Catastrophes—and our poor judgment
of their true risks
Civilization exists by geologic consent, subject to change without notice
For days now, the southeastern United States has been pounded by record tornado outbreaks. Just over a month ago, Japan experienced the third largest earthquake ever recorded by seismographs, followed by a devastating series of tsunamis that may have killed tens of thousands. Add this to the major recent earthquakes in Chile, New Zealand, and Haiti, or Hurricane Katrina in 2005, or the Boxing Day tsunami in the Indian Ocean in 2004, and the public is more aware than ever of the hazards of living on Planet Earth.
But before we overreact or act irrationally in fear of earthquakes or other natural disasters, we must step back a bit and put them in perspective. as I have done in my new book Catastrophes! Earthquakes, Tsunamis, Tornadoes, and other Earth-Shattering Disasters. Which ones cause the most damage? Which ones are the most dangerous to us in the short term? Which ones are dangerous in the long term? What things are most likely to kill you? What kinds of things are capable not only of hurting us as individuals, but destroying human civilization?
Psychologists have shown that human beings are notoriously poor at judging relative risks, and assessing which threats are really serious and which ones are exaggerated. For deeply-held psychological reasons, people are far more afraid of dying from a snake bite or in an earthquake, even though these are staggeringly improbable events for most people in the United States. Only 5 to 10 people die of snakebite each year. In the United States over the past century, earthquakes have killed an average of only 6 people per year. Yet because earthquakes are unpredictable and shatter our notion of “terra firma,” we are unjustifiably afraid of them. Because snakes trigger a primordial fear response in our brain, we are terrified of them. When we were small vulnerable hominids running across the African savanna, snakes were a real threat to us, because many African snakes, like mambas and cobras, are poisonous. But now that snakes are so heavily slaughtered in this country (despite the fact that most American snakes are non-poisonous), we are much more a threat to them than they are to us.
A more objective way of assessing real threats is to look at cold hard statistics, as an actuary or insurance adjuster does. A 2008 article by Borden and Cutter looked at deaths in the U.S. from all natural hazards from 1970 to 2004. Despite the fact there were several big California earthquakes (1971, 1987, 1994) and large hurricanes during that time window, you would never guess what the number one killer was. It was not even a topic that we think of as catastrophe, since it happens so often and so slowly. The top killers among natural hazards in this country are … heat waves, storms, and winter!
The ranked list of killer hazards and the percentage of the roughly 20,000 that died over that time window is:
- Heat and drought (19.6%)
- Severe weather (18.8%)
- Winter weather (18.1%)
- Flooding (14.0%)
- Tornado (11.6%)
- Lightning (11.3%)
- Coastal hazards, like drowning (2.3%)
- Hurricanes and tropical storm (1.5%)
- Earthquakes (1.5%)
- Landslides (1.9%)
- Wildfires (0.4%)
Yes, earthquakes, hurricanes, and other extreme weather events are terrifying disasters, but the biggest killers are slow and subtle: heat waves and drought. Likewise, we take severe storms and the bitter cold of winter for granted since they happen so often, but they kill a lot more people than more dramatic events like tornadoes, hurricanes, and earthquakes. Notice that hurricanes, earthquakes, and landslides are near the bottom of the list with less than 2% of the total deaths. Even though hurricanes and tornadoes are potentially very dangerous, we usually have some warning of when they are coming, and most people take shelter or evacuate when warned. Volcanic events did not even make the list, since the small Mt. St. Helens eruption was the only deadly volcanic event in this country for better than a century.
Borden and Cutter also plotted the risk on a map of the U.S., organized county by county. If you asked anyone where the most dangerous places to live in the U.S. are, they would probably guess California with its earthquakes and landslides and brush fires. But as the map shows, the opposite is true. Coastal California was one of the least hazardous in this regard (because we seldom get extreme killer weather of either hot or cold variety). The most deadly regions turned out to be the Deep South, where severe heat and humidity is common, and where occasional hurricanes and tornadoes only add to the carnage. Also dangerous was “tornado alley” in the southern Plains, with heat, drought, and tornadoes, and the southern Rocky Mountain region, with its desert heat and flash floods. The northern Plains and Rockies and Midwest were also death “hot spots” due to extreme cold and drought and occasional flooding. The rest of the country did not show any striking trends one way or another. What you don’t see is any strong correlation of high death risk with the fault zones map or even with the Gulf Coast-Florida hurricane zone.
