The Disaster Profiteers: How Natural Disasters Make the Rich Richer and the Poor Even Poorer (8 page)

Of all natural disasters, earthquakes (including those that induce tsunamis) and cyclones do the most damage, much more than landslides, lightning strikes, volcanic eruptions, and most floods.
8
Volcanic eruptions have a distribution much like earthquakes because they too result from plate boundary processes, but, with notable exceptions, they rarely cause damage on the scale of earthquakes.
9
None of the other disasters has the same distinct, and distinctly different, geographic patterns. Here we focus our discussions mostly on cyclones and earthquakes.

In some settings, droughts and floods may be very damaging and economically very consequential. Droughts are the leading natural disasters in poor countries, which typically have economies dependent on agricultural exports, and they can have long-term effects. In contrast, most settings can recover fairly quickly from the effects of floods. Floods, as I noted before, may have positive economic effects in the years that follow. For example, among other factors, floods can recharge groundwater reserves, which can benefit crop production in the following year or years by providing ample water for irrigation.

Science, mainly generated in the rich world, can tell us exactly why earthquakes happen in some places and not in others. It can also tell us how likely it is that earthquakes of different magnitudes will occur in different places. Happily, very large–magnitude earthquakes
don't occur often, but very tiny earthquakes occur all the time. For example, about a million magnitude 2 earthquakes occur every year. No one but seismologists knows that because they are rarely felt and cause little or no damage. Only about 20 magnitude 7 earthquakes occur yearly, and most of those are sufficiently far from populated areas that they also are not usually felt.
10
Truly huge earthquakes, around magnitude 9, are mercifully very rare. In the last 100 years or so, there have been only five earthquakes at the top end of the magnitude scale—two in Chile (1960 and 2010), one in Indonesia (2004), one in Alaska (1964), and most recently the 2011 event off Japan. The storied 1906 earthquake in San Francisco, at magnitude 7.8, wouldn't make it into an earthquake hall of fame.

Remarkably, a simple mathematical expression describes exactly how often small and medium-size earthquakes happen compared to the biggest ones. In fact, all earthquakes of all sizes scale to one another. The expression is called the Guttenberg-Richter relationship (the same Frank Richter who gives his name to the commonly used magnitude scale) and is in a class of formulas called power law relationships. The Guttenberg-Richter relationship tells us in mathematical form what we know already—that very large earthquakes don't happen very often—and it gives a quantitative measure of how often we can expect very large quakes to occur based on how often smaller ones occur. An extraordinary number of phenomena can be described by power law relationships. For example, if we know how many small cities a country has, we can reasonably estimate the number of large cities it has, because we know, statistically, that for every big number X of small cities, there is a smaller number x of large cities. The same is true for earthquakes. The surprising thing is that there is such a rigorous and dependable relationship between small and large earthquakes. The distribution is not random, with the most common value lying in the middle of the distribution, like people's height.

Science has a very good understanding about why earthquakes occur where they do but much less understanding of why magnitudes
should
obey a power law function. And the goal of predicting exactly when a large earthquake will happen and exactly where within the elongated bands of earthquakes has eluded science. Most seismologists have serious doubts that this goal will ever be met.

Meteorology provides an understanding
of why cyclones occur where they do and why they exhibit distinctly curved tracks that look a little like fish hooks. The prediction challenge here is to calculate the track a cyclone will take once formed and what strength it will attain. The critical issues to predict are where the cyclone will make landfall, what its sustained wind speed will be at that time, and the size of the storm surge along the coast so people can be warned and evacuated from areas most likely to be impacted. Storm surge height and maximum wind strength are related, but not in a strictly linear way; seafloor topography and coastal features like bays and inlets influence storm surge height as well. In most parts of the world, storm surge heights can be predicted reasonably well from wind speed data. In wealthier parts of the world, where the seafloor structures have been mapped, very accurate and detailed predictions are possible. Storm surge, rather than strong wind, is responsible for most cyclone-related damage and deaths. Winds are responsible for power outages, most often caused by trees falling on power lines.

Meteorology can explain the geographic distribution of cyclones and why they are sometimes very large and don't amount to much at other times. Cyclones almost always start off traveling from east to west, whether in the Northern or the Southern Hemisphere. The Earth rotates toward the east, and cyclones are obliged to travel in response to the rotation.

In the Atlantic Ocean basin, storm tracks usually look as if they originate from about the same place, a region just off the west coast of Africa. They don't all form there, however, especially in the early months of the hurricane season, when they form more often on the western side of the Atlantic. Hurricanes form where very warm air and very warm water combine; these conditions are present off the Atlantic coast of Africa when extremely hot winds blow off the Sahara and ocean waters are unusually warm. Those conditions create intense evaporation; clouds form, and a disorganized collection of small storms emerges. As they grow, these storms begin to be influenced by a natural phenomenon known as the Coriolis effect, which adds a spin to their motion.
11
The several smaller storms then build into one large storm that can eventually become a hurricane.

