A Markov chain approach in the prediction of severe pre-monsoon thunderstorms through artificial neural network with daily total ozone as predictor: XXXth URSI general assembly and scientific symposium to be held in Istanbul, Turkey, August 13–20, 2011

Purpose of the present paper is to examine the predictability of the occurrence of the severe pre-monsoon thunderstorm over Gangetic West Bengal. Instead of considering various meteorological predictors, the daily total ozone concentration is chosen as the predictor because of the influence of tropospheric as well as stratospheric ozone on the genesis of meteorological phenomena. Considering the occurrence/non-occurrence of thunderstorm in the pre-monsoon season (March-May) of the year 2005 as the dichotomous time series{X
_t
} that realizes 0 and 1 for non-occurrence and occurrence of TS respectively, a first order two state (FOTS) Markov dependence is revealed within this time series.