Tracing the environmental and human health impacts of oil and gas development

Restricted (Penn State Only)
Tasker, Travis Lindsay
Graduate Program:
Environmental Engineering
Doctor of Philosophy
Document Type:
Date of Defense:
November 16, 2018
Committee Members:
  • William D Burgos, Dissertation Advisor
  • William D Burgos, Committee Chair
  • Nathaniel Richard Warner, Committee Member
  • Franklin Lewis Dorman Jr., Committee Member
  • Matthew S Fantle, Outside Member
  • hydraulic fracturing
  • Marcellus Shale
  • Utica Shale
  • Oil and gas
  • isotopes
Advances in drilling and hydraulic fracturing technology over the last several decades have made it feasible to extract natural gas from unconventional oil and gas (O&G) formations. These techniques have shifted U.S. energy production to natural gas and have allowed O&G companies to develop shale reservoirs throughout the entire country. Two of the largest shale reservoirs in the U.S. are the Marcellus Shale and Utica-Point Pleasant (UPP) Shale formations located in the northeastern, Appalachian Basin which produced over 20% of the total domestic natural gas production in 2017. These formations have low permeability and therefore require directional drilling in combination with hydraulic fracturing techniques to release the trapped hydrocarbons. Historically, hydrocarbons in these formations were believed to be inaccessible and therefore O&G development was limited to formations that could be targeted with conventional drilling methods (i.e., vertical gas wells). Before unconventional drilling started in the Marcellus and UPP Shale plays in the early 2000s, there were already approximately 100,000 conventional O&G wells drilled throughout the Appalachian Basin. However, over the last decade, the number of active conventional wells has declined while the number of active unconventional wells continues to increase. As of 2018, there were over 10,000 unconventional wells in the Marcellus Shale and over 2,000 in the UPP Shale. The increase in unconventional shale gas extraction throughout the Appalachian Basin has raised concerns regarding the environmental and human health impacts of this practice and renewed interest in the potential impacts from conventional O&G development. One of the biggest challenges with O&G development is wastewater management. In Pennsylvania alone, over 0.5 billion liters and 8 billion liters of wastewater were generated from conventional and unconventional wells in 2017, respectively. This wastewater can have high salt and metal concentrations in addition to radioactive elements that require proper disposal. One disposal option that is allowed in several states is spreading O&G wastewater on roads for de-icing or dust suppression. However, very little is known about the extent of O&G wastewater spreading on roads throughout the U.S. and its potential environmental and human health implications. As discussed in the second chapter of this dissertation, this practice is permitted in at least 13 states throughout the U.S. but is mostly limited to wastewaters from conventional O&G formations. Collection and characterization of wastewaters spread on roads in the northeastern U.S. revealed that these fluids also have high salt, radioactivity, and organic contaminant concentrations that are often 2 to 3 orders of magnitude above drinking water standards. These contaminants were toxic to aquatic organisms like Daphnia Magna and also activated xenobiotic signaling pathways commonly associated with carcinogenesis. This is a concern as lab experiments demonstrated that nearly all of the metals in these wastewaters leach from roads after rain events, likely reaching ground and surface waters. Some contaminants such as Pb and diesel range organic compounds were almost entirely retained in the road material, potentially accumulating in fine dust particles that could be transported as fugitive dust emission from vehicular traffic. The release of Ra (a known carcinogen) from roads treated with O&G wastewaters has been largely ignored. Most studies on O&G development have focused on potential contamination events from spill events, hydraulic fracturing, or facilities that discharge partially treated O&G wastewater to surface water. However, from 2008-2014, spreading O&G wastewaters on Pennsylvania roads released approximately 4 times more Ra to the environment (320 millicuries) than facilities treating O&G wastewaters and 200 times more radium than spill events. While many states that permit this practice require a certificate of analysis for wastewaters spread on roads, the analyses are often limited to a few elements (often Cl, Ca, Mg, and Na) and do not report Ra. Methods for reducing the potential impacts of spreading O&G wastewaters on roads are discussed in more detail in Chapter 2 and can also be reviewed in the journal of Environmental Science and Technology where this work was published. One way that potential contamination from O&G wastewaters can be identified in the environment is by using novel tracers. Numerous tracers have been identified and used to fingerprint contamination events from Marcellus Shale or conventional O&G development. However, there are no studies that have identified tracers for UPP produced waters (i.e., wastewaters) despite the extensive development in Ohio and the projected future development throughout the Appalachian Basin. There are also no studies that