To keep growing, the storm needs warm water in its path as it works its way westward on the trade winds. Tropical oceans provide that water. If an Atlantic hurricane ventures into the Gulf of Mexico, it can quickly grow in strength because the waters of the Gulf are fairly confined and often very warm. The overall path of a hurricane—a westward initial track that gradually hooks north—is also a consequence of the Coriolis effect. A hurricane is influenced by the patterns of atmospheric pressure as well. The Bermuda High, for instance, is a quasi-stable feature in the Atlantic during hurricane season that essentially pushes or holds hurricane tracks to the south of where they would otherwise go; hurricanes head north only when they have escaped the influence of the Bermuda High. Superstorm Sandy's somewhat unusual track—taking a sharp landward turn after heading northward well offshore—was the result of a high-pressure ridge that blocked its northward track.

Hurricanes diminish in strength as they head northward because they encounter colder and colder water. When they make landfall, they tend to peter out fairly quickly as they are deprived of warm
water. They can still dump a tremendous amount of rain while traveling inland; freshwater flooding is common. In many wealthy countries today, inland flash flooding causes almost as much loss of life as storm surge because storm surge warnings are usually accurate and people can evacuate the affected areas. However, flash floods, as the name implies, develop very rapidly and much less forewarning is possible.

Just as there are places that seldom experience earthquakes, there are also places that seldom experience cyclones and others that experience more and stronger cyclones than elsewhere. South America hardly ever suffers from cyclones, and most of Africa is spared except for Mozambique and Madagascar. The whole South Atlantic is spared cyclone activity, as is the very middle equatorial belt.
12

The equivalent of the Richter scale for cyclone strength is the Saffir-Simpson scale, a five-level scale based on wind speed.
13
At category 5, sustained winds may exceed 157 miles per hour (mph). Wind speeds in Typhoon Haiyan in the Philippines in late 2013 had sustained winds of nearly 190 mph; Haiyan was one of the strongest storms on record. The strongest cyclones on record formed in the Pacific Ocean, where they are called typhoons. They can become so strong because they have such a large expanse of warm ocean to cross, which gives them plenty of time to build in strength. They often cross the Philippines, losing some power, then build again as they head for China. Because the Atlantic and Indian oceans are much smaller than the Pacific, cyclones that form there typically do not reach the same size.

The understanding of tropical cyclones is such that in most parts of the world and at most times, it is possible to warn of their approach in sufficient time to allow for societies to prepare in a way that minimizes fatalities, primarily through evacuation. In Bangladesh, where evacuation would be difficult and communications are imperfect, permanent cyclone shelters have been constructed. These are very simple places of refuge built on strong pilings that are taller than the
level of the highest historic storm surge. They proved their worth in 2007 during Cyclone Sidr; although 3,000 people were killed, many lives were saved. In comparison, Cyclone Bohla in 1970 is thought to have taken 500,000 lives in Bangladesh, making it one of the deadliest natural disasters of all time.

No warning is possible for earthquakes, although science does have a very good understanding of the type of quake that gives rise to tsunamis and how tsunami waves propagate across the ocean. Very large earthquakes beneath the ocean floor off Japan, Indonesia, and elsewhere can cause tsunamis because the quake rapidly displaces the ocean floor, causing the ocean waters to move too. Tsunami heights usually are predicted quite accurately, although, as with storm surges, local seafloor topography and coastal features are important factors and are not always known in detail. Warnings can be made up to several hours in advance of a wave's arrival, depending on the distance of ocean it crosses, and this can save lives. The survival strategy is simple in principle: move to high ground as quickly as possible. But, of course, these warnings cannot save capital assets, most of which are immovable and difficult to protect.

What makes earthquake prediction so hard and cyclone track prediction relatively easier? One way to think about it is in terms of what scientists would call realizations: the number of repeat events. Almost every year, somewhere in the world a cyclone will form and reach the maximum strength on the Saffir-Simpson scale. Due to cyclone frequency, over a period of just a few decades, a data set can be collected that covers all possible cyclone strengths and all parts of the oceans where cyclones occur. In fact, cyclone records have been kept for more than 100 years, and there are thousands of tracks—realizations—stored for analysis. Hundreds of these tracks are of category 5 cyclones.

Because the largest earthquakes happen so infrequently, we have only a handful of realizations.
14
Although that is lucky for human societies, it is unlucky for scientific analysis. Oddly enough, those few that have occurred affected wealthy to middle-income countries—the United States, Japan, Chile (twice), and Indonesia. Luckily, very few historic large earthquakes have occurred in very poor countries. It's hard to imagine what would have happened in Haiti had the 2010 earthquake been magnitude 9 rather than the much smaller 7.0. From a statistical standpoint, the data set of earthquakes is sparse at the top end. To have a suitable data population that includes very large quakes, we would need seismograph recordings for thousands of years. The modern seismograph network is only a little more than 60 years old.

The fundamental reason why earthquake prediction is extremely difficult is that it requires predicting several factors at once: the time when a rupture will take place, on which of the many thousands of faults the rupture will occur, how much of the fault will rupture, and how much the shift will displace the crust. It would also be good to know how deep in the Earth fault rupture will occur, because for an earthquake of any given magnitude, the shallower the rupture, the more ground motion occurs and the more damage potential there is. And the ability to predict relies heavily on knowing the current state of stress in Earth, and that is something we really don't know in enough detail to be very helpful.

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.)