The southeastern United States experiences some of the greatest tornado fatality rates in the world, with a peak in the western portion of the state of Tennessee. Understanding the physical and social characteristics of the area that may lead to increased fatalities is a critical research need. Residents of 12 Tennessee counties from three regions of the state (N = 1804) were asked questions about their perception of climatological tornado risk in their county. Approximately half of participants underestimated their local tornado risk calculated from 50 years of historical tornado data. The percentage of participants underestimating their climatological risk increased to 81% when using model estimates of tornado frequencies that account for likely missed tornadoes. A mixed effects, ordinal logistic regression model suggested that participants with prior experience with tornadoes are more likely to correctly estimate or overestimate (rather than underestimate) their risk compared to those lacking experience (β = 0.52, p < 0.01). Demographic characteristics did not have a large influence on the accuracy of climatological tornado risk perception. Areas where more tornadoes go unreported may be at a disadvantage for understanding risk because residents' prior experience is based on limited observations. This work adds to the literature highlighting the importance of personal experiences in determining hazard risk perception and emphasizes the uniqueness of tornadoes, as they may occur in rural areas without knowledge, potentially prohibiting an accumulation of experiences.
Keywords: Climatology; Population bias; Prior experience; Risk; Tornado.