Let’s put that in an even broader perspective. Many people are terrified of earthquakes and tornadoes and hurricanes, but these events are not something to lose sleep over except when there are clear warnings that a hurricane or tornado is coming. We should be more careful and worried about heat waves and severe winter storms, but we’re so accustomed to these each year that we don’t realize how deadly they are. But worrying about natural disasters looks absurd in the face of where the real risks come from: your cheeseburger and French fries, your car, cigarettes, and all sorts of things you encounter every day. Borden and Cutter point out that for the 20,000 people killed in the U.S. by natural disasters during the study period from 1970 to 2004, there were 652,000 deaths from heart disease alone (more than 30 times the natural disaster total)! There were 600,000 deaths from cancer (also 30 times the total from natural disasters). Of cancer deaths, almost a third were from lung and other cancers due to smoking. Colorectal cancer, pancreatic cancer, prostate cancer, and breast cancer were the other biggest killers. There were 143,000 deaths from stroke, 130,000 from chronic lower respiratory diseases (bronchitis, pneumonia), and even 117,000 killed in accidents (mostly car accidents). If we really took the issue of risk seriously and evaluated it objectively, we would do well to worry much more about our diet and exercise, get frequent health checkups, stop smoking, and modify our driving habits. We may fear death in an earthquake or hurricane, but lunch, cigarettes, and driving are much more deadly to you!
A Global Perspective on Natural Disasters
The story changes significantly when we look at natural hazards on a worldwide basis. In the U.S., the quality of our building construction in seismic zones, and our relatively good health care and emergency services, means that even big earthquakes result in relatively little loss of life. In many underdeveloped countries like Turkey, Iran, Armenia, Pakistan, Haiti, and China, and even European countries like Greece and Italy, earthquakes have a much higher death toll. The single biggest factor is the construction of their buildings. Most are built of simple stones or bricks held together by mortar, known as “unreinforced masonry”. Those buildings are death traps in even a mild earthquake, since they shake apart and then collapse with tons of weight on the trapped inhabitants, killing thousands. By contrast, in seismically risky areas like California, most older brick buildings have already shaken down in earlier earthquakes, and the codes forbid any masonry except reinforced bricks, where steel tie rods and rebar is threaded through the holes in the cinder blocks to hold them together when shaken. For much of the Mediterranean-Alpine-southern Asia earthquake zone, however, the population is too poor to afford more expensive but seismically safer construction (and their governments do nothing to prevent it). Even as the survivors dig out their dead, they are rebuilding the same death traps with the same old bricks in the same way. In some cases, like the 2008 Sichuan earthquakes in China, or the 1985 Mexico City earthquake, buildings may be built of modern safer materials but still become deathtraps because of political corruption and shoddy construction.
The global list of the deadliest natural disasters (see below) also includes a lot more flood deaths than we would find in the U.S. Most of these killer floods occurred in China. Their deadliness was exacerbated by political weakness or ineptitude in providing disaster relief, plus the fact that China has huge populations of peasant farmers living on floodplains with no place to go in event of a disaster. The same could be said for the effects of big typhoons, cyclones, and tsunamis in the Philippines, Bangladesh, China, and Burma. Events such as the recent Cyclone Nargis in Burma were made far deadlier than necessary by the huge populations of poor peasants living on low-lying ground and vulnerable to big storms, and governments that can’t or won’t provide warnings or timely evacuation plans, or disaster relief in any large-scale meaningful way.
| Rank | Event | Location | Date | Death Toll (Estimate) |
|---|---|---|---|---|
| 1. | 1931 China floods | China | July–November, 1931 | 1,000,000–4,000,000 |
| 2. | 1887 Yellow River flood | China | September–October, 1887 | 900,000–2,000,000 |
| 3. | 1556 Shaanxi earthquake | Shaanxi Province, China |
January 23, 1556 | 830,000 |
| 4. | 1970 Bhola cyclone | Bangladesh | November 13, 1970 | 500,000 |
| 5. | 1839 India Cyclone | India | November 25, 1839 | 300,000 |
| 6. | 526 Antioch earthquake | Antioch, Byzantine Empire | May 20, 526 | 250,000 |
| 7. | 1976 Tangshan earthquake | Tangshan, Hebei, China |
July 28, 1976 | 242,000 |
| 8. | 1920 Haiyuan earthquake | Haiyuan, Ningxia-Gansu, China | December 26, 1920 | 240,000 |
| 9. | 1975 Banqiao Dam flood | Henan Province, China |
August 7, 1975 | 90,000–230,000 |
| 10. | 2004 Indian Ocean tsunami | Indian Ocean | December 26, 2004 | 229,866 |
Table 1. Top ten deadliest natural disasters in world history (excluding disease and famine)
Once again, however, we must put these natural disasters in perspective. Hundreds of thousands of deaths in an earthquake or cyclone or flood sounds terrifying until we stack them up against the true killers: disease, famine, drought, and other slower but deadlier agents. If the table listed the mortality rates due to all deadly world events, all top ten events would be diseases and famines, and no natural disaster would make the top ten. The drought in India that led to the Great Famine of 1876-1878 killed at least 25 million people, at least six to 25 times as many as any natural disaster on the list in the table. The 1918–1920 outbreak of Spanish influenza killed at least 20 million to as many as 100 million people, far more than any natural disaster in human history. The great bubonic plague or “Black Death” outbreak of the 1300s may have killed even more, possibly as many as 75 million to 200 million, but the death estimates are very uncertain.
