Earthquake lights are among nature’s most strange, most ephemeral, and hard-to-explain phenomena. While rare, and only recently accepted by science, they have been reported for hundreds or even thousands of years from essentially every region of the globe where earthquakes occur.
Centuries ago, when the only artificial lights were campfires, burning torches and oil lamps, people must have looked awestruck at the pale cool flames flickering ankle-high out of the ground, or at the flashes of bright light shooting into sky, or the luminous balls of light drifting through the air before bursting silently, leaving in their wake nothing but an acrid smell.
It’s understandable how, in more primitive times before science, people made connections to the world of gods and spirits. Many apparitions from folklore and from historical records can probably be explained by earthquake lights. The story of Moses, who reportedly witnessed a “burning bush” on Mount Sinai, for example, or the ancient Greek Callisthenes who wrote of “immense columns of flame” that foretold the earthquake that destroyed the cities of Helike and Boura in 373 BC.
After the invention of electric lighting in the 19th century, followed by the development of seismology, scepticism towards these strange phenomena grew. These strange lights did not fit into the framework of classical mechanics on which seismology is built. Scientists were at a loss as to how to explain them, and tended to dismiss them entirely. Such scepticism reached its peak in the 1970s when, presented with the colour photos of earthquake lights taken during the 1965-1969 Matsushiro earthquake swarm, prominent seismologists sucha as Tsuneji Rikitake of Japan suggested the pictures were forgeries. They are still a divisive topic to this day.
Tackling earthquake lights scientifically
But recently there has been a change. In a critical review of hundreds if not thousands of reports of earthquake lights in Europe and the Americas over the past 400 years, Robert Thériault and France St-Laurent from Canada, together with John Derr, formerly of US Geological Survey’s Albuquerque Seismological Laboratory, selected 65 reports as most reliable.
For example, shortly before the 2009 earthquake in L'Aquila, Italy, flames of light several inches high were seen flickering above the stone-flagged streets of the town’s historic centre. In Pisco, Peru, as the 8.0 magnitude earthquake struck in 2007, a naval officer saw pale blue columns of light bursting out of the water four times, caught on security cameras. A seismometer record collected at the same site allowed for the exact timing of the light outbursts. In 1988, bright purple-pink globes of light were seen hovering along the St Lawrence River near Quebec City, Canada, 11 days before an earthquake. And just before one of the worst natural disasters in the US, the 1906 San Francisco earthquake, coloured flames hovered in the foothills west of the city and in nearby San Jose, just before the quake struck, perhaps similar to those recorded during the 2008 earthquake in Sichuan, China.
Examining the geology where these outbursts of light had been observed, the authors discovered a rather intriguing pattern. Out of the 65 earthquakes with trustworthy reports of lights, 63 had occurred at or near sub-vertical geological faults, as opposed to faults with more shallow angles. Strikingly, while earthquakes along these steep faults make up just 5% of seismic activity, they account for 97% of documented cases of earthquake lights.
Sub-vertical faults generally occur at rifts within continental plates, rather than along boundaries between them. These are typical of places where, some time in the distant past, the Earth’s crust was pulled apart, allowing magma from great depths of 100-150km, to rise up and intrude. When the magma solidified it left dark, blackish rocks of the gabbro family.
I joined the research team, which published our findings in Seismological Research Letters, to bring some insight to a physical explanation for earthquake lights. For many years I have studied defects in crystals, such as in magnesium oxide (MgO). I found that MgO crystals contain peroxy defects, which are distinct in the way their oxygen atoms are bonded together. Over the following years I found that peroxy defects are ubiquitous, occurring in nearly all rock-forming minerals in the Earth’s crust. The most remarkable property of the peroxy defects is that, when rocks are put under mechanical stress – such as during the build-up of tectonic stresses before a major earthquake or when a seismological wave strikes during an earthquake – they instantly break apart and generate electricity.
It’s like switching on a battery, generating electrical charges that can flow out of the stressed rocks into and through unstressed rocks. The charges travel fast, at up to around 200 metres per second, and far – easily metres in lab experiments, and up to tens of kilometres in the field.
As the charges flow through the Earth’s crust they emit ultralow frequency electromagnetic waves. On arriving at the Earth’s surface, they produce a range of secondary reactions such as infrared light, massive air ionization, and corona discharges, which in turn produce ozone and broad-band radio noise. These signals are indicators for the massive stress build-up deep below, and so could be considered pre-earthquake indicators – though not every tectonic stress build-up leads to a catastrophic rupture of the rocks, that is, to an earthquake.
Crystallising the physics
Producing earthquake lights, however, requires special conditions, which we are only now slowly beginning to understand.
In real situations, for earthquake lights to occur the rocks must contain many peroxy bond defects, as is true for gabbro rock, and they must be stressed very quickly. Under these conditions, so much electrical charge can be generated that the rocks enter a rare physical state known as solid state plasma.
At this point, when the positive or negative charge-carrying particles generated by the rapid stressing of the rock reach a critical density, they can move through the rocks as plasma. When this cloud of electrical charge reaches the Earth’s surface, it can burst through and discharge into the air.
And so voilà, the stage is set for the luminous phenomena of earthquake lights: variously reported as flickering, pale, cool flames reaching ankle-high, or different forms of bright, short flashes that shoot high into the air, or as free-floating plasma balls that detach from the ground and drift away, described by eyewitnesses as “luminous globes” or “ball lightning”.
Reading the warning signals
For more than a century early warnings of earthquakes have remained a dream. There are records of many different types of signals that the Earth seems to produce during the weeks, days and hours leading up to a major seismic event. But nobody has been able to “read” these signals, or say how they may be connected.
This has left a deeply divided community: mainstream seismologists who reject even the idea of pre-earthquake signals, pitched against a group of well-intended individuals with a wide range of scientific and non-scientific interests and backgrounds. But as is often the case in the history of science, once the basic physical processes responsible for a set of seemingly unconnected processes are deciphered, the pieces fall into place.
We are at this crossroads, having begun to understand the Earth’s signals. Current seismological earthquake prediction deals in windows of uncertainty of 30 years or more, or issues “early” warnings only when an earthquake is already underway and racing through the ground at kilometres per second. But with resources and effort, we could now bring together all the relevant scientific information with its basis in physics to build a global earthquake forecasting system that could provide early warning of at least a few days – a huge leap forward.