UW scientists take earthquake science closer to prediction

The advances for which University of Washington scientists have played key roles could make it easier to predict a devastating subduction earthquake long feared in the state.

Crosscut archive image.

Floating house after Japan earthquake and tsunami, 2011

The advances for which University of Washington scientists have played key roles could make it easier to predict a devastating subduction earthquake long feared in the state.

The Pacific Ocean floor is in a shoving match with North America. It is losing. The oceanic plate, pressing forcefully toward the center of the earth, trembles slightly as it loses ground.

These small shivers can’t be felt at the surface, but the delicate quivers may arise from the ocean floor battleground where the plates wrestle and could prove key to understanding how massive subduction zone earthquakes — like the one that recently devastated Japan — build up their potent energy.

The tremors were first discovered in 2002 after Japanese scientists installed a complex and costly — $500 million — system of monitors to study seismic activity in Japan, following a 1995 quake that devastated the city of Kobe and caused more than $100 billion in damage. Before tremors were discovered, scientists believed that faults moved very slowly but continuously most of the time and every once in a while, faults would slip fast, generating large earthquakes.

However, these tremor episodes may indicate the fault section undergoing this new found movement is slipping sporadically and slightly faster than normal.

Ken Creager, an earth and space sciences professor at the University of Washington, has been studying these tremors on the Cascadia Subduction Zone, since 2003. He said, “with this phenomenal data set, they discovered that they were seeing signals which looked like noise. We’ve been seeing that forever. But the noise looked similar on different stations.”

This completely puzzled scientists; they had no idea what it was.

The noise on the Japanese seismometers looked like vibrations produced by volcanoes as magma rises to the surface of the earth. But these events seemed to be occurring deeper — 25 to 30 miles below the earth’s surface — rather than 10 miles below, where volcanic tremors usually occur, said Mike Brudzinski, a geology professor at Miami University in Ohio who has been studying tremors for a decade.

Japanese scientists noticed that the nondescript noise on each seismometer looked similar to what was on other nearby seismometers.

In addition to the noise, scientists in the Pacific Northwest and Canada, using GPS, could see the plates move and the ground deform — and realized the tremors might be linked to this massive plate migration.

“As soon as we heard about [Japan’s] stuff and they heard about our stuff here, they put the two together,” Creager said.

Subduction zones are combat zones where one plate of the earth’s crust is thrust jerkily under another slab of rock. As tension builds between the two massive plates, they become more and more dangerous.

These types of faults can be found in Chile, Sumatra, and the Pacific Northwest. They have the potential to produce magnitude 9 earthquakes, which can be cataclysmic across huge areas. A Magnitude 9 earthquake struck Japan in March. The Kobe earthquake was a magnitude 7.2 earthquake and killed over 6,000 people; the Sumatra quake in 2004 was a magnitude 9.1 and generated a tsunami, killing over 213,000 people. The Chilean earthquake in 2010 was measured as a magnitude 8.8 quake, which killed over 500 people.

Despite the size of these quakes, scientists have little information about what led up to them or if tremors were involved. Any seismic data from Sumatra or Chile remains unavailable to UW scientists.

“Ideally, if we had instruments there, and watched the time before the big earthquake[s], we would have a sampling of whether there is a change in the pattern [of tremors],” said John Vidale, an earth and space sciences professor at the UW. “At least in Chile, they put in a lot of instruments afterwards but we already have ways of telling when a Magnitude 9 earthquake hits. We’d like to know when it is going to hit.”

Tracking these barely perceptible tremors hasn’t been easy and determining how deep they are has proven nearly impossible.

Until now, “we haven’t been able to locate tremor real well, especially at depth,” said Abhijit Ghosh, a doctoral student in the earth and space sciences department at the University of Washington who has spent fours years chasing tremors.

Ghosh has been studying these tremor events and can precisely locate where the tremors occur from the surface. However, the depths of the tremors have remained elusive.

