Markov chain models for pre-monsoon season thunderstorms over Pune

The probabilistic distribution of the thunderstorm phenomenon during the pre-monsoon season (1 March to 18 June) over Pune, a tropical Indian station, has been examined with the help of Markov chain models using daily thunderstorm data for a period of 11 years (1970-80). The data have also been tested using Akaike's information criterion. This test has clearly indicated that the first-order Markov chain model is the best fit model for thunderstorm forecasting, which has described the appropriate period (8 days) of occurrence of thunderstorm phenomenon over Pune. Further, the steady-state probabilities and mean recurrence time of thunderstorm days and non-thunderstorm days have also been calculated for the first- and second-order Markov chain models. These computations have revealed that the observed and theoretical values of steady-state probabilities are realistically matched. Copyright (C) 2002 Royal Meteorological Society.