Chemosphere 169 (2017) 627e635

Contents lists available at ScienceDirect

Chemosphere
journal homepage: www.elsevier.com/locate/chemosphere

Holistic risk assessment of surface water contamination due to Pb-210

in oil produced water from the Bakken Shale
Luisa Torres a, Om Prakash Yadav b, Eakalak Khan a, *
a
b

Department of Civil and Environmental Engineering, North Dakota State University, Fargo, ND 58108, USA
Department of Industrial and Manufacturing Engineering, North Dakota State University, Fargo, ND 58108, USA

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

A risk assessment process combining

technical and social aspects is
developed.

Social risk perceptions of different
stakeholders vary based on several
factors.

Risk of surface water contamination
by Pb-210 in produced water is not
negligible.

By including social risk perception,
the risk score increased by 25%e50%.

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 22 September 2016
Received in revised form
19 November 2016
Accepted 25 November 2016

A holistic risk assessment of surface water (SW) contamination due to lead-210 (Pb-210) in oil produced
water (PW) from the Bakken Shale in North Dakota (ND) was conducted. Pb-210 is a relatively long-lived
radionuclide and very mobile in water. Because of limited data on Pb-210, a simulation model was
developed to determine its concentration based on its parent radium-226 and historical total dissolved
solids levels in PW. Scenarios where PW spills could reach SW were analyzed by applying the four steps
of the risk assessment process. These scenarios are: (1) storage tank overow, (2) leakage in equipment,
and (3) spills related to trucks used to transport PW. Furthermore, a survey was conducted in ND to
quantify the risk perception of PW from different stakeholders. Findings from the study include a low
probability of a PW spill reaching SW and simulated concentration of Pb-210 in drinking water higher
than the recommended value established by the World Health Organization. Also, after including the
results from the risk perception survey, the assessment indicates that the risk of contamination of the
three scenarios evaluated is between medium-high to high.
2016 Elsevier Ltd. All rights reserved.

Handling Editor: Caroline Gaus

Keywords:
Holistic risk assessment
Lead-210
Simulation model
Social risk perception
Surface water contamination
Unconventional onshore oil production

1. Introduction
Oil produced water (PW) generated during the owback process

* Corresponding author. Department of Civil and Environmental Engineering,

North Dakota State University, Dept. # 2470, P.O. Box 6050, Fargo, ND 58108-6050,
USA.
E-mail address: [email protected] (E. Khan).
http://dx.doi.org/10.1016/j.chemosphere.2016.11.125
0045-6535/ 2016 Elsevier Ltd. All rights reserved.

and production stage contains different constituents including

chemical additives, salt, oil and grease, natural organic and inorganic compounds, and natural radioactive materials. Leakage, spills,
and discharges of PW present a threat to human health and the
environment because of the relatively high concentrations of these
contaminants (Guerra et al., 2011; Vengosh et al., 2014; Vidic et al.,
2013). Research on water contamination due to oil and gas production has been conducted on specic constituents, such as

628

L. Torres et al. / Chemosphere 169 (2017) 627e635

salinity and organic compounds, but studies about radioactive

materials are still scarce (Barbot et al., 2013; Nelson et al., 2016;
Paschoa, 1998). This is especially true for North Dakota (ND), the
second largest oil producing State in the United States (Argonne
National Laboratory, 2014; Scanlon et al., 2014).
Naturally occurring radioactive material (NORM) is found in
geologic formations and distributed naturally while the technologically enhanced NORM (TENORM) is the result of human activities, such as oil and gas (O&G) production, which transport the
radionuclides and increase the concentrations and the probability
of exposure (Rich and Crosby, 2013). Some underground geological
reservoirs exploited to produce O&G in the U.S. contain uranium238 (U-238) and thorium-232 (Th-232) which are sources for
TENORM (Argonne National Laboratory, 2014).
Uranium-238 is naturally found in the environment and its
decay series includes lead-210 (Pb-210) along with radon-222 (Rn222) and polonium-210. Stable isotopes of lead are found in both
terrestrial and freshwater ecosystems but their concentrations vary
based on the location (Tamponnet, 2009). Once the radioactive
material reaches the environment, there are different pathways to
human exposure such as surface and shallow groundwater
contamination. The fate and transport of Pb-210 in the environment have not been studied as widely as other radionuclides
(Argonne National Laboratory, 2014; Nelson et al., 2016;
Tamponnet, 2009).
Previous TENORM studies have focused mostly on radium-226
(Ra-226), Ra-228, and Rn-222, which are the predominant radioisotopes present in natural gas production liquid and solid wastes
(Godoy and Petinatti da Cruz, 2003; Rich and Crosby, 2013; Rowan
et al., 2011). The radioactivity levels in PW vary from background
values to thousands of picocuries per liter (pCi/L) and usually Ra226 is found in high concentrations (Chriss and Bursh, 2002).
Because of this, State regulations typically are based upon total
radium (Ra-226 plus Ra-228) (Argonne National Laboratory, 2014).
Previous research has reported Pb-210 and other radionuclides in
O&G production wastes. For example, a concentration of
1.7 1.2 pCi/g of Pb-210 in soil and water (sludge) collected from a
reserve pit in Texas was observed (Rich and Crosby, 2013) and up to
5135 pCi/L of Pb-210 in produced water and up to 35,000 pCi/g of
Pb-210 in sludge were reported (International Atomic Energy
Agency, 2003). High concentrations of Pb-210 in O&G production
waste increase the risk of human exposure. Very little research
related to Pb-210 in PW has been conducted despite that lead is
known for its harmful effects, especially on children, such as
damage to the brain and nervous system, slowed growth and
development, learning and behavior problems, and hearing and
speech problems (Center for Disease Control and Prevention, 2013;
Rich and Crosby, 2013).
Produced water quality varies according to several factors
including geological formation, location, depth, stimulation
method, and the chemicals applied. According to previous research,
there is a strong correlation between total dissolved solids (TDS)
concentration and radium activity in PW from reservoirs with
similar lithological characteristics (Fisher, 1998; Godoy and
Petinatti da Cruz, 2003; Nelson et al., 2016; Rowan et al., 2011).
Fisher (1998) found that TDS levels higher than ~35,000 mg/L result
in high radium activity of more than ~100 pCi/L. Produced water
from the Williston Basin, where the Bakken is located, is known to
have some of the highest TDS concentrations in the U.S. ranging
from a few hundreds to 632,700 mg/L (Gleason and Tangen, 2014).
Based on Fisher (1998), radioactivity in PW from the Bakken should
be high due to elevated salinity.
In 2014, the Argonne National Laboratory published a report
entitled Radiological Dose and Risk Assessment of Landll
Disposal of Technologically Enhanced Naturally Occurring

