After a small discussion in my previous post about hydraulic fracturing, I decided to look in to the subject a little more deeply, since the commenter focused strongly on concerns related to horizontally drilled wells. Horizontal wells weren’t really mentioned in the articles I linked to, but horizontal drilling is becoming the common method for extracting gas from tight shales and will likely be used in the development of the Marcellus Shale. As such, it’s more than fair to look at the practice and see if there are issues unique to it, or problems that occur more commonly in horizontal rather than vertical wells.
One thing I’ve noticed so far in researching is that papers that could answer that concern are few and far between. Looking at both GeoRef and GeoScience World, many reference hits are to expanded abstracts from meetings, and most of those are related to the effects of horizontal drilling on reservoir development rather than environmental impact. The EPA is also unhelpful on the topic; as the practice of hydraulic fracturing has been excluded from the Safe Drinking Water Act, the EPA isn’t in a position to look in to the safety from an environmental standpoint. The one study they have done relates to coal bed methane. In their coal bed methane study, the EPA concludes:
Although potentially hazardous chemicals may be introduced into USDWs when fracturing fluids are injected into coal seams that lie within USDWs, the risk posed to USDWs by introduction of these chemicals is reduced significantly by groundwater production and injected fluid recovery, combined with the mitigating effects of dilution and dispersion, adsorption, and potentially biodegradation. Additionally, EPA has reached an agreement with the major service companies to voluntarily eliminate diesel fuel from hydraulic fracturing fluids that are injected directly into USDWs for coalbed methane production.
However, that said, the coal bed methane situation is arguably not quite the same, since most coal beds used in that study were fairly shallow, and many in direct “communication” with adjacent formations or aquifers. As a side geological note, coal layers can often act as barriers to fluid (such as hydrocarbons) flow from lower formations to upper formations, since they’re only really permeable via fractures. Also, the conclusions have been called in to question in 2004 by an EPA employee named Weston Wilson:
While EPA’s report concludes this practice poses little or no threat to underground sources of drinking water, based on the available science and literature, EPA’s conclusions are unsupportable. EPA has conducted limited research reaching the unsupported conclusion that this industry practice needs no further study at this time. EPA decisions were supported by a Peer Review Panel; however five of the seven members of this panel appear to have conflicts-of-interest and may benefit from EPA’s decision not to conduct further investigation or impose regulatory conditions.
An apparent conflict of interest is certainly something to look in to. I found one site that said Mr. Wilson’s concerns were considered valid enough to prompt further investigation, though I have no idea if that investigation occurred or what the conclusion turned out to be. So I think for now, we’ll skip the coal bed methane study.
The commenter made two points that I looked in to, which I’ve paraphrased here:
1) That the fractures made by hydraulic fracturing can extend further than intended, and in to different zones than are desired, possibly putting the aquifer at risk.
I tend to think that subsurface, technically anything is possible; geophysics is a difficult field (certainly not one I’ve mastered) since the conditions are uncontrolled and it’s nigh impossible to know what the underground stress field is like precisely or what zones of weakness/minor faults may also exist. So is it possibly that a fracture created in a deep zone could “go rogue” (oh, I feel dirty just typing those words!) and go 4,000 feet up in to an aquifer? Possible, yes, but I would venture to say highly unlikely and would probably require an existing network of fractures/faults to do. The EPA coal bed methane study does actually say one thing of use about this matter, which holds true even if we’re not talking coal bed methane:
A hydraulic fracture will propagate perpendicularly to the minimum principal stress. In some shallow formations, the least principal stress is the overburden stress; thus, the hydraulic fracture will be horizontal. In deeper reservoirs, the least principal stress will likely be horizontal; thus, the hydraulic fracture will be vertical.
Basically, the generic stress you’d expect in a deep formation would cause the fractures to tend to propagate upward. However, the general stress field will change as you approach the surface, and there are going to be other stress factors that may redirect fractures along the way. Intervening formations with different properties and different zones of weakness also have an important effect. A basin is in no way homogeneous vertically – and often formations will change their properties over horizontal distance as well. The New York Department of Environmental Conservation SGEIS report is a little more specific:
ICF – citing PTTC, 2006 – concludes that: “In the Appalachian Basin, the stress state would be expected to lead to predominantly vertical fractures below about 2500 feet, with a tendency towards horizontal fractures at shallower depths.”
Depending on the depth of the aquifer, this conclusion makes the propagation of a vertical fracture in to it seem fairly unlikely.
At this point, the best that can be done is computer modeling, which has become increasingly sophisticated – though it can always be argued of course that we lack perfect knowledge of sub surface conditions. However, the fact of the matter is that it’s in the best interest of the people fracturing the formation to prevent such wild fracture propagation from happening. Going outside the intended zone tends to mean getting a lot of unintended fluid – normally salt water – which has to be separated and disposed of, not something that is cheap or convenient to do.
The New York Department of Environmental Conservation has a report that’s an overview of the historical and current practice of horizontal drilling and hydraulic fracturing here. I think the authors of the report make a valid point by stating:
Not only is fracture growth outside of the target formation discouraged relative to the potential of reduced production by production of fluids from non-productive zones, creating fracture size outside of the productive interval is more expensive and less cost beneficial to the well’s economics.
• The developable shale formations are vertically separated from potential freshwater aquifers by at least 1,000 feet of sandstones and shales of
moderate to low permeability.
• The amount of time that fluids are pumped under pressure into the target formation is orders of magnitude less than the time that would be required for fluids to travel through 1,000 feet of low-permeability rock.
• The volume of fluid used to fracture a well could only fill a small percentage of the void space between the shale and the aquifer.
• Any flow of fracturing fluid toward an aquifer through open fractures or an unplugged wellbore would be reversed during flowback, with any residual fluid further flushed by flow from the aquifer to the production zone as pressures decline in the reservoir during production.
