Earthquake Lights: Do They Exist?

by Brian Dunning
August 30, 2016

The ground rumbles, buildings quiver, items tumble from shelves. Facebook and Twitter light up with pictures of the damage, hair-raising tales of fear, and — in the past few years — a new addition to earthquake lore. Commenters and reporters have taken to calling them earthquake lights. They are said to be natural illumination somehow triggered by the same forces that cause earthquakes, and stories of them have quickly spread to all corners of the Internet. They're mentioned equally on science and pseudoscience websites, and have clearly captured the imagination of those interested in the mysteries of the Earth. Today we're going to dig in and see if earthquakes lights truly are an actual phenomenon requiring an explanation, or whether it might be even simpler.

Here's our first warning sign that earthquake lights (often abbreviated EQL) might not be a scientifically sound idea: Descriptions of them vary widely. Sometimes they're described as glowing patches up in the sky like an aurora, sometimes they're bright spark-like flashes on or under the ground, sometimes the name is attached to conventional lightning near an earthquake. There are also temporal contradictions. Some say they appear before an earthquake to warn that it's coming, and some say they happen during or after a quake. This lack of consistency tells us one thing very clearly: This is almost certainly not one, known, proven phenomenon. Earthquake light reports are totally disparate in location, duration, size, shape, color, behavior, time, and other metrics. Thus, if there is such a thing, most of these discordant reports aren't consistent with it, and are therefore misidentifications of something else. So now let's see if there actually is some single natural phenomenon that we can call earthquake lights.

Although attention has soared in recent years, EQL have been reported for a long time. Way back in 1913, the Bulletin of the Seismological Society of America published a survey of as much existing literature as the author could find. He began by criticizing the quality of the reports:

The descriptions lack precision, and many of them contain fantastic details. The time-intervals between the earthquake and the luminous phenomena are very variable, being some seconds, minutes or even hours before or after them. They are reported indifferently from localities of the epicentral area or from very distant ones, and they more frequently come from the atmosphere than from the ground.

These reports were of two basic kinds: lightning and shooting stars. In 1913, today's more exotic versions of EQL had not yet entered the public consciousness, which could be interpreted as evidence that it's more of a cultural phenomenon than a physical one. Supporting this interpretation is that the largest collection of earthquake information he reviewed, which was from China, never mentioned luminous phenomena at all — again suggesting that this might be more cultural than physical. And of the most reliable reports of lightning during earthquakes, the author found that such quakes were usually coincident with thunderstorms. He concluded:

At the present stage of our observations is it not scientific or rational either to affirm or to deny the existence of luminous earthquake phenomena; but all the arguments and facts at our command bear us out in a negative conclusion.

And thus, EQLs have resided in something of a scientific "twilight zone" ever since. A staggering volume of literature has been written. Science journals are full of proposed explanations for how such things might be. Legitimate journals, too; with articles co-authored by credentialed, serious academics and their similarly-adorned colleagues. Each is followed with pages of references. But when we look closer, we see that hardly any of these papers agree on anything; and that their proposed mechanisms for the lights are all over the map: bizarre, hypothetical if not fantastical, and not one has ever been conclusively observed. I'm forced to wonder how many of these eager researchers are familiar with Hyman's Categorical Imperative: "Do not try to explain something until you are sure there is something to be explained."

One example came in 1973 when "Earthquake Lights: A Review of Observations and Present Theories" was published in the Bulletin of the Seismological Society of America, and its discussion focused on a swarm of earthquakes in Matsushiro, Japan in 1966. A dentist took a series of photographs, shortly before midnight, showing the sky all lit up. They covered a period of about 20 seconds when it got bright, then dimmed back to normal. It's difficult to analyze the photos, because they are said to have been 4-second exposures using an F1.9 lens (and we'd expect the city sky to look pretty bright with such settings) but were said to have been taken only 2 seconds apart: an obvious impossibility. This author's favored explanation was:

Ultrashort-period air oscillations and generation of a large potential difference in quartz-bearing rock by the piezoelectric effect.

The piezoelectric effect is where physical pressure on certain crystals can produce a voltage, and it's often brought up in EQL articles, just as it is in articles on various ghost lights worldwide (see Hessdalen, Min Min, and Marfa). But neither voltage nor air oscillation is the same thing as a glowing sky, so these proposals are incomplete at best.

Three scientists from the US Geological Survey published in Nature in 1983, and went into great depth discussing how heat and an electrical charge could easily be produced by an earthquake and concluded "EQL can be generated and should be expected for at least some earthquakes". Their hypothesis for how this could make light was arcing and corona discharge, both incompatible with luminous sky EQLs, and — not incidentally — both unproven to be associated with earthquakes. But this explanation became as close to the mainstream as EQLs have ever had.

