Tuesday, November 8, 2011

How To Cause An Earthquake, And Other Assorted Ramblings

It’s been known for quite some time that human activities can induce seismicity. Dams create lakes behind them that can weigh down the crust (as can flooded quarries) Mines can collapse. Gas and oil extraction can do the same. Geothermal energy projects induce seismicity as well. This isn’t exactly controversial, and hasn’t been for several decades. On nature’s side, heavy rainfall can induce earthquakes as well.

Since natural gas extraction has lately become a serious environmental and partisan political issue, and big earthquakes worldwide have made big news over the last two years, I thought I’d write a brief review about what’s known, and what’s not.

First, I’d be remiss if I didn’t start this diary out by taking a scientific stand, and explicitly spelling out my stance on the phenomenon in relation to a series of moderate quakes that have rattled and caused significant damage to areas east of the Rockies over the last couple years. I wouldn’t want to be accused of being a shill for the gas companies, be questioned about my motives because of my employer (those who know me know who I work for, and my employer is rather incensed by the extra work and significant damage the drillers are doing to infrastructure in my state), and so on. This is a bit of meta but I think it needs saying, since the behavior is rather infamous here on a variety of controversial scientific issues. Genetically modified food is one of those, although we’re not talking about that here. On the other hand, what do you folks think corn/maize is? Sorry, couldn’t resist.

To put it simply: I like proof. Good old fashioned peer-reviewed Proof. It's sexy. This sort of proof is what we use for climate change and the ongoing disaster that has become. Pointing to a bunch of somewhat related studies (or more commonly, other blog posts commenting on university press releases) and then going “see?” isn’t proof to me. It’s a hypothesis, sure, and an entirely reasonable one given the research, but until it’s investigated and researched for the specific subject at hand, it isn’t proof. I also know, from all available evidence, that hydrofracking for natural gas extraction is a very bad thing from just about every environmental angle I can think of (even the earthquake one).

Are the recent earthquakes in Virginia, Oklahoma, and Colorado induced by oil and gas extraction activities? I very much lean toward no for all three based on available evidence , although a weak case for the August quake in Colorado I think can be made. In Virginia it’s a quake that occurred in a region that’s hosted infrequent large (and possibly, very large) earthquakes in the past several millennia, although I do think a (rather weak) case can be made that it was partially induced by events begun in 1971 relating to the creation of Lake Anna, which is nearby. In Oklahoma I have a bunch of ideas ranging from natural seismicity to induced seismicity to the recent uptick is a prelude to a much larger sequence of quakes, even if they don’t occur for several decades. This has likely happened in Oklahoma before, but when it did, no one who wrote down history was living there. For all three, I still submit that they’d have happened anyway at their magnitudes with or without human influence which hasn’t been absolutely proven, and I doubt was involved. All three occurred in areas that are seismically active relative to their surrounding areas.

But how would humans induce an earthquake?


Reservoir induced seismicity is well known, and has been both highly damaging and deadly worldwide.

In the 1940s, as Lake Mead filled, a series of quakes began, rattling nerves in the new and growing city of Las Vegas. Some were large, compared to what had been experienced in the local region to that point. Eventually, seismologists tied the rising lake to the quakes which incidentally, still occur today. It got deadly after that: an M6.5 earthquake in Koyna, India, in 1967 killed 180 people. (Seismicity still continues in that region to this date).

How does this happen? While the exact mechanism isn’t fully understood, it works like this: Water is heavy. Very heavy. It can weigh down the crust, deforming it, and increasing pressure on the ground below. Locally it can influence whatever regional stress field exists. Water can also force its way underground beneath a reservoir (or, a flooded quarry). Much like hydrofracking, the water creates fractures. If the water finds itself a fault that happens to both be 1. Stuck and 2. Oriented to the regional strain field, the fault can rupture. This is fairly well observed, all over the world.
In 1994 a fairly sizable earthquake for the region (M4.6, with an M4 foreshock) struck near Reading, PA. As it occurred two evenings prior to Northridge, it eventually got forgotten in the news. However research determined that a quarry in the area had ceased production, and flooded. When the quarry was abandoned and flooded, the strain in the local area changed suddenly enough to produce an earthquake on one of the many buried faults that riddle the foothills of the Appalachian Mountains. Well, series of earthquakes—the sequence began in 1993 and is apparently still ongoing. Minor quakes still shake the Reading suburbs.
I’ve read an argument that this is precisely what happened in Virginia. Lake Anna sits right in the middle of the Central Virginia Seismic Zone, and was filled in 1971. A retired geophysicist feels it was the causative agent behind August’s quake. I think a case can be made with further research, but I sincerely doubt it.

However this isn’t something you should expect with all dams, quarries, and reservoirs. A fault loaded to rupture already needs to be there.
This is the short version. A wealth of information can be found via Google Scholar, and a variety of websites on the internet.

Perhaps one of the best papers on the subject that I’ve found is “On the Nature of Reservoir-induced Seismicity” by Pradeep Talwani.


