Characterization and Analysis of Liquid Waste from Marcellus Shale

Jul 3, 2015 - Lyndon B. Johnson School of Public Affairs, University of Texas at Austin, P.O. Box Y, Austin, Texas 78713, United States. Environ. Sci...
1 downloads 16 Views 2MB Size
Download PDF
Policy Analysis pubs.acs.org/est

Characterization and Analysis of Liquid Waste from Marcellus Shale Gas Development Jhih-Shyang Shih,† James E. Saiers,‡ Shimon C. Anisfeld,‡ Ziyan Chu,† Lucija A. Muehlenbachs,†,§ and Sheila M. Olmstead*,†,∥ †

Resources for the Future, 1616 P Street NW, Washington, DC 20036, United States School of Forestry and Environmental Studies, Yale University, 195 Prospect Street, New Haven, Connecticut 06511, United States § Department of Economics, University of Calgary, 2500 University Dr. NW, Calgary, Alberta Canada T2N 1N4 ∥ Lyndon B. Johnson School of Public Affairs, University of Texas at Austin, P.O. Box Y, Austin, Texas 78713, United States ‡

S Supporting Information *

ABSTRACT: Hydraulic fracturing of shale for gas production in Pennsylvania generates large quantities of wastewater, the composition of which has been inadequately characterized. We compiled a unique data set from state-required wastewater generator reports filed in 2009−2011. The resulting data set, comprising 160 samples of flowback, produced water, and drilling wastes, analyzed for 84 different chemicals, is the most comprehensive available to date for Marcellus Shale wastewater. We analyzed the data set using the Kaplan− Meier method to deal with the high prevalence of nondetects for some analytes, and compared wastewater characteristics with permitted effluent limits and ambient monitoring limits and capacity. Major-ion concentrations suggested that most wastewater samples originated from dilution of brines, although some of our samples were more concentrated than any Marcellus brines previously reported. One problematic aspect of this wastewater was the very high concentrations of soluble constituents such as chloride, which are poorly removed by wastewater treatment plants; the vast majority of samples exceeded relevant water quality thresholds, generally by 2−3 orders of magnitude. We also examine the capacity of regional regulatory monitoring to assess and control these risks.

1. INTRODUCTION

The effectiveness of regional wastewater treatment processes in removing flowback and produced water constituents is limited, with dissolved solids of particular concern.6 Unconventional gas wells in the Marcellus increased total generation of oil-and-gas-related wastewater in the region by about 570% between 2004 and 2011.7 Surface water quality impacts from incomplete wastewater treatment have been demonstrated for chloride,8 bromide,9 and radionuclides.10,11 Road-spreading of shale gas wastewater is associated with accumulation of radionuclides in surface sediment.12 Partially treated shale gas waste in surface water also poses risks to finished drinking water from disinfection byproducts.13,14 PA state regulations regarding disposal of Marcellus wastewater since 2011 have changed.8,12,13 In 2011, PA banned shipments of this wastewater to municipal sewage treatment plants unless the waste is pretreated to new state effluent standards for total dissolved solids (TDS), chloride (Cl), barium (Ba), and strontium (Sr).15,16 These standards have also

The extraction of natural gas from shale formations is increasing rapidly, due to advances in hydraulic fracturing, horizontal drilling, and seismic imaging.1 Hydraulic fracturing injects water, sand, and chemical additives into a wellbore at very high pressure to fracture the shale and prop open pathways for natural gas to flow out of the well. Water quantities used in hydraulic fracturing vary with geology, the amount of recoverable gas, and other factors; wells in the Marcellus Shale require 7.6−15.1 million liters (ML) each.2 In December 2011, the end of our study period, 4908 unconventional wells had been drilled in Pennsylvania (PA); the well count in June 2015 was 9240.3 A significant but highly variable fraction (10−40% in the Marcellus) of water used returns to the wellhead as “flowback” within one month after fracturing.4 Later, gas flowing to the wellhead is accompanied by formation water (“produced water”), which emerges at a much lower flow rate for the life of the well. In PA through 2011, natural gas flowback and produced water are partially treated for recycling, shipped to underground injection wells in PA or Ohio, or treated and released to surface water by permitted wastewater treatment facilities.5 © 2015 American Chemical Society

Received: Revised: Accepted: Published: 9557

April 8, 2015 June 15, 2015 July 3, 2015 July 3, 2015 DOI: 10.1021/acs.est.5b01780 Environ. Sci. Technol. 2015, 49, 9557−9565

Policy Analysis

Environmental Science & Technology

scanned available Form 26Rs to portable document format (PDF), converted them to spreadsheets (using ABBY FineReader version 11), and read them into a statistical software program for data cleaning, quality assurance and quality control (QA/QC), and analysis. We focused on liquid waste types corresponding to four PADEP categories (Supporting Information text S2): 801drilling fluids and residuals; 802brine (produced water); 803drilling fluid waste; and 804fracturing fluid waste (fracturing fluid, flowback, and proppants). The Supporting Information (SI text S2) describes QA/QC procedures in detail. An important step was the calculation of a charge balance for each sample. Out of 211 samples, 51 (24%) failed the charge balance test, many by large amounts (Figure 1), and were excluded from further analysis. This indicated