These factors point to groundwater contamination without wild fractures going a thousand or more feet farther than intended being highly unlikely as well.
At this point, the question of the threat that the fracturing practice has on the ground water becomes a cost/benefit analysis. It’s in the best economic interest of the company developing the well to model their fractures properly to ensure that they do not move out of the intended zone. But it’s also fair to argue that modeling is imperfect and that the most bizarre accidents can happen. Is that amount of risk worth the economic benefit?
2) Horizontal wells carry a greater risk of inducing harmful seismic activity which can cause an array of problems.
I’m actually already familiar with the concept of just plain old vertical wells inducing seismic activity; I doubt horizontal wells would be at all different. It’s certainly not a new concept. Most of the reading I’ve done on this subject relates to disposal wells; in particular there was a case in Colorado where a deep disposal well was lubricating nearby faults and causing them to reactivate. The seismic activity in these cases is generally pretty low level (0 to 3 on the Richter Scale) and mostly an argument for being careful where one puts one’s disposal wells – and for not overpressurizing them.
Subsidence is another matter that goes with the extraction of any fluid from a reservoir. With the reservoir losing pressure and volume with the removal of fluid, it can compact, which destroys reservoir permeability and causes subsidence. Subsidence also occurs and is a significant threat when aquifers are overused for drinking water. Basically, any time fluid is taken out of the ground and not replaced, there will be subsidence. From the US Department of Energy:
The most famous early instance was in Wilmington, California, where the oil production triggered a series of damaging earthquakes. In this instance the cause of the seismicity was traced to subsidence due to rapid extraction of oil without replacement of fluids. Once this was realized the oil extraction was balanced with water injection to mitigate the seismicity. Ever since then the oil and gas industry has adopted these practices to mitigate seismicity , but also mitigate damage to the oil wells in the producing field (wells would be sheared off in the subsurface as subsidence occurred).
The practice of reinjecting fluid isn’t exactly perfect – if nothing else there are cases where the replacement fluid can’t be returned to the right location and overpressurization can occur.
With the reading, I’ve done, I tend to agree with the DoE assessment:
Overall the impact of induced seismicity on the implementation of various different energy recovery and or disposal activities will depend on the risk associated with the activity and the cost-benefit ratio. All experience to date has shown that the risk, while not zero, has been either minimal or can be handled in a cost effective manner.
So once again, this comes to a question of cost/benefit analysis. Are the possible seismic dangers present worth the economic benefit? Are the means by which the concern about subsidence good enough at mitigating the problem? Since we live in a less than perfect world, it becomes a personal choice if the mitigation of the risk makes the development of the resource worth it. Though I think it is worth noting that in the DEC review of the pertinent geology that comes from their SGEIS, they state that in at least New York, the Marcellus and Utica shales underlie the areas of lowest seismic risk.
The DoE also makes a point of separating the induced seismicity associated with fluid injection and extraction from hydraulic fracturing:
Hydrofracturing is distinct from many types of induced seismicity because hydrofracturing is by definition only created when the forces applied create a type of fracture called a “tensile” fracture, creating a “driven” fracture… To our knowledge hydrofracturing to intentionally create permeability rarely creates unwanted induced seismicity large enough to be detected on the surface with very sensitive sensors, let alone be a hazard or annoyance.
I think the crux of this issue becomes the phrase “to our knowledge.” With what data and studies are available, it seems this is not something to be overly concerned about. However, I think a good argument can also be made in regards to needing more knowledge and more research. At this time, since the practice of hydraulic fracturing is in many ways proprietary and also exempt from government oversight, that makes the study of it in regards to environmental impacts difficult. I think that an NPR story on this topic sums it up nicely:
Critics of hydraulic fracturing suspect that the chemicals used in the process have somehow leaked into the groundwater supply. It has been difficult, however, to demonstrate a direct connection between these apparent instances of water pollution and the hydraulic fracturing procedures that have taken place nearby…It is also true, however, that state regulators have not been able to disprove a connection between hydraulic fracturing and water contamination.
At this time, it’s not something that could be directly proven or disproven. Off the top of my head, I think there are experiments that could be conducted that could tell us if:
a) Is there groundwater pollution associated with fracturing fluid?
b) Is this contamination associated with surface pollution or with the underground process?
For example, if you could add an inert “tracer” compound to fracturing fluid, you could then check if contaminated water had that tracer in it. As far as discerning between surface contamination and subsurface contamination, that might be more difficult. Though I imagine geologists who know far more about hydrology than I do may be able to determine the source just by looking at how quickly the contamination showed up. If you have contamination showing up in a reservoir two days after a nearby well has been fractured (and our mythical tracer compound shows it was from that well) then the next question is if a subsurface fracture could have possibly delivered fluid that quickly; if not, the more likely culprit is surface run off from an accident or inadequately observed safety practices.
Either way, this is all very pie in the sky. I’ll be interested to see more on this subject as more data is collected.
As a final note, the Department of Energy Induced seismicity primer is honestly a fascinating read in and of itself and I’d highly recommend it. I also recommend the New York Department of Environmental Conservation report that I linked to above as well, since it’s also a very good historical overview. The DEC also has a review of the pertinent geology that comes from their SGEIS that’s definitely worth a read. There is also a good description of the drilling process and equipment used from the SGEIS, which also includes a look at slickwater fracturing as used in the development of the Marcellus Shale.
Edit: An anonymous comment was left rather incongruously my post about Heene being sentenced. The commenter left a link for another bit of good reading on this topic, however: Impact Assessment of Natural Gas Production in the New York City Water Supply Watershed. The report covers quite a bit about the underlying fractures/faulting that was not in the other reports I read.