Wastewater management strategies for sustained shale gas production

Recent advances in shale gas development have largely outpaced efforts to manage associated waste streams that pose significant environmental risks. Wastewater management presents significant challenges in the Marcellus shale, where increasing fluid volumes concomitant with expanding development will threaten to overwhelm existing infrastructure over the next decade. In this work, we forecast growth in drilling, flowback, and produced fluid volumes through 2025 based on historic data and consider conventional and alternative disposal options to meet future demands. The results indicate that nearly 12 million m3 (74 MMbbl) of wastewater will be generated annually by 2025. Even assuming wastewater recycling rates in the region rebound, meeting increased demands for wastewater that cannot be reused due to poor quality or logistics would require significant capital investment to expand existing disposal pathways, namely treatment and discharge at centralized facilities or dedicated brine injection in Ohio. Here, we demonstrate the logistical and environmental advantages of an alternative strategy: repurposing depleted oil and gas wells for dedicated injection of wastewater that cannot otherwise be reused or recycled. Hubs of depleted wells could accommodate projected increases in wastewater volumes more efficiently than existing disposal options, primarily because the proximity of depleted wells to active production sites would substantially reduce wastewater transport distances and associated costs. This study highlights the need to reevaluate regional-scale shale wastewater management practices in the context of evolving wastewater qualities and quantities, as strategic planning will result in more socially and economically favorable options while avoiding adverse environmental impacts that have overshadowed the environmental benefits of natural gas expansion in the energy sector.


Introduction
The rapid development of unconventional resources has transformed US energy systems since the mid-2000s. Advances in horizontal drilling combined with high-volume hydraulic fracturing have rendered hydrocarbons trapped in low-permeability source rocks economically recoverable, leading to a resurgence in domestic oil production and a transition from coal-to gas-fired power generation that has lowered US CO 2 emissions to 1990 levels [1]. Despite clear economic and environmental benefits, shale gas production remains controversial due in part to impacts associated with the storage, transport, and disposal of fluids that return to the surface [2,3]. The quality and quantity of wastewater vary with geochemical characteristics of the shale play but present inevitable logistical and economic challenges that will only continue to escalate as the industry expands. Wastewater disposal practices have largely been driven by economics and logistics, but with shale gas expected to maintain a significant role in our energy systems for decades to come [4], new strategies will be crucial to minimize environmental and social impacts.
In this work we focus on a regional case study of the Marcellus shale in Pennsylvania, which is currently Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. the most prolific shale gas play in the US [5]. The Marcellus is relatively dry as low levels of formation brine along with strong capillary imbibition limit water recovery. Additionally, according to a state-wide analysis by Lutz et al [6], unconventional wells generate less wastewater per unit of gas produced than conventional wells in the Marcellus region. However, the sheer magnitude of unconventional gas production across the play has drastically increased statewide wastewater volumes, which quadrupled from 2006 to the peak of drilling activity in 2011 [6]. Managing such large quantities of fluid is further complicated by its low quality, as high concentrations of dissolved salts along with trace metals and radioactive species present significant treatment challenges [7]. The risk of spillage during transport or leakage from storage facilities also raises concerns because even low-volume spills of highly saline wastewater can significantly affect freshwater resources [8]. While prior work has emphasized the environmental consequences of past wastewater handling and discharge practices [3,6,9], continued development in the Marcellus warrants a reevaluation of existing disposal options to ensure future wastewater volumes can be sufficiently and safely accommodated.
The objective of this study was to evaluate current trends in Marcellus wastewater management and explore the economic and environmental implications of alternative strategies to meet growing demands. Specifically, we assess the feasibility and sustainability of converting depleted Pennsylvanian oil fields into storage hubs for dedicated wastewater injection and estimate costs relative to existing treatment and disposal options. While a complete life cycle analysis is beyond the scope of this work, we forecast statewide wastewater volumes through 2025 based on historic data to demonstrate challenges and opportunities associated with continued shale gas and concomitant wastewater production.