essentially eliminated treatment of Marcellus waste at industrial centralized waste treatment (CWT) facilities. However, wastewaters from hydraulically fractured wells in tight sandstone and other unconventional and conventional formations in PA continue to be partially treated at a small number of CWTs. Partial treatment and release to rivers and streams of these high-TDS wastewaters raises concerns similar to those of Marcellus waste,7,8 though their volume is much smaller. Potential spills of untreated effluent remain an additional environmental risk; spill rates have increased as incomplete treatment of Marcellus waste by CWTs has fallen.17 Most flowback from Marcellus wells is now recycled for new well completions.18 However, over time, with more wells in production and fewer being drilled, regional reuse capacity will decline, while recycling of produced water (typically higher in TDS) will remain challenging.19 These concerns are particularly relevant to the Marcellus Shale because (a) its formation water is unusually salty and radiogenic compared to other shale plays;19 and (b) opportunities for deep injection and evaporation are limited.7,10 Despite environmental concerns and a dynamic regulatory framework, Marcellus flowback and produced water have been infrequently characterized.20 Information on the chemical composition of additives used in hydraulic fracturing fluid is available through an online chemical disclosure registry, FracFocus (www.fracfocus.org), but public information on flowback and produced water constituents is sparse. The existing literature analyzes data from a small number of wells21−23 or reports summary statistics obtained from oil and gas operators;24 recent studies provide a few important exceptions.25−28 We analyze unique publicly available data on characteristics of wastewater from Marcellus Shale gas production in PA from 2009 to 2011. The data are the most comprehensive analyzed, in terms of waste types, water-quality parameters examined, and the number and spatial distribution of sources. Our analysis contributes sophisticated statistical treatment of analyte concentrations below laboratory detection limits. We also assess the dilution factors required to treat sample wastes to regulatory standards, and the degree to which current regional surface water quality monitoring is able to detect impacts from waste treatment and disposal. We also report on the quality of the laboratory processes employed in firms’ reporting of the data to state regulators.

Figure 1. Charge balance for original data set, showing acceptable samples (open symbols, n = 143) and unacceptable samples (solid gray symbols, n = 51). Not shown are 17 samples that did not have major ion concentrations reported but were still considered acceptable for the analytes that were reported. Line shown is 1:1 line.

potentially serious analytical problems with the Form 26R laboratory reports. The most common problem appeared to be erroneously high chloride (Cl) concentrations, perhaps because brine samples were outside the calibration range of standard methods at the receiving laboratories. The final data set comprised 160 samples in five waste categories (39 samples of drilling waste (categories 801 and 803); 58 samples of produced water (802); 61 samples of flowback (804); and 2 samples labeled “supply water”), transported to 32 treatment and/or disposal locations. In terms of spatial distribution, 9 samples were from the PADEP Northwest regional office, 58 from North Central, 3 from Northeast, and 90 from Southwest (Supporting Information, Figure S1). We identified the precise location of the 100 gas wells that were the source of the wastewater for 118 of the samples. Data for 418 different chemical analytes were available. We excluded analytes that had: (a) more than 75% of the reported values below detection limits (BDL) or (b) fewer than 5 detected values reported; this left 84 analytes in the final data set.

2. MATERIALS AND METHODS 2.1. Data Collection. Chemical analyses of shale gas wastewater were collected from the Pennsylvania Department of Environmental Protection (PADEP), representing samples drawn from Marcellus Shale wells from 2009 to 2011. In PA, firms that generate more than 2200 pounds of residual nonhazardous industrial waste from a single generating location in a single month must file an annual residual waste report, known as Form 26R. These reports include generator information, a waste description (including a chemical analysis performed by a state-certified laboratory), and the name and location of the receiving waste processing or disposal facilities. Form 26R is filed in hard copy with the relevant regional office of the PADEP. With the exception of a USGS study of the radium (Ra) content in 23 examples of Form 26R,21 these reports have not been analyzed in the scientific literature. Researchers traveled to the four regional PADEP offices with significant shale gas development during Fall/Winter 2011, 9558

DOI: 10.1021/acs.est.5b01780 Environ. Sci. Technol. 2015, 49, 9557−9565

Policy Analysis

Environmental Science & Technology Table 1. Summary Statistics for Selected Marcellus Shale Liquid Waste Constituents constituent

units

n

% BDLa

minb

medianc

max

NPDES facilities with permit limitse

PA STORET monitors (2000/2011)f

Major Anions Br Cl SO4

mg/L mg/L mg/L

142 144 145

8% 0% 50%

0.24 18 1.0

450 51 000 11

3300 200 000 1700

1 3 1

0/949 178/1213 0/0

General TDS EC

mg/L umhos/cm

145 143

0% 0%

2.8 64

88 000 110 000

390 000 480 000

2 0

52/1213 0/0

Major Cations Al, total Ba, total Ca, total Li, total Mg, total K, total Na, total Sr, total

mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L

142 149 146 139 146 33 142 147

37% 0% 0% 14% 3% 0% 0% 0%

0.010 0.061 16 0.018 0.25 0.074 8.0 0.063

0.30 430 6200 44 570 210 22 000 1200

860 12 000 40 000 630 3700 5000 82 000 7900

2 2 0 1 1 0 1 2

0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0

NORM gross alpha gross beta 226 Ra 228 Ra 234 U 235 U 238 U

pCi/L pCi/L pCi/L pCi/L pCi/L pCi/L pCi/L

104 104 95 94 9 19 19

4% 1% 2% 4% 0% 5% 5%

0 0 0.067 0 0.089 0 0

3200 1300 890 39 1.3 0.0000 0.67

41 000 220 000 17 000 2600 22 40 170

0 0 0 0 0 0 0

0/72 0/72 0/0 0/0 0/0 0/0 0/0

mg/L mg/L mg/L

146 138 141

72% 31% 45%

0.0010 0 3.0

NDd 0.20 5.4

1.3 240 1500

1 1 3

0/0 0/0 0/0

mg/L

19

74%

0.0047

0.021

0.11

0

0/0

mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L

146 145 147 122 142 138 146 142

60% 72% 0% 7% 4% 72% 62% 44%

0.00084 0.0065 0.073 0 0.010 0.0068 0.0050 0.0025

0.011 0.027 47 33 3.6 0.017 0.022 0.080

2.2 18 1400 350 73 2.0 3.2 17

1 3 3 2 1 1 2 2

0/0 0/0 0/0 0/0 0/0 0/0 0/0 0/0

Organics benzene MBAS oil and grease total (HEM) xylene Transition Metals Cr, total Cu, total Fe, total Fe, dissolved Mn, total Mo, total Ni, total Zn, total a

