Identification of ZDR columns for early detection of severe convection in southern England

Various studies in the UK, Great Plains and Southeastern USA have highlighted the presence of certain radar signatures prior to the onset of or during severe convection. One type of such radar signature is a differential reflectivity (Z_DR) column, which is defined as a vertical columnar region of enhanced Z_DR that extends above the freezing level. Several field campaigns synthesising radar and in-situ measurements confirmed that such columns contain large supercooled millimetre-sized droplets lofted into convective storms and are in, or near strong updrafts. Recent work using a single research radar in Oklahoma also exploited the usefulness of detecting Z_DR columns for informing nowcasters of severe convection.

The goal of this study is to identify potential severe convective events in the UK mostly for cases over the summer season using polarimetric radar measurements. The UK Met Office has fully upgraded all 18 C-band radars since January 2018 with full dual-polarisation operational capability. From this network, we derive a 3D radar composite, which provides large coverage on the order of 1000 km for monitoring potentially hazardous weather. Environmental conditions are also investigated prior to and during the onset of convection to understand the effectiveness in Z_DR columns as precursors of severe convection depending on synoptic regime.

Using past cases, we track storm cells using maximum reflectivity in the column and identify whether the cells contain Z_DR columns, where a Z_DR column is identified based on a 3D volume thresholded by reflectivity (Z_H) and Z_DR. For nowcasting of severe storms, with Z_H > 50 dBZ, we find optimal Z_H and Z_DR thresholds of around 30 dBZ and 2.0 dB respectively existing within Z_DR columns. This agrees with past literature and physical understanding indicating a low concentration of large super-cooled water droplets within Z_DR columns explained by condensation-coalescence processes, especially during early stages of convective development. In contrast, other works may show Z_DR columns associated with areas of high Z_H, suggesting detection of such columns in more mature stages of a storm. Algorithm performance in identifying Z_DR columns for early detection of severe convection and its optimal parameters vary with synoptic regime.