Wastewater trends and projected challenges
Shale gas production can generally be broken into three stages that each produce unique streams of wastewater. When drilling new wells, drilling mud consisting of water and clay is circulated as a lubricant and coolant [3]. These fluids along with rock cuttings from the wellbore must be separated at the surface into liquid drilling wastewater and solid cuttings that are typically sent to landfills [10]. Drilling wastewaters are typically low in volume but high in total dissolved solids (TDS) [6]. Following well completion, which involves perforating the wellbore and injecting large volumes of fluid to hydraulically fracture the surrounding shale, fluids are allowed to flow back to the surface. Initial flowback volumes are large, consisting primarily of injected frac fluids and formation brines with small amounts of entrained gases that increase over time. As a result, flowback wastewater typically also contains high levels of TDS as well as varying quantities of residual frac chemicals and trace metals leached from the formation [3]. Once the water content of the produced gas-fluid mixture is low enough to process and transport the gas, flowback ends and production begins. However, gas streams must still be separated from small (on the order of 1-2 m 3 d −1 ) and often intermittent volumes of produced fluids over the remaining lifetime of the well [3]. Produced fluids consist primarily of high-salinity formation brines and are generally stored on-site in lined pits or rigid tanks until volumes are large enough to justify transport for treatment or disposal [11]. Note that the flowback period is operationally defined in practice but is here assumed to comprise the first 30 days of fluid production after bringing a well online per operator reporting requirements [12].
Statewide volumes of wastewater associated with Marcellus shale gas production were projected through 2025 based on historic trends and natural gas price forecasts. Full details of this analysis are provided in the supporting information (SI), available online at stacks.iop.org/ERL/15/024001/mmedia. Wastewater data published by the Pennsylania Department of Environmental Protection (PADEP) from 2011 to 2018 was first aggregated by year and filtered by stage (i.e. drilling, flowback, and production). To assess trends in disposal practices, flowback and produced fluid volumes were also broken down by their destination: direct reuse at another frac site with no treatment; recycling, i.e. reuse following some treatment; treatment and discharge at a centralized waste treatment (CWT) facility; injection into a dedicated underground disposal well; or unspecified storage and transfer. Volumes of drilling, flowback, and produced fluids were forecast individually based on projected drilling rates, well completions, and fieldwide active well counts. Since its peak in 2011, drilling activity in the Marcellus has generally been trending downward, lagging a steep drop in natural gas prices in 2009. As noted in section S2 of the SI, our estimates assume drilling will continue to decline in the nearterm but will begin to slowly rebound in 2020 as gas prices slowly rise. Even with this anticipated growth, by 2025 our projected drilling rates fail to reach 2015 levels, which had already dropped 60% from their peak in 2011 (see SI, figure S1). This forecast may be realistic given the natural lag associated with remobilizing rigs in response to market shifts, especially if prices recover incrementally, but our estimates are likely conservative in light of the current political climate where a renewed focus on domestic fossil energy development could more rapidly restore drilling activity.
The resulting forecasts (figure 1(a)) indicate wastewater production will continue to rise in the near term despite marginal rebounds in drilling activity because the total volume consists primarily of produced waters that are generated over the lifetime of a well. Total annual wastewater volumes are projected to near 12 mM 3 (74 MMbbl) by 2025, which represents a 66% increase over the total volume generated in 2018 (the most recent full year of available data). Note that we do not account for any well closures during this time frame but acknowledge that wells drilled in the formative years of the Marcellus boom may be nearing the end of their anticipated lifetimes (up to 30 years) [13,14] by 2025. Managing increasing quantities of high-salinity produced waters within regulatory constraints will be a critical challenge as wastewater volumes threaten to overwhelm existing infrastructure [6]. Historically, most wastewater has been reused at other frac sites (either directly or with minimal treatment) to reduce costs of water sourcing and disposal [2,15]. However, direct reuse has been declining in recent years, and combined reuse and recycling rates fell from around 87% from 2011 to 2015 to 55% in 2016 following steep declines in drilling activity ( figure 1(b)). The combined reuse and recycling rate dropped from 94% (64% reused and 30% recycled) in the first half of 2016 to 37% (22% reused and 15% recycled) in the second half of the year. This decline in reuse was primarily offset by a dramatic increase in injection, which accounted for an unprecedented 44% of year-end wastewater disposal and 63% for the second half of 2016. While disposal trends returned to 'baseline' in 2017 and 2018, with reuse and recycling accounting for ∼85% of wastewater generated, the fate of unrecycled wastewater again shifted as the fraction of injected wastewater fell to less than 8% while the fraction held in storage rose from <0.1% in 2016 to 9% in 2017 and 8% in 2018. The 2017 reports also have a new category for 'unused fracturing fluids' that may further reflect declines in drilling and completions, but the volumes of reported unused fluids were <0.1% of the total wastewater generated.
