Tugas 3 (B)

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 12

Science of the Total Environment 717 (2020) 137085

Contents lists available at ScienceDirect

Science of the Total Environment

journal homepage: www.elsevier.com/locate/scitotenv

Can we beneficially reuse produced water from oil and gas extraction in
the U.S.?
Bridget R. Scanlon a,⁎, Robert C. Reedy a, Pei Xu b, Mark Engle c, J.P. Nicot a, David Yoxtheimer d,
Qian Yang a, Svetlana Ikonnikova a
a
Bureau of Economic Geology, Jackson School of Geosciences, The University of Texas at Austin, TX, United States of America
b
New Mexico State University, Civil Engineering Department, Las Cruces, NM, United States of America
c
Dept. of Geological Sciences, The University of Texas at El Paso, TX, United States of America
d
Earth and Environmental Systems Institute, College of Earth and Mineral Science, Penn State Univ., PA, United States of America

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• Irrigation demand exceeds produced


water volumes by 5 times.
• Produced water quality is variable with
salinity up to 7× seawater.
• Intensive treatment is required for pro-
duced water use outside of energy.
• Produced water volumes would not
substantially alleviate overall water
scarcity.
• Knowledge gaps related to produced
water quality preclude reuse outside of
energy.

a r t i c l e i n f o a b s t r a c t

Article history: There is increasing interest in beneficial uses of large volumes of wastewater co-produced with oil and gas extrac-
Received 20 December 2019 tion (produced water, PW) because of water scarcity, potential subsurface disposal limitations, and regional link-
Received in revised form 31 January 2020 ages to induced seismicity. Here we quantified PW volumes relative to water demand in different sectors and PW
Accepted 1 February 2020
quality relative to treatment and reuse options for the major U.S. shale oil and gas plays. PW volumes from these
Available online 3 February 2020
plays totaled ~600 billion liters (BL, 160 billion gallons, Bgal) in 2017. One year of PW is equal to ~60% of one day
Editor: Dr. Damia Barcelo of freshwater use in the U.S. For these plays, the total irrigation demand exceeded PW volumes by ~5× whereas
municipal demand exceeded PW by ~2×. If PW is reused for hydraulic fracturing (HF) within the energy sector,
there would be no excess PW in about half of the plays because HF water demand exceeds PW volumes in those
plays. PW quality can be highly saline with median total dissolved solids up to 255 g/L in the Bakken play, ~7×
seawater. Intensive water treatment required for PW from most unconventional plays would further reduce
PW volumes by at least 2×. Desalination would also result in large volumes of salt concentrates, equivalent to
~3000 Olympic swimming pools in the Permian Delaware Basin in 2017. While water demands outside the en-
ergy sector could accommodate PW volumes, much lower PW volumes relative to water demand in most regions
would not substantially alleviate water scarcity. However, large projected PW volumes relative to HF water de-
mand over the life of the play in the Permian Delaware Basin may provide a substantial new water source for

⁎ Corresponding author.
E-mail address: bridget.scanlon@beg.utexas.edu (B.R. Scanlon).

https://doi.org/10.1016/j.scitotenv.2020.137085
0048-9697/© 2020 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
2 B.R. Scanlon et al. / Science of the Total Environment 717 (2020) 137085

beneficial use in the future. Large knowledge gaps in PW quality, lack of appropriate regulations, and economic
factors currently preclude beneficial uses outside the energy sector in most regions.
© 2020 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://
creativecommons.org/licenses/by/4.0/).

1. Introduction basins (Ferguson et al., 2018). Previous studies address a variety of


risks related to PW management, including pollution from spills and
Co-production of large volumes of waste water or produced water leaks and casing failures (Meng, 2017; Torres et al., 2016).
(PW) with the expansion of energy production from shale or unconven- Various PW management approaches have been suggested to re-
tional oil and gas (UOG) plays within the past decade is becoming an duce adverse environmental impacts, such as water scarcity, induced
important topic because these plays are located mostly in the semiarid seismicity, and contamination. The most obvious approach is to reuse
western U.S. where water scarcity is a critical issue (Fig. 1) (Huang or recycle PW within the energy sector for hydraulic fracturing (HF) of
et al., 2012; Reig et al., 2014; Scanlon et al., 2012). Water is also pro- new wells in shale or tight oil plays. PW reuse would reduce water
duced when coal beds are dewatered to mobilize methane, often sourcing to support HF. A recent analysis compared PW supplies relative
termed coal bed methane (CBM) (NASEM, 2010). Previous analysis es- to HF water demand in major U.S. plays, showing that PW reuse should
timated the total PW volumes in the U.S. to be ~3.4 trillion liters (TL; alleviate many of the adverse impacts of subsurface disposal; however,
3.4 km3 ~0.9 trillion gallons, Tgal) in 2012 (Veil, 2015), similar to previ- some plays have PW volumes that exceed HF water demands (Scanlon
ous estimates from 2007 (Clark and Veil, 2009). Most PW is injected into et al., 2020). PW reuse for HF was facilitated by advances in fracturing-
the subsurface, with almost half of the 2012 volume injected into high fluid chemistry that shifted water-quality requirements for HF from
permeability conventional reservoirs, mostly for pressure maintenance freshwater during the early years of UOG development to use of
and enhanced oil recovery (EOR). In contrast, PW from UOG reservoirs “clean brines” with minimal treatment in many regions (Barnes et al.,
cannot be managed by disposal into the shale and tight rock reservoirs 2015; McMahon et al., 2015; Nichols et al., 2017). Although a large per-
because of the low permeability, but instead is injected into intervals centage of PW is reused for HF in the Marcellus and Fayetteville plays
that do not produce oil and gas using salt-water disposal (SWD) wells. (Greaves et al., 2017; Rassenfoss, 2011), the total volumes of PW in
This process affects the subsurface water budget, resulting in increased these plays are low and can more readily be accommodated through
pressures and has been linked to seismicity, particularly if disposal is ad- reuse for HF than in plays with much larger PW volumes, such as the
jacent to basement rocks (Scanlon et al., 2019; Walsh and Zoback, Permian Basin (Scanlon et al., 2017). New Mexico has changed its regu-
2015). Regulations were promulgated in Oklahoma and New Mexico, lations precluding landowners from requiring operators to purchase
restricting disposal of PW in certain units, such as the Arbuckle in Okla- water for HF when PW is available (New Mexico House Bill 0546). How-
homa near the basement and its geological equivalent in New Mexico, ever, even in shale gas plays with low PW volumes, downturns in dril-
the Ellenburger, to reduce actual or potential induced seismicity ling can temporarily limit the potential for reuse for HF, as seen in the
(Lemons et al., 2019; Scanlon et al., 2019). Additional adverse impacts Fayetteville play (Greaves et al., 2017). Lack of reporting requirements
of subsurface disposal include contamination, with a recent analysis for PW reuse/recycling volumes makes it difficult to assess the extent
suggesting that disposal wells may impact overlying aquifers in some of PW reuse within the energy sector.

Fig. 1. Comparison of water demand for irrigation (2015), produced water volumes and hydraulic fracturing water demand (2017) for shale oil and gas reservoirs (Bakken, Niobrara,
Permian [Midland and Delaware basins], Eagle Ford, Barnett, Oklahoma Area of Interest [AOI], Haynesville, Fayetteville, and Marcellus) and coal bed methane reservoirs (Powder River,
San Juan, Uinta, and Black Warrior basins). See Fig. S1 for units in billion gallons.
B.R. Scanlon et al. / Science of the Total Environment 717 (2020) 137085 3

