LAUSR

Tornadogenesis occurs in a variety of storm types, or convective modes, each having a unique climatology and challenges in their detection and warning. Some warnings result in false alarms, meaning that no tornado occurred within the warning polygon. We used a mixed-methods approach to assess how convective mode––discrete supercell, cell in cluster, cell in line, or quasi-linear convective system (QLCS)––affects the tornado climatology and National Weather Service (NWS) procedures within three County Warning Areas (CWAs): Memphis (MEG), Nashville (OHX), and Morristown (MRX), Tennessee. We used three data sets: tornadoes (2003–2014) categorized by convective mode, false alarms (2012–2016) categorized by convective mode, and eleven interviews of NWS forecasters. The CWAs had no significant difference in mode frequency when removing replication from multiple-tornado events. However, when outbreaks were included, discrete supercell and QLCS signals were identified in MEG and OHX, respectively. Convective mode, season, and time of day were strongly associated. Tornadic discrete supercells followed a traditional severe weather pattern of spring and daytime occurrences and caused fewer false alarms. More QLCS tornadoes happened at night and in winter. Cells in lines and clusters accounted for larger proportions of events in the false-alarm data set than the tornado data set. Forecasters noted challenges in detecting tornadoes in convective modes other than discrete supercells, including short-lived QLCS tornadoes. Key forecaster concerns other than convective mode included storm speed, outbreaks, and lack of ground-truthing at night. Forecasters differed in their motivation to either warn on every tornado or avoid false alarms. Key Words: climatology, false alarm, QLCS, supercell, tornado.