Before 2008, Oklahoma experienced roughly one noticeable earthquake per year. By 2014, that number had soared to almost one a day, and the state is not alone. Scientists have documented an astronomical rise in seismic activity across the central and eastern United States, linking it to wastewater pumped into the ground from burgeoning oil and gas production. Now, new research suggests that high rates of fluid injection—rather than other factors such as volume or depth—may be the root of the problem.
Since the 1960s, geologists have recognized that fluid injection can induce earthquakes by raising the hydraulic pressure along a fault, pushing the two sides apart to let the crust slide along it. But scientists don't know exactly which aspect of well operation matters most.
Until now, most studies have focused on individual earthquakes and their connection to nearby wells. But that approach can make it hard to spot the difference between wells that are linked to quakes and those that aren't, says Matthew Weingarten, a Ph.D. student at the University of Colorado, Boulder, and lead author of the new study, published in this week's issue of Science. “You would never know that cancer or other diseases are more likely in smokers if you don't look at the vast majority of other people who don't smoke,” he says.
So Weingarten's team compiled the first comprehensive data set of all injection wells operating in the central and eastern United States. They looked both at wells used for enhanced oil recovery—in which fluid is injected to flush lingering oil from a depleted reservoir—and at those used to dispose of wastewater from conventional oil and gas extraction or from hydraulic fracturing (fracking).
The scientists found that disposal wells were 1.5 times more likely to be associated with earthquakes, although the region contains far more enhanced recovery wells. The link was strongest at higher injection rates, above about 300,000 barrels per month. Other potentially important factors—such as the pressure at the well-head, the total volume of fluid injected, and whether fluid was injected near basement rock—did not appear to make much difference at a regional scale, the researchers say.
They propose that injecting fluid quickly may induce earthquakes by jacking up reservoir pressure more drastically and over a larger area than adding it more slowly would. The elevated pressures increase the chances of triggering slip on a nearby fault that is already under natural stress. “You're looking at a balance between the rate you're putting fluid in and the rate it can diffuse away,” says Katie Keranen, a seismologist at Cornell University who was not involved in the study. That would explain why wells used for enhanced oil recovery are less likely to trigger earthquakes: Pumping out oil as water is injected helps keep the pressure in check.
Weingarten says the team's results suggest that limiting injection rates may reduce earthquake hazard. Previous studies that showed a potential link have already spurred some states to take action. In Kansas, regulators issued an order in March to ramp down injection rates in three hard-hit areas. Since then, earthquake activity appears to have quieted down, says Rex Buchanan, the interim director of the Kansas Geological Survey in Lawrence, although it's too early to say for sure.
The Oklahoma Corporation Commission has also adopted a “traffic light” system to force operators to decrease or even halt injection at potentially quake-prone wells. “All options available to us to address this are on the table,” said Tim Baker, director of Oklahoma's Oil and Gas Conservation Division in Oklahoma City, in a statement issued last month.
Another new study, published this week in Science Advances, found that Oklahoma's earthquake activity has increased in areas where disposal rates have sky rocketed. However, it raises questions about the benefits of reducing the injection rates at individual wells. “A number of wells injecting right next to each other could have the same effect as one well injecting their combined volume,” says Rall Walsh, a Ph.D. student in geophysics at Stanford University in Palo Alto, California, and the lead author of the study. For neighboring wells, he says, the question is, “how close is close?”
Nor does the new research explain why large areas of the country—including North Dakota, the Gulf Coast, and the Michigan Basin—have experienced few earthquakes despite having disposal wells with high injection rates. Maybe fluids aren't reaching the bedrock there, or background stress levels in the crust are lower, says William Ellsworth, a geophysicist with the U.S. Geological Survey in Menlo Park, California. Understanding the reason could help well operators and regulators “assess the hazard of a project perhaps before it has begun,” Ellsworth says.
The new study sets the stage for work that may provide more answers about the relatively new problem of induced earthquakes. “I'm very happy they did it in a time frame that actually makes sense so people can discuss it,” says Murray Hitzman, a geologist at the Colorado School of Mines in Golden, who chaired the National Research Council's committee on induced seismicity. “All this is happening really fast.”