ORIGINAL PAPER A composite stability index for dichotomous forecast of thunderstorms Sutapa Chaudhuri & Anirban Middey Received: 3 December 2011 / Accepted: 20 March 2012 / Published online: 5 April 2012 # Springer-Verlag 2012 Abstract Thunderstorms are the perennial feature of Kol- kata (22° 32′ N, 88° 20′ E), India during the premonsoon season (April–May). Precise forecast of these thunderstorms is essential to mitigate the associated catastrophe due to lightning flashes, strong wind gusts, torrential rain, and occasional hail and tornadoes. The present research provides a composite stability index for forecasting thunderstorms. The forecast quality detection parameters are computed with the available indices during the period from 1997 to 2006 to select the most relevant indices with threshold ranges for the prevalence of such thunderstorms. The analyses reveal that the lifted index (LI) within the range of -5 to -12 °C, convective inhibition energy (CIN) within the range of 0– 150 J/kg and convective available potential energy (CAPE) within the ranges of 2,000 to 7,000 J/kg are the most pertinent indices for the prevalence thunderstorms over Kolkata during the premonsoon season. A composite stabil- ity index, thunderstorm prediction index (TPI) is formulated with LI, CIN, and CAPE. The statistical skill score analyses show that the accuracy in forecasting such thunderstorms with TPI is 99.67 % with lead time less than 12 h during training the index whereas the accuracies are 89.64 % with LI, 60 % with CIN and 49.8 % with CAPE. The perfor- mance diagram supports that TPI has better forecast skill than its individual components. The forecast with TPI is validated with the observation of the India Meteorological Department during the period from 2007 to 2009. The real- time forecast of thunderstorms with TPI is provided for the year 2010. 1 Introduction Forecasting weather hazards is the prime concern of the operational meteorological centers around the globe. India is no exception. Severe thunderstorms are momentous weather phenomena which impact on various facets of national activities (e.g., civil and defense operations, particularly space vehicle launching, agriculture, etc.). Additionally, there may be damage potential to life and property on the ground and aviation aloft. Thunderstorms, in general, connote a perennial feature and occur over different parts of India during different seasons. However, the thunderstorms occurring over Kol- kata (22° 32′ N, 88° 20′ E) during the transition period of premonsoon season (April–May) are the most devastating when accompanied by lightning flashes, large hail, torrential rainfall, high wind gusts, and occasional tornadoes. These storms are usually termed as Nor ’westers because they mostly approach to the station from the northwesterly direc- tion (IMD 1941). The societal and economic impact by thunderstorms made accurate prediction of the weather nec- essary. The inadequate network of the observation systems made it a great challenge for the professional meteorologists of India to develop a forecast tool for thunderstorms. Thunderstorms are cloud scale phenomena but generated due to mesoscale instabilities which are embedded on synoptic scale environment. Formation of thunderstorms is thus, an interaction among the conditions on different scales (Doswell 1987). India is situated within 8° to 37°N latitudes. The tropical easterlies and the midlatitude westerlies interact and there- fore, influence the weather systems over the subcontinent. During the premonsoon period (April–May), there is no major synoptic system like, fronts, jets, special distribution of convergence field and wind shear over the land. Heat S. Chaudhuri (*) : A. Middey Department of Atmospheric Science, University of Calcutta, Kolkata 700-019, India e-mail: chaudhuri_sutapa@yahoo.com Theor Appl Climatol (2012) 110:457–469 DOI 10.1007/s00704-012-0640-z