A new paper explores the physics that drive big earthquakes along plate boundaries
New research from Caltech seeks to explain the size of the forces acting on so-called "mature faults"--long-lived faults along major plate boundaries like the San Andreas Fault in California-in an effort to better understand the physics that drive the major earthquakes that occur along them.
Major earthquakes in the range of magnitude 7.5 or greater are relatively rare, making them difficult for scientists to study. Using computer modeling, a team from Caltech has examined the relationships between the size of an earthquake, the energy it radiates out, and the heat generated by movement along the fault.
"Understanding the physics that govern major earthquakes on different types of faults will help us to generate more accurate forecasts for earthquake threats," says Caltech graduate student Valère Lambert (BS '14, MS '17), lead and corresponding author of a paper on Lambert collaborated with Nadia Lapusta , the Lawrence A. Hanson, Jr., Professor of Mechanical Engineering and Geophysics, and Stephen Perry (MS '14, PhD '18) of the Caltech Seismological Laboratory.
One challenge in understanding mature faults is the heat-flow paradox: over the past 50 million years, the Pacific Plate and North American Plate have slid past one another along the San Andreas Fault at an average rate of about 2 inches per year, a tectonic grinding that should produce a tremendous amount of heat from friction. However, no excess heat has been detected.
As such, seismologists have concluded that, during earthquakes, mature faults along plate boundaries slide at much lower levels of stress than would be expected based on the results of lab experiments.
Two competing models seek to explain the paradox. One suggests that the friction along the fault is high (preventing motion) when the ground is still, but, during an earthquake, the fault becomes what is known as dynamically weak. This can happen during an earthquake if, for example, fluid trapped along the fault vaporizes to create a counterforce to those keeping the fault clamped shut; this allows the two sides of the fault to more easily slide past one another.
The second model assumes that pressurized fluid is always present along the fault, making it weak all the time.
While these two models paint very different pictures for how faults move during large earthquakes, it is challenging to distinguish between them using motion on Earth's surface. The Caltech team turned to computer modeling to examine how seismological observations can be used to differentiate these two possible scenarios.
The modeling revealed that in the rupture of a persistently weak fault, ever-larger swaths of the fault would slip as the quake progresses, as occurs in the formation of a crack.
In contrast, the rupture of a dynamically weak fault would propagate as a narrow "self-healing pulse" traveling along the fault; in this scenario, a much larger amount of radiated energy would be released than would be generated by a crack-like rupture causing an earthquake of the same size (as measured by the total area of the fault that ruptures during the earthquake and the amount of fault slip).
A comparison of the amount of energy that would be released by these two scenarios against seismological observations showed that self-healing pulses are rare; an alternative explanation is that the amount of radiated energy generated by earthquakes along plate boundaries has been dramatically underestimated.
The team also found that the physics of large earthquakes on crustal faults located within continents, such as the San Andreas Fault, may be different than that of megathrust faults in subduction zones, where one tectonic plate is forced beneath another, such as along the Japan Trench.
A few measurements of radiated energy have been obtained from earthquakes on continental crust faults. The energy released is comparable to the estimated energy released in the models of self-healing pulses, but much larger than the energy released by subduction-zone earthquakes. Both types of faults yield large earthquakes, but the forces creating those earthquakes are different-so understanding the differences rather than lumping them together will be key to developing more accurate earthquake forecast maps.
"We have a lot of data from large earthquakes along subduction zones, but the last really major earthquakes along the San Andreas were the magnitude-7.9 Fort Tejon quake in 1857 and the magnitude-7.9 San Francisco Earthquake in 1906, both of them before the age of modern seismic networks," Lapusta says.
The findings will inform physics-based models that estimate shaking and seismic hazard from future earthquakes.
The paper is titled "Propagation of large earthquakes as self-healing pulses or mild cracks." This research was funded by the National Science Foundation, the U.S. Geological Survey, and the Southern California Earthquake Center (SCEC).
Image: Oblique aerial view of San Andreas Fault in southeastern Coachella Valley, near Red Canyon; view to the west. (Credit: Michael Rymer, U.S. Geological Survey)