A Los Angeles Times article published on 11 June 1952 tells of a successful new oil well at Wheeler Ridge in Kern County in California. The well operated for 98 days, but then, on 21 July at 4:52 a.m. local time, a 7.5-magnitude earthquake let loose beneath the well along the White Wolf fault. It was the second-largest earthquake in California in the 20th century, and it killed 12 people. A team of seismologists, reporting new research, thinks the oil drilling triggered the event. The work is the first to give a detailed explanation for how industrial activity could cause such a big earthquake, the researchers said.
Taking oil out of the ground likely destabilized the White Wolf fault, triggering the Kern County quake, explained Susan Hough, a seismologist at the U.S. Geological Survey in Pasadena, Calif., and lead author of a study published this month in the Journal of Seismology.
The work follows a 2016 Bulletin of the Seismological Society of America study in which Hough and a colleague suggest that oil drilling played a role in other historic southern California earthquakes, like the deadly 1933 6.4-magnitude Long Beach earthquake that killed 120 people. That study, however, lacked an explanation for how drilling could trigger such large quakes when modern experience shows that induced quakes rarely exceed a magnitude of even 5. This time, Hough and her colleagues propose a mechanism.
Putting the Pieces Together
Hough told Eos how she stumbled across old California state reports that give detailed accounts of oil drilling activity in southern California. The reports revealed evidence for a spatial and temporal association between oil industry activity and earthquakes. “From the industry data for the [oil] production volumes and the location of the well and the location of the [White Wolf] fault, we can show that the stress change on the fault would’ve been potentially significant,” she said.
The stress change Hough refers to happened as the well pumped oil out of the ground. This, Hough explained, likely triggered the quake by “unclamping” the underlying fault. In this case, picture the fault as a fracture along an inclined plane where crustal blocks on opposite sides stall as they try to move past one another. “The fault is locked because there’s friction on the fault, and part of the reason for that is there’s the weight of the overlying crust on the fault plane,” said Hough. “But if you take some of that weight off, it shifts; it’s going to reduce the confining pressure…depending on the faults that are there, that could just destabilize what had been a locked fault.”
Liquids like oil, however, typically lubricate faults, making them more prone to slipping. So how could removing oil help trigger an earthquake? The answer lies in the structure of the rock layers beneath the well, which, Hough explained, prevented the oil’s lubricating effects from reaching the White Wolf fault. This means it was only a matter of removing the oily overburden that led to the fault destabilization.
According to the team’s calculations, the amount of oil removed from above the fault generated a stress change of about 1 bar of pressure, a value that seismologists generally think of as the amount of stress change required to set an earthquake in motion, Hough explained. “After 80 days of drilling, the stress change was right at and exceeding that magic number that we think is significant,” she said.
“They’ve developed a very plausible geologic scenario for how the Kern County earthquake could’ve been induced,” said Gillian Foulger, a geophysicist at Durham University in the United Kingdom, who was not involved in the work. “They’re really putting flesh on the bones for this particular earthquake.”
Foulger also agrees that a modest change in the overlying weight could have been enough to set off the quake. “Earthquakes are a little bit like snow avalanches,” she said. “You can have a massive amount of snow pile up on a mountainside, and then you have a skier who skis across it and that’s just enough to trigger the disturbance that causes the whole lot to fall off.”
Hough presents a model for initiating a large earthquake based on just one case example, although she thinks her work can apply to induced earthquakes in general: “It highlights the possibility that inducing any initial [earthquake] nucleation in proximity to a major fault could be the spark that detonates a larger rupture,” she said.
“Nucleation” refers to the small change in stress needed to destabilize a fault—a stress change that could happen in oil-producing regions today. But the chances of producing another temblor in the manner of the Kern County earthquake are slim, according to Hough, mostly because oil fields tend not to sit above major fault lines. In addition, oil producers long ago changed to a standard practice of injecting water into the ground after oil removal, something that was not done at the Wheeler Ridge oil field and that could have restored much of the otherwise lost weight locking the fault.
Most induced earthquakes are small—usually no bigger than a 4 magnitude—although there is no reason to suspect that humans cannot induce a big quake, explained Hough. The reason most induced quakes tend to be relatively small, she added, is that most earthquakes, in general, tend to be small. “One school of thought argues that the size distribution is the same for induced and natural earthquakes,” she said. But whether there is a maximum size limit for induced earthquakes, seismologists still do not know, she added.
An important aspect of the new work, Foulger said, is that Hough presents a model that other scientists can test, which is a first for a large induced event like the Kern County earthquake. For Seth Stein, a geophysicist at Northwestern University in Evanston, Ill., who also had no part in the study, “the take-home is that for one of the largest earthquakes that we know of in the last hundred years in California, a reasonable case can be made that it was induced.”
—Lucas Joel (email: [email protected]), Freelance Writer
Correction, 20 October 2017: An earlier version of this article incorrectly omitted the region about which a source was making a statement. The region is now mentioned in the statement.