Gabriel Garfield

Submitted 3 May 2005

1. Introduction

For many years, it has been observed that a supercell thunderstorm can produce multiple tornadoes in a periodic fashion (e.g., Fujita, 1960; Agee et al., 1976). Traditionally, there has been an association of this phenomenon with the “series” model (Agee et al. 1976). However, it is of great interest also to examine the behavior of storms that are in “parallel” mode (Agee et al. 1976). That is, tornadic thunderstorms that exhibit behavior that resembles that of multiple vortex tornadoes. This will be examined in detail in the latter portion of this review.

The process that leads to this periodicity in tornado formation is referred to as “cyclic mesocyclogenesis” and is defined as “a periodic succession of low and/or mid-level mesocyclones” (Adlerman et al. 1999). A mesocyclone is defined as “3-9 kilometer diameter region of vertical vorticity of greater than 0.01 s^-1” (Doswell and Burgess 1993). However, low-level mesocyclones are not on the same scale as what is traditionally referred to as the mesocyclone (Agee et al. 1976). Within the larger scale circulation (mesocyclone; as defined by Doswell and Burgess 1993), smaller scale vortices known as “tornado cyclones” (also known as low-level mesocyclones) can produce tornadoes in a periodic manner, or even produce tornadoes simultaneously (Agee et al. 1976). Although the behavior of these vortices is analogous to multiple-vortex tornadoes, the scale of the “tornado cyclone” is somewhat larger than the tornado (Agee et al., 1976; see Figure 1). As with multiple-vortex tornadoes, the existence of cyclic tornadic behavior is strongly linked with the presence of a downdraft (Ward 1972).

The rear-flank downdraft (RFD) is believed to be strongly associated with the development of cyclic tornadic thunderstorms (Adlerman et al., 1998; Lemon and Doswell 1979). Observations of tornadic supercells support this conclusion (Rasmussen et al., 1982; Jensen et al., 1983). The RFD is thought to re-orient horizontal vorticity into the vertical by tilting and stretching, much in the same way that a mid-level mesocyclone forms in a developing supercell thunderstorm (Davies-Jones 1984). However, numerical simulations of supercell thunderstorms suggest that the tilting and stretching of ambient vorticity (due to vertical wind shear) is not sufficient to generate vorticity strong enough to induce low-level mesocyclogenesis (Davies-Jones, 1982; Davies-Jones and Brooks, 1993; Davies-Jones, 1996). For this reason, other sources of horizontal vorticity are required to produce the necessary vertical vorticity to produce a low-level circulation capable of producing a tornado. Burgess et al. (1982) suggested that the tilting and stretching of pre-existing vorticity in the low-levels might be sufficient for low-level mesocyclone generation. Adlerman et al. propound that baroclinic generation of vorticity (by way of a horizontal density gradient) is the most likely source for this ambient horizontal vorticity.

2. The Role of Downdrafts in Cyclic Mesocyclogenesis

A dual-Doppler radar analysis of the Kellerville cyclic tornadic thunderstorm of 8 June 1995 showed that low-level mesocyclogenesis did not occur until strong outflow enhanced convergence beneath its mesocyclone (Dowell and Bluestein 2002). This agrees with previous studies that have shown that tornadoes form in regions of strong vertical velocity gradients (Adlerman et al., 1999; Brandes, 1978). In addition to the ambient vorticity due to vertical wind shear, horizontal vorticity generated by baroclinicity (due to evaporative cooling/water loading) is present in the lower levels of the atmosphere in the supercell environment (Adlerman et al. 1999). As the downdraft descends, the vortex lines are turned upward through the tilting of vorticity (Adlerman et al. 1999). During descent, most of the environmental vorticity is in the direction of the crosswise component of vorticity. However, upon impacting the ground, the vortex lines are rapidly tilted upward and a significant streamwise component is established as the downdraft rotates cyclonically toward the mesocyclone (Adlerman et al. 1999). This agrees with the findings of Markowski et al. (2002) who have found that a rear-flank downdraft with positive buoyancy and significant upward vertical velocities (immediately following impact with the ground) is more likely to support tornadogenesis.

