Imaging from 22,000 small earthquakes reveals fault zone intricacies

New high-res 3D images of a fault zone show how fluid movement causes an earthquake swarm.
 

By Julie Pierce Onos (@JuliePierceOnos)
 

Citation: Pierce Onos, J. 2020, Imaging from 22,000 small earthquakes reveals fault zone intricacies, Temblor, http://doi.org/10.32858/temblor.099
 

 

Scientists are constantly monitoring fault zone activity throughout the country. Occasionally, they hear from members of the public who have noticed seismic activity and want explanations. Almost always, when the scientists receive these queries, they have an explanation or at least are already monitoring the activity. In early 2018, however, a member of the public contacted scientists in Southern California before the scientists noticed a growing pattern of low-magnitude seismicity known as a swarm. The query ultimately landed on the desk of Zachary Ross, a geophysicist at Caltech. As he and his colleagues started looking into the query, sure enough, they noticed an ongoing earthquake swarm that began in 2016 near Cahuilla, Calif., about halfway between the Elsinore Fault Zone and the San Jacinto Fault Zone.

 

 

Ross started tracking the swarm and applied a new machine-learning algorithm he had been developing to examine the architecture of faults and how earthquake swarms propagate. The resulting study, published last week in Science, provides the most detailed mapping of fault architecture to date.

 

A lengthy swarm

Most earthquake swarms last for short periods — from hours to months. When Ross started looking at it, this swarm, however, had already been continuously active for more than a year.

Swarms “are driven by processes like fluids moving through the earth or aseismic slip … and are not dominated by a large event or a stand-out event in the sequence,” Ross says. And if there are large earthquakes, he adds, they often do not occur at the beginning of the swarm. That is true of the Cahuilla sequence, which experienced its largest quake to date, a magnitude 4.4, in August 2018.

 

What started it?

Even though seismologists know that fluids are a potential cause of earthquake swarms and that unseen barriers and structures have an impact on fluid flow and seismicity, previous 2D models for general fault architecture have lacked details on constraints on processes that initiate, grow and stop the swarms.

According to journalist and earthquake researcher Kathryn Miles, who was not involved in this study, part of the reason seismologists cannot predict earthquakes is because there is much we do not know about fault zone structures, how they function or even where faults all are, since they are not always observable with the human eye.

With the new machine-learning algorithm, scientists can, in vastly greater detail, examine the characteristics and structures of fault zones. That information provides clearer insight into the effects that the physical characteristics of the fault zones have on earthquakes and aseismic processes, the changes that occur after the aseismic activity.

 

 

Pinpointing earthquake activity

The algorithm takes advantage of the more than 100 sensors placed near the Elsinore and San Jacinto fault zones that provide second-by-second information. The algorithm then detects and locates earthquakes and collects and automatically processes their data. This new process is more sensitive and therefore can detect exponentially smaller earthquakes than would have been noticed previously, Ross says. It can also precisely pinpoint the location of an earthquake and fault zone. Initially, Ross used his new methodology to evaluate the earthquake swarm starting in 2019. When it proved successful at precisely pinpointing the swarms’ location and providing more detailed imaging than what was previously available, he then used the algorithm to go back and evaluate all the data from 2016 to the present. The entire dataset included some 22,000 individual earthquakes ranging in size from magnitude 0.7 to 4.4.

The result was the most detailed 3D images of the structure of a fault zone to date. The swarm was based on a narrow, labyrinthian fault zone, about 50 meters wide with steep curves, Ross and his colleagues reported. These detailed images make it easier to distinguish whether seismic activity was triggered by seismic slip or fluid injection.

 

Fluid movement causes seismicity

The images revealed that the cause of the Cahuilla swarm was natural fluid injection from an underground reservoir and subsequent migration of the fluid along the curves of the fault zone once the reservoir started leaking. The data also suggest that all of the earthquake events have been migrating away from one single point at an average rate of 5 meters per day over the four-year period. Ross and his colleagues suggest this single point source is the fluid injection point.

 

 

The images allowed the team to see that the physical structure of the fault affected the earthquake swarm. In the past 18 months, the migration rate has slowed down considerably but it still has not stopped. “As of this point, over four years in, it’s the longest of any swarm that we have documented in Southern California going back 40 years or more,” Ross says.

Seismologists had speculated that seismicity moving through a fault zone did always not move in the same direction at the same speed, says John Vidale, a seismologist at the University of Southern California who was not involved in the new study. Since Ross and his colleagues could map this seismicity so well, they could observe the variations in movement and surmise when it was due to the physical structure of the fault zone.

With the improved resolution of images produced by Ross’ new algorithm, scientists can better observe the complex sequences that are a result of the fault zones’ dynamic physical characteristics, Vidale says. Additionally, he says, the information on stress drops — the difference between stress across a fault before and after an earthquake — will help researchers more clearly see when to attribute changes in stress drops to the physical structure of a fault zone, changes in fluid movement or other causes. These data add much more information about causes of earthquakes to the existing catalog of earthquake information that can be later used for hazard forecasting.

Further research, Vidale says, could include figuring out how fluid flows where there is not seismicity to track its progress. “We’d like to know where the fluid pressure is changing because that could tell us where there is increased likelihood of dangerous earthquakes.”

 

Want to know your earthquake risk? Check it at Temblor.
 

Further Reading

Ross, Z.E., Cochran, E.S., Trugman, D.T., and Smith, J.D., (2020). 3D fault architecture controls the dynamism of earthquake swarms. Science, 368(6497), 1357–1361. https://doi.org/10.1126/science.abb0779.