A paper (“Perspective”) published this week by Stanford University professors Mark Zoback and Steven Gorelick in the Proceedings of the National Academy of Sciences questions the viability of Carbon Capture & Sequestration (CCS) as a climate mitigation technology. A comprehensive report on the potential for seismicity from energy technologies more broadly was also published this week by the National Research Council (NRC). Zoback and Gorelick raise some valid issues that should be looked at, but reach sweeping conclusions without evidence or scientific basis. The NRC report presents a far more balanced analysis of the situation. For the public, some of the key questions that need to be answered are:
- whether CCS (or other technologies that inject fluids underground) can cause earthquakes;
- how large and damaging can these be;
- whether the risk can be managed;
- whether the technology can be deployed at a meaningful scale; and
- whether these earthquakes could have undesirable consequences such as leaks of the injected fluids.
Managing earthquakes caused by human activity is an issue that deserves more attention than it has received to date. It can and should be done with today’s tools, but it hasn’t been done everywhere. The NRC report is timely in that respect, and documents known earthquakes caused by human activities. None of these have been caused by CCS projects. The largest seismic event has been caused by an oil/gas extraction operation, while the more frequent sources are geothermal and waste water injection projects. No felt earthquakes are known to have been caused by enhanced oil recovery operations that inject CO2. In most cases, common sense by operators and regulators could have prevented these events. I agree with the NRC study on this point: further study and modeling are in order. Even though smaller earthquakes may not cause any damage, causing them is a profoundly bad idea. It betrays a lack of scrutiny over project operations, especially since they are avoidable.
Zoback and Gorelick however appear to have been causing undue alarm in the media. They state (p. 2) that their “principal concern is not that injection associated with CCS projects is likely to trigger large earthquakes; the problem is that even small to moderate earthquakes threaten the seal integrity of a CO2 repository”. They acknowledge that only slip on large faults can result in earthquakes large enough to cause damage to human environments, and that such faults are easily identified and avoided. No objections on that last point. The potential for slip on existing faults/fractures and seismicity can and should be taken into account during site selection. This is routinely done as part of a proper geomechanical assessment, and Federal Underground Injection Control Program regulations for geologic sequestration operations require “[i]nformation on the seismic history including the presence and depth of seismic sources and a determination that the seismicity would not interfere with containment”.[1] Large seismic events can be avoided in a straightforward way through proper siting and operations.
Zoback’s and Gorelick’s arguments against CCS hinge on the assertion that “[b]ecause laboratory studies show that just a few millimeters of shear displacement are capable of enhancing fracture and joint permeability, several centimeters of slip would be capable of creating a permeable hydraulic pathway that could compromise the seal integrity of the CO2 reservoir and potentially reach the near surface.” In plain English, the authors are saying that even a small earthquake can cause CO2 to escape all the way to the surface, without investigating the circumstances under which this might happen or their applicability to broad scale CCS. This creates the impression that it will happen in every case, and is a big logical leap and a gross simplification, for several reasons.
First, the laboratory studies they cite were performed on granite, which is extremely unlikely to be used as a sealing layer, or “caprock” in a real-life sequestration project. Almost certainly, the caprock will be shale or another low permeability sedimentary rock. The way that a strong but brittle rock like granite deforms in response to stress is very different from the way that softer and more ductile shales and other sedimentary rocks deform, and is therefore not a good analogue.[2]
Second, concluding de facto that joint and fracture permeability in the caprock(s) would increase in all cases, and that a pathway would be created that would result in the migration of CO2 to the surface, is wrong. The degree to which joint and fracture permeability is increased, if at all, depends on many factors, including rock type, stress state, and in-filling materials. This is well documented in a large body of literature on shear-induced behavior of fractures and faults (if you want a flavor, take a look here[3] for example). In fact, situations abound where many large faults that exhibit large slip act as seals and have no effect on permeability. Such is the case in California and Iran, where trapped oil and gas exists despite frequent large natural earthquakes. In these areas, in fact, faults themselves have acted as seals as opposed to pathways for fluid migration, and trapped hydrocarbons over geologic time. Another well-documented event is the magnitude 6.8 earthquake in Chuetsu, which did not result in any leaks in the nearby Nagaoka CO2 injection project. Despite frequent and large natural earthquakes therefore, CO2 and other fluids have remained trapped in the subsurface.
Additionally, assuming that CO2 will reach the surface implies that the fault in question extends from the injection zone to the surface. As the authors themselves note, such a large fault would be easy to identify and avoid. Even if a fault allows CO2 to migrate out of the injection zone, many sites also have multiple sealing layers that impede the motion of fluids to the surface as well as multiple permeable layers that can act as secondary containers. In fact, studies show that such layered systems can help prevent fluids from reaching the surface.[4] Assuming that a pathway will be created all the way to the surface is a huge leap of logic. Fluids can and do move along faults and fractures – but this does not mean that the containment “box” has been breached – fluids can simply move within the “box”, leaving the caprocks intact.
