Memorandum: Emissions Overview: Hazardous
Air Pollutants in Support of the Final Mercury
and Air Toxics Standard
-------
-------
EPA-454/R-11-014
November 2011
Memorandum: Emissions Overview: Hazardous Air Pollutants in
Support of the Final Mercury and Air Toxics Standard
Marc Houyoux
and
Madeleine Strum
Emission Inventory and Analysis Group
Air Quality Assessment Division
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Air Quality Assessment Division
Emissions Inventory and Analysis Group
Research Triangle Park, North Carolina
-------
MEMORANDUM
To: Toxics Rule docket, number EPA-HQ-OAR-2009-0234
From: Marc Houyoux, group leader Emission Inventory and Analysis Group
Madeleine Strum, Emission Inventory and Analysis Group
Subject: Emissions Overview: Hazardous Air Pollutants in Support of the Final Mercury and Air Toxics
Standard
Date: December 1, 2011
1 Introduction
The purpose of this document is to summarize the mercury (Hg) and non-Hg Hazardous Air
Pollutants (HAP) emissions from U.S. electric generating utilities (EGUs) associated with the final Mercury
and Air Toxics Standard (MATS). This information includes both background information about these
pollutants as well as emissions totals from several emissions cases developed for the final MATS. In this
memo, we summarize current HAP emissions estimates from U.S. EGUs, the estimates used in the national-
scale Hg risk assessment (Hg Risk TSD) for years 2005 and 2016, and the estimates for the future impact of
the rule when fully implemented (in 2016).
2 Hg Emissions
2.1 Hg Emissions Overview
Mercury is an element. There is a fixed amount of it in the world. As long as it is bound up, for
example in coal, it cannot affect people or the environment. Once it is released, for example via the
combustion process, it enters the environment and becomes available for chemical conversion. Once
emitted, Hg remains in the environment and can bioaccumulate in organisms or be remitted through natural
processes. Mercury is emitted through natural and anthropogenic processes, and previously deposited Hg
from either process may be re-emitted. The majority of natural Hg emissions arise from volcanoes,
geothermal activity, Hg-enriched topsoil, oceans/lakes, and vegetation. The category of natural Hg
emissions may or may not include biomass burning, due to the difficulty in determining whether biomass
burning events are started by anthropogenic or natural causes. Generally, natural Hg emissions are
dominated by the elemental form of mercury, and as a result have more impact on the global pool of Hg than
on deposition in the region of the emissions.
Mercury is known to exist in the atmosphere in three forms: (1) elemental mercury, Hg°, (2) gaseous
oxidized mercury, Hg+2, and (3) particulate bound mercury, Hgp. Hg° dominates total mercury composition
in the atmosphere (greater than 95%) and has a much greater residence time than Hg+2 or Hgp . Hg° ambient
concentrations are spatially uniform due to the long atmospheric residence time and a more uniform
distribution of sources due to re-emission of previously deposited mercury. Hg+2 and HgP concentrations are
locally elevated near sources due to their shorter atmospheric lifetime. Total mercury deposition is likely to
be dominated by all sources of Hg° including global sources, while deposition of Hg+2 and Hgp is primarily
from local and regional sources.2
1 Schroeder, W. H. and J. Munthe (1998). "Atmospheric mercury - An overview." Atmospheric Environment 32(5): 809-822.
2 Marsik, F. J., G. J. Keeler, et al. (2007). "The dry-deposition of speciated mercury to the Florida Everglades: Measurements and
modeling." Atmospheric Environment 41(1): 136-149.
1
-------
The categories for anthropogenic Hg emissions include the combustion of fossil-fuels, cement
production, waste incineration, metals production, and other industrial processes. Anthropogenic Hg
emissions consist of all three forms of Hg.
Mercury re-emissions include previously deposited Hg originating from both natural and
anthropogenic sources. At this time, it is not possible to determine the original source of previously
deposited Hg, i.e., whether its source is natural emissions or re-emissions from previously deposited
anthropogenic Hg.3'4'5 Current publications on this topic estimate that half of re-emitted Hg originates from
anthropogenic sources.6'7
Current estimates of total global Hg emissions based on a 2005 inventory range from 6,600 to 7,500
metric tons per year (mt/yr). The United Nations Environment Programme (UNEP) estimates of global Hg
emissions for 2005 are somewhat lower, at 5,600 mt/yr8'9. Global anthropogenic Hg emissions, excluding
biomass burning, have been estimated by many researchers. UNEP's 2005 estimate is approximately 1,900
mt/yr (with a range of 1,200 to 3,000 mt/yr)10 and the 2005 estimate by Pirrone, et al. is approximately 2,400
mt/yr. u Global fossil-fuel fired power plants total approximately 500 to 800 mt/yr, a large fraction (25 to 35
percent) of the total global anthropogenic emissions.
There are large uncertainties regarding projected mercury global inventories. To model Hg deposition
for the national-scale Hg risk assessment, initial and boundary conditions were based on a GEOS-CHEM
simulation using a 2000-based global inventory that includes approximately 7,000 mt/yr of mercury
emissions13. Recent research shows that Hg emissions from China were consistent between 2000 and 2006
(approximately 1,300 mt/yr)14, so no adjustments were made to the 2000-based global Hg emissions for the
2005 simulation. Recent research has shown that ambient mercury concentrations have been decreasing in
3 Lindberg, S., Bullock, R., Ebinghaus, R., Engstrom, D., Feng, X., Fitzgerald, W., et al. (2007). A Synthesis of Progress and
Uncertainties in Attributing the Sources of Mercury in Deposition. Ambio, 36(1), 19-33.
