Detection of Differential Reflectivity columns using operational c-band radar and their relation to hail probability and size in Switzerland

Differential reflectivity columns (ZDRC) have been observed during and at the onset of severe convection. From modelling studies, we know that they have the potential to characterise updrafts in convection and consequently may help in the nowcasting of hail falls. Here we investigate the occurrence and characteristics of ZDRC in the northern and southern boundary regions of the Swiss alps, two regions where convective storms are particularly frequent and storm-related damages, especially the impacts of hail, are of explicit concern.

We investigate the relation between ZDRC and the probability and size of hail on the ground by combining polarimetric data from the Swiss weather radar network with a wealth of ground-based hail information from crowd-sources hail reports and a network of automatic novel hail sensors. First, several manual case studies are performed and analysed. Then, in a second step, ZDRC are characterised for the whole convective season of 2021 with the assistance of an automated ZDRC detection algorithm.

Automated ZDRC detection in the alpine region provides challenges regarding visibility and data quality. Effects such as ground clutter and the shielding of the radar beam by topography must be considered. Further, differential reflectivity (ZDR) measurements suffer from the effects of differential attenuation and artefacts, such as three-body scattering, which have to be taken into account. The Swiss radar network consists of 5 radars, of which two are located at an altitude close to 3000 m above sea level, in an arrangement which allows for most geographical locations to be in the range of multiple radars. This configuration leads to good visibility in the Alps as well as the potential to use compositing to work around attenuation effects. We derive a 3D composite of ZDR in which we identify ZDRC with an algorithm using minimum thresholds of ZDR and ZDRC height and perform further filtering of the data using reflectivity and correlation coefficient information.