Exploring the Overlap: Comparing STORGIS and ONETOR data between 1950 and 1989

Abstract

Abstract The current understanding of tornado climatology centers on the Storm Prediction Center's tornado database (ONETOR) which dates back to 1950. To understand tornado climatology before this date, a secondary database (STORGIS) of digitized tornado records over the period 1880–1989 are used. Here, the ONETOR database and STORGIS data set are compared for individual tornadoes and outbreaks with six or more F2+ tornadoes during the overlapping years of 1950 to 1989. For the study period, there are more individual tornado reports in the ONETOR database relative to the STORGIS data set. While both databases suggest a similar broad-scale understanding of tornado climatology spatially and by decade, month, and year, there is a statistically significant difference between the two databases with regards to the total number of reports over the study period. For the study period, there are more tornado outbreaks in the ONETOR database relative to the STORGIS data set. In total, more than 94% of the missing outbreaks in the STORGIS data set are weaker outbreaks with ten or less tornadoes. While the general spatial and temporal patterns of the number and size of tornado outbreaks is similar between the STORGIS data set and ONETOR database, there is a statistically significant difference in the total number and average number of tornado outbreaks between each database. These results herein indicate that while the STORGIS data set is representative of large-scale patterns of F2+ tornado behavior in the United States, it cannot be used synonymously with the ONETOR database without additional statistical methods or context.