The common diseases kill even more than the time-constrained “events” and “epidemics.” Although the medical world is justifiably proud of containing it, smallpox killed at least 300 million in the last century alone, and nobody knows how many humans have died of smallpox over the centuries. Measles killed more than 200 million over the last 150 years, despite the fact that it has been virtually eliminated in the U.S. Malaria killed 80-250 million people in the last century, even though there are lots of ways of mitigating its spread and deadliness. Tuberculosis killed 40-100 million people in the last century, and its death rates are increasing as the world’s skies become more polluted, and cigarettes become more and more popular outside the U.S. Even AIDS, which is a relatively young epidemic (spreading only since the 1980s) has killed more than 25 million people worldwide, far greater than any natural disaster.
We are impressed with large and terrifying events like tornadoes and landslides and blizzards, but as we saw with U.S. statistics, they don’t even rank. On a worldwide basis, many of the more terrifying disasters in history are pikers when compared to death tolls such as those caused by diseases and famines. The world’s deadliest tornado killed only 1300 in Bangladesh in 1989. The world’s worst avalanche killed 96 in the U.S. in 1910. The world’s deadliest blizzard killed 4000 in Iran in 1972. The world’s deadliest landslide killed 20,000 in Venezuela in 1999. The world’s worst wildfire killed about 2000 in the U.S. in 1871. The world’s biggest historic volcanic eruption, Mt. Tambora in 1815, was a bit more impressive with 92,000 deaths, but still this doesn’t come close to cracking the Top Ten list. And the great heat wave that fried Europe in 2003 was the deadliest ever with over 23,000 deaths, but that still does not make the Top Ten list, either. However much we are impressed by spectacular catastrophic events like volcanoes and tornadoes and landslides and blizzards, let’s not mistake the terrifying power of these events with the mortality they cause. Disease and famine are a lot slower and less dramatic, but still far greater purveyors of death than any rapid catastrophic natural event.
If you’re in Southern California on May 15 and have the time, come hear me speak about the topic at the last Skeptic Society meeting of the year, at 2:00 in Baxter Auditorium at Caltech.
A small correction: There was no big earthquake in California in 1987. You are thinking of 1989.
No, I mean the Oct. 1, 1987 Whittier Narrows earthquake here in Southern California–it was big enough (M = 5.9) to rattle buildings all over southern California, killed 8 people and caused over $350 million in damage. I remember it vividly–it woke me at 7:42 a.m. and I quickly set aside my planned lecture in physical geology to speak about earthquakes instead. It was certainly the biggest and deadliest event in the LA area since the 1971 Sylmar quake. You are thinking of the 1989 Loma Prieta quake in the Bay Area and Santa Cruz, which was not felt here.
Exactly. I was getting ready to go to work, and could hear the cracking noise as the S made coming up the L. A. River channel where it crosses under De Soto Ave (San Fernando Valley), and then had to slide down at a doorway to keep my footing (only a 5.9? I thought it was larger). Definitely not as big a shaker as the 1971, but remarkable.
It wasn’t felt “here”??? You are aware that the bay area is California, right?
Correction: The Bay area is IN California. Hopefully, your mistake was also a typo, and not a belief that the largest earthquake in a major metropolitan area in California since 1906, one that killed 63 people and did 6 BIllION dollars in damage(the most expensive natural disaster in US history at the time) doesn’t count as a large California earthquake, because anything outside LA isn’t really in California.
You’re reading all sorts of stuff into what Prothero did and did NOT say.
He said the 1987 quake was the biggest in SoCal since 1971. (I think Northridge passed it, since then.)
He said the Loma Prieta quake was NOT felt in SoCal.
He did NOT, anywhere, say the Bay area is NOT in California. Pull the cotton out of your cyber-ears. And, stop acting like anything outside the Bay Area isn’t really in California, to riff on your own words.
“He did NOT, anywhere, say the Bay area is NOT in California.”