Radioactive Materials (TENORM) in North Dakota in response to a

request by the North Dakota Department of Health (ND DoH) to
support possible changes in the regulations regarding TENORM
(Argonne National Laboratory, 2014). Based on this report, regulations on solid waste could be safely modied as long as the
maximum exposure to landll workers does not exceed 100 mrem/
yr (Argonne National Laboratory, 2014). The report covers the human health risks of Ra-226, Ra-228, Pb-210, and Th-232 in solid
wastes but excludes PW.
This research focuses on assessing the risks of surface water
(SW) contamination caused by spills of PW from the Bakken formation containing Pb-210. To achieve this, historical data on PW
from ND was analyzed to simulate the radiation levels coming from
Pb-210 because this information is not publicly available. With this
data, a holistic risk assessment of SW contamination due to Pb-210
was performed considering three different scenarios: (1) storage
tank overow, (2) leakage in equipment, and (3) spills related to
trucks used to transport PW. Simulation was applied in this study
by using Palisade's decision tool @Risk and the Canadian Centre
for Environmental Modelling and Chemistry's Quantitative Water
Air Sediment Interaction (QWASI) model. In addition, social risk
perception and awareness of PW water risks of different stakeholders in ND were measured. The objectives of this study are:

To simulate average Pb-210 concentration in PW from the
Bakken.

To quantify the risks of SW contamination due to Pb-210 under
three different scenarios.

To perform a holistic risk assessment by incorporating the results of a survey that measures risk awareness and perception of
the issue.

2. Methodology
2.1. Method overview
A risk assessment focused on three scenarios that could result in
SW contamination was conducted. The rst scenario is storage tank
overow that reaches SW that is nearby the oil well pads. The
second scenario is leakage in pipes, storage units, and other
equipment, that reaches SW. The third scenario is accidents related
to trucks transporting PW which cause spills that reach SW. These
scenarios were selected because they were found to be common in
reported spills and media coverage in ND (Guerin, 2015; Kusnetz,
2012; ND DoH, 2015). Other scenarios that could contribute to
SW contamination such as runoff and melting snow were not
considered because of supporting data unavailability and high
variability of environmental conditions.
This research uses a combination of stochastic and deterministic
methods to analyze the risks of Pb-210 in each scenario. These
methods were combined because of the variability in the type of
data used. First, the historical TDS data, reported spills, and truck
crashes were analyzed using Palisade's @Risk tool. Second, due to
extremely limited data, the Ra-226 level in the Bakken was calculated based on a generalized linear regression model between
observed TDS and Ra-226 levels from different locations including
Pennsylvania, Texas, New York, and Wyoming.
Third, the Rutherford-Soddy's decay law was used to simulate
the decay of Ra-226 into Pb-210. @Risk was used to determine the
type of probability distribution followed by the simulated values of
Ra-226 and the parameters that describe that probability distribution. The probability distribution parameters (i.e. shape and
scale) of Ra-226 were included in the estimation of Pb-210 because
they capture the variability of the non-normal distribution of Ra-

L. Torres et al. / Chemosphere 169 (2017) 627e635

226 data due to different factors which could not have been
considered if point estimates were used. The predicted Pb-210
concentration was used to assess the exposure by oral ingestion.
The exposure concentration was estimated using the QWASI model
that simulates chemical fate in lakes. Furthermore, a survey was
distributed in several counties in ND to collect data from different
stakeholders including general public, oil eld and truck operators,
and emergency management personnel and/or individuals in
charge of PW management (hereinafter experts). The survey
collected data on the knowledge, opinions, and concerns of each
stakeholder related to PW management. The data was analyzed to
quantify awareness and perception of the risks of each scenario
where PW could reach SW. The results from the survey were
included in the risk characterization of each scenario.
Finally, a semi-quantitative risk assessment matrix was used to
translate the results obtained in the previous steps of the risk
assessment into quantitative measurement based on low, mediumlow, medium-high, and high risk. Also, a sensitivity analysis was
carried out to evaluate the variation in the Pb-210 levels if the
simulated Ra-226 concentrations were underestimated or overestimated by 25%e50% compared to actual levels. The entire
methodology is summarized in Fig. 1.

629

regarding truck crashes in the State (Figs. S1 and S2, SM) (ND DoT,
2015).
2.3. Determination of Pb-210 concentrations
There were two steps in determining Pb-210 concentration. Ra226 concentration was estimated and then Pb-210 was calculated
based on the decay of Ra-226. In ND, data on radium in PW has not
been reported in the USGS PW database and to our knowledge, only
one study by Lauer et al. (2016) has measured Ra-226 in PW from
the Bakken but the data was based on only three samples. Because
of this, it is necessary to use a simulation approach to determine the
radium concentration. Since there is not enough data on Ra-226
from the Bakken, three different datasets from previous studies
were analyzed (Table S3a, SM). The rst dataset, which comes from
Pennsylvania and New York, was obtained from Rowan et al., 2011.
The second set, which is from Texas, was found in Fisher (1995).
Lastly, the third set, which is from Pennsylvania and Wyoming, was
retrieved from the USGS PW database for the wells that were
monitored for TDS and Ra-226 levels. Using a total of 331 data
points from the three datasets, a linear regression model with an Rsquared value of 0.80 (Tables S3b and S3c, SM) was developed as
follows:

2.2. Sources of data

Log(226Ra) 1.12  Log(TDS)  3.33
The chemical quality, particularly TDS, of PW in the Bakken was
obtained from the U.S. Geological Survey Produced Water Database
(hereinafter USGS PW database) (Blondes et al., 2014). Please see
Table S1 in Supplementary Material (SM) for sample TDS data. Also,
the environmental incident reports available at the ND DoH website were used (ND DoH, 2015). This dataset is a collection of reported oileld environmental incidents that caused any discharge
of liquids or solids that might cause pollution of water bodies. The
information used in this study is associated with saltwater spills
from October 7, 2014 to October 5, 2015 (Table S2, SM). In addition,
the ND Department of Transportation (ND DoT) report called 2014
North Dakota Crash Summary was analyzed to collect information

This generalized model (hereinafter the Model) was used to

simulate the Ra-226 levels for the Bakken. By using the data from
four different locations, the Model is more representative than a
regression model created from only one dataset because the Model
captures variability introduced by several factors from different
locations and hence provides a more generalized pattern. It is
important to note that the results obtained with the Model are
estimates since its parameters were calculated from historical data
and for this reason, treated as random variables. The observed Ra226 data in ND by Lauer et al. (2016), which is extremely limited
(three PW samples), will be used to validate the Model results.