Another approach for managing PW is to reuse it outside of the en- list, with most contaminants listed N20 years ago. The limited under-
ergy sector, such as in irrigation, municipal, and industrial sectors, or standing of the toxicity of PW constituents underscores the risks and
to discharge treated PW to surface water or to recharge groundwater. hazards to humans and the environment from reuse outside of the en-
The feasibility of using PW for irrigation was evaluated for Colorado ergy sector (Danforth et al., 2019).
(Dolan et al., 2018). A recent study by the U.S. Groundwater Protection Projected exponential increases in PW from tight oil plays (Scanlon
Council (GWPC) evaluated the potential for beneficial use within and et al., 2020) raise the question about the adequacy of subsurface dis-
outside the energy sector, focusing on legal and regulatory issues and posal capacity to accommodate the PW increases. Recent projections
research needed to ensure safe use of PW in other sectors (GWPC, of PW over the life of the plays range from 1.1 TL (0.3 Tgal) in the
2019). New Mexico developed a Memorandum of Understanding with Eagle Ford to 49 TL (13 Tgal) in the Permian Basin (Scanlon et al.,
the U.S. Environmental Protection Agency (EPA) to assess regulatory 2020). The projected PW volumes in the Permian represent ~3× water
frameworks for PW reuse within and outside the energy sector, includ- use in the state of Texas in 2017 (17 TL, 4.6 Tgal). Therefore, potential
ing discharge to surface water (Danforth et al., 2019; USEPA, 2018). water scarcity, groundwater contamination, and induced seismicity
Beneficial uses of PW outside the energy sector will require much concerns underscore the need to assess the potential to develop benefi-
more intensive water treatment than that required to support HF cial uses for PW to partially mitigate these issues.
where minimal treatment (clean brine) is sufficient. However, charac- The objectives of this study are to address the following questions:
terizing all of the constituents in PW, including flowback water from
HF, is complicated because of difficulties with analytical techniques, • What is the potential for beneficially using PW from UOG reservoirs
problems with high salinity matrices, and lack of appropriate reference outside the energy sector based on volumetric water budgets?
materials (Oetjen et al., 2017; Tasker et al., 2019). Less than a quarter of • How feasible is beneficial use of PW considering water quality issues?
the ~1200 chemicals identified in PW have an approved analytical tech-
nique (Danforth et al., 2020). There is a lack of toxicological information
for the majority of chemicals found in PW (Danforth et al., 2020). This study builds on a previous analysis that focused on reuse of PW
Treatment technologies for PW are continually advancing, and selec- from UOG reservoirs within the energy sector (Scanlon et al., 2020). In
tion of appropriate technologies will depend on the PW quality, water- the current study, we examined a variety of potential beneficial uses
quality requirements for reuse options, and treatment economics. To for PW outside of the energy sector (Fig. 2). We quantified PW volumes
optimize PW reuse, fit-for-purpose treatment will be essential to mini- relative to water demand for different sectors, including irrigation, mu-
mize costs. Total Dissolved Solids (TDS) of PW from many UOG reser- nicipal, livestock, and industrial uses. We briefly discussed issues with
voirs may be too high (i.e. ≥~40 g/L) for traditional reverse osmosis discharge to surface water and recharge to groundwater. The analysis
(RO) approaches, and more complex thermal distillation approaches covers the major UOG and CBM plays in the U.S., leveraging off of previ-
may be required. Management of concentrates is also an important ous studies that quantified different components of the system in vari-
issue and can greatly increase treatment costs. Current water quality ous regions (Graham et al., 2015; Horner et al., 2016; Kondash et al.,
standards for different sectors, including irrigation and public water 2017; Nicot et al., 2014; Scanlon et al., 2017). We do not consider PW
supplies, are insufficient for assessing the feasibility of using PW in from conventional reservoirs because this is mostly reinjected into the
these sectors because the standards did not consider many of the con- high permeability reservoirs with minimal adverse environmental im-
stituents present in PW (GWPC, 2019). For example, the standards for pacts. However, we recognize that if incentives are created to reuse
public water supplies only include 90 contaminants in the EPA primary PW, operators may also consider PW from conventional reservoirs

Fig. 2. Beneficial use of produced water within the energy sector (hydraulic fracturing) and outside the energy sector (irrigation, municipal, industrial, livestock), surface water discharge
(evaporation ponds, stream discharge) and groundwater recharge.
4 B.R. Scanlon et al. / Science of the Total Environment 717 (2020) 137085

which could greatly increase the volumes of PW available for reuse. For we recognize that the TDS data alone are insufficient for assessing
example, PW volumes from the Permian conventional reservoirs were PW reuse outside of the energy sector. In particular, organics can
~10× greater than those from unconventional reservoirs (2005–2015) be toxic at low concentrations, and analytical tools to determine
(Scanlon et al., 2017). We also evaluated the lifespan of PW from se- presence and concentration of organics are limited. Data on major
lected plays (Bakken, Eagle Ford, Marcellus and Permian plays) based and trace element chemistry were also compiled, and we included
on projected PW volumes, which is important for evaluating the reliabil- analyses that had charge balances within ±15%. We reviewed treat-
ity of PW feedstock for developing treatment options. This study com- ment technologies based on PW quality and water quality require-
plements the recent GWPC report assessing the feasibility of beneficial ments for different sectors or for discharge to surface water or
use of PW in different sectors by providing quantitative data on the rel- recharge to aquifers. Options for managing concentrates were also
evant water volumes (GWPC, 2019). PW quality was evaluated using examined because they can greatly affect costs.
existing data from the USGS Produced Waters database (Blondes et al.,
2017) and literature studies. Treatment options for PW reuse were ex-
amined and salt concentrate management was evaluated. The quantita- 3. Results and discussion
tive data provided in this assessment will be valuable to regulators and
policy makers evaluating different options for managing PW. 3.1. Spatiotemporal variability in produced water supplies

2. Materials and methods PW volumes totaled ~600 BL (0.6 km3; ~160 Bgal) in 2017 from eight
major UOG reservoirs (Fig. 1, Table 1). These eight plays account for 88%
The primary emphasis of this study was on major shale oil and gas of tight oil and 84% of shale gas production in the U.S. based on 2018
plays within the U.S., often referred to as tight oil and shale gas data. PW volumes were much higher in western unconventional oil
(Fig. 1). The UOG plays evaluated include the Oklahoma Area of Interest plays than in eastern unconventional gas, with 50× higher PW in the
(AOI in terms of high seismicity), Bakken, Barnett, Eagle Ford, Fayette- Permian oil play relative to the Marcellus gas play in 2017. PW volumes
ville, Haynesville, Marcellus, and Permian (Midland and Delaware Ba- were much lower from CBM plays, totaling 46 BL (12 Bgal) in 2017.
sins) plays. Water issues related to selected CBM plays were also Beneficial use of PW would require a reliable supply of PW or feed-
examined, focusing on the Black Warrior, Powder River, Uinta, and stock. The only play with continuously increasing total PW volumes
San Juan plays. CBM plays were evaluated because they have a long his- over the past decade is the Permian Basin, where PW volumes increased
tory of PW management and they represent an end member in terms of by ~20× from 2011 to 2017 (Fig. 3a). PW volumes are not reported in
TDS because the PW TDS is generally much lower than that from UOG Oklahoma but are approximated by SWD volumes which have
reservoirs. remained fairly stable over the past several years at ~250 BL/yr (~65
Bgal/yr). PW volumes in many of the plays peaked in 2011 or 2014
2.1. Volumetric water budgets and generally declined since then with decreasing oil and gas prices or
resource depletion (2011 peak: Barnett, Fayetteville; 2014 peak: Eagle
The potential for beneficial use of PW depends on synergy between Ford, Bakken). PW from CBM plays peaked in 2009 and generally de-
PW supplies and sectoral water demands and alignment of PW quality clined since then because of the reduction of CBM production (Fig. 3b,
relative to quality requirements for different sectors. Data on PW vol- Fig. S2). For example, volumes of PW in the Powder River Basin (Wyo-
umes are available for ~100,000 UOG wells in the U.S. from state records ming) peaked in 2008 and declined by ~80% in 2017. The Permian
and commercial databases (IHS Enerdeq). These data do not consider Basin seems to provide the most reliable long-term PW feedstock
any reuse of PW for HF. Data on PW from 2017 were used in the analysis based on the increasing PW volumes.
because PW reporting lags by at least one year in many of the plays PW volumes available for beneficial use in other sectors decrease
(Texas plays: Barnett, Eagle Ford, Permian, and Haynesville). PW vol- substantially if the PW is first reused within the energy sector for HF
umes in excess of HF water demand in 2017 were also calculated for (Fig. 1, Table 1). Reuse within the energy sector should be maximized
the scenario with maximum reuse of PW for HF within the energy sector because it represents the lowest risks relative to reuse in other sectors.
prior to considering reuse outside the energy sector. Water demand for HF water volumes exceeded PW volumes in about half of the plays;
different sectors—including irrigation, municipal, and industrial sectors therefore, PW reuse for HF could potentially use up all of the PW de-
—was obtained from the U.S. Geological Survey (USGS) compilation pending on logistics. In contrast, PW exceeded HF water use in the re-
using the most recent data from 2015 (Dieter et al., 2018). PW volumes maining half of the plays (Bakken, Barnett, Oklahoma, and Permian
(2017) were compared to water demands for various sectors (2015) at Delaware basins). For example, PW in the Permian Delaware Basin
the play level and also at the county level, the lowest spatial unit in- was almost 2× HF water use in 2017. However, lack of reporting of PW
cluded in the USGS water use database. Projections of PW were com- reuse for HF precludes quantification of the extent of reuse within the
pared with projections of HF water demand based on previous energy sector. Other factors work against PW reuse for HF, such as land-
analysis of technically recoverable resource assessment over the life of owners and organizations requiring operators to purchase water from
the play, assuming all potential wells will be drilled using current tech- them as part of lease agreements in Texas. For example, HF water de-
nology (Scanlon et al., 2020). Data on projections are available for the mand in the Permian would represent ~$0.5 billion in 2017 based on
Bakken, Eagle Ford, Permian, and Marcellus plays. local water charges (~$2.2/1000 L; $0.35/barrel) (Scanlon et al., 2017).
The previous analysis focused on PW data from 2017. Projections of
2.2. Produced water quality PW and HF water are also important for assessing the potential for PW
reuse in the future. These data are available for the Bakken, Eagle Ford,
Data on PW quality were obtained from the USGS Produced Wa- Permian, and Marcellus plays (Scanlon et al., 2020). The potential for
ters database, which is based primarily on conventional oil and gas reuse outside of energy is similar for the projections and 2017 data for
reservoirs and CBM plays (Blondes et al., 2017). Details related to most plays except for the Permian Delaware Basin, where the ratio of
the data used in this study, including numbers of samples, types of PW/HF water doubles over the life of the play (ratio: 3.6) relative to
analyses (TDS, major and trace element chemistry), and time periods 2017 (ratio: 1.8). The increase in the ratio is attributed to the age of
covered are provided in Table S26. Additional data were obtained the well population, which is young in 2017 but increases substantially
from the literature based on field studies of sampling in selected over the life of the wells (~20 yr). Therefore, there should be increased
UOG plays. Most focus was placed on total dissolved solids (TDS) to opportunity for reuse of PW outside of HF water demand in the Permian
evaluate water treatment options and treatment goals; however, Delaware Basin within the next couple of decades.
B.R. Scanlon et al. / Science of the Total Environment 717 (2020) 137085 5

Table 1
Volumes (109 L) of produced water (PW), saltwater disposal (SWD), and hydraulic fracturing (HF) water during 2017 compared with irrigation (Irrig) and municipal (Muni) water use
during 2015 and projected remaining total PW volumes. Also listed are the remaining years of well completion activity based on the projected well inventory and historical annual max-
imum number of wells drilled per year for selected plays. For more details on the projections, see Scanlon et al., 2020. Data for additional sectors are provided in metric and English units in
Table S1. County level data for each play are provided in Tables S4 and S5.

