Hail detection with a C-Band dual Polarization radar in the Canadian Great Lakes region

Two methods of hail detection were investigated using data from the operational King City C-band dual polarized radar. The first algorithm estimates hail location and size, using volume scan data to examine storm structure, and is based on vertical integration of the cell's reflectivity profile (VIL) and the height of the 45dBZ level above the freezing level. The second is a polarimetric technique employing a fuzzy logic for hydrometeor classification. The particle classification algorithm (“iParCA”) produced polarimetric hail detections, which were measured directly from 0.5^o elevation scans. The data set consists of 74 storm cells for 21 days during the summers of 2005-2008. Observer reports of hail location and time, and the Buffalo NEXRAD hail algorithm output was used to supplement our verifications.

The standard skill score measures, critical success index (CSI) and probability of detection (POD), were better for the polarimetric detections. They were 88% and 94% respectively compared to 73% and 76% for the VIL method of detection. Next, 30 of the 74 cells were investigated in greater detail. The iParCA detections were determined subjectively to be better in 70% of these 30 cases when compared with the VIL algorithm. Specifically hail location, timing and extent were superior with iParCA. The VIL algorithm suffers from signal attenuation, the inability to discriminate between higher reflectivity rain and high reflectivity hail, and volume scan limitations on storm cell near the radar. The relative performance was seen to differ between multi-cellular storms and super-cells.