Examining Public Protective Actions during Simultaneous Tornado and Flash Flood Threats in the U.S. Southeast

Background: Traditionally, disaster planning and communication has focused on specific, singular events. Severe weather events can create multiple hazard threats, such as simultaneous tornadoes and flash floods. These multi-hazard events, known as TORFF events, create a complex decision-making process for the people receiving warning protocols, as the recommended protective actions for the two hazards are contradictory––sheltering below ground during a tornado and moving to high ground during flash flooding events. Furthermore, when multi-hazard events occur, dimensions of social vulnerability may further compound protective decision-making as many people do not have access to the resources (i.e., basement, storm shelter, vehicle, etc.) necessary to respond safely to one or both hazards. Public response to TORFF events, which are warned for approximately 400 times per year, has not yet been examined. In the current study, we utilized structural equation modeling (SEM) to examine the factors and mechanisms of protective decision-making in members of the public (N = 247) who experienced concurrent tornado and flash flood threats on the 25 and 27 of March, 2021 in the U.S. Southeast region.

Methods: Data analyses were completed using R statistical software and packages. A two-step SEM procedure was conducted to test the study’s hypotheses. First, a measurement model was estimated to examine and confirm the factor structure of the latent variables and indicators (e.g., tornado and flash flood risk perceptions, tornado and flash flood protective actions). Second, a structural model was estimated to examine the associations between the different factors that are hypothesized to influence people’s protective actions during dual hazards (e.g., hazard warning sources, risk perception, behavioral actions).

Results: SEM analyses found that for the public, having access to more hazard-related information sources increased respondents’ tornado risk perception (β = 0.202, p<.01) and flash flood risk perceptions (β = 0.434, p<.001). Results also found that tornado risk perception was found to increase tornado protective actions (β = 0.302, p<.01), but increased tornado risk perception was found to decrease flash flood protective action (β = - 0.182, p<.05). Finally, results found that and flash flood risk perception was found to increase flash flood protective actions (β = 0.315, p<.01).

Conclusion and Implications: Our findings highlight the complexity of decision-making when tornadoes and flash flood threaten concurrently. Specifically, our findings indicated that while tornado risk perception increased tornado protective action, it decreased protective action for flash flood events. Furthermore, our findings highlight that social vulnerability may further compound protective decision-making. Although TORFF events can occur anywhere across the U.S., these events presented a unique issue as they occurred in a region of the U.S. with a high percentage of homes that are mobile/manufactured. Overall, these results highlight that hazard risk communication should provide equal attention to both threats, along with dual protective guidance, particularly for socially vulnerable populations. Understanding the protective decision-making process within multi-hazard risk situations is essential for developing mitigation and resilience-building strategies that provide social responses and solutions to a changing environment.