The Science Behind OK’s Man-made Earthquakes, Part 1

By Ariel Conn, Seismologist and Science Writer with the Virginia Tech Department of Geosciences

On April 21, the Oklahoma Geological Survey issued a statement claiming that the sharp rise in Oklahoma earthquakes — from only a couple per year to thousands — was most likely caused by wastewater disposal wells associated with major oil and gas plays. This is huge news after years of Oklahoma scientists hesitating to place blame on an industry that provides so many jobs.

Now, seismologists from around the country — including Oklahoma — are convinced that these earthquakes are the result of human activity, also known as induced or triggered seismicity. Yet many people, especially those in the oil industry, still refute such an argument. Just what is the science that has seismologists so convinced that the earthquakes are induced and not natural?

Hidden Faults

Over the last billion years (give or take a couple hundred million), colliding tectonic plates have created earthquake zones, just as we see today in California, Japan, Chile and Nepal. As geologic processes occurred, these zones shifted and moved and were covered up, and the faults that once triggered earthquakes achieved a state of equilibrium deep in the basement rocks of the earth’s crust. But the faults still exist. If the delicate balance that keeps these fault systems stable ever shifts, the ancient faults can still move, resulting in earthquakes. Because these inactive faults are so deep, and because they can theoretically exist just about anywhere, they’re incredibly difficult to map or predict – until an earthquake occurs.

Thanks to historic reports of earthquakes in the central and eastern United States, we know there are some regions, far away from tectonic plate boundaries, that occasionally experience large earthquakes. Missouri and South Carolina, for example, suffered significant and damaging earthquakes in the last 200 hundred years, yet these states lie nowhere near a plate boundary. We know that fault zones exist in these locations, but we have no way of knowing about dormant faults in regions of the country that haven’t experienced earthquakes in the last couple hundred years.

What is induced seismicity?

As early as the 1930s, seismologists began to suspect that extremely large volumes of water could impact seismic activity, even in those regions where earthquakes weren’t thought to occur. Scientists found that after certain reservoirs were built and filled with water, earthquake swarms often followed. This didn’t happen everywhere, and when it did, the earthquakes were rarely large enough to be damaging. These quakes were large enough to be felt, however, and they represented early instances of human activity triggering earthquakes.[1]

Research into induced seismicity really picked up in the 1960s. The most famous example of man-made earthquakes occurred as a result of injection well activity at the Rocky Mountain Arsenal. The arsenal began injecting wastewater into a disposal well 12,000 feet deep in March of 1962, and by April of that year, people were feeling earthquakes. Researchers at the arsenal tracked the injections and the earthquakes. They found that each time the arsenal injected large volumes of water (between 2 and 8 million gallons per month, or 47,000 to 190,000 barrels), earthquakes would start shaking the ground within a matter of weeks (Figure 1).

Rocky Mountain Arsenal fluid injection correlated to earthquake frequency

Figure 1. Rocky Mountain Arsenal fluid injection correlated to earthquake frequency

South Carolina experienced induced earthquakes after filling a reservoir

Figure 2. South Carolina experienced induced earthquakes after filling a reservoir

When the injections ended, the earthquakes also ceased, usually after a similar time delay, but some seismicity continued for a while. The well was active for many years, and the largest earthquake thought to be induced by the injection well actually occurred nearly a year and a half after injection officially ended. That earthquake registered as a magnitude 5.3. Scientists also noticed that over time, the earthquakes moved farther and farther away from the well.

Research at a reservoir in South Carolina produced similar results; large volumes of water triggered earthquake swarms that spread farther from the reservoir with time (Figure 2).

When people say we’ve known for decades that human activity can trigger earthquakes, this is the research they’re talking about.

Why now? Why Oklahoma?

Class II Injection Well. Photo by Lea Harper

Injection Well in Ohio. Photo by Ted Auch

Seismologists have known conclusively and for quite a while that wastewater injection wells can trigger earthquakes, yet people have also successfully injected wastewater into tens of thousands of wells across the country for decades without triggering any earthquakes. So why now? And why in Oklahoma?

The short answers are:

  • At no point in history have we injected this much water this deep into the ground, and
  • It’s not just happening in Oklahoma.

One further point to clarify: General consensus among seismologists is that most of these earthquakes are triggered by wastewater disposal wells and not by hydrofracking (or fracking) wells. That may be a point to be contested in a future article, but for now, the largest induced earthquakes we’ve seen have been associated with wastewater disposal wells and not fracking. This distinction is important when considering high-pressure versus high-volume wells. A clear connection between high-pressure wells and earthquakes has not been satisfactorily demonstrated in our research at the Virginia Tech Seismological Observatory (VTSO) (nor have we seen it demonstrated elsewhere, yet). High-volume wastewater disposal wells, on the other hand, have been connected to earthquakes.

At the VTSO, we looked at about 8,000 disposal wells in Oklahoma that we suspected might be connected to induced seismicity. Of those, over 7,200 had maximum allowed injection rates of less than 10,000 barrels per month, which means the volume is low enough that they’re unlikely to trigger earthquakes. Of the remaining 800 wells, only 300 had maximum allowed injection rates of over 40,000 barrels per month — and up to millions of barrels per year for some wells. These maximum rates are on par with the injection rates seen at the Rocky Mountain Arsenal, and our own plots indicate a correlation between high-volume injection wells and earthquakes (Figure 3-4).

Triangles represent wastewater injection wells scaled to reflect maximum volume rates. Wells with high volumes are located near earthquakes.

Figure 3. Triangles represent wastewater injection wells scaled to reflect maximum volume rates. Wells with high volumes are located near earthquakes.

Triangles represent wastewater injection wells scaled to reflect maximum pressure. Wells with high pressures are not necessarily near earthquakes.

Figure 4. Triangles represent wastewater injection wells scaled to reflect maximum pressure. Wells with high pressures are not necessarily near earthquakes.

This does not mean that all high-volume wells will trigger earthquakes, or that lower-volume wells are always safe, but rather, it’s an important connection that scientists and well operators should consider.

Starting in 2008 and 2009, with the big oil and gas plays in Oklahoma, a lot more fluid was injected into a lot more wells. As the amount of fluid injected in Oklahoma has increased, so too have the number of earthquakes. But Oklahoma is not the only state to experience this phenomenon. Induced earthquakes have been recorded in Arkansas, Colorado, Kansas, New Mexico, Ohio, West Virginia and Texas.

In the last four years, Arkansas, Kansas, Ohio and Texas have all had “man-made” earthquakes larger than magnitude 4, which is the magnitude at which damage begins to occur. Meanwhile, in that time period, Colorado experienced its second induced earthquake that registered larger than magnitude 5. Oklahoma may have the most induced and triggered earthquakes, but the problem is one of national concern.


[1] Induced seismicity actually dates back to the late 1800s with mining, but the connection to high volumes of fluid was first recognized in the 1930s. However, the extent to which it was documented is unknown.