Subtle Signs May Have Hinted at Earthquake

Scientists taking a first pass at analyzing data from last month’s magnitude 6 earthquake in Parkfield have found preliminary evidence that the San Andreas fault may have given off early hints that it was about to rupture.

There were none of the obvious signs that scientists had hoped for: no foreshocks or cracking of the ground in the hours before the quake. But in the 24 hours before the quake, there were subtle, slow movements of the ground and water deep below the surface.

Some scientists think these may have been the first tiny rumblings of the bigger rupture.

If these early hints -- called precursors -- exist, they could serve as an early warning system for earthquakes.

“If you had an observation of a precursor, you have reason to believe there is a preparation process that happens before an earthquake,” said Susan Hough, a seismologist with the U.S. Geological Survey in Pasadena. “That gives you something to hope for and reason to believe short-term predictions of earthquakes are possible.”

Researchers remain cautious about the data. The signals were so faint -- 100 to 1,000 times weaker than the quake that came after them -- that they could have been background noise caused by small events, such as changes in atmospheric pressure.

“The issue is to convince ourselves this was not just noise,” said Malcolm Johnston, a geophysicist with the USGS in Menlo Park. “It’s right on the edge, so we don’t want to make it a big deal. On the other hand, if they’re there, it’s very important.”

Scientists have been hunting unsuccessfully for decades to establish the existence of precursors. Many phenomena thought to precede quakes -- such as strange animal behavior or sudden magnetic shifts -- have never panned out scientifically.

For years, geologists couldn’t talk about prediction or precursors without being laughed at.

“You couldn’t say the P-word in polite company,” Hough said.

But seismologists also couldn’t quite give up the search.

The best hope to snare precursors is Parkfield. Because moderate earthquakes shake the area every few decades, USGS scientists have wired the area extensively.

Parkfield has also been of special interest to those hunting for precursors because of two reports from the region’s last magnitude 6 quake in 1966. A few hours before that quake, an irrigation pipe lying across the fault burst. In the days beforehand, small cracks were observed in the ground.

Nothing so obvious happened at the surface before the Sept. 28 Parkfield quake. The clues were deeper, said Johnston.

Strain meters embedded 600 to 800 feet underground detected a slow, subtle shifting deep in the earth in the 24 hours before the earthquake, he said.

Meters put in sealed wells at the same depth showed changes in water pressure, indicating that fluids had moved during the same time period. Such fluid movement is expected during a quake and perhaps beforehand; anecdotes of water levels shifting in wells and streams before quakes are common lore.

Were these signals detected by Johnston a whispering or perhaps a stuttering by the fault before it let loose with a shout?

“The answer to the question is a guarded yes,” he said.

As can be expected, with evidence so scant and stakes so high, the arguments have already begun.

“It’s really too early to talk about precursors. There was nothing that jumped off the screen,” said Bill Ellsworth, chief scientist of the earthquake hazards group at the USGS in Menlo Park. “Parkfield told us something we already knew: If earthquakes are predictable, it’s not going to be easy.”

Parkfield has already shattered many expectations of how the San Andreas fault should behave.

Bob Nadeau, a research geophysicist at UC Berkeley who runs a network of very high resolution seismic monitors in Parkfield, was expecting to see a cluster of micro-earthquakes caused by rocks breaking under intense pressure.

“We were hoping the earthquake would essentially start cracking -- little cracks -- a few days before,” he said.

Nadeau has analyzed data from three days before the quake and has found no micro-earthquakes. “It’s a little bit of a letdown in that regard,” he said, adding that he would continue to look at the weeks leading up to the quake for signs of any suspicious activity.

The quiet may be a signal in itself. Although Nadeau normally sees two to three micro-earthquakes a day on his network, in the three days before the quake, he saw zero.

“They call it quiescence or quietness,” he said. “It may be that the rocks hold tighter before the quake.”

Or it may be that scientists just have to look deeper. Other seismologists, he said, are trying to track shaking known as “nonvolcanic tremors” that occurs deep underground. Some of these tremors have been detected before earthquakes in Japan and in the Pacific Northwest, Nadeau said.

USGS seismologist Steve Hickman is co-director of the San Andreas Fault Observatory at Depth, or SAFOD, a 3-mile-long borehole being drilled into the fault in Parkfield.

Hickman was atop the drill rig, 31 feet above the ground, when the Parkfield quake hit.

“We’ve been looking at earthquakes from very far away, from the surface,” he said. “With SAFOD, we’re going to get into the heart of the fault zone.”

The borehole was not finished in time to capture last month’s quake, but repeating quakes of about magnitude 2 occur on a stretch of the fault in Parkfield every few years.

The good news is that the Parkfield quake has sped up the cycle of these repeating quakes, so the team will be able to monitor them sooner than they had anticipated -- and get a head start on teasing out any precursors that may exist within the fault zone.

Even though Parkfield is considered one of the best behaved stretches of the San Andreas because of its regularity, last month’s earthquake has left plenty of new puzzles.

Unlike the past two quakes in Parkfield, this one was not preceded by a magnitude 5 foreshock 17 minutes beforehand. It traveled backward, compared to the past two quakes. And, luckily, it was not followed by a massive Southern California quake a few hours afterward, as happened in 1857, when a 7.9 quake hit Fort Tejon.

“It shows you how mercurial earthquakes are,” said Hough, who thinks the jury is still out on whether precursors were detected at Parkfield. “You think you understand, and then the earth does something else.”