodore Fujita's Northern Illinois Meteorological Research On Downbursts project (otherwise cleverly known as project NIMROD) observed Illinois thunderstorms in the 1970s (Fujita 1979). The project's observations consisted of several storm interaction cases. In a study using NIMROD observations, Westcott and Kennedy (1989) revealed that interaction events could occur in a number of ways, including via a "bridging" or "bonding" of mid-level reflectivity as the distance between two merging cells decreased. Following these "bridges", rapid intensification of the dominant echo occurred (Figure 4).

Bluestein and Parker (1993) studied convective storm initiation along Texas drylines and subsequent evolution modes giving way to the first attempt at classifying storm interaction types. These types included "pair", "line-segment", "cluster", and "merger". A particular example shared in their study featured a "merger" type interaction that occurred immediately prior to the genesis of a weak tornado on 9 May 1982 (Figure 5).

---

The Modelers and DOWs

1990s-2000s: Investigation of the storm interaction problem through numerical modeling techniques began to emerge in the 1990s and 2000s following the development of idealized simulation systems in the 1970s and 1980s*. To my knowledge, the first study to purposefully simulate the interaction of severe convective storms was Bluestein and Weisman (2000), which served as an idealized modeling extension of Bluestein and Parker (1993). The authors used the Klemp-Wilhelmson cloud model with 1000m horizontal grid spacing to simulate four thermal bubble spaced 30km apart along a "line" (resembling a dryline boundary). They studied the effects and evolution of interactions in varying wind shear.

* see the landmark studies of Klemp and Wilhelmson (1978) & Weisman and Klemp (1982)

Though Bluestein and Weisman (2000) was the first to simulate interactions on purpose, others accidently simulated and studied storm interactions, such as Kulie and Lin (1998) and Finley et al. (2001). In Kulie and Lin's study, they used the Terminal Area Simulation System (TASS) model at 750m horizontal grid spacing to investigate a 1988 Raleigh, North Carolina tornado event. Their experiment simulated numerous flanking-line interactions (similar to the phenomenon documented by Lemon 1976) that increased updraft strength of their parent supercell. In the study by Finley et al., a nested-grid Colorado State University Regional Modeling System (RAMS) simulation of a high-precipitation* supercell experienced drastic changes in storm behavior following five interaction events. These "interaction events" were associated with increases in cold-pool strength, intensification of the low-level mesocyclone, and in some instances tornadogenesis. The supercell eventually evolved into a "bow-echo" following the strengthening cold pool and stronger RFD surges.

* The reader is referred to Moller et al. 1994 for supercell classifications

Later, a series of studies by Brian Jewett (and co-authors) between 2002 and 2006 were the first to employ the Weather Research and Forecasting (WRF) model to investigate storm interactions. The authors were also the first to employ a systematic ensemble-based experiment design. These simulations, using a central thermal bubble and an auxiliary thermal bubble, simulated the outcomes of storm interactions by changing the location of the auxiliary bubble within the domain*. Jewett et al. (2002) simulated two cyclonic cells at once while Jewett et al. (2006) placed a thermal bubble in the top half of the domain and another in the bottom half of the domain. Because of the initial conditions profile they used, a right mover split** (from the top half) and a left mover split (from the bottom half) would converge on each other and interact near the domain's center. Results of these studies (a few hundred simulations) revealed that in most cases, the interaction was destructive and resulted in weaker parent supercells. However, in the case of the constructive interactions, surface vertical vorticity was larger than the control simulations (a discrete supercell without interactions) -- especially when the auxiliary cell interacted with the parent's rear flank.

* thus changing the "storm interaction type"; see Hastings and Richardson (2016) or Nixon et al. (2024) for more details on that.

** storm "splitting" has been observed since the 1960s and modeled since the 1980s, see Weisman and Klemp (1984)

Storm interactions have been observed by mobile Dopper radar units (DOWs, Doppler-On-Wheels) during the 2000s and 2010s and were found to have varying implications on a parent supercell's behavior. Wurman et al. 2007 presented DOW observations a supercell thunderstorm near Kiefer, Oklahoma on 26 May 1997 that produced "repeated tornadogenesis" following several storm interactions (Figure 7). The authors documented

---

There's something to this

lee, hastings

---

The big ones

Several studies, especially recently, have focused on specific events and identified that storm interactions may have played some role in the event's evolution. In fact, some of the more notable tornado events of the last two decades featured parent supercells undergoing storm interactions, such as the May 2011 EF5 Joplin, Missouri tornado (Van Leer, 2013), the so-called April 2011 “super outbreak” event (Lyza & Flournoy, 2023), the May 2011 El-Reno-Piedmont, Oklahoma EF5 (Tanamachi et al., 2015), the May 2013 Newcastle-Moore, Oklahoma EF5 (Kurdzo et al., 2015) and the May 2013 El-Reno, Oklahoma 2.6 mile-wide EF-3 tornado (Wurman et al., 2014)

---

Today

2020s: Research prior to the last decade have, other than the studies included in the section There's something to this, simply observed storm interactions and briefly described their impact and speculated on why these impacts occur. In the current decade, more nuanced and controlled investigations into this problem have emerged. In particular, Broyles et al. (2023), Fischer and Dahl (2023), and Nixon et al. (2024) have applied new approaches and goals, such as identifying where most auxiliary cells are located prior to tornadogenesis (Figure ...) and constructing a pattern-recognition system for storm-interaction events.

