Most hail detection algorithms for single-polarization radar are based on the analysis of the vertical profiles of reflectivity. These profiles are estimated using the volume data generated by a scanning at multiple elevations. Beside the usual sources of errors affecting the radar reflectivity data, the estimated vertical profiles are also affected by the uncertainties in the trajectories of the radar beams. This results in uncertainties in the height assignment of the measured reflectivities, and thus introduces errors in the measured vertical profiles.

To get a first idea of the effect of these errors on the estimated probability of hail, we have compared the results obtained from two radars for a number of thunderstorm cases with hail observed in Belgium and The Netherlands. The data are provided by the radar of Wideumont located in the south of Belgium (operated by RMI) and the radar of De Bilt in The Netherlands (operated by KNMI). First, the probability of hail is estimated from both datasets using the modified method of Waldvogel which is operationally used at KNMI. This method is based on the difference between the 45 dBZ echotop and the height of the freezing level. Then, the differences resulting from height assignment errors will be highlighted by a comparison of range height indicators of reflectivity factor from the two radars.