Improving hail identification in the Ebro Valley region using radar observations: Probability equations and warning thresholds Manuel Ceperuelo Mallafré a, ⁎, Tomeu Rigo Ribas b , Maria del Carmen Llasat Botija a , José Luis Sánchez c a Department of Astronomy & Meteorology, Faculty of Physics, University of Barcelona, Barcelona, Spain b Meteorological Service of Catalonia, Barcelona, Spain c Laboratory for Atmofpheric Physics, University of León, León, Spain article info abstract Article history: Received 1 April 2008 Received in revised form 8 August 2008 Accepted 24 September 2008 In order to identify hail into thunderstorms identified with radar data, different kinds of techniques are use based. This paper will evaluate some of these methodologies for the case of the Ebro valley (NE Spain), in order to obtain the best method to identify hail at surface. To achieve this end, an analysis of the 2004 and 2005 hail seasons has been undertaken using C-band radar, MM5 meteorological model outputs and ground observations provided by two hailpad networks. These data were integrated, identifying, characterizing and tracking the convective cells, and obtaining for each one different hail probability equations by means of various radar techniques. Kinetic energy flux was found to be the best parameter for distinguishing between hail and no-hail precipitation, although there was found to exist no significant difference between the various methods used. Moreover, the high correlations between radar parameters obtained by means of cell analyses led us to reduce the initial number of variables in new radar parameters. These new variables are defined and provide new improved models of the intensity of the storm. © 2008 Published by Elsevier B.V. Keywords: Hail Radar Convective cell Probability of hail 1. Introduction In order to avoid casualties and high damages associated with hail storms, there exist advanced warning systems that identify hail precipitation and help risk management. Different kinds of techniques are used for this purpose. One of the most common methods consists in finding relation- ships between environmental conditions and radar observa- tions (Stumpf et al., 2004), with the objective of identifying hail at the surface. Other common methods are: the use of radar data combined with radiosonde observations to find the relationship between environmental conditions, convective cells and hail precipitation (Edwards and Thompson, 1998; Waldvogel et al., 1979); the hail detection algorithm, which produces estimations of the probability of hail (any size), probability of severe-size hail (diameter ≥ 19 mm), and max- imum expected hail size for each detected storm cell (Witt et al., 1998); or, the use of logistic functions aimed at minimising the false-alarm ratio of hail precipitation (Billet et al., 1997). Finally, it is necessary to cite that the use of combinations of different hail indicators has allowed an improvement over previous methodologies (Kessinger and Brandes, 1995). In summary, most of those algorithms try to identify hail by means of the echoes observed in the radar images, searching for patterns where strong vertical updrafts are produced (intense radar echoes at high levels of altitude). The cost of the damages produced by hail in the agrarian production of Spain is between 650 and 700 million Euros. In particular, some previous studies showed how the Ebro valley region (in the NE part of Spain) presents the highest frequency of hailstorm (Font, 1983; Pascual, 2002), with 9.4 Atmospheric Research 93 (2009) 474–482 ⁎ Corresponding author. Meteosim SL, Barcelona Science Park, Baldiri Reixac 10-12, 08028 Barcelona, Spain. Tel.: +34 934487265; fax: +34 934490010. E-mail addresses: mceperuelo@meteosim.com (M.C. Mallafré), tomeur@meteocat.com (T.R. Ribas), carmell@am.ub.es (M. del Carmen Llasat Botija), jl.sanchez@unileon.es (J.L. Sánchez). 0169-8095/$ – see front matter © 2008 Published by Elsevier B.V. doi:10.1016/j.atmosres.2008.09.039 Contents lists available at ScienceDirect Atmospheric Research journal homepage: www.elsevier.com/locate/atmos