Contentious baselining: The politics of “pre-drilling” environmental measures in shale gas territory

Introduction

Environmental harm, which “assumes a baseline unharmed environment against which some event or change can be compared,” is a fundamental concept in environmental law (Aagaard, 2011: 1516). In theory, a baseline provides hope of proving that industrial activity is inflicting environmental harm, enabling a community to halt or mitigate a project. The relevant science seems simple: compare environmental conditions before and after the activity began. But baselines are often unknown, unclear, or controversial. Indeed, ignorance about baseline conditions can be advantageous to companies that are suspected of causing environmental harm, because it allows them to claim that perhaps the degradation was a pre-existing problem. In the case of Marcellus Shale natural gas extraction in Pennsylvania (USA), gas well operators have advanced “no proof” defenses (Proctor, 2008: 12) against accusations that their activities pollute well water and degrade watersheds. Drawing on this case, this paper posits that one of the ways that companies avoid accountability is by creating strategic ignorance (McGoey, 2012) about and through baselines. This can take at least two forms: (1) drawing attention to weaknesses in the documentation of past environmental states and (2) controlling access to data that was collected in the past.

This case further reveals that ordinary individuals and environmental organizations have taken an active interest in producing environmental baselines, separate from any formal regulatory process that might call for them. Indeed, in Pennsylvania, rural residents have paid for costly baseline well water tests, while volunteer watershed groups have carried out baseline studies of numerous streams across the state. A key motivation for these actions is to counter anticipated gas industry efforts to mobilize strategic ignorance. In other words, the gas industry's use of ambiguous baselines to mount a “no proof” defense is now so well known that concerned residents seek to “get ahead of it” by documenting baseline conditions as accurately and credibly as possible. However, as this article will show, public participation in baselining does not occur on a level playing field, and it is a weak strategy for holding polluters accountable. ¹ Furthermore, there are some signs that by focusing their energies on documenting “pre-drilling” conditions, concerned citizens may be distracted from broader and more urgent normative questions about the use, distribution, and protection of water.

Pennsylvania is one of the USA's biggest producers of natural gas, thanks to the recent development of the Marcellus Shale (using methods informally known as “fracking”). ² Water quality impacts can arise from many steps in the production of natural gas from shale (Brantley et al., 2014; Rahm and Riha, 2012; Vidic et al., 2013). High-volume hydraulic fracturing uses millions of gallons of water, which is injected underground under high pressure to create fractures in the shale to release the gas. This water either stays deep underground or comes back to the surface containing salts, hydrocarbons, drilling chemicals, and in some cases radioactive material (Wilson and Vanbriesen, 2012). Understandably, some people living in the vicinity of gas drilling operations are worried about the fate of these fluids and how the drilling activities overall might affect their water supplies. There have been numerous complaints to regulators. From 2004 to 2016, the Pennsylvania Department of Environmental Protection (DEP) received nearly 10,000 complaints about environmental impacts of oil and gas development in the state. According to the reporters who obtained and analyzed the agency's records of these complaints, 4108 were related to water-quality problems (Phillips and Hurdle, 2017).

Industry actors have strategically used ignorance to deflect accusations that their activities are causing water pollution. For example, the American Petroleum Institute and America's Natural Gas Alliance, two industry lobby groups, commissioned a study to examine baseline water quality in five areas that were under study by the US Environmental Protection Agency (EPA). The Battelle study, released in 2013, is indicative of an industry preparing to cast doubt on any findings that might implicate them in water pollution. Their method was to use baseline or “background” environmental data to suggest prior causes of water pollution, as well as to identify where there was insufficient data to make claims about prior environmental conditions (Battelle, 2013). But industrial interests are not alone in pointing out the gaps in baseline water quality data. Reviews of existing data by government scientists have also concluded that, because of gaps in monitoring programs, there is insufficient baseline data to detect a statistically significant change in water quality in some areas impacted by shale gas development. ³ These knowledge gaps can work to the industry's advantage.

While there is growing scholarly and popular attention to the controversial scientific claims about the impacts of shale oil and gas exploration (Griswold, 2018; Pearson, 2017; Wylie, 2018), the crucial role of baselines in these debates has generally been overlooked. Nor have baselines been deeply considered in the broader literature on industry strategies to cultivate ignorance and doubt (McGoey, 2012; Markowitz and Rosner, 2003; Michaels, 2008; Oreskes and Conway, 2010; Proctor, 2011). In this paper, I will remedy this gap by examining the dynamics of contentious baselining surrounding Marcellus Shale natural gas development in Pennsylvania. First, I will introduce the concept of contentious baselining and its relevance for understanding ignorance and environmental governance. I will then show how natural gas drilling companies sought to avoid accountability for the impacts of their activities by strategically creating and mobilizing ignorance about baseline conditions. Next I will turn to the growth of public participation in water quality baselining, exploring how private well testing and volunteer stream monitoring projects anticipated industry efforts to use unknown baseline conditions to create strategic ignorance. Finally, I will consider how law and policy generated these particular processes of baselining, and begin to imagine how alternative policies might create more inclusive and equitable approaches to environmental decision-making.