Fraction of samples that were below detection limit (BDL). bMinimum detected concentration. cAs calculated by the Kaplan−Meier method. dND: median concentration could not be estimated owing to large fraction of BDL values. eNumber of NPDES permitted facilities facing an effluent limit for that particular constituent (out of 8 facilities for which effluent limit data were available). fNumber of PA STORET water quality monitors (in 2000/2011) reporting observed concentrations of each contaminant.

Ciflip = M − Ci

2.2. Data Analysis. Out of the 84 analytes, 52 contained left-censored observations (those with concentrations BDL). To accurately compute the empirical distribution function of concentration for each analyteincluding those with up to 75% BDLwe employed the nonparametric Kaplan−Meier (KM) method (implemented in MATLAB), which compares favorably to other methods for left-censored water quality data.29,30 The KM method accepts right-censored data, which we constructed by subtracting observations, Ci, of left-censored data sets from a flipping constant, M, according to the approach described by Helsel (2012):31

(1)

The empirical cumulative density of the flipped observations, F(Cflip i ), was computed with the KM method, whereupon the empirical distribution of the original concentrations, F(Ci), was determined by the following retransformation: F(Ci) = 1 − F(Ciflip)

(2)

We report the 10th, 25th, 50th (median), 75th, and 90thpercentile concentrations from the empirical distributions. The complete F(C) could not be determined for 11 analytes for 9559

DOI: 10.1021/acs.est.5b01780 Environ. Sci. Technol. 2015, 49, 9557−9565

Policy Analysis

Environmental Science & Technology

CWTs), data on effluent limits for at least one year (2009− 2011) were available in ICIS for 8 facilities. The remaining 2 (“minor” facilities with design flows smaller than 1 million gallons−about 3.8 mL per day) had no data available. We also compared the list of wastewater constituents to the list of water quality parameters tracked by surface water monitoring systems contributing to EPA’s Storage and Retrieval (STORET) database, the largest public repository of such information.

which the fraction of censored observations was greater than 0.25 and the minimum detected concentration equaled or exceeded the minimum detection limit. These conditions precluded estimation of (i) the 10th-percentile concentration for six analytes, (ii) the 10th- and 25th-percentile concentrations for four analytes; and (iii) the 10th-, 25th-, and 50thpercentile concentrations for one analyte. For each analyte, empirical concentration distributions were estimated for the entire data set and for the following subsets: (a) samples grouped by waste type and (b) samples sent to wastewater treatment facilities permitted to release treated waste to rivers and streams. To identify subset (b), we parsed the list of 32 treatment and disposal locations. Of these, 9 were for “dust suppression” in PA townships and boroughs; 5 were injection wells in eastern Ohio; 1 was a facility injecting waste into an abandoned PA coal mine; and 17 were waste transport, treatment, and/or disposal facilities. Of the latter category, 7 had no identifiable National Pollutant Discharge Elimination System (NPDES) permit number, and were presumably not releasing treated waste to rivers and streams. The remaining 10 did have NPDES permit numbers, and comprised subset (b) above. To help assess the geochemical origin of the samples, we plotted major ion concentrations alongside samples from the literature: the seawater evaporation curve;32 in situ brine samples from the Marcellus reported by Osborn and McIntosh (2010);33 flowback samples from the PA Bureau of Oil and Gas Management (BOGM), reported by Haluszczak et al. (2013);27 conventional oil and gas well brines from PA reported by Dresel and Rose (2010);34 and Marcellus flowback samples analyzed in Hayes (2009) and Llewellyn (2014).35,36 To compare median concentrations of analytes between flowback and produced waters, the Kruskal−Wallis test was used. The null hypothesis was that all samples were drawn from the same population (or equivalently, from different populations with the same distribution). We adopted the critical p-value of 0.05 and treated censored data by setting all observations below the highest detection limit to the same value.31 To elucidate the human and ecological health risks, we compared our wastewater pollutant concentrations to PA effluent standards (for Ba, Cl, Fe, oil and grease, Sr, and TDS) applicable to wastewater treatment plants receiving Marcellus wastes and to Nuclear Regulatory Commission (NRC) standards for radionuclides in effluent.37 For each sample, the pollutant that was present in highest concentration relative to the standard was identified as the limiting pollutant. We then calculated the ratio of the sample concentration (of the limiting pollutant) to the effluent standard; this represents the dilution or removal factor required before discharge in order to meet effluent standards. We repeated this exercise using PA ambient water quality criteria (for acetone, Ag, As, Ba, benzene, Cd, Cr, Cu, Hg, Ni, Pb, Sb, Se, Sr, toluene, and Zn). In this case, the ratio that we report (calculated as above) represents the dilution factor that would be required before achieving acceptable instream water quality (assuming there are no other sources of these pollutants). Lastly, we used the treating facilities’ NPDES permit numbers to obtain information from the U.S. EPA’s Integrated Compliance Information System (ICIS) database, via the Discharge Monitoring Report Pollutant Loading Tool. Among these 10 facilities (3 municipal sewage treatment plants and 7