4 Lohman, K., Seigneur, C., Gustin, M, & Lindberg, S. (2008). Sensitivity of the global atmospheric cycle of mercury to
emissions. Applied Geochemistry, 23(3), 454-466.
5 Seigneur, C., Vijayaraghavan, K., Lohman, K., Karamchandani, P., & Scott, C. (2004). Global Source Attribution for Mercury
Speciation in the United States. Environmental Science and Technology^?,), 555-569.
6 Mason, R., Pirrone, N., & Mason, R. P. (2009). Mercury emissions from natural processes and their importance in the global
mercury cycle. InMercury Fate and Transport in the Global Atmosphere (pp. 173-191): Springer U.S.
7 Selin, N. E., Jacob, D. I, Park, R. I, Yantosca, R. M., Strode, S., Jaegle, L., et al. (2007). Chemical cycling and deposition of
atmospheric mercury: Global constraints from observations. J. Geophys. Res, 112, 1071-1077.
8
UNEP (United Nations Environment Programme), Chemicals Branch, 2008. The Global Atmospheric Mercury
Assessment: Sources, Emissions and Transport, UNEP Chemicals, Geneva.
9 The 5,600 metric ton estimate is from the Figure on page 37 of the report. The report also provides ranges in the executive
summary.
10 Study on Mercury Sources and Emissions and Analysis of the Cost and Effectiveness of Control Measures "UNEP Paragraph 29
study", UNEP (DTIE)/Hg/INC.2/4. November, 2010.
11 Pirrone, N., Cinnirella, S., Feng, X., Finkelman, R. B., Friedli, H. R., Leaner, J., et al. (2010). Global mercury emissions to the
atmosphere from anthropogenic and natural sources. Atmospheric Chemistry and Physics Discussions, 1 0(2), 47 1 9-4752.
12 Study on Mercury Sources and Emissions and Analysis of the Cost and Effectiveness of Control Measures "UNEP Paragraph 29
study", UNEP (DTIE)/Hg/INC.2/4. November, 2010.
13 Selin, N.E., Jacob, D.J., Park, R.J., Yantosca, R.M., Strode, S., Jaegle, L., Jaffe, D., 2007. Chemical cycling and deposition of
atmospheric mercury: Global constraints from observations. Journal of Geophysical Research-Atmospheres 1 12.
14 Streets, D.G., Zhang, Q., Wu, Y., 2009. Projections of Global Mercury Emissions in 2050. Environmental Science & Technology 43, 2983-
2988.
-------
the northern hemisphere since 200015, so no adjustments were made to the boundary condition inflow for the
2016 simulations.
2.2 Estimating Current Hg Emissions
In 1990, Hg emissions for coal and oil EGUs in the U.S. were 59 tons16 out of a total of 246 tons17,
and by 1999 Hg emissions for U.S. EGUs were 4918 out of 115 tons. In 2005, they were 5319 tons out of a
total of 105 tons. As part of the development of the MATS, EPA completed in 2010 an Information
Collection Request (ICR) that gathered the latest available emissions data from coal- and oil-fueled EGUs.
The ICR data have been used in several ways, including to calculate EPA's best estimate of current
emissions, called the "current base" from U.S. EGUs. At proposal, the current base Hg emissions estimate
(labeled "2010" at proposal) was 29 tons, and after the revisions made and described here for the final
MATS, the current base estimate remains 29 tons of Hg from U.S. EGUs.
While the current base Hg emissions estimate is not a basis for the MATS rule, the risk assessment,
or the Regulatory Impact Analysis (RIA), EPA received numerous comments on its value as part of the
proposed rule. These comments compared the current base estimate to EPA's future year Hg emissions
estimate from U.S. EGUs and called into question the future-year estimate since it was the same as the
current base estimate. Since the current base estimates use a 2007-2009 average heat input, it did not seem
likely that the current base estimate would be the same as future years that include additional controls from
other rules. As a result, EPA has reviewed and updated its methods for the current base emissions, and these
updates are described in this section. These changes resulted in a decrease of 0.3 tons/year, which rounds to
the same 29 ton/yr value included with the proposed MATS.
The calculations underlying the current base inventories are available in the MATS docket in an
Excel workbook entitled "MATS_Final_Current_Base_HAP_inven.xlsx".
EPA updated the current base national inventories for the final rule in the following three ways:
• Updated list of units assumed to be subject to the rule for the purposes of calculating the national
inventory values. Within a given facility, some units are subject to the rule and some are not
depending on each unit's capacity and fuels.
• Revised EFs for untested units. These revisions included outlier statistical tests of the underlying
stack test data for each pollutant to improve the quality of the estimates. This approach was also
consistent with the EF method used to estimate emissions for untested units in the non-Hg risk case
studies.
• Revised the heat input to use the 2007-2009 average CEM-based heat input for units reporting to the
CEM program, rather than ICR-based heat input. For other units, EPA used the same approach as at
proposal, which was based on the ICR data and also represents a 2007-2009 average.
Slemr, F., Brunke, E.G., Ebinghaus, R., Kuss, J., 2011. Worldwide trend of atmospheric mercury since 1995. Atmospheric Chemistry and
Physics 11, 4779-4787.
