Objective forecast of thundery/nonthundery days using conventional indices over three northeast Indian stations

Abstract

The preconvective atmosphere of the four months (March, April, May and June) for the year 1980-84 have been studied using the Radiosonde data (RS/RW) of the three stations of northeast India namely Guwahati, Dibrugarh and Agartala. The objective is to identify a suitable index for these stations with a prescribed threshold value that can be used as the predictor for forecasting thundery/nonthundery days. Total eleven indices are considered in this study. The thermodynamic and kinematic parameters required for calculation of the indices are derived from the RS/RW data of 0000 and 1200 UTC and taking three stations together. Probability distribution curves are plotted to make a qualitative comparison on the ability of the indices to differentiate the thundery and nonthundery days. Mean and standard deviations are calculated and these are used to apply Z-statistics. The indices that have shown Z values at 90% or more significant level are considered for further analysis and rest are rejected. Threshold values of the selected indices are assigned by an iterative process where the mean of value of the index is used as the first guess and skill scores are calculated. Next the first guess value is changed and the skills are recalculated. The process is continued till the best score is attained for each selected index. The threshold values of these indices are used to predict the thundery and nonthundery days of the verification year 1985. It is found that HI has the best skill in forecasting thundery/nonthundery days at 0000 and 1200 UTC for Guwahati, Dibrugarh and Agartala stations.