Subscriber access provided by AUBURN UNIV AUBURN
Article
Power plant bromide discharges and downstream drinking water systems in Pennsylvania Kelly D Good, and Jeanne M. Vanbriesen Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.7b03003 • Publication Date (Web): 25 Sep 2017 Downloaded from http://pubs.acs.org on September 29, 2017
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 31
Environmental Science & Technology
1 2 3 4
Power plant bromide discharges and downstream drinking
5
water systems in Pennsylvania
6 Kelly D. Gooda, Jeanne M. VanBriesenb*
7 8 9
a
Graduate Research Assistant, Department of Civil and Environmental Engineering, Carnegie
10
Mellon University, 5000 Forbes Ave., Pittsburgh, PA 15213, USA. Email:
[email protected] 11
b
12
Systems (Water-QUEST), Department of Civil and Environmental Engineering and Department
13
of Engineering and Public Policy, Carnegie Mellon University, 5000 Forbes Avenue, Pittsburgh,
14
PA 15213, USA. Email:
[email protected] Duquesne Light Company Professor, Director of Water Quality in Urban Environmental
15 16
ACS Paragon Plus Environment
Environmental Science & Technology
Page 2 of 31
17
ABSTRACT
18
Coal-fired power plants equipped with wet flue gas desulfurization (FGD) systems have been
19
implicated in increasing bromide levels and subsequent increases in disinfection by-products at
20
downstream drinking water plants. Bromide was not included as a regulated constituent in the
21
recent steam electric effluent limitations guidelines and standards (ELGs) since the U.S. EPA
22
analysis suggested few drinking water facilities would be affected by bromide discharges from
23
power plants. The present analysis uses a watershed approach to identify Pennsylvania drinking
24
water intakes downstream of wet FGD discharges and to assess the potential for bromide
25
discharge effects. Twenty-two (22) public drinking water systems serving 2.5 million people
26
were identified as being downstream of at least one wet FGD discharge. During mean August
27
conditions (generally low-flow, minimal dilution) in receiving rivers, the median predicted
28
bromide concentrations contributed by wet FGD at Pennsylvania intake locations ranged from
29
5.2 to 62 µg/L for the Base scenario (including only natural bromide in coal) and 16 to 190 µg/L
30
for the Bromide Addition scenario (natural plus added bromide for mercury control); ranges
31
depend on bromide loads and receiving stream dilution capacity.
32 33 34
KEYWORDS: source water, disinfection by-products, bromide, mercury control, coalfired power plants
2 of 31
ACS Paragon Plus Environment
Page 3 of 31
35
Environmental Science & Technology
TOC/ABSTRACT ART:
36 37
3 of 31
ACS Paragon Plus Environment
Environmental Science & Technology
Page 4 of 31
38
INTRODUCTION
39
Bromide in surface waters is of concern due to its interaction with applied disinfectants in
40
drinking water treatment systems. Elevated bromide concentrations increase the rate and extent
41
of formation of disinfection by-products (DBPs)1–4 and shift speciation towards brominated
42
DBPs, which are more toxic.5–7 Thus, drinking water sources containing bromide may lead to
43
formation of DBPs in treated drinking water that present higher health risks to consumers.8,9
44 45
In surface water sources, background bromide levels may be affected by nonpoint runoff sources
46
such as road salt or brine10–12 or by soil releases.13,14 Historically, elevated bromide in drinking
47
water sources was associated with coastal groundwater and estuary sources, which are subject to
48
saltwater intrusion.15–19 Recently, rising bromide concentrations have also been associated with
49
anthropogenic point sources (e.g., flame retardants for plastics or textiles,20 municipal waste
50
incineration21), as well as water resource changes associated with altered climate, including sea
51
level rise increasing bromide in estuaries and near shore groundwater16,22,23 and drought affecting
52
dilution capacity of surface waters that receive bromide loads.24,25
53 54
Of particular note, point source discharges associated with energy extraction and utilization
55
activities have been implicated in increased surface water concentrations of bromide in the
56
United States (U.S.) since oil and gas (O&G) wastewater,26,27 coal mine discharges,28–30 coal
57
combustion residue (CCR) effluent,31 and coal-fired power plant wet flue gas desulfurization
58
(FGD) wastewater32–35 are elevated in bromide. The rapid expansion of unconventional gas
59
drilling and subsequent increase in produced water discharges in Pennsylvania (PA) led to
60
elevated bromide levels that were well-documented,26,33,36–38 prompting the PA Department of
4 of 31
ACS Paragon Plus Environment
Page 5 of 31
Environmental Science & Technology
61
Environmental Protection (PADEP) in 2011 to request that drillers voluntarily stop using surface
62
discharging plants for unconventional produced water disposal.39
63 64
These recent bromide management challenges from O&G wastewater highlight the importance
65
of understanding bromide sources and options for managing discharges; however, due to limited
66
surface water monitoring of bromide, few studies have evaluated anthropogenic bromide source
67
contributions at the regional scale.21 Several recent studies have looked specifically at the
68
Allegheny River Basin in Pennsylvania,26,33,40 where O&G produced water and wastewater
69
associated with coal-fired power plant discharges have been implicated in elevated bromide
70
concentrations at downstream drinking water plants.
