Attempts have been made to track large earthquake occurrences into the distant past using historic information. Both the Japanese and the Chinese keep detailed records of the damage done by major earthquakes, and these records, plus descriptions of ground shaking, can provide estimates of what is called the “recurrence time” of large earthquakes. The devastating Tohoku earthquake and tsunami of 2011 was at first thought to be unprecedented in magnitude for the region at 9.1, but after examining the historic records, seismologists found evidence of a very similar event 1,000 years earlier. Just how helpful that is seems somewhat debatable: How should a society prepare for a once-in-1,000-year event?
Another reason earthquake prediction is so difficult arises from the very nature of quakes. They occur when Earth’s outer crust, or shell, fractures. Forces that arise beneath the crust (mostly due to the great heat of the inner Earth) constantly exert themselves on the planet’s rigid outer shell. When these forces overcome the strength of the crust to resist movement, a fracture occurs to relieve the stress. These fractures occur mostly at weak points in the crust. Faults are some of those weak points—places that have already failed before. Like a dinner plate with a crack in it, chances are that if the crust breaks, it will find a fault to break along.
Large, active faults are often very easy to see in the landscape, as shown in the photo in figure 2.6, which is a satellite image of the Enriquillo-Plantain Garden Fault, which moved in 2010 in Haiti. (Actually, a splay, or secondary offshoot, of this fault moved, not the main fault itself.)
Figure 2.6
“Quake Aftershock Damage Seen by NASA Satellite,” a NASA Image of the Day. The image, taken on January 21, 2010, reveals damage from the aftershocks of the quake. The Enriquillo-Plaintain Garden Fault line, indicated by the black arrow, is highly visibile in this image.
Source: NASA/GSFC/METI/ERSDAC/JAROS, and U.S./Japan ASTER Science Team.
The Haitian capital of Port-au-Prince is located on the natural harbor to the right in the image. A heavy black arrow guides your eye to the location of the fault. Once you know where to look, the fault is quite obvious. It runs just south of the city. There appears to be a road running almost east-west on top of a ridge that is part of the expression of the fault. The San Andreas Fault in California is similarly easy to see at this scale. People without training in geology would not recognize these features at ground level, and the danger they pose is not especially evident. A fault that has not moved in hundreds of years is even more difficult to recognize from ground level.
Geologists, however, do know the locations of almost all faults. Even in very poor places like Haiti, where few, if any, local scientists are mapping faults, their pattern is usually reasonably well known because at some point a colonizing power sent geologists to assess the land’s potential for minerals and other raw materials. Faults appear on maps made in the earliest years of the British colonization of Africa and India.
It is very difficult to forecast when a fault will slip to generate an earthquake because we don’t know much about either the state of stress in the Earth or how much stress faults can endure before rupturing.15 Weather forecasts use computer models to see forward in time, and these require data that describe the state of the atmosphere now: atmospheric pressure, wind speed and direction, temperature, humidity, and many other factors. These data have been acquired over a long period of time from satellites, weather balloons, and ground-based measurements. It is fairly easy to interrogate this long record to build up an understanding of how weather systems develop and to use that understanding to make predictions.
The same idea applies to forecasting what will happen in Earth’s crust in the future, but we don’t have detailed measurements of what is going on in the crust beneath us. So seismologists are like weather forecasters before there were satellite measurements or even weather balloons. If we could somehow measure stress weather—the buildup and movement of stress in a broad region around a fault—we could reconstruct the conditions that evolved in Earth as a fault gradually inched toward rupture. But getting these data would require thousands of boreholes embedded with thousands of stress-measuring devices and continuous measurement of their output before and after every earthquake for years to come. Then maybe we would have a better sense of what happens in the prelude to earthquakes and be able to make predictions. Even then several accurate predictions of damaging earthquakes would be required before a prediction scheme would be thought reliable and could be used operationally.
Right now, there are no reliable precursors—signals from Earth that warn it is about to rupture—though many people claim they had some sort of premonition of when an quake would occur. You may have heard of foreshocks and aftershocks, particularly the latter, which very commonly follow the main shock. Aftershocks represent the continued release of built-up stress, because the main shock doesn’t release all the stress. The magnitude of aftershocks decreases with time, and they become less frequent in a very predictable way.16 Those that happen soon after the main shock can have a magnitude near to that of the main shock and can be damaging.
Foreshocks are much more poorly understood. It might be reasonable to suppose that as a fault is stressed nearer and nearer to its moment of failure, it sends out some recognizable signals that we could read as warnings. But no such signals have been found yet. In fact, typically we don’t recognize foreshocks that have been recorded until after the main shock occurs. Earthquakes often come in swarms of tremors, many smaller quakes over a period of days or weeks, sometimes building in magnitude, sometimes not. It is essentially impossible to tell if a swarm of tremors is or is not the portent of a main shock. About 1 in 100 tremor swarms is a rehearsal for a main shock; in other words, even if a tremor swarm is recorded, the likelihood that a large earthquake is imminent rises by only 1 percent.
