Seismic Monitors Detect Physical Changes Deep Within Faults

The new findings, published in the Dec. 4 issue of the journal Nature, are based on an extensive study of data collected between 1987 and 1997 by ultrasensitive borehole seismometers along the Parkfield segment of the San Andreas fault in central California. By comparing seismograms from a series of minor earthquakes that occurred both before and after a "slow" earthquake at Parkfield in 1993, the researchers were able to detect subtle changes in the level of stress along the fault zone that were caused by this event.

"Our study focused on S-waves, also known as shear waves, which are created during every earthquake," said study co-author Fenglin Niu, assistant professor of Earth Science at Rice University. "S-waves bounce off of deep fractures filled with fluid, and we believe our data show how these fluids were redistributed as a result of the aseismic event in 1993."

An aseismic event occurs when there is a significant amount of movement along a fault line. Unlike earthquakes, where movement occurs within a few seconds, aseismic movements can take place over days, weeks, months or even years. For this reason, they are sometimes called slow earthquakes.

Parkfield, a small town that lies on the San Andreas fault, was chosen as the site of a focused earthquake experiment by the U.S. Geological Survey because of the town's history of magnitude-6 earthquakes. Such quakes occurred in 1857, 1881, 1901, 1922, 1934, and 1966. Believing that another magnitude-6 quake was likely to occur before 1993, the USGS began the Parkfield Experiment in 1985, positioning a dense network of instruments in an effort to capture unprecedented and detailed information about an earthquake as it happened. The long-expected magnitude-6 quake has yet to occur at Parkfield, though the annual probability remains around 10 percent.

For Niu and his colleagues, the dense network of instruments around Parkfield provided the critical data needed to prove that structural changes in faults can be detected with seismic instruments. Using data collected by borehole seismometers positioned about 200 meters below ground, the group looked at a series of areas that scattered the S-waves produced by minor quakes along the fault.

"With the seismic data from only one seismic station, it is very difficult to determine whether the physical properties of the material in the fault zone have changed or the positions of the minor earthquakes have shifted," said Niu.

Niu and colleagues Paul G. Silver of the Carnegie Institution in Washington, D.C., and Robert M. Nadeau and Thomas V. McEvilly, both of the University of California, Berkeley, were able to correct for movement errors by studying data from seismometers at various positions throughout the region.

Niu said the research could become increasingly important in coming years because it provides a basis for understanding the structural changes that could be viewed with active seismic monitoring systems. Given recent improvements in seismic technology, seismologists are now considering how to design active seismic systems that monitor fault behavior continuously.

The study was funded by Rice, the Carnegie Institution of Washington, NASA and the USGS.