OFFICE OF INSPECTOR GENERAL
Catalyst for Improving the Environment
Evaluation Report
Monitoring Needed to Assess
Impact of EPA's Clean Air
Mercury Rule on Potential Hotspots
Report No. 2006-P-00025
May 15, 2006
-------
Report Contributors:
Rick Beusse
Carolyn Blair
Hilda Canes
Susan Charen
Sarah Fabirkiewicz
James Hatfield
Erica Hauck
James Van Orden
Abbreviations
CAMR
CMAQ
EPA
Hg
IPM
km
MACT
MDN
mg/kg
OIG
Clean Air Mercury Rule
Community Multiscale Air Quality
Environmental Protection Agency
Mercury
Integrated Planning Model
kilometer
Maximum Achievable Control Technology
Mercury Deposition Network
milligram per kilogram
Office of Inspector General
Cover photo: A fisherman holding a walleye, a predator fish for which mercury
contamination is a concern (photo courtesy EPA).
-------
I
I
U.S. Environmental Protection Agency
Office of Inspector General
At a Glance
2006-P-00025
May 15, 2006
Catalyst far Improving the Environment
Why We Did This Review
In support of its Clean Air
Mercury Rule (CAMR), the
Environmental Protection
Agency (EPA) conducted a
detailed analysis of mercury
emissions and deposition.
EPA concluded that "utility-
attributable" hotspots would
not occur after implementation
of CAMR's mercury trading
program. This evaluation
assesses the basis for EPA's
conclusion.
Background
About 40 percent of U.S.
man-made airborne mercury is
emitted from coal-fired
utilities. EPA revised a
previous finding that mercury
emissions from coal-fired
utilities be regulated with a
Maximum Achievable Control
Technology standard. Instead,
EPA adopted a cap-and-trade
program to reduce mercury
emissions. Several State
agencies and environmental
groups objected to these
actions. One concern was that
a cap-and-trade program could
result in localized areas with
unacceptably high levels of
mercury, or "hotspots.",
For further information,
contact our Office of
Congressional and Public
Liaison at (202) 566-2391.
To view the full report,
click on the following link:
www.epa.qov/olq/reports/2006/
20060515-2006-P-00025.pdf
Monitoring Needed to Assess Impact of EPA's
dean Air Mercury Rule on Potential Hotspots
What We Found
EPA brought significant scientific, technical, and modeling expertise to bear in
developing a specific methodology to consider the potential for mercury hotspots.
Several uncertainties associated with key variables in the analysis could affect the
accuracy of the Agency's conclusion that the Clean Air Mercury Rule (CAMR)
will not result in "utility-attributable" hotspots. We noted:
gaps in available data and science for mercury emissions estimates,
limitations with the model used for predicting mercury deposition,
uncertainty over how mercury reacts in the atmosphere, and
uncertainty over how mercury changes to a more toxic form in waterbodies.
Two recent studies support the need for additional monitoring to ensure that EPA's
analysis has properly estimated the contribution of local, regional, and global
sources on U.S. deposition. These studies are "Mechanisms of Mercury Removal
by O) and OH in the Atmosphere," published in Atmospheric Environment in June
2005; and "Sources of Mercury Wet Deposition in Eastern Ohio, USA," submitted
for publication in a scientific journal in February 2006. Results of both studies
were not available until after EPA issued CAMR in March 2005, and thus could
not have been considered in EPA's deliberations on CAMR. Although EPA
indicated in CAMR that it would monitor the impact of the cap-and-trade rule on
mercury deposition, the Agency has not yet developed a monitoring plan for this
purpose. Without field data from an improved monitoring network, EPA's ability
to advance mercury science will be limited and "utility-attributable" hotspots that
pose health risks may occur and go undetected.
Based on our interpretation of CAMR, EPA could not take action to mitigate a
mercury hotspot unless the Agency first determined that me hotspot was solely
"utility-attributable." Therefore, EPA could not require additional utility emission
reductions if utilities contributed significantly, but not solely, to a mercury
hotspot. This could limit EPA's ability to mitigate human health hazards by
reducing potentially harmful levels of mercury in waterbodies and fish tissue.
This could also limit EPA's ability to reduce the number of waterbodies with fish
consumption advisories.
What We Recommend
We recommend that EPA develop and implement a mercury monitoring plan to
(1) assess the impact of CAMR, if adopted, on mercury deposition and fish tissue;
and (2) evaluate and refine mercury estimation tools and models. Further, if
CAMR is adopted after the rule reconsideration process is complete, we
recommend that EPA clarify in the final rule that the "utility-attributable" hotspot
definition does not establish a prerequisite for making future revisions to CAMR.
In response to the draft report, the Agency agreed that additional mercury
monitoring is needed and explained that CAMR does not establish the "utility-
attributable" hotspot definition as a prerequisite for future changes to CAMR.
