Fracking’s Hidden Hazards

Late on a Saturday evening in November 2011, Sandra Ladra was reclining in a chair in her living room in Prague, Oklahoma, watching television with her family. Suddenly, the house started to shake, and rocks began to fall off her stone-faced fireplace, onto the floor and into Ladra’s lap, onto her legs, and causing significant injuries that required immediate medical treatment.

The first tremor that shook Ladra’s home was a magnitude-5.0 earthquake, an unusual event in what used to be a relatively calm state, seismically speaking. Two more struck the area over the next two days. More noteworthy, though, are her claims that the events were manmade. In a petition filed in the Lincoln County District Court, she alleges that the earthquake was the direct result of the actions of two energy companies, New Dominion and Spress Oil Company, that had injected wastewater fluids deep underground in the area.

House damage in central Oklahoma from a magnitude 5.7 earthquake on November 6, 2011

Ladra’s claim is not as preposterous as it may seem. Scientists have recognized since the 1960s that humans can cause earthquakes by injecting fluids at high pressure into the ground. This was first established near Denver, Colorado, at the federal chemical weapons manufacturing facility known as the Rocky Mountain Arsenal. Faced with the thorny issue of how to get rid of the arsenal’s chemical waste, the U.S. Army drilled a 12,044-feet-deep disposal well and began routinely injecting wastewater into it in March 1962.

Less than seven weeks later, earthquakes were reported in the area, a region that had last felt an earthquake in 1882. Although the Army initially denied any link, when geologist David Evans demonstrated a strong correlation between the Arsenal’s average injection rate and the frequency of earthquakes, the Army agreed to halt its injections.

Since then direct measurements, hydrologic modeling, and other studies have shown that earthquakes like those at the Rocky Mountain Arsenal occur when injection increases the fluid pressure in the pores and fractures of rocks or soil. By reducing the frictional force that resists fault slip, the increased pore pressure can lubricate preexisting faults. This increase alters the ambient stress level, potentially triggering earthquakes on favorably oriented faults.

Although injection-induced earthquakes have become commonplace across broad swaths of the central and eastern U.S over the last few years, building codes—and the national seismic hazard maps used to update them—don’t currently take this increased hazard into account. Meanwhile, nagging questions—such as how to definitively diagnose an induced earthquake, whether manmade quakes will continue to increase in size, and how to judge whether mitigation measures are effective—have regulators, industry, and the public on shaky ground.

Surge in Seismicity

The quake that shook Ladra’s home is one example of the dramatic increase in seismicity that began across the central and eastern U.S. in 2001. Once considered geologically stable, the midcontinent has grown increasingly feisty, recording an 11-fold increase in the number of quakes between 2008 and 2011 compared with the previous 31 years, according to a study published in Geology in 2013.

The increase has been especially dramatic in Oklahoma, which in 2014 recorded 585 earthquakes of magnitude 3.0 or greater—more than in the previous 35 years combined. “The increase in seismicity is huge relative to the past,” says Randy Keller, who retired in December after serving for seven years as the director of the Oklahoma Geological Survey (OGS).

Yesterday, Oklahoma finally acknowledged that the uptick in earthquakes is likely due to wastewater disposal. “The Oklahoma Geological Survey has determined that the majority of recent earthquakes in central and north-central Oklahoma are very likely triggered by the injection of produced water in disposal wells,” the state reported on a new website. While the admission is an about-face for the government, which had previously questioned any link between the two, it doesn’t coincide with any new regulations intended to stop the earthquakes or improve building codes to cope with the tremors. For now, residents of Oklahoma may be just as vulnerable as they have been.

This live-updated map shows all earthquakes magnitude 2.5 and greater for the last 30 days.

This surge in seismicity has been accompanied by a spike in the number of injection wells and the corresponding amount of wastewater disposed via those wells. According to the Railroad Commission of Texas, underground wastewater injection in Texas increased from 46 million barrels in 2005 to nearly 3.5 billion barrels in 2011. Much of that fluid has been injected in the Dallas area, where prior to 2008, only one possible earthquake large enough to be noticed by people had occurred in recorded history. Since 2008, the U.S. Geological Survey (USGS) has documented over 120 quakes in the area.

The increase in injection wells is due in large part to the rapid expansion of the shale-gas industry, which has unlocked vast new supplies of natural gas and oil that would otherwise be trapped in impermeable shale formations. The oil and gas is released by a process known as fracking, which injects a mix of water, chemicals, and sand at high enough pressure to fracture the surrounding rock, forming cracks through which the hydrocarbons, mixed with large volumes of fluid, can flow. The resulting mixture is pumped to the surface, where the hydrocarbons are separated out, leaving behind billions of gallons of wastewater, much of which is injected back underground.