16 2008 EPA Report on the Environment available at http://www.epa.gov/ncea/roe/index.htm
17 This value was transposed as 264 tons in the version of this memorandum included with the MATS proposal.
18 ftp://ftp.epa.gov/pub/EmisInventory/finalnei99ver3/haps/summaries
19 Emission Inventory Technical Support Document (TSD) for the proposed MATS rule
3
-------
To update the list of assumed affected units in the current base inventory, EPA compared the primary
fuels listed for each unit in the NEEDS v4.0 and NEEDS v4.1020 databases as well as 2005 data available
from the Energy Information Administration. The updated list of affected units is provided in the current
base workbook, including the reason a unit was added to, or excluded from, the original list of units
developed for the MATS ICR.
Mercury EFs were developed based on reported data from 323 units (boilers), with data from 6 of
these units excluded as outliers. Using these data, EPA computed unit-specific EFs and industry-wide
average EFs associated with boiler configuration, which considers combinations of fuel type, boiler type, and
control devices. The average EFs were assigned to untested boilers based on their configuration. The boiler-
specific EFs from the test data and the average EFs were applied to the units reporting data. The EFs were
combined with average heat input for 2007-2009 for all units subject to the rule. The approaches and data
are available in "Emission Factors Supporting Documentation for the Final Mercury and Air Toxics
Standards," which is available in the MATS docket.
For the final rule, EPA revised the approach for the operating data (heat input). For proposal, the
operating data gathered as part of the ICR was used, which included calculating an average throughput using
the 3-year average capacity factor for 2007 to 2009 and the boiler maximum heat input. For the final rule,
heat input data were compiled from the CEM program for 2007, 2008, and 2009 when available for units
included in that program. For units not reporting to the CEM program, the same approach as at proposal was
used: the operating data collected as part of the rule were used by calculating the 3-year average capacity
factor for 2007 to 2009 and applying to the boiler maximum heat input.
One goal of using the average 2007-2009 approach to estimate the current base emissions was to
limit the impact of the economic downturn on Hg emissions estimates, since lower throughput of EGUs
caused by decreased electricity demand is not a long-term or enforceable decrease in emissions. To evaluate
this impact, EPA computed demand-adjusted U.S. EGU emissions for 2002 through 2010 based on the
current, ICR-based boiler configurations (fuels, technologies, controls). The heat input data were based on
the year-specific CEM data where available. For units that did not report to EPA using CEMs for a given
year, the heat input was held constant at the 2007-2009 average. Table 1 shows the resulting demand-
adjusted emissions. The emissions estimate for 2009 represents a low point in Hg demand-adjusted
emissions assuming current boiler configurations for all years, whereas 2007 and 2008 are relatively
consistent with historical values. Demand-adjusted increases start to occur in 2010, which is about 1
ton/year higher than 2009 based on demand impacts alone. Thus, a 2007-2009 average will include some
impact of temporary decreased demand since it includes 2009. Since CEM data were used only where
available and a 2007-2009 average heat input is used for other units, this approach will tend to somewhat
dampen the resulting trend; however, this is a reasonable guide to help understand the impact of demand
decreases on emissions levels.
20 http://www.epa.gov/airmarkets/progsregs/epa-ipm/BaseCasev410.html#needs
4
-------
Table 1: Demand-adjusted ICR-based U.S. EGU Hg emissions
Year
2002
2003
2004
2005
2006
2007
2008
2009
2010
Total Mercury using ICR-based boiler
configurations (tons/year)
29.5
30.1
29.8
30.3
29.9
30.5
29.6
25.5
26.8
To further evaluate EPA's current base Hg emissions estimate for MATS, EPA reviewed the data
available from industry as reported into the Toxics Release Inventory (TRI). Table 2 provides the TRI-based
emissions and illustrates that while industry-reported emissions in TRI were lower than EPA estimates in
2005, the 2010 TRI compilation of 33.5 tons/year is higher than the EPA current base estimate for MATS of
29 tons/year. Since facilities estimate and report their emissions to the TRI Program, facilities are
responsible for estimating emissions using a variety of method reporting codes21. These methods include
continuous monitoring data, periodic or random monitoring data or measurements, mass balance
calculations, using published general emission factors or site-specific emission factors, and other approaches.
Since the methods used by industry will be a complex composite of these different methods, and EPA's
approach for MATS uses a single method based on site-specific and general emission factors using the latest
source test data, it is not surprising that the estimates are different.
Table 2: Industry-reported emissions from the Toxics Release Inventory
Year
2005
2006
2007
2008
2009
2010
Total mercury (tpy)
48.2
46.7
46.5
44.8
35.6
33.5
60.00
50.00
• 40.00
§ 30.00
20.00
10.00
0.00
2005 2006 2007 2008 2009 2010
2.3 Hg Emissions Trend and Projections
Table 3 shows past U.S. emissions of Hg, the revised current emissions estimates, and future
emission projections for the year 2016. In 2005, 53 out of 105 tons of Hg came from combustion of coal and
oil at U.S. EGUs, or approximately 50%. The current estimate shows a decrease in these emissions to 29
tons, which includes reductions resulting from the installation of Hg controls to comply with state Hg-
