The ability to predict earthquakes could save thousands of lives every year. But for most scientists, knowing in advance when and where such events will happen is little more than a pipe dream. Jon Cartwright tells the story of one physicist who believes that such warnings could soon be possible.
Down the road from his lab at NASA’s Ames Research Center just south of San Francisco Bay, California, Friedemann Freund is a regular visitor at the local memorial mason. Among the rows of ready-chiselled gravestones, he likes to browse the sundry stacks of raw, unfinished rock imported from as far afield as Norway and China. “Just by looking at a rock I can say, yes, this one looks good for us,” Freund explains. “The black ones are the best. They will tend to conduct the charge well.”
Electrical conductivity is not a property that is often associated with rocks, which are insulators under normal conditions. Freund, however, is interested in rocks under extraordinary conditions. His lab experiments involve studying what happens when high mechanical stresses are applied to igneous rocks — materials that solidified from magma and that are common deep in the Earth’s crust, where earthquakes form. The routine is straightforward: he puts big slabs of rock under two pistons and crushes them while probing for associated electromagnetic effects. “In the beginning, I was always breaking them,” Freund says. “Now we are much gentler, and a rock can last for weeks or months. We can do tens of experiments on a rock without ever breaking it.” He is hoping that one day his work will help save thousands of lives. Friedemann Freund wants to understand how earthquakes can be predicted.
Shaky understanding
Earthquakes are the only natural disasters that scientists are unable to predict with any reliability. Institutions like the US Geological Survey (USGS) monitor the strain in the Earth’s surface through movement sensors in the ground and, together with historical records of seismological activity, they can usually forecast the long-term prospects of earthquakes occurring in active regions, typically within a period of 30 years. Predictions, which need to specify the exact time, place and magnitude of an impending tremor, have proved hard to come by. Part of the problem is that seismologists do not have a clear picture of how the ground fractures. Although the study of plate tectonics has allowed researchers to isolate the most quake-prone regions, the current thinking is that each tiny fracture in the Earth’s crust spreads in a chaotic fashion. This means that it is difficult to say which cracks will stop short, and which will rupture into an Earth-shattering event. Most seismologists believe that impending earthquakes send no reliable warning signals.
Freund has a different opinion. Deep down in the crushing boundaries between the Earth’s tectonic plates — where earthquakes form — the conditions are far from normal. As the plates struggle to grind past each other, the stresses grow until the plates finally slip with a devastating release of energy. Freund thinks that this huge stress build-up prior to an earthquake can flood the surrounding rock with electric charge. Indeed, he believes that in the hours or days before an earthquake the ground could brim with so much charge that it generates a host of visible effects above the surface, such as infra-red emissions and vivid corona discharges. These electromagnetic phenomena could be earthquake precursors.
But despite living within walking distance of the infamous San Andreas Fault, Freund was not always involved in earthquake science. The interest spawned from his research in the early 1980s, when he was studying the properties of simple crystals like magnesium oxide (MgO). He found that MgO always absorbs infra-red light at the characteristic wavelengths of hydrogen molecules. The only way for hydrogen to be present, he thought, would be if water crept into the structure during crystallization as defect OH– groups among the Mg2+ and O2– ions. Then, pairs of the OH– groups could combine to form H2, leaving the remaining O– ions to bond into more stable O22– groups, known as peroxy links.
Peroxy links, according to Freund, are key in turning MgO and other insulators into conductors. With a little heat, the wavefunctions of a peroxy link’s constituent O– ions “loosen up” and spread over hundreds of neighbouring ions; increase the temperature further and the O– ions completely dissociate from each other. In this state, each of the O– ions is missing an electron that it would need in order to be stable. But the missing electron or positive “hole” — which Freund prefers to call a “phole” — is able to hop to a nearby, non-defect O2– ion. Indeed, the heated crystal acts like a pure semiconductor in which the pholes repel one another through the sea of O2– ions to form a blanket of positive charge on the surface.
By 1994 Freund had collected strong evidence for pholes through measurements of electrical conductivity and other properties, and was beginning to think of other ways to recreate the phenomenon. As he recalls, “The logic was, what does it take to break the peroxy bond? When you heat the crystal, you’re really just increasing the amplitude of vibration of the ions. It was then that I began to wonder whether dislocations that are produced by mechanical deformation could also do the job.”