Misunderstanding the meaning of tremors was probably at the heart of a tragedy in L’Aquila, Italy, in 2009. In the early morning, a 5.8-magnitude earthquake—not very large—struck, causing 309 deaths. Earlier there had been a series of tremors that frightened the residents, many of whom chose to sleep outdoors. Earthquakes are at their most dangerous at night when people are asleep. Italy’s Major Risks Committee was called in to assess the situation. Part of what its members said was correct but unhelpful: They saw “no reason to suppose a sequence of small earthquakes could be the prelude to a strong event” and believed that “a major earthquake in the area is unlikely but can’t be ruled out.”17 That’s completely correct, but the committee did not tell the people of L’Aquila how they should behave. They left it up to the people themselves to decide whether to continue to sleep outdoors.
A week before the quake, one statement was reported that people did act on because it was much clearer. Dr. Bernardo De Bernardinis—deputy head of the Department of Civil Protection and acting as spokesman for a group of government scientists that included the six seismologists who formed the Major Risks Committee—is said to have told a member of the press that “the scientific community tells me there is no danger [of a large earthquake] because there is an ongoing discharge of energy.”18
The tragedy of the L’Aquila quake hasn’t ended. A year later, De Bernardinis and the six members of the scientific panel were charged with manslaughter on behalf of 29 residents who allegedly had taken those words of assurance to heart. All seven were convicted and faced significant jail sentences and large fines, not for failing to make an accurate prediction (which almost all seismologists think they were convicted of) but rather for doing a shoddy job in assessing the situation and making the public aware of the risks. In effect, they were convicted of professional malpractice.
On appeal, six committee members had their convictions overturned; De Bernardinis was only partially exonerated. His sentence was reduced but not overturned.
The take-away message has at least two parts. First, it is important to get the science right. De Ber
nardinis was close. He may have had in mind the fact that sections of some faults creep, moving as we hope all faults would do: little by little, releasing stress in small packages instead of one large event.19 Second, accurate and clear communication is critical. Most people have a very difficult time assessing risk and expect experts to spell out clearly what risks they face. In this case, De Bernardinis’s statement was clear but wrong.
It does, however, have a ring of plausibility to it. It is easy enough to imagine that the bit-by-bit release of energy could indeed be making things safer. I introduced the idea of scienciness earlier. De Bernardinis’s statement is an example of scienciness committed by a scientist that may have had deadly consequences.
A simple comparison of the two maps in figure 2.1 and the map of earthquake distribution in figure 2.4 shows that there isn’t an obvious relationship between past earthquake occurrences and present-day poverty and wealth. Afghanistan and Pakistan have high earthquake risks, as does Bangladesh. Northern India and the Himalayan countries are at high risk also, but it would be quite a stretch to suggest that earthquakes gave rise to these countries’ low levels of development or are currently inhibiting their development.
But that’s not really the point. Roger Bilham, a seismologist at the University of Colorado in Boulder, worries most about what might happen in the future and paints a very grim picture. In an influential paper titled “The Seismic Future of Cities,” he argues that rapid population growth, particularly in poorer parts of the world, will require vast additions of building stock that are unlikely to be seismically safe, given the history of building practices in those regions. He thinks the world will soon see a million-fatality earthquake.20
Bilham emphasizes that disasters are worse killers if they affect cities, and many cities in poorer areas are growing rapidly and often chaotically, with little, if any, attention paid to construction standards or enforcement. Bribery and corruption are rife. The 2010 earthquake in Haiti might serve to make the point. So too might the earthquake in the province of Sichuan, China, in 2008. Both of these rapidly growing regions are noted for the poor quality of their buildings, especially public buildings. In China, the tragedy was amplified because so many schools were destroyed; the fatalities among schoolchildren (and their teachers) were appalling. China even has a term to describe structures of such poor construction quality: they are “tofu dreg” projects, referring to the remains that are left over after making tofu.21 The term refers to both the building materials—such as cement building blocks in which the sand fraction is far too high, for instance—and the construction process itself, in which too few steel-reinforcing elements (rebar) might be used in columns.
The message here is simple. As the adage familiar to all seismologists and well known to the citizens of L’Aquila says, earthquakes don’t kill people; buildings kill people. Sometimes earthquakes trigger landslides that bury people, but the point is that, Superman movies notwithstanding, the ground does not open up and swallow people. Death comes when your home, workplace, or a shopping center collapses and crushes you. Buildings are more deadly at some times of day than others. For example, schools are more deadly during the daytime hours when students are inside.
In homes, earthquakes are deadliest at night, when people are asleep. A shaking home may wake residents, but when half-awake and in the dark, people have trouble thinking quickly enough to find somewhere safe to shelter or to get out into the open. The L’Aquila earthquake struck at 3:32 in the morning; most victims were found in their beds, crushed to death by their collapsed homes.
