Over the course of 1 month in the spring of 2016, scientists at the U.S. Geological Survey (USGS) detected about 15,000 tiny earthquakes in an area in northwest Oklahoma about the size of New York City (about 800 square kilometers). These tiny earthquakes all had magnitudes of less than 3.
Many of the teeny tremors were too small to feel, but when the scientists analyzed the quakes, they got a more detailed picture of what happens when petroleum operations dispose of their wastewater by reinjecting it into the ground. They already knew that the resulting seismic activity is the major cause of the increase in earthquakes in the central United States, but the miniearthquakes allowed them to track the seismic activity in time and space in much greater detail. Sara Dougherty, a former postdoc at the USGS who is now at the California Institute of Technology, presented these findings last week at AGU’s Fall Meeting 2018 in Washington, D. C.
Oil drilling operations extract from the ground a mix of petroleum, salt water, and anything else that gets carried along. After filtering out the petroleum, the wastewater gets reinjected into the ground: hundreds of millions of barrels each year.
More than 10,000 reinjection wells in the state penetrate Earth’s crust, reaching layers of porous rock such as limestone or sandstone. Through these wells, wastewater—salty, oily water that may contain toxic contaminants—is forced into the soft rock, causing shifts in the surrounding layers and prying apart existing faults. The latter can trigger quakes, a phenomenon known as induced seismicity.
Digging into the Details
Dougherty wanted to see in more detail what seismic activity looked like after a wastewater injection event, but the existing earthquake data collected by the U.S. Geological Survey recorded only earthquakes of magnitude 3 and above. A magnitude 3 quake is strong enough to rattle dishes in your kitchen cabinet but still small enough that some observers might mistake it for a truck passing by on the highway. These earthquakes definitely indicated a trend in induced seismicity, but Dougherty wanted more information on the quakes themselves. She wondered, What about earthquakes smaller than magnitude 3?
“Really, what we were trying to do was to get into details,” she said. “We have all these really, really tiny earthquakes that no one had really looked at before, so we decided do that.”
Using an array of more than 1,800 sensors laid out over street grids in Grant County, on the Oklahoma-Kansas border north of Oklahoma City, Dougherty and her coauthors created their own catalog of around 15,000 microearthquakes that rattled the area in the spring of 2016.
The sensors, white cylinders about the size of coffee cans, tracked seismic activity for a month. Normal earthquake detection systems might use fewer than 100 instruments, but Dougherty’s study used input from 1,800: N equals 1,800, in statistical lingo.
“Our large-N [sensor] array has served as a sort of test case for what can be accomplished with such a large number of sensors and what kind of data we can get out of it,” she said.
With the information they gathered, the researchers were able to look in more depth at how earthquakes traveled out from a specific site and to map clusters that occurred over the month.
Zach Rosson, a graduate research assistant at the Oklahoma Geological Survey who also studies induced seismicity in Oklahoma, said that Dougherty’s catalog of microearthquakes is a valuable resource for learning about earthquake behavior.
“The earthquake catalog built up in Oklahoma for the past decade has been very ad hoc,” he said. “We lose out on a lot of information about small earthquakes, and small earthquakes—particularly the ones below magnitude 2—are very important to discerning how earthquakes cluster and how they evolve in space in time.”
Earthquake Sweet Spot
Dougherty’s data allowed her to track waves of tiny quakes as they radiated out from certain points. These, Dougherty said, might be wastewater disposal sites or just small fault lines.
During a microearthquake cluster, quakes started out at shallower depths and then moved deeper into Earth. Over the next 24 hours, they also migrated farther away from the site. The following 24 hours, Dougherty found, were generally quiet. Then quakes started up again.
“You have this 24-hour sweet spot—a burst of activity in 24 hours followed by a 24-hour pause, then another burst of activity in 24 hours—and then it goes back to the background rate,” she said.
During Dougherty’s experimental period, there were no recorded hydraulic fracturing events—the petroleum companies weren’t injecting fluids into the ground to extract oil or gas. That means that these results are most likely tied to wastewater disposal. Dougherty’s next steps, she said, are to see if the tiny quakes are, in fact, the direct result of wastewater injection.
“One thing we’re focusing on now is identifying any characteristic patterns to the sequences of earthquakes that we’ve identified and exploring if these patterns can be correlated to any specific wastewater injection or oil and gas production behavior,” she said.
In 2016, after studies revealed the extent of the human activity–induced earthquakes in Oklahoma, policy changed to tighten regulations on wastewater injections. Still, seismicity levels remained high.
Dougherty said the new level of detail achieved in her research could further influence oil and gas regulations in the area. For example, she noted, the work can help pinpoint wells that trigger high rates of microearthquakes.
“We were able to find new pockets of seismicity that [the USGS] was not able to pick up on the regular network,” she said. “They may not know that these wells are also causing earthquakes; the [earthquakes] are just not big enough. So maybe it’s something they should watch more closely.”
—Eva Frederick ([email protected]), Science Writing Student, Massachusetts Institute of Technology, Cambridge