The observations reported in this paper are from research I conducted from 2009 to 2014. This research included interviews and focus groups with residents of two gas producing counties of northeastern Pennsylvania during the early part of the gas boom (2009–2010); a survey of watershed monitoring organizations throughout the state in 2012; interviews with citizen scientists, researchers, and government scientists involved in watershed protection (2012–2013); and observation of multiple trainings and other meetings for volunteer watershed monitors (2010–2014). These observations are supplemented with an array of primary and secondary documents, including regulatory documents, news reports, scientific publications, and industry websites. My relationship to one of the most intensely developed shale gas-producing counties of Pennsylvania affected how I gathered and interpreted data. I lived in a rural part of the state for most of my life until I left for college, and my parents lived there during most of the time that I was conducting this research, during the Marcellus Shale gas boom. ⁴ My perception of the gas boom is therefore inflected by direct personal experiences and through professional and personal relationships with individuals who were living through that experience as well. My personal knowledge of life in a boom town also, I believe, strengthened my credibility with some people I interviewed in other parts of the state, whether they were watershed advocates, scientists, or regulatory officials. In general, I approach this study with a critical perspective on fossil fuel extraction because of the urgency of the climate crisis and because the enormous power of the oil and gas industry remains largely unchecked. That said, this study was not designed to conclude that one group is factually right or wrong, when it comes to the specific questions about water pollution. Instead, my aim has been to examine the unintended consequences of baseline-centered governance in a context where actors have different degrees of access to knowledge resources.

Contentious baselining: Strategic ignorance and public participation

The conceptual toolkit for studying ignorance sociologically has grown considerably in recent years, with the publication of edited volumes such as Agnotology: The Making and Unmaking of Ignorance (Proctor and Schiebinger, 2008) and the Routledge International Handbook of Ignorance Studies (Gross and McGoey, 2015). Among the useful advances of this research is the expanded understanding of how and why certain actors, such as corporations or government agencies, seek to preserve ignorance rather than dispel it (Frickel and Vincent, 2011; Galison, 2004; Gross, 2007; McGoey, 2012; Richter et al., 2018). McGoey (2012: 555) offers the concept of “strategic ignorance”—the mobilization of unknowns in order to “deny liability in the aftermath of disaster” and to “assert expert control in the face of both foreseeable [and] unpredictable outcomes.” For those accused of causing environmental harm, “unknowns” can be amplified or manipulated to maintain that there is no proof that a product or activity is to blame (similar to the well-documented strategies of the tobacco industry) (Proctor, 2008). Industrial strategies for “producing doubt” and “manufacturing uncertainty” include questioning the relevance of laboratory studies or animal models, calling for media “balance,” maintaining secrecy, commissioning studies to keep a controversy open, and other activities meant to give the impression that the evidence against them is weak (Markowitz and Rosner, 2003; Michaels, 2008; Oreskes and Conway, 2010; Proctor, 2011).

Corporate actors do not mobilize unknowns in a vacuum, however; their institutional settings contribute to the production of ignorance. For example, the US pesticide industry has cast doubt on evidence of a link between their products and the death of honeybees, and their success has depended on the research norms and practices that have been formalized in EPA evidential standards for assessing harms caused to insects. Suryanarayanan and Kleinman (2016: 111) surmise that the colony collapse disorder controversy will remain unsettled “as long as the norms that govern what counts as acceptable methods and standards are not revised.” Other studies, such as Richter et al.'s (2018) recent examination of regulatory inaction on per- and polyfluorinated alkyl substances, support the argument that government policy contributes to ignorance about the harms caused by industrial chemicals. In that case, DuPont discovered that a product might cause birth defects, but kept the information secret for decades. The authors argue that the case

illustrates the risks of a regulatory structure that assumes chemicals are safe until proven harmful, and places the burden of proof of harm on residents and agencies ill-equipped to produce scientific knowledge on chemical compounds only substantively known by their private producers. (Richter et al., 2018: 708)

A small number of previous studies focusing on the mining and energy industries suggest several different ways that baselines and ignorance are linked. In Mining Capitalism, Stuart Kirsch (2014) observes that mining companies strategically reduce their accountability for environmental impacts by “failing to conduct proper baseline studies for new mining projects” (145). In a study of an environmental impact assessment (EIA) process for the oil industry in Ecuador, Amelia Fiske (2017) observes that some of the things the EIA accomplishes are to document the environmental degradation for which the oil company is not responsible and to circumscribe the kinds of “impacts” that can be legitimately considered. In Chile, a small hydropower company used outdated measurements from a different river to estimate baseline water availability, ultimately leading to significant water loss affecting the indigenous community that relied on the dammed river (Kelly, 2019). And in Peru, a process of public participation in a gold mine EIA seemed to make the mining company more accountable. However, it framed the problems to be addressed in the corporation's terms, excluding some public concerns, such as the disappearance of water springs, and long-term effects of the mine's cyanide leach pads (Li, 2009). Each of these examples illustrates aspects of how mining and energy companies create strategic ignorance through regulatory processes that involve baselining.