Play Play name 2017 2015 Projected


Type
PW SWD HF Irrig Muni PW HF water Well invent. Max. wells Rem. yrs
per yr

TO Permian 264.4 378.6 210.2 841.7 65.1 49,300 18,200 320,000 5,000 64
TO Delaware 164.2 217.7 90.8 708.7 54.0 39,400 10,800 207,000 2,600 80
TO Midland 100.3 160.9 119.4 132.9 11.1 9,900 7,400 113,000 2,400 47
TO OK AOI 195.3 195.3 28.0 108.7 270.1 – – – – –
TO Bakken 54.6 56.7 29.3 747.5 18.8 2,500 1,700 69,000 2,700 25
TO Eagle Ford 35.1 38.9 66.2 219.1 144.6 1,100 3,200 105,000 4,000 26
TO Niobrara 6.7 6.4 28.0 867.4 362.8 – – – – –
SG Barnett 42.5 42.5 1.4 32.9 134.2 – – – – –
SG Marcellus 5.3 – 26.0 3.9 161.6 2,200 5,200 124,000 1,700 73
SG Haynes. 2.2 45.0 3.6 2.6 76.4 – – – – –
Total 606 763 393 2824 1234 55,100 – – – –
CBM PRB 26.1 – – 1,172.5 18.4 – – – – –
CBM Raton 7.8 – – 103.8 7.3 – – – – –
CBM San Juan 5.7 – – 855.5 36.5 – – – – –
CBM BW 4.9 – – 16.9 171.7 – – – – –
CBM Uinta 1.7 – – 227.6 9.9 – – – – –
Total 46.2 – – 2,376 243.8 – – – – –

TO: Tight Oil; SG: Shale Gas; CBM: Coal Bed Methane; Haynes: Haynesville Play; PRB: Powder River Basin; BW, Black Warrior Basin; PW volumes are not reported for the Oklahoma AOI but
are approximated by SWD volumes. PW and SWD volumes are assumed equal in the Barnett because of uncertainties in PW volumes. The Permian data represent the sum of data from the
Midland and Delaware basins and do not include development outside of these basins.

How long will PW be available for beneficial use? The PW volumes projected well inventory over the life of the plays (Scanlon et al.,
are projected to last ~25 yr for the Bakken and Eagle Ford plays, ~50 yr 2020) divided by the historical maximum drilling rate. Therefore, PW
for the Permian Midland Basin, and ~70 to 80 yr for the Permian Dela- supplies would provide the most reliable feedstock in the Permian Del-
ware and Marcellus plays (Table 1). These estimates are based on aware Basin.
Most of the PW from UOG reservoirs has been managed by subsur-
face injection using SWD wells. Although PW and SWD volumes are re-
ported separately without any direct linkage between the two, the
reported SWD volumes provide a check on PW volumes (Table 1).
SWD volumes are high in the Permian Basin and Haynesville plays but
include PW from both unconventional and conventional reservoirs. Cu-
mulative PW and SWD volumes are similar in the Bakken and Eagle Ford
plays. PW from CBM plays has not been managed using subsurface dis-
posal but mostly surface disposal in ponds, as described in a later section.

3.2. Assessing water demand in different sectors relative to produced water


supplies

3.2.1. Irrigation sector


The sector with the largest water demand is irrigation (2015 data),
exceeding PW volumes in UOG plays (2017 data) by ~5× and exceeding
PW volumes from CBM plays by ~50× (Fig. 1). Irrigation is concentrated
mostly in the western U.S. (Figs. 4, S3). Box plots of water use in differ-
ent sectors are provided in Figs. S4 and S5 and Tables S4 through S13.
The much higher irrigation volumes relative to PW volumes at the
play level means that the irrigation sector should be able to accommo-
date the PW volumes. If PW is reused for HF within the energy sector,
there would be no excess PW in about half of the plays because HF
water demand exceeds PW volumes in those plays (Table 1). The irriga-
tion to PW ratio is similar in the remaining plays (ratio: 5). Irrigation
could accommodate excess PW volumes in these plays, including the
Permian Delaware (irrigation/excess PW ratio: ~10×) and Bakken
(ratio: 30×) plays but not in Oklahoma AOI (ratio: ~0.6). There is also
a temporal disconnect between PW, which is generated throughout
the year, and irrigation demand, mostly restricted to summer months.
Irrigation at the play level was highest in the Niobrara (870 BL, 230
Bgal) and Permian Basin (~840 BL, ~220 Bgal) plays in 2015 (Table 1).
Although most (96%) PW in the Niobrara is from a single county
Fig. 3. Time series of produced water (PW) volumes for a) the major tight oil and shale gas
(Weld County), HF water use exceeds PW in this county by ~4× and
plays and b) the coal bed methane plays. Data for Oklahoma represent statewide values. could reuse all of the PW (Fig. S10, S19; Table S18). If PW was used
The data are tabulated in Tables S2 and S3. for irrigation, it would contribute only ~1% to irrigation in this county,
6 B.R. Scanlon et al. / Science of the Total Environment 717 (2020) 137085

Fig. 4. Distribution of irrigated lands in the U.S. based on composite satellite images for 2002, 2007, and 2012 based on Moderate Resolution Imaging Spectroradiometer (MODIS) Irrigated
Agriculture Dataset for the United States (MIrAD-US) (https://earlywarning.usgs.gov/USirrigation). The various play regions discussed in this study are highlighted. The area of focused
unconventional oil and gas development located in a 19-county region of the Permian Basin is also outlined within the basin, showing most intensive irrigation north of the UOG
development.

with little impact on water scarcity. These results are consistent with 3.2.2. Municipal, livestock, industrial, and mining sectors
previous findings (Dolan et al., 2018; Walker et al., 2017). Irrigation Many of the UOG reservoirs are located in rural areas with minimal
exceeded PW volumes by ~3× in the Permian Midland and Delaware municipal water use. Total municipal water demand in 2015 in the
Basins (Table 1). The most intensive irrigation is generally north of the area of UOG reservoirs exceeded PW volumes in 2017 by ~2×
UOG development (~2500 BL; 660 Bgal), ~3× irrigation in the Midland (Table 1). Municipal demand was highest in the Niobrara (30% of total
and Delaware Basins (Fig. 4). PW represented 1%, 12%, and 17% of irriga- water demand) and Oklahoma AOI (22%), followed by the Barnett,
tion demand in counties with the highest irrigation (Eddy, Lea, and Eagle Ford, and Marcellus (11%–13%). PW would represent b0.1% of mu-
Pecos counties (2017) in the Permian Delaware basin) (Fig. S6, nicipal demand in Niobrara counties with high municipal demand
Table S14). These percentages would decrease to bb1%, 5%, and 11% if (Fig. S10). About 50% of the municipal demand in Oklahoma AOI is in
PW was reused to meet HF water demand in these counties. Counties Oklahoma County, where PW would constitute ~3% of the municipal de-
with large volumes of PW relative to irrigation (e.g. Midland County: mand (Fig. S7). Similarly in the Marcellus, municipal demand was
PW/irrigation ~30) could support expansion of irrigation, but PW in highest in a couple of counties, but PW would only represent ≤3% of
this county represents ~3% of irrigation in the Permian. Irrigation in this demand because of the low PW volumes in this play (Fig. S12). Mu-
the Eagle Ford exceeded PW volume by ~6× (Table 1), with most irriga- nicipal demand in CBM plays was ~5× PW volumes, mostly in the Ala-
tion in a few counties where PW would contribute b1% to irrigation de- bama portion of the Black Warrior Basin (Table 1). PW would
mand. However, HF water use exceeded PW in most counties by up to represent ≤8% of municipal demand in the Black Warrior Basin. In sum-
2–3×; therefore, reuse of PW for HF would eliminate the PW source mary, PW would contribute minimally to municipal demand in most
for irrigation. Irrigation in the Bakken/Three Forks is mostly restricted cases from a volumetric standpoint. In the past, water has generally
to a few counties, with PW accounting for b1% of irrigation in most been transferred from the municipal sector to UOG reservoirs for HF in
counties. Total irrigation in the Oklahoma AOI is less than the PW vol- the Barnett, Niobrara, and Permian plays, either from freshwater or
ume but varies spatially (Table 1). Irrigation is generally much higher treated municipal waste water (Nicot et al., 2014; Scanlon et al., 2017;
than PW volumes in the western CBM plays, ~45× PW in the Powder Walker et al., 2017).
River Basin (Table 1), with PW contributing b5% to irrigation in most Water use for livestock was ~25% of total PW volume in UOG reser-
counties (Fig. S14, Table S23). voirs in 2017 and was highest in the Haynesville and Eagle Ford plays
In summary, irrigation could accommodate PW volumes in various (Table 1). Counties with high livestock water demand generally did
regions of western plays; however, the high ratios of irrigation to PW not coincide with relatively high PW volumes in these plays. Livestock
in many counties suggest that PW volumes would not substantially re- water use was negligible in CBM plays.
duce water scarcity in these regions. In many cases, HF water demand Total volumes for industrial water uses were generally low in UOG
exceeds PW volumes and would eliminate the PW source for irrigation and CBM regions (Table S1). The industrial water use category refers
if PW was reused within the energy sector. In the past, water transfers to self-supplied withdrawals for the industrial sector (Dieter et al.,
generally occurred from the irrigation sector to the oil and gas sector 2018).
for HF in some plays, e.g., Bakken/Three Forks, Niobrara, and Permian Thermoelectric cooling for power plants requires large water with-
plays (Horner et al., 2016; Kurz et al., 2016; Scanlon et al., 2016; Shuh, drawals in once-through cooling systems, but consumption equals
2010). However, lack of reporting of water sourcing for HF makes it dif- only a few percent of withdrawal (Dieter et al., 2018). This study focused
ficult to track water transfers among sectoral users. on the consumptive water use for recirculating cooling towers, which
B.R. Scanlon et al. / Science of the Total Environment 717 (2020) 137085 7

