Recently the UK government has announced it is to push ahead with the exploration of shale gas seams (fracking) throughout the country. This news has been met with strong protests, notably from environmental lobby groups. The debate however has become clouded (and not for the first time) by both the lack of, and lack of understanding of scientific evidence on the potential environmental effects of fracking.
Two of the more widely mentioned effects in the media are potentially groundwater contamination and an increase in earthquake risk. Frustratingly neither of these issues have yet been treated to an accepted and comprehensive scientific review and thus the debate is largely an exercise in trading anecdotal evidence. An angle which receives surprisingly little attention in the news is the potential effects of fracking on surface water resources and thus the implications for river management.
As the name suggests “hydraulic” fracturing is dependant on using water to create fractures in rock seams in order to extract the shale gas. (For a detailed overview of the process you can read here and here). The “hydraulic” part has two potential implications for rivers, the first is the sourcing of fresh water which is needed to make the injection fluid and the second is the waste water (or “flowback”) has to be disposed of.
The precise volume of water required per fracking well is very hard to predict, and in various technical and environmental reports from the US it is hard to even pin down a consistent range. The figure to the right for water usage intensity in US wells with depth illustrates some of the problems even where data is available, with a large amount of scatter around the optimistically fitted regression lines. The Marcellus Shale Region of the US appears to require an average of 4.5 million litres (M.l.) of fresh water per well (2012 report from The Pacific Institute) with a reported range estimated to be 3-5 M.l. per well (2010 report ALL Consulting). However a report from The University of Texas at Austin reports a maximum usage of 13 M.l. per well in that state. Conversely Kargbo et al (2010) in Environmental Science and Technology report 7.7-38 M.l. of water per well taking into account the entire process of drilling & extraction. Indeed in the UK government adoption report (pg. 21) an estimate of 10 M.l. was used, however in consultation this was highlighted as potentially being too low so it was changed to 10-25 M.l. I can find no references for the basis of either of these figures.
The key issue with water abstraction for fracking is not the total amount per se, but the time over which the water is abstracted and where it is abstracted from. Due to the nature of fracking, the abstraction will occur chiefly over a matter of days and this will usually occur from a single source. Due to the expense of setting up pipelines and pumping water from distant locations or transporting water via tankers water is usually extracted from sources very local to the well site. Often the water will be stored in tanks or open air reservoirs to act as a ready supply. Where a number of wells are to be drilled in a relatively small area it can often prove more economical and practical to modify watercourses or install pumps to create and maintain a larger reservoir which can act as a reliable local source for a number of wells.
The obvious implications of this activity is the impact of short duration, high volume, discontinuous abstraction events upon river flows in terms of low flows; in effect artificial “micro droughts” within the river. Typically where stream biota are able to adapt to low flow/drought conditions they will be experiencing a gradual onset of lower flows, allowing adaptation, the impacts of a sudden reduction in stream flows on biota are unknown. Particularly in the case of ecologically sensitive streams during summer months this would need careful monitoring to say the least. In such cases, where land is available a more practical option maybe to utilise existing reservoirs or create new ones in order to better manage low flows during summer months. The loss of land/habitat for reservoir construction maybe be outweighed by the ability to manage low flows in the river by effectively spreading abstraction out over a long time period.
In the South of the UK, where water supply is already stretched increased demand for water abstraction may have interesting connotations. In Colorado licences for abstraction are auctioned and in recent years farmers needing abstraction licences for crop irrigation have found themselves outbid by fracking companies leaving some agriculture at risk. As pressure on water resources increases these conflicts and competing demands are likely to become more commonplace.
The reverse of the abstraction challenges is what to do with the flowback once the water has been used. Again available data on this is sparse. It is not possible to know the fraction of the water pumped in which is recovered for two reasons, firstly the water returns to the surface over a period of days to weeks, and secondly groundwater stored within the shale reservoir (the amount of which is unknown) will also flow to the surface after fracking (known as “produced water”) and it is not possible to distinguish this from the fracking fluid. Thus simply measuring the amount of water coming out tells you nothing about how much of that water was what you put in.
Whichever way you cut it this flowback is dirty. It contains a mix of the residual introduced chemicals within the fracking fluid, heavy metals, salts & radioactive materials. The precise mix of containments will depend on the local geology, but has been found in some cases to have total dissolved solids (TDS) exceeding 200,000 mg/l (3 times that of sea water).
The problem is what to do with this waste water. In the US the majority of flowback is pumped into disposal wells deep underground (which is the primary source of concern regarding groundwater contamination). Of more concern to surface water managers is the practice of either treatment through evaporation or through conventional waste water treatment plants. Typically flowback will be stored in open pits for some time after it has been recovered, occasionally it can be left in open pits to evaporate and the resulting precipitate recovered and disposed of as solid waste. Given the climate this seem unlikely in the UK. The other option is to deliver the flowback to normal water water treatment plants. Although waste water treatment is an eminently sensible route, most waste water processing facilities will not be able to cope with the nature of containments possibly present in flowback. Of particular note is the high TDS content which conventional plants are not equipped to deal with and has resulted in highly elevated level of TDS in streams which received “treated” flowback (and other treated waste water) in Pennsylvania.
There are emerging technologies which could potentially be used to treat flowback on site, such as sequential precipitation and reverse osmosis however these have not been demonstrated in field applications and for very high TDS concentrations (>100,000 mg/l) evaporation maybe the only current viable method of removal/treatment. Although re-use of flowback for further fracking is widely discussed and advocated it is almost always impractical due to the aforementioned TDS levels which lead to precipitates (barium,calcium,iron,magnesium,manganese,and strontium) forming in equipment, this is sometimes even an issue in pumping flowback into disposal wells.
Given the potential challenges to surface water resources presented by local fracking it is necessary for river managers and stakeholders to keep abreast of planned fracking exploration. Of particularly concern is the poor predictive ability for water usage, the figure on the right is from a 2011 report in which projected and actual water use at a district (county) level is compared on a log-log scale. Anything below the line is a well where water use for a region was underestimated, there are quite a few points on the chart where water use has been underestimated by an order of magnitude. For example the point ringed in red the county was projected to use ~150,000 m³ of water but actually used ~1 million m³. The UK government adoption report reads as frankly blasé about some of the challenges discussed above, variously putting the emphasis on the fracking companies themselves to be responsible, the water companies to manage abstraction and brushes off one concern with “The industry is also not expected to be at substantial scale before the 2020s. This will allow time for any necessary new investment” (pg.26). The emphasis is therefore very much on stakeholders to monitor submissions for fracking wells and to ensure, particularly during the planning consultation phase that environmental and practical concerns related to river management are adequately addressed and that predictive evidence is strongly supported.
Heather Cooley & Kristina Donnelly (2012) Hydraulic fracturing and water resources: separating the frack from the fiction, Pacific Institute, California. ISBN:1-893790-40-1. http://pacinst.org/wp-content/uploads/2013/02/full_report5.pdf
Kargbo, D.M., Wilhelm, R.G., Campbell, D.J., 2010. Natural gas plays in the Marcellus shale: Challenges and potential opportunities. Environmental Science & Technology, 44(15), 5679-5684.