The case of Marcellus Shale natural gas development in Pennsylvania is notably different from previous studies of baselining that focus on formal EIA procedures. The organization of baselining practices in this case results not from policies that overtly require baselines, but rather from a combination of different rules and government programs that tend to encourage certain baselining practices. Indeed, this case provides a unique opportunity to examine informal and “extramural” (Lave, 2015) processes of setting environmental baselines. For example, gas well operators are not required to document pre-drilling groundwater quality; the law simply makes it in their interest to do so, by making them responsible for any contamination that occurs within a certain distance of a gas well, unless they can prove it was pre-existing. At the same time, the Pennsylvania DEP has provided guidance to landowners on how to sample and test their own well water, to establish a pre-drilling baseline. Additionally, federal and state agencies have encouraged concerned citizens and environmental organizations to form volunteer watershed monitoring groups, emphasizing the necessity of establishing baseline stream conditions.

What are the effects of opening up baselining to greater participation? One important possibility is that it will lead to more “faithful accounts” (Haraway, 1988) of the environments that people have inhabited by incorporating multiple ways of knowing (Karpouzoglou and Zimmer, 2016). However, participation is rarely fully inclusive. Although media outlets often describe citizen science as “democratizing science,” social factors such as child-care obligations, income, comfort level interacting with experts, and gendered expectations about appropriate use of free time are likely to limit participation in such projects. For instance, a study of one statewide volunteer stream monitoring program, Alabama Water Watch, found that the program mostly appealed to “more educated people with discretionary time and wealth,” which left monitoring gaps in disadvantaged rural areas (Deutsch et al., 2009: 647). Therefore, understanding the politics of baselining requires close attention not only to the social process of setting a baseline, but also the social exclusions that occur.

Another possibility of opening up baselining to participation is that it becomes contentious, as in other cases of citizen science (Allen, 2003; Brown, 2007; Frickel et al., 2009; Ottinger, 2013). By contentious, I mean it “involve[s] making claims that bear on someone else's interests,” such as their finances or well-being (Tilly and Tarrow, 2015). Ureta (2018) observes that baselining commonly involves controversies and power struggles. Indeed, baselining can develop the character of an “arms race,” a competitive buildup of baseline knowledge that is expected to lead to victory in a future contest over environmental pollution (Kinchy, 2017). In the context of Marcellus Shale development, claims about the prior quality of water have implications for the industry's financial interests, public interests in conserving natural resources, and private landowners' interests in having safe water to drink. Other actors' interests are also implicated, such as state legislators' interests in maintaining a flow of tax dollars from industrial activity or winning elections on the strength of their environmental positions.

The contentious nature of baselining stems, ironically, from the presumed neutrality and objectivity of baselines. A baseline's distance from the present disagreement makes it a trusted arbiter. People on all sides of an environmental conflict have an interest in establishing a baseline, because a baseline appears to provide a point of comparison, occurring before (and ostensibly untainted by) the moment of conflict. The neutrality of baselines is illusory, however. Scientists may view baseline studies as “pure” research (Barandiarán, 2015), but many decisions involved in creating a baseline are inflected by unstated or unrecognized cultural and political judgments. For example, in Ureta's (2018) ethnographic study of soil sampling in Chile, decisions about where to sample were affected by cultural frames of reference about what constituted “natural” soil. In the present case, people make similar decisions about which parameters represent “water quality” (though further details about this are beyond the scope of this article). All baselines are the product of baselining—they never arrive from nature in an unmediated way—yet as the following case will show, many hopes are pinned on the purity of baselines.

Mobilizing unknown baselines

Pennsylvania has more than one million private wells supplying water to 3.5 million rural residents (Penn State Extension, 2018). Many of these residents live in places where the gas industry is now active. Pennsylvania's approach to governing potential impacts of the shale gas industry on these water wells is unique (Richardson et al., 2013: 30). Under the 2012 Oil and Gas Act, gas well operators are presumed responsible for any well water pollution occurring within 2500 yards and within 12 months of completion, drilling, stimulation, or alteration of a gas well. No pre-drilling water tests are mandated, but “if tests are not done before development, operators are barred from claiming in future legal action that any alleged groundwater contamination was preexisting” (Richardson et al., 2013: 30).

