Super High-Resolution Simulation of the Fine-Scale Tornado Structure

Detailed structure of tornadoes remains unclear because of difficulties in collecting observational data with fine spatial and temporal resolutions. Most of previous studies have been conducted by photogrammetric analyses, laboratory experiments and idealized simulations such as a large eddy simulation, which contains calculational errors or some obvious unrealistic aspects. In this study, we performed downscale experiments with realistic conditions for Tsukuba supercell tornado (2012) using nested grids with as small as 10-m horizontal grid spacing to resolve the fine-scale tornado structure. The model contains 4001 x 3001 grid points in the horizontal and 250 vertical levels with grid intervals of 10 m near the surface.

Minimum pressure of a simulated tornado reaches 937 hPa (pressure deficit; 65 hPa), and maximum ground-relative surface wind speeds exceed 70 m s-1. During the rapid intensifying stage, the vortex core region accompanying large vertical vorticity contracted and was gradually occupied by downdraft. After that, the central downdraft intensified, and multiple vortices formed with an increase of horizontal dimension of a tornado. Thus, it is evident that the simulated tornado evolved from one-celled to two-celled tornado and subsequently exhibited multiple vortices, which are consistent with a tornado-like vortex evolution in laboratory experiments. There exist two prominent cyclonic subvortices associated with pressure deficit when most significant multiple vortices formed. Although subvortices locally intensify winds owing to the superposition of the velocity field associated with the small-scale subvortex and the larger-scale tornado, the strongest wind is found in the shrinking stage prior to multiple vortices.