have lower water withdrawals and similar rates of water consumption 3.4. General quality of produced water
than once-through systems. Water use for recirculating cooling was
low, ~20% of PW, highest in the Oklahoma AOI (Table S5). PW would Produced water can contain oil and grease droplets, suspended
represent 23%–45% of cooling water use in counties with the highest solids, major elements, transition metals, naturally occurring radioac-
water demand for these plants (Noble and Oklahoma counties). tive material (NORM), organic compounds, and microbes. Chemicals
The mining water use sector includes water use for extraction of present in HF fluids as reported in FracFocus (FracFocus.org) may also
rocks and minerals (e.g. coal, sand, and gravel) and fossil fuels be contained in PW, particularly during the early stages of production.
(Table S1) (Dieter et al., 2018). However, the mining values are difficult Data on PW quality are dominated by analyses from conventional oil
to interpret because they do not seem to be internally consistent. The and gas reservoirs from the USGS PW database, with limited data from
mining values should include HF water but are less than HF water in UOG reservoirs (Blondes et al., 2017). TDS provide a general indication
the Marcellus and Niobrara, are similar to HF water in the Bakken, and of the mineral content of the PW with implications for water treatment
are similar to HF water + PW in Texas plays (Barnett, Eagle Ford, and and for salt management. Median TDS is highest in Bakken tight oil
Permian). Therefore, the mining sector does not seem to provide an op- (244 g/L), ~7× that of sea water (Figs. 5, 6). TDS is moderately high in
portunity for reuse of PW. There may be some local cases where PW the Permian tight oil play (median: 154 g/L) and the Appalachian (Mar-
could be used in another mining sector. For example, PW could be cellus) shale gas play (166 g/L), with a lower value in the Eagle Ford
used for potash mining in New Mexico; however, the potash mines shale play (57 g/L). The Permian TDS data include PW from unconven-
are currently selling water to UOG operators because potash prices are tional units, such as the Wolfcamp and Cline Shales, along with wells
low (Hayden, 2017). specifically designated as unconventional in the USGS database. The
In summary, the most obvious sector for reuse of PW from a volu- highest TDS (99th percentile) in shale oil and gas plays ranges from
metric perspective is irrigation. With irrigation greatly exceeding PW 300 to 370 g/L, with a lower value in the Eagle Ford play (~180 g/L)
volumes, it can accommodate the PW volumes in semiarid oil plays and Niobrara (95 g/L). No systematic difference exists between TDS in
where PW volumes are highest. However, the PW volumes represent conventional and unconventional reservoirs as seen from the USGS da-
a small percentage of irrigation demand in most counties, particularly tabase. There can be substantial variability in TDS within plays, both ar-
if PW is first reused for HF; therefore, PW will not substantially help re- eally and vertically (Fig. 5). Recent analysis of Permian data reveals the
duce water scarcity concerns in these plays. highest TDS in shallower zones near salt deposits and decreasing TDS
with depth (Chaudhary et al., 2019). In the Eagle Ford play, a salinity re-
versal and freshening of PW (a factor of 10 reduction in TDS from
3.3. Surface water discharge and managed aquifer recharge ~200 g/L at 2.5 km [1.6 m] depth to ~20 g/L at 3.5 km [2.2 m] depth)
was noted and attributed to clay conversion from smectite to illite
Additional beneficial uses of PW include discharge to surface water with interlayer water release (Nicot et al., 2018). PW from CBM plays
and recharging groundwater. Most PW from CBM plays has been is generally much fresher than that from the shale oil and gas plays,
discharged at the surface into unlined impoundments in Wyoming with median TDS ranging from ~1 g/L in the Powder River Basin to ~
(~4000 permits by 2007) (Healy et al., 2011) or discharged to rivers, 11 g/L in the Uinta Basin (Fig. 5).
such as the Black Warrior River in Alabama, without any or with mini- PW from tight oil and shale gas plays is dominated by Na (median:
mal treatment (e.g. settling ponds). Comparing PW volumes with flow 15–76 g/L) and Cl (median: 22–150 g/L) (Fig. S31), consistent with pre-
in the Black Warrior River, which is in a humid region, suggests that vious studies indicating that nearly all basinal waters N10 g/L TDS are
river discharge exceeds rates of PW generation by ~50 times during dominated by Na\\Cl, with other ions existing only as minor or trace
low flows and up to 400 times during normal flows (SI, Section 1a). constituents (Hanor, 1994). Levels of Ca are generally much lower
Treated PW has also been discharged to surface water in the Marcellus (~0.3–13 g/L). Sulfate levels are mostly low (median:
play; however, there were contamination issues during the early years 0.020–0.675 g/L). PW from CBM plays is dominated by Na and Cl in
(SI, Section 1b). Incentives in the form of a tax credit ($0.13/barrel; the Black Warrior and Uinta Basins but Na and HCO3 in the Powder
$0.01/L) were put forward by New Mexico to promote discharge of River, San Juan, and Raton Basins, reflecting marine versus terrestrial
treated PW into the Pecos River (SI, Section 1c). However, PW discharge depositional environments (Fig. S32). Low sulfate concentrations in
from oil and gas wells within a 50 km corridor of the Pecos River (arid PW from CBM plays are attributed to methanogenesis. Median sodium
setting) would exceed annual stream discharge mostly by factors of adsorption ratios (SARs), important for irrigation, are generally high in
4–10× but up to 20× in dry years and by much greater values during ex- PW from CBM plays but vary considerably among the plays studied,
tremely dry years (Fig. S30b). The natural water quality in the Pecos with medians ranging from 7 to 294 in different basins (Fig. S33).
River is quite variable, with total dissolved solids ranging from 1.6 to SARs N3 generally require freshwater to flush the salts in irrigated
17 g/L in different gages along the river based on USGS data (https:// lands, and SARs N13 can degrade soil texture (SI, Section 1d).
waterdata.usgs.gov/nwis). In summary, river discharge in humid re- Data on minor or trace elements in PW are limited. Here we only
gions greatly exceeds PW volumes, increasing the assimilative capacity mention naturally occurring radioactive material (NORM) in PW,
of the rivers for treated PW. In contrast, rivers in semiarid regions are which mostly consists of 226Ra (half life: 5.75 yr) and 228Ra (half life:
mostly ephemeral and discharge is much less than the PW volumes, 1600 yr) derived from U and Th in the reservoir (Guerra et al., 2011).
complicating discharge of treated PW. Total Ra (226Ra + 228Ra) levels greatly exceed the EPA regulatory limit
The feasibility of groundwater recharge and aquifer storage and re- of 5 pCi/L (pico curies/L) for drinking water but are higher in the Marcel-
covery (ASR) is also being considered in some plays, such as the Perm- lus shale (median: 1980 pCi/L) and the Bakken (1,200 pCi/L) than in the
ian Basin. However, the risks related to such a practice seem high, and Permian (535 pCi/L) or Eagle Ford (284 pCi/L) (Fig. S31).
potential unintended consequences large. Storing excess PW in the sub-
surface would provide a means to resolve PW supply relative to water 3.4.1. Sources of produced water
demands for different sectors. An alternative to aquifer storage would Combinations of compositional and isotopic data provide significant
be to use deeper geologic units currently used for salt water disposal, insight into the origin of PW from oil and gas reservoirs. The presence of
such as the Delaware Mountain Group in the Permian Delaware Basin Cl as the dominant anion in nearly all PW with salinity N10 g/L and max-
or the San Andres Formation in the Permian Midland Basin (Lemons imum TDS concentrations several times seawater (35 g/L TDS) suggest
et al., 2019). Pumping water from these units to support HF would re- that the dominant source of PW and ions is evaporated seawater and/
duce overpressuring from SWD and potential contamination of overly- or dissolution of Cl-bearing evaporites. Studies of UOG plays, including
ing aquifers (Ertel and Bogdan, 2017; Landis et al., 2016). the Bakken, Marcellus, and Permian Basin, conclude that these same
8 B.R. Scanlon et al. / Science of the Total Environment 717 (2020) 137085

