What Is Ball Lightning, and Why Can't Science Explain It?

Background

Reports of a glowing, mobile sphere of light appearing during thunderstorms go back centuries, scattered through diaries, newspaper accounts, and scientific correspondence long before anyone attempted to study the phenomenon systematically. French physicist François Arago made the first serious scientific compilation in 1837, coining the term "foudre en boule" (ball lightning) to group together dozens of historical accounts he judged credible enough to warrant explanation. Later tallies built on his work: by the mid-20th century, researchers had catalogued more than two thousand reported sightings spanning three centuries, and a widely cited 1960 American Physical Society survey of roughly ten thousand respondents suggested that as many as 5% of people may have witnessed the phenomenon at some point, though a more methodologically careful later survey put the figure closer to 3%.

Eyewitness descriptions are strikingly consistent. The typical report describes a sphere or ellipsoid roughly 10 to 20 centimetres across, most often orange, yellow, or white, appearing at or near ground level during or immediately after a lightning strike. It drifts, sometimes horizontally, sometimes seemingly at random, for anywhere from a couple of seconds to, in rarer accounts, over a minute, before disappearing silently or ending with a small explosion and a smell often described as sulphurous or acrid. A 1966 survey of NASA Lewis Research Center employees found that 95% of self-reported witnesses placed the sighting during a thunderstorm, reinforcing the phenomenon's strong statistical association with lightning even though the exact mechanism connecting the two had no accepted explanation.

Main Theories

The silicon-vaporisation model is currently the theory with the strongest direct observational support. First proposed in outline by chemist John Abrahamson and physicist James Dinniss in 2000, it holds that when a lightning strike hits soil, it vaporises silicon compounds in the ground, creating a cloud of silicon nanoparticles that then slowly oxidise in the surrounding air, releasing their stored chemical energy as heat and light over several seconds, similar to a very slow-burning flare. The model gained significant credibility in July 2012, when researchers from Northwest Normal University, working on the Qinghai Plateau in China, accidentally recorded both high-speed video and, for the first time, a full optical spectrum of a natural ball lightning event, formed roughly 900 metres away immediately after a cloud-to-ground lightning strike. The spectrum showed prominent emission lines from neutral silicon, iron, and calcium, exactly the elements the vaporisation model predicted would be present, persisting through most of the event's roughly 1.6-second lifetime.

Electromagnetic and plasma-confinement theories form an older and more varied family, dating to Nobel laureate Pyotr Kapitsa's 1955 proposal that ball lightning is a pocket of ionised air sustained by continuous microwave energy fed in from the parent thunderstorm, rather than a self-contained object burning through its own stored fuel. Later refinements by other physicists proposed variants involving standing electromagnetic waves, "maser" energy storage in the rotational states of water molecules, or self-organising magnetic structures sometimes described as electromagnetic knots. These models can, in principle, explain reports of ball lightning passing through window glass without breaking it or appearing indoors far from any obvious source of vaporisable soil, cases the silicon model struggles with, but no version has been confirmed observationally, and the specific mechanism by which a lightning strike would generate the sustained microwave field these theories require has never been identified.

Neither family of theories has been able to account for every category of report on its own, which is a large part of why the subject remains open: researchers increasingly suspect that eyewitness accounts labelled "ball lightning" may not all describe the same underlying physical process.

Evidence For and Against

The 2012 Qinghai spectrum is the single strongest piece of direct physical evidence in the phenomenon's research history, because it is the only occasion on record where scientific instruments, rather than a human witness, captured a natural event's optical signature in real time. It gave the silicon-vaporisation theory a specific, testable prediction that was then confirmed, a rare occurrence for a subject otherwise reliant almost entirely on eyewitness testimony collected after the fact.

Set against that, ball lightning research has a long history of retracted or reinterpreted photographic and video "evidence." A 2025 study in the Quarterly Journal of the Royal Meteorological Society by Texas State University researcher Karl Stephan examined a widely circulated video recorded in Bozeman, Montana, in August 2023 that appeared to show glowing spheres near the camera; Stephan's analysis found the objects consistent with burning debris ejected from a nearby power-line arc rather than atmospheric ball lightning, and he published the study specifically, in his own words, "to show how easy it is to be tricked by images that look like ball lightning but aren't." That finding does not disprove the phenomenon generally, since the 2012 spectroscopic case and the broader historical eyewitness record are independent of any single disputed video, but it illustrates why researchers now treat photographic and video claims with particular caution, and why instrumentally captured cases like the Qinghai event carry disproportionate evidentiary weight.

