Assessing a Tornado Climatology from Global Tornado Intensity Distributions

Abstract

Recent work demonstrated that the shape of tornado intensity distributions from various regions worldwide is well described by Weibull functions. This statistical modeling revealed a strong correlation between the fit parameters c for shape and b for scale regardless of the data source. In the present work it is shown that the quality of the Weibull fits is optimized if only tornado reports of F1 and higher intensity are used and that the c?b correlation does indeed reflect a universal feature of the observed tornado intensity distributions. For regions with likely supercell tornado dominance, this feature is the number ratio of F4 to F3 tornado reports R(F4/F3) = 0.238. The c?b diagram for the Weibull shape and scale parameters is used as a climatological chart, which allows different types of tornado climatology to be distinguished, presumably arising from supercell versus nonsupercell tornadogenesis. Assuming temporal invariance of the climatology and using a detection efficiency function for tornado observations, a stationary climatological probability distribution from large tornado records (U.S. decadal data 1950?99) is extracted. This can be used for risk assessment, comparative studies on tornado intensity distributions worldwide, and estimates of the degree of underreporting for areas with poor databases. For the 1990s U.S. data, a likely tornado underreporting of the weak events (F0, F1) by a factor of 2 can be diagnosed, as well as asymptotic climatological c,b values of c = 1.79 and b = 2.13, to which a convergence in the 1950?99 U.S. decadal data is verified.