21 See TRI 2010 reporting instructions for form R, http://www.epa.gov/tri/report/rfi/rv20lOrfi 061511 .pdf. pg 46-47.
5
-------
specific rules, voluntary reductions, as well as from the co-benefits of Hg reductions from control devices
installed for the reduction of 862 and PM as a result of state and federal actions, such as New Source Review
(NSR) enforcement actions. The 2016 Hg estimate of 27 tons for U.S. EGUs was projected using the
Integrated Planning Model (IPM), which uses a different methodology to project emissions.22 The 2016
base-case scenario has been updated since the proposal to reflect known federal measures for all sectors,
including the recently finalized Cross-State Air Pollution Rule (CSAPR), but without state Hg rules.23. It
also reflects projected economic changes and fuel usage for the EGU and mobile sectors. Based on this
modeling, the estimated U.S. EGU proportion of Hg emissions in 2016 is 42%. More information on IPM is
available in the "Documentation Supplement for EPA Base Case v.4.10_MATS - Updates for Final Mercury
and Air Toxics Standards (MATS) Rule," which is available in the MATS docket.
Table 3: U.S. anthropogenic Hg emissions trends for U.S. EGUs.
U.S. EGUs
Non-EGU Anthropogenic
U.S. Hg
Total U.S. Anthropogenic
Hg
19901
(tons)
59
205
264
19992
(tons)
49
66
115
20053
(tons)
53
52
105
Current Base4
(tons)
29
Not Available6
Not Available6
Final 2016 Base Case5
(tons)
27
35
64
Percent of U.S. Anthropogenic Hg Emissions from U.S. EGUs
1990
22%
1999
43%
2005
50%
Current Base
Not Available6
2016
42%
1. 2008 EPA Report on the Environment available at http ://www. epa. gov/ncea/roe/index. htm, which also has 2002
emissions that used a consistent approach as 1990.
2. ftp://ftp.epa.sov/pub/EmisInventorv/fmalnei99ver3/haps/summaries/. The 1999 estimates used different methods
than the 1990 and 2005 methods, and so trends shown here also reflect changes in methods.
3 . Emission Inventory Technical Support Document (TSD) for the proposed MATS rule
4. The estimate of the current base emissions of Hg may underestimate Hg emissions from U.S. EGUs due to targeting
of the 2010 ICR on the best performing EGUs.
5. The 2016 base case scenario represents predicted emissions including known federal measures for all sectors,
and excludes state Hg rules for EGUs. It reflects projected economic changes and fuel usage for the EGU
and mobile sectors. The term "Final" is used to indicate that the EGUs were projected utilizing the final
version of the IPM version 4. 10 MATS developed for the final MATS rule as opposed to the Interim
version. "Base" indicates that these are the projected emissions in absence of the Final MATS policy.
6. Information on recent U.S. EGU emissions was obtained using an ICR for EGUs only. This same
information is not available for other sources, which were not covered by the ICR.
Table 4 below provides the projections for non-EGU emission sources that include expected future
reductions from existing and proposed regulations for HAPs that are expected to be implemented prior to
IPM is a multi-regional, dynamic, deterministic linear program model of the U.S. power sector that determines the least cost
solution to meeting a set of environmental constraints while still meeting specified electric demand. For more detail on IPM see:
http://www.epa.gov/airmarkets/progsregs/epa-ipm/index.html
23 2016 base Hg emissions from U.S. EGUs were estimated to be 29 tons/year at proposal. The future-year emissions are
described further in Section 4.
6
-------
2016. These regulations include the NESHAPs that affect the following industries: Portland Cement;
Industrial, Commercial and Institutional Boilers and Process Heaters; Gold Ore Mining and Production,
Electric Arc Furnaces, Hazardous Waste Incineration, and Mercury Cell Chlor-alkali facilities. The
projections also include the replacement of a smelter and a pulp and paper plant, which will result in the
elimination of Hg emissions from those sources, along with other known plant closures. This projection
shows that U.S. EGU emissions will continue to comprise a significant, dominant portion of the total U.S.
anthropogenic inventory in 2016. In 2016, Hg emissions from U.S. EGUs are projected to comprise 27 tons
out of a total of 62 tons. More information on the non-EGU projections of Hg is available in the Emission
Inventory Technical Support Document (TSD) for the proposed MATS rule. This document has not been
updated for the final rule.
Table 4: Anthropogenic Hg emissions and projections in the Continental U.S.
2005 2016 Base
Mercury Mercury
Category (tons) (tons)
Electric Generating Units
Portland Cement Manufacturing
Stainless and Non-stainless Steel Manufacturing: Electric Arc Furnaces
Industrial, Commercial, Institutional Boilers & Process Heaters
Chemical Manufacturing
Hazardous Waste Incineration
Mercury Cell Chlor-Alkali Plants
Gold Mining
Municipal Waste Combustors
Sum of other source categories (each of which emits less than 2 tons)
Total
53
7.5
7.0
6.4
3.3
3.2
3.1
2.5
2.3
17
105
27
1.1
4.6
4.6
3.3
2.1
0.3
0.7
2.3
16
62
3 Non-Hg HAP Emissions
Fossil-fuel fired boilers emit a variety of metal HAP, organic HAP and HAP that are acid gases.
Acid gas and metal HAP emissions are discussed below.
The same three updates made for the Hg national inventory were made for the acid gases and non-Hg
metal emissions: a slightly updated list of units are assumed to be subject to the rule, improved heat input
data based on the CEM program is used where available, and improved EF estimation methods are used to
exclude statistical outliers. The emissions factors (EFs) used to estimate current non-Hg HAP emissions are
consistent with the EFs used to calculate chromium, nickel, and arsenic emissions for EGUs without
available facility-specific stack test data in the revised Non-Hg Case Study Chronic Inhalation Risk
Assessment.
