To find a Black Swan event, think about the unthinkable

Rare catastrophic events prove hard to predict, but catastrophist Gordon Woo posits a methodical way to think through just how bad future disasters might be.
 

By Alka Tripathy-Lang, Ph.D., science writer (@DrAlkaTrip)
 

Citation: Tripathy-Lang, A, 2020, To find a Black Swan event, think about the unthinkable, Temblor, http://doi.org/10.32858/temblor.141
 

 

On December 26, 2004, more than 750 miles (1200 kilometers) of the Sunda Trench ruptured off the western coast of northern Sumatra, causing the third largest earthquake in recent memory. This quake, and the catastrophic tsunamis it produced, killed more than 200,000 people around the Indian Ocean’s rim, with Indonesia recording the most deaths, followed by Sri Lanka and India.
 

 

This calamitous pair constitutes a Black Swan Event, a term originally coined by Nassim Nicholas Taleb in his book The Black Swan. Taleb threw down the gauntlet by defining a Black Swan as rare to the point of unfathomable, catastrophic in impact and explicable and predictable only in hindsight. A Black Swan is, in Taleb’s original framework, simply unpredictable. Gordon Woo, a theoretical physicist and resident catastrophist at Risk Management Solutions, Inc., picked the gauntlet right back up in a paper published in Frontiers in Earth Science, and argued that finding future Black Swans requires thinking about a historical event and asking, “How could it have been worse?” in a systematic way. “Most events have happened before, nearly happened before or could have happened before,” says Woo. But, having the foresight to imagine the unimaginable by thinking about high-risk, extremely low-probability events raises the question of what scientists, governments and emergency managers around the world can do even if Black Swans can be systematically identified.
 

The downward counterfactual

In psychology, research on counterfactual thinking explores the effects of imagining how a situation could have been different — counter to the reality. In applying this concept to disasters, Woo says, “most [people’s] thoughts [are about] how things could have been better,” but those who work in hazard management should focus on the question of how things could have been worse — the downward counterfactual.

Rick Murnane, CEO of Kinetic Analysis Corporation, a company that provides information on tropical cyclone hazards, explains that modeling risk posed by big earthquakes and other Black Swan events presents serious challenges because such events occur infrequently, and there may be no analog in the historical record. He says, “We don’t have a good sense for what all the possibilities are, so [researchers use statistics] and expert opinion to develop a catalog of events that might happen over a given period of time.”

Because we lack information about these sparse or nonexistent events in our catalog, understanding their effects becomes difficult. “If you had thousands of [big] events … simple statistics gives you the answer,” Woo says. But, those statistics don’t exist because the catalog is incomplete, rendering us effectively blind. Woo proposes using what he calls downward counterfactual analysis to study historic events by reimagining them to be worse than they were.
 

The analysis

Beginning Woo’s analysis requires selecting a historical event to reimagine. The goal, he says, is to develop foresight by studying past near-misses — events that could have turned out much worse. By increasing foresight, he says, “it’s not that you’ll know when something will happen; you just know that something is a possibility.”

After selecting a near-miss event, Woo describes four aspects of the event to reimagine — location, energy, time and human response. Evaluating the consequences of each change helps determine how much worse the event might have been.

One historic near-miss occurred on Nov. 25, 1833, when an earthquake rocked South Sumatra at approximately 10 p.m. local time. The resulting tsunami flooded parts of the island. This estimated magnitude-8.8 to magnitude-9.2 quake was one of the largest global earthquakes of the past two centuries. The number of deaths on Sumatra is unknown. Because this earthquake nucleated further south than the 2004 event, the west-directed tsunami waves largely dissipated, sparing India and Sri Lanka, with only a single report of wave damage reported in the Seychelles.

Had the 1833 earthquake nucleated northward, closer to the vicinity of the 2004 quake, India and Sri Lanka may have sustained numerous casualties when the tsunami crashed ashore, according to Woo.

