December 2022 California earthquake ruptured unknown fault: An analysis

The quake produced severe shaking in a lightly populated area, brought the southern tip of the Cascadia Subduction Zone slightly closer to failure, and modestly inhibited rupture on the San Andreas Fault.
 

By Ross S. Stein, Ph.D., Temblor, Inc., Shinji Toda, Ph.D., Tohoku University, Chris Rollins, Ph.D., GNS Science New Zealand, and Volkan Sevilgen, M.Sc., Temblor, Inc.
 

Citation: Stein, R. S., Toda, S., Rollins, C., and Sevilgen, V., 2023, December 2022 California earthquake ruptured unknown fault: an analysis, Temblor, http://doi.org/10.32858/temblor.294
 

At 2:34 a.m. local time on Dec. 20, 2022, a magnitude-6.4 earthquake hit Ferndale, Calif. Severe shaking lasted for 15 to 20 seconds, leading indirectly to the deaths of two people, extensive building and infrastructure damage, and power blackouts throughout the region. As geologists began to analyze the event, they found landslides and cracks in roads — but no signs of liquefaction or, intriguingly, an obvious rupture at the surface (California Geological Survey, 2022). The earthquake triggered an Earthquake Early Warning alert, issued by the U.S. Geological Survey (USGS). These alerts ideally provide several precious seconds of notice to those who could experience strong shaking. Whether advance warning reached those within about 28 miles (45 kilometers) from the epicenter (those who experienced strong shaking) is not yet clear.

Scientifically, the quake was not a surprise: According to the Temblor model, there’s a 5% chance that a magnitude-6.4 earthquake will occur once per year in this region. This hyperactive seismic region has the strongest “once-in-a-lifetime” quake that can be expected compared to anywhere in the United States (even Alaska!): magnitude 7.2. That quake has about a 1% chance of occurring per year. But the severe rupture of a hidden, unmapped fault in this seismically active region nevertheless raises concern over current hazard risk assessments and methods that are based on mapped faults.
 

Why is the Cape Mendocino area so seismically active?

Ferndale and the surrounding 60-by-60-mile (100-by-100-kilometer) area make up the most seismically active region of California, with one magnitude-6.4 or larger quake every decade, on average. Thankfully, it is not densely populated, and its residents and governments are well prepared for earthquakes.

Ferndale sits atop a geologic “triple junction,” where three tectonic plates collide — in this case, the North American, Juan de Fuca, and Pacific Plates are colliding in the Mendocino Triple Junction, shown in the figure below. The collisions lead to Earth’s crust being compressed and stretched, and that distortion is exacerbated by two factors: the Juan de Fuca Plate subducting (sliding underneath) unevenly under the North American Plate, and the 80-degree bend the San Andreas Fault makes when it juts out into the Pacific. The Ferndale and surrounding Cape Mendocino areas are laced by faults.
 

The magnitude-6.4 quake struck within the Mendocino Triple Junction, where three tectonic plates meet. Because the junction is distorted and compressed, deformation is distributed across a network of faults, one of which ruptured on Dec. 20, 2022. Historically, the junction has been extremely seismically active, but fortunately it is more lightly populated and better prepared than most areas in the United States. The arrows give the sense of motion of the faults. The color ramp gives the magnitude (within 100 kilometers) that’s likely in a lifetime. Note that likely maximum magnitude within 100 kilometers of the Mendocino region in a lifetime is magnitude 7.2.
The magnitude-6.4 quake struck within the Mendocino Triple Junction, where three tectonic plates meet. Because the junction is distorted and compressed, deformation is distributed across a network of faults, one of which ruptured on Dec. 20, 2022. Historically, the junction has been extremely seismically active, but fortunately it is more lightly populated and better prepared than most areas in the United States. The arrows give the sense of motion of the faults. The color ramp gives the magnitude (within 100 kilometers) that’s likely in a lifetime. Note that likely maximum magnitude within 100 kilometers of the Mendocino region in a lifetime is magnitude 7.2.

 

While the San Andreas Fault might have broad name recognition, over the past 58 years, the Mendocino Triple Junction has actually experienced six quakes of the 21 in the U.S. that topped magnitude 6.4 (large gray circles, which do not scale with magnitude), while the San Andreas has only experienced one.  So, despite the prevalence of the San Andreas in hazard assessments, it has taken a back seat to the Mendocino Triple Junction.
While the San Andreas Fault might have broad name recognition, over the past 58 years, the Mendocino Triple Junction has actually experienced six quakes of the 21 in the U.S. that topped magnitude 6.4 (large gray circles, which do not scale with magnitude), while the San Andreas has only experienced one. So, despite the prevalence of the San Andreas in hazard assessments, it has taken a back seat to the Mendocino Triple Junction.