Nixon et al. (2024) found that over 90% of the EF2+ tornado-producing supercell cases analyzed in their study were accompanied by external features (auxiliary cells, systems, or boundaries -- of that 90%, 60% of the external features were some form of auxiliary convection). Broyles et al. 2023, Fischer & Dahl 2023, and Nixon et al. 2024, using slightly different methodologies, concurrently identified that auxiliary cells seem to be most-frequently located in a parent-supercell's rear-flank prior to tornadogenesis. Such a finding is consistent with those early observations nearly half-a-century earlier that found "flanking-line mergers" or "rear-flank storm interactions" to be important to constructively altering a parent supercell's behavior.

Where do we go from here? Its hard to say. What I dream of is at least some degree of reliable predictability of storm-interaction outcomes. Perhaps we simply increase warning lead time or maybe down the line we begin to understand what role the environment plays in this phenomenon making it predictable before storms even fire.

My PhD work at the University of North Dakota will continue to investigate how two storm's actually interact with each other and how a storm interaction can constructively alter storm behavior and hazard production.

---

References and Resources

• See Dr. Cameron Nixon's website for a detailed walk-through of storm interaction types and their implications on storm behavior, hazard production, and storm-spotting. Check out his YouTube too!
• Check back to my site over the coming months for additional articles detailing my work and other facets of the storm-interaction problem!