solute and water sources also dominate most shale reservoirs (Engle
et al., 2016; Lauer et al., 2016; Rowan et al., 2015). Black shale specific
processes, such as hydrocarbon maturation, clay diagenesis, and
water-clay interactions, have also been shown to control the composi-
tion of PW from unconventional plays (Engle et al., 2016; Nicot et al.,
2018; Phan et al., 2016; Stewart et al., 2015).
Sources of water and solutes in CBM plays vary depending on geo-
logic history and depth. Low TDS (median: 2.9 mg/L) Na-Cl-type waters
found in much of the Black Warrior Basin indicate a marine origin
(Pashin et al., 2014), with even lower TDS (median ~1 g/L) Na-HCO3-
rich water from the Powder River Basin reflecting meteoric water that
has undergone a series of geochemical reactions within the coal beds
from which they are produced (Brinck et al., 2008).

3.4.2. Evaluating water quality requirements for different sectors relative to


produced water quality
There is a long history of assessments of potential reuse of PW from
CBM plays for irrigation and surface water discharge (Arthur et al.,
2005; Guerra et al., 2011). CBM PW discharges are managed outside of
national programs and are subjected to local permitting requirements.
PW from CBM plays is of much higher quality than that from shale
UOG plays, with median TDS values of 1–11 g/L (Fig. 6). In addition,
CBM PW does not contain the chemicals from HF. However, CBM PW
does contain polycyclic aromatic hydrocarbons from the associated
coal (Orem et al., 2014). For example, PW in the Black Warrior Basin
(median TDS: ~4 g/L) is discharged directly to the Black Warrior River Fig. 6. Total dissolved solids (TDS) for tight oil (TO) and shale gas (SG) reservoirs
in Alabama where it is diluted (SI, Section 1a). However, some issues re- (unconventional oil and gas reservoirs), conventional oil and gas reservoirs (Conv), and
coal bed methane (CBM) reservoirs. The Bakken unconventional tight oil includes the
lated to use of CBM PW for irrigation include SAR and negative impacts
Bakken and underlying Three Forks units. Coal bed methane reservoirs include Powder
on soil infiltration and other toxic constituents, such as boron etc. (SI, River, Raton, Black Warrior, and Uinta basins. The numbers at the base refer to the
Section 1d). Use of high quality PW from CBM for irrigation can leach number of analyses. For locations of reservoirs, see Fig. 1. The USGS data are provided in
salts from semiarid soils, resulting in poor quality water reaching Table S26. Additional data were obtained from Nicot et al. (2018).

Fig. 5. Total dissolved solids (TDS) of produced water from the USGS Produced Waters database (version 2.3) with supplemental data for the New Mexico region of the Permian Basin
provided by the New Mexico Institute of Mining and Technology (NMIMT) Petroleum Research and Recovery Center (PRRC), and data from the USGS in the Eagle Ford Play. Labeled values
represent median produced water TDS concentrations within each play area of wells classified as either shale gas, tight oil, or coal bed methane, with the exception of the Permian value,
which includes wells that are classified as conventional hydrocarbon wells completed in unconventional formations (i.e., Wolfcamp, Bone Spring, Cline, Spraberry, and Dean). Data are
provided in Table S26.
B.R. Scanlon et al. / Science of the Total Environment 717 (2020) 137085 9

underlying aquifers, with up to 100 g/L TDS in groundwater near an im- (2) the quality of the water that is generated relative to the require-
poundment in Wyoming (Healy et al., 2011). ments for beneficial use.
Water quality standards and regulations for water use in irrigation, Recovery efficiency and waste generation from treatment are also
livestock, and municipal sectors (drinking water maximum contami- important factors.
nant levels and irrigation and land application standards) do not con-
sider many of the risks related to use of PW from UOG reservoirs and 4.1. Desalination of produced water for industrial, agricultural, and potable
are not appropriate for assessing such use (GWPC, 2019). For example, reuse
the EPA primary drinking water regulations include ~65 contaminants
within the chemical contaminant rules, including inorganics and vola- Removal of dissolved solids (desalination) is required for applications
tile and synthetic organics (https://www.epa.gov/dwreginfo/chemical- that require high water quality. Two primary technologies currently used
contaminant-rules). Assessing treatment of PW by comparing with for water desalination include those based on (1) membrane technologies
drinking water standards is insufficient because the regulations were or (2) thermal technologies (Fig. 7). Additional information on treatment
not designed to consider PW. technologies is provided in SI, Sections 2 and 3.
Surface discharges of PW from UOG wells are regulated under the Membrane technologies include electrodialysis (ED), electrodialysis
National Pollution Discharge Elimination System (NPDES) and related reversal (EDR), nanofiltration (NF), and reverse osmosis (RO). NF and
national effluent-limitation guidelines or corresponding state regula- RO use high hydraulic pressure to diffuse pure water through a dense
tions (GWPC, 2019). The EPA prohibits discharge of PW from UOG non-porous membrane and retain solutes on the feed water side of the
plays to public-owned treatment works, but PW can be discharged to membrane. In general, RO is capable of treating PW with TDS up to
Centralized Water Treatment (CWT) facilities. ~40 g/L due to the limitation of pressure vessels; however, RO becomes
Because of the large number of potential organic contaminants in too energy intensive and expensive at much lower levels of TDS
PW and difficulties in analyzing many of these elements within the con- (~15–35 g/L). Water recovery varies with salinity, ranging from 30% to
text of high salinity matrices (Danforth et al., 2019; Luek and Gonsior, 60% for seawater (35 g/L) to 60% to 85% for brackish water (≤10 g/L)
2017; Nelson et al., 2014; Oetjen et al., 2017), it seems infeasible to con- (Igunnu and Chen, 2014). RO treatment is effective in removing almost
duct a comprehensive evaluation of PW chemistry. Detailed risk assess- all inorganic contaminants, including NORM; however, additional ana-
ment will be required to evaluate potential human health and lytical techniques are required to determine whether some organics re-
environmental impacts of beneficial PW reuse outside of the energy sec- main in the treated water. To minimize membrane scaling and fouling,
tor. The GWPC report outlines many of the aspects of a suitable risk as- RO requires extensive pretreatment to remove sand, silt, clay, algae, mi-
sessment, building on previous studies (NRC, 2009). Whole effluent crobes, colloidal particles and large molecular organics (e.g., petroleum
toxicity (WET) assessment is proposed to address known and potential hydrocarbons), and sparingly soluble salts. RO has been applied to PW
unknown contaminants (GWPC, 2019). from the Marcellus, Barnett, and Fayetteville plays (ALL, 2010).
Thermal technologies (e.g., multiple-effect distillation [MED], me-
chanical vapor compression [MVC] and recompression [MVR]) are al-
4. Treatment options for produced water reuse applications most independent of source water salinity (Fig. 7). Thermal distillation
technologies involve heating and evaporating feed water followed by
Selection of the appropriate treatment technologies depends pri- condensation of pure water. Typical water recoveries range from 20%
marily on two factors: to 35% for MED to 40% for MVC (Igunnu and Chen, 2014). These low
(1) the quality of the input water or feed water, particularly salinity water recoveries result in large volumes of concentrate that need to
and other inorganic and organic constituents, and be disposed of. Biocides are not required because of the elevated