The effect of this law has been to incentivize natural gas companies to routinely collect pre-drilling data, documenting any pre-existing contamination, so they will not be forced to take responsibility for water problems that pre-date their activities. This approach to baselining clearly differs from the examples of EIA processes seen in other studies of the mining and energy industries, discussed earlier, because they are private and voluntary, and are not intended to aid decision-making prior to development. The well operators hire independent certified laboratories to test well water prior to drilling, providing a copy of the results to the owner of the water source (Boyer et al., 2011). Under regulations put into place in 2016, they must also submit the results to the DEP, which does not make them available in a database, but holds the PDF documents on file. ⁵ The testing companies analyze water for dozens of potential contaminants (Pennsylvania Department of Environmental Protection, 2016).

Early in the development of the Marcellus Shale, however, there was a major controversy over groundwater in which an unknown baseline was mobilized to create strategic ignorance about environmental harm. The events took place in the small village of Dimock, located in the northeast corner of the state. In 2009, an elderly woman's drinking water well filled with methane gas, causing an explosion. Several other Dimock residents complained that their well water was fizzy, discolored, and had a bad smell. Pennsylvania's DEP determined that Cabot Oil and Gas, a company that had drilled several natural gas wells in Dimock, had constructed its wells poorly, allowing methane to migrate into groundwater, affecting the water supplies for 13 homes within 1300 feet of one or more Cabot natural gas wells. The DEP concluded that Cabot had allowed “the unpermitted discharge of natural gas, a polluting substance, into the groundwater,” in violation of multiple state laws. ⁶ In the 2009 Consent Order and Agreement, the DEP also listed additional violations, such as several incidents of discharging industrial waste onto the ground and into surface water. For these violations, Cabot was ordered to halt all drilling and hydraulic fracturing activities in the affected area, pay a fine, correct its violations, and provide potable water to the homes whose water supplies were contaminated with methane.

Cabot strategically used gaps in knowledge about baseline groundwater quality in its efforts to defend itself against these charges. Federal environmental law—the Safe Drinking Water Act—protects the quality of groundwater sources that serve municipalities, but it does not address the safety of water in individual wells. As a consequence, there is no public record of well water quality. The company claimed the groundwater already had fluctuating levels of naturally occurring methane, contesting the DEP's assumption about baseline water quality and the cause of the methane pollution. There was little data to support this claim, because of the absence of baseline data about private well water quality. However, many members of the community came to the defense of Cabot. As Jerolmack and Walker (2018) have documented, property owners in this region often mobilized support for the gas industry, even when their personal benefits were limited, because they believed in individual property rights, were cynical about government regulation, and viewed critics as “liberal” outsiders. A common refrain on pro-gas industry websites and blogs was that the “anti-frackers” in Dimock were lying about changes to their water and that methane has always been present in Dimock-area wells. ⁷

By the end of 2010, the DEP determined that Cabot had engaged in unlawful conduct because it had still not taken the required remedial action. ⁸ Under threat of litigation, Cabot offered each resident a whole-house gas mitigation water treatment system and, for each of the 19 households, created an escrow account worth twice the owner's property value. In order to resume drilling activities, the company also had to show for two straight years that dissolved methane in the affected wells was below a threshold of 7 milligrams/liter (below the concentration that poses a safety risk). Baseline conditions again had an important role in this settlement agreement. As journalist Laura Legere (2012) pointed out, “the state can waive those requirements if it finds that the methane levels in the water now reflect natural conditions unrelated to shale development - a tricky task because no methane samples were taken before drilling began.”

Cabot's actions did not satisfy many of the affected residents, who claimed that the water treatment systems were not effective and that the payments Cabot offered were based on underestimates of the property value (Public Herald, 2011). They worked together to pursue a class action lawsuit, but most of the plaintiffs eventually accepted settlements. Two families continue their battle with Cabot in the courts. In the meantime, scientific studies of water contamination in Dimock have been controversial, with some investigations declaring the water safe to drink and others identifying hazardous contaminants such as arsenic (Agency for Toxic Substances and Disease Registry, 2016). To date, Cabot and its supporters continue to argue that we simply do not know that the water was good quality before the drilling began. While this strategy has not absolved the company of consequences, it has helped the industry as a whole to maintain that there is no proof that its activities affect groundwater.

Controlling access to baseline data

Of course, because gas well operators are now paying for pre-drilling well water tests, a tremendous amount of baseline data is being collected, but this information is not accessible to scientists or the public. Controlling access to baseline data is the industry's second use of strategic ignorance about how shale gas development might be affecting groundwater.

