Probabilities of Daily Hail and Thunder Occurrences

Abstract

The nonhomogeneous Poisson distribution is applied to the probabilities of the number and dates of days experiencing thunder and had in southern Saskatchewan, Canada. The seasonal nature of hail and thunder occurrence is adequately represented by the Gaussian distribution. Use of the resultant simple three-parameter model provides a theoretically sound and flexible technique by which accurate predictions of the probabilities of events in any time period of interest may be calculated without recourse to the reanalysis of the original data. The success of the methodology in modeling the event frequencies both over, and within, traditional monthly time periods and other arbitrarily selected periods reflect the general applicability of the technique.