Hail detection using a polarimetric algorithm at C band: impact on attenuation correction

Discrimination between rain and hail radar echoes is a long-standing objective for meteorological, hydrological, aviation and agricultural applications. Hail detection using single polarized radar began in the late 1950s with techniques based on reflectivity measurements considering the echo intensity, its structure and time evolution within the hailstorm (Cook, 1958; Douglas and Hitschfeld, 1958). With the development of radar polarimetry the differential reflectivity Zdr became the key radar parameter for hail detection (Aydin et al. 1984; Bringi et al., 1984). Since the late 1990s, fully polarimetric algorithms based on the Fuzzy Logic have been proposed for S-band radar classification of hydrometeors (Vivekanandan et al. 1999; Zrnić et al., 2001). Recent studies have confirmed the higher performance of Fuzzy Logic algorithms with respect to methodologies employing radar reflectivity only for the diagnosis of hail using radar measurements at S band (Heinselmanand Ryzhkov, 2006). Almost all European weather radars operate at C band, mainly due to cost and size constraints (e.g., Alberoni et al., 2002). However, at frequencies higher than S band, path attenuation effects due to rainfall can be significant and need to be compensated for any quantitative applications (Vulpiani et al. 2005). Robust polarimetric algorithms for rain path attenuation do exist (Bringi et al., 1990; Testud et al., 2000; Bringi et al. 2001); nevertheless mixed precipitation might cause large attenuation that traditional methodologies are unable to deal with. An iterative correction procedure based on preliminary classification of the observed hydrometeor types is applied to solve this problem. It relies on a long-term statistical characterization of the relationship between attenuation and differential phase in mixed phase (rain / hail) precipitation. Fuzzy logic based algorithms are fairly established approaches for the discrimination of hydrometeors because of the partial overlapping of radar signatures. A fully polarimetric version of the classification technique proposed by Marzano et al. (2006) is applied in the present work. The proposed correction methodology, applied to an intense hailstorm observed in the Paris area, has shown a remarkable reduction of attenuation effects. In situ hail observations are used to check the effectiveness of the hail detection algorithm.