IN TANGSHAN, China, residents were woken one night in 1976 by fireballs and flashes in the sky. The next night, an earthquake registering around magnitude 7.8 on the Richter scale killed over 240,000 people and destroyed the city. The stories of the mysterious lights joined local lore, but were they related to the earthquake, or simply rumours invented by a terrified populace?
It’s not the first time people have seen strange lights before a quake. Throughout history, eyewitnesses have reported orange glows, bluey-white luminescence, fireballs or flashes in the days before and during an earthquake. As far back as 1755, after the great Lisbon quake set church bells ringing as far away as Sweden, the philosopher Immanuel Kant wrote of warning signs. “Eight days before the concussion the ground near Cadiz was covered by a multitude of worms that had crept out of the earth. Of several other earthquakes, violent lightning in the air and the fear that one notices in animals have been the precursors.” Then in 1968 the first photographs of “earthquake lights” were taken by Yutaka Yasui of the Kakioka Magnetic Observatory in Ibaraki during a series of quakes at Matsushiro. Some showed red streaks across the sky, like low-lying aurora. Others looked like a low blue dawn from a distance. In 1999, floating balls of light in the sky were broadcast on Turkish television, reportedly filmed the night before the devastating earthquake of 7.4 on the Richter scale that killed 15,000 people in Izmit, Turkey.
Mysterious or not, repeated sightings of earthquake lights confirm their existence. “It has to be said that earthquake lights are a fairly well known phenomenon,” says Chris Marone, who works on the physics of rock deformation at Pennsylvania State University in University Park. But what we don’t know is what they mean, or what causes them. Seismologists have struggled for years to find a reliable quake predictor. Could the lights hold the key?
Advertisement
The US Federal Emergency Management Agency (FEMA) has begun asking that very question. During the past few years they have funded NASA to study earthquake lights, using weather satellites and the MODIS research satellite. The studies followed claims by Russian and Chinese scientists in the early 1990s that published US weather satellite data shows infrared effects in the region around an earthquake, a few weeks beforehand. Mainstream geologists had dismissed the earlier claims as coincidental.
The main problem facing FEMA is that earthquake lights still don’t have an accepted scientific explanation. But that might be about to change. When Nevin Bryant, a remote-sensing expert at NASA’s Jet Propulsion Laboratory in Pasadena, began looking into earthquake lights, he stumbled on a theory put forward by physicist Friedemann Freund of San Jose State University in California. Freund’s theory was described in a paper in 2000 (Journal of Geophysical Research, vol 105, p 11,001). Although controversial, it is also the most promising explanation for the thermal emissions seen on US satellite images.
“Until our discussions with Freund, we had found no plausible explanation for the rapid appearance and disappearance of thermal anomalies we noticed on the weather satellite data,” wrote Bryant in a report to the Carnegie Institution of Washington in Maryland. Stuart Nishenko, former senior seismologist at FEMA, now at the Pacific Gas and Electric Company in San Francisco, agrees: “It did help focus thinking about what it is we were looking at.”
Freund attempts to explain how earthquakes could generate both the lights captured on Yasui’s camera and the strange thermal emissions seen by the satellites. His idea is that the immense pressures generated prior to an earthquake cause igneous rocks, which normally act as insulators, to briefly behave like “p-type” semiconductors, meaning that they contain mobile positive charges that can conduct electrical charge. Crystals in volcanic rocks contain paired oxygen atoms, called peroxy groups, which can snap under stress. Freund speculates that once a peroxy group is snapped, a negative oxygen ion will remain trapped in the lattice of the rock, while a positive charge – or hole – will be free to flow through it (see Graphic). Since like charges repel, all the positive holes will swiftly flow outwards. Even rock that is not under stress should be able to transport these holes, so some will reach the surface above.
Freund envisions two ways a build-up of charge could cause strange phenomena at the surface. In one scenario, positive holes crowded at the surface could recombine with electrons in the surface layer to reform a peroxy bond. “This will be very excited at the moment of its birth,” says Freund. The bond will then release energy in an infrared emission like those seen by Bryant and Nishenko in their satellite data.
In a second scenario, which could overlap with the first, the positive charges spread out on the surface to a very thin layer, which behaves just like a huge, flat electrode on top of the ground. As the charge layer thins to about 10 nanometres, its relatively weak surface potential of 0.4 volts is concentrated over a very small distance, so the electric field produced is huge, around 400 kilovolts per centimetre, although the field could be even higher. This would be enough to ionise the air several metres above the ground, causing the luminous plasma known as earthquake lights.
In Freund’s view, whether the positive charges produce visible light as well as infrared emissions depends on how many charges arrive at the rock’s surface, and also on the topography of the surface. Ionisation, producing visible light, is most likely to occur over peaks or crags of rocks where the surface area is greater and the charge layer therefore thinner. A thinner layer of the same amount of charge produces a higher electric field within the layer, making electric breakdown of the air more likely. This would explain why anecdotal reports of earthquake lights seem more common in mountainous regions.
