Along the coast of California, a sleeping giant lays in wait. The San Andreas Fault System, a massive 750-mile long strike-slip fault system, forms the boundary between the Pacific Plate and the North American Plate. Over the last 30 million years, it has spawned the transverse mountain ranges of Southern California. It moves an average of ~30 mm per year — which sounds slow, but accommodates a staggering amount of tectonic stress. And periodically, this stress is relieved, generating a massive earthquake, typically greater than magnitude 7.    

We know these earthquakes happen — they’re recorded in both the geologic and recent instrumental record. But a large earthquake hasn’t happened in Southern California for nearly 170 years, prompting seismologists to wonder what amount of stress is sufficient to wake the southern San Andreas up. 

A new JGR Solid Earth study led by a team from the University of Hawaii has gotten one step closer to understanding the San Andreas’ tipping point. The team modeled stress accumulation along parts of the southern San Andreas Fault and neighboring San Jacinto Fault, finding that these stress levels are at their highest of the last 1000 years.

A Piece in the Puzzle

Over the past twenty years, earthquake research in Southern California has strongly suggested an earthquake of magnitude 7 or greater is on the horizon. The paleoseismic record shows that the major faults associated with the plate boundary tend to rupture every 100 to 300 years, depending on the segment. 

There are many pieces of the puzzle that seem to suggest the timing is about right for the San Andreas’ next big shift — from estimation of strain accumulation from InSAR data to pressure exerted on fault planes by giant paleolakes.

The last large earthquake in Southern California was the Fort Tejon earthquake in 1857 — and compared to the past, it seems that the San Andreas and its neighboring faults have been unusually quiet since then. This relative quiescence has prompted suspicion that the region is likely overdue. But earthquakes don’t operate on a schedule. They’re notoriously difficult to predict, and deviation from a long-term average is typical. 

Instead of counting down the years to the next earthquake, scientists can try to build models that estimate the conditions sufficient to trigger the next large rupture on the fault. And geodetic data that measures the current rate of plate motion has a role to play in strengthening this understanding, too.

All across the western United States, a dense network of GNSS stations has been established to measure tectonic deformation caused by the complex plate motion along the coast. This network, called the Network of the Americas, has provided earthquake scientists with highly accurate geodetic slip rates of the San Andreas, making it possible for research to use these rates for estimations of stress and strain accumulation. In fact, NOTA data has already been used in earthquake “forecasts” by the USGS to estimate the 30-year risk of large earthquakes all along the San Andreas. So naturally, these slip rates could be used to help understand how past large earthquakes came to be.

Grounded in Reality

To create a comprehensive understanding of the San Andreas’ earthquake cycle grounded in reality, the University of Hawaii team used a physics-based model of earthquake stress loading coupled with paleoseismological records of past earthquakes to understand how the San Andreas Fault and its neighbor, the San Jacinto Fault, are linked. The team hypothesizes that the junction of the two faults, at Cajon Pass near Los Angeles, could act as an “earthquake gate” — certain levels of stress allow an earthquake to occur across the junction and others stop the rupture in its tracks. 

Building on a previous rupture model constrained by a robust 31-site paleoseismic record, the team incorporated this history into a four-dimensional earthquake stress model to estimate how much stress could be built up and released across these large faults in a joint rupture. The model creates a simplified sliver of the tectonic plates and underlying upper mantle. The plates and associated fault systems slip at the rate recorded by NOTA GNSS stations across the state. When an earthquake ruptures in the model, stress is released at the surface. The underlying plate motion continues, which continues to fuel stress loading in the crust, priming the faults for their next big quake.

While accurate to a degree by incorporating general behaviors of plate and fault motion, the model is still a simplification. It’s designed to capture some characteristics of the earth that have not typically been used in similar models, like the plasticity of the asthenosphere below the crust and gravity’s effect on restoring crust. The team’s model more realistically simulates the physics of crustal deformation, so plate motion is more accurately factored into earthquake processes.  

But as in any good model, assumptions must be made for this one to work. The team assumes that slip rate along portions of each fault are uniform, and that the earthquake record of Southern California is complete and accurate. Neither of these things are true. Slip rate along the San Andreas is not the same across every segment, with different sections creeping while others are locked, causing stress to build. 

The model also does not account for the different seismicity rates of the San Andreas and San Jacinto faults. For example, while the southern portion of the San Andreas moves at roughly the same speed as the San Jacinto, the San Andreas has more small earthquakes per year than the San Jacinto. Seismologists aren’t sure why this is the case, and this missing understanding could change the model’s results. 

Additionally, the paleoseismic record, while extensive, is prone to uncertainty. While large earthquakes are pretty good at leaving behind evidence in the geologic record, much of the last 1000 years of seismic activity was not instrumentally recorded, so estimates of rupture length and dates are not exact. 

Lastly, there is no set periodicity in earthquakes of the last millennium. But, compared to the historical record in Southern California, it has indeed been more than double the usual amount of time between earthquakes. It is this timing that drives the higher stress accumulation in the model, all other things constant.  

While the results show that stress along the faults seems to be at an historic high, it doesn’t tell us that now is the time for the next big SoCal earthquake — nor was this the aim of the study. When analyzing the stress states modeled before each past earthquake, the team noted that stress along both the San Andreas and the San Jacinto faults had to be similar enough for the “earthquake gate” to open and allow a rupture along both. Though conditions seem to be right, it still remains a mystery as to what exactly will cause the tipping point to be reached.

No Reason for Alarm

The team’s results are consistent with other lines of evidence suggesting that Southern California should prepare for a violent temblor, but modeled high stress levels do not spell immediate disaster. It’s important to understand the difference between model and reality — while the Hawaii team’s results are bracketed with real, empirical evidence from the geologic record and high-precision geodetic measurements, they are not real in situ measures of fault stress. Rather, they are a highly educated estimate, reminding all of us who live in proximity to a major fault zone that we shouldn’t be scared, but we should be ready.