It’s been more or less conclusively proven that these activities can induce seismicity.
In the early 1960s, at the Rocky Mountain Arsenal, a program to dispose of waste fluid began. The fluid was injected down very deep wells drilled deep into the earth. What could go wrong, the Army thought.

Now, earthquakes had rattled the Front Range in Colorado in the past. A fairly significant one struck in 1882, its magnitude assessed at probably 6.6. A fault in Colorado has been recently identified (PDF). But these probably weren’t exactly known in the 1960s. Human beings have a habit of forgetting and losing their awareness of hazards over time.

At any rate, the ground started rumbling.
(PDF) It built to a point in 1966 and 1967, where three earthquakes over M5 caused some damage in and around Denver. After that, injection ceased. Then, the earthquakes more or less ceased. Further investigation proved that the injection activated a fault, which then ruptured. Would it have done so without our help? Perhaps, or perhaps not.

This is what happened near Guy, Arkansas. The region had been active before, a swarm rattled a nearby area in 2001, so when the quakes started again in 2010, that’s what people thought was going on. However, once injection ceased, so did (for the most part) the earthquakes. Arkansas has placed a moratorium on waste fluid disposal in wells. I assume the fluid is now going to sewage treatment plants, like it is in Pennsylvania, and that is a whole other kettle of fish.
It’s also occurred in Ohio and other places.

Basically, it works like this: You have your well. Your well is drilled perhaps 2, 3, even 4 kilometers into the earth’s crust. Unknown (or perhaps known, if one assumes a level of dishonesty on the part of the geophysicists who work for energy companies) to you, there’s a buried fault there, right under or adjacent to your disposal well. The fault, of course, has no noticeable surface expression, especially if you’re east of the Rocky Mountains. You pump the waste fluid from your gas/oil extraction operation into this well. It has to be disposed of somehow. The fluid works its way into the fault, lubricating it. In most cases, nothing happens, but this fault happens to be under strain, close to rupture geologically speaking, and aligned to the regional compressive stress field. And, it does. And it doesn’t stop until you stop injecting fluid down the well. When you do, most stop right away, although the seismicity may linger for a time after that.

You can also modify this for similar processes used for geothermal energy. The Geysers in northern California is infamous and After a series of quakes in Basel, Switzerland, a geothermal plant was shut down after it was linked to them. It’s not just the oil and gas industry that can induce seismicity. The very green geothermal industry can too. But it’s more or less the same here too, as it is in seismicity generated by other human activities. You have to have a fault right under your well, or very close to it, and it has to be a fault that’s already stressed just about to the point of failure. Conditions have to be just right (well, wrong, if it’s a populated area) for a quake of size to happen. This is the short version, of course: this site, one of many, has a lot more primary source information.

I want to go back to something I mentioned at the beginning of the diary, as to an idea I’ve kicked around since the earthquakes in Oklahoma began this weekend. Dr. Seth Stein of Northwestern University, in Chicago, and his team have developed a series of intriguing theories to help explain why earthquakes occur in the middle of continental plates/. The USGS has created a seismic hazard map, which it is in the process of updating, that I feel isn’t entirely accurate. Scattered across the eastern and Midwestern part of the nation are various little zones where seismicity is elevated compared to surrounding regions. Central Virginia is one of those areas. Charleston, South Carolina is another. A region centered at New Madrid, MO is the most famous. Another is located in central Connecticut at Moodus. We don’t entirely understand why these regions are the way they are---some may be related to exceptionally ancient rifting, others are related to post-glacial rebound and others on ancient continental collisions. Various pieces of continents attached themselves to what is now the eastern third of the US, and perhaps they’re not all knit neatly together. All of this is being pushed, generally, westward. Dr. Stein hypothesizes that these little patches of elevated seismicity are actually aftershocks of much larger quakes, and that large quakes can migrate all over the place—zones can turn on and off over time. We sort of see this happening in China.

For Oklahoma (and perhaps elsewhere) I’ve come up with a few hypotheses:
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 (PDF), 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 mitigated somewhat, because the strain was partially released due to hypothesis 2.

I guess we won’t know. I lean toward hypothesis one based on all available evidence (and there's none at this point that points the blame at nothing but the Earth itself), but hypothesis 3 is interesting too. It's not like Oklahoma is aseismic.

This is by far an incredibly cursory review of what I know, and what science knows. Now I didn’t spend this time to disprove my point, which is I think the quakes mentioned at the beginning of the diary were natural. I wrote all of this to show how it could be proven otherwise, and done so scientifically (which rather is the only proof I’ll really accept). I’m in no way saying gas drillers should keep fracking. I absolutely think they should stop.Their danger to water supplies is documented. Their damage to roads is well known to transportation departments all over the place. I also think that if they're inducing quakes to occur, this has to be investigated and proved.

In short, prove it first before you blame the drillers. This isn't out of concern for the drillers. It's out of concern for the scientific process.