Freund realized that MgO crystals are too brittle to sustain much deformation, so he turned his attention to stronger crystalline materials that could allow peroxy-link defects — rocks. For his initial experiments he took an unconventional approach: using a modified toy crossbow, he fired pea-sized steel pellets at 100 m s–1 into small rock cylinders. His hunch paid off. Not only did he measure a positive surface potential of about 400 mV spreading from the impact area across the rock, but he also recorded a concurrent burst in infra-red emission (2002 J. Geodynamics 33 543). Freund attributed the latter event to the pholes liberating their stored energy as they recombined into peroxy links at the surface. “It was what you call a serendipitous discovery,” he says.
Bad omens
Freund believes that his discovery can explain some of the bizarre events that are said to signal an imminent earthquake, such as eerie lights and strange animal behaviour. In 1966 in Matsushiro, Japan, ghostly lights were photographed during a string of tremors. Last year in the UK, following a moderately strong earthquake that rippled through the small Lincolnshire town of Market Rasen, The Times reported one frightened woman’s account of a “grapefruit-sized glowing sphere” that materialized in her bedroom and floated towards her, and others who claimed to have seen lightning flashes even though there were no storms.
In the winter of 1975 in Haicheng, China, there were widespread reports of peculiar animal behaviour: dogs growing very agitated; cattle running amok; and even snakes suddenly waking up from hibernation only to die because of the freezing conditions. Encouraged by seismologists who had also started registering an increase in low-amplitude seismic activity, the authorities decided to evacuate the region. A couple of days later, a quake with a magnitude of 7.3 struck the region, killing over 2000 people. That figure could have been 100 times higher had the population not been evacuated.
Seismologists, however, doubt the significance of these precursors. For over a hundred years they have tried in vain to correlate such events with seismic activity, and found them to be unreliable warning signals, especially since most were reported after the event. Successful predictions like that at Haicheng they deem to be flukes because there is no consistent pattern of accurate predictions.
Freud agrees that we still do not have a fully fledged prediction technique, but he thinks researchers are missing the big picture. He is confident that he has an underlying mechanism that will indicate where to look for precursors — so confident, in fact, that he has been backing his work with more than one million dollars of his own cash.
Freund’s idea is that, kilometres underground, the stress of an earthquake nucleation could produce a cloud of pholes that surges to the surface, creating electromagnetic disturbances such as earthquake lights. He has already seen related phenomena in the lab. By positioning pistons above and below a slab of rock to inflict concentrated loads, he has found that above a mass of few tonnes, a nearby, negatively biased metal sheet can draw a 10–25 nA current of positive ions across a 5 mm air gap. On the other hand, if the sheet is positively biased, it can cause electrons to shower onto the rock in a fleeting 100 nA current. This electrical breakdown also produces a flash of visible light, or what is known as a corona discharge.
Freund thinks most other supposed earthquake precursors, too, have their origin in the propagation of pholes. He points to past clinical tests indicating that positive ions can distress animals — among other things causing respiratory problems and a heightened sensitivity to pain — which might be why they are sometimes seen to behave oddly. He says positive ions could also attract or repel regions of the ionosphere, an effect that researchers apparently recorded in the majority of earthquakes that occurred around Taiwan between 1999 and 2002. And then there is the infra-red emission. Several satellites have recorded what are deemed “thermal anomalies” above the epicentres of major shocks, including some before the 6.2 magnitude quake that struck the county of Zhangbei, China, in 1998. Freund thinks these anomalies have the same source as the infra-red emission in his experiments, namely the recombination of pholes into peroxy links (Earth and Planetary Science Letters submitted).
All this may be a lot to take in, but that is the point. In the past almost all those trying to search for signs of earthquakes have only had the facilities to monitor a single type of precursor, whereas Freund claims that his mechanism could show them how the precursors are all related, and thus where to look. “There are people who analyse the ionosphere, for example, and if they see a bump in the data they claim that this is an indication of an impending earthquake,” he explains. “Then people rightly say that this is too much — you can’t draw a one-to-one correlation with just one parameter. My work could enable people to look at several parameters, each of which could be an indicator, to search for an early warning.”
Stressful work
One of the more common questions directed at Freund is whether there could be any other explanation for the rocks’ conductivity and the related electromagnetic phenomena. The most obvious would be piezoelectricity, in which certain materials — notably quartz — build up a charge imbalance when they are stressed. But while it is true that many of the rocks chosen by Freund, such as the “Sierra White” granite sourced from within California, contain a third or more quartz, those from further afield, such as the black gabbro from northern China, are quartz-free.
