Estimating the occurrence probability of Spring Thunderstorms using Markov chain, Case Study: Tabriz

Thunderstorms are one of the climatic phenomena that cause numerous damages in different parts of the world, due to the accompaniment with thunder, high winds, hail and heavy precipitation. In this study, the probability of thunderstorm days in Tabriz in the spring is analyzed using probabilistic rules and Markov chain model. For this purpose, the daily data related to thunderstorms (codes 95 to 99) was used for 65 years (1951-2015). At first, the daily data is classified into the normal days (code 0) and thunderstorm days (code 1). Then the frequency matrix is formed and the probability matrix is created accordingly based on maximum likelihood method. The Markov chain properties such as empirical probability and equilibrium probability, Frequency of occurrence, mean time periods and weather cycle were investigated. Finally, Occurrence Probability and return period of these spells were determined. The results show that the shortest weather cycle is in May, which has the highest frequency of thunderstorms. Conversely, the longest weather cycle is in April, which has the lowest frequency of thunderstorms. Also, in 65 years period, the one-day and two-day stormy sequences have the highest frequency. And For longer sequences, the frequency of thunderstorm days is reduced. The return period of one-day and two-day stormy spells is 1.5 and 5 days. Regression relation between the observed and estimated values of n-step periods of thunderstorm days shows that the considered accuracy and reliability for all months is more than 99%.