Extended Abstract (1.4M)10A.4
Hail detection and quantification with a C-band polarimetric radar : challenges and promises
Pierre Tabary, Météo France, Toulouse, France; and B. Fradon, A. J. Illingworth, and G. Vulpiani
Hail detection and quantification in altitude and at ground-level is one of the often mentioned perspectives of polarimetric radars. While many works have been conducted and published at S-band, little has been done so far at C-band. We present here a thorough and objective evaluation of polarimetric hail detection techniques at C-band. The study relies on the French C-band operational Toulouse polarimetric radar and on a dense network of hailpads (more than a hundred located within 100 km of the radar). Each hailpad provides the surfacic hail stone size distribution (with a 2 mm diameter bin size and a minimum diameter of 5 mm) integrated over the entire episode. From those data, several severity indices can be computed such as the total kinetic energy, the total mass of hailstones and the maximum diameter. A first result is that the hail stone size distributions follow remarkably well an exponential law.
On the radar side, several conventional and polarimetric hail indicators are included in the comparison : 1) a simple threshold (varied between 45 and 60 dBZ) on the raw reflectivity PPIs, 2) same as 1) but with fDP –based attenuation-corrected PPIs, 3) the Probability Of Hail (POH) indicator based on the difference between the 45dBZ height and the Freezing Level Height and 4) several polarimetric indicators obtained from the outputs of a fuzzy logic classification scheme (Hail alone, Hail plus Rain and Hail, Hail plus Rain and Hail plus Graupel, …). The fuzzy logic classification scheme relies on attenuation-corrected reflectivity (ZH) and differential reflectivity (ZDR), correlation coefficient (rHV), estimated specific differential phase (KDP) and a temperature profile reconstructed from the polarimetric bright band identification algorithm. The PPI-based indicators are computed for several tilts (0.8 and 1.5°).
The main difficulties of polarimetric hail detection at C-band are clearly illustrated : Non Uniform Beam Filling, Three-Body Scattering, “Hot Spots” (unusually high differential reflectivity and differential attenuation caused by large melting hailstones), … The attenuation correction was improved to handle cases of large melting hailstones causing resonance effects and large differential attenuation. The fDP-based correction for attenuation is demonstrated to improve the performance of ZH-only-based estimators. Polarimetric estimators are shown to overperform conventional estimators most notably for echoes in the 50 – 60 dBZ reflectivity range where heavy rain can be confused with hail without the polarimetric information. The next step, currently underway, is to relate the severity of hail (revealed by the kinetic energy or maximum diameter) to the polarimetric variables.
Session 10A, Polarimetric Radar I
Thursday, 8 October 2009, 8:00 AM-10:00 AM, Auditorium
Previous paper Next paper
Recorded presentation