The mechanism for the formation of the RFD is a subject of some debate. Most agree that evaporative cooling plays at least a minor role in the origin of the rear-flank downdraft (Adlerman et al., 1998; Lemon and Doswell, 1979). Kamburova and Ludlam (1966) initially stated that RFDs form by evaporational cooling alone. However, Clark and List (1971) hypothesized that the RFD is dynamically forced (i.e., a blocking updraft) and is only aided by evaporative cooling. Newton (1959) was the first to suggest that the updraft of a supercell thunderstorm acts as a “rock in a stream” (because of the lack of environmental entrainment) by diverting the mid-level flow downward (as a result of the dynamically induced pressure gradient force) at the edge of the updraft. An analysis of Doppler radar velocity measurements of tornadic thunderstorms by Lemon et al. (1977) confirmed this hypothesis.

In addition to evaporative cooling caused by falling precipitation, Adlerman et al. (1998) also refer to a second source of outflow by which the rear-flank downdraft is enhanced. In a numerical model simulation, Adlerman et al. (1999) showed that a secondary surge of outflow is initiated by a downward directed vertical perturbation pressure gradient force (due to ehanced rotation induced by the initial rear-flank downdraft surge) in the low levels. This secondary surge of outflow is known as the “occlusion downdraft”, and is not present in all models of cyclic mesocyclogenesis. Dowell and Bluestein (2002) surmised that evaporative cooling in the lower levels of the troposphere might result in locally enhanced outflow. In any case, low-level mesocyclogenesis did not occur without the aid of a significant downdraft. Lemon and Doswell (1979) showed that tornadogenesis did not occur in supercells until the mesocyclone became divided (creating a strong updraft/downdraft interface).

Interestingly, the same process that initiates the formation of the low-level mesocyclone is also the process that is thought to bring about its demise. According to Lemon and Doswell (1979), a tornado cyclone will continue until the circulation becomes “cut off” from its inflow by the rear-flank downdraft. Agee et al. (1976) describe two methods by which the low-level mesocyclone dissipates. First, the strong vertical velocity gradient (responsible for the formation of the low-level vortex) propagates downstream leaving the low-level circulation without a source of convergence. Second, the negative buoyancy associated with precipitation loading may cause low-level air to become too stable to rise to its level of free convection (Dowell and Bluestein 2002). Analysis of the Kellerville tornadic storm by Dowell and Bluestein (2002) showed that the storm’s outflow did not have any lack of convective available potential energy (CAPE); thus, the tornado did not dissipate because it ingested “cold” air, but because the low-level mesocyclone associated with the tornado lost strength (due to lesser values of vertical vorticity being ingested into the low-level mesocyclone).

As the RFD begins to “wrap” around the initial low-level mesocyclone, a new mesocyclone is initiated by the surging gust front (Adlerman et al. 1999; Brandes 1984; Dowell and Bluestein 2002). The new circulation then generates a new dynamically driven gust front surge as the older circulation begins to fall apart (Adlerman et al. 1999). This process may continue through multiple cycles and produce several intense low-level mesocyclones (including tornadoes).

3. Cyclic Mesocyclogenesis: Observations and Behavior

a. Observations

Observations of tornadic supercells indicate that, after the occlusion of the first low-level mesocyclone, the gust front surges eastward and a new mesocyclone forms where shear and convergence are maximized (Brandes 1984, Dowell and Bluestein 2002). The first low-level mesocyclone forms in a “vorticity rich” environment, which allows for the rapid formation of the second low-level mesocyclone (Burgess et al. 1982). In studies by Fujita (1963), Darkow and Roos (1970), and Darkow (1971), a statistical examination of the periodicity of mesocyclogenesis showed that the average time between tornado formation for each mesocyclone is approximately 45 minutes.

Like the tornado cyclone (low-level mesocyclone) that produces it, a tornado can only be maintained with a continual juxtaposition of vertical vorticity and horizontal convergence (Dowell and Bluestein 2002). Thus, the low-level mesocyclone will weaken if it moves into the core of precipitation because it will lose its source of buoyancy and subsequently, its source of forcing for stretching of vertical vorticity (Adlerman et al. 1999).

