Horses, barns and earthquakes
Well, it appears that the state of Oklahoma finally bought into the connection of earthquakes to deep injection wells as the recent M5.6 earthquake led them to shut down injection wells in the vicinity of the epicenter [and once again we learn the national media still cannot discern between fracking, which is not the cause here, and injection of waste water, which is the likely culprit]. Interestingly, there are two views on how Oklahoma seismicity is varying: Dan McNamara of the USGS argues that seismicity is still on the rise, while Oklahoma Geological Survey director Jeremy Boak is quoted by the Tulsa World that “I still expect to see declining figures over the rest of the year just because we’ve decreased the (wastewater) injection so much.”
Given how long the Oklahoma survey dragged its feet on acknowledging the problem, their credibility is kind of at a low point. McNamara in November said that more M5s were likely, and two more have happened since. McNamara made a plot of seismic moment over time that is pretty damning:
The big decrease in seismicity Boak was excited about is the somewhat shallower slope of moment increase in early 2016, a decrease now obliterated by this latest quake.
The problem is that fluid injection of this magnitude over this amount of time has probably not reached any kind of equilibrium yet. The overall upward concavity of this plot suggests that we aren’t at the end of increasing rates of moment release. Hopefully it will come as a boatload of small-impact M4s and low M5 events, but M6 events don’t seem implausible. If you look at a much smaller example, the likelihood of earthquakes continuing for decades is substantial–even if injection stops.
Back in the 1960s the Army injected wastewater at the Rocky Mountain Arsenal into basement. This caused a bunch of earthquakes and eventually the injection was stopped–but two years later some of the largest quakes in the sequence happened. At the time the interpretation was that the pressure wave from the injection was propagating outward and so could have a substantial time lag. Regardless of mechanism, it should concern Oklahoma residents that in a similar case with much, much smaller volumes of water being injected that earthquakes continued long after injection ceased.
It is great that the Corporation Commission in Oklahoma has acted to shut down a number of injection wells. Too bad some of this didn’t come before the billions of barrels of produced water were injected into the Arbuckle Formation. We will see if this closing of a barn door caught the horse or not. The problem may be that the pasture gates need closing too: the production of oil is not likely to shut down at the same rate as injection well capacity; produced water will probably be rerouted to wells that have not yet been shut down. And while many wells probably pose no risk of inducing earthquakes, some probably do. So this might simply migrate the problem even farther afield.
One reality is that the duration of time needed to really see if this helps–probably on the order of years at this point–is almost certainly beyond the ability of government overseers to keep operators from applying political pressure to resume operations at some level. The only really good solution is some kind of processing of these waters so they can be released at the surface, but such purification is expensive and would require creation of infrastructure that doesn’t yet exist.
Well, of course, there is another solution: quit pumping oil. Don’t hold your breath wait for that one. And if you live in Oklahoma, you might just want to see how much that earthquake insurance is. And find those webpages Californians have perused for years on how to make your house more quake-proof.