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
OFFICE OF
INSPECTOR GENERAL
MEMORANDUM
SUBJECT:
TO:
May 15, 2006
Monitoring Needed to Assess Impact of EPA's
Clean Air Mercury Rule on Potential Hotspots
Report No. 2006-P-00025
William L. Wehrum
Acting Assistant Administrator for Air and Radiation
This is our report on the subject evaluation conducted by the Office of Inspector General (OIG)
of the U.S. Environmental Protection Agency (EPA). This report contains findings that should
help EPA to better monitor the impact of the Clean Air Mercury Rule and refine performance
standards under the rule, if necessary. This report represents the opinion of the OIG and the
findings in this report do not necessarily represent the final EPA position. Final determinations
on matters in the report will be made by EPA managers in accordance with established
procedures.
Action Required
In accordance with EPA Manual 2750, as the action official, you are required to provide a
written response within 90 days of the final report date. The response should address all
recommendations. For the corrective actions planned but not completed by the response date,
please describe the actions that are ongoing and provide a timetable for completion. Where you
disagree with a recommendation, please provide alternative actions for addressing the findings
reported.
We appreciate the efforts of EPA managers and staff in working with us to develop this report.
If you or your staff have any questions regarding this report, please contact me at 202-566-0847,
or Rick Linthurst at 919-541-4909.
Sincerely,
lenci
Acting Inspector General
-------
Table of Contents
At a Glance
Chapters
1 Introduction 1
Purpose 1
Background 1
Scope and Methodology 5
2 Monitoring Plan Needed to Address Uncertainties
In EPA's Hots pots Analysis 6
EPA Analyzed Potential for "Utility-Attributable" Hotspots 6
Data and Science Gaps Exist for Mercury Emissions Estimates 8
CMAQ Model Uncertainties and Limitations 9
Study Finds Different Rates of Atmospheric Chemical Reactions 10
Uncertainties Noted with Methylation and Bioaccumulation 10
Study Shows Significant Deposition from Local Sources 12
Uncertainties Underscore Need for Mercury Monitoring Plan 13
Conclusion 15
Recommendation 15
Agency Comments and OIG Evaluation 15
3 EPA Needs to Clarify Conditions Under Which CAMR
Performance Standards Can be Tightened 17
EPA Provides "Utility-Attributable" Hotspot Definition 17
"Utility-Attributable" Definition Could be Interpreted to Limit EPA's
Ability to Mitigate Hotspots 18
CAMR and Revision Must be Read Together 18
Conclusion 20
Recommendation 20
Agency Comments and OIG Evaluation 20
i
Appendices
A Details on Scope and Methodology 21
B Models Used in CAMR Analysis 24
C Agency Response to Draft Report 25
D Distribution 28
-------
Chapter 1
;»,,'t~~,. rfM ! - Iv, "', «
Introduction
Purpose
A prior Environmental Protection Agency (EPA) Office of Inspector General
(OIG) report cited concerns about EPA's limited assessment of the potential for
mercury hotspots resulting from a cap-and-trade program under the Clean Air
Mercury Rule (CAMR). We issued this prior report, Additional Analyses of
Mercury Emissions Needed Before EPA Finalizes Rules for Coal-Fired Electric
Utilities (Report No. 2005-P-00003), on February 3, 2005. In support of CAMR,
EPA conducted a detailed analysis of mercury emissions and deposition and
concluded that "utility-attributable" hotspots would not occur after
implementation of the mercury emissions trading program.
EPA's Water Quality Trading Assessment Handbook defines hotspots as
"localized areas with unacceptably high levels of pollutants." In this evaluation
report, however, a hotspot is a waterbody containing consumable fish with
elevated levels of methylmercury in their tissues.
We conducted this evaluation to assess the basis for the Agency's determination
that CAMR would not result in "utility-attributable" hotspots.
Background
Mercury (Hg) is released into the atmosphere through natural processes and
through human activities, such as combustion processes. Once emitted,
atmospheric mercury undergoes several chemical and physical processes and can
then be deposited to the ground or waterbodies via wet or dry processes. In wet
deposition, mercury is deposited by precipitation, such as rain or snow. In dry
deposition, mercury settles to the earth's surface and sticks to or is absorbed by
trees, soil, water, or other surfaces. The largest source of airborne mercury
emissions in the United States is the coal-fired electric utilities industry,
representing an estimated 40 percent of total U.S. man-made airborne mercury.
Although airborne mercury is generally not considered to be a serious health
concern, once mercury enters freshwater and salt-water bodies, it can
bioaccumulate in fish and other animal tissues in its more toxic form,
methylmercury. As methylmercury bioaccumulates in the food chain, its
concentration becomes increasingly higher in animals at the top of the food chain
(such as larger predatory fish) that consume smaller, contaminated organisms.