71 72
Bromine is naturally present in coal in trace amounts,41,42 and bromide addition to coal has been
73
proposed to enable compliance with the Mercury and Air Toxics Standards (MATS).43–46 During
74
coal combustion, bromine is primarily converted to volatile hydrogen bromide (HBr) and
75
subsequently to bromine gas (Br2) upon cooling, with minimal transfer to residue streams such as
76
fly ash and bottom ash (typically >90% remains in the vapor or gas phase).47–51 Halogen species
77
such as bromine, which would otherwise be released into the atmosphere out of the stack,50 are
78
reduced to soluble ionic forms and are commonly observed in flue gas desulfurization (FGD)
79
wastewater.52,53 Additional details on bromide at power plants are provided in SI Section A.
80 81
Discharges of bromide have historically been unregulated34,54 since bromide is not expected to
82
harm humans or ecosystems at concentrations typically observed in surface waters.13,55 However,
83
due to its effect on drinking water DBPs, the U.S. Environmental Protection Agency (EPA) has
5 of 31
ACS Paragon Plus Environment
Environmental Science & Technology
Page 6 of 31
84
recently increased its attention to bromide by including it in the Safe Drinking Water Act
85
(SDWA) fourth Unregulated Contaminant Monitoring Rule (UCMR 4)56 and by discussing the
86
potential for FGD discharges to affect drinking water quality in the Steam Electric Power
87
Effluent Limitations Guidelines and Standards (ELGs) final rule57 and supporting
88
documentation.58 As part of its analysis for the ELGs, the EPA evaluated the proximity of steam
89
electric power plant discharges to public drinking water resources, including intakes, wells, and
90
sole-source aquifers using an 8 km (5 mi) geographic buffer,58 and concluded that “only a
91
fraction of steam electric power plants have downstream drinking water intakes.”59
92 93
The ELGs focused predominately on controlling discharges of toxic metals and nitrogen and did
94
not set numeric limits for bromide.57 Instead, the EPA included recommendations for permitting
95
authorities to consider regulation of bromide discharges on a permit-by-permit basis, depending
96
on how the discharges affect downstream drinking water plants.57 The ELGs also recommended
97
that permitting authorities require monitoring and reporting of bromide, which is currently not
98
required and which would enable improved quantification of wet FGD bromide contributions.
99 100
Our recent work, focused on the Allegheny River Basin in PA, demonstrated that multiple power
101
plants operating wet FGD currently contribute approximately one third of the total bromide
102
concentration at a drinking water treatment plant located a significant distance downstream
103
(12.8-142 km).40 This suggests that the use of a geographic buffer for evaluating the effects of
104
wet FGD discharges on surface water sources is not appropriate and that power plant
105
contributions are relevant even in the context of other sources (oil and gas produced water
106
accounted for approximately half of the total concentration predicted). The present work uses a
6 of 31
ACS Paragon Plus Environment
Page 7 of 31
Environmental Science & Technology
107
watershed approach to identify drinking water intakes that have wet FGD discharges upstream
108
(regardless of the distance). Then bromide load and concentration contributions from wet FGD at
109
each intake are evaluated. The analysis includes consideration of the population served by public
110
drinking water that is potentially affected by wet FGD discharges. It further demonstrates how
111
concentration-based goals for bromide at drinking water intakes could be used to evaluate the
112
effect of bromide discharges from power plants on drinking water treatment plants.
113 114
METHODS
115
Bromide is nonreactive under typical environmental conditions, and thus the magnitude of the
116
discharge and the dilution capacity of the surface water at the intake – not geographic proximity
117
– govern the concentration at the intake. The purpose of this work was to identify any intakes
118
potentially affected by at least one wet FGD discharge and to estimate the bromide concentration
119
contribution from upstream wet FGD discharge(s) at each intake location.
120 121
The analysis used a watershed approach to identify Pennsylvania drinking water intakes (for
122
systems primarily using rivers and serving greater than 10,000 people) that were downstream of
123
a wet FGD discharge, regardless of whether the discharge was located in Pennsylvania. The
124
analysis was limited to large systems as defined by SDWA,60 following Rice and Westerhoff
125
(2015),61 and captures 60% of the total PA population and 70% of the PA population served by
126
community water systems.