Declines in direct reuse combined with an increase in storage may reflect degrading qualities of flowback and produced waters along with reduced drilling activity, where fluids are held either in anticipation of resuming well completions or due to a lack of viable disposal options. Even if drilling activity recovers, reuse of flowback waters will likely decline with continued development as operators move out of shale gas hotspots and coordinating successive reuse becomes less feasible among distanced wellpads. Furthermore, while the generation of large volumes of water in a short time frame makes flowback waters amenable to reuse in subsequent wells or frac sites, the produced waters that constitute the majority of total wastewater production are typically not reused due to logistical challenges associated with storing and transporting incremental and variable wastewater volumes [2]. Produced fluid reuse may also be precluded by inadequate water quality, such as excessive levels of TDS or divalent cations that promote scale formation. Considering that produced fluids will increasingly dominate future wastewater flows with declining well completions and are generated more intermittently as wells age, strategic planning is needed to manage variable wastewater qualities and quantities with minimal environmental consequences.

Reevaluating wastewater management options
Existing options to handle wastewater that cannot be reused due to logistics, lack of demand, or poor quality primarily fall under two categories: treatment and discharge at a centralized facility or deep underground injection [2]. Beneficial reuse in other municipal or industrial operations (which falls under 'other' in figure 1(b)) is limited due to lack of economic incentives and potential liability issues for operators. Since municipal treatment facilities stopped accepting fracking waste in 2011, a practice that was officially prohibited by the EPA in 2016 [16], treatment options have been restricted to CWT facilities that are equipped to handle oil and gas wastewaters and hold NPDES permits for discharge to the environment. However, increasing stringency of effluent standards for TDS and trace metals have limited the amount of wastewater sent to treatment and discharge [17]. As a result, most excess wastewater is injected into dedicated brine disposal wells. Because PA only has 10 permitted disposal wells [18] (not all of which are currently active or accept oil and gas waste), this practice predominantly involves shipping wastewater to Ohio, which currently operates over 200 brine disposal wells [2,19]. However, recent links between brine injection and induced seismicity in OH along with growing competition for wastewater disposal sites from Utica shale development will likely reduce storage space accessible to PA operators, particularly as current disposal fields reach capacity [3,20]. Even assuming the combined reuse and recycling rate rebounds to its peak of 88% (2014-2015), meeting the projected wastewater demands in the year 2025 would require 25 new dedicated wells injecting at a rate of 1000 bbl/d or 5 large CWT's processing 5000 bbl/d of exclusively fracking wastewater. Expanding existing disposal options will likely be insufficient considering the lead time and capital investment involved in building new treatment facilities; enduring concerns over the release of fracking waste streams, even with strict effluent standards; and the social and environmental implications of increasing waste shipments to Ohio.
