Performance Characteristics of Distinct Lightning Detection Networks Covering Belgium

Abstract This study reports results from electric field measurements coupled to high-speed camera observations of cloud-to-ground lightning to test the performance of lightning location networks in terms of its detection efficiency and location accuracy. The measurements were carried out in August 2011 in Belgium, during which 57 negative cloud-to-ground flashes, with a total of 210 strokes, were recorded. One of these flashes was followed by a continuing current of over 1 s—one of the longest ever observed in natural negative cloud-to-ground lightning. Lightning data gathered from the lightning detection network operated by the Royal Meteorological Institute of Belgium [consisting of a network employing solely Surveillance et Alerte Foudre par Interférométrie Radioélectrique (SAFIR) sensors and a network combining SAFIR and LS sensors], the European Cooperation for Lightning Detection (EUCLID), Vaisala’s Global Lightning Detection network GLD360, and the Met Office’s long-range Arrival Time Difference network (ATDnet) are evaluated against this ground-truth dataset. It is found that all networks are capable of detecting over 90% of the observed flashes, but a larger spread is observed at the level of the individual strokes. The median location accuracy varies between 0.6 and 1 km, except for the SAFIR network, locating the ground contacts with 6.1-km median accuracy. The same holds for the reported peak currents, where a good correlation is found among the networks that provide peak current estimates, apart from the SAFIR network being off by a factor of 3.