-------
3.1 Acid Gases
Acid gas emissions from U.S. EGUs include hydrogen chloride (HC1), hydrogen fluoride (HF),
chlorine (Cb) and hydrogen cyanide (HCN). Table 5 shows emissions of certain acid gases from EGUs,
based on the 2010 ICR data by fuel type. Although Cb was not included in the data collected from the 2010
ICR, C\2 gas is expected to be emitted at significantly lower amounts than HC1 (ratios of HCl-to- C\2 can be
as high as 200:124).
Similar to the Hg estimates, the current base emissions estimates for the acid gases are based on
emission factors generated from boiler-specific emission tests. These emission factors are based on test data
from 330 boilers for hydrogen chloride, 324 boilers for hydrogen fluoride, and 314 boilers for hydrogen
cyanide. Unit-specific emission factors were developed and applied from unit test averages where available;
otherwise, average emission factors were applied. Current-year estimates are not available for other U.S.
anthropogenic emission sources, so the latest available complete set of data, 2005, were used to approximate
the percentage contribution of U.S. EGUs to total U.S. anthropogenic emissions. As illustrated by Table 5,
U.S. EGUs are estimated to emit the majority of HC1 and HF nationally. Additional documentation on the
inventory calculations and emission factors used for each unit is available in the inventory workbook
identified in Section 2.2. Additional information on the approach to calculate binned emission factors is
available in "Emission Factors Supporting Documentation for the Final Mercury and Air Toxics Standards",
available in the docket.
Table 5: Summary of acid gas emissions from U.S. EGU sources
Hydrogen
Cyanide5
Hydrogen
Chloride
Hydrogen
Fluoride
Total
Selected Nationwide Acid HAP current base
emissions (tons/yr) from U.S. EGUs
Coal
5,459
99,565
25,657
130,680
IGCC2
0
1
1
3
PC3
84
81
7
172
Oil
83
329
67
480
Total
Current
Base EGU4
5,626
99,976
25,732
131,334
2005 Acid HAP
Emissions (tons/yr)
from the inventory
used for the National
Air Toxics
Assessment (NATA)1
Total
2005
EGU
1,200
350,000
47,000
398,200
2005
Non-
EGU
14,000
78,000
28,000
120,000
Percent of U.S.
Anthropogenic
Emissions
from EGUs,
based on the
inventory used
for the 2005
NATA
8%
82%
62%
77%
1 2005 NATA: http://www.epa.gov/ttn/atw/nata2005/ EGU emissions were extracted from the total using the MACT
code field (1808)
2 IGCC = integrated gasification combined cycle
3 PC = petroleum coke
4 The estimate of the current base emissions for acid gases may underestimate total EGU emissions due to targeting of
the 2010 ICR on the best performing EGUs.
5 Used cyanide emissions for hydrogen cyanide
For the final rule, average emission factors for the untested units were recomputed using improved
24 200:1 is based on testing done at the EPA's pilot facility in Research Triangle Park, NC. Personal communication, Nick
Hudson.
-------
methods. The approaches and underlying data are available in "Emission Factors Supporting Documentation
for the Final Mercury and Air Toxics Standards," which is available in the docket. For acid gases, the data
and calculations used for the average emission factors are available in the spreadsheet
"MATS_Final_Utility_Boiler_Mercury_and_Acid_Gases.xlsx".
3.2 Metal HAP
Metals are emitted primarily because they are present in fuels. Table 6 shows selected metals emitted
by EGUs and emission estimates based on data from the 2010 ICR. Similar to the Hg and acid gas estimates,
the current base emissions for the metal HAPs are based on emission factors generated from boiler-specific
emission tests. The number of tests used for each metal was different depending on the available data and
the impact of outlier tests applied to the analysis. Unit-specific emission factors were developed and applied
from unit test averages where available; otherwise, average emission factors were applied. Current-year
estimates are not available for other U.S. anthropogenic emission sources, so the latest available complete set
of data, 2005, was used to approximate the percentage contribution of U.S. EGUs to total U.S. anthropogenic
emissions.
Table 6: Summary of metal emissions from U.S. EGU sources.
Antimony
Arsenic
Beryllium
Cadmium
Chromium
Cobalt
Lead
Manganese
Nickel
Selenium
Total
Selected Nationwide Metal HAP current-
base emissions(tons/yr) from EGUs
Coal
8
55
2
3
236
10
43
137
238
230
962
IGCC2
0
0
0
0
0
0
0
0
0
0
0
PC3
0
1
0
0
0
0
2
0
1
0
4
Oil
15
1
0
0
4
7
3
12
161
1
204
Total
Current
Base
EGU4
23
57
2
3
240
18
48
149
399
231
1170
2005 Metal HAP
Emissions from the
inventory used for the
National Air Toxics
Assessment (NATA)1
Total 2005
EGU
19
200
10
25
120
54
21
270
320
580
1619
2005 Non-
EGU
83
120
13
38
430
60
1,194
1800
840
120
4698
Percent of U.S.
