Can Earthquakes Be Predicted?
If we could predict earthquakes reliably, communities could be prepared, people could be evacuated from dangerous locations, and many aspects of the impending disaster might be averted. How well can we predict earthquakes?
Predicting an earthquake means specifying its time, location, and size. By combining plate tectonic theory with detailed geologic mapping of regional fault systems, geologists can reliably predict which faults are likely to produce earthquakes over the long term. However, specifying precisely when a particular fault will rupture in a large earthquake has turned out to be very difficult.
Long-Term Forecasting
Ask a seismologist to predict the time of the next large earthquake at a particular location and the response is likely to be, “The longer the time since the last big quake, the sooner the next one will be.” As we have seen, the recurrence interval—the time required to accumulate the strain that will be released by fault slipping in a future earthquake—can be calculated from the rate of relative plate movement and the expected fault slip, as estimated from the displacements observed in past earthquakes. Geologists can also estimate the intervals between large earthquakes up to several thousand years in the past by finding and dating soil layers that were offset by fault displacements (Figure 13.32).
Although these two methods usually give similar results, the uncertainty of the predictions turns out to be large—as much as 100 percent of the recurrence interval. In Southern California, for example, the recurrence interval for the San Andreas fault is estimated to be 110 to 180 years, but the observed intervals between individual earthquakes can be appreciably shorter or longer than this average value. One part of this fault experienced a large earthquake in 1857, whereas another part (the southernmost) appears to have remained locked since a large earthquake that occurred around 1680 (see Figure 13.2). Therefore, an earthquake can be expected there at any time—tomorrow, or decades from now.
Because the prediction intervals are decades to centuries, these methods of earthquake prediction are referred to as long-term forecasting to distinguish them from what most people would really want: a short-term prediction of a large rupture on a specific fault accurate to within days or even hours of the actual event.
Short-Term Prediction
There have been a few successful short-term earthquake predictions. In 1975, an earthquake was predicted only hours before it occurred near Haicheng, in northeastern China. Chinese seismologists used what they considered to be precursors to make their predictions: swarms of tiny earthquakes and a rapid deformation of the ground several hours before the mainshock. Almost a million people, prepared in advance by a public education campaign, evacuated their homes and factories in the hours before the quake. Although many towns and villages were destroyed and several hundred people were killed, it appears that many were saved. The very next year, however, an unpredicted earthquake struck the Chinese city of Tangshan, killing more than 240,000 people. Obvious precursors such as those seen in Haicheng have not been repeated in subsequent large events.
Although many schemes have been proposed, we have not yet found a reliable method of predicting earthquakes minutes to weeks ahead of time. We cannot say that short-term earthquake prediction is impossible, but seismologists do not expect that it will be feasible in the near future.
We do have some useful guidelines about how the earthquake probabilities change over time. We know that earthquakes tend to cluster together in both space and time—for example, large earthquakes have nearby aftershocks—and seismologists have shown how the chances of a potentially damaging earthquake tend to go up during periods of increased seismic activity. Interpreting this type of information can be tricky, however, because, even when the seismic activity is high, accurate predictions of large earthquakes are still not possible. During seismic crises, it is easy for the public to become confused about how the hazard is changing. For example, a miscommunication of short-term earthquake probabilities before the damaging L’Aquila earthquake of April 6, 2009, led to the criminal prosecution of scientific advisors to the Italian government on charges of manslaughter (see Earth Issues 13.3). Forecasts based on earthquakes clustering are now being deployed to help Italians assess how the seismic hazards are changing. These short-term forecasting methods are also being developed in other regions, including California.
376
Earth Issues: 13.3 Italian Scientists Convicted of Manslaughter for Miscommunication of Risk Before 2009 L’Aquila Earthquake
On April 6, 2009, a magnitude 6.3 earthquake devastated the mountain city of L’Aquila, Italy, killing 309 people, injuring more than 1500, and leaving tens of thousands homeless. In the wake of this disaster, a local prosecutor indicted the vice-director of the Italian Department of Civil Protection (DCP) and six scientific advisors from Italy’s Major Risk Commission, a high-level advisory body, on charges of criminal manslaughter for statements made before the earthquake.
The case quickly became a cause célèbre among scientists. The indictments appeared to blame the scientists for not alerting the local population of an impending earthquake—for a “failure-to-predict.” It is well known that large earthquakes cannot be accurately predicted in the short term. Why would an Italian court punish scientists for not doing something they didn’t (and still don’t) know how to do?
Scientific organizations from around the world sent letters of protest to the Italian president. Nonetheless, after a year-long trial, an Italian court found all seven guilty as charged; it sentenced them to six years in prison and levied fines totaling more than 10 million euros.
So what really happened in L’Aquila?
Seismic activity in this part of Italy increased in January 2009. A number of small shocks, part of a “seismic swarm,” were widely felt and prompted school evacuations and other preparedness measures. In February and March, media coverage was inflamed by a series of earthquake predictions issued by a L’Aquila resident named Gioacchino Giuliani, who worked as technician in a national physics laboratory. These predictions had no official auspices and turned out to be false alarms, but they were widely reported by the media and caused some people to panic and evacuate their homes.