Research scientists have been frustrated by the lack of access to the industry's trove of well water tests. In 2011, researchers at Duke University published a study that found a correlation between methane levels in water wells and proximity to natural gas wells in northeastern Pennsylvania and a part of upstate New York where shale gas extraction was taking place. Many of the 68 drinking water wells sampled were in the county where Dimock is located (Osborn et al., 2011). The study was attacked by the industry lobby group, Energy in Depth, which claimed the study was invalid because the researchers lacked baseline data for the region—there were no pre-drilling samples. The paper was so controversial that one month after its publication, the authors released a “Frequently Asked Questions” document, responding to questions about and critiques of the study (Jackson et al., 2011). First on the list was the comment that the study lacked baseline sampling. The authors responded with an analogy to epidemiological research:

Because we were unable to travel back in time and sample every home pre-drilling, our initial experiment used the next-best approach to evaluate systematic variations in methane concentrations and isotopic fingerprints. Here is an analogy to describe why. What were the earliest tests examining a possible link between smoking and lung cancer? Scientists didn't follow people for years through their lives, sampling the people before they started smoking and then tracking their health for decades afterwards. That would have been the ideal experiment to prove cause and effect. Instead, scientists asked a simpler question first: “If you smoke, are you more likely to get lung cancer?” That the answer was “yes” didn't prove cause and effect. It did suggest that something important was occurring and that follow-up research was needed to identify the mechanism of any link between smoking and cancer. (Jackson et al., 2011)

How could scientists follow up to identify the causal mechanism? While pre-drill tests carried out by the gas industry constituted the most comprehensive collection of groundwater data ever amassed in Pennsylvania, it was not disclosed to the public or shared with researchers (Cusick, 2013). Speaking to a journalist, the lead researcher on the Duke methane study, Robert Jackson, said “The industry is sitting on hundreds of thousands of pre and post drilling data sets … I asked them for the data and they wouldn't share it” (Lustgarten, 2011). I observed this sentiment repeated at an interdisciplinary conference on shale gas, where academic researchers confronted gas industry scientists and engineers about their industry's unwillingness to share data. Industry representatives said they were protecting landowners' privacy.

Rather than release the data, companies commissioned their own analyses, which countered accusations that the company was not making use of its dataset for scientific research. Chesapeake Energy Corporation (one of the biggest operators in the region) contracted with a geology professor at Syracuse University, working with current and former Chesapeake employees and consultants, to carry out an analysis of the groundwater data the company had collected from over 20,000 wells in Appalachia. The results, published in the journal Applied Geochemistry, supported the perspective that drinking water quality in the region was already poor, prior to Marcellus Shale development. The authors reported: “What we find confirms what others have historically reported: there is wide variability in water quality in potable groundwater in the Appalachian Basin, and much of this water naturally exceeds regulatory standards” (Siegel et al., 2015: 54). A similar study was conducted by scientists at a private consulting firm, GSI Environmental, Inc., using data collected by Cabot Oil and Gas (Molofsky et al., 2013). This study concluded that methane was common in water wells in Susquehanna County (where Dimock is located), unrelated to shale gas extraction. The results of these studies lent credibility to the industry's longstanding strategy of casting doubt on claims that gas drilling was to blame for unsafe drinking water.

This particular use of baselines to create strategic ignorance has been contested in the scientific community. Initially, when the study using Chesapeake data was published, other scientists raised questions about its methodology and about the authors' undisclosed ties to the company (Banerjee, 2015). Subsequently, some scientists have secured access to the gas companies' “pre-drilling” datasets. One study analyzed the well water test results that gas companies submitted to the Pennsylvania DEP (Wen et al., 2018). Their analysis of data for one Pennsylvania county did not reveal any significant change in groundwater since the start of shale gas development. Yet they did find evidence of a small number of sites of methane contamination near gas wells (Wen et al., 2018). Another study made use of the data used in the GSI Environmental study cited above, because that dataset was published along with the paper (Yan et al., 2017). The new analysis of that data found that about 20% of the “pre-drill” samples were actually taken within 1 kilometer of an already-drilled gas well, making it possible to treat them as “post-drilling” samples. Using that data, they found increased concentrations of some contaminants in drinking water near sites where hydraulic fracturing had taken place.

These examples suggest that controlling access to baseline data has been difficult for the industry to sustain, in part because of requirements to provide the data to regulators (though not in a form that is easily analyzed), and in part because it can enhance credibility to adhere to scientific norms such as publishing peer-reviewed studies and providing the companion dataset. Nevertheless, there remain significant barriers to gaining access to baseline well water data. This leaves scientists without access at a disadvantage in a context where a lack of baseline sampling can leave them vulnerable to extensive criticism.

