On the role of descending rain curtains in tornadogenesis

The archetypal tornado vortex signature (TVS) descends from aloft toward the surface, preceding tornadogenesis by up to tens of minutes. However, descending TVSs account for only about half of all verified tornadoes also observed by Doppler radar. Non-descending TVSs comprise the other half, and form either simultaneously below cloud base or from the ground upwards. They are observed nearly concurrently with tornadogenesis, greatly reducing warning lead-time. Non-descending TVSs form in environments with vertical vorticity and radial inflow either constant beneath cloud base, or maximized at the surface.

We demonstrate a mechanism in which the rear flanking downdraft (RFD) and hook echo advect sufficient vertical vorticity at low levels to form a tornado strength vortex. These experiments are an extension of a previous study on a purely barotropic mechanism of tornadogenesis. Using Beltrami flow as the initial condition we simulate a mature mesocyclone, with a cyclonically rotating updraft maximized at midlevels surrounded by an anticylonically rotating downdraft. We perturb the flow from Beltrami by releasing hydrometeors above the updraft, which then spread out to the edges of the updraft in the divergent flow and descend. The hydrometeors fall through the downdraft as a rain curtain, representing the hook echo and transporting angular momentum to the surface. Near the surface, the hydrometeors converge inwards, producing and maximizing both vertical vorticity and radial inflow at the lowest levels as well as tangential wind. As there is no buoyant force sustaining the storm updraft, modeled storms collapse after tornadogenesis due to the downward-directed non-hydrostatic pressure gradient force associated with the low-level vortex. Thus, this model design can represent tornadogenesis but not tornado maintenance.

In a series of three-dimensional experiments, a circular plane of hydrometeors is released above the updraft. The maximum mixing ratio of the hydrometeors is varied between 1 g/kg to 9 g/kg in order to analyze the effect of varying precipitation amounts on low-level convergence of vertical vorticity and tornadogenesis. The results have implications for tornadogenesis in low precipitation (LP), classic (CL), and high precipitation (HP) supercells.

Many tornado simulations documented in the existing literature have been performed using axisymmetric models. We increase realism in a series of three-dimensional experiments with specified asymmetric rain curtains. In these experiments, the maximum mixing ratio of the hydrometeors is set to 5 g/kg, and the plane of hydrometeors released above the updraft is successively reduced by one-eighth slices. In axisymmetric simulations, horizontal vorticity cannot be tilted into the vertical. However, the asymmetric design of some of these experiments allows horizontal vorticity to be tilted vertically. The evolution of vertical vorticity in these experiments is analyzed in order to gain insight into the dynamics of tornadogenesis.