It is crucial for scientists to understand where these tremors are happening. If the tremors are scattered at depth, they may give no insight into mega-thrust earthquakes.

If Ghosh and his colleagues can prove their theory that tremors happen at the same depth as the fault, they could determine whether the subduction zone is slowly, but continually moving.

Brudzinski says that some tremors have been related to other natural events like volcanism, which would be meaningless to scientist’s understanding of seismic activity. However, he disagrees with these other theories, saying, “There has been some suggestion that these tremors and slips could trigger these big earthquakes.”

To solve this mystery, Ghosh is using something he calls “the array of arrays,” multiple clusters of instruments, each containing about 20 seismometers and taking up about a square mile of space. Eight of these arrays were placed near the Juan de Fuca Strait in 2009. Located on national forest land, private property, and areas owned by the state Department of Natural Resources, the equipment sits unobtrusively in the Olympic foothills near Sequim and Port Angeles.

Ghosh created a technique called “multi-beam back projection” that traces the path of the tremor from the seismogram to the original location. He developed an algorithm of codes and set up a program that can locate tremors on a flat surface within an hour of when they occur.

Developing the algorithm was time consuming but once the code is running and the tremor is located, Ghosh said it only takes a half an hour to compile a day's worth of tremor movement. Ghosh now has data for an entire year and is beginning to analyze it.

However, Ghosh admits it has been “hard to detect and locate [a] tremor,” because the vibrations are a low amplitude signal and are hard to detect.

While this sounds difficult, it is a lot like radar. Much like an airplane uses radio waves to detect a tower on a runway, the tremor signal reaches the seismometer and the signal bounces back to the source. This has allowed Ghosh to locate tremors on a flat surface.

Unfortunately, techniques available to detect and trace tremors have drawbacks.

“The initial results are that we can see tremor locations on a two-dimensional plane in more detail,” Ghosh said. “We can see there are different kinds of migrations, which tells us something about the physics of the slip.”

He believes that it is crucial to understand the physics of the tremor because if the silent slip is happening at fault depth, “it is happening because the fault is moving and it is generating seismic energies.”

What scientists believe is the general movement of a tremor follows the fault line as the plate subducts. Ghosh’s ability to locate tremors on a latitudinal-longitudinal axis, and trace its path over time, could mean that the Cascadia subduction zone is moving continuously, and not just in pulses, which is something scientists could only speculate about before they knew about the tremors.

Geologists do not even know how far inland a large earthquake in the Cascadia Subduction Zone could reach or the exact depth of the rupture. However, the western boundary of the tremor zone could indicate where the subducting slab of earth is stuck and building pressure.

“[We have] better information than what we had and unfortunately that [subduction] line is a lot closer to Seattle than we thought,” Vidale said.

Some clues were found that might indicate the location of the Cascadia subduction zone. Vidale pointed out “a sharp line where the tremor ends as [tremor] gets shallower,” meaning that “sharp line” could show were the locked part of the fault is.

What complicates the matter is that tremors can happen within different sections of the fault. This means that different areas of the fault would have distinctive levels of stress. If scientists, like Ghosh, can prove that these tremors are happening at a certain depth, they can begin to understand how seismic stress would build up at a fault and how it could suddenly give way to a huge earthquake.

Finding the exact depth of these tremors will take more work. However, scientists, including Ghosh, are optimistic that their information on tremors and continual research can teach them more about the Cascadia Subduction Zone, which could — and most likely will — wreak havoc on the Pacific Northwest.

"If I find that a majority of the tremors are occurring at the fault plane, it means that the fault is slipping … whenever tremors are occurring," Ghosh said. "That would mean that it is transferring stress to the locked section of the fault, making the big earthquake one step closer to happening."


Please support independent local news for all.

We rely on donations from readers like you to sustain Crosscut's in-depth reporting on issues critical to the PNW.


About the Authors & Contributors