127 128
To do this, the following data were mapped in ArcGIS: watersheds intersecting PA, rivers
129
within the watersheds, wet FGD discharge locations, and PA drinking water intakes. Then,
130
drinking water intakes downstream of wet FGD discharges were identified by visual inspection 7 of 31
ACS Paragon Plus Environment
Environmental Science & Technology
Page 8 of 31
131
(Figure 1), and loads and concentrations were modeled following the methods described in Good
132
and VanBriesen (2016).40 A schematic summarizing the watershed analysis and modeling
133
approach used here is provided in Figure S1.
134 135
Data sources
136
The watershed-level analysis required the following geospatial data: watersheds intersecting PA,
137
rivers within the watersheds, coal electricity generating units (EGUs) associated with wet FGD,
138
and PA river segments providing source water to community water systems serving greater than
139
10,000 people. For estimation of bromide loading, coal-fired power plant EGU data (i.e., coal
140
consumption) and characteristics (i.e., the presence of wet FGD) were required. Streamflow data
141
enabled estimation of wet FGD concentration contributions from the modeled bromide loads. In
142
order to compare the identified drinking water utilities with those described in the proximity
143
analysis results discussed in the ELGs, data were requested and received from EPA. This
144
included the power plants identified as having at least one drinking water system within 5 miles
145
of a discharge that would be regulated under the ELGs.62 These data sources and detailed
146
methods are described in SI Section B.
147 148
Modeling bromide load and concentration contributions from wet FGD at intakes
149
Once the downstream intakes were identified, bromide loads from each upstream wet FGD plant
150
were modeled for August and on an average annual basis for two scenarios: Base (natural Br in
151
coal) and Br Addition (natural plus Br added for mercury control) following the method from
152
Good and VanBriesen (2016).40 August was selected because it is a low-flow period in
153
Pennsylvania40,63,64 and because it is a summer month when elevated temperatures typically
8 of 31
ACS Paragon Plus Environment
Page 9 of 31
Environmental Science & Technology
154
require increased disinfectant dosing,65,66 leading to higher DBP concentrations during third
155
quarter compliance sampling.9,67,68 A summary of load model input and data are provided in SI
156
Section C; the method was described in detail previously.40
157 158
This method does not use concentration-based discharge data from the NPDES permitted
159
locations as power plants have not typically been required to monitor bromide in their effluent.
160
Rather, these load estimates are based on wet FGD-associated coal consumption, a range of
161
bromide content in coal, a range of percent capture of bromide in wet FGD wastewater, and a
162
range of bromide addition dosage; uncertainty is incorporate using Monte Carlo simulations.
163
Then, in-stream wet FGD bromide concentration contributions at each intake site were calculated
164
using the estimated loads under mean August and mean annual flows (estimated from the
165
National Hydrography Dataset [NHD],69 as described in SI Section B).
166 167
To provide context for these wet FGD contributions, total in-stream bromide loads were
168
calculated using monthly 2015-2016 bromide concentration data from the PADEP Water Quality
169
Network (WQN)70 and corresponding streamflow data for each sample date from the U.S.
170
Geological Survey (USGS).71 The WQN sites were selected based on proximity to the modeled
171
intakes. These calculated in-stream total loads were compared to the wet FGD loads at the
172
nearest intake site. Additional details are provided in SI Section D.
173 174
RESULTS AND DISCUSSION
175
Identification of potentially affected PA drinking water systems
9 of 31
ACS Paragon Plus Environment
Environmental Science & Technology
Page 10 of 31
176
The present analysis identified 21 river locations (numbered sites in Figure 1) for 23 intakes
177
providing source water for 22 drinking water systems in Pennsylvania that are downstream of at
178
least one wet FGD discharge from 9 power plants in Pennsylvania and West Virginia (see Table
179
S2 for more details). These 22 drinking water systems serve 2.5 million people, or 20% of the
180
PA population. Figure S2 shows the service areas for all public water systems in Pennsylvania,
181
and the service areas for the drinking water systems identified in this analysis are highlighted to
182
show the geographic extent of potentially affected populations.
