Hail detection during the Joint POLarization Experiment (JPOLE)

One objective of the National Severe Storm Laboratory’s (NSSL) Joint POLarization Experiment (JPOLE) is to evaluate the capability of classification of meteorological and nonmeteorological scatterers using three polarimetric variables: radar reflectivity factor Z, cross-correlation coefficient r_HV, and differential reflectivity Z_DR. This paper investigates an important aspect of this objective: assessing how well the detection of hail can be made without linear depolarization ratio LDR, the variable that was historically the first polarimetric parameter utilized for such purposes. The accuracy and skill of two NSSL prototype hydrometeor classification algorithms (HCAs) are assessed by first comparing HCA’s output to ground truth observations from five hail storms, and then computing several contingency table statistics. These algorithms include the version used during JPOLE and a recently upgraded version. Results demonstrate that accurate and skillful hail detection is possible without including LDR. The upgraded version of HCA shows improvement in measures of accuracy and skill compared to the version used during JPOLE. For comparison purposes, contingency table statistics are computed also for previously designed polarimetric hydrometeor classification algorithms. Results will be reported in the extended abstract.