The basic idea behind Freund’s theory – that cracking and stressing materials should lead to broken and dangling chemical bonds and wandering charges – is not unreasonable, say mineral physicists. According to mainstream theory, however, the positive holes should dissipate quickly inside the material, instead of building up on the surface.
But Freund says he’s proven positive charges do build up by recreating earthquake lights in the lab. He crushed half metre blocks of typical igneous rocks, such as granite and anorthosite, and blocks of upper mantle minerals such as olivine, using a 1500-tonne hydraulic press funded by NASA’s Goddard Space Flight Center, Maryland, and the Carnegie Institution. Each sample was monitored for infrared emissions or the accumulation of positive charges. Freund found that the spectrum of infrared light emitted by the stressed samples had a discrete peak that could correspond to the energy emitted by newly formed peroxy bonds on the surface. This means the emission is not due to heating by friction inside the rock. “The bulk of the rock does not get hot,” says Freund.
Infrared peaks have also been recorded in real earthquake data. Bryant’s group looked at satellite records of infrared emissions from two earthquake zones – the 1999 Izmit quake and the 2001 Bhuj India quake, which measured 7.9 on the Richter scale and killed 20,000 people. In the week or two before each quake, the team found strong infrared emissions at a wavelength of around 10 micrometres, similar to the peaks Freund saw in the rock-crushing experiments. “It is strong observational evidence that here is a mechanism to produce what we are seeing in the satellite band,” says Nishenko.
If, as Bryant and Nishenko originally thought, the infrared light on the satellite data came from the earthquake heating the ground, it would correspond to a temperature rise of between 2 °C and 4 °C. But there is very little data to confirm this. Before the data can be interpreted it is important to know how Bryant’s team takes account of hourly and daily variations in ground surface temperature, says Malcolm Johnston who co-runs a network of crustal monitoring stations from the US Geological Survey (USGS) in Menlo Park, California. Bryant will not comment while his work is pending review for publication.
The rock-crushing experiments support Freund’s ideas. For a start, the greatest infrared emission occurs at the edges and corners of the rock, not at the point of rupture. Once the rock ruptures, the emission slows and stops, just as happens in the satellite data. The samples also emit visible light and flashes, which Freund believes are due to positive charges on the surface collecting in a thin enough layer to create a strong field that ionises the air – just as might happen before an earthquake.
But none of the evidence is conclusive – critics say that a place where internal deformation causes a rock to heat need not be the same place where it eventually gives way. And there could be an alternative explanation for the lights and flashes. “You can generate light by crushing sugar grains,” says Johnston. In the case of sugar, the light is due to electrons recombining with atoms that have had their electrons stripped away by the applied force that snapped the bonds between them. Freund’s methodology makes it hard to distinguish this mechanism from the one he proposes, says Johnston. “You need some way to separate different processes.”
Another problem for Freund is that the Earth’s crust is saturated with water. As a good conductor, water should quench the build-up of charge, by providing a conduit for holes to travel back to the site they were stripped from. “Water is an ionic conductor. It’s a short circuit,” says Steve Constable, who studies the electrical conductivity of rocks and minerals at the University of California, San Diego. But this logic might not apply to Freund’s novel mechanism, as the water could help the positive charges to flow freely. “The water could be a conduit for getting the charges into the atmosphere,” says Marone. Freund intends to try his experiment with wet rock later this summer to find out who is right.
Even if the wet-rock experiments go his way, Freund faces a greater challenge: demonstrating that the lab mechanism really applies in nature. Most earthquakes occur at plate boundaries, where one plate slides beneath another hundreds of kilometres below the Earth’s surface. Many geologists find it hard to believe that positive holes liberated so deep down could flow to the Earth’s surface and collect there without being reabsorbed. Then again, earthquake lights are a real phenomenon, and some kind of mechanism must be creating them. Whatever it is, says Marone, it will involve maintaining charge over surprisingly large distances. “This is a very, very hard problem.”
Freund himself is hoping that the FEMA and NASA attention will encourage other groups to try to repeat his rock-crushing experiments – especially in wet rock. Meanwhile Bryant’s group is continuing to try to put the study of earthquake lights and other electromagnetic phenomena on a sound statistical footing. More work is needed to find out if the lights and radiation are reliable indicators, or whether they only appear before some quakes, or can appear when a quake does not follow.
But if the group succeeds, satellite data could eventually be far more useful than seismic data. Today’s earthquake prediction models use seismic information to produce a probability of a quake over several decades. For example, the USGS calculates that there is a 67 per cent chance of an earthquake registering 6.7 or greater on the Richter scale occurring in the San Francisco Bay area by 2032. This is a far cry from an earthquake forecast for concerned residents. If mysterious earthquake lights turn out to be a reliable phenomenon, the faintest hint of ghostly light on the horizon could be considered a good reason to pack up and go.