Anthropogenic
Emissions from
EGUs, based on the
inventory used for the
2005 NATA
19%
62%
44%
39%
22%
47%
2%
13%
28%
83%
26%
1 2005 NATA: http://www.epa.gov/ttn/atw/nata2005/ EGU emissions were extracted from the total using the MACT code
field (1808)
2 IGCC = integrated gasification combined cycle
3 PC=petroleum coke
4 The estimate of the current base emissions for metals may underestimate EGU metal emissions due to targeting of the
2010 ICR on the best performing EGUs.
As shown by the table, U.S. EGUs are estimated to be a significant source of emissions nationally for
these metals.
-------
For the final rule, average emission factors for the untested units were recomputed using improved
methods as compared to the proposed rule. The approaches and underlying data are available in "Emission
Factors Supporting Documentation for the Final Mercury and Air Toxics Standards," which is available in
the docket. For Chromium, Nickel, and Arsenic, the data and calculations used for the average emission
factors are available in the spreadsheet "EPA_Ar_Cr_Ni_Bin_EFs.xlsx".
3.3 Expected Impact on Non-Hg HAP Emissions from Existing Clean Air Act (CAA) and Other
Regulations and Programs
It is expected that acid gas and metal emissions will be reduced in the future due to existing CAA and
other regulations and programs. Acid gases and metals are emitted by a number of other industries already
regulated or scheduled for regulation under the National Emissions Standards for Hazardous Air Pollutants
(NESHAP) or the Risk and Technology Review (RTR) program. In addition, because acid gases respond
somewhat to the same types of controls used for SO2, it is expected that acid gas emissions will be reduced at
both EGU and non-EGU sources as a result of national, state, and local control programs to control 862.
Likewise, metals respond to the same types of controls used for PM, and as a result it is expected that metal
emissions will be reduced at both EGU and non-EGU sources due to national, state, and local control programs
to control direct PM.
4. Hg and non-Hg HAP Emissions Projections
This section summarizes the base and future-year emissions for Hg and HC1 used as part of the public
record for this rule. Future projections of other HAPs were not developed; therefore, they are not included in
this section. The current base emissions estimate has been described in Section 2.2. This section includes
further information about the emissions created for (a) the Hg Risk TSD conducted to inform the appropriate
and necessary finding, and (b) the base and final policy cases using the final version of IPM for the final co-
benefit estimates in the final RIA25.
For the 2016 emissions estimates, different phases of analysis resulted in slightly different emissions
estimates for the appropriate and necessary finding and the final base and policy cases. The purpose of this
section is to explain the source of these differences and document the different Hg and HC1 emissions values.
The subsections below provide additional information on emissions for the Hg Risk TSD and the final RIA.
Table 7 summarizes the origin information about all of the emissions cases, and Table 8 summarizes the
emissions estimates of Hg and HC1 for each, including speciated Hg.
4.1 Emissions for the Hg Risk TSD
The Hg Risk TSD is based on modeling Hg deposition associated with total 2005 Hg emissions and
total 2016 projected Hg emissions, as well as modeling Hg deposition resulting from scenarios that zeroed
out Hg emissions from U.S. EGUs. While the Hg Risk TSD was revised in response to comments received
during the peer review, EPA did not remodel Hg deposition since the MATS proposal.
The "2005 base year" emissions case in Table 7 uses emissions of criteria pollutants, Hg, HC1, and
Cb from all source categories. All emissions except Hg come from the 2005 version 4.1 modeling platform,
25 While the final MATS included an updated RIA to capture the final co-benefits of SO2 and PM2.5 changes, the Hg emissions
and benefits for that analysis were kept identical to those used for the MATS proposal.
10
-------
which is derived from the 2005 National Emissions Inventory (NEI). The HC1 and Cb emissions have a
slight impact on the modeled ozone and are therefore included in the standard modeling configuration used
for this effort. 2005 Hg emissions for all sources (EGU and non-EGU) come from the 2005 National Air
Toxics Assessment (NATA) inventory, a revised version26 of the 2005 NEI. EPA further revised the 2005
NATA inventory for non-EGU point and nonpoint sources based on data collected from the NESHAP for
industrial, commercial and institutional boilers. The current base case contains emissions only for U.S.
EGUs, and the Hg, HC1, and metal emissions in this inventory are based on the 2010 ICR data. As
previously described, initial and boundary conditions for the photochemical and deposition modeling were
based on a GEOS-CHEM simulation using a 2000 based global inventory that includes approximately 7,000
mt/yr of mercury emissions27.
Also summarized in Table 7, the "Proposal interim 2016 base case" emissions projections were
developed for use in modeling based on the latest available (at the time of the proposed rule) projection
techniques and they account for emissions reductions of all modeled pollutants from federal measures. In
addition, the "Proposal interim 2016 policy case" emissions projections are an interim version of the
expected emissions from U.S. EGUs after reductions associated with the proposed CSAPR and interim
proposed MATS. The interim 2016 EGU emissions for both the base and MATS policy cases were created
using an interim version of the IPM because the modeling needed to be completed in advance of the IPM
revisions associated with the proposed MATS. EPA concluded that the deposition modeling for the Hg Risk
TSD was sufficient; therefore, all of the emissions reflected below for the proposal interim 2016 base case
are unchanged from proposal. The change in the 2016 base Hg emissions estimate for U.S. EGUs of two
tons of Hg between proposal and final would not have a significant impact on risk results. The non-EGU
source emissions were created for the interim 2016 base projected emissions, and these are described in the
Emission Inventory TSD for the proposed MATS. These emissions in Table 8 were used for all other 2016
emission cases.