 

The magnitude-6.4 quake struck on a northeast-trending left-lateral strike-slip fault, approximately 12 miles (20 kilometers) long. (Left-lateral means if you’re standing on one side of the fault, the other side appears to move to the left; right-lateral, like the San Andreas, is opposite.) The fault cuts across the grain of the mapped surface faults in this region, which are right-lateral and trend northwest (Kelsey and Carver, 1988; Clarke and Carver, 1992). The left-lateral fault probably does not extend to Earth’s surface, consistent with the apparent absence of surface rupture.
 

The big surprise: The aftershock trend, which strikes northeast, is almost perpendicular to the mapped faults, which strike northwest. The aftershock alignment is also remarkably straight, whereas the mapped faults are sinuous and discontinuous.
The big surprise: The aftershock trend, which strikes northeast, is almost perpendicular to the mapped faults, which strike northwest. The aftershock alignment is also remarkably straight, whereas the mapped faults are sinuous and discontinuous.

 

The inconsistency between the mapped faults and the likely fault that slipped in this quake underscores how difficult it is to assess seismic hazard on the basis of mapped faults — even in superbly mapped California.
 

The currently available evidence, including seismic waveforms recorded worldwide, the GPS (‘GNSS’) displacements (shown here as red and black arrows) and the aftershock alignment (shown in the first figure), suggests that the magnitude-6.4 quake struck on an northeast-trending left-lateral strike-slip fault. The slip on that fault is shown here in dark red and orange (USGS Event Page). Note that the rupture is about about 9 miles (15 kilometers) long — much shorter than the length of the white rectangle.
The currently available evidence, including seismic waveforms recorded worldwide, the GPS (‘GNSS’) displacements (shown here as red and black arrows) and the aftershock alignment (shown in the first figure), suggests that the magnitude-6.4 quake struck on an northeast-trending left-lateral strike-slip fault. The slip on that fault is shown here in dark red and orange (USGS Event Page). Note that the rupture is about about 9 miles (15 kilometers) long — much shorter than the length of the white rectangle.

 

Very strong shaking where people live

The magnitude-6.4 earthquake produced strong shaking and damage over an area of about 600 square miles (about 1,000 square kilometers). Recall that 1 g is the acceleration that we feel because of Earth’s gravity. In simplified terms, more than 1 g means you’re pushed down, whereas less than 1 g means you’re being pulled upward. If, on the freeway, you slam on the brakes of your car, you experience about 100% g, which is equal to 1 g. Roughly speaking, a peak ground acceleration greater than or equal to 20% g is delineated by the orange contour in the next figure.

Although not densely populated, homes and buildings are concentrated along the coastal plain near Eureka and along the Eel River Valley. Both are sites of young, unconsolidated (loose) sediments, in which severe shaking occurred. Seismic stations in those areas measured shaking of 30% to 140% g, whereas the three closest seismic stations to the epicenter and peak slip registered only between 12% and 26% g. This means shaking in the loose sediments was greatly amplified, despite being farther from the epicenter. For context, in California, buildings are typically susceptible to damage beginning around 10% g; damage becomes widespread at 20% g and severe at 50% g.
 

In these maps that are zoomed in on the area around Ferndale, both show different metrics for shaking. The distribution of instrumentally observed shaking for this event is exceptionally well recorded, and the most reliable indicators of shaking are seismic station recordings, shown on the left (USGS). Here the contours indicate Peak Ground Acceleration (PGA); in California, building damage begins at about 10% g, becomes widespread at 20% g, and highly destructive at 50% g. Notice that the greatest shaking is not closest to the epicenter or site of peak slip. The colored circles show “Did You Feel It”  estimates for shaking. On the right figure, which shows Temblor’s site amplification model (STAMP ) the river valleys and coastal estuaries, which are filled with unconsolidated sediments, all show up as red areas that amplify seismic waves, which means more shaking.
In these maps that are zoomed in on the area around Ferndale, both show different metrics for shaking. The distribution of instrumentally observed shaking for this event is exceptionally well recorded, and the most reliable indicators of shaking are seismic station recordings, shown on the left (USGS). Here the contours indicate Peak Ground Acceleration (PGA); in California, building damage begins at about 10% g, becomes widespread at 20% g, and highly destructive at 50% g. Notice that the greatest shaking is not closest to the epicenter or site of peak slip. The colored circles show “Did You Feel It” estimates for shaking. On the right figure, which shows Temblor’s site amplification model (STAMP) the river valleys and coastal estuaries, which are filled with unconsolidated sediments, all show up as red areas that amplify seismic waves, which means more shaking.

 

What could happen next?

Rollins and Stein (2010) argued that since the 1970s, earthquakes in the Mendocino Triple Junction have occurred in a kind of chain reaction in which each large quake transferred stress to the site of a subsequent event. The interactions are clear when compared to a control group in which the quakes were assumed to be independent of each other. But the myriad right-lateral, left-lateral and thrust faults in this area make anticipating the next quake very difficult. Nevertheless, we attempt to assess future possibilities.