---

---

• Byers, H. R., & Braham, R. R. (1949). The thunderstorm project. U.S. Government Printing Office
• Bluestein, H. B., & Parker, S. S. (1993). Modes of isolated, severe convective storm formation along the dryline. Monthly Weather Review, 121(5), 1354- 1372.
• Bluestein, H. B., & Weisman, M. L. (2000). The interaction of numerically simulated supercells initiated along lines. Monthly Weather Review, 128(9), 3128- 3149.
• Broyles, J., Potvin, C., Dial, G., Murnan, J., Shores, S., Lyons, A., … Cook, A. (2023). Tornadogenesis in high-end tornadic supercells (part 1)- the rfd surge, rfd occlusion and cell mergers.
• Finley, C. A., Cotton, W. R., & Pielke Sr., R. A. (2001). Numerical simulation of tornadogenesis in a high-precipitation supercell. part i: Storm evolution and transition into a bow echo. Journal of the Atmospheric Sciences, 58(13), 1597–1629.
• Fischer, J., & Dahl, J. M. L. (2023). Supercell-external storms and boundaries acting as catalysts for tornadogenesis. Monthly Weather Review, 151(1), 23- 38.
• Fischer, J., Flournoy, M. D., & Lyza, A. W. (2023). Comments on ”a climatology of cell mergers with supercells and their association with mesocyclone evolution” and ”the influence of cell mergers on supercell characteristics and tornado evolution on 27-28 april 2011”. Monthly Weather Review, 151(9), 2541- 2545.
• Flournoy, M. D., Lyza, A. W., Satrio, M. A., Diedrichsen, M. R., Coniglio, M. C., & Waugh, S. (2022). A climatology of cell mergers with supercells and their association with mesocyclone evolution. Monthly Weather Review, 150(2), 451- 461.
• Fujita, T. T. (1979, October). Objectives, operation, and results of project nimrod. In Preprints, eleventh conference on severe local storms. Kansas City, Missouri: American Meteorological Society
• Gillett et al
• Hamilton, R. E. (1969). A review of use of radar in detection of tornadoes and hail (Vol. 34; Tech. Rep.). Washington, D.C.: U.S. Weather Bureau, Eastern Region. (Technical Memorandum)
• Hastings, R., & Richardson, Y. (2010, October). Storm mergers. part i: Preliminary numerical investigations of merger events. In Preprints, 25th conference on severe local storms. Denver, Colorado: American Meteorological Society.
• Hastings, R., & Richardson, Y. (2016). Long-term morphological changes in simulated supercells following mergers with nascent supercells in directionally varying shear. Monthly Weather Review, 144(2), 471- 499.
• Hastings, R., Richardson, Y., & Markowski, P. (2012a, November). Mergers in supercell environments. part i: Conceptual models of mechanisms governing merger outcomes. In Preprints, 26th conference on severe local storms. Nashville, Tennessee: American Meteorological Society.
• Hastings, R., Richardson, Y., & Markowski, P. (2012b, November). Mergers in supercell environments. part ii: Tornadogenesis potential during merger as evaluated by changes in the near-surface low-level mesocyclone. In Preprints, 26th conference on severe local storms. Nashville, Tennessee: American Meteorological Society.
• Jewett, B. F., Lee, B. D., & Wilhelmson, R. B. (2000). Initiation and evolution of severe convection in the 19 april 1996 illinois tornado outbreak. In Preprints, 20th conference on severe local storms. Orlando, Florida: American Meteorological Society. (Paper P2.4, presented 12 September 2000, Session: Convective Initiation)
• Jewett, B. F., Przybylinski, R. W., & Wilhelmson, R. B. (2006). Numerical simulation of the 24 april 2002 storm merger between a left-moving storm and supercell. In Preprints, 23rd conference on severe local storms (p. P11.3). St. Louis, MO.
• Jewett, B. F., Wilhelmson., R. B., & Lee, B. D. (2002). Numerical simulation of cell interaction. In Preprints, 21st conference on severe local storms (p. P6.6). San Antonio, TX.
• Klees, A. M., Richardson, Y. P., Markowski, P. M., Weiss, C., Wurman, J. M., & Kosiba, K. K. (2016). Comparison of the tornadic and nontornadic supercells intercepted by vortex2 on 10 june 2010. Monthly Weather Review, 144(9), 3201
• Kulie, M. S., & Lin, Y.-L. (1998). The structure and evolution of a numerically simulated high-precipitation supercell thunderstorm. Monthly Weather Review, 126(8), 2090- 2116.
• Kurdzo, J. M., Bodine, D. J., Cheong, B. L., & Palmer, R. D. (2015). High-temporal resolution polarimetric x-band doppler radar observations of the 20 may 2013 moore, oklahoma, tornado. Monthly Weather Review, 143(7), 2711- 2735.
• Lee, B. D., Jewett, B. F., & Wilhelmson, R. B. (2006). The 19 April 1996 Illinois tornado outbreak. part ii: Cell mergers and associated tornado incidence. Weather and Forecasting, 21(4), 449- 464.
• Lemon, L. R. (1976). The flanking line, a severe thunderstorm intensification source. Journal of Atmospheric Sciences, 33(4), 686- 694.
• Lyza, A. W., & Flournoy, M. D. (2023). The influence of cell mergers on supercell characteristics and tornado evolution on 27-28 april 2011. Monthly Weather Review, 151(6), 1551- 1569.
• Nixon, C. J., Allen, J. T., Wilson, M. B., Bunkers, M. J., & Taszarek, M. (2024). Cell mergers, boundary interactions, and convective systems in cases of significant tornadoes and hail. Weather and Forecasting, 39(10), 1435- 1458.
• Pearl, E. W. (1976). Feeder clouds associated with tornado-producing mesocyclone of march 26, 1976 [SMRP Research Paper]. Satellite and MesoMeteorology Research Project(143).
• Rogers, J. W. (2012). Significant tornado events associated with cell mergers. In 26th conference on severe local storms. Nashville, TN: American Meteorological Society.
• Rogers, J. W., & Weiss, C. C. (2008). The association of cell mergers with tornado occurrence. In 24th conference on severe local storms. Savannah, GA: American Meteorological Society.
• Staats, W. F., & Turrentine, C. M. (1956). Some observations and radar pictures of the blackwell and udall tornadoes of may 25, 1955. Bulletin of the American Meteorological Society, 37(10), 495- 505.
• Stout, G. E., & Hiser, H. W. (1955). Radar scope interpretations of wind, hail, and heavy rain storms between may 27 and june 8, 1954. Bulletin of the American Meteorological Society, 36(10), 519–530. \
• Tanamachi, R. L., Heinselman, P. L., & Wicker, L. J. (2015). Impacts of a storm merger on the 24 may 2011 el reno, oklahoma, tornadic supercell. Weather and Forecasting, 30(3),501- 524.
• Van Leer, K. W. (2013). Storm mergers and their role in tornado genesis during the 2011 joplin storm (M.S. thesis). University of Illinois at Urbana–Champaign, Urbana, Illinois.
• Westcott, N. (1984). A historical perspective on cloud mergers. Bulletin of the American Meteorological Society, 65(3), 219- 226.
• Westcott, N., & Kennedy, P. (1989). Cell development and merger in an illinois thunderstorm observed by doppler radar. Journal of Atmospheric Sciences, 46(1), 117-131.
• Wurman, J., Richardson, Y., Alexander, C., Weygandt, S., & Zhang, P. F. (2007). Dual doppler and single-doppler analysis of a tornadic storm undergoing mergers and repeated tornadogenesis. Monthly Weather Review, 135(3), 736- 758.