MYS0D40A0KP2QRS·Mystery · sighting

Ball lightning

Unknown author · Public domain

What: Luminous ball seen during thunderstorms, lasting seconds⁠[2]
Published reports: About 5,000⁠[1]
Frequency: Reported in a 1966 survey nearly as often as nearby cloud-to-ground strokes⁠[2]
Instrumental record: One natural event: Qinghai Plateau, China, July 2012, two slitless spectrographs at 0.9 km⁠[5]
Spectrum: Lines of silicon, iron and calcium throughout its life⁠[4]
Optical power: About 4.2 kW (400–690 nm) for most of its life⁠[6]
Status: Partly explained
Questions: 0 of 4 answered

People John AbrahamsonCo-author of the vaporised-silicon model Warren D. RayleNASA Lewis Research Center; author of the 1966 survey Alexander G. KeulResearcher matching ball-lightning reports with lightning networks Ping YuanPhysicist, Northwest Normal University; led the team that recorded the 2012 ball

Catalog keyMYS0D40A0KP2QRS

Ball lightning is the name for luminous spheres reported during thunderstorms, often near the ground and sometimes indoors, that drift, hover or roll for a few seconds before fading or bursting. The reports go back centuries and now number about 5,000⁠[1]. A survey of NASA staff in 1966 found it reported nearly as often as seeing ordinary lightning strike nearby, and collected 112 detailed descriptions⁠[2]. Yet for most of that history there was no instrumental record, and some physicists doubted it was a physical object at all⁠[3].

In July 2012 a team from Northwest Normal University in Lanzhou, running spectrographs on the Qinghai Plateau to study ordinary lightning, recorded a ball that formed where a cloud-to-ground stroke hit about 900 m away⁠[4]. Its spectrum showed lines of silicon, iron and calcium for its whole life of about a second or more, the signature expected if lightning had vaporised soil⁠[5]⁠[4]. Later analyses of the same data estimated a luminous power of about 4 kW and temperatures that differ by element, from about 2,700 K in the continuum to about 8,750 K in oxygen lines⁠[6]⁠[7].

That record supports the best-known conventional model, in which a strike vaporises silica in the soil and the resulting silicon nanoparticles glow as they oxidise⁠[8]. It does not settle the question. It is one event, observed from a distance, and other models predict different things: microwaves trapped in a plasma bubble⁠[9], chemiluminescent plasmoids of the kind made over water in the laboratory⁠[10], or perceptual effects of lightning's electromagnetic pulses⁠[3]. Recent work matches eyewitness reports against lightning-location networks to see whether balls form where strokes land⁠[11]⁠[1]. Of the 4 questions on this site, 0 are answered. The most likely answer to “Is ball lightning a physical object, distinct from afterimages, hallucinations and misperceived ordinary lightning?”, not yet confirmed: ball lightning is a physical luminous object produced in or near thunderstorms.

Key questionsset by the editors

Is ball lightning a physical object, distinct from afterimages, hallucinations and misperceived ordinary lightning?Most likely: ball lightning is a physical luminous object produced in or near thunderstorms

What forms and powers ball lightning?Most likely: lightning vaporises soil silica; silicon nanoparticle networks glow as they slowly oxidise

All 4 questions

What happened

First page of NASA Technical Note D-3188, Ball lightning characteristics — NASA Technical Note D-3188, Ball lightning characteristicsWarren D. Rayle, NASA Lewis Research Center, 1966. Public domainArtifact

The phenomenon is defined by its reports. In the 1966 NASA survey, balls had a log-normal distribution of diameters and no strong correlation between brightness, size, duration and colour; Rayle tentatively separated two kinds, one that appears after a stroke to ground and stays near the ground, and one first seen in mid-air that vanishes without disturbance⁠[2]. Statistical work on the large Russian SKB data bank of eyewitness reports finds lifetime related to diameter for balls that explode or decay, but not for those that simply go out⁠[12]. The standard reviews and monographs set out these regularities and the range of reported behaviour, including appearances inside buildings and aircraft⁠[13]⁠[14]⁠[15].