b. Behavior

It has been noted that the process of cyclic mesocyclogenesis resembles the life-cycle of a synoptic wave on a smaller scale (Brooks 1949). Like the synoptic cold front that accelerates and eventually merges with the warm front (i.e., occlusion), the rear-flank gust front rotates cyclonically around the low-level mesocyclone until it completely “cuts off” the inflow from the low-level mesocyclone (Lemon and Doswell 1979). As the RFD gust front occludes the original circulation, a new circulation forms downstream of the initial mesocyclone along the leading edge of the RFD gust front (Lemon and Doswell 1979). This circulation will be maintained as long as the outflow does not “outrun” the circulation (Dowell and Bluestein 2002). If the outflow is in relative “balance” with the inflow, this indicates the potential for the low-level mesocyclone to produce a long-track tornado (Dowell and Bluestein 2002). Stretching and advection of horizontal vorticity are the dominant terms in the lowest levels; thus, a low-level mesocyclone will tend to propagate in the direction of the local maxima in these quantities (Dowell and Bluestein 2002). The low-level mesocyclone will continue until the sources of enhanced vertical vorticity are reduced or eliminated (Dowell and Bluestein 2002).

As a result of cyclic mesocyclogenesis, multiple tornadoes (known as tornado “families”) can be produced by a single supercell storm. The track of each tornado depends upon the translational velocity of the storm, the translational velocity of the tornado cyclone, and the rotational velocity of the tornado cyclone about the larger scale mesocyclone (Agee et al. 1976). Tornado tracks viewed from above will have cycloidal damage paths that are similar in shape (but not in scale) to multiple-vortex tornadoes (Agee et al. 1976; see Figure 2).

Agee et al. (1976) identified three separate modes of cyclic mesocyclogenesis (see Figure 3). The first mode, known as the “series” mode, is characterized by tornado cyclones that form on roughly the same azimuth angle. Since the heading of the tornado family is relatively constant, continuous tornado paths from different tornado cyclones may appear as a single tornado (Agee et al. 1976). The second mode, known as the “parallel” mode, is characterized by behavior similar to multiple-vortex tornadoes (Agee et al. 1976). That is, individual tornado cyclones rotate about a common center (Agee et al. 1976). Finally, the third mode is a combination of the series and parallel mode (Agee et al. 1976).

Agee et al. (1976) hypothesize that the majority of long-track tornadoes form from a storm in series mode. This is because low-level mesocyclones in series mode have minimal interaction with the precipitation core, which would result in the weakening of the circulation (Agee et al. 1976). Fujita (1973) speculated that the periodicity of the internal dynamics of the updrafts is connected to the intensity of the low-level mesocyclones, which would be especially true for tornadic storms in series mode (because there is a smaller effect due to precipitation loading in series). Low-level mesocyclones in parallel mode will rotate around the center of the larger scale circulation, producing tornadoes when tilting and stretching are at their maximum (outside of the main precipitation core) and dissipate when precipitation loading overwhelms the low-level circulation (Agee et al. 1976).

4. Summary

In summary, cyclic mesocyclogenesis is primarily driven by the formation of the rear-flank downdraft (Dowell and Bluestein, 2002; Lemon and Doswell; 1979). The rear-flank downdraft is generated by a combination of dynamic forcing (Newton 1959) and evaporative forcing (Kamburova and Ludlam 1966). As this downdraft wraps cyclonically around the large-scale mesocyclone, it gains a significant component of streamwise vorticity (Adlerman et al. 1999). This vorticity is then stretched and tilted into the vertical in the vicinity of strong vertical velocity gradients (Adlerman et al. 1999). As the RFD gust front occludes this circulation, a new circulation will form downstream of the older circulation (Adlerman et. al., 1999; Dowell and Bluestein, 2002; Lemon and Doswell, 1979). The older circulation will die if either the source of its vorticity is “cut off” from the circulation (Lemon and Doswell 1979) or significant precipitation falls into the circulation (Agee et al., 1976; Dowell and Bluestein, 2002). The new circulation will continue until the conditions that brought it into existence (i.e., presence of streamwise vorticity being tilted and stretched into the vertical) are lessened or removed (Dowell and Bluestein 2002). Cyclic mesocyclogenesis will occur as long as conditions for developing new circulations are favorable.

References

Agee, E. M., J. T. Snow, and P. Clare, 1976: Multiple vortex features in the tornado cyclone and the occurrence of tornado families. Mon. Wea. Rev., 104, 552–563.

Adlerman, E. J., Droegemeier K. K., and R. Davies-Jones, 1999: A numerical simulation of cyclic mesocyclogenesis. J. Atmos. Sci., 56, 2045–2069.

Brandes, E. A., 1978: Mesocyclone evolution and tornadogenesis: Some observations. Mon. Wea. Rev., 106, 995–1011.