While repurposing depleted oil and gas wells is not a novel concept, Pennsylvania has an immense untapped disposal repository in its aging and depleted northwestern oil fields that could accommodate nearterm Marcellus wastewater fluxes more sustainably than existing options. Injection may be the preferred long-term disposal strategy as the Marcellus ages because the produced waters driving overall wastewater generation exhibit high TDS levels that increase with production time and, as previously noted, treating high-salinity wastewaters for beneficial reuse or discharge is costly and energy-intensive [21]. Additionally, investing in new centralized treatment capacity to handle future wastewater volumes may pose the risk of stranded assets when production declines and demand fades for plants designed to rigorously treat high volumes of low-quality fracking waste. Compared to dedicated brine disposal wells drilled explicitly for the purpose of wastewater injection, 'flipping the switch' on existing depleted wells may pose a lower risk of induced seismicity because the formation is essentially being re-pressurized. Furthermore, repurposing existing wells near zones of high production activity may reduce the amount of produced water stored in on-site surface impoundments that are subject to leakage. One critical challenge will be locating and remediating compromised wells within injection horizons, which could create potential leakage conduits. However, rigorous site characterization and remediation to enable reinjecting produced brines in their original formations may be more economically and environmentally efficient than injecting fludis across state lines or developing extensive infrastructure for treatment and discharge. Figure 2 maps the spatial distribution of active Marcellus gas production wells, existing wastewater disposal options, and depleted oil wells that could prospectively be repurposed for wastewater injection. Details of this geospatial analysis are provided in section S6 of the SI. With respect to existing options, only active brine disposal wells and CWT facilities that currently accept Marcellus wastewater and hold NPDES permits for discharge are shown. The size of each marker for waste disposal sites corresponds to its capacity, taken as the maximum permitted design flow for CWT facilities and the average injection rate for disposal wells (which are typically regulated by injection pressure). For wastewater unsuitable for reuse or recycling, Ohio injection wells are convenient to operators in southwestern Pennsylvania but disposal options in the Northeast are more limited. While CWT facilities are fairly accessible, they currently treat little fracking waste and expanding or retrofitting existing structures sufficiently to treat high-salinity, trace metal-bearing wastewaters to strict effluent standards is both costly and time-intensive.
PA has over 3000 depleted oil wells classified as 'plugged or abandoned', which means they have either been plugged by the PADEP or inspected with the PADEP maintaining the right to complete the well closure in the future (see SI, section S1.6). Most of these wells are located in the southwest to northwest regions of the state and are more accessible to Marcellus production activity than wells in Ohio or scattered CWT facilities in PA (figure 2). To streamline the permitting process and minimize costs, suitable fields of depleted wells could be clustered into centralized 'hubs' that would receive wastewater loads and distribute them to repurposed wellheads. Depleted conventional and unconventional gas fields could also be viable wastewater disposal repositories, particularly as production wells from the Marcellus boom are phased out in the coming decades. Selecting high-capacity formations that optimize the spatial distribution of injection and active production wells will further reduce disposal costs.

Operational cost comparison
To determine the economic viability of converting depleted production wells to wastewater injection wells, we compared high-level cost estimates with the leading disposal alternatives: reuse or recycling at a subsequent site; treatment for discharge at a permitted centralized treatment facility; or injection into dedicated brine disposal wells in Ohio. Since permitting and regulatory overhead costs are strongly site-dependent and not readily quantifiable, only major operational costs are considered: (a) transportation from wellheads to the nearest facility in each management category and (b) disposal via treatment or injection. Low, average, and high estimates are provided for each disposal option to account for ranges in transport distances and available cost data. Wastewater was assumed to be trucked between frac and disposal sites and transport distances were estimated through closest facility analyses using ArcGIS with the Network Analyst extension (see SI, section S7). The results are also separated into the two distinct production regions in the Northeast and Southwest because the most costefficient approach is strongly dependent on well location (e.g. trucking wastewater to Ohio is more favorable closer to the border). Details of this cost analysis are provided in section S7 of the SI and the results are summarized in figure 3 below. Note that these results are averaged over each region to provide a sense for cost-competitiveness but the cheapest options for a given individual wellhead will naturally vary by its proximity to disposal sites. Neglecting reuse, which will become logistically constrained with declining drilling activity, our calculations indicate that wastewater disposal in repurposed oil wells would be operationally competitive with (and likely cheaper than) treatment or injection in Ohio. While the average cost of wastewater treatment is double that of subsurface injection, transportation distances dictate the most efficient disposal pathways in each region. Among existing options, injection in Ohio is cheaper than centralized treatment in the Southwest but more expensive in the Northeast where trucking distances are much greater. In both regions, converted wells are most cost-effective because they are abundant in active oil and gas production regions so average trucking distances are generally shorter. In-state injection is also slightly cheaper at the wellhead because Ohio levies a $0.20/bbl fee on out-of-state loads.
