If a Tornado Occurs in a Field and No One Sees it, Will it Get Recorded? Quantifying Tornado Underreporting

Not all tornadoes that occur are reported, which means tornado climatologies will inevitably be incomplete. The incomplete tornado record introduces error to research results based on climatologies. This underreporting is commonly attributed in the literature to tornadoes occurring at night or low population density. However, other environmental and social reasons for tornado underreporting exist. For example, meteorologists in communist Romania (1947—1989) did not acknowledge, despite historical observations, that tornadoes occurred in their country. Instead, windstorms were blamed for tornado damage. Research into tornado underreporting helps in two ways. First, understanding causes for underreporting helps put tornado climatologies into context to recognize, for example, that a multi-year minimum in tornadoes is due to war, not a meteorological cycle of tornado activity. Second, developing a framework for underreporting can lead to measures to quantify the degree of underreporting. By providing reliable estimates for underreporting, future research using tornado climatologies can quantify error in the analysis. Additionally, policy makers, insurance companies, and society can have a better idea of how many tornadoes actually occur worldwide to better understand and prepare for tornado impacts. In this presentation, we propose a framework for estimating the degree of tornado underreporting. The framework for tornado underreporting presented here has three parts: the probability of the tornado being seen, the probability of the seen tornado being reported, and the probability of the tornado report being archived. The multiplicative probability can be used to estimate how many tornadoes go unreported in a given location. Finally, a demonstration of the tornado underreporting framework is provided using Romania as an example.