ABSTRACT
A detailed examination of the role played by preexisting airmass boundaries in a prototypical high-CAPE, low-shear environment was undertaken in this work. The first phase of the analysis focused on a long-lived, tornadic storm complex that affected central Texas on 27 May 1997. This event was notable because the storm complex developed in a high-CAPE, low-shear environment and propagated against the mean environmental wind along a dryline near its intersection with a cold front and the storm.s gust front.
Observational analysis of the 27 May 1997 central Texas event revealed that storm maintenance was driven by both quasi-continuous cell maintenance and discrete cell redevelopment. Both processes were dependent upon the preexisting boundaries; therefore, the preexisting boundaries ultimately controlled the longevity of the storm complex. Despite the small values of vertical shear associated with the environment of the 27 May 1997 event, mid-level mesocyclones were found in 6 of the 8 longest-lived, and well-sampled cells. However, no evidence of storm propagation governed by supercellular dynamics was found.
Observational analysis also revealed that all but one of the well-sampled tornadoes and all of the tornadoes that produced damage rated greater than F0 occurred on the portion of the storm-generated gust front that had been distorted by the preexisting cold front (the distorted gust front), not on the preexisting boundaries. However, because the cold front was responsible for the development of 1/3 of the observed cells as well as the distortion of the gust front it was concluded that tornadoes might not have developed in this event without the cold front.
Both mesocyclonic and non-mesocyclonic tornadoes were observed during this event however, the mesocyclonic tornadoes were found to be larger and more destructive than their non-mesocyclonic counterparts. Interestingly, all of the observed mesocyclonic tornadoes were found to develop near the time an existing updraft passed over the distorted gust front - a characteristic more commonly associated with non-mesocyclonic tornadoes.
In the second phase of this research numerical experiments were used to examine the open issues that emerged from the observational analysis. Simulations were conducted using an idealized cloud model with a horizontally heterogeneous initialization that included the dominant preexisting airmass boundaries of the 27 May 1997 event.
Numerical experiments conducted in the absence of preexisting boundaries demonstrated that a localized initialization procedure produced a simulated storm that was short-lived and therefore consistent with the high-CAPE, low-shear archetype. However, a linear initialization produced a line of initial cells that, through their reinforcement of a collective cold pool, insured cell redevelopment for the duration of the simulation (nearly 3 hrs).
A simulation conducted with only the dryline produced long-lived, back-building deep convection, thereby demonstrating that the cold front was not required for simulated long-lived deep convection. Cell redevelopment in this simulation was found to occur near the intersection of the dryline and the storm-generated gust front. Two modes of cell initiation were identified at this boundary intersection. The first mode cell development occurred as the deepened boundary layer, and associated reduced CIN magnitude, produced by the low-level convergence just east of the dryline was lifted by the storm generated gust front. This mechanism for cell development did not depend on the residual dryline secondary circulation (RDSC) as proposed by Weiss and Bluestein to explain preferential cell development at dryline-outflow boundary intersections. The second mode of cell redevelopment did appear to depend on the RDSC since the lifting of the boundary layer air by the storm-generated gust front was insufficient to initiate cell development. The dyline-only simulation also demonstrated that strong ( >0.01 s-1), temporally coherent, near-surface rotation could develop without the cold front but required the distortion of the storm-generated gust front. Furthermore, when the low-level winds behind the dryline were modified to resemble the winds northwest of the cold front, a more extensive gust front distortion developed and stronger near-surface rotation ensued. Thus, strong, temporally coherent near-surface vertical vorticity required the horizontal shear associated with either the dryline or the cold front.
When both the dryline and cold front were included in the initialization, the overall behavior of the simulated storm complex and boundaries closely resembled the early stages of the 27 May 1997 event. The quasi-continuous maintenance/propagation of the simulated storm was found to be a product of the flow of air through the zone of enhanced vertical motion along the front and distorted gust front where parcels gradually but continuously ascended to their level of free convection. It was concluded that the cold front ultimately controlled this process and therefore controlled the quasi-continuous maintenance/propagation of the simulated storm. Moreover, the role of the cold front in modifying cell discreteness also demonstrated that cell periodicity could be modulated by preexisting boundaries.
Simulated mid-level mesocyclones (defined as mesocyclones near the level of peak vertical velocity values) were dependent on the preexisting boundaries: when preexisting boundaries were excluded from the initialization, updrafts were too unsteady and short-lived to support temporally coherent regions of strong vertical vorticity. The direct contribution to the vertical vorticity of simulated mid-level mesocyclones from low-level vertical vorticity was negligible. Instead the simulated mid-level vertical vorticity generation was found to resemble the model for the generation of supercellular mid-level mesocyclones.
The weak low-level mesocyclone (defined as a mesocyclone just above cloud base) observed in the multiple boundary experiment were found to rely, in part, on the tilting of ambient vertical shear and the vertical transport and stretching of boundary layer vertical vorticity generated along the distorted gust front. It did not depend on the tilting of solenoidally generated vertical shear. Moreover, some of the processes responsible for the generation of the low-level mesocyclone appeared to be uniquely favored in the presence of the distorted gust front and, by extension, the cold front.
The production of near-surface vertical vorticity simulated in the multiple boundary experiment was the result of tilting of ambient vertical shear and the (downward) tilting of vertical shear generated in the buoyancy gradient along the extension of the distorted gust front west of the dryline. This latter contribution appeared to be uniquely favored in the presence of the distorted gust front and cold front.