3. RESULTS AND DISCUSSION 3.1. Liquid Waste Characterization and Comparison of Waste Types. Cl, Na, Ca, Sr, Ba, Mg, and Br were the primary constituent ions in the samples (Table 1), with typical median concentrations of hundreds to 10s-of-thousands of mg/ L. (See Supporting Information, Table S2, for summary statistics for all analytes). TDS concentrations ranged from less than 10 mg/L to nearly 400 000 mg/L, or 10 times that of typical seawater, with a median concentration of 88 000 mg/L. The low observed sulfate (SO4) concentrationin contrast to the elevated concentrations of most major ionsis typical of brines with high Ba concentrations and reflects the low solubility of BaSO4. A log(Br)−log(Cl) plot (Figure 2) illustrates the major-ion relationships in our data set. Most of our samples followed the

Figure 2. Log chloride concentrations (mg/L) plotted against log bromide concentrations (mg/L) for our data set (filled symbols; grouped by waste type). Also shown for reference are the seawater evaporation curve (black line; McCaffrey, 198732); Marcellus wastewater samples (open circles from Bureau of Oil and Gas Management (BOGM) in Haluszczak et al., 201327 and open triangles from Hayes, 200935); Western PA conventional oil and gas brine (open circles; from Dresel and Rose, 201034); and Marcellus in situ brines (pink X’s; Osborn and McIntosh, 201033). Not shown are samples of waste type 803 (4 samples) and 7 low-concentration samples (6 from our data set, 1 from BOGM).

pattern previously observed,24,38 in which fracking waste appears to be derived from the dilution of highly evaporated seawater brines. This is indicated by samples with lower Cl:Br ratios than seawater, reflecting a history of (a) brine formation with preferential loss of Cl during halite precipitation (the flat 9560

DOI: 10.1021/acs.est.5b01780 Environ. Sci. Technol. 2015, 49, 9557−9565

Policy Analysis

Environmental Science & Technology

Figure 3. Distributions of selected constituent concentrations in flowback and produced water from PA Marcellus Shale Gas Wells, 2009−2011, as calculated by the Kaplan−Meier method.

concentrated as the in situ Marcellus brines reported by Osborn and McIntosh.33 Samples from drilling fluid and flowback (categories 801 and 804, respectively), although generally more dilute than produced water (802), largely followed the dominant brinedilution line, indicating that major ion concentrations were determined by formation waters, rather than by the chemistry of drilling or fracking fluid (with the exception of a handful of dilute samples whose Cl:Br ratio diverged from the pattern). Statistical comparison of flowback and produced waters confirmed these observed differences in major ion concentrations (Figure 3). Median concentrations in produced water were significantly higher than in flowback for 7 major ions, as well as for TDS, Mn, and NORMs (Figure 3). The median combined concentration of Ra-226/228, the dominant naturally occurring radioactive material (NORM) in the Marcellus, was 929 pCi/L. Though fewer samples were tested for other radionuclides, tested samples indicated the presence of radioisotopes of uranium (U), thorium (Th), strontium (Sr), and lead (Pb) (Supporting Information, Table S2). Median gross alpha radiation was 3200 pCi/L, and gross beta was 1300 pCi/L. Mean values of Ra-226/228 activity (1700 pCi/L), gross alpha radiation (5,300 pCi/L), and gross beta radiation (6000 pCi/L) were greater than median values,

portion of the seawater evaporation line), followed by (b) dilution of the brine with freshwater or drilling fluid (movement along a diagonal line downward and to the left). Thus, the majority of our samples plot along a dilution line originating in the flat part of the seawater evaporation curve, as do most previously examined Marcellus fracking samples (e.g., BOGM27 and Hayes35 samples) and PA conventional oil and gas waste samples (e.g., Dresel and Rose34 samples, Figure 2). Of note, some of our samples were more concentrated than previously reported fracking waste samples. A total of 22 samples (all classified as waste type 802−produced water) had Br concentrations higher than the highest values reported in previous studies.27,33,35 In addition, our most concentrated samples diverged from the brine-dilution pattern generally seen for unconventional (and conventional) wastewater samples. While previously reported samples intersected the flat portion of the seawater evaporation curve, our highest-Br samples remained below the curve, indicating an even lower Cl concentration than expected from seawater evaporation. These samples are most likely derived from dilution of an even more concentrated brine, with log(Br) ≥3.6 and a seawater evaporation factor of ≥70.32 This hypothetical source brine would have to be about three times as 9561

DOI: 10.1021/acs.est.5b01780 Environ. Sci. Technol. 2015, 49, 9557−9565

Policy Analysis

Environmental Science & Technology

Table 3. Limiting Pollutants Based on Ambient Standardsa

indicating a positive skew in the empirical distributions. Compared to means reported by Barbot et al.,25 our mean estimates of Ra-226/228 and gross-alpha radiation were twoand 4-fold greater, respectively, while our mean estimate of gross-beta radiation was 7-fold smaller. Median NORM concentrations, especially Ra-226, were higher in produced water than in flowback (Figure 3). Form 26R requires reporting of only a small number of organic pollutants (Supporting Information, Table S1), though firms must test for additional constituents expected or known to be present in their wastewater, resulting in a long list of organic analytes reported on individual forms. The fraction BDL was high for each of these (Table 1, Table S2), likely because: (a) fracking fluid ingredients vary widely by operator and well; and (b) additives are likely used at concentrations that are low relative to typical detection limits (though not necessarily relative to toxicity thresholds). Two groupings of organic pollutants (methylene-blue-active substances (MBAS) and oil and grease) were detected in more than half of the samples. These may originate from in situ petroleum materials (including naturally occurring surfactants) or from fracking additives.39 The presence of transition metals in shale gas waste has received little attention in the literature, though their potential for accumulation in soils and sediments is significant.20 With the exception of iron (Fe) and manganese (Mn), which were both present in essentially all of the liquid waste samples, the remaining transition metals were frequently BDL. Concentration distributions of transition metals (as estimated by the KM method) were generally similar between waste types. 3.2. Treatment and/or Dilution Required to Meet Relevant Water Quality Standards. Depending on the fate of the wastewater, comparisons to both effluent standards (Table 2) and ambient standards (Table 3) may be relevant to

