The evaluation of tornado hazards at nuclear power plants uses past data in the area to estimate
the frequency of a tornado affecting the site. One of the correction factors that must be included
in this calculation is related to misclassification of tornados. One commonly used approach for
this misclassification factor assigns similar apportionment factors to under-classification and
over-classification. Recent tornado data from the National Oceanic and Atmospheric Association
(NOAA) indicates that fewer EF5 tornados have occurred than would be expected from previous
trends. This data could be indicative of a true shift in the tornado frequency or may indicate an
increase in under-classification of tornados, especially in the higher damage categories. This EF5
“drought” has resulted in no EF5 tornados from June 2013 through May 2025, with the drought
finally breaking in June 2025 with one EF5 tornado; this was the longest gap of F5/EF5 tornados
estimated since 1950 (with the second longest being an 8-year gap from June 1999 through May
2007). Probabilistic analyses of tornado hazards should address this observation as a source of
uncertainty.

This paper seeks to characterize the potential impact via at least two sensitivity studies: (1)
assuming the trend is correct, how would a representative tornado hazard curve shift, given the
new data? (2) Exploring alternative factors for under- and over-classification that more accurately
represent recent trends in classification. The results will be focused on the potential impact on
risk-informed applications.