Good results on rain pattern tracking are achieved by merging the information extracted from a weather radar data sequence, using both a correlation and shape analysis approach. The two methods, namely the tracking radar echoes by correlation corrected by continuity (COTREC) and the modal matching shape representation are both applied to the problem. This choice allows us to exploit the advantages of both a global and a more local approach to a motion-based interpretation of rain events. By combining the two methods, we may work on a global frame analysis based on correlation matching and add a local analysis of the shape of the cells to find a more refined motion vector field. Some results are shown, extracted from data sequences recorded by a C-band weather radar in Northern Italy.