n

median ratio

max ratio

noneb Cl Ba TDS Sr Ra226 oil and grease

11 82 34 21 9 2 1

NA 198 529 67 387 234 14

NA 800 1150 789 789 282 14

n

median ratio

max ratio

noneb Ba Sr benzene Cu Pb Cr Hg Zn Cd As

8 76 43 14 5 5 3 3 1 1 1

NA 788 185 208 253 360 17 13 117 10 21

NA 4792 1973 1121 2000 3472 22 16 117 10 21

a

Ratios refer to the ratio of sample concentration (of the limiting pollutant) to the value of the ambient standard. bEight samples met ambient standards for all pollutants.

samples with Cl or TDS as the limiting pollutants would require dilution by up to 800-fold, with a median of 158-fold. Comparison to ambient standards (Table 3) showed somewhat different results. Ba was the most common limiting pollutant, followed by Sr, benzene, and a number of other metals. Sample:standard concentration ratios ranged up to 4792, with an overall median of 525. 3.3. Characterization of Liquid Waste Treated by Active NPDES-Permitted Facilities. According to waste shipment data reported to the PADEP by unconventional gas operators,40 recycling and reuse of flowback have increased over time (60% of liquid waste was recycled and reused in July 2014 to March 2015, compared to 3% in 2009). However, according to these operator-reported data, NPDES discharge from CWTs still remains a disposal mechanism for unconventional gas waste (4.5% (more than 159 ML) from July 2014 to March 2015, compared to 81% in 2009). Some CWTs may not sufficiently remove halides from their discharged effluent,13 thus it is useful to characterize the wastes routed to NPDES-permitted facilities. A subset of 88 samples were treated by 10 NPDES-permitted facilities. We repeat the KM analysis for these samples, with full results available in the Supporting Information (Table S4). Given the smaller number of samples and left-censoring, the list of analytes for which medians could be calculated was smaller. However, because a firm’s choice of disposal method might be driven by wastewater chemical composition, the smaller sample may be more representative of the waste stream treated and released by CWTs. For 2 of the 32 analytes in Table 1 (234U and benzene), the NPDES-only sample was too small (and/or the fraction BDL too high) to estimate a median using the KM method. Of the remaining analytes, Ba had a median concentration more than 2.5 times higher in the active NPDES samples than in the full set of samples, the lithium (Li) median concentration was about 10% higher, Cl was about 12% higher, and TDS was about 14% higher. The observed higher concentrations of TDS and Ba may be due to regional differences in Ba concentrations, higher in the Marcellus formation underlying north-central PA than in southwestern PA. Liquid wastes from the southwestern region are more likely to be transported to deep injection wells in Ohio, with less treatment by CWTs, so the NPDES-only sample may be weighted toward samples from wells in the north-central part of PA. With the exception of Ba, Li, Cl, and TDS, most analytes had median concentrations in the NPDES-only sample that were

Table 2. Limiting Pollutants Based on Effluent Standardsa limiting pollutant

limiting pollutant

a

Ratios refer to the ratio of sample concentration (of the limiting pollutant) to the value of the effluent standard. bEleven samples met effluent standards for all pollutants.

the analyte concentrations described in Section 3.1. Ambient standards, of course, are more stringent, given that effluent is expected to mix with cleaner water before such standards are met. The vast majority of our samples exceeded both of these types of thresholds, generally by 2−3 orders of magnitude (Tables 2 and 3). Comparison of our samples to effluent standards (Table 2) revealed that Cl was the most common limiting pollutant (82 out of 160 samples), with Ba, TDS, and Sr also important for a subset of samples. Sample:standard concentration ratios ranged up to 1150, with an overall median of 227. Of the pollutants shown in Table 2, Cl and TDS are difficult to remove from wastewater, so their sample:standard ratio can be thought of as the dilution factor required to meet effluent standards. The 103 9562

DOI: 10.1021/acs.est.5b01780 Environ. Sci. Technol. 2015, 49, 9557−9565

Policy Analysis

Environmental Science & Technology

comprises the majority of waste over the life of a well19 and will be generated by producing wells even after the play is fully developed and reuse by additional wells is not feasible. Our data were collected from the only PA regulatory reporting process regarding shale gas wastewater disposal, and yet they suggest serious deficiencies in the required chemical analysis. One-quarter of the samples had charge balances that were in error by more than 25%, indicating failures by statecertified laboratories to accurately measure major ion concentrations. These failures were not exclusively associated with any region, laboratory, or method. If this failure is due to extrapolation of calibration curves to samples of very high ion concentrations, state-certified laboratories should be equipped to take into account the expected high concentrations in these samples. We show that the 2011 PA effluent standards for TDS, Cl, Ba, and Sr would have been binding for 157/160 samples in our data. Thus, since 2011, these standards have potentially played a key role in maintaining surface water quality. The limiting pollutants for meeting ambient standards are different from the limiting pollutants for meeting effluent standards, suggesting that additional effluent standards (e.g., for benzene, Cu, and Pb) may be necessary in order to ensure that natural gas wastewater is appropriately treated by CWTs. In the Marcellus Shale through 2011, POTWs and CWTs were primary treatment and disposal mechanisms for shale gas wastewater. Our analysis suggests that monitoring of key shale gas waste constituents in both wastewater facility effluent permits and ambient surface water was insufficient to assess potentially important impacts on rivers and streams at that time. Ambient water quality monitoring expanded during the study period for five key constituents associated with shale gas development (Br, Cl, TDS, gross alpha, and gross beta). For these pollutants and others, research is needed to determine the optimal spatial distribution of monitors to ensure compliance with state and federal ambient water quality standards. Marcellus Shale wastewater is no longer partially treated by CWTs and released to surface water, and treatment by CWTs occurs infrequently in other U.S. shale plays.41 Accidental releases do occur,20,42 however, and regions that face similar challenges in managing natural gas wastewater can learn from the PA experience.43,44 Our analysis may also be relevant in the ongoing development of federal rules regarding treatment of high-TDS unconventional oil and gas waste by wastewater treatment facilities nationwide.45