Figure 1-1 illustrates the exposure pathway of mercury.
-------
Figure 1-1: How Mercury Enters the Environment
Lake Ocean
Atmospheric
deposition
Emissions
From Power
Plants and Cither
Sources
Wet and Dry
Deposition
Mercury transforms into melhylmercury
in soils and water, then can
bio accumulate in fish
Fishing I
commericalH
recreational H
subsstenceH
Humans and f
wildlife affected I
primarily by eating
fish containing
mercury Impacts
Best documented
impacts on the
developing fetus:
impaired motor and
cognitive skills
* Possibly olher impacts
Trans-port and
Deposition
, and
H6atturtiulalior»
Consumption
Patterns
_ Dote
Rasponsa
Source: EPA
Fish consumption is the main route by which methylmercury harms human health.
Excessive human exposure to methylmercury has been associated with severe
detrimental neurological and developmental health effects. Research has shown
that the developing fetus is at risk for impaired motor and cognitive skills. Thus,
exposure to mercury by women of child-bearing age is of particular concern.
Most U.S. Fish Advisories Due to Mercury Contamination
When levels of chemical contamination in fish are considered unsafe, States,
tribes, and territories can issue consumption advisories that may recommend that
people limit or avoid eating certain species of fish caught in certain places. Each
State sets its own criteria and decides which bodies of water to monitor.
Monitored waterbodies may vary from year to year. Fish advisories are voluntary
State recommendations not governed by Federal regulations. In 2004, 44 States
issued fish advisories for mercury. The number of mercury-related fish advisories
continues to rise as States increase fish tissue testing.
EPA recently reported in its 2005 Performance and Accountability Report that the
Agency did not meet its goal of reducing the number of overall fish advisories by
at least 1 percent from 2002 levels. From 2003 to 2004, the number of mercury
advisories rose from 2,362 to 2,436, or 3.1 percent. According to the 2004
National Listing of Fish Advisories, the vast majority (68 percent) offish
advisories in the United States are due to mercury contamination, as illustrated in
Figure 1-2.
-------
Figure 1-2: Percent of Fish Advisories for Each of the Top Five
Bioaccumulative Contaminants in 2004
2% 3% 2%
25%
Mercury PCBs D Chlordane D Dioxins DDT & Metabolites
Source: 2004 National Listing of Fish Advisories
CAMR First Rule for Mercury Emissions from Coal-Fired Utilities
On March 15, 2005, EPA issued CAMR, which established the country's first
regulation of mercury emissions from coal-fired power plants. CAMR uses a
declining cap-and-trade approach to regulating coal-fired utilities under
Section 111 of the Clean Air Act by setting a fixed national cap. Utilities can buy
and sell credits among one another in a national emissions market. Utilities that
cannot cost-effectively reduce emissions may buy allowances from units that
reduced emissions below established allowance limits. Under CAMR, an interim
national cap of 38 tons per year becomes effective in 2010 and a final annual cap
of 15 tons becomes effective in 2018. EPA's first cap is based on mercury
reductions expected to be achieved as a co-benefit of implementing the Clean Air
Interstate Rule, issued in March 2005. That rule requires utilities to take actions
to reduce emissions of sulfur dioxide and nitrogen oxides, and those actions are
also projected to reduce mercury emissions.
EPA Revised its Prior Regulatory Finding Regarding Utilities
To use a cap-and-trade program to regulate coal-fired utilities, EPA first had to
revise a December 2000 regulatory finding' that indicated it was appropriate
and necessary to regulate coal-fired utilities under Section 112 of the Clean Air
Act. This finding required EPA to regulate utilities using a Maximum
Achievable Control Technology (MACT) standard. MACT standards are
1 Regulatory Finding on the Emissions of Hazardous Air Pollutants From Electric Utility Steam Generating Units,
December 20, 2000; Vol. 65, No. 245.
-------
industry-specific, technology-based standards designed to reduce hazardous air
pollutant emissions. These standards can require facility owners/operators to
meet emission limits, install emission control technologies, monitor emissions
and/or operating parameters, and use specified work practices. In March 2005,
EPA issued a Revision of December 2000 Regulatory Finding,2 stating that the
Agency no longer found it appropriate or necessary to regulate utilities under
Section 112. This released the Agency from the requirement to regulate
utilities using a MACT standard. EPA issued the finding the same day it
issued CAMR, which established a mercury cap-and-trade program under
Section 111.
For its Revision of December 2000 Regulatory Finding, EPA interpreted
Section 112(n) to mean that utilities alone had to be the sole cause of a health
hazard in order to be regulated under Section 112 and subject to MACT
standards. Specifically, EPA developed the following "utility-attributable"
hotspot definition for its revision: "... a waterbody that is a source of
consumable fish with Methylmercury tissue concentrations, attributable solely
to utilities, greater than EPA's Methylmercury water quality criterion of
0.3 milligrams per kilogram (mg/kg)."