If the 1833 event had released more energy, an already large earthquake would have ruptured a greater surface area, increasing the reach of devastation. This would likely produce more tsunamis that could have impinged on India, Sri Lanka and other parts of southeast Asia.

Determining the downward counterfactual when it comes to time poses a quandary. With modern buildings in the U.S., “the safest place to be, statistically, is in bed,” says Brian Terbush, an emergency manager in the state of Washington. Nighttime earthquakes are considered less hazardous because more people are already in the safest part of their home, rather than commuting or walking outside where bricks or other debris can fall. However, in the early 1800s, structures in Indonesia may have been more likely to collapse. Perhaps a 2 a.m. quake would have wrought more devastation because more people would have been asleep.

Finally, the human aspect of this downward counterfactual proves difficult to determine. The region jolted by this earthquake appears to have been under Dutch colonial rule. Did the Dutch authorities provide the colonizers and Indigenous peoples with assistance after the calamity? If they did, removing that assistance would likely increase the number of people affected.

Woo posits that if anyone had interrogated the 1833 event following his framework prior to 2004, they would have concocted the very Black Swan event that unfolded that year. “50,000 people died in Sri Lanka and India, and they had two hours of warning between the earthquake and the time that the tsunami struck,” he says. “If someone had done this kind of downward counterfactual thinking, they would have been able to provide some risk assessment to advise the [affected populations] that this could happen,” says Woo. Those two hours in 2004 could have saved numerous lives.
 

Limited resources

“Just because you are aware of something doesn’t necessarily translate into a response or proactive action,” says Murnane. The paper that identified the magnitude of the 1833 event came out in 1999, only six years before the 2004 catastrophe. Terbush, whose job includes disaster preparation, says, “unfortunately, getting the necessary backing from governments to move that quickly is hard, especially when there hasn’t been an event causing loss of life recently,”

Even after such a catastrophe, keeping the momentum to fund measures to protect the population may wane because governments lose interest after their initial enthusiasm, says Murnane. A robust governmental response would include a tsunami warning system, along with funding for maintenance and a communications network. Such a system has been established in the Indian Ocean, but alerting the population of impending disaster has proved to be another complicated hurdle.
 

 

“What do you do in advance to prepare prudently for a Black Swan event? asks Susan Hough, a seismologist at the U.S. Geological Survey. For example, she explains, there’s a very small chance of a magnitude-6.0 event hitting Boston. “In Boston,” Hough says, “you’re not going to retrofit all those old buildings … so what you want to do is focus on the future building codes.” It turns out that the cost to construct buildings resilient to moderate earthquakes is only a few percent of the total. “But, everything is a matter of resources,” she says, “and there’s so many needs in the world.”

For possible hazards that have already been identified, “individual-level preparedness activity is the most cost effective thing we can do with limited resources,” says Terbush. However, he says, “our memories are pretty short when it comes to disasters, and when it comes to history.”
 

Further reading

Natawidjaja, D. H., Sieh, K., Chlieh, M., Galetzka, J., Suwargadi, B. W., Cheng, H., … & Ward, S. N. (2006). Source parameters of the great Sumatran megathrust earthquakes of 1797 and 1833 inferred from coral microatolls. Journal of Geophysical Research: Solid Earth, 111(B6).

Okal, E. A., & Synolakis, C. E. (2008). Far-field tsunami hazard from mega-thrust earthquakes in the Indian Ocean. Geophysical journal international, 172(3), 995-1015.

Woo, G. (2019). Downward Counterfactual Search for Extreme Events. Frontiers in Earth Science, 7, 340.

Zachariasen, J., Sieh, K., Taylor, F. W., Edwards, R. L., & Hantoro, W. S. (1999). Submergence and uplift associated with the giant 1833 Sumatran subduction earthquake: Evidence from coral microatolls. Journal of Geophysical Research: Solid Earth, 104(B1), 895-919.