Portland faces a one-in-five chance of a major earthquake in the next 50 years.Jeff Vinnick / Reuters

For about the last 30 million years, a small tectonic plate named Juan de Fuca has been sliding under the far vaster North American plate into the Earth’s mantle. Today, this mostly happens without anyone’s notice—even though it causes minor, near-undetectable earthquakes about every 300 days—but sometimes the pressure pent up is released suddenly and catastrophically.

This is what happened on January 26, 1700. The plate slipped, and a magnitude-9.0 earthquake resulted, devastating the coast of modern-day Oregon and Washington. According to one story, an entire First Nation on Vancouver Island, the Pachena Bay people, died in flooding overnight. And the quake triggered a tsunami that rode across the Pacific Ocean for 10 hours before slamming the east coast of Japan, where merchants and samurai recorded flooding and damage.

As hundreds of thousands of Americans now know, this could happen again—except now, millions more people inhabit the Pacific Northwest. The existence of the Cascadia subduction zone, and its power to jolt the region with a “really big one,” was revealed to mass audiences last year by the writer Kathryn Schulz in a barn-blazing story for The New Yorker.

However, it now seems these coastline-altering events happen more frequently than previously thought. A team of researchers led by Chris Goldfinger, a geologist at Oregon State University, has found evidence that at least 43 major earthquakes have occurred in the last 10,000 years. That number is slightly larger than previously estimated, which means that—over the long time period—it significantly alters the likelihood of any one event occurring.

In terms of sheer numbers, previous reports estimated that Washington state would have a major quake every 500 years on average. Goldfinger’s team now shows them to occur about every 430 years. And northern Oregon, which used to receive a quake about every 430 years, instead faces one every 350 years.

This means that there is about a 20 percent chance that northern Oregon—a region that includes Portland and Astoria—will be hit by a magnitude-8.0-or-higher quake in the next 50 years. (Previously, the area was estimated to have about a 12 percent chance of facing such a quake.) Washington state has between a 14 and 17 percent chance of facing a big one, up from an 8 to 14 percent chance.

These numbers still pale in comparison to earthquake risks to the south. For instance, the San Francisco Bay area has a 50 percent chance of experiencing a magnitude-7.0 earthquake in the next 30 years. Los Angeles faces a 93 percent chance. But scientists expect a major rupture in Cascadia would be much more powerful than one of these Californian quakes.

“Geology is a badly edited tape, and we lose lots of evidence along the way,” says Mika McKinnon, a geophysicist and disaster researcher who was not associated with the study. “We lose evidence of drowned forests, we lose evidence of faults moving, we bury everything in new volcanic floes, we run glaciers over everything 15,000 years ago and completely erased the older history.”

This means that geologists looking for clear evidence turn to the Holocene, the past 10,000 years that followed the end of the last Ice Age.“Yet even in that little, limited time span, we found [earthquakes] as close together as 100 years, and as far apart as 1,000 years. Averaging doesn’t work very well when you’re using very small numbers,” she said.

“It is not surprising that we are finding more evidence of more earthquakes,” she added.

Goldfinger’s team conducted this research by taking core soil samples from underwater sites off the Pacific coast that showed evidence of submarine landslides. By comparing how those core sample hold up to a variety of tests—including density, resistivity, and radiocarbon aging—they can identify whether multiple landslides occurred simultaneously. If different samples from different parts of the ocean bear the same signature, it’s a give-away for an earthquake.

This isn’t a new technique in the hunt for Cascadian earthquakes, but Goldfinger and the team deployed it much more broadly than before. Previous studies had only examined 12 core samples; their research analyzed 195 different deposits.

For geologists and geophysicists, the headline of the report isn’t so much the newly discovered earthquakes—it’s the increased confidence around the basic science of the Cascadian subduction zone. By going back 10,000 years, the study makes it even clearer that enormous ruptures in the Pacific Northwest happen regularly and strongly.

“Almost all the angles that have been tried result in almost exactly the same answer,” Goldfinger told me. “You have to go back to the 1970s to find a viable alternative hypothesis. In the early days of plate tectonics, there was an idea for a short time that Cascadia was very recently a subduction zone but that it was no longer. But then in 1980, Mount St. Helen’s went off—and evidence was found that other volcanoes had erupted in relatively recent time periods, too—and we realized that theory probably wasn’t a great one.”

“The sure end came in the 1990s, when GPS was invented and we could easily measure plate motion. And, yep, North America and the Juan de Fuca plate were converging at the rate we thought,” he said.

Knowledge of the most recent quake formed even more recently. Seismologists, anthropologists, and Japanese historians only pieced together the story of that rupture last decade, when they produced the magnificent book The Orphan Tsunami of 1700. (That report, available for free online, was updated last year). And general knowledge only followed last year with Schulz’s story.

Goldfinger had thought most people knew about the Cascadia fault: Stories about it had appeared in National Geographic and local papers, and NOVA had done three episodes.  “Then this article came out in the New Yorker, and we learned hardly anyone knew about this,” he said, laughing. (Goldfinger is a major character in the story.)

“Now it’s turned into a regional semi-panic, and that’s not entirely a bad thing,” said Goldfinger. “We do have a big problem, and we do have a long way to go, and if we don’t panic a little bit, we’ll never get anywhere.”

We want to hear what you think. Submit a letter to the editor or write to letters@theatlantic.com.