Abstract Two winter snowstorms in the Toronto area are investigated with an X-band Doppler radar. The scan strategy involved alternating volume scans and a vertically pointing mode. The Extended Velocity Azimuth Display (EVAD) technique was applied to the volume scan data to document mesoscale wind patterns. Moments of the Doppler power spectrum, namely the fall velocity ( V t) and the spectral width (σ w), were derived from the vertically pointing data. The variation of these moments in height and with time was used to deduce precipitation formation processes. In both cases low level convergence in the easterly flow across Lake Ontario ahead of the main disturbance was found. Here, accretion was very active. In the first case, strong convergence near the centre of the low as it passed overhead the radar was discerned. In this part of the storm, aggregation was the dominant precipitation mechanism. To the rear of the low, with neither an off-lake component to the flow nor consequential convergence aloft, diffusion increased in significance. In the second case, warm air overrunning an approaching warm front produced strong mid-level convergence. This resulted in significant accretion in the mid-levels such that a period of ice pellets was observed without the benefit of an above 0°C layer aloft. Where convergence was weaker, aggregation became the dominant process. The approach used in this study enabled variations in the microphysics on the mesoscale to be deduced. The EVAD technique in the circumstances of almost complete radar coverage and slowly evolving fields was able to distinguish between low level local effects and synoptic scale aspects of the storms. Furthermore, the moments of the Doppler power were able to identify the occurrence of diffusion, aggregation, and accretion. Regions of the storm in which V t > 1.75 m s −1or σ w > 0.60 m s −1 were associated with the accretion process. V t > 1.25 m s −1 was associated with aggregation; V t < 1.00 m s −1 was associated with diffusion. Spectral width was not generally a good discriminator between diffusion and aggregation. When 1.00 V t 1.25 m s −1, however, values of σ w above 0.40 m s −1 were more commonly associated with aggregation than diffusion. The profiles of V t and σ w within the different growth regimes give a more complete description of the precipitation process and can provide the basis of an improved understanding of the hydrometeor size spectrum.