Study Finds Correlation Between Injection Wells and Small Earthquakes

Aug. 6, 2012

AUSTIN, Texas —

Map of earthquakes and injection wells in Barnett Shale of North Texas

Map showing earthquake epicenters determined in this study (red circles), injection wells (squares and + symbols) in use since October 2006, seismograph stations (white triangles), and mapped faults (green lines). Circle sizes indicate quality of epicentral location, with large, medium and small sizes indicating qualities A, B and C. For injection wells, yellow squares are wells with maximum monthly injection rates exceeding 150,000 barrels of water per month (BWPM); white squares, exceeding 15,000 BWPM; + symbols, exceeding 1,500 BWPM. Credit: Cliff Frohlich/U. of Texas at Austin.

Most earthquakes in the Barnett Shale region of North Texas occur within a few miles of one or more injection wells used to dispose of wastes associated with petroleum production such as hydraulic fracturing fluids, according to new research from The University of Texas at Austin. None of the quakes identified in the two-year study were strong enough to pose a danger to the public.

The study by Cliff Frohlich, senior research scientist at the university’s Institute for Geophysics, appears this week in the journal Proceedings of the National Academy of Sciences.

"You can't prove that any one earthquake was caused by an injection well," says Frohlich. "But it's obvious that wells are enhancing the probability that earthquakes will occur."

Frohlich analyzed seismic data collected between November 2009 and September 2011 by the EarthScope USArray Program, a National Science Foundation-funded network of broadband seismometers from the Canadian border to the Gulf of Mexico. Because of the high density of instruments (25 in or near the Barnett Shale), Frohlich was able to detect earthquakes down to magnitude 1.5, far too weak for people to feel at the surface.

He found that the most reliably located earthquakes — those that are accurate to within about 0.9 miles (1.5 kilometers) — occurred in eight groups, all within 2 miles (3.2 kilometers) of one or more injection wells. Before this study, the National Earthquake Information Center had only identified two earthquake groups in the area strongly associated with specific injection wells. This suggests injection-triggered earthquakes are far more common than is generally recognized.

The Barnett Shale is a geological formation in North Texas bearing a large amount of natural gas that was difficult to recover prior to recent technological advances such as hydraulic fracturing. The formation lies beneath Dallas and Fort Worth and extends over several counties, mostly to the west of those cities. Development of the Barnett Shale and other unconventional plays — such as the Haynesville Shale in Texas and Louisiana and the Marcellus Shale in Pennsylvania, New York and West Virginia — have spurred dramatic growth in domestic natural gas production.

This study comes as some policymakers and members of the public are expressing concern about possible environmental and health effects of hydraulic fracturing. Most earthquakes identified in the study ranged in magnitude from 1.5 to 2.5, meaning they posed no danger to the public.

"I didn't find any higher risks from disposal of hydraulic fracturing fluids than was thought before," says Frohlich. "My study found more small quakes, nearly all less than magnitude 3.0, but just more of the smaller ones than were previously known. The risk is all from big quakes, which don't seem to occur here."

All the wells nearest to the eight earthquake groups reported high injection rates (maximum monthly injection rates exceeding 150,000 barrels of water). Yet in many other areas where wells had similarly high injection rates, there were no earthquakes. Frohlich tried to address those differences.

Location of Barnett Shale and area covered in accompanying map

Texas map showing the Barnett Shale (gray) and rectangle indicating region mapped in figure 2. Credit: Cliff Frohlich/U. of Texas at Austin.

"It might be that an injection can only trigger an earthquake if injected fluids reach and relieve friction on a nearby fault that is already ready to slip," says Frohlich. "That just isn't the situation in many places."

Hydraulic fracturing is an industrial process in which water and various chemicals are pumped deep underground in order to fracture rock, allowing oil or gas to more easily flow to a well. As petroleum is produced at the surface, most hydraulic fracturing fluids return to the surface too. Frohlich is careful to point out that he did not evaluate the possible correlation of earthquakes with the actual hydraulic fracturing process, but rather the effects of disposing of fracturing fluids and other wastes in these injection wells.

Support for this study came from the U.S. Geological Survey and the Jackson School of Geosciences at The University of Texas at Austin. The author has no financial ties to the hydraulic fracturing industry. Frohlich has consulted for the construction industry on seismic risks for projects including dams, power plants and pipelines. He plans to participate in a future study relating to hydraulic fracturing in the Barnett Shale by the university's Energy Institute.

For more information, contact: Marc Airhart, College of Natural Sciences, 512 232 1066.

13 Comments to "Study Finds Correlation Between Injection Wells and Small Earthquakes"

1.  Matt Dameron said on Aug. 6, 2012

This could be a boon to the local economy. We will need to build earthquake-tolerant infrastructure. Jobs!

2.  Bob Wernly said on Aug. 6, 2012

Nowhere did I see the confidence level used in the test and the relevance of the two time periods of the earthquakes. This study seems very shallow. Earthquakes are a function of the time being monitored at the a location so that has to be factored into their study. Something so important should be done more accurately.

3.  Drew said on Aug. 6, 2012

There's no way to tell that the fracturing process is causing earthquakes. If fracturing occurs in places that have stratified rock, then those locations probably had earthquakes before. Any historical data to compare earthquake activity before fracturing?