Public participation in baselining

The valorization of “baseline data” has extended beyond the natural gas industry and scientific community and has become an imperative for volunteer groups and private landowners seeking to protect water. Many landowners distrust the drilling companies to provide truthful reports on their well water quality and have hired water testing consultants to do baseline testing. This is a widespread practice, facilitated by the state. The Pennsylvania DEP provides a fact sheet with a list of recommended parameters to test, advice on finding a laboratory to perform the testing, and what to ask the laboratory before sending samples. ⁹ Penn State Extension (a public outreach branch of the state university) also provides advice on pre-drilling well testing, in the form of a webinar and other online information for landowners. Complete testing can be expensive—in 2011 they could cost up to $1640 per sample—which is prohibitive for many residents (Boyer et al., 2011: 11). Yet one study found that 28% of water supply owners had paid for pre-drilling water tests while also receiving results from tests purchased by gas well operators (Boyer et al., 2011: 11).

There are presumably thousands of Pennsylvania residents who have participated in groundwater baselining, but participation has been private and individualized, as landowners pay for well water testing, or interpret the test results provided to them by gas well operators. The fate of this data is unclear. If the reports are used in legal conflicts with gas well operators, it might not become public knowledge, since legal settlements about well water pollution often come with gag orders (Alagood, 2015).

While groundwater was at the center of the debates and controversies over fracking throughout this time period, many local organizations also raised questions about potential impacts on surface water. Beginning around 2009, communities affected by gas development across the state of Pennsylvania launched citizen science projects to gather information about their local streams. These activities grew out of decades of volunteer capacity-building efforts. Since the 1980s, the US EPA has encouraged states to recruit volunteer “stream watch” teams to gather surface water data (USEPA Office of Water and USEPA Office of Policy, Planning, and Evaluation, 1987). Efforts to facilitate volunteer watershed monitoring included the development of testing methods that are relatively easy to teach to volunteers and inexpensive to carry out (Pfeffer and Wagenet, 2007). In Pennsylvania, a state-wide volunteer water monitoring initiative was launched in 1996. Although that program was phased out in 2009, the community groups that had previously participated in water monitoring had the skills, knowledge, and networks to start monitoring streams for the impacts of shale gas development (Jalbert et al., 2014). In some cases, community groups seeking to monitor the effects of fracking were motivated by a desire to maintain the water quality that they had worked hard to attain, after decades of pollution resulting from coal mining and other industries.

The policy context for volunteer stream monitoring differs from that of groundwater testing in several ways. The Clean Water Act mandates that states clean up surface water and provide annual reports on surface water quality. There are several federal, state, multi-state, and local government agencies that routinely collect surface water data in a limited number of locations across the state (Kinchy et al., 2016). The presence of (some) ongoing regulatory monitoring contrasts with the absence of government monitoring of private well water supplies. Another difference with groundwater is that Pennsylvania law does not require or incentivize natural gas companies to establish a surface water quality baseline prior to drilling. These policy differences mean that different actors are involved in baselining surface water quality. In particular, nonprofit organizations and volunteer groups play a more prominent role, and they sometimes see themselves as filling gaps in regulatory science.

In the fall of 2010, the Alliance for Aquatic Resource Monitoring, located at Dickinson College in Carlisle, Pennsylvania, released its “Marcellus Shale Volunteer Monitoring Manual.” ALLARM's protocol uses total dissolved solids (TDS) and conductivity as “red flag” parameters to indicate possible contamination. These are easily measured by volunteers using a hand-held electronic device. Volunteers establish a schedule for taking measurements of TDS at chosen stream sites. These measurements constitute the baseline conditions of the stream. Subsequent monitoring is always compared against the baseline, and unusual readings trigger the collection of samples for laboratory testing for “signature chemicals”—barium, strontium, and total alpha. If these are found, it is a sign that contaminated fluids from hydraulic fracturing operations have polluted the waterway. ALLARM went on to partner with a variety of organizations to conduct at least 60 workshops for volunteers across the state (Alliance for Aquatic Resource Monitoring, 2018).

When I interviewed volunteer stream monitors in 2012 and 2013, many anticipated that their baseline data would be contentious if a pollution event occurred. A common sentiment was that baseline data would be “ammunition” in some unspecified future confrontation with a polluter. Most stream monitoring groups focused on gathering long-term data and saw themselves as establishing a baseline to which future water quality could be compared. Volunteers often conceptualized baseline data as protection against industry lies and manipulation. Having an established baseline was expected to “keep people from trying to ‘snow’ you that nothing is wrong.” ¹⁰ Some believed that having good baseline data would compel regulators to respond to pollution incidents. In general, they presumed that a factual baseline is the key to proving environmental harm.