Fig. 7. Treatment technologies for produced water, including minimal treatment of PW for hydraulic fracturing (clean brine), desalination for beneficial uses in various sectors, surface
water discharge and groundwater recharge, and posttreatment technologies. TDS: total dissolved solids; ED: electrodialysis; NF: Nanofiltration; BWRO: brackish water reverse
osmosis; SWRO: seawater reverse osmosis; MED: multiple effect distillation; MVC: mechanical vapor compression; MVR: mechanical vapor recompression, emerging technologies
including FO: Forward Osmosis; MD: membrane distillation. AOP: advanced oxidation processes.
10 B.R. Scanlon et al. / Science of the Total Environment 717 (2020) 137085

temperature. MVR has been applied to waters with up to 300 g/L TDS be injected into a SWD well. The low recoveries of thermal distillation
and can generate high quality water (Nasiri et al., 2017). MVR can be techniques result in up to 80% concentrate that is generally injected
used as a crystallizer for systems with zero liquid discharge (ZLD). How- into SWD wells. In some cases, the concentrate is solid (ZLD) and the
ever, managing solids is often challenging; therefore, systems designed products are either marketed (salts) or disposed of in landfills (Ertel
to generate a concentrated brine are often preferred (GWPC, 2019). and Bogdan, 2017). For example, PW from the Delaware Basin in 2017
Thermal approaches have high energy requirements and are generally (160 BL, 43 Bgal) would result in 16 × 109 kg of salt, or 16 million
used where waste heat is available. tons assuming a TDS of ~100 g/L (SI, Section 3). This would correspond
In addition, emerging technologies are being developed to improve to a volume of solids approximating ~3000 Olympic swimming pools.
certain aspects of the performance of existing desalination processes Disposal of solid wastes laden with contaminants transferred and con-
(e.g., increasing recoveries, reducing fouling, decreasing energy con- centrated from PW may be cost prohibitive and may not be a viable op-
sumption and capital and operating costs). These new technologies tion to reach ZLD or high water recovery for PW treatment.
can be classified into three categories: thermal (membrane distillation,
MD), physical (forward osmosis, FO), and chemical (capacitive deioni- 4.2. Implications for produced water management
zation). MD uses a heat source to enhance mass transport through
membranes. One of the advantages of this approach is its ability to use There is considerable interest in beneficially using PW outside of the
any level of TDS in the feed solution. Solutes, including Na, SiO2, B, and energy sector because of the perception that PW represents huge water
heavy metals, are rejected at nearly 100%. Water recovery can be im- volumes and reuse would retain the water within the hydrologic cycle,
proved when paired with crystallizer technologies. A research example whereas subsurface disposal removes it from the active hydrologic
based on treating PW with ~250 g/L TDS showed this approach to be cycle. The total volume of PW from UOG reservoirs in 2017 (~600 BL,
cost competitive (Macedonio et al., 2014). Forward osmosis is an os- ~160 Bgal) corresponds to ~60% of fresh water use in the U.S. in one
motic pressure driven membrane process (Hickenbottom et al., 2013; day, excluding thermoelectric water use. However, most of the PW is
Xu et al., 2013). Water diffuses from a feed stream with low osmotic generated in semiarid regions where water scarcity is a big concern. In
pressure through a semi-permeable membrane to a draw solution gas plays (Haynesville, Marcellus) and some oil plays (Eagle Ford, Nio-
with high osmotic pressure. No external high pressure is required by brara, and Permian Midland), HF water demand exceeds PW volumes;
an FO system. Other hybrid configurations are also being developed by therefore, much of this PW could be reused within the energy sector
different companies, such as the Veolia OPUS™ system. Development and reduce water scarcity caused by pumping water to supply HF.
of cost-effective desalination technologies represents the new frontier Reuse within the energy sector represents much less risk with minimal
of PW treatment research. Although these emerging technologies have treatment and related lower costs and energy use than beneficial use of
shown promise in highly saline PW treatment, improvements in mem- PW in other sectors.
brane properties, membrane design, and module hydrodynamics are The most likely sector to reuse PW in semiarid western U.S. is irriga-
expected to further increase system efficiency (Shaffer et al., 2013). tion. Irrigation exceeds PW volumes in many plays, except the Okla-
A limited number of treatment plants have been developed for PW. homa AOI, Barnett play, and Permian Delaware Basin. If we assume
Antero Resources built a large desalination plant (Antero Clearwater Fa- that PW is first reused within the energy sector, only half of the plays
cility) in the Marcellus to treat up to 10 million liters of PW/day (GWPC, would have excess PW relative to HF water demand. Considering
2019). The estimated cost was ~$300 million; however, the facility was water treatment requirements and recovery factors would further re-
recently closed for evaluation. Eureka Resources developed three central- duce the PW volumes. PW as a percent of irrigation would represent
ized wastewater treatment facilities to generate water fit for discharge 2% of irrigation in the Bakken, 5% in the Permian Delaware Basin, 63%
from PW in the Marcellus (Mueller, 2017). Treatments include mechani- in the Barnett, and 77% in Oklahoma AOI, assuming 50% recovery factors
cal vapor recompression distillation and crystallization. ALL Consulting for treatment. The percentages are more variable at the county scale,
provided a treatment technology tool that includes treatment types and showing counties with large potential in the Oklahoma AOI and Barnett.
vendors for specific basins: Marcellus, Fayetteville, Haynesville, Some point to examples in California where PW is reused for irrigation;
Woodford, and Barnett; however, the information is generally applicable however, the water quality is quite high and treatment is minimal (SI,
to other basins (http://www.all-llc.com/projects/produced_water_tool/). Section 4). Other sectoral uses represent more localized demands,
Besides salts, PW and flowback water contain complex organic con- such as municipal and industrial uses. Detailed site specific studies
stituents such as oil and grease, BTEX (benzene, toluene, ethylbenzene would be required before discharging PW to surface water or recharging
and xylene), PAHs (polycyclic aromatic hydrocarbons), biopolymers, groundwater would be considered. Therefore, it does not seem that PW
and humic substances (Danforth et al., 2020; Khan et al., 2016). These reuse will mitigate water scarcity concerns in most regions. Water scar-
organic constituents in high salinity water present unique challenges city related to oil and gas development would be more readily ad-
to most technologies. Organic matter must be removed first to minimize dressed by reusing PW within the energy sector in different plays.
fouling of membranes or other surfaces, and to prevent environmental Development of treatment plants for beneficial reuse of PW would
impacts during disposal or reuse. Biological processes have been suc- require a reliable feedstock. Maintaining PW volumes requires contin-
cessfully used to remediate water contaminated by petroleum hydro- ued drilling because of exponential declines in PW volumes with time
carbons, solvents, and other dissolved organic chemicals commonly from UOG and CBM plays (Fig. 8).
observed in PW. The effectiveness of biological processes varies depend- Many factors currently do not support reuse of PW outside of the en-
ing on the properties of chemicals present and their respective concen- ergy sector. From a volumetric perspective, reuse of PW within the en-
trations, as well as the salinity level in PW. For example, organic ergy sector would eliminate half of the plays because HF water demand
compounds such as BTEX and PAHs are readily biodegradable in aerobic exceeds PW in these plays. In the remaining plays, PW represents a
processes, whereas halogenated organic compounds or highly chlori- small fraction of water demand for irrigation mostly at the play and
nated compounds are more refractory. Physico-chemical treatment, county level. In the future, high projected PW volumes in the Permian
such as adsorption and advanced oxidation processes, can be used to re- Delaware Basin, exceeding projected HF water demand, would support
move organic contaminants. Because the treatment technologies may PW reuse outside of energy in future decades. From a water quality per-
not be highly effective in removing contaminants, risk assessment spective, the following limitations restrict the potential for PW reuse
should also be conducted on the treated PW to evaluate the impact of outside of energy: poor knowledge of PW chemistry, inability to accu-
PW reuse on public health and environment. rately measure the PW quality because of high salinity matrix and inter-
Limited information on concentrate management indicates that PW ference issues, lack of acceptable measurement techniques, absence of
is often treated to a point where the residual is still sufficiently liquid to suitable standards, and lack of regulations for various sectors to consider
B.R. Scanlon et al. / Science of the Total Environment 717 (2020) 137085 11