The emissions used for the Hg Risk TSD are described in more detail in the three TSDs associated
with this rule: (1) the 2005 version 4.1 Platform TSD, (2) the Emission Inventory TSD for the proposed
MATS, and (3) the Air Quality Modeling TSD: EGU Mercury Analysis. These documents and the IPM results
are available in the MATS docket.
4.2 Emissions for the Regulatory Impact Analysis (RIA) for the final rule
EPA updated the emissions estimates of the 2016 base and policy case emissions for the final rule. In
Tables 7 and 8, the resulting emissions are labeled as the "Final 2016 base" and "Final 2016 policy" cases.
The final IPM emissions are cited in the preamble for the MATS and in any summaries stating the final 2016
emissions associated with the rule generally. The final base and policy case emissions were not used in the
modeling. Instead, the PM-related health co-benefits estimated for the final interim scenario were scaled
using benefit-per-ton estimates, which is fully described in the RIA for the final rule.
26 Mercury revisions for NATA primarily affected non-EGU sources such as cement kilns and hazardous waste combustors.
27 Selin, N.E., Jacob, D.J., Park, R.J., Yantosca, R.M., Strode, S., Jaegle, L., Jaffe, D., 2007. Chemical cycling and deposition of
atmospheric mercury: Global constraints from observations. Journal of Geophysical Research-Atmospheres 112.
11
-------
Table 7: Emission sources and basis for current and future-year emissions HAP inventories.
Purpose
Hg Deposition modeling for the Appropriate and
Necessary Finding and Hg benefits analysis
Current
significance of
EGUs
^-B
5 1
o a>
+•• 0 1?
*8 •« td
+j 6D T3
£;§ g
a o o
.§"8*
T3 S
S 0
E S
Emissions Case
2005 base year
Proposal Interim
2016 Base Case
projections
Proposal Interim
20 16 Policy
Case projections
Current base
Emissions
Final 20 16 Base
Case
Final 20 16
Policy Case
Pollutants
Hydrogen
chloride, and
chlorine
Mercury
mercury and
hydrogen chloride
Criteria pollutants,
mercury and
hydrogen chloride
Mercury, acid
gases, and metals
Criteria pollutants,
mercury and
hydrogen chloride
Criteria pollutants,
mercury and
hydrogen chloride
Sector
EGU
Non-EGU point
Non-point
Mobile
EGU
Mobile
Non-EGU point
Non-point
EGU
Non-EGU point
Nonpoint
Mobile
EGU
Non-EGU point
Nonpoint
EGU
EGU
Non-EGU point
Nonpoint
Mobile
EGU
Non-EGU point
Nonpoint
Mobile
Emissions source
2005 emissions from the version 4. 1
modeling platform, derived from the 2005
National Emission Inventory (NEI) data
2005 National Air Toxics Assessment
(NATA) data
2005 NATA data, adjusted to account for
Boiler MACTICR data
Emissions from IPM v.4.10, including the
Transport Rule Proposal
Emissions from the version 4.1 modeling
platform: projected 2005 emissions to
account for all known national rules,
consent decrees, and plant closures
Emissions from IPM v.4.10, including the
Transport Rule Proposal and interim MATS
Proposal
Same as interim 2016 base projections
EGU emissions inventory based on unit-
specific emission factors for tested units and
average emission factors for untested units.
2007-2009 average CEM heat input data
used to represent current conditions where
available, and ICR-based 2007-2009 heat
input estimates otherwise.
IPM 4.10_MATS, including the final Cross-
State Air Pollution Rule
Interim 2016 Base Case projections
IPM 4.10_MATS, including the final Cross-
State Air Pollution Rule
Interim 2016 Base Case projections
Table 8 summarizes the Hg and HC1 emissions for the anthropogenic sources of these emissions
across the various cases. All EGU emissions presented in Table 8 represent units greater than 25 MW. The
2005 base case includes 105 tons of Hg and 430,000 tons of HC1 from all sources, of which 53 tons of Hg
and 350,000 tons of HC1 are from EGUs. The current base emissions include the Hg and HC1 emission
estimates from U.S. EGUs are lower than in 2005, at 29 tons and 100,000 tons respectively. Speciated Hg
emissions were not computed for the current base case. The interim 2016 base total Hg emissions from all
sources are 64 tons and HC1 emissions are 140,000 tons, with 29 tons of Hg and 74,000 tons of HC1 from
EGUs. For the interim 2016 policy case, total Hg was estimated to be 42 tons, with 6.8 tons from EGUs.
For HC1, the interim 2016 policy case shows a total of 75,000 tons from all sources with EGU emissions
emitting 8,800 tons.
12
-------
The final 2016 base case EGU emissions were 27 tons of Hg out of a total of 62 tons, and 45,000 tons
of HC1 out of a total of 115,000 tons. In the final 2016 MATS policy case, EGU emissions of Hg are
estimated to be reduced to 6.6 tons and emissions of HC1 reduced to 5,500 tons.
Table 8. Summary of Hg and HC1 emissions for emissions cases.
