Lighting up earthquakes: How scientists watched ruptures evolve in milliseconds

"We've essentially opened a window into the heart of earthquake mechanics," said Sylvain Barbot, associate professor of earth sciences at the USC Dornsife College of Letters, Arts and Sciences and principal investigator of the study. "By watching how the real contact area between fault surfaces evolves during the earthquake cycle, we can now explain both the slow buildup of stress in faults and the rapid rupture that follows. Down the road, this could lead to new approaches for monitoring and predicting earthquake nucleation at early stages."

For decades, scientists have relied on empirical "rate-and-state" friction laws to model earthquakes -- mathematical descriptions that work well but don't explain the underlying physical mechanisms. "Our model reveals what's actually happening at the fault interface during an earthquake cycle."

Barbot says the discovery is a deceptively simple concept: "When two rough surfaces slide against each other, they only make contact at minuscule, isolated junctions covering a fraction of the total surface area." This "real area of contact" -- invisible to the eye but measurable through optical techniques -- turns out to be the key state variable that controls earthquake behavior.

Laboratory earthquakes: Lighting earthquakes in real time

The study utilizes transparent acrylic materials that allowed the researchers to literally watch earthquake ruptures unfold in real time. Using high-speed cameras and optical measurements, the team tracked how LED light transmission changed as contact junctions formed, grew and were destroyed during laboratory earthquakes.

"We can literally watch the contact area evolve as ruptures propagate," Barbot said. "During fast ruptures, we see approximately 30% of the contact area disappear in milliseconds -- a dramatic weakening that drives the earthquake."

The laboratory results revealed a previously hidden relationship: The empirical "state variable" used in standard earthquake models for decades represents the real area of contact between fault surfaces. This discovery provides the first physical interpretation of a mathematical concept that has been central to earthquake science since the 1970s.

From simulation to prediction

The researchers analyzed 26 different simulated earthquake scenarios and found that the relationship between rupture speed and fracture energy follows the predictions of linear elastic fracture mechanics. The team's computer simulations successfully reproduced both slow and fast laboratory earthquakes, matching not only the rupture speeds and stress drops but also the amount of light transmitted across the fault interface during ruptures.

As contact areas change during the earthquake cycle, they affect multiple measurable properties including electrical conductivity, hydraulic permeability and seismic wave transmission. Since the real area of contact affects multiple physical properties of fault zones, continuous monitoring of these proxies during earthquake cycles could provide new insights into fault behavior.

The implications extend far beyond academic understanding and laboratory experiments. The research suggests that monitoring the physical state of fault contacts could provide new tools for earthquake short-term systems and potentially for reliable earthquake prediction using the electric conductivity of the fault.

"If we can monitor these properties continuously on natural faults, we might detect the early stages of earthquake nucleation," Barbot explained. "This could lead to new approaches for monitoring earthquake nucleation at early stages, well before seismic waves are radiated."

Looking ahead

The researchers plan to scale up their findings outside controlled laboratory conditions. Barbot explained: The study's model provides the physical foundation for understanding how fault properties evolve during seismic cycles.

"Imagine a future where we can detect subtle changes in fault conditions before an earthquake strikes," Barbot said. "That's the long-term potential of this work."

About the study

In addition to Barbot, Baoning Wu, formerly at USC and now at the University of California, San Diego, authored the study.

The study was supported by National Science Foundation award number EAR-1848192 and the Statewide California Earthquake Center proposal number 22105.