Many scientists, including Keller, believe there is a correlation between the two increases. “It’s hard to look at where the earthquakes are, and where the injection wells are, and not conclude there’s got to be some connection,” he says. Rex Buchanan, interim director of the Kansas Geological Survey (KGS), agrees there’s a correlation for most of the recent tremors in his state. “Certainly we’re seeing a huge spike in earthquakes in an area where we’ve also got big disposal wells,” he says. But there have been other earthquakes whose cause “we’re just not sure about,” Buchanan says.

Diagnosing an Earthquake

Buchanan’s uncertainty stems in part from the fact that determining whether a specific earthquake was natural or induced by human activity is highly controversial. Yet this is the fundamental scientific question at the core of Ladra’s lawsuit and dozens of similar cases that have been filed across the heartland over the last few years. Beyond assessing legal liability, this determination is also important for assessing potential seismic hazard as well as for developing effective methods of mitigation.

One reason it’s difficult to assess whether a given earthquake was human-induced is that both types of earthquakes look similar on seismograms; they can’t be distinguished by casual observation. A second is that manmade earthquakes are unusual events; only about 0.1 percent of injection wells in the U.S. have been linked to induced earthquakes large enough to be felt, according to Arthur McGarr, a geologist at the USGS Earthquake Science Center. Finally, scientists have comparatively few unambiguous examples of induced earthquakes. That makes it difficult to create a yardstick against which potential “suspects” can be compared. Like a team of doctors attempting to diagnose a rare disease, scientists must examine all the “symptoms” of an earthquake to make the best possible pronouncement.

To accomplish this, two University of Texas seismologists developed a checklist of seven “yes” and “no” questions that focus on four key characteristics: the area’s background seismicity, the proximity of an earthquake to an active injection well, the timing of the seismicity relative to the onset of injection, and the injection practices. Ultimately, “if an injection activity and an earthquake sequence correlate in space and time, with no known previous earthquake activity in the area, the earthquakes were likely induced,” wrote McGarr and co-authors in Science earlier this year.

Oilfield waste arrives by tanker truck at a wastewater disposal facility near Platteville, Colorado.

These criteria, however, remain open to interpretation, as the Prague example illustrates. Ladra’s petition cites three scientific studies that have linked the increase in seismicity in central Oklahoma to wastewater injection operations. A Cornell University-led study, which specifically examined the earthquake in which Ladra claims she was injured, concluded that event began within about 200 meters of active injection wells—closely correlating in space—and was therefore induced.

In a March 2013 written statement, the OGS had concluded that this earthquake was the result of natural causes, as were two subsequent tremors that shook Prague over the next few days. The second earthquake, a magnitude-5.7 event that struck less than 24 hours later, was the largest earthquake ever recorded in Oklahoma.

The controversy hinged on several of the “symptoms,” including the timing of the seismicity. Prior to the Prague sequence, scientists believed that a lag time of weeks to months between the initiation of injection and the onset of seismicity was typical. But in Prague, the fluid injection has been occurring for nearly 20 years. The OGS therefore concluded that there was no clear temporal correlation. By contrast, the Cornell researchers decided that the diagnostic time scale of induced seismicity needs to be reconsidered.

Another key issue that has been raised by the OGS is that of background seismicity. Oklahoma has experienced relatively large earthquakes in the past, including a magnitude-5.0 event that occurred in 1952 and more than 10 earthquakes of magnitude 4.0 or greater since then, so the Prague sequence was hardly the first bout of shaking in the region.

The uncertainty associated with both these characteristics places the Prague earthquakes in an uncomfortable middle ground between earthquakes that are “clearly not induced” and “clearly induced” on the University of Texas checklist, making a definitive diagnosis unlikely. Meanwhile, the increasing frequency of earthquakes across the midcontinent and the significant size of the Prague earthquakes are causing scientists to rethink the region’s potential seismic hazard.

Is the Public at Risk?

Earthquake hazard is a function of multiple factors, including event magnitude and depth, recurrence interval, and the material through which the seismic waves propagate. These data are incorporated into calculations the USGS uses to generate the National Seismic Hazard Maps.

Updated every six years, these maps indicate the potential for severe ground shaking across the country over a 50-year period and are used to set design standards for earthquake-resistant construction. The maps influence decisions about building codes, insurance rates, and disaster management strategies, with a combined estimated economic impact totaling hundreds of billions of dollars per year.

