The rate of earthquakes has increased in central Oklahoma since 2009, culminating in the damaging quake of November 5, 2011 near Prague, OK.
The US Geological Survey and the Oklahoma Geological Survey are studying the problem, as there's a correlation between the quakes and human activity associated with the petroleum industry.
“While it’s been known for decades that Oklahoma is ‘earthquake country’, we hope that this new advisory of increased hazard will become a crucial consideration in earthquake preparedness for residents, schools and businesses in the area,” said Dr. Bill Leith, Senior Science Advisor for Earthquakes and Geologic Hazards at USGS. “Building owners and government officials should have a special concern for older, unreinforced brick structures, which are vulnerable to serious damage during sufficient shaking.”
USGS statistically analyzed the recent earthquake rate changes and found that they do not seem to be due to typical, random fluctuations in natural seismicity rates. Significant changes in both the background rate of events and earthquake triggers needed to have occurred in order to explain the increases in seismicity, which is not typically observed when modeling natural earthquakes.
The analysis suggests that a likely contributing factor to the increase in earthquakes is triggering by wastewater injected into deep geologic formations. This phenomenon is known as injection-induced seismicity, which has been documented for nearly half a century, with new cases identified recently in Arkansas, Ohio, Texas and Colorado. A recent publication by the USGS suggests that a magnitude 5.0 foreshock to the 2011 Prague, Okla., earthquake was human-induced by fluid injection; that earthquake may have then triggered the mainshock and its aftershocks. OGS studies also indicate that some of the earthquakes in Oklahoma are due to fluid injection. The OGS and USGS continue to study the Prague earthquake sequence in relation to nearby injection activities.
With the large increase in petroleum and gas extraction in the eastern US, there has been an increase of earthquakes larger than M3 since 2010. The vast majority of these quakes have occurred in Oklahoma, Arkansas, Colorado, Texas, and Ohio. A curious question comes to mind: Why these states and not North Dakota, Pennsylvania, or other states where fracking is occurring? The answer may reveal a complex picture.
The November 5, 2011 quake was felt across much of the central US. It locally damaged a number of homes in the epicentral region, as well as a few large buildings at a college nearby. No one was injured or killed.
It was part of a sequence that day. Earlier, a smaller quake shook the region. The main shock was Mw5.7, and a large aftershock followed. It is one of the few eastern quakes that can be attributed to a known fault system. The Wilzetta is now a household name in central Oklahoma.
It appears likely that this sequence was induced by wastewater injection in an oil field not far from the quake's epicenter. The first quake was induced by 18 years of water injection. The first quake then triggered the larger main quake. That quake then triggered a vigorous sequence of aftershocks that continue to this day, and it perhaps triggered a great many more to the south.
Induced. Triggered. Fracking. Injection. All of these sound pretty confusing, right? To get to the bottom of this, and to explain in a way I think may help, we have to look at an analog clock.
The clock will represent the stress that builds up in the upper crust of the Earth, everywhere. This stress acts on all faults that cross through the crust and along the Earth's surface. Every fault has a clock. Sometimes, several clocks, as many faults act in segments. One segment's clock may "tick" at a different rate than others. The fault-clocks of the earth all tick at different time-scales compared to others. When the clocks reach midnight, that means the shear stress has overcome the fault's resistance to slip. An earthquake occurs.
Please keep in mind that I'm greatly simplifying some very complex physics, and I'm probably not doing it the proper justice.
The San Andreas, for example, has perhaps four large segments--a Northern one, two Central ones, and the Southern one. The Northern one last unzipped in 1906. There's evidence its clock is connected to Cascadia--except the 1906 quake was out of sequence. The Parkfield segment had an earthquake in 2004. Its clock runs pretty fast, perhaps every 22 to 36 years, and the quakes rarely exceed Mw 6. The next segment on ruptured in 1857 and the southern segment ruptured perhaps in 1680. The southern segment's clock may be close to "midnight." That means an earthquake, perhaps a very large one, could occur relatively soon. (I'm told the phrase being used is "10 months pregnant." It's...an interesting way to present risk and probability.) There's also evidence the Southern segment's clock is "connected" to the filling of the lake that occasionally fully fills the valley the Salton Sea now sits in. The lake fills. The earthquake happens. Humans prevented the lake from fully filling by preventing the Colorado River from doing so. Like I said, it's complicated!
That's the other thing about the clocks. We have no idea what time it is on them. Because earthquake prediction is not possible, and likely never will be, we can only make statistical forecasts at what time it might be on any given fault segment.