Perform risk perception

survey

Perform risk assessment

Hazard identification
Hazard quantification
Exposure assessment
Risk characterization

Analyze each dataset

Predict Pb-210
concentration based on
Ra-226 distribution and
decay law

Sensitiviy analysis

Calculate Ra-226 levels

based on TDS
(correlation)

Simulate Ra-226
probability distribution

Analysis of results

Collect data

(1)

Fig. 1. Process followed to perform a holistic risk assessment of Pb-210 surface water contamination.

630

L. Torres et al. / Chemosphere 169 (2017) 627e635

According to the Rutherford-Soddy's decay law, the atoms of a

radioactive substance change randomly from a parent element to
another daughter element at a specic rate (Braun, 1983). The decay
constant l indicates how fast the substance decays or disintegrates
which can be measured by the half-life (t1=2 ) of the isotope, or the
time it takes for half of the original quantity of isotope to decay
(Braun, 1983). The fraction of atoms that disintegrate after time t
can be expressed as:
N(t) N0elt

(2)

Where N0 is the amount of atoms at t 0 and l tln2

1=2
Depending on the activity relationship between the parent
compound and the daughter product, there can be no equilibrium,
transient equilibrium, or secular equilibrium (Prince, 1979). When
Ra-226 decays, it does not disappear but changes into new nuclear
species including Pb-210. Because the half-life of Ra-226 (1600
years) is much longer than the half-life of Pb-210 (22 year), they are
in secular equilibrium (Prince, 1979). During the lifespan of an oil
well in the Bakken, which is approximately 30 years, there is no
signicant change in Ra-226 and it is constantly feeding Pb-210
(Braun, 1983; North Dakota Petroleum Council, 2012). However,
Ra-226 is unsupported because its parent, U-238, remains in the
reservoir (Smith, 2011). Because of the signicant difference between the half-lives of Ra-226 and Pb-210, the activities of both are
the same over time meaning that the Pb-210 concentration at time t
Nt can be determined using the concentration of Ra-226 at t 0
N0 .
2.4. Social perception
Risk assessment from an engineering or technical point of view
alone might deliver biased results (Perry, 2012). This is because risk
measurement is usually narrowed down into probabilities and
consequences, leaving aside the public awareness of risk
(Leiserowitz, 2005; Slovic, 2000). Produced water spills are mostly
caused by human errors. As a result, the perception of risk can be a
critical factor for quantifying risk level. A holistic approach to
analyze risk incorporates risk perceptions which are created based
on different dimensions including technical knowledge, psychological and social factors, and personal experiences (Leiserowitz,
2005; Slovic, 2000). It is believed that risk perception inuences
risk behaviors, meaning that the more aware a person is about a
specic risk the more protective actions the person is willing to
take and vice versa (Ding et al., 2010; Slovic and Peters, 2006).
A survey was developed to collect risk perception of different
stakeholders in ND including the general public, operators, and
experts. The survey was used to obtain data regarding the knowledge, opinions, and concerns of different stakeholders affected
directly and indirectly by the oil production in the State. The survey
questions were focused on determining the risk awareness and
perception, and inuence of PW management on human behavior
or action. The survey was conducted completely online where the
participants were asked to answer 21 to 25 questions, depending
on their responses, in a time period of 5e10 min. The regions of
interest were Cass county (where Fargo, ND is located) and the 17
oil-producing counties in the State. These counties are: Billings,
Bottineau, Bowman, Burke, Divide, Dunn, Emmons, Golden Valley,
McHenry, McKenzie, McLean, Mountrail, Renville, Slope, Stark,
Ward, and Williams (North Dakota Petroleum Council, 2015).
2.5. Holistic risk assessment
The steps followed for the assessment include hazard

identication, hazard quantication, exposure assessment, and

nally, risk characterization. Hazard identication set the relation between scenarios where PW could be discharged and the
initiating events or hazard that could cause those scenarios.
Hazard quantication established the probability of each scenario occurring. The exposure routes to Pb-210 can be through
oral ingestion, dermal contact or inhalation (Emmanuel et al.,
2007). In general, about 80% of the Pb-210 internal exposure in
humans is through ingestion and 20% through inhalation but this
can vary depending on the location (Jaworowski, 1969;
Tamponnet, 2009).
When Pb-210 reaches a lake, several interactions, which change
the initial amount of contaminant, occur. A lake system consists of
three different compartments which are water, sediments, and air
(Mackay and Diamond, 1989). Environmental fate models have
been developed and used for the past decades to understand the
fate and transport potential of chemicals in each of the compartments (Woodne et al., 2000). For this analysis, the QWASI
spreadsheet model version 1.00 was used due to its simplicity to
illustrate the variation of Pb-210 concentration once it reaches the
environment.
The QWASI model uses the fugacity concept, which is suitable
for chemicals with measurable vapor pressure, but since metals,
such as lead, have zero or negligible vapor pressure some modications are required (Mackay et al., 2014). As a basic explanation,
fugacity is the tendency of a chemical of escaping from the phase it
is in (i.e. solid, liquid, or gas) (Valsaraj and Melvin, 2000). The
processes simulated with QWASI include advective ow, sediment
deposition, resuspension, burial, and sediment-water diffusion
(Woodne et al., 2000). Details on the fugacity model and the
equations used in QWASI are available in Mackay et al. (2014),
Webster et al. (2005), Woodne et al. (2000), and Mackay and
Diamond (1989).
In the risk characterization, the data obtained in the rst three
steps was linked and the social risk perceptions and awareness
about PW from different stakeholders were included.

3. Results and discussion

3.1. Analysis of data
The historical data of TDS (n 7907) in the Bakken was
analyzed using Palisade's @Risk which showed that the data
follows a distribution with a mean equal to 264,512 mg/L and a
standard deviation of 76,754 mg/L. From all the spills that
occurred from October 7, 2014 to October 5, 2015, there were 609
contained and 112 uncontained spills and the volume ranged from
42 gallons to 2,940,000 gallons (1e70,000 barrels) (ND DoH,
2015). The average volume for uncontained spills is 2757 gallons
without taking into account a major incident which released
almost 3 million gallons of PW in Williams County in January of
2015. This incident was treated as an extremely large rare event
and thus, not included to avoid the inuence of extreme rare
events or outliers.
From 2009 to 2014, there were 3283 crashes involving trucks
and 6029 crashes involving truck tractors with average annual increments of 17.4% and 16.6%, respectively (Fig. S1 in SM) (ND DoT,
2015). These numbers include other accidents besides those associated with PW transportation. The ND DoT was contacted to obtain
data regarding the distribution of these accidents in different
Counties in the State. From the total amount of crashes occurred in
the State in 2014, 5.5% involved truck units and more than 60% of
these accidents took place in the major oil-producing counties in
the State (Fig. S2 in SM).