Variable TDS of PW from UOG reservoirs (median TDS: 244 g/L in the
Bakken, up to 7× seawater), would mostly require thermal distillation
approaches to treat the PW. PW from CBM reservoirs is much higher
quality (median TDS: 1–11 g/L) and does not include HF water. Evalua-
tion of PW from CBM plays provides an example of PW that can be con-
sidered for potential reuse outside the energy sector. It is difficult to
characterize the PW quality from UOG reservoirs because of problems
with measurements, interferences caused by the high salinity matrix,
and lack of suitable standards resulting in many unknowns. Uncer-
tainties in treatment efficiency and problems with contaminants
leaking through the treatment process further increase risks of reuse.
The current regulations for various sectoral uses and discharge require-
ments were not designed to address PW issues. Therefore, large uncer-
tainties related to water quality issues currently preclude PW reuse
outside of the energy sector.
In summary, water scarcity issues are more readily addressed by
reusing PW within the energy sector rather than beneficial use of PW
outside of the energy sector. Much more research is required to safely
reuse PW in other sectors or discharge to surface waters or recharge
aquifers.
Fig. 8. Produced water decline curves for the major tight oil (TO: Bakken, Midland,
Delaware), shale gas (SG: Marcellus), mixed TO and SG, (Eagle Ford) plays and for the
Powder River Basin coal bed methane play (PRB). Data represent the median monthly Declaration of competing interest
produced water volumes expressed as a percentage of first-month production for wells
completed during 2015–2017, except the PRB (1989–2018), in which few wells produce
The authors declare that they have no known competing financial
longer than 10 years. Note that the Delaware and Marcellus declines are virtually identical.
interests or personal relationships that could have appeared to influ-
ence the work reported in this paper.
the complexity of PW. In addition, suitable treatment technologies, such
as thermal distillation, are expensive and have high energy demands. Acknowledgments
The efficiency of different treatment technologies in removing certain
contaminants, particularly those with a boiling point similar to water, We are very grateful for financial support for this study from the
for example, is also not known. Because of these many uncertainties, Mitchell Foundation, Sloan Foundation, ExxonMobil, and Jackson School
more emphasis is placed on risk assessment, whole effluent toxicity, of Geosciences Endowment. We very much appreciate access to the IHS
and related factors. The risks and liabilities associated with PW reuse Enerdeq database.
outside the energy sector are high, and much more data are required
to address uncertainties in this field before such reuse should be Appendix A. Supplemenary data and Information
considered.
Raw data on water volumes and water quality used in this study are
archived in Mendeley Data (10.17632/jjy5mtfkfk.2) and described in an
5. Conclusions associated data article (Scanlon et al., submitted). Supplementary data
to this article can be found online at https://doi.org/10.1016/j.
Although interest in beneficially reusing the large volumes of PW scitotenv.2020.137085. Additional figures and tabulated data are pro-
generated from UOG reservoirs is increasing in the U.S., quantitative vided in the Supporting Information.
analysis of the volumetric and water quality issues does not support
reuse outside of the energy sector.
References
PW volumes in 2017 totaled ~600 BL (160 Bgal) from UOG reser-
voirs, primarily from tight oil reservoirs in the semiarid western U.S., ALL, 2010. Water Treatment Technology Fact Sheet: Reverse Osmosis. ALL Consulting.
and ~45 BL (12 Bgal) from CBM reservoirs. Beneficial reuse outside of Arthur, J.D., Langhus, B.G., P., C., 2005. Technical summary of oil & gas produced water
treatment technologies. Report Prepared by ALL Consulting for the National Energy
the energy sector would favor irrigation, the largest water user. Irriga- Technology Lab. (NETL) and the Dept. of Energy (DOE).
tion exceeds PW from UOG reservoirs by ~5× and PW from CBM reser- Barnes, C.M., Marshall, R., Mason, J., Skodack, D., DeFosse, G., Smigh, D.G., et al., 2015. The
voirs by ~50×. Reuse of PW for HF water demand within the energy New Reality of Hydraulic Fracturing: Treating Produced Water is Cheaper than Using
Fresh. SPE-174956-MS. Society of Petroleum Engineering 29 p.
sector would reduce the number of UOG plays with excess PW by Blondes, M.S., Gans, K.D., Engle, M.A., Kharaka, Y.F., Reidy, M.E., Sarraswathula, V., et al.,
about half. PW requires intensive treatment for reuse outside of energy. 2017. U.S. Geological Survey National Produced Waters Geochemical Database v2.3
If we assume an average recovery factor from treatment of ~50%, then (Provisional). https://energy.usgs.gov/EnvironmentalAspects/
EnvironmentalAspectsofEnergyProductionandUse/ProducedWaters.aspx#3822349-
the ratio of irrigation to PW would be doubled. Considering all of data.
these factors, plays with potential for PW reuse for irrigation include Brinck, E.L., Drever, J.I., Frost, C.D., 2008. The geochemical evolution of water coproduced
Oklahoma AOI, Barnett, and the Permian Delaware Basin. PW volumes with coalbed natural gas in the Powder River Basin, Wyoming. Environ. Geosci. 15,
153–171.
in the Permian Delaware Basin would represent ~5% of the irrigation de-
Chaudhary, B.K., Sabie, R., Engle, M.A., Xu, P., Willman, S., Carroll, K.C., 2019. Spatial vari-
mand. PW projections relative to HF water demand would double in the ability of produced-water quality and alternative-source water analysis applied to the
Permian over the life of the play. Other sectoral users may provide local Permian Basin, USA. Hydrogeol. J. 27, 2889–2905.
Clark, C., Veil, J., 2009. Produced water volumes and management practices in the United
potential for reuse, including municipal, livestock and industrial uses
States. Rep. ANL/EVS/R-09/1, Argonne Natl. Lab., Argonne, Ill.
but are generally limited based on volumetric analysis. Discharging Danforth, C., McPartland, J., Blotevogel, J., Coleman, N., Devlin, D., Olsgard, M., et al., 2019.
PW to surface water is also being considered, but the PW would not Alternative management of oil and gas produced water requires more research on its
be substantially diluted in western streams. Recharging depleted aqui- hazards and risks. Integr. Environ. Assess. Manag. 15, 677–682.
Danforth, C., Chiu, W.A., Rusyn, I., Schultz, K., Bolden, A., Kwiatkowski, C., et al., 2020. An
fers is also a consideration; however, water quality issues may preclude integrative method for identification and prioritization of constituents of concern in
this option. produced water from onshore oil and gas extraction. Environment Intl 134, 677–682.
12 B.R. Scanlon et al. / Science of the Total Environment 717 (2020) 137085