Map of the worldOpen the full map

Region World

+−

Lanzhou group's lightning observation site, Datong County, QinghaiField observation site (about 2,530 m)

Within 1 km

Germany, Austria, Czech Republic and SwitzerlandRegion

A region

Pruno, Stazzema, TuscanyHamlet

Within 1 km

NASA Lewis Research Center (now Glenn Research Center), Cleveland, OhioResearch centre

Within 1 km

Within 1 kmA regionNatural Earth

Timeline

Timeline

Squeezed where little happened. Each mark is a published claim, by how it stands:AcceptedMost likelyStill possibleDisputedRuled out

The accounts

Most of the evidence is testimony, collected by surveys, by national databases, and more recently by websites that ask witnesses for times and locations accurate enough to compare with lightning data⁠[2]⁠[1]. In central Europe, 34 cases from 1994 to 2016 were matched with the EUCLID lightning-location network; in the close cases a cloud-to-ground stroke lay on average 0.42 km from the reported ball, and higher peak currents went with shorter distances⁠[11]. In the United States, 31 reports from a defunct website were compared with the National Lightning Detection Network and radar; three of nine well-located reports had plausible nearby strikes, but for the best-located case the nearest strike was 2.3 km away⁠[1]⁠[16].

What was recorded

The Qinghai event of July 2012 is the only natural ball lightning with spectra. Two slitless spectrographs with high-speed cameras, set up to record ordinary lightning, caught a ball that appeared after a cloud-to-ground stroke 0.9 km away and moved horizontally while it glowed⁠[5]. The recorded glow was about 5 m across, though the ball itself was smaller; it drifted about 10 m and rose about 3 m, changing from white to reddish⁠[4]. The spectrum, in the 400–1000 nm range, showed silicon, iron and calcium for the whole lifetime; aluminium, expected from soil, has no lines in that range⁠[4]. Re-analysis found the light stable but with a periodic oscillation carried by the nitrogen and oxygen lines, and mean temperatures of about 8,750 K from oxygen, 4,330 K from silicon, 4,600 K from iron and 2,700 K from the continuum⁠[7]. The luminous intensity peaked at about 1.24 × 10⁵ cd and was about 5.9 × 10⁴ cd for most of its life, with about 4.2 kW of optical power⁠[6]. An earlier video of a probable ball in Italy was analysed photometrically but without spectra⁠[17].

Laboratory analogues reproduce parts of the picture. Arc discharges into silicon wafers make luminous balls that last seconds⁠[18]; microwaves focused into silicates eject buoyant fireballs⁠[19]; and discharges over water make plasmoids up to 0.2 m across lasting a few hundred milliseconds, lit mainly by calcium hydroxide emission⁠[10]. None matches every reported property.

The explanations

Burning soil vapour. Abrahamson and Dinniss proposed that a strike on soil vaporises silica, carbon reduces it to silicon, and the vapour condenses into nanoparticle networks that oxidise slowly, giving light and heat for seconds⁠[8]. It is supported by the silicon, iron and calcium lines in the Qinghai spectrum, by the ball's origin at a ground strike, and by the silicon-arc experiments⁠[5]⁠[18]. Reports of balls inside aircraft, which rival models set out to explain, are hard to reconcile with a soil source⁠[9].

Trapped microwaves. Wu's model has a relativistic electron bunch at the stroke tip generate microwaves that ionise air and are trapped in a plasma bubble; particle simulations support the mechanism, and the model is claimed to cover appearances in aircraft and indoors⁠[9]. It predicts a microwave signature, which no observation has yet recorded.

Chemical plasmoids. The water-discharge plasmoids show that long-lived luminous balls can be made from a discharge over a wet surface, with light from chemical reactions⁠[10].

Perception. Peer and Kendl proposed that lightning's magnetic pulses could induce phosphenes in observers' visual cortex, explaining many reports without any object⁠[3]. Their corrected figures confine this to strikes within metres, or about 50 m for short pulses, which limits the reading to a subset of cases⁠[3]. It cannot account for the Qinghai recording or for balls seen by several people or on video.

Exotic physics, from dark-matter nuggets to magnetic monopoles, has also been proposed without observational tests⁠[20].

