NWC REU 2023
May 22 - July 28

 

 

Photo of author

Using the Tornado Tales Survey to Measure Warning Reception and Response During Evening and Nocturnal Events

Lauren A. Harvey and Justin Sharpe

 

What is already known:

  • Nocturnal tornadoes are responsible for a disproportionately high number of fatal events.
  • The proportion of deaths due to nocturnal tornado events has increased between the 1960s and early 2000s, despite advances in tornado forecasting and warning systems.
  • People have far less confidence in their ability to receive tornado warnings overnight
  • Warning reception is needed to start the protective action decision-making process.
  • The challenge is to identify where the breakdown in the decision-making process occurs between warning issuance and protective action.

What this study adds:

  • An introduction to the Tornado Tales web application as a tool to measure warning reception and response during real-world events.
  • Those respondents who experienced evening storms were more likely to receive at least one warning, and were more likely to get warnings from multiple sources, than those who experienced early-morning storms.
  • Automated text alerts were the most common source of tornado warnings, followed by sirens.
  • There was little difference in warning response rates during each event among those who received warnings.
  • Moving oneself and one’s family to the most secure location in the same building were the most common responses to receiving a warning.
  • In these examples, it appears that the issue lies between warning issuance and warning reception, rather than between warning reception and protective action taking. This indicated that public complacency towards taking appropriate protective actions was not a reported issue.

 

Abstract:

Nocturnal tornadoes are extremely hazardous weather events. Despite this, little is known about how people receive and respond to tornado warnings during the overnight hours. This survey uses data from the Tornado Tales web application, a voluntary online survey of people who have experienced a tornado, tornado warning, or severe thunderstorm capable of producing a tornado. Three events were selected from this data based on timing of the events and statistical significance of the response number. Quantitative data were analyzed using the chi-square test, while qualitative data were analyzed for common themes related to the Protective Action Decision Model, milling, and aspects of transformative learning. The proportion of people receiving warnings was greater during the evening events compared to the early morning event. This may be due to a majority of people being asleep in the early morning. Out of all warnings received, automated text alerts were the most commonly reported warning source, followed closely by sirens. The automated nature of these warning sources could increase the likelihood that they are received. When warning reception is accounted for, there is little difference in the proportion of people taking protective action, suggesting that complacency was not an issue during these events. Furthermore, moving oneself and one’s family to a secure location were the most commonly reported protective actions, which indicates that the public understands proper sheltering techniques. Overall, these findings suggest that lack of warning response could be attributed to issues with warning reception, rather than issues with complacency.

Full Paper [PDF]