Dieter, C.A., Maupin, M.A., Caldwell, R.R., Harris, M.A., Ivahnenko, T.I., Lovelace, J.K., et al., flowback water from Marcellus Shale. Environmental Science & Technology Letters
2018. Estimated Use of Water in the United States in 2015. 1441. U.S. Geological Sur- 1, 204–208.
vey Circular 65 p. Nichols, K., Sawyer, J., Bruening, J., Halldorson, B., Madhavan, K., 2017. Development of a
Dolan, F.C., Cath, T.Y., Hogue, T.S., 2018. Assessing the feasibility of using produced water large scale water recycling program for the Delaware Basin, New Mexico. Soc. of Pet-
for irrigation in Colorado. Sci. Total Environ. 640, 619–628. rol. Engin., SPE-186086-MS 12 p.
Engle, M.A., Reyes, F.R., Varonka, M.S., Orem, W.H., Ma, L., Lanno, A.J., et al., 2016. Geo- Nicot, J.P., Scanlon, B.R., Reedy, R.C., Costley, R.A., 2014. Source and fate of hydraulic frac-
chemistry of formation waters from the Wolfcamp and “Cline” shales: insights into turing water in the Barnett shale: a historical perspective. Environmental Science &
brine origin, reservoir connectivity, and fluid flow in the Permian Basin, USA. Chem. Technology 48, 2464–2471.
Geol. 425, 76–92. Nicot, J.P., Gherabati, A., Darvari, R., Mickler, P., 2018. Salinity reversal and water
Ertel, D.J., Bogdan, J.J., 2017. A Sustainable Choice for Unconventional Oil and Gas Waste- freshening in the Eagle Ford Shale, Texas, USA. ACS Earth and Space Chemistry 2,
water Management/Treatment When Options are Limited (EM, August RO). 1087–1094.
Ferguson, G., McIntosh, J.C., Perrone, D., Jasechko, S., 2018. Competition for shrinking win- NRC, 2009. Science and Decisions: Advancing Risk Assessment. National Research Council,
dow of low salinity groundwater. Environ. Res. Lett. 13. National Acadmey Press, Washington DC, p. 422.
Graham, E.J.S., Jakle, A.C., Martin, F.D., 2015. Reuse of oil and gas produced water in south- Oetjen, K., Giddings, C.G.S., McLaughlin, M., Nell, M., Blotevogel, J., Helbling, D.E., et al.,
eastern New Mexico: resource assessment, treatment processes, and policy. Water 2017. Emerging analytical methods for the characterization and quantification of or-
Int. 40, 809–823. ganic contaminants in flowback and produced water. Trends in Environmental Ana-
Greaves, R., Hartstein, R., Lincicome, D., Beck, P., Boothe, M., Olson, K.E., 2017. Fresh water lytical Chemistry 15, 12–23.
neutral: managing water use and giving back to the environment. SPE Annual Tech- Orem, W., Tatu, C., Varonka, M., Lerch, H., Bates, A., Engle, M., et al., 2014. Organic sub-
nical Conference and Exhibition. Society of Petroleum Engineers, San Antonio, Texas, stances in produced and formation water from unconventional natural gas extraction
USA, p. 27. in coal and shale. Int. J. Coal Geol. 126, 20–31.
Guerra, K., Dahm, K., Dundorf, S., 2011. Oil and gas produced water management and ben- Pashin, J.C., McIntyre-Redden, M.R., Mann, S.D., Kopaska-Merkel, D.C., Varonka, M., Orem,
eficial use in the western United States. Bureau of Reclamation Science and Technol- W., 2014. Relationships between water and gas chemistry in mature coalbed meth-
ogy Report No. 157 113 p. ane reservoirs of the Black Warrior Basin. Int. J. Coal Geol. 126, 92–105.
GWPC, 2019. Produced Water Report: Regulations, Current Practices, and Research Phan, T.T., Capo, R.C., Stewart, B.W., Macpherson, G.L., Rowan, E.L., Hammack, R.W., 2016.
Needs. Groundwater Protection Council (GWPC) 310 p. Factors controlling Li concentration and isotopic composition in formation waters
Hanor, J.S., 1994. Origin of saline fluids in sedimentary basins. In: Parnell, J. (Ed.), and host rocks of Marcellus Shale, Appalachian Basin. Chem. Geol. 420, 162–179.
Geofluids: Origin, Migration and Evolution of Fluids in Sedimentary Basins. 78. Geo- Rassenfoss, S., 2011. From flowback to fracturing: water recycling grows in the Marcellus
logical Society Special Publications, pp. 151–178. Shale. SPE Journal of Petroleum Technology 63 (7), 48–51.
Hayden, M., 2017. Local potash dipping toes into water business. Carlsbad Current Argus, Reig, P., Luo, T., Proctor, J.N., 2014. Global Shale Gas Development: Water Availability and
June 16, 2017. Business Risks. World Resources Institute www.wri.org Jan 2013, 80 p.
Healy, R.W., Bartos, T.T., Rice, C.A., McKinley, M.P., Smith, B.D., 2011. Groundwater chem- Rowan, E.L., Engle, M.A., Kraemer, T.F., Schroeder, K.T., Hammack, R.W., Doughten, M.W.,
istry near an impoundment for produced water, Powder River Basin, Wyoming, USA. 2015. Geochemical and isotopic evolution of water produced from Middle Devonian
J. Hydrol. 403, 37–48. Marcellus shale gas wells, Appalachian basin, Pennsylvania. AAPG Bull. 99, 181–206.
Hickenbottom, K.L., Hancock, N.T., Hutchings, N.R., Appleton, E.W., Beaudry, E.G., Xu, P., et Scanlon, B.R., Faunt, C.C., Longuevergne, L., Reedy, R.C., Alley, W.M., McGuire, V.L., et al.,
al., 2013. Forward osmosis treatment of drilling mud and fracturing wastewater from 2012. Groundwater depletion and sustainability of irrigation in the US High Plains
oil and gas operations. Desalination 312, 60–66. and Central Valley. Proc. of Natl. Acad. of Sciences 109, 9320–9325.
Horner, R.M., Harto, C.B., Jackson, R.B., Lowry, E.R., Brandt, A.R., Yeskoo, T.W., et al., 2016. Scanlon, B.R., Reedy, R.C., Male, F., Hove, M., 2016. Managing the increasing water foot-
Water use and management in the Bakken Shale oil play in North Dakota. Environ- print of hydraulic fracturing in the Bakken Play, United States. Environmental Science
mental Science & Technology 50, 3275–3282. & Technology 50, 10273–10281.
Huang, Y., Scanlon, B.R., Nicot, J.P., Reedy, R.C., Dutton, A.R., Kelley, V.A., et al., 2012. Scanlon, B.R., Reedy, R.C., Male, F., Walsh, M., 2017. Water issues related to transitioning
Sources of groundwater pumpage in a layered aquifer system in the Upper Gulf from conventional to unconventional oil production in the Permian Basin. Environ-
Coastal Plain, USA. Hydrogeol. J. 20, 783–796. mental Science & Technology 51, 10903–10912.
Igunnu, E.T., Chen, G.Z., 2014. Produced water treatment technologies. J. Low Carbon Scanlon, B.R., Weingarten, M.B., Murray, K.E., Reedy, R.C., 2019. Managing basin-scale fluid
Technol. 9, 157–177. budgets to reduce injection-induced seismicity from the recent U.S. shale oil revolu-
Khan, N.A., Engle, M., Dungan, B., Holguin, F.O., Xu, P., Carroll, K.C., 2016. Volatile-organic tion. Seismol. Res. Lett. 90, 171–182.
molecular characterization of shale-oil produced water from the Permian Basin. Scanlon, B.R., Ikonnikova, S., Yang, Q., Reedy, R.C., 2020. Will water issues constrain oil and
Chemosphere 148, 126–136. gas production in the U.S.? Env. Science and Technol. https://doi.org/10.1021/acs.
Kondash, A.J., Albright, E., Vengosh, A., 2017. Quantity of flowback and produced waters est.9b06390.
from unconventional oil and gas exploration. Sci. Total Environ. 574, 314–321. Scanlon, B.R., Reedy, R.C., Xu, P., Engle, M., Nicot, J.P., Yoxtheimer, D., et al., 2020. Datasets
Kurz, B.A., Stepan, D.J., Glazewski, K.A., Stevens, B.G., Döll, T.E., Kovacevich, J.T., et al., 2016. Associated With Investigating the Potential for Beneficial Reuse of Produced Water
A review of Bakken water management practices and potential outlook. Final Con- From Oil and Gas Extraction Outside of the Energy Sector (Data in Brief submitted).
tract Rept. Prepared for the Members of the Bakken Production Optimization Pro- Shaffer, D.L., Arias Chavez, L.H., Ben-Sasson, M., Romero-Vargas Castrillon, S., Yip, N.Y.,
gram. Univ. of N. Dakota Energy adn Environmental Research Center 42 p. Elimelech, M., 2013. Desalination and reuse of high-salinity shale gas produced
Landis, M.S., Kamal, A.S., Kovalcik, K.D., Croghan, C., Norris, G.A., Bergdale, A., 2016. The water: drivers, technologies, and future directions. Environ Sci Technol 47,
impact of commercially treated oil and gas produced water discharges on bromide 9569–9583.
concentrations and modeled brominated trihalomethane disinfection byproducts at Shuh WM. Water Appropriation Requirements, Current Water Use, & Water Availability
two downstream municipal drinking water plants in the upper Allegheny River, for Energy Industries in North Dakota—A 2010 Summary: Response to House Bill
Pennsylvania, USA. Sci. Total Environ. 542, 505–520. 1322, Section 2 of the 61st Legislative Assembly of North Dakota, Water Resources In-
Lauer, N.E., Harkness, J.S., Vengosh, A., 2016. Brine spills associated with unconventional vestigation No. 49, North Dakota State Water Commission, August 2010, www.swc.
oil development in North Dakota. Environmental Science & Technology 50, state.nd.us/4dlink9/4dcgi/GetContentPDF/PB-1800/W&E%20RPT%20FinalR.pdf
5389–5397. (accessed 2014). 2010.
Lemons, C., McDuid, G., Smye, K.M., Acevedo, J.P., Hennings, P.H., Banerji, D.A., et al., 2019. Stewart, B.W., Chapman, E.C., Capo, R.C., Johnson, J.D., Graney, J.R., Kirby, C.S., et al., 2015.
Spatiotemporal and stratigraphic trends in salt-water disposal practices of the Perm- Origin of brines, salts and carbonate from shales of the Marcellus Formation: evi-
ian Basin, Texas and New Mexico, United States. Env. Geosciences 26 (4), 107–124. dence from geochemical and Sr isotope study of sequentially extracted fluids. Appl.
Luek, J.L., Gonsior, M., 2017. Organic compounds in hydraulic fracturing fluids and waste- Geochem. 60, 78–88.
waters: a review. Water Res. 123, 536–548. Tasker, T.L., Burgos, W.D., Ajemigbitse, M.A., Lauer, N.E., Gusa, A.V., Kuatbek, M., et al.,
Macedonio, F., Ali, A., Poerio, T., El-Sayed, E., Drioli, E., Abdel-Jawad, M., 2014. Direct con- 2019. Accuracy of methods for reporting inorganic element concentrations and
tact membrane distillation for treatment of oilfield produced water. Sep. Purif. radioactivity in oil and gas wastewaters from the Appalachian Basin, US based on
Technol. 126, 69–81. an inter-laboratory comparison. Environmental Science-Processes & Impacts 21,
McMahon, B., Mackay, B., Mirakyan, A., 2015. First 100% reuse of Bakken produced water 224–241.
in hybrid treatments using inexpensive polysaccharide gelling agents. Society of Pe- Torres, L., Yadav, O.P., Khan, E., 2016. A review on risk assessment techniques for hydrau-
troleum Engineers, SPE-173783-MS. Society of Petroleum Engineers . https://doi.org/ lic fracturing water and produced water management implemented in onshore un-
10.2118/173783-MS https://www.onepetro.org/conference-paper/SPE-173783-MS. conventional oil and gas production. Sci. Total Environ. 539, 478–493.
Meng, Q.M., 2017. The impacts of fracking on the environment: a total environmental USEPA. EPA signs MOU with New Mexico to explore wastewater reuse options in oil and nat-
study paradigm. Sci. Total Environ. 580, 953–957. ural gas industry. [accessed 2018 Jul 30]. https://www.epa.gov/newsreleases/ 2018.
Mueller, D., 2017. Water Management Associated With Oil and Gas Development and Veil, J., 2015. U.S. produced water volumes and management practices in 2012. Report
Production. EM, August 2017. . Prepared for the Groundwater Protection Council, April 2015.
NASEM, 2010. Management and Effects of Coalbed Methane Produced Water in the West- Walker, E.L., Anderson, A.M., Read, L.K., Hogue, T.S., 2017. Water use for hydraulic fractur-
ern United States. National Academies of Sciences, Engineering, and Medicine ing of oil and gas in the South Platte River Basin, Colorado. J. Am. Water Resour. Assoc.
(NASEM), National Acadmey Press., Washington DC. 53, 839–853.
Nasiri, M., Jafari, I., Parniankhoy, B., 2017. Oil and gas produced water management: a re- Walsh, F.R., Zoback, M.D., 2015. Oklahoma’s recent earthquakes and saltwater disposal.
view of treatment technologies, challenges, and opportunities. Chem. Eng. Commun. Sci. Adv. 1.
204, 990–1005. Xu, P., Cath, T.Y., Robertson, A.P., Reinhard, M., Leckie, J.O., Drewes, J.E., 2013. Critical re-
Nelson, A.W., May, D., Knight, A.W., Eitrheim, E.S., Mehrhoff, M., Shannon, R., et al., 2014. view of desalination concentrate management, treatment and beneficial use. Environ.
Matrix complications in the determination of radium levels in hydraulic fracturing Eng. Sci. 30.

You might also like