Hg°
(tons)
Hg2+
(tons)
Hgp
(tons)
Total Hg
(tons)
Hydrogen
chloride (tons)
2005 Base Case Emissions (tons)
EGU*
NonEGU Point
Nonpoint
Mobile
All sources
30
30
3.1
0.79
64
21
11
1.06
0.29
33
1.6
6.2
0.65
0.13
8.5
53
46
4.8
1.2
105
350,000
49,000
29,000
430,000
Current Base Emissions (tons)
EGU
n/a
n/a
n/a
29
100,000
Proposal interim 2016 base case projections (tons)
EGU""
NonEGU Point
Nonpoint
Mobile
All sources
21.1
16.9
3.1
0.79
42
6.9
7.9
1.06
0.29
16
0.67
4.4
0.65
0.13
5.9
28.7
29.2
4.8
1.2
64
74,000
38,000
29,000
140,000
Proposal interim 2016 policy case projections (tons)
EGU"
NonEGU Point
Nonpoint
Mobile
All sources
4.8
16.9
3.1
0.79
26
1.6
7.9
1.06
0.29
11
0.36
4.4
0.65
0.13
5.6
6.8
29.2
4.8
1.2
42
8,800
38,000
29,000
75,000
Final 2016 base case projections (tons)
EGU
NonEGU Point
Nonpoint
Mobile
All sources
20.6
16.9
3.1
0.79
41
5.5
7.9
1.06
0.29
15
0.37
4.4
0.65
0.13
5.5
26.5
29.2
4.8
1.2
62
45,300
41,000"""
29,000
115,300
Final 2016 policy projections (tons)
EGU
NonEGU Point
Nonpoint
Mobile
All sources
5.2
16.9
3.1
0.79
26
1.2
7.9
1.06
0.29
10
0.15
4.4
0.65
0.13
5.3
6.6
29.2
4.8
1.2
42
5,500
41,000"""
29,000
75,500
13
-------
* For this table, the EGU Sector is defined as follows: (1) For all pollutants other than mercury (Hg): 2005 NEI v2 point source
EGUs mapped to the Integrated Planning Model (IPM) model using the National Electric Energy Database System (NEEDS)
2006 version 4.10 database. A few revisions were made to the 2005 NEI v2 annual emission estimates as discussed in the 2005-
based, Version 4.1 platform document. (2) For Hg: 6/18/2010 version of the inventory used for the 2005 National Air Toxics
Assessment (NATA) mapped to IPM using NEEDS version 4.10. The NATA inventory is an update to the 2005 NEI v2 and was
divided into EGU and non-EGU sectors consistent with the other pollutants. We additionally removed Hg from sources from the
NESHAP for Industrial, Commercial, and Institutional Boilers and Process Heaters (aka "Boiler MACT") Information Collection
Request (ICR) database because we included these emissions in the non-EGU sector. For more information, see the 2005-based
Platform Documentation, version 4.1.
** EGU Sector for 2016 is consistent with the 2005 base case discussed above. The future year emissions were generated by the
Interim version of IPM4.10 and adjustments were made to remove Hg associated with the Boiler MACT ICR (which is accounted
for in the Non-EGU sector), and to apply the impact of the final Boiler MACT for the other pollutants.
Since proposal, EPA has compiled additional documentation on the speciation approach applied to
U.S. EGUs. The speciation approach applied for the final MATS was developed as part of the Clean Air
Mercury Rule (CAMR) based on a 1999 ICR. Eighty mercury emissions tests were collected from 69
facilities. These data include inlet and outlet speciated mercury concentration measurements (|ig/dscm) for
Hgp, Hg2+, and Hg° mercury. Only the outlet data (measurements after the last control device) were used to
develop the mercury speciation profiles. Forty-three mercury speciation profiles were developed from these
data based on the same bins developed for the Emission Reduction Factors used in IPM. Additional
information on these profiles is available in the "Electric Generating Utility Mercury Speciation Profiles for
the Clean Air Mercury Rule" and the workbook "Hg_speciation_data_CAMR.xlsx", which are available in
the docket.
Table 9 highlights the U.S. EGU emissions for the various cases shown in Table 8. It also includes
the average percent of divalent gaseous and particulate Hg. The divalent gaseous Hg ranges from 20% to
40% of the total Hg emissions, and the particulate Hg ranges from 1% to 5% of total Hg emissions.
Table 9: EGU emissions of Hg and HC1 for emissions cases in support of the final MATS.
2005 Base Case
Emissions
Current base
Emissions
Proposal interim
20 16 base case
projections
Proposal interim
2016 policy case
projections
Final 20 16 base case
projections
Final 2016 policy
case projections
Hg°
(tons)
30
N/A*
21.1
4.8
20.6
5.2
Hg2+
(tons)
21
N/A*
6.9
1.6
5.5
1.2
Hgp
(tons)
1.6
N/A*
0.67
0.36
0.37
0.15
Total Hg
(tons)
53
29
28.7
6.8
26.5
6.6
Hydrogen
chloride
(tons)
350,000
100,000
74,000
8,800
45,300
5,500
Hg2+
Percent of
total Hg
40%
24%
24%
21%
18%
Hgp
Percent of
total Hg
3%
2%
5%
1%
2%
Speciation data were not collected for Hg in the 2010 ICR and CAMR-based speciation profiles were not applied to the current
base emissions estimate.
14
-------
-------
o o
^ (N
>—' s-T
' (L)
PH
W
O
o
o
fin
C ,-H
£ o
™ 'S3 O
§ S^
^
o
s
O
o
'•§
(L)
s
fin
* 2
+J P
00 § >,
T3 O o
5 CH X
.-£ •> §
e c oo
^ pq
------- |