Despite enormous progress in our understanding of tornadic supercells obtained over the last several decades, many aspects regarding the evolution of these storms and their attendant tornadoes require further investigation. High-resolution numerical simulations of tornadic supercells represent an important tool to shed light on a number of physical processes associated with supercell tornadogenesis. In recent years, numerical simulations have been used successfully to better understand the role of surface friction, strong low-level updrafts, internal momentum surges, and other types of storm-scale processes and/or structures in tornadogenesis. In this dissertation, idealized, high-resolution (horizontal grid spacing ranging from to 30 to 100 m) numerical simulations of a tornadic supercell are performed using the Advanced Regional Prediction System (ARPS) in order to better understand some important aspects of the complex evolution of tornadic supercells at fine scales. These analyses focus on two aspects of tornado evolution. First, the cyclic nature of tornadogenesis is investigated, with emphasis on understanding how a supercell evolving in a horizontally homogenous, steady-state environment can produce significantly different tornadoes in each tornado cycle. Second, the evolution of horizontal vortex tubes sometimes observed near real tornadoes is also addressed. Given that surface friction can have a significant impact on the evolution of simulated tornadoes and surrounding storm structures, its effects are included in the simulations.
A 50-m grid spacing experiment is employed to address the questions regarding cyclic tornadogenesis. To avoid constant modification of the base-state wind profile by surface friction, a three-force balance between the horizontal pressure gradient, Coriolis, and frictional forces is enforced using the Geotriptic Wind Balance (GWB) technique, such that the wind profile remains quasi-steady over the course of the experiment away from storm-induced perturbations. The simulated supercell produces four tornadoes in relatively regular periods during its life span, three of which attain Enhanced Fujita 5 (EF5) intensity, while one briefly attains EF3 winds. All tornadoes develop under intensifying low-level updrafts and lowering pressure aloft, but their ensuing evolution differs considerably. The first tornado, also the strongest one, moves along the interface between a rear-flank downdraft (RFD) and a forward-flank convergence boundary (FFCB), while highly tilted to the tornado. When the tornado’s parent updraft sheds from the main updraft to the east, it moves under the midlevel updraft, strengthening to peak intensity until it is overtaken by cold outflow. After its dissipation, large amounts of precipitation in the rear-flank of the storm cause the subsequent tornadoes to have shorter life spans, as they tend to become “wrapped in rain” and detach from their parent low-level updraft too quickly. All tornadoes are preceded by a low-pressure lobe (LPL) associated with accelerating inflow into the tornado’s parent updraft. A band of enhanced near-surface streamwise vorticity in conjunction with the LPL and enhanced inflow also develops and appears to feed into the low-level updraft, potentially intensifying upward motions dynamically. Unlike previous conceptual models of cyclic mesocyclogenesis, where midlevel updrafts move rearward relative to the storm and decay completely, occluded midlevel updrafts merge with newly developed updrafts and produce convoluted downdraft distribution at middle levels and near the rear-flank of the storm. This setup is at least partially responsible for inducing the transition of the supercell from a “classic” morphology into an HP mode, a condition that accounts for most of the individual differences among tornadoes in this simulation. The interactions between HV and tornadoes are analyzed in a 100-m and a 30-m grid spacing simulations. For the 100-m grid spacing experiment (which is an early version of the 50- and 30-m simulations), visualizations of the three-dimensional (3D) flow field based on direct volume rendering aided by visual observations of HVs in a real tornado reveal the existence of a complex distribution of 3D vortex tubes surrounding the tornadic flow throughout the simulation. A distinct class of HVs originates in two key regions at the surface: around the base of the tornado and in the RFD outflow and are believed to have been generated via surface friction in regions of strong horizontal near-surface wind. HVs around the tornado are produced in the tornado’s outer circulation and rise abruptly in its periphery, assuming a variety of complex shapes, while HVs to the south-southeast of the tornado, within the RFD outflow, ascend gradually in the updraft. A combination of visual observations of a violent tornado and 3D visualizations of the vorticity field near the tornado in the 30-m simulation (which based on the same observed tornado case) are used to document a distinct type of HV, which persistently trails the right flank of the tornado very close to the ground, hereafter referred to as “trailing HV”. The analysis shows that trailing HVs are larger, stronger, and last longer than their small-scale counterparts. Still, their vorticity matches that of other smaller HVs previously documented in the literature, which is consistent with generation via frictional torques and baroclinity along warm RFD internal boundaries. Interestingly, in some instances, trailing HVs display smaller spiral vortices circulating their periphery, which may evolve into complex structures. Visualizations of the 3D vorticity field show that the trailing HV arises as an entanglement along an RFD internal boundary of large and small HVs originated in the RFD outflow during a period of tornado intensification. The RFD internal boundary also serves as focus for stretching of vorticity that is exchanged from originally crosswise into streamwise vorticity at the location of trailing HV, causing strengthening of the trailing HV. The spiral vortices result from the same entangling processes that gives rise to the trailing HV. Moreover, the analysis suggests that trailing HV may act as a rotor that reinforces the surface wind speed in the right flank of the tornado.