Government scientists responded to this chaotic situation by stating that there were no accurate methods for earthquake prediction, that earthquake swarm activity was common in this part of Italy, and that the probability of substantially larger earthquakes remained small. But these assurances did not dispel public concern caused by Giuliani’s continuing predictions, so the government hastily convened its Major Risk Commission in L’Aquila on March 31. The commission concluded that “there is no reason to say that a sequence of small-magnitude events can be considered a sure predictor of a strong event.” This statement was scientifically correct—most seismic swarms in Italy do not lead up to a much larger earthquake—but it underplayed a fact accepted by most seismologists: the chances of larger earthquakes do increase during a swarm.
At a press conference following the meeting, the DCP vice-director, who was not a seismologist, said that “the scientific community tells us there is no danger, because there is an ongoing discharge of energy. The situation looks favorable.” This statement was not scientifically correct, because, even during an intense seismic swarm, small earthquakes cannot relieve the regional tectonic stress that leads to large earthquakes (see the Practicing Geology exercise at the end of this chapter).
The tremors continued into April, prompting more school evacuations. Shortly before 11 p.m. on April 5, just a few hours before the mainshock, a strong, magnitude-3.9 earthquake shook the city. In an interview in Nature Magazine, Vincenzo Vittorini describes how he debated with his wife and his terrified nine-year-old daughter whether to spend the rest of the night outside—a customary response to seismic activity in this part of Italy. Recalling official statements claiming that each shock diminished the potential for a major earthquake, he persuaded his family to remain in their apartment building. The building collapsed in the mainshock at 3:32 a.m., and his wife and daughter and five others were killed. Nearly everyone in L’Aquila, including the prosecutor, lost relatives or friends. Tragic testimony like Vittorini’s constituted much of the prosecution’s case, which charged that the Major Risk Commission had given “incomplete, imprecise, and contradictory information about the nature, causes, and future developments of the seismic hazards.”
With hindsight it is clear that the Italian scientists got trapped by a simple yes-no question, “Will we be hit by a major earthquake?” From what the scientists could have known a week before the earthquake, a big shock was not very likely, probably less than a 1-in-100 chance. Even so, seismic activity had increased the probability of a large earthquake above the long-term average—large earthquakes are more likely during seismic swarms than in times of no seismic activity. Distracted by Giuliani’s predictions, the authorities did not emphasize this increase in hazard, nor did they focus on advising the people of L’Aquila about preparatory measures warranted by the seismic crisis. Instead, they tried to calm the population by making reassuring statements that were widely interpreted as a firm prediction: “no major earthquake will occur.”
Few scientists would argue the merit of prosecuting public servants who were trying in good faith to protect the public under chaotic circumstances. With hindsight the failure of the defendants to highlight the increased hazard may be regrettable, but the inactions of a stressed risk advisory system and misstatements by nonscientists representing that system can hardly be construed as criminal acts on the part of individual scientists. The L’Aquila verdicts are currently under appeal. One can only hope that judicial sanity will prevail.
A few weeks after the L’Aquila disaster, the government appointed an international panel of experts, which one of us (THJ) chaired, to suggest guidelines for improving earthquake forecasting procedures in Italy. Our report reaffirmed the infeasibility of high-probability earthquake prediction by any known method and addressed how short-term forecasts—in which the probabilities of large local shocks are invariably low—could be publicly utilized. Authoritative statements about what is, and is not, known about the current hazard are needed to educate the public and fill information vacuums that lead to informal predictions and misinformation. Alert protocols should be standardized to facilitate decisions at different levels of government, based in part on objective analysis of costs and benefits but also on the less tangible aspects of value-of-information, such as gains in psychological preparedness and resilience.
Our review found the Italian system wanting, but we could point to no country where operational earthquake forecasting was done much better. Regions that face high seismic risk can learn lessons from L’Aquila. Among them is the need to separate the role of science advisors, whose job is to provide objective information about natural hazards, with that of civil decision makers, who must weigh the social, economic, and political benefits of protective actions against the costs of mistakes. The L’Aquila prosecution has misconstrued these roles.
Medium-Term Forecasting
Uncertainties in long-term forecasting can be reduced by studying the behavior of regional fault systems. One strategy is to generalize the elastic rebound theory. The simple version of the theory depicted in Figure 13.3 describes how the tectonic stress that builds steadily on an isolated fault segment is released in a periodic sequence of fault ruptures. However, as we have seen in the case of Southern California (see Figure 13.18), faults are rarely isolated. Instead, they are connected to one another in complex networks. Thus, a rupture on one fault segment changes the stresses throughout the surrounding region (see Figure 13.4). Depending on the geometry of the fault system, these changes can either increase or decrease the likelihood of earthquakes on nearby fault segments. In other words, when and where earthquakes happen in one part of a fault system influences when and where they happen elsewhere in the system.
377
If Earth scientists can understand how variations in stress raise or lower the frequency of small seismic events, they might be able to predict earthquakes over time intervals as short as a few years, or maybe even a few months, although still with substantial uncertainties. Monitoring of such events on networks of seismographs could then provide a regional “stress gauge.” Someday you might hear a news report that says, “The National Earthquake Prediction Evaluation Council estimates that, during the next year, there is a 50 percent probability of a magnitude 7 or larger earthquake on the southern segment of the San Andreas fault.”
378
The ability to issue such medium-term forecasts would raise some difficult questions, however. How should society respond to a threat that is neither imminent nor long-term? A medium-term forecast would give the probability of an earthquake only on time scales of months to years—not precisely enough to evacuate regions that might be damaged. False alarms would be common. What effect would such predictions have on property values and other investments in the threatened region? These questions would have to be addressed by both policy makers and scientists.