American Scientist

On Christmas Day 2004, early in the afternoon, geophysicists at the Pacific Tsunami Warning Center in Ewa Beach, Hawaii, realized that a powerful temblor had occurred just west of the Indonesian island of Sumatra—one they feared might cause a significant tsunami when the disturbance reached shore. Unlike the waves that propagate through the solid earth, tsunamis can take hours to travel between their source region and some of the faraway places they damage. So, in principle, early detection and the dissemination of warnings can save lives. Yet the geophysicists at the Pacific Tsunami Warning Center found themselves ill-prepared to alert officials in nations bordering the Indian Ocean. One reason is that the network of tide gauges and ocean-bottom pressure sensors these geophysicists rely on is, for the most part, limited to the Pacific, so they did not have the means to check whether a tsunami had indeed been generated. Given the huge number of people that perished in this horrific disaster, it seems fitting that world leaders have called for a tsunami warning system, presumably one similar to that in the Pacific, to be set up in the Indian Ocean. But is this really the best response?

One concern is that such a move does nothing to improve tsunami readiness in communities bordering the Atlantic, an ocean that is typically ignored in this context. Big tsunami-generating earthquakes are most likely to take place in the Pacific, which is surrounded by subduction zones (places where two tectonic plates grind together as one plunges downward), or to a lesser extent in the Indian Ocean, which is bordered by the Sumatran subduction zone on the east. But, in fact, the Atlantic also has regions where subduction takes place. One is situated to the east of the Caribbean Sea, just outside the arc of the Antilles islands. Subduction also takes place to the east of the South Sandwich Islands, which are found in the South Atlantic. Although these earthquake-prone zones are smaller than those bordering the Pacific, neglecting their potential to cause destructive tsunamis would be irresponsible—especially given that an even more subtle area of subduction in the Atlantic produced a devastating tsunami in the not-so-distant past.

On November 1 of 1755—All Saint's Day—at about 9:30 in the morning, a powerful earthquake struck in the easternmost Atlantic. The shaking was felt throughout Europe, with the most destructive effects focused along the coast of the Iberian Peninsula and Morocco. Lisbon in particular was hard hit, and many of its 250,000 residents lost their lives. Most perished in the collapse of buildings. Others succumbed to the fires that immediately broke out. Some sought refuge from the toppling buildings and growing inferno along the city's waterfront, only to become victims some 40 minutes later when the tsunami reached shore. The waves grew about 6 meters tall in Lisbon and reached a height of more than 15 meters at Cape San Vicente in southwest Portugal. They ran up to 10 meters or more when they hit the coast of Morocco.

Casualty estimates vary considerably, but it is generally thought that the earthquake, fire and tsunami together killed some 60,000 people. This death toll amounted to a substantially greater fraction of the world's population (then probably about a half a billion) than did the toll in the recent Indian Ocean calamity. The scale of this 18th-century disaster and its devastating impact on Lisbon—then an important center of Catholicism—engendered much philosophical questioning about the nature of divine justice. It inspired, for example, Voltaire's satirical novel Candide, in which the titular protagonist is constantly asking his tutor, the ever-optimistic Dr. Pangloss, about how to reconcile the woes of the world with the existence of a benevolent God.

So the lesson from history is clear enough: Subduction-zone earthquakes in the Atlantic can also create destructive tsunamis. Another reason not to exempt the Atlantic from future planning is that dangerous tsunamis are not all triggered by large subduction-zone earthquakes. In 1929, for example, a relatively modest quake triggered an undersea landslide on the continental slope off Canada. That landslide in turn gave rise to a tsunami, which killed 29 people. And marine geologists are aware of similar spots on the other parts of the continental shelf that must have raised even more havoc in the distant past. In 2000, Neal Driscoll of Woods Hole Oceanographic Institution and two colleagues published a study of the sea floor off the coasts of Virginia and North Carolina, suggesting that this area of the continental slope may be subject to failure, something that could conceivably spawn a devastating tsunami along the Eastern Seaboard of the United States.

Is the proper answer, then, to expand the tsunami-warning system now in place in the Pacific to the whole planet? Perhaps. But this strategy doesn't make for a complete solution. The problem, as many commentators have noted, is that registering the threat and communicating warnings to country officials does not guarantee appropriate action on the ground. The Indian Ocean disaster provides an illuminating example: After the quake struck, the geophysicists in Ewa Beach sent an e-mail warning to authorities in Thailand, who had given them no other contact information. But the message was ignored. (Accounts differ: Some say the critical e-mail was never read; others contend that Thai officials were fearful of jeopardizing the tourist trade and decided not to issue a warning, expecting that it would turn out to be a false alarm.) But more vigilance on the part of civil-defense officials may not be the answer either. In Japan, where tsunami warnings are effectively communicated, the citizenry tends to get jaded—in a recent episode in Hokkaido, only half the people warned to evacuate the danger zone bothered to leave.

What then must we do? Better education about tsunamis for anyone living close to the ocean would probably go a long way—even without any formal warning system. Steven Ward, a geophysicist and tsunami modeler at the University of California, Santa Cruz, says that "the first arrivals are not usually all that big." He further points out that the period of these waves can be 10 or 15 minutes, meaning that after observing unusual changes in the level of the ocean" you'd have time either to walk to high ground or climb up a good, strong building."

Expanding on the Pacific warning system would, of course, be helpful too. But in doing so, planners should pay considerable attention to the inherently weak link between warning and response. The system shouldn't allow lackadaisical authorities to veto alarms by whim; nor should it be set up with such sensitive triggers that populations eventually become complacent. To fix such problems, geophysicists and civil-defense planners will probably need to take heed of psychology as much as technology. In the best of possible worlds, clever minds could be expected to come up with a system that always worked flawlessly. But, as Voltaire was so good at pointing out, ours is far from the best of all possible worlds.—David Schneider