Posted on behalf of Alexandra Witze.

The 18 great earthquakes that have struck Earth in the past decade hold ominous lessons for western North America, a top seismologist has warned. Many of these large quakes — including the 2004 Sumatra quake that spawned the Indian Ocean tsunami, and the 2011 Tohoku disaster in Japan — were surprisingly different from one another despite their similar geologic settings.

That variety implies that almost any scenario is possible in another part of the Pacific Rim where quake risk is thought to be high — along the Cascadia subduction zone offshore of Washington, Oregon, and other parts of the western United States and Canada.

“We do not fully understand the limits of what can happen,” says Thorne Lay, a seismologist at the University of California, Santa Cruz. “We have to be broadly prepared to respond.”

Lay spoke on 21 October at the Geological Society of America meeting in Vancouver, Canada, a city on the front lines of Cascadia earthquake risk.

“The recent spate of great subduction-zone quakes, of magnitude 8 or larger, began with the 2004 Sumatra earthquake”

The last great quake in the region happened in 1700. Conventional wisdom holds that the next one, perhaps as large as magnitude 9, could strike at any time in the next several hundred years. Geologically speaking, Cascadia is a classic subduction zone, where one plate of Earth’s crust plunges beneath another, building up stress and occasionally relieving it in large earthquakes.

The recent spate of great subduction-zone quakes, of magnitude 8 or larger, began with the 2004 Sumatra earthquake. On average, each year since then has brought 1.8 great quakes, more than twice the rate of the previous century.

In large part, they happened where and when seismologists expected them. “The quakes are basically filling in a deficiency of activity,” Lay says. But their details have been surprising.

The 2004 Sumatra quake, for instance, ruptured unexpected portions of a subduction zone off Indonesia, where the fault zone bends as opposed to running straight. That implies that areas in Cascadia with unusual geometry might also be at risk, Lay says.

In 2007, in Peru, a major earthquake began to happen, then essentially stopped for 60 seconds before picking up again and eventually generating a large tsunami. That start-stop-start pattern raises challenges for Cascadia because seismologists are trying to develop an accurate earthquake early warning system there.

And in April 2014, a Chilean quake ruptured a far shorter portion of a subduction zone than scientists had expected. That suggests that researchers can’t be complacent about thinking they know which parts of Cascadia might break, Lay says. (The worst-case scenario for Cascadia involves a rupture of approximately 1,000 kilometres.)

That’s not to say scientists aren’t preparing. The recently launched M9 project, coordinated out of the University of Washington in Seattle, aims to help officials cope with the risk of a great Cascadia quake. At the Vancouver meeting, Arthur Frankel of the US Geological Survey in Seattle showed early results of calculations of where the ground might shake the most. Enclosed basins, like Seattle, amplify the shaking, he reported.