While capital costs are not quantified here, repurposed oil wells likely require fewer upfront investments than competing disposal pathways. Converted wells may need remedial cementing to prevent leakage through the casing and meet EPA requirements for brine (class II) injection wells but would avoid the more intensive capital costs of well or treatment facility construction. Targeting wells that have already been inspected and plugged by the PADEP should also reduce remediation and site assessment costs. Moving forward, in-state injection will be more attractive than out-of-state injection as PA operators compete with rising demands for wastewater disposal wells in Ohio, especially as drilling and production activity expands in the Utica shale. Building additional CWT capacity would also be economically challenging because treatment companies generally desire long-term contracts before investing in new facilities, which producers hesitate to enter into given the instabilities associated with profit in the oil and gas industry [22]. The risk of stranded assets when well completions and concomitant wastewater production slow may also be a deterrent from investments in retrofitting or constructing CWT facilities that are exclusively designed to treat high-volume, low-quality brines associated with shale gas production.

Environmental implications
Management and disposal of wastewater remains a critical environmental concern in shale resource production. In addition to high salinity (up to 300 000 ppm TDS) [3], Marcellus wastewater often contains metals (e.g. barium, strontium) and naturally occurring radioactive material (e.g. radium) that were leached from the shale [23,24]. Concentrations of these elements tend to increase over the lifetime of a well, indicating stricter monitoring may be necessary with increasing volumes of produced fluids as the field ages [25]. Elements in the host rock and brine can also react with frac fluid additives to form other aqueous compounds or precipitates (e.g. scaling minerals such as barite and celestite) that may be difficult to anticipate or predict [26]. Despite an increasing stringency of effluent standards targeting such contaminants, treated wastewater may still contain dissolved organics, trace metals, or radionuclides that are then released to surface waters, which are a significant source of public water supplies, profitable fisheries, and industrial cooling water in the Marcellus region [25]. This can be particularly problematic in dry seasons, when some waterways can become dominated by fracking wastewater discharge and thus would not be able to reduce concentrations through dilution [9]. Release of treated fracking wastewater has also been linked with accumulation of radium isotopes in sediments along streambeds, which creates longterm ecological and health hazards [23].
Direct wastewater injection avoids the risk of impairing surface waters but technically results in net consumption of injected frac fluids that were initially sourced from surface water. Water availability is less of an issue in the Marcellus than in arid shale plays such as the Eagle Ford and Permian, where frac fluid sources are limited but nearly all wastewater from oil and gas production is injected [27]. In many water-stressed regions of the western US where baseline water withdrawals across basins exceed 40% of annual replenishment rates, treatment for recycling or release back into the watershed may be prioritized over injection to support water conservation efforts. However, injection remains the most economically and logistically favorable disposal option in western plays where reservoir geology is also conducive to wastewater storage, and expanding subsurface disposal in the Marcellus by leveraging existing infrastructure can serve as a bridge until more efficient means of beneficial reuse or enhanced treatment technologies become economically viable. Furthermore, reinjecting produced brines that originated from the formation is inherently closed loop and not necessarily wasteful as these low-quality fluids have limited value.