When the latest version of the maps was released in July, the USGS intentionally excluded the hazard from manmade earthquakes. Part of the reason was the timing, according to Nicolas Luco, a research structural engineer at the USGS. The maps are released on a schedule that dovetails with building code revisions, so they couldn’t delay the charts even though the induced seismicity update wasn’t ready, he says.

Such changes, however, may take years to implement. Luco notes that the building code revisions based upon the previous version of the USGS hazard maps, released in 2008, just became law in California in 2014, a six-year lag in one of the most seismically-threatened states in the country.

Instead, the USGS is currently developing a separate procedure, which they call a hazard model, to account for the hazard associated with induced seismicity. The new model may raise the earthquake hazard level substantially in some parts of the U.S. where it has previously been quite low, according to McGarr. But there are still open questions about how to account for induced seismicity in maps of earthquake shaking and in building codes, Luco says.

McGarr believes that the new hazard calculations will result in more rigorous building codes for earthquake-resistant construction and that adhering to these changes will affect the construction as well as the oil, gas, and wastewater injection industries. “Unlike natural earthquakes, induced earthquakes are caused by man, not nature, and so the oil and gas industry may be required to provide at least some of the funds needed to accommodate the revised building codes,” he says.

But Luco says it may not make sense to incorporate the induced seismicity hazard, which can change from year to year, into building codes that are updated every six years. Over-engineering is also a concern due to the transient nature of induced seismicity. “Engineering to a standard of earthquake hazard that could go away, that drives up cost,” says Justin Rubinstein, a seismologist with the USGS Earthquake Science Center. A further complication, according to Luco, is that building code changes only govern new construction, so they don’t upgrade vulnerable existing structures, for which retrofit is generally not mandatory.

The occurrence of induced earthquakes clearly compounds the risk to the public. “The risk is higher. The question is, how much higher?” Luco asks. Building codes are designed to limit the risk of casualties associated with building collapse—“and that usually means bigger earthquakes,” he says. So the critical question, according to Luco, is, “Can we can get a really large induced earthquake that could cause building collapses?”

Others are wondering the same thing. “Is it all leading up to a bigger one?” asks Keller, former director of the OGS. “I don’t think it’s clear that it is, but it’s not clear that it isn’t, either,” he says. Recalling a magnitude-4.8 tremor that shook southern Kansas in November, KGS’ Buchanan agrees. “I don’t think there’s any reason to believe that these things are going to magically stop at that magnitude,” he says.

Coping with Quakes

After assessing how much the risk to the public has increased, our society must decide upon the best way to cope with human-induced earthquakes. A common regulatory approach, one which Oklahoma has adopted, has been to implement “traffic light” control systems. Normal injection can proceed under a green light, but if induced earthquakes begin to occur, the light changes to yellow, at which point the operator must reduce the volume, rate of injection, or both to avoid triggering larger events. If larger earthquakes strike, the light turns red, and further injection is prohibited. Such systems have recently been implemented in Oklahoma, Colorado, and Texas.

But how will we know if these systems are effective? The largest Rocky Mountain Arsenal-related earthquakes, three events between magnitudes 5.0 and 5.5, all occurred more than a year after injection had ceased, so it’s unclear for how long the systems should be evaluated. Their long-term effectiveness is also uncertain because the ability to control the seismic hazard decreases over time as the pore pressure effects move away from the well, according to Shemin Ge, a hydrogeologist at the University of Colorado, Boulder.

Traffic light systems also rely on robust seismic monitoring networks that can detect the initial, very small injection-induced earthquakes, according to Ge. To identify hazards while there is still sufficient time to take corrective action, it’s ideal to identify events of magnitude 2.0 or less, wrote McGarr and his co-authors in Science. However, the current detection threshold across much of the contiguous U.S. is magnitude 3.0, he says.

Kansas is about to implement a mitigation approach that focuses on reducing injection in multiple wells across areas believed to be underlain by faults, rather than focusing on individual wells, according to Buchanan. He already acknowledges that it will be difficult to assess the success of this new approach because in the past, the KGS has observed reductions in earthquake activity when no action has been taken. “How do you tease apart what works and what doesn’t when you get all this variability in the system?” he asks.

This climate of uncertainty leaves regulators, industry, and the public on shaky ground. As Ladra’s case progresses, the judicial system will decide if two energy companies are to blame for the quake that damaged her home. But it’s our society that must ultimately decide how, and even if, we should cope with manmade quakes, and what level of risk we’re willing to accept.