While volunteers have observed some negative impacts on shale gas development on surface water over the years, baseline data have not been the ammunition that many anticipated. As of 2015, ALLARM volunteers' chemical tests of water quality (sensing of TDS and conductivity) had not revealed any fracking-related pollution (Wilderman and Monismith, 2016). They cautioned that this should not be taken to mean that shale gas development does not cause pollution, since most of the volunteers' measurements were aimed at establishing a baseline for watersheds that did not have a shale gas well. In contrast, on at least 44 occasions, volunteers trained by ALLARM visually observed and photographed pollution linked to natural gas development. Such incidents—mostly related to soil erosion—were apparently resolved without baseline data. For example, “One pipeline observer captured mudslides in Tioga County, Pennsylvania where the County Conservation District took the lead in addressing the situation with the company” (Wilderman and Monismith, 2016).

In these examples, baselining is a strategy for the public to hold industry accountable for the pollution it causes, but there are at least two major weaknesses. One is that existing baselining practices do not allow for the kind of aggregation that would support critical assessment of the industry's impacts and collective action to protect valued environments. Patterns of environmental change can be masked by aggregating or disaggregating data in particular ways (Frickel and Kinchy, 2015), and the current organizational arrangements for baselining both groundwater and surface water hinder aggregation of data at scales that might give affected communities leverage. In the case of surface water, volunteers are collecting data, but they typically lack the infrastructure and expertise to combine and analyze it themselves. This problem is being partially addressed by university-led projects that have sought to compile citizen science data into regional and state-wide water quality databases, but it remains unclear what kind of analytical leverage these databases will provide for the communities themselves. ¹¹ Private well water testing hinders aggregation for different reasons: data are individualized and distributed across numerous testing companies. This approach to baselining, which centers on the individual property owner, discourages collective representations of water quality problems.

The second weakness is that participation in baselining is highly unequal. Notably, we see the disproportionate influence of the natural gas industry, which resonates with the situation described by Fabiana Li (2009) in Peru: “the use of science to produce counter-information inherently creates an uneven playing field where small organizations with limited resources face corporations that spend incomparable amounts of money conducting scientific studies” (229). In the case of groundwater, the industry appeared to have far greater influence than public interest groups in establishing or questioning baselines. To the extent that residents of the affected areas had a role in establishing baseline groundwater quality, they did so as private individuals, often paying large sums for third-party tests. The gas drilling companies that possess thousands of water quality data points are strengthened by information asymmetry, since they alone possess the aggregated data to represent the region's water quality.

In the case of surface water monitoring, the industry has been less directly involved in establishing a baseline; yet threats of industry challenges have an obvious influence on citizen monitoring strategies. Water monitoring groups were typically in a defensive posture, preparing to protect themselves from future harms they expected the industry to cause—and to deny. Yet some volunteers did not want to come into direct conflict with industry, and, according to ALLARM, may have held back from sharing their observations. The authors of the report from ALLARM noted that “Volunteers are extremely cautious about reporting violations because of the contentious nature of the situation [and] actually may have under-reported probable incidents” (Wilderman and Monismith, 2016). It seems that volunteers have anticipated conflict and chosen to avoid or delay it, by focusing on streams where there is little industrial activity and refraining from reporting violations.

Beyond the lack of parity between the industry and local residents who participate in baselining, there are disparities between communities that result in uneven monitoring of watersheds and disparities among individuals that result in uneven access to baseline well water quality tests. By 2012, over 30 organizations were monitoring the impacts of Marcellus Shale development on surface water in New York and Pennsylvania, covering a broad geographic territory. These efforts extended to roughly half of the watersheds in Pennsylvania's Marcellus Shale region (Kinchy et al., 2016). This was a remarkable feat of community organizing, but participation was geographically uneven. Some of the areas with intensive drilling activity never developed volunteer stream monitoring programs. Kinchy et al. (2016) mapped the geographic distribution of government and volunteer surface water monitoring, finding that there were notable gaps in some of the areas where the natural gas industry is active. This implies that communities have varying levels of scientific and organizational resources to make baseline measurements of surface water quality. With respect to groundwater, disparities across households are likely because of the high cost of well water testing. If baseline data are believed to be protection against an industry's use of strategic ignorance, this protection is not equally shared.

Toward policy alternatives

Is gathering baseline data to demonstrate environmental change the best way to problematize shale gas development? In addition to the weaknesses noted above, focusing on quantitative measures of past water quality seems to sideline more significant normative questions about the use and value of water, now and in the future. The accuracy of the baseline measurements appears to be prioritized over the fairness of the relationship between the industry and the residents of these regions. More concretely, there is some evidence that the work of gathering baseline water quality data can have a depoliticizing effect on volunteers. For instance, as described above, focusing on “pre-drilling” conditions has drawn volunteer stream monitors' attention away from places where environmental contamination—and broader social harms—may already be happening. For these reasons, it is worth taking a second look at the public policies that have drawn focus toward baselines, and consider alternative possibilities.