lower than or approximately equal to those reported for the full sample of liquid wastes in Table 1. The most significant example was SO4, with a median concentration 91% lower than the full sample. 3.4. Analysis of Regulatory and Monitoring Capacity. Regulatory monitoring of the major constituents present in the Form 26R samples, in terms of both limits in treated effluent, and ambient water quality limits, was sparse through 2011 (Table 1). Regarding ambient water quality, PA ambient water quality monitors in the federal STORET database report concentrations for only five of the 32 major waste constituents in Table 1. Thus, public ambient water quality monitoring during this time may have been insufficient to assess the impacts of PA shale gas development. However, PA’s ambient monitoring capacity for dissolved solids (Cl, Br, TDS) increased significantly by 2011 (Table 1). Twenty monitoring stations sampled for radionuclides in 2011, though there were no reporting stations in 2000, an additional improvement in regional monitoring capacity. Regarding wastewater effluent standards, regional monitoring capacity for Cl, Br, and NORM is of particular interest, since their environmental impacts have been documented.8−11 Cl was limited by NPDES permits in three of eight treating facilities between 2009 and 2011, Br was limited at a single facility, and no receiving facilities faced NPDES limits for NORM. (The full lists of NPDES effluent limits in place for these 8 facilities are reported in the Supporting Information (Table S6)). When NPDES permits are reissued every five years, the PADEP now establishes monitoring requirements for TDS, Cl, Ba, Sr, and NORM for facilities treating natural gas waste, and has done so since 2011, even if facilities are not limited for these pollutants.16 Thus, current state regulatory monitoring capability may be stronger than is indicated by available information on permit limits. However, the data suggest that the primary tool for monitoring wastewater effluent, the NPDES, did not provide information sufficient to assess the water-quality implications of the rapid expansion in PA natural gas waste treatment through 2011. Since 2011, CWTs have been required to meet new PA standards for Cl, TDS, Ba, and Sr, and shipments to municipal sewage treatment plants (POTWs) have stopped due to a voluntary ban, unless the waste is pretreated at a CWT that meets the new standards. Though these changes initially exempted a set of CWTs, since 2011 partial treatment of Marcellus waste has gradually been eliminated. Partial treatment and release to the environment of high-TDS waste from hydraulic fracturing operations in other PA formations (e.g., tight sandstone) continues at a small number of CWTs. These CWTs are under legally binding consent decrees to adopt technologies that will meet the 2011 standards, over time.

■

ASSOCIATED CONTENT

S Supporting Information *

Data collection and QA/QC description, additional supporting tables and figures, and full data set used for the analysis. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.est.5b01780.

4. CONCLUSIONS AND RECOMMENDATIONS The composition of raw shale gas wastewater has important implications for accidental releases20,41 and for assessing local and regional treatment capacity, technology, and regulation. The unique data in this study add to the small literature on concentration distributions of shale gas wastewater constituents. The observed concentrations illustrate the need for significant treatment before release to the environment. The analysis suggests that typical shale gas wastewater may be even more concentrated (in Br, Cl, and NORM) than previously reported. This is particularly true for produced water, which

