Objective forecast of thundery and non-thundery days using conventional indices over Bangalore during pre-monsoon season

Abstract

Prediction of thunderstorms is a challenging task for a forecaster as their spatial scale is small and lifespan is short. This study aims at evaluating the skill of stability indices with specific thresholds in relation to forecasting thundery and non-thundery days over Bangalore. A total of nine indices derived from radiosonde/radiowind data of Bangalore during 1991-2003 are treated as predictors and their statistical parameters were computed. The Z test (Zxy) was applied to find out significant predictors and the skill score method was used for finding the threshold value amongst the significant predictors. The threshold value of selected predictors is obtained by maximising normalised skill score. The results show that significant indices K (> 33 oC) and SHIm (< 3 oC) are most efficient predictors for 0000 UTC while SLI (< - 4 oC) and DCI (> 44 oC) are most efficient predictors for 1200UTC. These results were verified for pre-monsoon period of 2004-2006 over Bangalore.