Predicting Earthquakes

Oregon State University Professor Finds Link Between West Coast Earthquakes

By Elaina M. Jorgensen

It struck during Game 3 of the 1989 World Series and lasted 15 seconds. In that time, the Loma Prieta earthquake in the San Francisco area caused 67 deaths, almost 4000 injuries, and over $6 billion in property damage. Earthquakes can occur worldwide, but in our region and the rest of the countries on the "Ring of Fire,” they are a fact of life; not a question of if, but when and how bad.

Now, a new facet of West-coast earthquakes has been revealed: quakes along the two major fault systems may be linked. Results of a new study show a strong correlation between the timing of earthquakes in the San Francisco region and those of the Juan de Fuca Plate in the Northwest. The work was published in the April 2008 issue of the Bulletin of the Seismological Society of America by Chris Goldfinger, associate professor of marine geology at Oregon State University.

These plates are the reason we have earthquakes in the first place. Our spherical planet has an internal core surrounded by a molten mantle which is, in turn, surrounded by a hard crust. The crust, or lithosphere, consists of several plates which move around on the more malleable mantle. The plates are always in motion, though typical rates are on the order of millimeters a year.

The plates move in three general ways: away, toward, and by one another. Where the plates move away from each other, new crust is formed when the magma rises to the surface; this is called a spreading center. Where the plates collide results in either one plate sliding under the other, called a subduction zone, or both plates being uplifted, forming mountain ranges. Where plates slide by one another, strike-slip faults are formed.

The West Coast of North America, on the North America Plate, is bordered by two plates, the Juan de Fuca Plate off British Columbia, Washington, Oregon, and northern California and the Pacific Plate off the southern portion of California.

The Juan de Fuca Plate is being subducted under the North American Plate while the Pacific Plate is sliding along it, creating the San Andreas Fault.

Goldfinger's results indicate that seismic activity along the subduction zone of the Juan de Fuca Plate and North American Plate may trigger earthquakes on the San Andreas Fault.

Similar results have been found by other geologists, including Ross Stein of USGS, when he successfully predicted earthquakes in northern Turkey. The theory he used to predict the earthquakes was that as the plates are forced into one another, one part of a plate will move abruptly while the rest of the plate remains unmoved. The portion of the plate that does not move with this initial displacement is now under increased stress, which can only be relieved by another earthquake.

While Stein researched historical records of earthquakes due to two plates sliding by one another, Goldfinger's work focuses on the marine geological record for two plates that appear not to be in contact. "The principle is the same,” says Stein, "the difference is that one fault is vertical [subducting] and one is horizontal [strike-slip].” Stein explains that, "faults don't have to be physically connected for stress to transfer. The earth's crust behaves like stiff rubber so stress can be transmitted in the crust.”

Goldfinger and his team assembled the earthquake history for the West Coast by sampling marine sediments off the coast. "The [geology] community is excited because Chris has found a way to sample a long history of earthquakes and the results are consistent,” says Stein. They used steel core tubes, dropped vertically into the seafloor to retrieve a sample of sediments that then are analyzed and radiocarbon dated. They look for layers of sediments that meet certain criteria to establish their earthquake origin.

When an earthquake strikes in our area, it causes landslides off the coast. During these landslides, the marine sediments are liquefied into a muddy slurry. Because the slurry is heavier and more dense than the surrounding water, it has the ability to more easily move larger sized particles than just water alone. When the sediments from these landslides finally settle they leave a distinctive layer for geologists to find, called turbidites.

Turbidites caused by earthquakes have a large spatial range whereas those created by storms or small, natural landslides are more localized. To determine the spatial distribution of the turbidite layer, the team assembled hundreds of cores from the subduction zone and from the San Andreas Fault. They found 15 distinct turbidite layers over the past 3000 years in the San Andreas Fault samples. Of those, 13 were coincident in time with turbidite layers from the Cascadia subduction zone (the southernmost subducting portion of the Juan de Fuca Plate). "There is no other place on the San Andreas Fault with such a long history,” notes Stein. The Cascadia subduction zone turbidite layers appeared slightly earlier than those from the San Andreas Fault region.

Cascadia subduction zone earthquakes are generally larger and more frequent so are believed to be the trigger, not vice versa.

The method for measuring time scales cannot offer an accurate measurement of how offset the timing of the events were. But it appears that a major earthquake in the Cascadia zone is followed by a San Andreas Fault earthquake by about 25 to 45 years, averaged over 13 earthquakes. In the 3000 years covered in this study, the Cascadia zone earthquakes occurred about every 530 years in the northern portion and every 220 years in the southern portion. As Stein points out, "the information on the timing of these events is profoundly important to understanding our hazards.”

Elaina Jorgensen is a doctoral student at the University of Washington School of Oceanography.

Image: An illustration of the tectonic plate boundaries of the West Coast of the United States (indicated by red line). Image: Elaina M. Jorgensen/UW