The Dec. 20, 2022, magnitude-6.4 quake imparted stress to surrounding faults, which can act to promote or inhibit subsequent earthquakes. One approach is to resolve the Coulomb stress changes on the major nearby faults (left panels, next graphic). We find that failure is modestly promoted on a portion of the Cascadia megathrust, a 600-mile (1,000-kilometer)-long fault zone known for its seismic risk to the Pacific Northwest (top left). Failure is also promoted on a 15.5-miles-long (25-kilometer-long) section of the Mendocino Fault Zone near its junction with the San Andreas, and more strongly along a 15.5-mile-long (25-kilometer-long) section of California’s Little Salmon and Bear River Faults (bottom left).
 

Calculations, via Coulomb 3.3 (Toda et al., 2011), of the stress imparted by the magnitude-6.4 earthquake to major mapped faults (left), and to the faults on which past earthquakes struck (right). The focal mechanisms in the upper right panel are from the Global CMT catalog, from 1976 to 2022. Here, we assume the fault friction is 0.4.
Calculations, via Coulomb 3.3 (Toda et al., 2011), of the stress imparted by the magnitude-6.4 earthquake to major mapped faults (left), and to the faults on which past earthquakes struck (right). The focal mechanisms in the upper right panel are from the Global CMT catalog, from 1976 to 2022. Here, we assume the fault friction is 0.4.

 

Another way to assess the stress changes is to resolve them on past earthquake focal mechanisms (the “beachballs” in the right panels, above). This calculation lends additional support to our inference that the easternmost Mendocino Fault was brought closer to failure, as were the Little Salmon Fault and the offshore portion of the Bear River Fault.
 

The Bottom Line

That the Ferndale earthquake occurred was not surprising scientifically, but it is surprising that the fault that ruptured bears no resemblance to the faults mapped at the surface. This is not good news for fault-based seismic hazard assessments. The last large quake in California, the 2019 magnitude-7.1 Ridgecrest quake, also ruptured a heretofore unknown fault (Toda and Stein, 2020). Both point to a potential weakness in seismic hazard assessment.

Looking ahead, there are several sections of right-lateral faults, including the Mendocino, Bear River and Little Salmon, that we calculate were brought slightly (about 0.3 bar) closer to failure by the Ferndale quake. A 186-square-mile (300-square-kilometer) section of the Cascadia megathrust fault was also brought closer to failure by about the same amount, and so these are possible sites for future quakes. But the good news is that the northern tip of the San Andreas is calculated to be inhibited (by about 0.25 bar) by the Ferndale rupture.

In short: It’s hard to know exactly how one earthquake will affect future quakes, but our analyses suggest the Ferndale quake could have made quakes along other small faults around the Mendocino Triple Junction slightly more likely. It also could have made a fault failure along the large Cascadia Subduction Zone slightly more likely. Rupture along the northernmost San Andreas Fault may be slightly less likely.
 

Acknowledgments. We thank Jay Patton of the California Geological Survey for sharing with us insights about the field reconnaissance.
 

References

California Geological Survey (2022), California Earthquake Clearinghouse:
https://learningfromearthquakes.org/2022-12-20-ferndale-ca/resources

Samuel H. Clarke, Jr., and Gary A. Carver (1992), Late Holocene tectonics and paleoseismicity, southern Cascadia Subduction Zone, Science, 255, 188-192, doi: 10.1126/science.255.5041.188

Harvey M. Kelsey and Gary A. Carver (1988). Late Neogene and Quaternary tectonics associated with northward growth of the San Andreas transform fault, Northern California, J. Geophys. Res., 93, 4797–4819. https://doi.org/10.1029/JB093iB05p04797

Carolyn P. Nuyen and David A. Schmidt (2022). Strain partitioning among forearc faults in southern Cascadia inferred from GNSS, J. Geophys. Res., 127, e2022JB024236. https://doi. org/10.1029/2022JB024236.

John C. Rollins and Ross S. Stein (2010), Coulomb stress interactions among M ≥ 5.9 earthquakes in the Gorda deformation zone and on the Mendocino Fault Zone, Cascadia subduction zone, and northern San Andreas Fault, J. Geophys. Res., 115, doi:10.1029/2009JB007117.

Shinji Toda and Ross S. Stein (2020), Long- and short-term stress interaction of the 2019 Ridgecrest sequence and Coulomb-based earthquake forecasts, Bull. Seismol. Soc. Am., 110, 1765–1780, doi: 10.1785/0120200169.

Shinji Toda, Ross S. Stein, Volkan Sevilgen, and Jian Lin (2011), Coulomb 3.3 Graphic-rich deformation and stress-change software for earthquake, tectonic, and volcano research and teaching, U.S. Geological Survey Open-File Report 2011–1060, 63 p., https://temblor.net/coulomb/

U.S. Geological Survey Event Page, M 6.4 Ferndale Earthquake, https://earthquake.usgs.gov/earthquakes/eventpage/nc73821036/executive