Another thing about the clocks: they're easily disturbed. An earthquake in one location can "wind the clocks forward" at a distance. So, one clock reaches midnight: boom, an earthquake. It then can wind the clocks all the way down a long fault so that each segment fails in sequence as the stress is transferred to already almost-critical faults. These sequences do happen naturally on a large scale; it's been happening over the last several decades in Turkey.
And yes, people can disturb those clocks too. Dams, mining, groundwater extraction for drinking and irrigation, quarrying, geothermal energy production, oil and gas extraction and fluid injection all can and do and have "disturbed the clocks" of faults nearby. I would love to talk about ALL of them, but we're going to focus on the petroleum industry and most specifically, fluid injection, since that is what seems to be a factor in a good many of the quakes in Oklahoma.
Even though the entire process is fairly complicated, it can be explained pretty easily. The injection of fluid into a well, especially under pressure, raises the pressures beneath the ground. It isn't that the fluids are lubricating the faults. That's not what is happening. Instead the fluid injection and its pressures change the stress conditions that are acting on any possible subsurface faults. Because all of the continents have an overall stress and strain field due to their motions relative to each other, a fault that is oriented in a favorable fashion may find its clock advanced significantly forward if the local stress and strain field is changed.
This is what happened in Ohio through much of 2011.
A series of quakes jolted the Youngstown area through most of 2011, an area of Ohio that was generally aseismic. The quakes eventually illuminated a previously unknown fault, and were very closely correlated with an injection well. The well was injecting wastewater from Marcellus Shale operations in Pennsylvania. When the injection stopped and the well was plugged, the quakes eventually stopped. Because there was very clear data (Ohio requires injectors to document very closely the amount of fluid being injected along with the pressures--Oklahoma does not), it was not difficult to confirm that the injection induced the swarm of quakes, the largest of which was Mw 4.1. The paper about this will be released sometime soon, so stay tuned.
Ohio is one of a few states that actually has regulations regarding induced seismicity. They recently shut down an operation southeast of Youngstown when they proved it was causing a number of tiny earthquakes. Oklahoma, to date, does not,preferring a "wait and see" approach.
Now, there is a difference between induced and triggered seismicity, in the case of people causing earthquakes. Induced quakes probably would not have occurred otherwise if not for some anthropogenic cause. Triggered quakes occur when small changes generated by some anthropogenic activity are sufficient enough to move a fault system from a quasi-critical state to an unstable one---in effect, winding the clock that was already close to midnight forward. The difference probably means little to the average person whose house is shaking when they otherwise thought it wouldn't be, but it matters scientifically, and very likely legally too. In the coming months I expect you're going to see a lot of people wrangling with this very difference.
They already are in Italy.
A sequence of quakes in the Emilia region occurred from May 20th to May 29th 2012. This sequence, while of moderate magnitude, was very damaging and it was deadly. Moderate earthquakes and Italian construction don't seem to mix.
The Emilia region lies in the Po River Valley and this region is, according to seismic hazard maps of Italy, not really at risk. Historical records say otherwise, and there are blind but active faults that underlie the valley. The region's tectonics are fairly complicated and I won't go into them here, but suffice to say the active blind folds and thrusts beneath the valley trap a lot of gas and oil.
There is a field where natural gas is being stored. This is a common practice worldwide--it is why Pennsylvania exports its wastewater from the Marcellus Shale that isn't being reused or dumped at sewage plants. Most available wells in Pennsylvania are used for natural gas storage, and under some conditions, this can and has induced or triggered quakes. A commission that studied the quakes concluded the gas storage field did not induce or trigger the Emilia sequence.
The commission did cast its eye at an active field at Mirandola and concluded it was possible that the activities there triggered the sequence, or more importantly, they triggered the first quake, which then wound the clocks all the way down a fault whose clock was already close to midnight.
“The Commission believes that it is highly unlikely that the activities of hydrocarbon exploitation at Mirandola and the geothermal activity at Casaglia have produced sufficient stress change to generate an ‘induced’ seismic event. While it cannot constitute proof, the current state of knowledge and all the processed and interpreted information does not allow the ruling out of the possibility that the actions involved in hydrocarbon exploitation in the Mirandola field may have contributed to ‘trigger’ the Emilia seismic activity."
The Economist notes that the Italian government sat on this report for two months, and it eventually was leaked. The Italian government has previously prosecuted seismologists for failing to do the impossible (that is, predict an earthquake), so I can't say I'm at all surprised.
