Shale Gas and Groundwater Contamination: Thoughts on a Recent Study
Alan Krupnick, Lucija Muehlenbachs, and Karlis Muehlenbachs
June 29, 2011
A recent article by Duke University researchers on methane contamination in drinking water wells in northeastern Pennsylvania and upstate New York has brought further attention to the risks to groundwater in areas exploited for shale gas. The study shows that methane concentrations are higher in most water wells located 1 km or closer to a shale gas drilling site compared to those in wells located beyond 1 km, and that the signature of the gas in these close-in areas is consistent with a deep (thermogenic) source. Those opposed to shale gas drilling see this article as confirming their fears, while the industry and others point out problems with the study.
The research has several important flaws and related ambiguities. First, without baseline data, there is no definitive way of eliminating the possibility that the closer wells had high concentrations before the drilling began. Because there is such a strong relationship between distance to the shale well and water well concentrations, however, this possibility could only happen if one of two conditions is present: 1) companies drill in “sweet” spots which also happen to be places of intense methane migration, or 2) the geology and hydrology varies enough that the study’s results are happenstance. We find the second option to be implausible, except that there are a significant number of close-in wells with low methane levels.
More information is needed to know where these wells are and what could explain their low readings. Data on more wells would also help, as only 68 were included in the study. The first condition is plausible, however, but if true, then much more information is needed on baseline concentrations of methane in groundwater and whether fracking makes migration in such areas worse. The critics’ argument that the study focused on an area which is known to have had trouble (namely Dimock, Pennsylvania, for which Pennsylvania’s Department of Environmental Protection in 2009 concluded faulty well casings and drilling operations contaminated the groundwater) doesn’t hold up—if that were the case, methane concentrations would be high in wells located farther from the shale gas site as well.
Making progress, in our understanding, is easy. Well concentration in active areas in Pennsylvania that border New York could be compared to concentrations in areas across the border in New York that are targeted for drilling but have not seen activity because of the ban. This way, the location of drilling would be determined by regulation (the ban), and not by geological or hydrogeological factors, which could have also influenced preexisting methane levels in water wells. In general, the study highlights the necessity of starting baseline water well testing prior to drilling. Such baseline tests are currently required in Alberta, Canada before drilling for shallow, coal-bed methane.
The second big issue with the study concerns the source of the gas. The water wells could be contaminated with shallow (biogenic) gas; induced deep (thermogenic) gas from various pockets of gas that the wellbore traverses before reaching the hydraulic fracking zone; deep (thermogenic) gas from the hydraulic fracking zone (that is, the production gas); or gas released from anywhere by microseismic activities related to fracking or old bore holes. Although the study is able to rule out biogenic gas (from shallow sources) that could have been previously present in wells, it does not rule out methane from deep sources unrelated to the production wells. Nor does it identify—assuming the gas was not present in the well water before drilling began—whether the gas got there from poor cementing around the casing, upward migration from the fracked seam, old wells, or from microfractures created farther above the seam that allow methane in other areas to migrate to the wells.
Moreover, for decades, research by universities and industry has documented that natural gas from different sources has a different fingerprint (from their carbon and hydrogen isotope ratios). Often the isotope ratios are diagnostic of individual gas sources and can be used for forensic fingerprinting. For example, engineers now use this fingerprint to pinpoint at what depth a well is leaking, in order to know what portion of the casing to repair. The most prolific deep shale gases in Texas, New York, and Pennsylvania are known to have a very unique isotopic fingerprint. The Duke study could have been even more specific as to the source of the methane contamination with a more rigorous analysis of the isotope ratios.
In general, a careful study of “fugitive” gases may identify whether they come from the ultimate depth of the hydraulic fracturing or whether they come from intermediate depths. Apart from requiring a baseline sample of water wells prior to drilling and during drilling, a baseline database of production gases is also needed. Industry can collect gas samples while drilling a new well, in order to have a mapping of the types of gases present at various depths from surface to the pay zone. This information would be constructive in establishing fracking risks and where they occur in the development process and could potentially lead to improved industry practices and identification of liability. To address issues of existing wells, more information is needed about the location, condition, and depth of abandoned wells and on the implications of seismic activity on risks of methane migration.
Alan Krupnick, RFF Research Director, Senior Fellow, and Director of RFF’s Center for Energy Economics and Policy; L. Muehlenbachs, RFF Fellow; K. Muehlenbachs, Professor, Department of Earth and Atmospheric Sciences, University of Alberta.