ST-AR (STorm-ARchive): A project developed to assess the ground effects of severe convective storms in the Po Valley

An analysis of remote sensing data related to convective precipitation events in Northern Italy is presented in order to evaluate the ability of the logistic regression models to calculate the probability of particular ground effects as hail, wind damage or small floods. In order to carry out and test the method, a database containing both remote sensing data and ground effects reports about convective thunderstorms in 2003 and 2004 summers was developed. Numerous parameters were extracted from satellite, radar and lighting location systems data of the 2003 sample and were used to determine probability models. The sample of 2004 data was used for a preliminary test of the method. The method of logistic regression seems to be a useful tool to evaluate a binary response (presence or absence of a particular effect) on the basis of continuous or binary variables. The calculated indexes of success are also presented and analysed. The best results are obtained with the model to establish the presence of wind damage and the one for determining the absence of important ground effects. All the models present moreover the advantage of having low false alarm rates. The obtained results encourage the use of this statistical technique for further statistical analysis on larger data set.