Here, like everywhere, the difference between induced and triggered is important. The report in question is in fairly plain English and can be downloaded here. It's a PDF, so rightclick and save.
That is what's being asked in Oklahoma, which unlike the Po Valley, is far from a plate boundary. It's still complicated. Did people induce quakes? The answer in some cases, like the Eola Fieldand like the Jones, OK swarm, seems to be yes. An abstract presented at the Seismological Society of America meeting in Anchorage last week reads:
Earthquake relocations and hydrogeologic modeling show that the Jones earthquake swarm, occurring near Oklahoma City since 2008, is linked to disposal wells injecting high volumes of water along the Nemaha Fault. Felt and recorded earthquakes in the Jones swarm began in 2008, approximately 15 km from four high-volume wastewater disposal wells. These wells dispose of ~2-3 million barrels per month (4-5 million barrels per month cumulatively) in two adjacent locations on the downthrown side of the Nemaha fault. Earthquakes are observed to migrate away from these high-volume disposal wells up the structural dip and down hydraulic gradient. Hydrogeologic modeling shows that the increase in subsurface pore pressure resulting from the fluid injection is of sufficient magnitude to trigger slip on pre-existing faults. The region of increased pore pressure grows outward through time with injection. The larger, mapped faults in the subsurface may act as conduits or guides to fluid flow, and may transmit fluid pressure into basement. Our results demonstrate that wastewater disposal can raise fluid pressure and trigger earthquakes at tens of kilometers from the wellbore on existing faults.
Or, translated, these wells dispose of a lot of water, in locations along a fault. The fluids travel along the faults away from the wells, and they change the stress fields to wind the clocks along pre-existing faults so that they reach midnight.
Elsewhere, the answer appears to be that one may have been triggered, one that may have happened otherwise, and then that quake then set off even more because the clocks were already at two to midnight. On top of that, the fluids can travel along faults, because they act as conduits (and anyone who lives in an arid region knows this. You find a fault--you find water). This is one of the many reasons, human triggered, induced, or otherwise, that the USGS will raise Oklahoma's overall seismic hazard in the next National Seismic Hazard Map release. That map and its documentation will be released this year.
This is what I keep thinking:
1. The earthquakes are entirely natural
2. They're induced entirely by human activity
3. They're natural, but indicative (warning) of a coming, larger and more significant quake i.e., the Wilzetta Fault, which appears to be the culprit, has turned itself on, perhaps due to stress transfer from the New Madrid quakes that occurred to its east 200 years ago.
4. Hypothesis 3, but the threat perhaps increased, because the "clock was advanced" due to hypothesis 2. (I've changed this from the last time I wrote about this.)
I don't know what's going to happen, but if I lived in Oklahoma, I'd purchase earthquake insurance if I could do so, just in case, even though you may never use it. It's complicated, but some fault's clock may be at two to midnight, and getting closer because of anthropogenic activity.
Here are some links to follow so that you can educate yourself and then perhaps educate some legislators so there can be some action taken on this.
The Oklahoma Geological Survey has collected a great many documents. The two white papers linked, if you click the second link, are in plain English, very informative, and also link to actual presentations that you can download. I did laugh reading through the one presented by an Exxon representative though.
I've liked the NPR station in Oklahoma City's coverage on this.
The US Department of Energy has a webpage on induced seismicity.
A report titled Induced Seismicity Potential in Energy Technologies is available for free here.
Seek out scientific groups on facebook if you use facebook and if they're open to the public. An example is the Pacific Northwest Seismic Network.
The EPA (which is also preparing a report on induced and triggered quakes) has a program that monitors injection wells. You can find its website here. The wells in question are mostly Class II wells.
A lot of these studies are behind paywalls. That in itself is a subject for another diary but there are ways around it. If you don't have access to an academic library and your local library system can't help, #ICanHazPDF, if you have twitter, is one way. Simply post the abstract link, and an email address along with the hashtag. I've found that people in the geological sciences are very responsive, happy to teach and share, and they have a large presence on twitter relative to the other sciences. When you get your paper, delete the tweet. Do not thank your academic angel in public.
Or even better, email the authors and ask them. They usually are happy to share.
The lack of federal regulation around this problem is a big problem, and even if and after a large and damaging quake occurs in Oklahoma, I don't know if there will be any action taken. That shouldn't stop people from getting informed and then making sure the people they or their neighbors elected are informed too, because this is an environmental and social problem.