Brooks, E.M., 1949: The tornado cyclone. Weatherwise, 2, 32-33.

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

Clark, T.L., and R. List., 1971: Dynamics of a Falling Particle Zone. Journal of the Atmospheric Sciences: Vol. 28, No. 5, pp. 718–727.

Darkow, G.L., 1971: Periodic tornado production by long-lived parent thunderstorms. Preprints 7^th Conf. Severe Local Storms, Kansas City, AMS, Boston, Mass., 214- 217.

____, and J.C. Roos, 1970: Multiple tornado producing thunderstorms and their apparent cyclic variations in intensity. Preprints 14^th Radar. Meteorology Conference, AMS, Boston, Mass., 305-308.

Davies-Jones, R., 1982: Observational and theoretical aspects of tornadogenesis. Intense Atmospheric Vortices, L. Bengtsson and J. Lighthill, Eds., Springer-Verlag, 175–189.

____, 1984: Streamwise vorticity: The origin of updraft rotation in supercell storms. J. Atmos. Sci., 41, 2991–3006.

——, and H. E. Brooks, 1993: Mesocyclogenesis from a theoretical perspective. The Tornado: Its Structure, Dynamics, Prediction, and Hazards, Geophys. Monogr., No. 79, Amer. Geophys. Union, 105–114.

——, 1996: Formulas for the barotropic and baroclinic components of vorticity with applications to vortex formation near the ground. Preprints, Seventh Conf. on Mesoscale Processes, Reading, United Kingdom, Amer. Meteor. Soc., 14–16.

Doswell, C.A. III, and D.W. Burgess, 1993: Tornadoes and tornadic storms: A review of conceptual models. The Tornado: Its Structure, Dynamics, Prediction, and Hazards, Geophys. Monogr., No. 79Amer. Geophys. Union, 161-172.

Dowell, D.C., and Bluestein, H.B., 2002: The June 8, 1995 McLean, Texas, Storm. Part II: Cyclic Tornado Formation, Maintenance and dissipation. Mon. Wea. Rev., 130, 2649-2670.

Fujita, T.T., 1973: Proposed mechanism of tornado formation from rotating thunderstorms. Preprints 8^th Conf. Severe Local Storms, AMS Boston, Mass., 191-196.

____, 1960: A detailed analysis of the Fargo tornadoes of June 20, 1957. U.S. Weather Bureau Research Paper No. 42, Washington, D.C., 67 pp.

Jensen, B., E. N. Rasmussen, T. P. Marshall, and M. A. Mabey, 1983: Storm scale structure of the Pampa storm. Preprints, 13th Conf. on Severe Local Storms, Tulsa, OK, Amer. Meteor. Soc., 85–88.

Kamboruva, P.L., and F.H. Ludlam, 1966: Rainfall evaporation in thunderstorm downdrafts. Quart J. Roy. Meteo. Soc., 92, 510-518.

Lemon, L. R., and C. A. Doswell III, 1979: Severe thunderstorm evolution and mesocyclone structure as related to tornadogenesis. Mon. Wea. Rev., 107, 1184– 1197.

_____, Donaldson R.J., Burgess, D.W. and R.A. Brown, 1977: Doppler Radar Application to Severe Thunderstorm Study and Potential Real-Time Warning. Bulletin of the American Meteorological Society: Vol. 58, No. 11, pp. 1187–1193.

Markowski, P.M., Straka, J.M. and E.N. Rasmussen, 2002: Direct Surface Thermodynamic Observations within the Rear-Flank Downdrafts of Nontornadic and Tornadic Supercells. Monthly Weather Review: Vol. 130, No. 7, pp. 1692– 1721.

Newton, C.W., and Newton, H.R. 1959: Dynamical interactions between larger convective clouds and environment with vertical shear. Journal of the Atmospheric Sciences: Vol. 16, No. 5, pp. 483–496.

Rasmussen, E.N., R. E. Peterson, J. E. Minor, and B. D. Campbell, 1982: Evolutionary characteristics and photogrammetric determination of wind speeds within the Tulia outbreak of tornadoes 28 May 1980. Preprints, 12th Conf. on Severe Local Storms, San Antonio, TX, Amer. Meteor. Soc., 301–304.

Ward, N. B., 1972: The exploration of certain features of tornado dynamics using a laboratory model. J. Atmos. Sci., 29, 1194–1204.