Another possibility is that the conductivity is caused by a phenomenon known as a streaming potential. This type of voltage is sometimes generated in machines when weakly conducting fluids such as fuel or transformer oil are pumped through pipes, though in rocks it can also occur if there is salt water present. The water seeps through pores in the rock, picking up ions of one charge while leaving aside ions of the opposite charge. Freund points out that the charge in his impact experiments flows at between 100 to 300 m s–1, which is too fast to be a streaming potential in rocks. Furthermore, when Freund later upgraded his crossbow to a canon at NASA’s Ames Research Center that is known unofficially as the “Big Gun” — a research tool typically used to study the formation of meteorite craters — the shock waves resulting from impacts at 1.5 km s–1 appeared to activate charges throughout the rock instantaneously.
But the flip side of water, according to Tony Fraser-Smith, a geophysicist at Stanford University, California, is that it might actually stem electrical current by “shorting out” any charges present. Freund admits that the pholes could react with water, although he thinks that the process would actually complete the circuit to keep the charges moving (Earth and Planetary Science Letters at press). “Fraser-Smith is right in saying that water may do something to the currents,” he says. “But it is not as destructive as he thinks.”
In any event, Freund is not the only researcher to have noticed the effect. Al Duba, a retired geophysicist who used to work at Lawrence Livermore National Laboratory, California, spent the better part of his career investigating anomalous conductivity in rocks. However, he concluded that the conductivity it is due to contamination, and he managed to destroy it by heating samples above 700 °C in a mixture of carbon dioxide and carbon monoxide. Freund argues that this process only serves to react away the crucial O– ions. “We used to have friendly discussions about it,” he says. “Duba thought that it must be junk, and that you should get rid of the junk. But I said, ‘No, you’re getting rid of the golden egg!’.”
Don’t mention the “p” word
Earthquake prediction is a pejorative term — at least among most seismologists, who make up the bulk of earthquake researchers. Although Freund prefers not to use the word prediction, saying it is “too strong a word”, his research inevitably falls under that banner because of his claims that it could lead to an early-warning system.
The trouble with prediction is that it has a long history of failure. For most of last century seismologists devoted themselves to finding statistically reliable precursors, but failed to find any that occurred consistently before major quakes. Now the nearest thing to a consensus is that prediction is an unlikely goal, at least in the short term, and that researchers first need to get a better understanding of how fractures nucleate and spread through the Earth’s crust.
But for some, prediction research should be outlawed altogether. Robert Geller, a seismologist from Tokyo University, dislikes the fact that certain countries, like Japan, give disproportionate funding to those trying to hone prediction methods, which he says will never work. “My stance is as follows,” he says. “Anyone who wants to do earthquake-prediction research should send his or her proposal to the normal funding system where it should get reviewed in competition with all other research in geophysics — that is, treated neither favourably nor unfavourably. All work that passes such a normal review should be funded.”
Freund sees nothing fair about the US funding system. He says that he has sent grant proposals to the USGS annually for the past five years only to have each rejected on what he insists are “unscientific” grounds. Although in the early days of his research NASA lent him modest support, he has since had to finance himself. “I have essentially been blacklisted by the seismology community,” he says.
Tom Heaton, a geophysicist from the California Institute of Technology, thinks the main problem is that the seismology community has been “betrayed” too often in the past by those who believed that their observations in the lab would scale up to the real world. This happened in the 1970s, when many seismologists became excited that rocks under stress in the lab appeared to swell as a result of numerous microfractures. However, subsequent attempts to exploit the effect to predict earthquakes failed. “Now when we monitor the stress at the beginning of a big earthquake and the beginning of a small earthquake, we don’t see a difference,” he explains. “So even if people could predict earthquakes, they might be making predictions for all the hundreds of small earthquakes as well as the big earthquakes. What would we do with a hundred or so predictions?”
But Freund insists that this does not rule out the possibility of useful precursors. “When seismologists talk about stress and strain, they mean putting meters into bore holes that are between 200 and 1000 m underground,” he says. “But most earthquakes nucleate in the 10 to 30 km range. The seismologists have to rely on extrapolation and linear models, which they know aren’t much good.”
Recently, though, the USGS appears to have had a change of heart: it has invited Freund to give a talk next month. However, for the man who has made such a huge personal and financial commitment to his work, entering what he calls “the lion’s den” is not a matter of pride but the sole opportunity to mend the perceived fault line separating his work from established science, and perhaps give stability to the many millions of people living on uncertain ground.
