Bowing
California’s San Andreas Fault is renowned for its large and destructive earthquakes. But some stretches of the fault slide in slow motion, with minimal or no earthquake activity. These “creeping” fault segments slip steadily, releasing stresses that would otherwise accumulate and trigger large earthquakes.
“This has very important implications for where to expect earthquakes versus where to not expect earthquakes.”
Geologists generally think that creep happens on faults that slice through particularly slippery rocks. But a recent study reports that the overall structure of a fault network might also dictate why some faults creep rather than intermittently slip dramatically.
“This has very important implications for where to expect earthquakes versus where to not expect earthquakes, as well as for predicting where the most damaging earthquakes will be,” Victor Tsai, a geophysicist at Brown University and study coauthor, said in a statement.
Although the study focused on California’s faults, which include the San Andreas, the results could apply to other faults around world, according to the researchers. Their findings were published in Nature.
Complex Faults
Faults are rarely simple, planar features. They have wiggles and gaps and crisscrossing stands—complexities known to influence earthquake hazards.
A bend in a fault can, for instance, stop an earthquake in its tracks or cause it to hop to another branch and grow larger. “There’s been a lot of work on fault geometry in the past,” said Julian Lozos, a seismologist at California State University, Northridge, who was not involved in the study. “Fault complexities are one of the factors that go into hazard assessments.”
Tsai and his colleagues investigated how geometry affects creep. They measured the orientation of surface fault traces on U.S. Geological Survey maps. They split the maps into 30-kilometer-wide (19-mile-wide) areas, assigning each area a value between 0 and 1, indicating whether it comprised parallel fault strands or random fault orientations, respectively. They then compared that dataset with previously measured creep rates, established using instruments that record displacement across a fault and other techniques such as observations of offset roads and fences.
The researchers found that complex fault zones have slower creep rates and experience more earthquakes than faults that run in parallel, suggesting that these complexities might play a guiding role in creep.
Until now, seismologists have been studying how creep correlates with certain rock properties, Lozos said. For decades, geologists have found a green metamorphic rock, serpentinite, on creeping sections of the San Andreas. Serpentenite is inherently slippery, but under certain conditions it can alter to talc, which is even weaker. These low-friction rocks allow the sides of the fault to slip slowly and steadily. These rock properties have been demonstrated in the lab on samples hauled up from deep on the San Andreas.
“How much faults are misaligned seems to correlate well with where faults are creeping or have earthquakes,” Tsai said. He and his colleagues think that the jagged, interlocking interfaces between misaligned faults may cause the rocks to snag against each other. They suggest that rock properties might play a lesser role in creeping than previously thought. They wrote in the study, “Our findings contradict the previous understanding that creep occurs along faults with high roughness, heterogeneous structure and certain compositions.”
Combined Factors
“I don’t think it’s just down to geometry. Creep must be caused by a combination of factors.”
“I think they have an interesting and useful result here,” Lozos said. But the researchers’ interpretation is too black-and-white, he said, “I don’t think it’s just down to geometry. Creep must be caused by a combination of these things.”
Tsai agreed. “Our work demonstrates that fault geometry is important,” he wrote in an email, “but not necessarily that it must be the most important factor.” He added that the lack of data on rock friction at depth makes determining the relative roles of geometry and rock properties difficult.
A way to test whether one factor predominates would be to use a model to simulate a complex fault with the added rock properties of a creeping fault, Lozos said. Where the new study comes into its own is in explaining why some parts of the San Andreas are locked when, in fact, they have the right rocks for creep, he explained.
“There’s a freeway in the Bay Area that follows the path of the San Andreas,” Lozos explained. “You can see the green rocks in the road cutting. But that part of the fault is locked; it was involved in the 1906 San Francisco Earthquake.” In contrast, the nearby Hayward and Calaveras Faults, which branch off the San Andreas and run roughly parallel to it, have serpentinite on them but are creeping, as would be expected.
The San Andreas is just a tiny bit more misaligned with the plate boundary than the other two faults in this area, Lozos said. “They’re off by just enough that this argument [about orientation controlling creep] is still tantalizing.”
—Erin Martin-Jones, Science Writer