183 184
Figure 1. Map of Pennsylvania drinking water intakes (for public water systems utilizing
185
primarily surface water and serving more than 10,000 people) and coal-fired power plants
186
operating wet FGD. 10 of 31
ACS Paragon Plus Environment
Page 11 of 31
Environmental Science & Technology
187 188
EPA ELG national proximity analysis
189
The EPA drinking water proximity analysis evaluated 222 discharge locations from 195 steam
190
electric power plants in the U.S.58 From this analysis (described in more detail in SI Section B),
191
the EPA identified 113 drinking water intakes or reservoirs as being potentially affected by
192
discharges from 67 coal-fired power plants.62 These 113 intakes or reservoirs were associated
193
with 92 public drinking water systems, and 14 (15%) of the 92 systems were located in
194
Pennsylvania. The map in Figure S3 uses color gradation to show the distribution of these 92
195
buffer-identified drinking water systems in each state, where darker shading indicates a higher
196
number of systems. Coal EGUs with wet scrubbers are also shown on this map for reference.
197 198
Four of the 14 EPA-identified drinking water plants in Pennsylvania are within 5 miles of power
199
plants that do not use wet FGD (stars inside the circles with dashed outlines in Figure S3). These
200
plants would discharge bromide only if they added it as a biocide in the cooling towers and are
201
outside the scope of the present analysis. The remaining 10 Pennsylvania drinking water systems
202
identified by EPA were within EPA’s geographic buffer around a wet FGD power plant
203
discharge and were compared to the present watershed analysis. The population served by these
204
drinking water systems was not discussed in the ELGs or its supporting documentation; however,
205
using the latest available SDWIS data (2016 Quarter 4),72 the population served by these 10
206
drinking water systems was determined to be 150,000. Additional details on the EPA ELG
207
proximity analysis and comparison to the current watershed analysis are provided in SI Section
208
B.
209
11 of 31
ACS Paragon Plus Environment
Environmental Science & Technology
Page 12 of 31
210
Wet FGD bromide loads
211
Wet FGD bromide loading depends on coal consumption, which can vary by month. Results are
212
shown at the watershed level in Figure 2 (and Table S10), with additional tables and figures
213
provided in SI Section D. Wet FGD bromide loads are higher in the Allegheny and Monongahela
214
Basins in western PA (August Base median of 610 kg/day and 590 kg/day, respectively)
215
compared with the Susquehanna Basin in eastern PA (August Base median of 310 kg/day). This
216
is because there are more power plants in the western basins and because they had higher coal
217
consumption in the modeled period (2015-2016). The Allegheny and Monongahela Rivers meet
218
in Pittsburgh, PA to form the Ohio River, which therefore receives wet FGD bromide loads from
219
those river basins (August Base median of 1200 kg/day). These loads could affect additional
220
drinking water plants on the Ohio River outside of Pennsylvania (not considered in this analysis).
221
Dilution of the Allegheny and Monongahela bromide loads as well as additional bromide loads
222
from coal-fired power plants in the Ohio Basin would affect bromide concentrations at Ohio
223
River drinking water plants.
12 of 31
ACS Paragon Plus Environment
Page 13 of 31
Environmental Science & Technology
224 225
Figure 2. Boxplot of watershed-level August and average annual total predicted bromide
226
loads from wet FGD discharges for Base and Br Addition Scenarios. Boxes extend from the
227
first to third quartile with a line at the median; whiskers extend 1.5 times the interquartile
228
range. Summary statistics and power plants contributing to each watershed are provided
229
in Table S10.
230
Using 2015-2016 observed bromide and flow data for the rivers, the wet FGD discharges were
231
estimated to contribute 31% (Allegheny), 30% (Monongahela), 38% (Ohio), and 14%
232
(Susquehanna) of total in-stream bromide under Base conditions (see SI Section D for more
233
details). Wet FGD bromide loads estimated for the power plants in the present analysis are lower
234
than if estimated using the approach provided by McTigue et al. (2013),34 which utilized power
235
plant capacity, and thus, represented a theoretical maximum bromide load, rather than an
236
anticipated load. The present work considers reported coal consumption, variable bromide
13 of 31
ACS Paragon Plus Environment
Environmental Science & Technology
Page 14 of 31
237
addition rates, and observed bromide capture rates for FGD. A power plant-by-power plant
238
comparison of the two methods is provided in Table S12.
239 240
Wet FGD bromide concentration contributions at drinking water intakes
241
The effect of bromide loads on in-stream concentrations depends on the magnitude of the load
242
and the dilution capacity of the receiving stream at the downstream location of interest (i.e., for
243
this analysis intake sites labeled 1 through 21 in Figure 1). The contributed concentrations from
244
wet FGD discharges presented here cannot be directly compared with observed bromide
245
concentrations as in-stream concentrations include multiple sources. However, for context,
246
information collection rule (ICR) bromide concentrations at large Pennsylvania surface water
247
intakes (in 1996-1997) ranged from below detection (