The amount of shaking needed to cause a building to collapse depends on the integrity of the building itself and of the surface on which it was built. In vast sections of blighted areas of Detroit and other US cities that have fallen on hard times, the buildings often are partly collapsed. Water has invaded the structures, causing critical beams to rot and be unable to bear the weight of roofs and walls. The Rana Plaza building in Bangladesh collapsed on its own in 2013 without any earthquake, killing 1,129 people, mostly female garment workers on the four upper floors—unauthorized additions to the original building, which consisted of a bank, stores, and apartments on the lower floors. Cracks in the building the previous day had caused all but the garment factories to be evacuated. The garment workers were forced to work under threat of being fired or having their pay withheld.22 After a power outage in the morning, a diesel generator was started up on an upper floor. Soon after the building collapsed. The upper floors of the building were so badly built and contained so much heavy machinery that the vibration of the generator may have been enough to shake it into rubble.
A contributing factor in the collapse was the ground on which the building was constructed. The area included a pond that had been roughly filled in. The building was partly located on the in-filled pond, so part of the structure rested on very soft ground and part on firmer ground. This is a recipe for serious problems. Part of the overweight building would have slowly sunk into the soft ground while part would have resisted. The cracks that opened in the building the day before its collapse were most likely the result of differences in the underlying support that caused parts of the building to subside while other parts stayed firm. Rana Plaza was ready to topple at any moment.
To get a sense of how important ground conditions are in determining the highly variable nature of damage that can be done by an earthquake, look at the image in figure 2.7.23
Figure 2.7
Evidence of liquefaction during an earthquake in Christchurch, New Zealand, in 2011.
Source: NZ Raw. Photograph by Mark Lincoln of marklincoln.co.nz. Used with permission.
In the figure, it looks like the back end of a car has been sawn off at a bizarre angle and dumped on the ground, but the car has actually sunk into the ground. On the window, you can actually see a high-water mark, suggesting it was even more deeply submerged at one point. You might think someone had the misfortune of driving into a deep pool of mud, but the mud was actually created by shaking during the Christchurch earthquake in New Zealand in 2011. The process, called liquefaction, commonly happens when very soft, wet soil is shaken during an earthquake. Liquefaction typically is seen in portside areas that have been built out with landfill. Much of the Marina District in San Francisco and the southern tip of Manhattan were formed that way using material excavated for the foundations of tall city skyscrapers. Houses in the Marina District were all but submerged in the 1906 earthquake and again during the 1989 Loma Prieta earthquake.
The image in figure 2.8 appears in almost every textbook description of liquefaction. It shows a group of buildings in Niigata, Japan, that look identical in construction type. They all experienced the same earthquake on June 16, 1964, but one building is flat on its side, one is tilted half over, another is tilted just a little, and others remain upright. Soil liquefaction is the reason for the earthquake’s variable effects, which are obviously very localized. It wasn’t bad luck that the one building toppled; it was located on especially weak soil.24
There are many causes for the very uneven distribution of damage in most earthquakes, but when a fully intact building topples, as shown in figure 2.8, it is a sure sign that ground conditions were the issue.
Figure 2.8
Liquefaction during an earthquake is illustrated in this photo of the aftermath of a quake in Niijata, Japan in 1964.
Source: U.S. Geological Survey
Look back at figure 2.7, the photograph of the sunken car. In the background, next to an apparently undamaged house, is a car in a driveway that is in perfectly good shape, sitting high on what must be solid ground. The fact that it is perhaps 30 yards away from the vehicle in the foreground indicates that the effects of liquefaction and soil conditions in general can be highly localized.
Just as scientists can map faults, they understand how to assess ground conditions. The equipment required is neither
expensive nor difficult to operate. Operators do not require high-level training. So like faults, ground conditions can be mapped in almost any setting around the world.
When earthquake waves propagate through the Earth, they shake different types of ground by varying amounts with distinct results. Even if the ground doesn’t fully liquefy, it responds very differently according to the type of rock layers, their age, the extent of weathering, and many other factors. Very loose soil shakes more than very solid ground.25 This variable effect often accounts for much of the difference in damage from place to place. Liquefaction is an extreme expression of this effect.
Topography makes a difference too. Just as the shape of a glass lens in a telescope or magnifying glass focuses light energy, the shape of Earth’s surface and the rock layers beneath can focus seismic energy. Sharply peaked ridges, for instance, can show particularly severe fracturing (sometimes described as shattering by geotechnicians). Small depressions of softer rocks can cause seismic energy to reverberate and amplify, causing more severe local shaking than in surrounding areas.
The net result is that the damage caused by an earthquake can appear capricious, with completely collapsed buildings next to others that show little if any damage. Older buildings may remain in good condition while newer ones are severely damaged. Of course, more damage might be expected in poor countries where governance is weak and/or corruption levels are high. But good governance and low corruption are no guarantees of suffering less damage. New Zealanders were fairly astonished when the relatively modern Canterbury Television Building collapsed, causing more than half the total deaths in the 2011 quake. It had been inspected several times after earlier earthquakes and was found to be in good condition. The construction was sound and up to code. The ground was solid and suitable for the size of the structure. No one was bribed. No one cut corners. No one cheated. Yet the building collapsed. After a royal commission investigation, no one was charged.
The Disaster Profiteers: How Natural Disasters Make the Rich Richer and the Poor Even Poorer Page 7