have reported the chemistry of these produced waters, which could have implications on understanding the origin of the contaminants and developing best management practices for handling and disposing these fluids. As discussed in the third chapter of this dissertation, UPP produced waters were collected from 26 mature wells (>90 days of production) throughout Ohio, Pennsylvania, and West Virginia and documented their elemental and isotopic chemistry. Compared to the Marcellus Shale and other conventional O&G formations in the Appalachian Basin, the UPP produced waters had very similar elemental chemistry (i.e., Na, Ca, and Cl) but often higher total activities of Ra (226Ra + 228Ra) and fluid chemistries that varied depending on well location. Specifically, produced water chemistry from eastern UPP wells (i.e., northeast PA, southwestern PA, and West Virginia) had higher Na/Cl and lower Ca/Cl ratios than western UPP wells (i.e. northwest PA and Ohio) that were largely explained by exchange reactions between Na+ and Ca2+. From a waste management perspective, current and future concerns with handling the fluids from UPP development will likely be related to the proper disposal of radioactive sludge generated from facilities treating the fluids, faulty well casings that allow fluid migration, or potential produced water spills at the surface. In a suspected contamination event, many of the same tracers that work for identifying produced waters from the Marcellus Shale also work for identifying UPP Shale. For instance, UPP produced waters had Sr/Ca molar ratios (0.07 to 0.13 based on the 95% confidence interval) and 87Sr/86Sr isotope ratios (0.7088 to 0.7114 based on the 95% confidence interval) that were similar to Marcellus Shale produced waters but were distinct from conventional O&G produced waters. These tracers can be used to fingerprint small fractions (~0.1%) of UPP produced water in freshwater. However, in regions of the play where there is both Marcellus and UPP development (i.e., PA and WV), the similarities between the Marcellus and UPP produced waters make it challenging to distinguish these two fluids if they mix with freshwater. If there is a need to determine if suspected contamination is from a Marcellus or UPP Shale well, the higher 228Ra/226Ra ratios (>1) in UPP produced waters can differentiate these two sources; although the sensitivity of this tracer is limited to approximately 1% admixing with freshwater before the two sources are indistinguishable. The management of wastes from O&G development, the identification and use of novel tracers for tracing O&G contamination in the environment, and the interpretation of O&G geochemistry from produced water chemistry all hinge on the accuracy of the methods used to characterize solid and liquid wastes from O&G development. As discussed in the fourth chapter of this dissertation, an inter-laboratory comparison with 15 laboratories was conducted to evaluate the accuracy of elemental and radioactivity measurements for characterizing O&G wastes. Three O&G wastewaters from the Appalachian Basin and four solid materials consisting of sediments impacted by O&G development, sediments mixed with barite sludges from O&G treatment facilities, and shale collected from a outcrop in central Pennsylvania were analyzed for a suite of inorganic and radioactive elements including Na, K, Mg, Ca, Sr, Ba, Li, B, Al, Fe, Mn, Cr, Ni, Cu, Zn, As, Pb, Cl-, Br-, SO42-, 226Ra, and 228Ra; only 226Ra and 228Ra were measured for solid samples. Analytes in the wastewaters with high concentrations (i.e., > 5 mg/L) were easily detectable with relatively high accuracy, often within ±10% of the most probably value (MPV). The good agreement between labs supports the use of Sr/Ca and B/Cl ratios, among other ratios, for tracing O&G wastewaters. However, the range in reported Br/Cl ratios indicate that geochemical explanations using this ratio could have uncertainty that should be acknowledged when interpreting data sets with values reported by multiple laboratories. Often less than 7 of the 15 labs were able to report detectable trace metal concentrations (i.e., Cr, Ni, Cu, Zn, As, and Pb) with accuracies of approximately ±40%. Method detection limits for these trace elements were frequently above or near the drinking water standards, which could make it difficult to regulate O&G wastewaters based on trace metal concentrations. The majority of the labs were able to measure detectable Ra radioactivity in the wastewaters, but the reported values had more variability than the major and minor elements. The range in the reported values suggests that individual laboratories could be over-reporting or under-reporting radium activities by ±50%. These large differences could be attributed to several sources of bias including radon leakage, self-attenuation, or calibration errors. Radium analyses of solid samples were more accurate than liquid measurements, deviating by ±20% from the MPV but could likely be improved by using certified radium standards and accounting for self-attenuation that results from matrix interferences or density differences between the calibration standards and the unknown samples. Overall, this inter-laboratory comparison illustrated that major element, minor element, and anion analyses of O&G wastewaters have relatively high accuracy while trace metal and radioactivity analyses may often be over 20% different from the MPV.