4.  Francis Milholland said on Aug. 6, 2012


5.  John Wittmaak said on Aug. 7, 2012

There is an interesting point raised here...albeit common sense when you think about it. Hydraulic fracturing can can relieve stress in rock formations resulting in these tiny earthquakes. Large earthquakes are caused by a buildup of stress in rock formations that release all at once. Could hydraulic fracturing be used to gradually relieve major fault lines around the world resulting in more small earthquakes rather than a single large earthquake every 10 years?

6.  Mike Scott said on Aug. 7, 2012

People this study is not about hydraulic fracturing it is about waste water injection. Hydraulic fracturing occurs in a few minutes to a few hours and is over. Injection wells run continuously for years and routinely dispose in a day more than most frac jobs ever use. I know that whatever journalist wrote up this report wanted to find something sexy to write about. But there is little new here. It has nothing to do with fracking, period.

7.  Katherine said on Aug. 7, 2012

@Francis: Drinking water is actually not being harmed by the fracking process. Any contamination of drinking water is caused by mishandling of materials used in fracking or by other non-fracking related causes. There is no scientifically proven connection between the process of fracking and water contamination. Additionally, the "poster-child" of the cause you seem to champion, Dimock, PA, has been re-evaluated by the EPA. In case you missed the news that day, the EPA found that there was hardly any water contamination in the area and what contamination there was did not occur due to fracking.

8.  Carolyn Deason said on Aug. 7, 2012

We live on FM 3009, 4 mi west of IH 35 North, towards Natural Bridge Caverns. Our old house is on 1/2 acre, has been here for 40 years. I am at home and have noticed frequent 'shaking' more so than ever. Blame has been for the most part, due to dynamite use in the Hanson Quarry, but due to very pricey homes in the area, that was supposed to be stopped. We still feel the shaking, sometimes the dog barks. We are on a fault line so maybe that's why we feel tremors, our old windows shake, you can feel these funny sensations in the air. I wish someone would investigate the source because it's happening too often, sometimes the shakes are worse than others, lasts a few seconds. Thank you.

9.  Eideard said on Aug. 7, 2012

#5, John Wittmaak, raises a point that has been suggested before in areas of large slippage, serious earthquakes.

I don't recall if any experiments or pilot operations ever happened - mostly because of lack of availability several years ago of water for such projects - lack of money for science [as usual]. This study may be a boon towards getting research like that a boost.

As for the whines about drinking water - when we're discussing wells 13K' in depth - ignorance is bliss, eh?

10.  Jefferson Williams said on Aug. 9, 2012

Nowhere did I see the confidence level used in the test and the relevance of the two time periods of the earthquakes. This study seems very shallow. Earthquakes are a function of the time being monitored at the a location so that has to be factored into their study. Something so important should be done more accurately.

That's probably because you read a reporter's synopsis of the article rather than the original article itself. How often do you see reporters writing for the lay public mention 95% confidence interval ? Most of the public doesn't know what that means.

11.  GogogoStopSTOP said on Aug. 10, 2012

1. The well locations appear as a scattergram vs/ the epicenters.
2. The 'proof' of a correlation would require a 3 dimensional analysis.
3. Any positive or negative conclusions would have to include a temporal correlation, showing "I drilled a well, I fracked the well... I felt a quake!"

But, my conclusion is better, more simple, wrt the scattergram. I conclude: There is utterly no positive or negative correlation between fracking, fracking pressure or fracking location... None, zero, nada!

The only 'proof' that fracking is hazardous is that it's done by capitalistic corporations, it involves CO2 & it produces a cheap energy source... Q. E. D.

12.  Diane Falk said on Aug. 13, 2012

Earthquakes and fracking

13.  Roger Matson said on Aug. 26, 2012

Injection caused quakes are very old news, From the Colorado Geologic Survey's newsletter, RockTalk:
"In the late 1950s, liquid waste was stored in ponds at the U.S. Army’s Rocky Mountain Arsenal, famous for its store of nerve gas during the Cold War. In order to alleviate environmental concerns, they decided to inject the liquid into a two-mile deep well. Less than a year after injection began earthquakes began occurring in the vicinity.* Thousands of small earthquakes (in 1967 two earthquakes over magnitude 5.0) were recorded near the
Arsenal. The largest caused an estimated $1 million in damage in Commerce City and north Denver. After a couple of years of this earthquake activity a geologist in Denver claimed that the volume of liquid being injected into the Arsenal disposal well correlated with the number of earthquakes occurring in the area; the greater the volume of injected liquid, the higher the number of earthquakes (Figure 7). The Army denied it, many geologists doubted it, and the USGS set out instruments to prove that he was wrong. Instead, they proved that this Denver geologist was correct. Fifteen years later the President’s Council on Environmental Quality recognized the geologist, David Evans, with a $50,000 award."
(Evans associated the quakes with the Front Range fault west of Denver. The disposal well was sited east of Denver near Stapleton airport.

Evans, David M. 1966. “The Denver Earthquakes and the Rocky Mountain Arsenal Disposal Well.” The Mountain Geologist. Vol. 3, No. 1. p. 23