A comparison of radar observations to real data simulations of axisymmetric tornadoes

The interaction between a tornado and its immediate environment is critical in determining tornado vortex dynamics. This research was the first to integrate tornado-scale observations into a numerical model of a tornado vortex. As such, this allowed for the direct comparison between observed and modeled tornado vortex structure. To enable this comparison, two methodologies were utilized: (1) an axisymmetric retrieval technique was developed to extract, from single-Doppler radar data, the 3D wind distribution and tornado vortex structure; and (2) a tornado-scale model was created that allowed for the specification of observed winds at the lateral and top boundaries.;The axisymmetric analysis of the 12 May 2004 DOW data revealed different vortical structures at different times in the tornado's evolution. Correspondingly, differences existed in the peak velocities and low-level angular momentum transport. As such, the second portion of this research addressed the issue of how well these observations can be modeled as well as the subsequent insight gained from analysis of the model results. Axisymmetric analyses from both the 12 May 2004 Harper, KS and 30 May 1998 Spencer, SD tornadoes were used as lateral boundary conditions on the numerical model. These boundary conditions were chosen during different times of the tornadoes' life cycles. The model results showed variability in vortex structure between each of the different sets of boundary conditions. Despite the analysis of fully two-celled vortices in the radar data, the model was unable to replicate a fully two-celled flow. To some degree this was shown to depend upon the chosen roughness values. The boundary conditions were found to exert control over tornado vortex intensity. The differences between the radar analyses and model results elucidated the critical properties of the flow necessary to replicate observations.