The first case occurred during the late afternoon of April 13, 1998 when a broken line of thunderstorms formed west of St. Louis and moved through the metro area. As the line passed immediately east of the KLSX radar, mid-level rotation was detected with a convective cell within the line of storms. The mid-level rotation intensified and deepened quickly to mesocyclone criteria within several volume scans, while strong low-level convergence developed along the storm's gust front. Just prior to tornado formation, the mid-level rotation increased, while the strong convergence along the gust front became rotationally covergent. Low-level velocity data during the time of the tornado indicate several velocities peaks: a gate-to gate TVS feature probably directly associated with the tornado, and a slightly larger rotation believed to be a tornado cyclone. Structurally, the storm exhibited reflectivity characteristics of a High-Precipitation (HP) Supercell.
The second case occurred during the late afternoon of May 23, 1998 when a small linear convective cluster which formed in southwest Missouri moved through western portions of the St. Louis metro area. The evolution of this case was more complex and involved the interaction of the linear storm cluster with a storm scale boundary associated with another thunderstorm just northwest of the KLSX RDA. Velocity data revealed deep convergence along the gust front as the storm moved towards and over the RDA. As the storm passed immediately over the RDA and interacted with the storm scale boundary, deep convergent rotation develop along the northern portion of the gust front and throughout the observable radar data. The convergent signatures quickly evolved into a tight pure rotational signature as the storm moved to the northeast with gate-to-gate TVS shear at the time of the tornado. The evolution of this tornado appeared to be more consistent with the non-supercell variety.