Water resource impacts during unconventional shale gas development: The Pennsylvania experience

Susan L. Brantley, Dave Yoxtheimer, Sina Arjmand, Paul Grieve, Radisav Vidic, Jon Pollak, Garth T. Llewellyn, Jorge Abad, Cesar Simon

Research output: Contribution to journalArticlepeer-review

234 Scopus citations


Improvements in horizontal drilling and hydrofracturing have revolutionized the energy landscape by allowing the development of so-called "unconventional" gas resources. The Marcellus play in the northeastern U.S.A. documents how fast this technology developed: the number of unconventional Marcellus wells in Pennsylvania (PA) increased from 8 in 2005 to ~. 7234 today. Publicly available databases in PA show only rare evidence of contamination of surface and groundwaters. This could document that incidents that impact PA waters have been relatively rare and that contaminants were quickly diluted. However, firm conclusions are hampered by i) the lack of information about location and timing of incidents; ii) the tendency to not release water quality data related to specific incidents due to liability or confidentiality agreements; iii) the sparseness of sample and sensor data for the analytes of interest; iv) the presence of pre-existing water impairments that make it difficult to determine potential impacts from shale-gas activity; and v) the fact that sensors can malfunction or drift.Although the monitoring data available to assess contamination events in PA are limited, the state manages an online database of violations. Overall, one fifth of gas wells drilled were given at least one non-administrative notice of violation (NOV) from the PA regulator. Through March 2013, 3.4% of gas wells were issued NOVs for well construction issues and 0.24% of gas wells received NOVs related to methane migration into groundwater. Between 2008 and 2012, 161 of the ~. 1000 complaints received by the state described contamination that implicated oil or gas activity: natural gas was reported for 56% and brine salt components for 14% of the properties. Six percent of the properties were impacted by sediments, turbidity, and/or drill cuttings. Most of the sites of groundwater contamination with methane and/or salt components were in previously glaciated northern PA where fracture flow sometimes allows long distance fluid transport. No cases of subsurface transport of fracking or flowback fluids into water supplies were documented. If Marcellus-related flowback/production waters did enter surface or groundwaters, the most likely contaminants to be detected would be Na, Ca, and Cl, but those elements are already common in natural waters. The most Marcellus-specific "fingerprint" elements are Sr, Ba, and Br. For example, variable Br concentrations measured in southwestern PA streams were attributed to permitted release of wastewaters from unconventional shale gas wells into PA streams through municipal or industrial wastewater treatment plants before 2011. Discharge has now been discontinued except for brines from a few plants still permitted to discharge conventional oil/gas brines after treatment. Overall, drinking water supply problems determined by the regulator to implicate oil/gas activities peaked in frequency in 2010 while spill rates increased through 2012. Although many minor violations and temporary problems have been reported, the picture that emerges from PA is that the fast shale-gas start may have led to relatively few environmental incidents of significant impact compared to wells drilled; however, the impacts remain difficult to assess due to the lack of transparent and accessible data.

Original languageEnglish (US)
Pages (from-to)140-156
Number of pages17
JournalInternational Journal of Coal Geology
StatePublished - Jun 1 2014

All Science Journal Classification (ASJC) codes

  • Fuel Technology
  • Geology
  • Economic Geology
  • Stratigraphy


Dive into the research topics of 'Water resource impacts during unconventional shale gas development: The Pennsylvania experience'. Together they form a unique fingerprint.

Cite this