In other words, jumping to the conclusion that a small induced earthquake would result in surface leakage is wrong. That’s not to say that it cannot happen, but the problem with the authors’ assertion is that they then postulate that not enough sites for sequestration can be found that avoid this scenario to meaningfully deploy CCS at scale. Although they acknowledge that certain geological settings are ideally suited to secure sequestration of CO2, such as in the case of the Sleipner project in Norway (which features a highly porous and permeable reservoir consisting of weak, poorly cemented sandstone that is laterally extensive), they then extrapolate that not enough sites like Sleipner can be found around the U.S. to house the necessary volumes of CO2 to mitigate climate change. This extrapolation is based on speculation and comes with no scientific justification. The authors do not study the potential for sites like Sleipner – i.e. with sufficient porosity and permeability to accommodate injected CO2 without giving rise to unacceptable stresses – to be found around the country. This can only be done with a rigorous geologic assessment, and there is no evidence to suggest that such sites cannot be found in sufficient numbers.
Not all sequestration sites need to be slam-dunk cases with porosity and permeability like Sleipner’s in order to safely accommodate CO2. Of course – wouldn’t it be nice if things were ideal everywhere, but a wide range of geological settings can also accommodate CO2 safely without causing unacceptable seismicity risk. The regulation of maximum allowable pressure, evaluation of seismic risk, and of the conditions in which transmissive faults would threaten groundwater is central to Federal regulations under the Underground Injection Control Program. Industry and regulators should take note, however: even though smaller earthquakes caused by injection may cause no physical damage or human harm, the public may reject the idea of CO2 injection if these quakes and perceptible.
Zoback and Gorelick’s assertions were met with skepticism by expert scientists. Sally Benson (Stanford professor of Energy Resources Engineering and Director of Stanford's Global Climate and Energy Project, and Lead Coordinating Author of the Underground Geological Storage Chapter in the IPCC Special Report on CCS) said “of course, you need to pick sites carefully, but finding these kinds of locations does not seem infeasible”. I think Rob Finley hit the nail on the head when he compared Zoback and Gorelick's analysis to early criticisms of the Wright brothers and the notion at the time that airplanes would never work at scale. Rob is the principal investigator of the Midwest Geological Sequestration Consortium, which is now operating a large CO2 injection project in Decatur, Illinois, and has spent considerable time and money investigating the geology of the Illinois Basin. Julio Friedmann at Lawrence Livermore National Lab points out that “[b]y 2020, we're going to have somewhere between 15 and 20 projects around the world. That will be a good time to assess what we've learned and whether [CCS] can be scaled up more.” The last in the series of international conferences on the subject attracted 1,500 people. None of them appear to have voiced the seeming impossibilities for CCS that Zoback and Gorelick describe in their “Perspective”.
Should we therefore be alarmed by the prospect of CO2 injection in terms of earthquakes? My view is “no” – we should however be vigilant. Improperly conducted CCS does have the potential to cause earthquakes, due to the volumes of CO2 injected. But preventing and predicting these is within our capabilities. Avoiding the large ones is straightforward. It is worth noting that large natural earthquakes have not compromised the storage security in natural and man-made sites that trap CO2 and hydrocarbons. This does not mean, of course, that we should tolerate CCS projects that could cause earthquakes. Avoiding smaller quakes that may not cause harm but may alarm the public and local communities will require will careful site operation and regulation. And that can and must be done. Regulators and prospective injectors, do your homework.
[1] See Class VI regulations: 40 C.F.R. § 146.82(a)(1)(iii)(v)
[2] The technically minded among you may wish to read on… It is an established concept in rock mechanics that application of shear stress to a fracture will result in dilatancy (opening of the fracture). The amount of dilatancydepends on many factors, including the magnitude of the stress applied normal to the fracture, the strength of the rock, roughness of the surfaces of the fracture, and what kind of material is present in the fracture. If a fracture dilates, its permeability can increase. Granite is at one end of the spectrum of possible outcomes. It is strong, and fractures are often rough, so permeability increases can be large. At the other end of the spectrum are soft shales where dilatancy can be much smaller, or even negligible. Active faults, which see relative movement over geologic time, are filled with all sorts of materials representing a spectrum of hydraulic properties. But, often, they are filled with "gouge", which is essentially clay, which can sustain large shear movement without large dilatancy.
[3] JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 114, B05409, 18 PP., 2009, doi:10.1029/2008JB006089.
[4] Nordbotten, J. M., M. A. Celia, and S. Bachu (2004), Analytical solutions for leakage rates through abandoned wells, Water Resour. Res., 40, W04204, doi:10.1029/2003WR002997.