Sure he did:
Quote one:Despite the fact there were several big California earthquakes (1971, 1987, 1994)
Quote two(defending quote one):you are thinking of the 1989 Loma Prieta quake in the Bay Area and Santa Cruz, which was not felt here.
Where exactly is the Socal only exclusion? I must have missed it. Because it seems to me that he said that a 5.9 earthquake that killed 8 people and did $350 million in damage ranks as a major California earthquake, but a 7.0 that killed 63 people and did $6 billion in damage doesn’t. He said the quake wasn’t felt “here”. But here, according to that quote, was California, not Socal only. How could you read it any other way? The only way I could see is if you are one of those people who feels the strange urge to fluff and suck up to anyone with even the slightest tinge of authorit….Oh, right. Nevermind.
You’re still wrong and digging the hole deeper, Grungor.
First, he never says his “here” is all of California. Seems pretty clear to me that, in context he meant SoCal. It’s easy to read it exactly the way I read it, and I’m sure others did.
Second, if you think you can identify me as a suck-up to authority after one comment, you:
A. Missed my “handle.”
B. Are a twit who needs to read Chris Mooney’s MoJo article about self-justifying “reasoning” by one’s self.
End of discussion from my end.
Putting risks in perspective, the safest way to travel is the space shuttle, because in 30 years and 530 million miles, only 2 shuttles carrying only 14 astronauts were lost.
In case you didn’t get that, if you’re sitting in a space shuttle, you worry about the space shuttle, and if you’re sitting on a big earthquake fault, you worry about earthquakes. Kansans generally don’t worry about earthquakes, and Californians don’t worry about tornadoes.
Though a tornado did hit Point Mugu near Malibu once. That place gets hit with everything: wildfires, earthquakes, mudslides. They must’ve made a deal with the devil.
2 out of 133 shuttle missions have ended catastrophically (not counting Apollo 13), about 1.5%. How does that compare to driving?
I’ve heard elevators are the safest form of travel. Although in Japan a few years ago there was a cluster of limbs and heads getting chopped off.
Hey, you’re right. Also, 2 out of 5 shuttles were destroyed. So what you’re saying is that the wrong statistics can make things appear safe when they really aren’t. Great point.
It would be more accurate to say that NASA has experienced a 1.5% mission failure rate. Mortality rate would be better calculated by the number of crew deaths versus successful returns.
Percentages are a great means of comparison by putting everything on the same scale, but the problem is often with the inputs used, which many seldom pay attention to, which is in turn abused by salespeople or sensationalists.
So how would one go about unifying the percentage inputs for natural disasters? Perhaps using the average number of people per square mile occupying the affected areas?
A few scattered thoughts– Another aspect of bad risk assessment that is destined to become inflated over coming years is that caused by the inflating human population. More people in a given area, especially in hazard-prone areas, will on average tend to lead to more deaths and greater media coverage because of them. The benefits of safety-improving technology should ameliorate this in some areas of course, but I expect to see the news media spend more and more time on disasters in our lifetimes. The hazards may not become more frequent or dangerous by some metrics, but simply because there are more people in the firing line they may seem more dangerous (and be more costly in absolute terms of human lives). Goes not just for natural hazards but mass murders and other crimes; more billions means more lunatics and malfeasants (unless their proportion in the population declines…).
Increased ability to detect and report on disasters, enhanced by globalization and the internet, will help spread that doom-laden message. Relative deaths (e.g. as % of population) could sometimes be a useful metric (especially for broad meta-analyses)– by that normalization, the 526 Antioch quake would truly be epic.
Mr. Prothero, you left out FEMA death camps. I don’t know about the commenters on your previous blogs, but here we expect thorough analysis from the bloggers, this isn’t amateur hour.
J/K sir, it was another great piece. Keep ‘em coming!
Geological and meteorological disasters are part and parcel of living on a planet. It’s not going to change, so get educated.
That’s the whole purpose of my book!
So explain the TSA’s job again? Billions to protect us against terrorist attacks that couldn’t work again anyway?
To protect against terrorist attacks that could work.
To protect us from terrorist underwear models, duh.
According to the CDC, the 600K rates for heart disease and for cancer are in one year, not over the multiple years counted for natural disasters. This makes the point even more emphatically so you might want to clarify.
Just adding more perspective for this excellent post. I recently generated this tweet:
“2010 US budget for defense – 700b. NIH – 32b. US terrorism deaths since 1968 – 50k, deaths from cancer in 2010 – >500k.”
First, you’re comparing the entire defense budget to one research agency of the DHHS.
Second, if I didn’t know any better, I could interpret these numbers as proof that the NIH is a waste of money.
Third, if you’re going to single out risks, there were zero US nuclear war deaths, so I guess it was never a threat.
