Masters Theses

Date of Award


Degree Type


Degree Name

Master of Science



Major Professor

Kelsey N. Ellis

Committee Members

Henri D. Grissino-Mayer, Lisa Reyes Mason


Tennessee resides in the Southeastern United States, a region prone to violent tornadoes on a year-round basis. With one of the highest tornado fatality rates in the country, and a climatology that varies across the state, analysis of storms resulting in Tennessee tornadoes is necessary for improving forecasting techniques and decreasing loss of life. This study analyzed convective modes responsible for Tennessee tornadoes from 2003 to 2014 to determine an association with fatalities, seasonality, day and night, tornado magnitude, regionality, and multiple-tornado days. Chi-squared tests were conducted to determine if these patterns were significant. National Weather Service forecasters from the Morristown, Tennessee, Weather Forecast Office (WFO) were interviewed to gain insight into how convective mode affects tornado forecasting and warning procedures.

Discrete supercells were the overwhelming producer of tornado-related fatalities, higher-magnitude (≥ EF2) tornadoes, and multiple-day tornadoes. Quasi-linear convective systems (QLCSs) produced only non-fatal tornadoes with magnitudes of ≤ EF2 during the period; however, QLCS tornadoes were more frequent at night and in winter, when the public may have been more vulnerable. Spring was the most tornadic season, but approximately 37% of tornadoes occurred outside of this season. Multiple-tornado days were major contributors to tornado totals, with just over half of the 427 tornadoes occurring on ten days. I found no clear longitudinal gradient of convective mode or tornado characteristics across Tennessee. Chi-squared results indicated a relationship between convective mode tornado production and fatalities, seasonality, day and night, magnitude, and region of Tennessee. Forecasters commented on the relative ease associated with warning for discrete supercells, resulting in higher probability of detection, lower false alarm ratios, and longer lead times when compared to QLCSs. Forecasters shared invaluable information concerning staffing considerations and warning decisions during severe weather events. This mixed-methods approach provided a comprehensive assessment of how convective mode affects tornado production and warning procedures, contributing to the emerging field of critical physical geography. Future work will include interviews with forecasters from the Memphis and Nashville WFOs, leading to a more comprehensive discussion of how modes differentially affect warning and forecasting procedures across Tennessee.

Radar_Log.pdf (5076 kB)
Log of manually assigned convective modes

NWS_Interview_Guide.pdf (66 kB)
National Weather Service interview guide

Files over 3MB may be slow to open. For best results, right-click and select "save as..."