L. Torres et al. / Chemosphere 169 (2017) 627e635

3.2. Pb-210 simulation

The TDS data points obtained (n 7907) from the Bakken were
plugged in the Model to calculate the Ra-226 concentrations
(n 7907) and then, this dataset was tted using @Risk best t
tool (Table S4, SM). The results from the analysis indicate that the
simulated Ra-226 data follows a distribution with a mean equal to
557 pCi/L and a standard deviation of 179 pCi/L (Fig. S3, SM). The
average Ra-226 value falls within the range (527-1210 pCi/L) found
by Lauer et al. (2016) based on three PW samples in ND. Because the
mean and the standard deviation capture the Ra-226 variability in
our data, it was important to include these parameters in the
Rutherford-Soddy's decay law to calculate the Pb-210 concentration. For this analysis, the initial Pb-210 concentration was assumed
to be equal to the Ra-226 concentration N0 226 Ra due to secular
equilibrium and after one year of PW generation t 1, half-life of
22 years for Pb-210, and l ln2
22 then, equation (2) is now:
210

Pb

Ra  el(1)

226

(3)

Equation (3) was used to simulate Pb-210 in @Risk (Figs. S4 and

S5, SM) which estimated an average concentration of 540 pCi/L. The
concentration is the same for the three scenarios under evaluation
assuming all PW spills had the same chemical composition. In
addition, the @Risk tool was used to create a probability distribution (Fig. S6, SM) of Pb-210 using the results from 100,000 Monte
Carlo iterations and 1000 bootstrap samples. The Monte Carlo
method and bootstrapping were used to obtain better estimations.
3.3. Risk assessment
The rst step is hazard identication. Hazard events identied
for each scenario under evaluation are:

Storage tank overow: Storage capacity limitation, human error,
and heavy rain or storm.

Equipment leakage: Equipment failure and human error.

Truck accident during transportation of PW: Insufcient storage
capacity, human error, road conditions, and weather conditions.
Incident reports available on the ND DoH website were used for
hazard quantication. From all the incidents related to PW spills,
contained and uncontained, reported on the website from October
7, 2014 to October 5, 2015, 70% was due to leakage, 15% was due to
storage unit overow, 9% was unspecied incident, and 7% was
related to truck incidents (ND DoH, 2015). The leakage incidents
were mostly caused by equipment failure including pipeline
integrity, valve/piping connections, stufng box, and treater leaks.
Only one of the spill reports indicated that the PW spill was caused
by a truck rolling into a ditch while the rest of the spills classied as
truck incident were caused by truck overow and operator error.
Based on the ND DoH reports, the total number of spills that
reached SW was very low. Less than 3% indicated that the spilled
water was in contact with SW. From this fraction of spills, less than
6% was caused by storage tank overow, almost 89% by leakage, and
6% by truck incident. These numbers are based on historical data
and should not be treated as predictions. The average volume of PW
that reached a water body in each scenario was 420 gallons due to
storage unit overow, 23,562 gallons due to leakage (without taking into account the spill in Williams County in January 2015), and
1680 gallons by truck incident. However, it is important to clarify
that among the water bodies that were in contact with the PW were
small pools, rife streams, stock ponds, and slough water which are
not considered sources for human drinking water. It should be

631

noted that the reports do not indicate whether these small water
bodies are directly connected to drinking water sources. Only the
incident in Williams County in January 2015, where almost 3
million gallons were spilled, impacted a drinking water source. The
spill affected the Blacktail Creek and Little Muddy River which ow
into the Missouri River, a drinking water source (Lauer et al., 2016).
In the second step of the risk assessment, the exposure to Pb210 by oral ingestion was calculated based on the average initial
concentration of 540 pCi/L determined previously. The exposure
assessment was based on the scenario where PW is accidentally
spilled and reaches a SW body that serves as human drinking water
source. The actual exposure to Pb-210 is the concentration that
remains after the contaminant undergoes different chemical and
physical processes, which happen naturally, and after the SW is
treated by different methods.
For the exposure scenario where PW reaches a SW body, the
characteristics of Lake Sakakawea were used. This lake was chosen
because it is a major source of SW for several cities in ND including
Dickinson, Williston, Garrison, and Parshall (City of Williston
(2014); ND DoH, 1999), and because it is located close to the regions with high density of oil wells. Table S5 in SM summarizes the
inputs and the sources used in the simulation. Some required inputs were not found in the literature and in these cases the default
values given by QWASI were applied (Figs. S7 and S8, SM). In
addition, processes in the air compartment were ignored because
Pb atmospheric deposition in ND is not signicant (Fig. S9, SM) (U.S.
Environmental Protection Agency, 2011).
The results from QWASI indicated that the Pb-210 concentration
that remains in the water compartment is 1.81  103 ng/L or
approximately 140 pCi/L. For detailed results, please see Figs. S10a
and S10b in SM. Among methods to remove radioactive materials
from drinking water is lime softening, ion exchange, and activated
carbon (U.S. Environmental Protection Agency, 2015). The average
removal efciency of these three methods is approximately 88%
(U.S. Environmental Protection Agency, 2015) which means that the
nal concentration after treating the water is 17 pCi/L.
Becquerels per liter (Bq/L) is another unit for radioactivity and
since 1 pCi/L is equal to 0.037 Bq/L (Keith et al., 2012) then, 17 pCi/L
is equivalent to 0.63 Bq/L. For comparison, some of the maximum
allowable levels of lead in drinking water are 0 mg/L established by
the U.S. Environmental Protection Agency, 1 pCi/L or less based on
the National Interim Primary Drinking Water Regulations, 0.1 Bq/L
according to the World Health Organization (WHO), and 0.2 Bq/L in
Canada (Shammas and Wang, 2016; U.S. Environmental Protection
Agency, 2012; Velten and Jacobs, 1982; World Health Organization,
n.d.). The concentration of Pb-210 to which a person could be
exposed to in the simulated case exceeds these standards.
To characterize the risk of each scenario, a 4  4 risk matrix
(Food and Agriculture Organization of the United Nations, n.d.)
shown in Fig. 2 was used. The risk matrix, or C  L matrix, combines
consequences (C) and the probability or likelihood (L) of those
consequences happening. As seen in Table 1, C and L are divided
into four different levels. The L levels are remote, unlikely, possible,
and likely which are dened based on the probability of the event
occurring (Table 1) (Food and Agriculture Organization of the
United Nations, n.d.; Mosleh and Bari, 1998; Stouffer et al., 2015).
The four levels of C are minor, moderate, major, and extreme.
They are based on the maximum allowable concentration of the
chemical under study (Brereton and Alenbach, 1998; Stouffer et al.,
2015). For this risk assessment, the limit value used for Pb-210 in
drinking water was 0.1 Bq/L established by the WHO (n.d.). This
recommended value was used because a limit in terms of radioactivity has not been established in the U.S. (Velten and Jacobs,
1982). Table 1 describes the different consequence levels.
Each of the three scenarios that could result in SW