■

AUTHOR INFORMATION

Corresponding Author

*Phone: 512-471-2064; fax 512-471-4697; e-mail: sheila. [email protected]. Notes

The authors declare no competing financial interest. 9563

DOI: 10.1021/acs.est.5b01780 Environ. Sci. Technol. 2015, 49, 9557−9565

Policy Analysis

Environmental Science & Technology

■

NPDES Permitting of Discharges of Total Dissolved Solids (TDS) 25 Pa. Code §95.10, Document #385-2100-002, 12 November 2011. http://www.elibrary.dep.state.pa.us/dsweb/Get/Document-85967/ 385-2100-002%20tech%20guidance.pdf. (17) Brantley, S. L.; Yoxtheimer, D.; Arjmand, S.; Grieve, P.; Vidic, R.; Pollak, J.; Llewellyn, G.; Abad, J.; Simon, C. Water resource impacts during unconventional shale gas development: the Pennsylvania experience. Int. J. Coal Geol. 2014, 126 (1), 140−156. (18) Maloney, K. O.; Yoxtheimer, D. A. Production and disposal of waste materials from gas and oil extraction from the Marcellus Shale play in Pennsylvania. Environ. Pract. 2012, 14, 278−287. (19) Vidic, R. D.; Brantley, S. L.; Vandenbossche, J. M.; Yoxtheimer, D.; Abad, J. D. Impact of shale gas development on regional water quality. Science. 2013, 340 (6134).123500910.1126/science.1235009 (20) Vengosh, A.; Jackson, R. B.; Warner, N.; Darrah, T. H.; Kondash, A. A critical review of the risks to water resources from unconventional shale gas development and hydraulic fracturing in the United States. Environ. Sci. Technol. 2014, 48, 8334−8348. (21) Rowan, E. L.; Engle, M. A.; Kirby, C. S.; Kraemer, T. F. Radium Content of Oil- And Gas-Field Produced Waters in the Northern Appalachian Basin (USA): Summary and Discussion of Data, USGS Scientific Investigations Report 2011−5135; U.S. Geological Survey: Reston, VA, 2011; http://pubs.usgs.gov/sir/2011/5135/. (22) Blauch, M. E.; Myers, R. R.; Moore, T.; Lipinski, B. A.; Houston, N. A. Marcellus Shale post-frac flowback waters − where is all the salt coming from and what are the implications?, Paper #125740; presented at the Society of Petroleum Engineers Eastern Regional Meeting, Charleston, WV, 23−25 September, 2009. (23) Thacker, J. B.; Carlton, D. D., Jr.; Hildenbrand, Z. L.; Kadjo, A. F.; Schug, K. A. Chemical analysis of wastewater from unconventional drilling operations. Water 2015, 7, 1568−1579. (24) Revised Draft Supplemental Generic Environmental Impact Statement on the Oil, Gas and Solution Mining Regulatory Program; Well Permit Issuance for Horizontal Drilling and High-Volume Hydraulic Fracturing to Develop the Marcellus Shale and Other Low-Permeability Gas Reservoirs; New York State Department of Environmental Conservation: Albany, NY, 2011; http://www.dec.ny.gov/data/dmn/ rdsgeisfull0911.pdf. (25) Barbot, E.; Vidic, N. S.; Gregory, K. B.; Vidic, R. D. Spatial and temporal correlation of water quality parameters of produced waters from Devonian-age shale following hydraulic fracturing. Environ. Sci. Technol. 2013, 47 (6), 2562−2569. (26) Chapman, E. C.; Capo, R. C.; Stewart, B. W.; Kirby, C. S.; Hammack, R. W.; Schroeder, K. T.; Edenborn, H. M. Geochemical and strontium isotope characterization of produced waters from Marcellus Shale natural gas extraction. Environ. Sci. Technol. 2012, 46, 3545−3553. (27) Haluszczak, L. O.; Rose, A. W.; Kump, L. R. Geochemical evaluation of flowback brine from Marcellus gas wells in Pennsylvania, USA. Appl. Geochem. 2013, 28, 55−61. (28) Abdulfaraj, N.; Gurian, P. L.; Olson, M. S. Characterization of Marcellus shale flowback water. Environ. Eng. Sci. 2014, 31 (9), 514− 524. (29) Antweiler, R. C.; Taylor, H. E. Evaluation of statistical treatments of left-censored environmental data using coincident uncensored data sets: I. summary statistics. Environ. Sci. Technol. 2008, 42, 3732−3738. (30) She, N. Analyzing censored water-quality data using a nonparametric approach. J. Am. Water Resour. Assoc. 1997, 33 (3), 615− 624. (31) Helsel, D. R. Statistics for Censored Environmental Data Using Minitab and R, 2nd ed.; John Wiley and Sons: Hoboken, NJ, 2012. (32) McCaffrey, M. A.; Lazar, B.; Holland, H. D. The evaporation path of seawater and the coprecipitation of Br- and K+ with halite. J. Sediment. Res. 1987, 57, 928−937. (33) Osborn, S. G.; McIntosh, J. C. Chemical and isotopic tracers of the contribution of microbial gas in Devonian organic-rich shales and reservoir sandstones, northern Appalachian Basin. Appl. Geochem. 2010, 25 (3), 456−471.

ACKNOWLEDGMENTS This work was supported in part by a grant from the Alfred P. Sloan Foundation. We thank Alan Krupnick for helpful input, Dietrich Earnhart for data on wastewater treatment facility NPDES permits, staff in the PADEP regional offices for assistance with Form 26R data collection, and four referees for excellent comments.