While underground disposal also avoids the risk of impairing surface waters, the real and perceived risks of induced seismicity are one of the greatest barriers to expanding wastewater injection [20,28]. The recent rise of seismic activity connected to wastewater disposal in Oklahoma, where earthquakes have caused injury and damage to infrastructure, has particularly stigmatized shale oil and gas production [29]. Earthquakes typically occur when an increase in pore pressure from fluid injection triggers critically-stressed faults [28]. However, depleted hydrocarbon wells likely pose lower risks than freshly drilled dedicated injection wells because seismic events are more probable during injection into undisturbed formations due to greater pore pressure buildup [30]. Thus, risks associated with induced seismicity can be minimized by limiting injection rates and volumes such that reservoir pressures never exceed initial (i.e. pre-production) levels. Distributing volumes across multiple wells, which would be less of an added investment in depleted formations where wells are already abundantly spaced across aquifers, can further reduce pressurization [28]. Risks should also be managed through pre-development seismic characterization to inform injection locations and depths that avoid fault zones followed by real-time seismic monitoring while the disposal wells are in operation [15].

Social implications
The current under-utilization of depleted fields for wastewater storage has not been for lack of viability or interest from industry. Unfavorable geology is commonly cited as the reason for a dearth of injection wells in PA [6], but permitting and economics are the primary limiting factors. While repurposing depleted oil and gas reservoirs would avoid the high capital costs of well drilling, plugging adjacent wells in an area of review can be cost-prohibitive, and PA's extensive permitting process disincentizes applications. In recent years, operators seeking new in-state injection wells have faced public backlash culminating in legal battles that serve as further deterrents. For example, the Pennsylvania General Energy (PGE)'s permit to convert one of their gas production wells into a wastewater injection well was approved by the EPA and subsequently by the PADEP in October 2015. A month later, the DEP rescinded approval following appeals from a concerned citizen group [31,32]. While PGE was able to reapply and was ultimately reapproved after the PADEP re-evaluated and clarified its review criteria [33], lengthy legal battles impose significant time and financial investments that most companies cannot afford when more accessible disposal options are still available. Additionally, the depleted fields most amenable to reinjection are also amenable for natural gas storage, which may create competition for underground resources. Realizing the economic and environmental benefits of this management strategy will require coordinated efforts among regulators and industry stakeholders to streamline the site selection and permitting process.
Despite broad public aversion to subsurface fluid injection, wastewater disposal in abandoned oil fields could provide several social benefits over existing options. Depleted wells are generally located in regions that already have significant industrial activity, which could mitigate public backlash over creating new disposal sites that are generally perceived as undesirable, unsafe, and detrimental to surrounding property values. More notably, this strategy significantly reduces wastewater transport distances, providing distinct public benefits in addition to reducing associated costs. Trucking is often the most visible and socially objectionable aspect of shale gas production but also poses health risks, including spillage, emissions of particulate matter, and increased risk of motor vehicle accidents [34]. Improved communication with surrounding communities regarding activities and impacts associated with shale resource development is critical to attain public acceptance of technologies like wastewater injection that may be low-risk and lowcost but are stigmatized by media dramatization and misinformation.

Conclusions
Wastewater management is an underrated challenge associated with the continued development of shale gas resources. Leveraging depleted oil wells for Marcellus wastewater injection could alleviate the costs and environmental risks associated with expanding wastewater treatment capacity or increasing transport to Ohio, but will require strategic planning along with external financing that is beyond the scope of this work. In a broader context, this study highlights the need for regional-scale wastewater management planning tailored to the specific challenges and opportunities in a given shale play. Developing long-term strategies with provisions for infrastructure development can prevent inadvertent environmental consequences, which have historically occurred because waste management was considered an afterthought and adverse impacts of the most logistically straightforward approaches were only exposed retroactively. Optimal wastewater disposal pathways will depend on several location-specific factors, including subsurface properties that drive fluid composition and transport; the accessibility and capacity of disposal options at the surface; and constraints on water availability that may prioritize conservation and recycling over economics. Reevaluating current disposal options and capacities in light of future demands is necessary to ensure that sufficient infrastructure is in place to control the inevitable growth in wastewater generation associated with continued shale gas production.