632

L. Torres et al. / Chemosphere 169 (2017) 627e635

a)

b)
4

12

16

12

Scenarios 1 and 3
Scenario 2

10

15

20

25

12

16

20

12

15

10

Scenario 2 =

Scenario 1 = 11
Low

Medium-high

Medium-low

High

Scenario 3 = 9

Fig. 2. Scores for each scenario using a: a) 4  4 risk matrix for technical risk and b) 5  5 risk matrix for combined technical and social risks.

Table 1
Measurements of risk likelihood and risk consequence or impact (Pb-10 concentration in contaminated water) used for the risk matrix.
Level

1
2
3
4

Likelihood

Consequence or Impact

Description

Characteristics

Description

Characteristics (Contaminant Level)

Remote
Unlikely
Possible
Likely

It has never been heard of but it is not impossible. Probability <2%.

Not expected to occur but it has been known to occur else- where. Probability of 2e10%.
Evidence suggests that it may occur in some circumstances. Probability of 10e35%.
It will probably occur. Regular and strong evidence. Probability of 40e100%.

Minor
Moderate
Major
Extreme

<0.1 Bq/L
0.1e10 Bq/L
10 - 100 Bq/L
>100 Bq/L

contamination due to Pb-210 was analyzed using the risk matrix.

The C is the same for the three scenarios in which people could be
in direct contact with Pb-210 by drinking water containing 0.63 Bq/
L. This concentration is 6 times higher than the guidance level
established by the WHO. This means that the C is level 2 or moderate impact. The L of each scenario happening (Table S6, SM) is
about 6% for storage unit spill, 89% for equipment leakage, and 6%
due to truck incident. By looking at Table 1, the L of scenario 1 is
considered unlikely (level 2), for scenario 2 is likely (level 4), and for
scenario 3 is unlikely (level 2).
After multiplying the C and L in each case, scenarios 1 and 3 have
a score of 4 while scenario 2 is 8. Based on Fig. 2, the risk can be
classied low if the score is 1e2 (green), medium-low if it is 3e4
(yellow), medium-high if it is 6, 8 or 9 (orange), and high if the score
is 12 or 16 (red) (Food and Agriculture Organization of the United
Nations, n.d.). In accordance with the numerical score and color
code, the risks of scenarios 1 and 3 are considered medium-low
while scenario 2 is classied as medium-high risk (Fig. 2a).

3.4. Social risk perception survey results

The survey was developed to capture the general risk perception
that oil PW from ND causes to different stakeholders including
general public, operators, and experts. The total number of respondents (self-selected) that completed the survey was 191 with a
19.4% dropout rate which is within the average range of 15%e30%
for online surveys (Galesic, 2006). From all the respondents, 35%

did not indicate their place of residence within the State, 46% lived
in Cass County while the remaining 19% were in other counties
including 11 oil-producing counties. The majority of participants
was 18e24 years old and had a 4-year college degree (38% and 39%,
respectively). For the purpose of this study, only 7 questions out of
25 were used because they were focused on measuring the level of
risk perception and awareness of PW. The remaining questions will
be analyzed and presented in a different study.
The participants were asked to rank their level of concern of
each scenario where PW could reach SW. For the storage tank
overow scenario, 41% responded to be not at all to slightly concerned. For the equipment failure and truck accident scenarios, the
majority indicated to be moderately to extremely concerned (57%
and 49%, respectively). Some questions in the survey were focused
on obtaining the opinion of experts in ND which represented 8% of

Table 2
Parameters used to calculate the holistic risk score of each scenario evaluated. The
technical risk scores were obtained from the 4  4 risk matrix. The perceived risk
scores were obtained from the survey and converted to PAF values.
Parameter

Scenario 1

Scenario 2

Scenario 3

Technical risk
Perceived risk (PAF):
Level of concern with PW
Level of familiarity with PW-MGMT
Level of awareness with PW-content
Experts' opinion-risk perception

1.625
2
1.625
1.75

0.5
2
1.625
1

0.5
2
1.625
1

L. Torres et al. / Chemosphere 169 (2017) 627e635

the total number of participants. The experts were asked to rank

the awareness of general public, oil eld operators, and truck operators regarding PW. The respondents indicated that the general
public is slightly aware (38%) while oil eld operators (56%) and
truck operators (44%) are extremely aware. The experts were also
asked to rank the level of risk perception of each stakeholder in
each scenario. According to the experts, the general public seems to
have the highest risk perception on each of the scenarios. On the
other hand, the oil eld operators and truck operators have lower
risk perceptions except on the scenarios where they are in direct
contact with PW, namely equipment and trucks, respectively.
The answers from all the participants were quantied in order
to integrate them with the results from the risk matrix in the
previous section. There are different methods to interpret raw data
from surveys including percent agree, rating and ranking average,
top box, and Z-score to percentile rank (MeasuringU, 2011;
SurveyMonkey, 2016). For the interpretation of our survey results,
a combination of percent agree and rating and ranking was applied
along with a modied version of the methodology proposed by

633

it was assumed that the less concern/aware/familiar a person was,

the higher the score. These survey scores were transformed into
PAF-values with a range of 0.5e2. For details, please see
Tables S7eS9 in SM. For example, most of the participants indicated
to be not familiar at all (survey score 5) with PW management
then, the PAF-value assigned was 2. Likewise, most of the people
indicated to be slightly aware (survey score 4) with the PW
content so the PAF-value in this case is 1.625. The opinion of experts
on risk perception of each stakeholder on each scenario was also
transformed into PAF-value and then the average was determined
(Table S10, SM). The transformed values for each risk level are 2,
1.25, and 0.5 for low, medium, and high, respectively. In scenario 1
the results were: general public e medium (1.25), oil eld operator
e low (2), and truck operator e low (2). In this case the average is
1.75. For scenarios 2 and 3 the average is 1. Table 2 summarizes the
values of each parameter for each scenario.
The next step was to calculate the risk of each scenario using the
information from Table 2. For scenario 1 the overall risk was
calculated as follows:



1:625  0:15 2  0:20 1:625  0:35 1:75  0:30
z11
Risksc1 4 4 
1

Plattner et al. (2006).

