STM-D-0487Paper2018Published and peer-reviewed

Assessment of ball lightning cases by correlated LLS data

Alexander G. Keul · Gerhard Diendorfer

Summary and citation · read the original at the source

In one page

Ball lightning has almost no instrument record. It arrives without warning, lasts seconds, and is gone before anything can be pointed at it. Alexander Keul and Gerhard Diendorfer found a way to put an instrument behind the eyewitness after the fact. Europe is blanketed by EUCLID, a lightning-detection network that logs every stroke’s position, time, polarity and peak current. Take a ball lightning report with a known place and time, open the network’s records for that spot and that minute, and see what ordinary lightning was doing. They did it for thirty-four cases from Germany, Austria, the Czech Republic and Switzerland between 1994 and 2016. Nineteen of the matched strokes were positive and fifteen negative; the positive ones ran from four to three hundred and seventy kiloamperes. The pattern is the interesting part — the larger the peak current, the closer the ball appeared to the stroke. That is what a real trigger should look like, and it is the first statistical handle on the moment these objects are born.

Why it matters hereChapter 9 needs ball lightning treated as an instrumented phenomenon rather than a story, and this is the paper that first attaches an existing continental measurement network to the case record. Chapter 1 gets a clean rung on its ladder from it: the authors state exactly what the network can and cannot deliver, and end with the replication that would confirm the pattern — which they then ran.

What it claims

01Thirty-four ball lightning case reports from the period 1994 to 2016 were matched against EUCLID lightning-location records — nineteen from Germany, ten from Austria, three from the Czech Republic and two from Switzerland. Twenty-eight occurred in the summer months and six in winter, and their time-of-day pattern follows the diurnal thunderstorm frequency of Central Europe. The seasonal and daily pattern of the reports is the pattern of thunderstorms themselves.Abstract, ICLP 2018, sentences on case distribution
Published and peer-reviewed
02Of the correlated EUCLID strokes, nineteen were positive and fifteen negative. The positive strokes ranged from four to three hundred and seventy kiloamperes, the negative from minus three to minus thirty-seven kiloamperes. Twenty-eight were classified by the network as cloud-to-ground and six as intra-cloud. Positive polarity in a majority of matches is notable in itself, because positive strokes are the minority of ordinary lightning.Abstract, ICLP 2018, sentences on polarity and peak current
Published and peer-reviewed
03For the close events — those with the ball reported less than one kilometre from the located stroke — the mean stroke-to-ball distance was 0.42 kilometres with a range of zero to 0.8 kilometres. For the distant events, one to ten kilometres, the mean was 5.7 kilometres with a range of 1.4 to ten. The network’s own positional uncertainty is stated alongside: the mean semi-major axis of the EUCLID location confidence ellipse was 0.45 kilometres, which is the same size as the close-event distances themselves.Abstract, ICLP 2018, sentences on distance statistics and the confidence ellipse
Published and peer-reviewed
04The load-bearing result is a correlation: across this sample, higher peak currents go with smaller ball-to-stroke distances, for both negative and positive cloud-to-ground events. The authors read that as a physical link between the stroke of maximum peak current and the ball lightning process that follows it, and as a reason to treat the distant matches as unlikely triggers, since a stroke’s physical effects at several kilometres are marginal.Abstract, ICLP 2018, closing three sentences
Published and peer-reviewed
05The method is stated with its own limits attached. The working idea is the initial flash hypothesis — that the electromagnetic pulse of a detected stroke near the reported position could be the triggering event. A lightning location system supplies stroke position, time, polarity and peak current, and no more: the authors state plainly that no electromagnetic field values for individual strokes can be computed from a peak current in kiloamperes, and that intra-cloud strokes have no defined striking point, so the distance derived for those six cases is a different quantity from the twenty-eight cloud-to-ground ones.Abstract, ICLP 2018, method sentences
Published and peer-reviewed
06What to watch: the paper ends by saying that replications with more cases will show whether the pattern holds — and the same authors ran it. Keul, Pichler and Diendorfer reported a follow-up at the 2024 conference with fifty-eight European cases from 1993 to 2023, searching the records sixty minutes either side of each report within a ten-kilometre radius. Fifty-three per cent of determinable distances came in under one kilometre, eighty-three per cent of the negative trigger strokes were under minus thirty kiloamperes, and seventy-one per cent of the positive ones were over fifty kiloamperes, with forty-three per cent over one hundred. Thirty-five per cent of the triggers were isolated strokes.Abstract, ICLP 2018, final sentence; ALDIS entry for Keul, Pichler and Diendorfer, ICLP 2024
What to watch

The way in

https://doi.org/10.1109/iclp.2018.8503422WHAT THIS PAGE IS WRITTEN FROM. Presented at the 34th International Conference on Lightning Protection, Rzeszów, Poland, 2 to 7 September 2018, and published in the proceedings as a six-page paper. IEEE holds the copyright and IEEE Xplore declines automated retrieval; Unpaywall, OpenAlex and Semantic Scholar all report no open copy, the ResearchGate entry is a request-only page, and the ALDIS publications list, which carries the paper, offers no download for it. The paper itself could therefore not be read for this sheet, and none of it is reproduced here. Everything below is written from two things that could be read on 2026-09-08: the authors’ published abstract, which is long and carries the whole result set with its numbers, and the ALDIS entry for the authors’ own follow-up study — Keul, Pichler and Diendorfer, Lightning location system strokes as ball lightning triggers, a follow-up study, 37th International Conference on Lightning Protection, Dresden, 2024 — which restates the 2018 finding and extends it. Locators say which of the two a number comes from. When the six pages themselves can be read, this sheet should be rewritten from them. EUCLID is the European Cooperation for Lightning Detection network; ALDIS, the Austrian Lightning Detection and Information System, is Diendorfer’s institution and one of its nodes.

How to cite it

Alexander G. Keul, Gerhard Diendorfer (2018) Assessment of ball lightning cases by correlated LLS data. doi:10.1109/iclp.2018.8503422

Where it sits in the curriculum

Plasmoids, charge clusters and the orbs The evidence ladder