So you don’t think the budget should reflect the impact of an issue? NIH is the largest source of funding for basic cancer research. Research that corps find to risky to do.
But I’ll bite. Why do you think funding NIH is a waste of time?
The NIH spent billions of dollars on cancer research, yet deaths from cancer remain high, so one might conclude that the spending was wasteful.
Liberals made the same accusation against the JIEDDO. It spent billions of dollars on defeating IEDs, yet IEDs still cause most coalition casualties.
You could try to estimate the number of lives saved, but that’s a lot trickier and harder to validate.
“The most deadly regions turned out to be the Deep South, where severe heat and humidity is common”
One minor correction here. If you look at the Hazard Induced Mortality by FEMA Region, you’ll see that heat/drought is more a problem in the Southwest and Midwest. Hazard mortality in the Southeast seems to link more to our frequent thunderstorms, with lightning, tornadoes, and “severe weather” being the prominent items.
Mr. Prothero:
Some of the death statistics quoted in your article do not ring true to me (and I see I am not the only one who is suspicious). Particularly, you implied strongly that “almost” 117,000 people died in car accidents in the U.S. from 1970 to 2004. But I checked the National Highway Traffic Safety Administration’s web site, and according to the statistics found there, over 413,000 people died in motor vehicle crashes in the U.S. just between 1994 and 2004, an average of over 37,000 per year. Perhaps the Borden and Cutter data you quoted was only for California.
Anyway, I don’t disagree with your main point. Some of the data needs to be cleaned up, however.
You’ll have to read Borden and Cutter carefully, and see exactly what they were citing. If they made an error, that’s between you and them. I’m just citing their statistics, not vouching for their accuracy. It’s just a blog–I don’t have time to check every detail!
Ouch! That comes off sounding like accuracy in blogging is not very important.I will give you the benefit of the doubt and chalk that remark up as a ‘shot from the hip’.
Thanks for the reply. I certainly intended no offense by pointing out the error, which is at least an order of magnitude. Also, I happened to already be familiar with the motor vehicle death rates in this country, as I have a teenaged daughter who I have tried to scare into compliance with my safe driving “suggestions”.
Well I can say that’s not a response I was expecting.
Where do they say anything about accidents and heart disease?
http://www.ij-healthgeographics.com/content/7/1/64
Where’d you get those numbers?
Going back to the 1970s, about 50,000 ppl per year died in car crashes, until it was cut to the upper 30,000s per year through air bags, tougher DWI laws and other things.
Your numbers are off, Don. Or theirs are.
It’s “teh Google.” Give it a whirl.
Yupp, that number seems impossibly low, that would mean only 3000 deaths due to car accidents per year, the true value should be around 10x as high, that number realy stands out.
I’m not sure all the data is that verifiable. Notably, the Antioch earthquake. Just knowing the population of a city in the ancient world with any degree of precision is dicey, let alone how many people may have died in an earthquake there.
The vast majority of deaths in developed countries are caused by old age, so that’s what we should REALLY worry about.
No, not really. Lives cut short are more tragic, so we should consider the age at death. A 25-year-old is most likely to die in a car accident, but cancer catches up by age 40. And people do fear cancer, which is why they fear radiation, and even avoid saying “the C word.”
Great stuff! I’m always trying to convince my friends that they worry too much about all the wrong things.
Donald, do you have earthquake insurance?
No, it’s ridiculously expensive in California, with outrageous deductibles, so the only way you could make a claim is if your house was totaled–and if that is the case, the insurance companies would be broke anyway and you’d never collect.
Exactly. See, they know the risk.
I still think people are justified to be more worried about earthquakes and hurricanes than drought or winter. If winter only lasted as long as the 2004 tsunami it wouldn’t kill anyone. If hurricane Katrina struck one area for months, as droughts do by definition, it would have been far more deadly. Apples and oranges.
The sad thing is that people did NOT worry enough about the tsunami and hurricane Katrina. They counted on their seawalls and levees to protect them, they built houses in low-lying areas, and they didn’t evacuate when they had the chance, thinking it’s a false alarm.
I don’t follow your reasoning here – in fact, I think you’ve got it backwards.
There are three factors here: frequency, intensity, and duration.
e.g. Harsh winters are annual in some places, quite long, not that intense.
e.g. Katrina-grade hurricanes are quite infrequent, brief, very intense.
Maybe a Katrina-grade hurricane could last months, but I suspect that it is wildly unlikely. So trying to defend against that is a poor use of your time and resources.
Short version of the above: take account of duration when you’re weighing up risks. Because “winter” lasts a long time, so you should count it as more dangerous, not less.