As described above, public policies have created two distinctive approaches to baselining for groundwater and surface water in Pennsylvania. It is important to recognize that these differences in the organization of baselining practices result not from policies that overtly define how baselining should proceed, but rather from a combination of different rules that tend to encourage certain ways of baselining. This is a notable difference from previous studies of baselining that focus on formal EIA procedures, and suggests that contentious baselining is probably occurring in other environmental disputes where there is no official requirement for a baseline. This merits further research.

Beyond research implications, this study suggests a need to reconsider the widespread dependence on baselines for resolving environmental disputes. Legal scholar Todd Aagaard (2011) has criticized baselines as a basis for establishing environmental harm, because baselines may be factually unclear, are easily manipulated, and “entail hidden normative assumptions about the problems at issue” (1523). These critiques are clearly relevant to the present case and should inform attempts to reform policies surrounding the natural gas industry. Consider the possible outcomes of one policy idea, to require public disclosure of all the well water data that natural gas companies are collecting. This would certainly be opposed by the gas industry, but even if this disclosure were to occur, it would probably not fundamentally change the conflictual dynamics surrounding water quality baselines. For example, questions might be raised about whether samples were representative of the region, whether too much time had passed between the test and the gas drilling activities, whether the right parameters were chosen, and so on.

As an alternative, Aagaard argues in favor of a “use-conflict” framework for environmental law. Rather than relying on proving harm compared to a baseline, environmental law premised on use conflicts would provide a way to manage conflicts among the various benefits people derive from environmental resources. Aagaard (2011) claims that this approach “enables fair comparisons of competing normative claims—whether they are claims about a specific policy issue or competing general normative perspectives—thereby facilitating more thoughtful deliberation about the often difficult underlying issues of environmental lawmaking” (1509). Could we begin to reform the specific aspects of the state and federal law and policy that hinder thoughtful deliberation about the past, present, and future of water quality? Are there policy alternatives that more effectively bring to light the competing normative perspectives at stake in debates about shale gas?

To give just one possible policy idea, what if towns and villages were required to have a community-backed plan for the use and care of their shared streams and groundwater, before any subterranean development (i.e. drilling and fracking gas wells) could occur? Negotiating such a plan would require the articulation of social values, economic interests, and visions of human interdependence with nonhuman nature. To create a community water plan, people would need to draw upon scientific knowledge, but any facts about the “baseline” would play a supporting role to more significant normative debates, rather than determining the outcome. For example,

There may be scientific uncertainty about the health effects of the substance [that is discharged into water], in which case there are normative questions to be answered about how much evidence should be required to prove or disprove a causal link and who should bear the burden of proof. (Aagaard, 2011: 1537)

Conclusion

Baselining can involve making claims that bear on others' interests; in other words, it can be contentious. Yet the politics of baselining are hard to recognize because baselines are often presumed to be neutral representations of past conditions. One aim of this article has been to bring visibility to this arena of political struggle. Through a focus on Marcellus Shale natural gas development in Pennsylvania, this article has explored how companies avoid accountability by creating and mobilizing ignorance about and through baselines, and how informal public participation in baselining may or may not work to increase corporate accountability. In this case, suspected polluters drew attention to weaknesses in documentation of past conditions and controlled access to documentation of environmental measurements taken prior to gas drilling.

I have suggested that to understand the production of ignorance via baselines, it is necessary to examine the institutional setting, not just the strategies of the actors seeking to avoid accountability. I showed that a combination of state and federal policies created incentives for gas well operators to collect baseline groundwater data—but not surface water data—and to control access to that data. In the absence of public policy requiring government monitoring of groundwater that supplies private drinking water wells—or aggregation of data supplied by the gas industry—there are widespread unknowns about the quality of these water sources. Furthermore, while groundwater and surface water are connected in the water cycle, an entirely different set of policy arrangements condition what is known and not known about how surface water is changing.

If this case is not unique, then it suggests that industry uses of baselines to evade accountability are likely to generate counter-baselining reactions from concerned residents, in order to resist industry “no proof” defenses. This reaction is understandable, particularly at a time of increasing accessibility of “citizen science.” However, the case examined above highlighted the weaknesses of baselining as a strategy to hold industry accountable for environmental harm. This critique should not be read as a rejection of baselines as comparison points for making knowledge about change; my point is that expanding participation in the kind of baselining that is currently encouraged by the regulatory state does not substantially shift power to communities. It is often assumed that public participation in environmental testing and monitoring is democratizing, but we have seen that significant power imbalances remain, even when many individuals are participating in baselining. Addressing these inequalities will require more significant transformations to how we collectively make environmental decisions.

Highlights

In the case of Marcellus Shale natural gas development, water quality baselines have been contentious.