Common Misconceptions

Ball lightning is frequently confused with other atmospheric electrical phenomena that already have settled explanations. St. Elmo's Fire, the faint blue-violet glow that appears around ships' masts or aircraft wingtips in strong electric fields, is a well-understood corona discharge, not a free-floating object, and has no genuine connection to ball lightning beyond both occurring during electrically charged weather. Sightings of ball lightning are also occasionally reported as unidentified flying objects, since a fast-moving, self-luminous, unexplained aerial light matches several UFO-report criteria on its face; investigators researching UFO sightings routinely list ball lightning as one of several mundane atmospheric explanations to rule out before treating a report as unexplained. The US Air Force's Project Blue Book invoked exactly this explanation for the 1957 Levelland, Texas car-stalling case, attributing roughly fifteen witnesses' reports of stalled engines and failed headlights to ball lightning during a severe electrical storm, an identification its own scientific consultant later rejected once he found no storm had actually been recorded that night.

The Hessdalen lights, an unexplained luminous phenomenon recorded in a Norwegian valley since the 1930s and studied continuously since the 1980s by the ongoing Project Hessdalen, are frequently mentioned in the same breath as ball lightning because both involve free-floating, physically unexplained light. The comparison has some merit: both are genuine, instrumentally documented anomalies rather than hoaxes, and some Hessdalen researchers have floated mechanisms involving ionised dust that are conceptually related to atmospheric plasma theories of ball lightning. But the two are not confirmed to be the same phenomenon. Hessdalen sightings are not reliably tied to thunderstorms, can last for over an hour, far longer than any reported ball lightning event, and occur in a specific valley with unusual local geology, making direct comparison difficult.

Current Consensus

Mainstream physics accepts that ball lightning is a real, physically documented phenomenon rather than a myth or a purely psychological artefact, a position the 2012 Qinghai spectrum did more than any other single piece of evidence to secure. What remains genuinely unresolved is which mechanism, or combination of mechanisms, produces it: the silicon-vaporisation model currently has the strongest direct observational support for lightning strikes on soil, but it cannot easily explain reports of ball lightning appearing indoors, passing through glass, or occurring in settings with no exposed ground to vaporise, cases that electromagnetic and plasma-confinement theories were originally designed to address without themselves having comparable direct evidence. Researchers increasingly treat "ball lightning" as a descriptive label for an eyewitness pattern that may turn out to have more than one distinct physical cause, rather than assuming every report describes the same underlying process.

Why This Mystery Endures

Ball lightning has resisted a settled explanation for nearly two centuries for a reason most of this site's other open scientific questions share: the phenomenon is too fast, too rare, and too unpredictable to study under controlled conditions, leaving researchers dependent almost entirely on eyewitness testimony collected after the fact and the occasional stroke of instrumental luck, like the 2012 Qinghai team's accidental spectrum. That evidentiary gap, a genuine physical phenomenon that keeps outrunning scientists' ability to capture it directly, is what has let so many competing theories persist side by side for decades without any single one being ruled definitively in or out. Earthquake lights sit in a similar position: a leading, peer-reviewed mechanism with real statistical support, yet individual reported cases that remain genuinely difficult to confirm against mundane alternatives like arcing power infrastructure.

The comparison to the Pioneer anomaly is instructive precisely because the two cases diverge. Both are real, precisely documented physical puzzles that briefly or repeatedly attracted proposals bordering on exotic physics, but where the Pioneer anomaly yielded completely to patient archival detective work once the right decades-old data could be recovered, ball lightning's central obstacle is different: the data simply does not exist to recover, because almost no natural event has ever been captured by scientific instruments in the first place. Zero-point energy offers a different kind of contrast from elsewhere in this site's physics coverage: it is a phenomenon physicists understand precisely but that attracts persistent pseudoscientific claims built on top of solid physics, whereas ball lightning is the opposite case, a phenomenon whose basic reality is not in serious dispute but whose actual mechanism remains authentically unsettled science. This page is part of this site's physics coverage.