The model developed by Plattner et al. (2006) combines technical risk measurement and a weighted mean of different perception affecting factors (PAF). Equation (4) denes risk as the sum of
hazard (objective or technical risk) and outrage (perceived risk)
(Plattner et al., 2006; Sandman, 1999):
Risk hazard outrage

(4)

Where hazard (Rtech) is equal to C  L and outrage (Rperc) can be

calculated with the following equation (Plattner et al., 2006):
Pn
paf $a
Pn i i
(5)
Rperc Rtech  i1
a
i1 i

For this study, four PAF (pafi ) were considered, each one with an
assigned weight (ai ). The values the PAF can take vary from 0.5 to 2
so that the Rtech can be reduced by half, increased by 2 folds, or in
between. The weight assigned to each PAF ranges from 0 to 1 so that
the total sum of the weights equals 1. The weights were assigned
based on the importance of the survey question or PAF to capture
the risk perception of each stakeholder. The four PAF and their
respective weights are: 1) level of concern with PW e 15%; 2) level
of familiarity with PW management e 20%; 3) level of awareness
with PW content e 35%; and 4) opinion of experts on risk perceptions e 30%.
Likert scales were used in the survey to measure the level of
concern, awareness, and familiarity. Each scale has 5 levels and
each one has a score (i.e. 5, 4, 3, 2, 1). For the quantication process,

Using the same method, the calculated risks for scenario 2 and 3
are approximately 19 and 9, respectively (see the SM for further
detail). The results indicate that by adding the quantied social risk
perception the risk of each scenario incremented. Based on the
color code from a 5  5 matrix the risk is low if the score is 1e3
(green), medium-low if it is 4e6 (yellow), medium-high if it is 8e12
(orange), and high if the score is between 15 and 25 (Fig. 2b). Then,
scenarios 1 and 3 are medium-high risk and scenario 2 is considered high risk.
The data collected with the survey indicate that the public is not
getting enough objective information about PW coming from the
oil wells in ND. This has caused overestimated and biased risk
perceptions. According to the results, people overestimated the risk
of each scenario by 25%e75%. Poor social perception or lack of
awareness about risk involved poses more risk on people that do
not take preventive measures.
3.5. Sensitivity analysis
In this study, the main input to simulate Pb-210 concentration
was the Ra-226 initial concentration. If the real average level of
radium in PW was underestimated or overestimated, the average
concentration of Pb-210 would be different. To know how much the
output would change if actual Ra-226 level varied between 25% and
50%, a sensitivity analysis of four scenarios (Table 3) was performed.
Table 3 shows that the mean Pb-210 concentration varied
signicantly compared to the mean base (540 pCi/L) when the Ra226 input changed. To understand the effects of these variations,

Table 3
Four different sensitivity analysis scenarios where the initial concentration varied 25% and 50% were evaluated. The nal Pb-210 remaining in treated water is shown (in Bq/
L) for each sensitivity analysis scenario.
Scenario
Ra-226
Ra-226
Ra-226
Ra-226

25%
25%
50%
50%

underestimation
overestimation
underestimation
overestimation

Initial Pb-210 (pCi/L)

Pb-210 predicted by QWASI (pCi/L)

Pb-210 in treated water (Bq/L)

674.54
404.81
809.52
269.85

173.41
104.35
208.70
69.36

0.77
0.46
0.93
0.31

634

L. Torres et al. / Chemosphere 169 (2017) 627e635

each sensitivity scenario was evaluated in the QWASI model. In

addition, the Pb-210 level (in Bq/L) that would remain in the water
after treatment was calculated and compared with the 0.63 Bq/L
that would remain if the base concentration was considered.
The concentration level of Pb-210 after treatment in each scenario was almost 3 to 9 times higher than the recommended limit
of 0.1 Bq/L established by the WHO. This range falls in the level 2
category (or moderate) of the measurement of risk consequence or
impact shown in Table 1, the same category obtained using the base
concentration. This indicates that even if our simulated Pb-210
concentration varies 25%e50% from the actual value, the risk of
water contamination in scenarios 1 and 3 would still be considered
medium-high while scenario 2 would still be high risk.
4. Conclusions
This study analyzed from an exploratory perspective the impacts of Pb-210 in PW from unconventional oil production in ND.
We developed a method to predict Pb-210 in the water environment based on Ra-226 in PW. Also, we created a holistic process
that incorporates technical and social risk. With this study,
although with some limitations, we found that the risks of SW
contamination with Pb-210 are not negligible and call for attention.
Undoubtedly, the holistic method developed in this study has some
limitations and more research on this subject is necessary to rene
and complement our ndings. Some of the limitations are: 1) accuracy of the Model due to the variability in the produced water
characteristics from different locations, 2) one-year period analysis
of spill incidents related to PW in ND and the decay of Ra-226, and
3) self-selection of the survey participants and narrow range of age
of the respondents. Despite these limitations, our results have
several signicant implications for assessing the real impacts of the
incorrect management of PW. First, the lack of data indicate that it
is imperative to request oil operators to report radioactivity levels
in PW. Second, improving the spill reporting system could reduce
the social risk perception since more and better information would
be available. Finally, our study offers the opportunity to utilize the
holistic process in future work that could be expanded to other
scenarios such as runoff and melting snow which were not covered
in this research.
Acknowledgment
The authors declare no competing nancial interest. Funding for
this research was partially provided by the North Dakota Water
Resources Research Institute. Any opinions, ndings, and conclusions or recommendations expressed in this material are those of
the author(s) and do not necessarily reect the views of the North
Dakota Water Resources Research Institute.
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.chemosphere.2016.11.125.
References
Argonne National Laboratory, 2014. Radiological Dose and Risk Assessment of
Landll Disposal of Technologically Enhanced Naturally Occurring Radioactive
Materials (TENORM) in North Dakota. https://www.ndhealth.gov/EHS/Tenorm/
ArgonneStudy/ANL-NDDH TENORM Landll Study (ANL EVS-14_13) Final
Report.pdf (Accessed 25 May 2015).
Barbot, E., Vidic, N.S., Gregory, K.B., Vidic, R.D., 2013. Spatial and temporal correlation of water quality parameters of produced waters from devonian-age shale
following hydraulic fracturing. Environ. Sci. Technol. 47, 2562e2569. http://
dx.doi.org/10.1021/es304638h.
Blondes, M., Gans, K., Thordsen, J., Reidy, M., Thomas, B., Engle, M., Kharaka, Y.,

