Assessment of ball lightning cases by correlated LLS data

Abstract

Ball lightning (BL) research lacks instrumental records and field experiments. Today, widely available networks for lightning detection offer a chance to assess alleged BL reports with located lightning and its physical parameters. This assessment was realized with BL case data and EUCLID lightning data for Germany, Austria, the Czech Republic, and Switzerland. According to the initial flash hypothesis, the electromagnetic pulse of a detected stroke near the geocoded BL position could be the triggering event of the BL process. The lightning location system (LLS) provides stroke location, time, polarity (positive/negative), and peak current in kA. In this assessment, 34 BL case reports over the period from 1994 to 2016 suggested a physical link of the stroke with maximum peak current in kA to the subsequent BL process. No electromagnetic field values for individual strokes can be computed with the kA peak current. As a first step in this analysis, geocoded distances of less than 1 km between stroke location and BL observation site were called \quotedblbase close event`` (CE) and distances 1-10 km \quotedblbase distant event`` (DE). For cloud-ground (CG) strokes the stroke-to-BL distance values were obtained. LLS-provided coordinates for intra-cloud (CC) strokes were also used to obtain a distance, although CC do not have a defined striking point. 28 BL cases occurred in the summer months, 6 cases in winter. Their time pattern follows the diurnal thunderstorm frequency in Central Europe. 19 cases were from Germany, 10 from Austria, 3 from the Czech Republic and 2 from Switzerland. 19 of the correlated EUCLID strokes were positive and 15 negative. Positive strokes ranged from 4 to 370 kA, negative from -3 to -37 kA. 28 were classified by the LLS as CG, 6 as CC strokes. The mean value of the semi-major axis of the EUCLID location confidence ellipse was 0.45 km. The BL-to- stroke distance in km for CG events had a mean of 0.42 km and a range of 0.0-0.8 km for CE cases, and a mean of 5.7 km and a range of 1.4-10 km for DE cases. The case statistics of this sample showed smaller BL-stroke distances for higher kA peak currents for both negative and positive CG events. This result leads to the interpretation that most DE strokes are uncorrelated with BL since their physical effects are marginal. Replications with more BL cases will show whether the pattern holds.