Most hail detection algorithms for single-polarization radars are based on the analysis of the vertical pro les of reectivity. These pro les are estimated ectivity. These pro les are estimated from the volumic data generated by a scanning at multiple elevations. Beside the usual sources of errors a ecting the radar reectivity, the estimated vertical pro les are also a ected by the uncertainties in the trajectories of the radar beams. This results in uncertainties in the height assignment of the measured reectivities, and thus introduces errors in the measured vertical pro les. To get a rst idea of the e ect of these errors on the estimated probability of hail (POH), RMI (Belgium) and KNMI (The Netherlands) have started a study in which the volumic data from the radars of Wideumont and De Bilt are compared, as well as the derived probability of hail. Both radars are Gematronik C-band Doppler radars. They perform a volumic scan every 15 minutes. It includes 10 elevations between 0.5 and 17.5 degrees for the radar of Wideumont and 15 elevations between 0.3 and 12 degrees for the radar of De Bilt. The distance between the two radars is 244 km. We here present rst results concerning thunderstorm cases with hail observed between the two radars in the summer 2002.