The very short forecasting or nowcasting of convective storms is a highly challenging problem due to the wide variety of spatial scales and processes involved. The state of the art numerical weather prediction models generally fail in predicting convective storms initiation and development with reasonable accuracy. Weather radars are intensively used by forecasters during convective situation. A radar provides instantaneous reflectivity measurements of precipitation at a spatial scale of about 1km. Convective storms, which occur at a spatial scale of a few kilometers, can then be detected by the radar. Moreover, if the time resolution is short enough, the convective storm evolution can be monitored across successive radar images. Based on those snapshots, the motion of a convective cell can be estimated and used to forecast its future position. However, this simple method does not take into account storm growth, decay and direction changes. For this purpose, a better knowledge of storm evolution characteristic is required. The Royal Meteorological Institute of Belgium has several years of archived volume data from a C-band radar. To process those data, the cell tracker TITAN (NCAR) has recently been installed at RMI. The behavior of this tracking system has been investigated through a sensitivity study to some parameters. The storm tracks provided by TITAN have been analysed by a suitable statistic tool. The interest of this analysis to obtain a robust climatology of convective storms is clear (Han, 2009 ; Tucker and Li, 2009). Furthermore it is worth investigating whether the analysis of convective storm characteristics can be used to improve its nowcasting. In this study, storm tracks from six years of volume data are analysed with different statistical methods. The frequency distribution of the storm tracks properties and their space-time variations are showed. Pairs of storm track properties are tested for a possible linear relation. The characteristics of the kinematic of the storms such as the storm initiation location are also analysed using spatial point pattern statistics.