Forecasting peak surface gust wind in association with thunderstorm activity during pre-monsoon season at Delhi

Abstract

In this study, an attempt is made to develop an objective method for forecasting the direction and speed of the gusty winds associated with thunderstorms at Delhi. For this purpose, surface and upper-air data for April, May and June (AMJ) for the years 1985-90 are utilized. Multiple regression equations are developed for forecasting the direction and speed of the gusty winds, using stepwise screening method, for which a total of 181 potential predictors are utilized. The developed dynamical-statistical models are tested with independent data sets of 1994 and 1995 for April, May and June. The dynamical-statistical models give satisfactory results with the developmental as well as the independent data sets. The root mean square error of the direction vary between 58° and 77° and the speed forecast vary between 9 and 12 knots. Possible reasons for large deviations of the forecast, noticed on a very few occasions, have also been examined.