Earthquake lights made the news quite famously in 2008, when the daytime skies lit up above China with a brilliant display of glowing rainbow-colored auroras, thirty minutes before the magnitude 8.0 Sichuan quake that killed a staggering 69,000 people. People posted their videos of it to YouTube, and the event firmly fixed the idea of earthquake lights preceding major quakes into the public consciousness.

This particular display did not impress atmospheric scientists, as the rainbow patches above Sichuan were (to those familiar with the phenomenon) simply iridescent clouds. These are patches of rainbow caused by water droplets or ice crystals in a cloud. If you do a Google image search for "rainbow in clouds" you'll see that these are often photographed. The one over Sichuan just happened to be on the same day as a major earthquake, and people invented the connection on their own.

The latter part of 2013 saw a number of significant deadly earthquakes, such as a 7.7 that killed over 800 in Pakistan and a 7.1 that killed over 200 in the Philippines. With the public's attention captured, reporters looking for grist trumpeted an article published at the beginning of 2014 in Seismological Research Letters. It was titled "Prevalence of Earthquake Lights Associated with Rift Environments" and concluded with the following proposed mechanism:

Mobile charge carriers, termed positive holes or pholes, flow along stress gradients and eventually accumulate at the surface, ionizing air molecules and leading to the generation of luminosities, among other phenomena.

Sufficiently energized, air can indeed emit light, just like the gas inside a fluorescent tube. This can happen in nature: we call it St. Elmo's Fire, best known for its occasional appearance as a faint blue/violet glow around the masthead of a tall ship during a thunderstorm. A field of approximately 100kV/m is needed, which is really strong. Pointed tips of ship spars are ideal geometries for St. Elmo's Fire, but the ground is quite the opposite. If an earthquake did generate enough electrical potential for St. Elmo's Fire, it would almost certainly dissipate throughout the ground (after all, the ground is already grounded) much faster than it could build up a sufficient potential in some prominence like a treetop or flagpole. It is perhaps noteworthy that in our whole history of studying St. Elmo's Fire, it's only ever described as appearing up high on a pointed prominence, and never down on the flat ground. Does that mean it can't? Certainly not. But you'd think somebody might have noticed by now.

Hot on the heels of this article, and still riding the wave of the late-2013 quakes, came another slew of popular media reports discussing the work of Troy Shinbrot at Rutgers. By agitating containers of granular material, Shinbrot managed to create electrical voltage, no matter what type of material the grains consisted of. (The current was near zero, but at as much as 400 volts.) This is called the triboelectric effect, in which materials pick up an electrical charge by friction. A related effect is triboluminescence, where the material emits tiny flashes of light. Neither is completely understood, but both are reproducible.

Other similar effects are often proposed as the mechanism for EQLs. Fractoluminescence produces the tiny light flashes when material is fractured — easily demonstrated when you bite into a Wint-O-Green Life Saver. The piezoelectric and piezoluminescence effects appear when certain materials are mechanically stressed but not broken. Those luminescence effects sold the media. If you see an article on the web saying that earthquake lights are now explained, it's likely to cite either of these 2014 authors; and the article almost certainly takes for granted that earthquake lights are a thing.

It was getting murky: more explanations that seemed to be a poorer and poorer fit for the reported observations. Nobody could agree on whether earthquake lights look like lightning, meteors, specks of light dancing on the ground, or a great glowing region in the sky; and yet everyone was putting forth the same basic explanation: electrical voltage in the ground, though produced by all sorts of different mechanisms, some real and some theoretical.

So I spoke with Prof. Shinbrot directly and asked him. His response was measured and scientific:

Earthquake lights are not an unknown phenomenon: they have long been reported, but their cause, or causes, are unconfirmed... The relation, if any, between our lab charging experiments and earthquake lightning also needs to be confirmed... Regarding earthquake lights themselves, it is possible that there are disparate causes, and some reports are fanciful.

But have there actually been any confirmed observations? He provided numerous studies, and there are a lot of cases where measuring equipment has been set up along earthquake-prone fault zones; and, sure enough, voltages have been detected before, during, and after quakes. It's highly inconsistent, but it does happen. Links to a few such papers are in the references below. As far as reliably observed lights, though? Still zilch. The same number as the Chinese recorded in their 19th century database.

My take is that most of the EQL phenomenon is sufficiently anecdotal that it should be dismissed pending decent evidence. What remains is plausible, though still backed up by only incomplete theory and poor evidence, and that's conventional lightning striking near active faults during quakes. It's a plausible idea, in search of better observations with better theory to explain them. When you hear claims of anything more, like the sky lighting up for minutes, or glowing colorful clouds portending quakes, or ball lightning bouncing around, you have very good reason to be skeptical.

By Brian Dunning
Follow @BrianDunning