Published claims

Published claims and their status

Year	Who	Claimed	Status	Reason
1966	Warren D. Rayle (NASA Lewis Research Center)⁠[2]	Ball lightning is reported about as often as seeing ordinary lightning strike nearby, and the reports fall into two rough kinds.	Accepted	Later surveys and data banks find the same frequency and the same statistical regularities⁠[12]⁠[14].
1969	Roger Jennison⁠[21]	A glowing sphere about 20 cm across drifted down the aisle of the airliner he was travelling on in 1963, after a lightning discharge.	Still possible	A single account by a physicist; it is the standard aircraft case that soil-vapour models struggle to explain⁠[9].
1995	Anvar Amirov and Vladimir Bychkov⁠[12]	In the Russian SKB data bank, a ball's lifetime rises with its diameter for balls that explode or decay, but not for those that simply go out.	Accepted	A statistical result on the largest report bank, not challenged since, though the bank itself is not public⁠[12].
1999	Mark Stenhoff⁠[14]	Ball lightning is a real atmospheric phenomenon with no accepted explanation.	Accepted	The 2012 instrumental record confirmed a physical ball, and no model yet covers every reported property⁠[5]⁠[15].
2000	John Abrahamson and James Dinniss⁠[8]	Lightning vaporises silica in soil, and the silicon nanoparticles that condense glow for seconds as they slowly oxidise.	Most likely	The only natural spectrum shows silicon, iron and calcium, but indoor and aircraft reports are hard to fit to a soil source⁠[5]⁠[9].
2006	Vladimir Dikhtyar and Eli Jerby⁠[19]	Microwaves focused into silicates eject buoyant fireballs from the molten hot spot, a laboratory analogue of ball lightning.	Accepted	A reproducible laboratory result; whether it resembles natural ball lightning is a separate question⁠[19].
2007	Gerson Paiva and colleagues⁠[18]	Arc discharges into silicon wafers make luminous balls that last seconds, as the silicon model predicts.	Accepted	A laboratory demonstration consistent with the Qinghai spectrum, though the balls are smaller and shorter-lived than most reports⁠[18].
2008	A. Versteegh and colleagues⁠[10]	Discharges over water make plasmoids up to 0.2 m across, lit mainly by calcium hydroxide, which may explain some ball lightning.	Still possible	The plasmoids are real but live a few hundred milliseconds, well short of most reports, and no natural spectrum shows their bands⁠[10].
2010	Josef Peer and Alexander Kendl⁠[3]	Lightning's magnetic pulses can induce phosphenes in the brain, so many ball-lightning reports may involve no object at all.	Disputed	Their own erratum cut the induced fields tenfold, confining the effect to strikes within metres, and a ball has since been recorded by instruments⁠[3]⁠[5].
2011	Stefano Sello and Paolo Viviani⁠[17]	A video made at Pruno, Italy, in June 2010 shows a ball lightning, on photometric and dynamical analysis.	Still possible	The analysis is careful but has no spectrum, so the object cannot be identified with certainty⁠[17].
2014	Jianyong Cen, Ping Yuan and Simin Xue⁠[5]	A ball recorded at 0.9 km after a ground stroke on the Qinghai Plateau emitted lines of silicon, iron and calcium throughout its life.	Accepted	Three later analyses of the same data agree on the soil-element reading⁠[6]⁠[7].
2016	Hui-Chun Wu⁠[9]	Microwaves made by relativistic electrons at the stroke tip are trapped in a plasma bubble, which explains balls inside aircraft and buildings.	Still possible	Supported by simulation only; its predicted microwave burst has never been measured⁠[9].
2018	Alexander Keul and Gerhard Diendorfer⁠[11]	Matched against the EUCLID lightning network, European ball-lightning reports lie on average 0.42 km from a cloud-to-ground stroke.	Still possible	A US match found the best-located report 2.3 km from any strike, so the two network studies point different ways⁠[1].
2020	Richard Sonnenfeld, Karl Stephan and colleagues⁠[1]	Matched against the US lightning network, the best-located ball-lightning report had no strike closer than 2.3 km.	Still possible	Only nine reports were well located, too few to separate a stroke origin from a storm-field one⁠[16].
2022	Tingting An, Ping Yuan and colleagues⁠[7]	The Qinghai ball's temperature differed by element, about 8,750 K in oxygen lines and 4,300 to 4,600 K in silicon and iron lines.	Accepted	Derived from the published spectra and consistent with a burning cloud of soil vapour around a hotter core⁠[7].
2025	Ariel Zhitnitsky⁠[20]	Ball lightning is triggered by dark-matter particles of the axion quark nugget type.	Disputed	No observation calls for new physics, and the one recorded spectrum shows ordinary soil elements⁠[5].

What would settle it

More instrumental records, with spectra, from different settings. Lightning research sites that already run spectrographs and high-speed cameras are the likeliest source; a second spectrum, especially from a ball away from soil or indoors, would test whether one mechanism covers all cases. A microwave receiver co-located with such cameras would test the microwave model. Statistically, a large set of well-located reports matched against lightning-location networks with a proper null (random times and places during storms) would show whether balls form at stroke points or merely near storms.