■

REFERENCES

(1) Annual Energy Outlook 2015; DOE/EIA-0383(2015); U.S. Energy Information Administration: Washington, DC, 2015; http:// www.eia.gov/forecasts/aeo/pdf/0383(2015).pdf. (2) Veil, J. Water Management Technologies Used by Marcellus Shale Gas Producers, Final report; DOE Award No. FWP 49462; U.S. Department of Energy, Argonne National Laboratory: Argonne, IL, 2010; www.ipd.anl.gov/anlpubs/2010/07/67463.pdf. (3) Pennsylvania Department of Environmental Protection. SPUD Data Report; http://www.depreportingservices.state.pa.us/ ReportServer/Pages/ReportViewer.aspx?/Oil_Gas/Spud_External_ Data. (4) Gregory, K. B.; Vidic, R. D.; Dzombak, D. A. Water management challenges associated with the production of shale gas by hydraulic fracturing. Elements 2011, 7, 181−186. (5) Rahm, B. G.; Bates, J. T.; Bertoia, L. R.; Galford, A. E.; Yoxtheimer, D. A.; Riha, S. J. Wastewater management and Marcellus Shale gas development: trends, drivers, and planning implications. J. Environ. Manage. 2013, 120, 105−113. (6) Ferrar, K. J.; Michanowicz, D. R.; Christen, C. L.; Mulcahy, N.; Malone, S. L.; Sharma, R. K. Assessment of effluent contaminants from three facilities discharging Marcellus Shale wastewater to surface waters in Pennsylvania. Environ. Sci. Technol. 2013, 47, 3472−3481. (7) Lutz, B. D.; Lewis, A. N.; Doyle, M. W. Generation, transport, and disposal of wastewater associated with Marcellus Shale gas development. Water Resour. Res. 2013, 49, 647−656. (8) Olmstead, S. M.; Muehlenbachs, L. A.; Shih, J.-S.; Chu, Z.; Krupnick, A. Shale gas development impacts on surface water quality in Pennsylvania. Proc. Natl. Acad. Sci. U. S. A. 2013, 110 (13), 4962− 4967. (9) Wilson, J. M.; VanBriesen, J. M. Oil and gas produced water management and surface drinking water sources in Pennsylvania. Environ. Pract. 2012, 14 (4), 288−300. (10) Zhang, T.; Gregory, K.; Hammack, R. W.; Vidic, R. D. Coprecipitation of radium with barium and strontium sulfate and its impact on the fate of radium during treatment of produced water from unconventional gas extraction. Environ. Sci. Technol. 2014, 48, 4596− 4603. (11) Warner, N. R.; Christie, C. A.; Jackson, R. B.; Vengosh, A. Impacts of shale gas wastewater disposal on water quality in Western Pennsylvania. Environ. Sci. Technol. 2013, 47, 11849−11857. (12) Skalak, K. J.; Engle, M. A.; Rowan, E. L.; Jolly, G. D.; Conko, K. M.; Benthem, A. J.; Draemer, T. F. Surface disposal of produced waters in western and southwestern Pennsylvania: Potential for accumulation of alkali-earth elements in sediments. Int. J. Coal Geol. 2014, 126, 162− 170. (13) Harkness, J. S.; Dwyer, G. S.; Warner, N. R.; Parker, K. M.; Mitch, W. A.; Vengosh, A. Iodide, bromide, and ammonium in hydraulic fracturing and oil and gas wastewaters: environmental implications. Environ. Sci. Technol. 2015, 49, 1955−1963. (14) Parker, K. M.; Zeng, T.; Harkness, J.; Vengosh, A.; Mitch, W. A. Enhanced formation of disinfection byproducts in shale gas wastewater-impacted drinking water supplies. Environ. Sci. Technol. 2014, 48, 11161−11169. (15) Pennsylvania General Code. Chapter 95, section 10. Treatment requirements for new and expanding mass loadings of Total Dissolved Solids (TDS); rev. 2010; http://www.pacode.com/secure/data/025/ chapter95/s95.10.html. (16) Pennsylvania Department of Environmental Protection, Bureau of Water Standards and Facility Regulation, Policy and Procedure for 9564

DOI: 10.1021/acs.est.5b01780 Environ. Sci. Technol. 2015, 49, 9557−9565

Policy Analysis

Environmental Science & Technology (34) Dresel, P. E.; Rose, A. W. Chemistry and Origin of Oil and Gas well Brines in Western Pennsylvania, Open-File Report OFOG 10-01.0; Pennsylvania Geological Survey, 4th series, 2010. (35) Hayes, T. Sampling and analysis of water streams associated with the development of Marcellus Shale gas. Report prepared for the Marcellus Shale Coalition by the Gas Technology Institute, 31 December, 2009; http://energyindepth.org/wp-content/uploads/ marcellus/2012/11/MSCommission-Report.pdf. (36) Llewellyn, G. T. Evidence and mechanisms for Appalachian Basin brine migration into shallow aquifers in NE Pennsylvania, USA. Hydrogeol. J. 2014, 22, 1055−1066. (37) Standards for Protection against Radiation, Appendix B, Radionuclide Table-Radium-226, 10 CFR, 20; U.S. Nuclear Regulatory Commission: Rockville, MD, 2013. (38) Warner, N. R.; Jackson, R. B.; Darrah, T. H.; Osborn, S. G.; Down, A.; Zhao, K.; White, A.; Vengosh, A. Geochemical evidence for possible natural migration of Marcellus Formation brine to shallow aquifers in Pennsylvania. Proc. Natl. Acad. Sci. U. S. A. 2012, 109 (30), 11961−11966. (39) Orem, W.; Tatu, C.; Varonka, M.; Lerch, H.; Bates, A.; Engle, M.; Crosby, L.; McIntosh, J. Organic substances in produced and formation water from unconventional natural gas extraction in coal and shale. Int. J. Coal Geol. 2014, 126, 20−31. (40) Pennsylvania Department of Environmental Protection. July 2014−March 2015. Liquid Waste Disposal from Unconventional Wells in Pennsylvania, https://www.paoilandgasreporting.state.pa.us/ publicreports/Modules/DataExports/DataExports.aspx. (41) Assessment of the Potential Impacts of Hydraulic Fracturing for Oil and Gas on Drinking Water Resources, External Review Draft, EPA/600/ R-15/047a; U.S. Environmental Protection Agency: Washington, DC, 2015; http://cfpub.epagov/ncea/hfstudy/recordisplay.cfm?deid= 244651. (42) Papoulias, D. M.; Velasco, A. L. Histopathological analysis of fish from Acorn Fork Creek, Kentucky, exposed to hydraulic fracturing fluid releases. Southeast. Naturalist. 2013, 12 (4), 92−111. (43) Olsson, O.; Weichgrebe, D.; Rosenwinkel, K.-H. Hydraulic fracturing wastewater in Germany: composition, treatment, concerns. Environ. Earth Sci. 2013, 70, 3895−3906. (44) Rivard, C.; Lavoie, D.; Lefebvre, R.; Séjourné, S.; Lamontagne, C.; Duchesne, M. An overview of Canadian shale gas production and environmental concerns. Int. J. Coal Geol. 2014, 126, 64−76. (45) Technical Development Document for Proposed Effluent Limitations Guidelines and Standards for Oil and Gas Extraction, EPA-821-R-15003; U.S. Environmental Protection Agency: Washington, DC, 2015; http://water.epa.gov/scitech/wastetech/guide/oilandgas/upload/ UOG-Proposal-TDD-2015.pdf.

9565

DOI: 10.1021/acs.est.5b01780 Environ. Sci. Technol. 2015, 49, 9557−9565