Rowan, E., 2014. U.S. Geological Survey National Produced Waters Geochemical
Database v2.1. http://energy.usgs.gov/EnvironmentalAspects/Environmental
Aspects
ofEnergyProductionandUse/ProducedWaters.aspx#3822349-data
(Accessed 18 August 2015).
Braun, M., 1983. Differential Equations and Their Applications: an Introduction to
Applied Mathematics. Springer Science & Business Media.
Brereton, S., Alenbach, T., 1998. In: Risk Ranking Methodology for Chemical Release
Events. Probabilistic Safety Assessment and Management 4 International Conference on Probabilistic Safety Assessment and Management (New York City,
NY).
Centers for Disease Control and Prevention, 2013. Lead - Infographic. http://www.
cdc.gov/nceh/lead/infographic.htm (Accessed 9 October 2015).
Chriss, D., Bursh, T.P., 2002. A Study of the Relationship of Geological Formation to
the NORM. Southern University, (United States), Pittsburgh, PA, and Morgantown, WV. http://dx.doi.org/10.2172/826008.
City of Williston, 2014. Annual Drinking Water Quality Report, p. 2013. http://
cityofwilliston.com/usrles/news/2013 ANNUAL WATER REPORT.pdf (Accessed
22 March 2016).
Ding, X., Hartog, J., Sun, Y., 2010. Can We Measure Individual Risk Attitudes in a
Survey? Germany, Bonn.
Emmanuel, E., Angerville, R., Joseph, O., Perrodin, Y., 2007. Human health risk
assessment of lead in drinking water : a case study from Port-au-Prince Haiti.
Int. J. Environ. Pollut.. 31, 280e291. doi:10.1504/IJEP.2007.016496.
Fisher, R., 1995. Naturally occurring radioactive materials (NORM) in produced
water and scale from Texas oil, gas, and geothermal wells: geographic, geologic,
and geochemical controls. Geol. Circ. 95.
Fisher, R., 1998. Geologic and geochemical controls on naturally occurring radioactive materials (NORM) in produced water from oil, gas, and geothermal operations. Environ. Geosci. 5, 139e150. http://dx.doi.org/10.1046/j.15260984.1998.08018.x.
Food and Agriculture Organization of the United Nations, n.d. Qualitative Risk
Analysis (consequence X likelihood). http://www.fao.org/shery/eaf-net/
eaftool/eaf_tool_4/en (Accessed 5 Febraury 2016).
Galesic, M., 2006. Dropouts on the Web: effects of interest and burden experienced
during an online survey. J. Off. Stat. 22, 313.
Gleason, R.A., Tangen, B.A., 2014. Brine Contamination to Aquatic Resources from Oil
and Gas Development in the Williston Basin. http://dx.doi.org/10.3133/
sir20145017. United States: U.S., Scientic Investigations Report 2014e5017.
Reston, Virginia.
Godoy, J.M., Petinatti da Cruz, R., 2003. 226Ra and 228Ra in scale and sludge
samples and their correlation with the chemical composition. J. Environ.
Radioact. 70, 199e206. http://dx.doi.org/10.1016/S0265-931X(03)00104-8.
Guerin, E., 2015. State Ofcials Misrepresent North Dakota's Spill Problem. Insid.
Energy. http://insideenergy.org/2015/02/16/state-ofcials-misrepresent-northdakotas-spill-problem/ (Accessed 16 November 2016).
Guerra, K., Dahm, K., Dundorf, S., 2011. Oil AndFood and Agriculture Gas Produced
Water Management and Benecial Use in the Western United States. https://
www.usbr.gov/research/AWT/reportpdfs/report157.pdf
(Accessed
18
December 2015).
International Atomic Energy Agency, 2003. Radiation Protection and the Management of Radioactive Waste in the Oil and Gas Industry. Safety Reports Series No.
34. Vienna, Austria.
Jaworowski, Z., 1969. Radioactive lead in the environment and in the human body.
At. Energy Rev. 7, 3e45.
Keith, S., Doyle, J.R., Harper, C., Mumtaz, M., Tarrago, O., Wohlers, D.W.,
Diamond, G.L., Citra, M., Barber, L.E., 2012. Toxicological Prole for Radon.
http://www.ncbi.nlm.nih.gov/books/NBK158780/ (Accessed 5 Febraruy 2016).
Kusnetz, N., 2012. North Dakota's Oil Boom Brings Damage along with Prosperity.
ProPublica.
https://www.propublica.org/article/the-other-fracking-northdakotas-oil-boom-brings-damage-along-with-prosperi
(Accessed
16
November 2016).
Lauer, N.E., Harkness, J.S., Vengosh, A., 2016. Brine spills associated with unconventional oil development in North Dakota. Environ. Sci. Technol. 50,
5389e5397. http://dx.doi.org/10.1021/acs.est.5b06349.
Leiserowitz, A.A., 2005. American risk perceptions: is climate change dangerous?
Risk Anal. 25, 1433e1442. http://dx.doi.org/10.1111/j.1540-6261.2005.00690.x.
Mackay, D., Diamond, M., 1989. Application of the QWASI (Quantitative Water Air
Sediment Interaction) fugacity model to the dynamics of organic and inorganic
chemicals in lakes. Chemosphere 18, 1343e1365. http://dx.doi.org/10.1016/
0045-6535(89)90027-1.
Mackay, D., Hughes, L., Powell, D.E., Kim, J., 2014. An updated Quantitative Water Air
Sediment Interaction (QWASI) model for evaluating chemical fate and input
parameter sensitivities in aquatic systems: application to D5 (decamethylcyclopentasiloxane) and PCB-180 in two lakes. Chemosphere 111,
359e365. http://dx.doi.org/10.1016/j.chemosphere.2014.04.033.
MeasuringU, 2011. How to Interpret Survey Responses: 5 Techniques. http://www.
measuringu.com/blog/interpret-responses.php (Accessed 27 April 2016).
Mosleh, A., Bari, R., 1998. In: Probabilistic Safety Assessment and Management.
Proceedings of the 4th International Conference on Probabilistic Safety
Assessment and Management (New York City, NY).
Nelson, A.W., Johns, A.J., Eitrheim, E.S., Knight, A.W., Basile, M., Bettis III, E.A.,
Schultz, M.K., Forbes, T.Z., 2016. Partitioning of naturally-occurring radionuclides (NORM) in Marcellus Shale produced uids inuenced by chemical
matrix. Environ. Sci. Process. Impacts 18, 456e463. http://dx.doi.org/10.1039/
C5EM00540J.

