Individuals' Responses to Tornado Warning Polygons

In the United States, the National Weather Service has replaced its original county-wide warnings with graphical displays (i.e., tornado polygons) to more clearly communicate tornado risk to the public. However, it is still unclear how recipients interpret this warning message and, in turn, respond to the warning. To better understand this issue, this study conducted an experiment asking 155 students at Texas A&M University to rate the likelihood of being struck by the tornado and their expected responses to 15 hypothetical tornado warning polygons. The results indicated that most of the participants were likely to respond appropriately to the tornado strike probabilities (p_s) they infer from the warning polygons. Specifically, p_s ratings are highest at the polygon’s centroid, lower just inside the edges of the polygon, still lower (but not zero) just outside the edges of the polygon, and lowest in locations far from the edges of the polygon. When the scenario positioned the respondents inside the warning polygons, they were more likely to expect to monitor additional information via television or the Internet. However, these scenarios elicited bipolar distributions on seeking warning confirmation from environmental cues, a desk clerk, or other guests. That is, people information search patterns were contingent on the content of the tornado warning and personal beliefs. Finally, respondents were more likely to seek a shelter and less likely to evacuate based on the warning. The study results will help meteorologists to better understand how people interpret the uncertainty associated with tornado warning polygons and, thus, improve tornado warning systems.