L. Torres et al. / Chemosphere 169 (2017) 627e635

North Dakota Department of Health, 1999. Source Water Assessment Strategic Plan
- Strategic Plan. http://ndhealth.gov/WQ/GW/publications2.htm (Accessed 22
March 2016).
North Dakota Department of Health, 2015. Oileld Environmental Incidents. http://
www.ndhealth.gov/ehs/spills/ (Accessed 12 October 2015).
North Dakota Department of Transportation, 2015. North Dakota Crash Summary,
p.
2014.
http://www.dot.nd.gov/divisions/safety/docs/crash-summary.pdf
(Accessed 12 October 2015).
North Dakota Petroleum Council, 2012. Oil and Gas Frequently Asked Questions.
http://www.nd.gov/ndic/ogrp/info/g-015-033-faq.pdf (Accessed 27 October
2015).
North Dakota Petroleum Council, 2015. Bakken by the numbers. http://www.ndoil.
org/image/cache/Facts_Figures.pdf (Accessed 09 December 2015).
Paschoa, A.S., 1998. Potential environmental and regulatory implications of naturally occurring radioactive materials (NORM). Appl. Radiat. Isot. 49, 189e196.
http://dx.doi.org/10.1016/S0969-8043(97)00239-X.
Perry, S.L., 2012. Environmental reviews and case studies: addressing the societal
costs of unconventional oil and gas exploration and production - a framework
for evaluating short-term, future, and cumulative risks and uncertainties of
hydrofracking. Environ. Pract. 14, 352e365. http://dx.doi.org/10.1017/
S1466046612000336.
Plattner, T., Plapp, T., Hebel, B., 2006. Integrating public risk perception into formal
natural hazard risk assessment. Nat. Hazards Earth Syst. Sci. 6, 471e483.
Prince, J.R., 1979. Comments on equilibrium, transient equilibrium, and secular
equilibrium in serial radioactive decay. J. Nucl. Med. 20, 162e164.
Rich, A.L., Crosby, E.C., 2013. Analysis of reserve pit sludge from unconventional
natural gas hydraulic fracturing and drilling operations for the presence of
technologically enhanced naturally occurring radioactive material (TENORM).
New Solut. 23, 117e135. http://dx.doi.org/10.2190/NS.23.1.h.
Rowan, E.L., Engle, M.A., Kirby, C.S., Kraemer, T.F., 2011. Radium Content of Oil- and
Gas-Field Produced Waters in the Northern Appalachian Basin (USA): Summary
and Discussion of Data. Reston, Virginia.
Sandman, P., 1999. Risk Hazard Outrage: Coping with Controversy about Utility
Risks. Eng. News-Record. http://www.psandman.com/articles/amsa.htm
(Accessed 10 May 2016).
Scanlon, B.R., Reedy, R.C., Nicot, J.-P., 2014. Comparison of water use for hydraulic
fracturing for unconventional oil and gas versus conventional oil. Environ. Sci.
Technol. 48, 12386e12393. http://dx.doi.org/10.1021/es502506v.
Shammas, N., Wang, L., 2016. Water Engineering: Hydraulics, Distribution and
Treatment. John Wiley & Sons.
Slovic, P., 2000. The Perception of Risk. Risk, Society, and Policy Series. Earthscan
Publications, London, England.
Slovic, P., Peters, E., 2006. Risk perception and affect. Curr. Dir. Psychol. Sci. 15,
322e325. http://dx.doi.org/10.1111/j.1467-8721.2006.00461.x.

635

Smith, A.L., 2011. First Correlation of NORM with a Specic Geological Hypothesis.
SPE European Health, Safety and Environmental Conference in Oil and Gas
Exploration and Production. Society of Petroleum Engineers, Viena. http://
dx.doi.org/10.2118/138136-MS.
Stouffer, J., King, M., Appy, M., Rogers, G., Schottke, J., Dargan, K., Schaenman, P.,
Senchyna, M., Crawford, J., 2015. Community risk assessment, a guide for
conducting a community risk assessment. http://strategicre.org/wp-content/
uploads/2015/04/Community -Risk-Assessment-Guide-v1.11.pdf (Accessed 5
Febraury 2016).
SurveyMonkey, 2016. Rating & Ranking Average Calculations. Help Center. http://
help.surveymonkey.com/articles/en_US/kb/What-is-the-Rating-Average-andhow-is-it-calculated (Accessed 27 April 2016).
Tamponnet, C., 2009. Behaviour of lead-210 in continental environment: comparison with stables isotopes of lead. Radioprotection 44, 285e289. http://
dx.doi.org/10.1051/radiopro/20095055.
U.S. Environmental Protection Agency, 2011. State and county emission summaries.
Air emissions sources. https://www3.epa.gov/cgi-bin/broker?_servicedata&_
debug0&_ programdataprog.state_1.sas&pol7439921&stps38#content
(Accessed 27 April 2016).
U.S. Environmental Protection Agency, 2012. Edition of the Drinking Water Standards and Health Advisories, p. 2012. Washington, DC.
U.S. Environmental Protection Agency, 2015. TENORM: Drinking Water Treatment
Wastes.
http://www.epa.gov/radiation/tenorm-drinking-water-treatmentwastes#tab-2 (Accessed 5 Febraury 2016).
Valsaraj, K.T., Melvin, E.M., 2000. Elements of Environmental Engineering: Thermodynamics and Kinetics. CRC Press.
Velten, R., Jacobs, B., 1982. Determination of Lead-210 in Drinking Water Method
909.0 (Cincinnati, Ohio).
Vengosh, A., Jackson, R.B., Warner, N., Darrah, T.H., Kondash, A., 2014. A critical
review of the risks to water resources from unconventional shale gas development and hydraulic fracturing in the United States. Environ. Sci. Technol. 48,
8334e8348. http://dx.doi.org/10.1021/es405118y.
Vidic, R.D., Brantley, S.L., Vandenbossche, J.M., Yoxtheimer, D., Abad, J.D., 2013.
Impact of shale gas development on regional water quality. Science 340,
1235009. http://dx.doi.org/10.1126/science.1235009.
Webster, E., Lian, L., Mackay, D., Harvey, C., 2005. Application of the Quantitative
Water Air Sediment Interaction (QWASI) Model to the Great Lakes. Report No.
200501, Report to the lakewide management plan (LaMP) committee.
Woodne, D.G., Seth, R., Mackay, D., Havas, M., 2000. Simulating the response of
metal contaminated lakes to reductions in atmospheric loading using a modied QWASI model. Chemosphere 41, 1377e1388. http://dx.doi.org/10.1016/
S0045-6535(00)00006-0.
World Health Organization, n.d. Radiological Aspects. http://www.who.int/water_
sanitation _ health/dwq/GDW9rev1and2.pdf (Accessed 5 Febraury 2016).