EPA Response to Petitions for Reconsideration
Several State agencies and other organizations oppose EPA's adoption of a
cap-and-trade program for mercury. These groups separately petitioned for
reconsideration of the Revision of December 2000 Regulatory Finding. Among
other things, they asserted that, in its analysis, EPA underestimated the impact of
deposition resulting from local and regional sources and overestimated the impact
of emissions from global sources. Thus, they argue, some mercury hotspots
already exist, and requiring sources to comply with MACT standards would
immediately reduce deposition in those areas. Further, these opponents to the
cap-and-trade program believe the program could result in new mercury hotspots
if some utilities bought excess emission credits instead of reducing emissions.
On October 21,2005, EPA reopened for public comment certain aspects of its
CAMR and, in a separate action, reopened for public comment certain aspects of
its Revision of December 2000 Regulatory Finding. The action to reopen
comment on CAMR was taken in response to petitions filed by 14 States,
5 environmental groups, a public utility, and a waste services association. The
Agency stated that it agreed to reconsider several aspects regarding CAMR. The
action to reopen comment on EPA's Revision of December 2000 Regulatory
Finding was based on two petitions, one from 14 States and a second from 5
environmental groups and 4 Indian tribes. The Agency agreed to reconsider the
legal issues underlying the decision as well as the methodology used to assess the
2 Revision of December 2000 Regulatory Finding on the Emissions of Hazardous Air Pollutants from Electric Utility
Steam Generating Units and the Removal of Coal- and Oil-fired Electric Utility Steam Generating Units from the
Section 112(c) List; Final Rule, March 15,2005.
-------
amount of "utility-attributable" mercury levels in fish tissue and the public health
implications of those levels. The Agency also agreed to reconsider how it defined
a utility hotspot for the purposes of its finding concerning regulation of Utility
Units under Clean Air Act Section 112. Comments regarding this reconsideration
were accepted until December 19, 2005. The Agency was still evaluating
comments at the time our field work ended.
Scope and Methodology
We conducted our review from September through December 2005, in
accordance with Government Auditing Standards issued by the Comptroller
General of the United States. We performed field work at EPA's Office of Air
and Radiation in Washington, DC; the Office of Air and Radiation's Office of
Air Quality Planning and Standards in Research Triangle Park, North Carolina;
the Office of Research and Development in Research Triangle Park; and the
Office of Water in Washington.
To answer our evaluation's objective, we examined: (1) the basis for the
Agency's "utility-attributable" hotspot definition and the consistency of this
definition with any prior Agency decisions regarding hotspots; (2) the key
attributes, assumptions, and limitations of the models used to assess the impact
of mercury emissions from coal-fired electric utility units under CAMR; and
(3) the key variables used as inputs to the models as well as the basis for
selecting these variables.
To gain an understanding of the definition of "utility-attributable" hotspots, the
modeling and analyses EPA used to determine the potential for "utility-
attributable" mercury hotspots after CAMR, and the inputs and assumptions
associated with the Agency's analyses, we interviewed EPA staff involved in the
development of CAMR or knowledgeable about the processes and models used
in EPA's analyses. We also interviewed officials from State agencies and
external organizations familiar with CAMR's development and EPA's hotspots
analysis. We reviewed data and analyses developed in support of the rule, and
selected public comments included in the rulemaking docket. We also reviewed
related information provided by both EPA and non-EPA officials.
Our analysis focused on the key assumptions and limitations of the Community
Multiscale Air Quality model, which was used to estimate mercury transport and
deposition. We did not review in detail the assumptions, limitations, and
uncertainties associated with the other models used in the Agency's analyses.
Appendix A provides additional details on scope and methodology.
-------
Chapter 2
' * ,-"""', \-i '*-.& -, -"
Monitoring Plan Needed to Address Uncertainties in
EPA's Hotspots Analysis
As with any modeling assessment, uncertainties may exist. Uncertainties
regarding EPA's analysis and conclusion that CAMR will not result in "utility-
attributable" hotspots include:
gaps in available data and science for mercury emissions estimates,
limitations with the model used for predicting mercury deposition,
uncertainty over how mercury reacts in the atmosphere, and
uncertainty over how mercury changes to a more toxic form in
waterbodies (i.e., methylation) and accumulates in fish tissue.
Two recent studies support the need for additional monitoring to ensure that
EPA's hotspots analysis has properly estimated the contribution of local, regional,
and global sources to U.S. deposition. These studies are:
(1) "Mechanisms of Mercury Removal by Oi and OH in the Atmosphere,"
Calvert, J.G., Lindberg, S.E., (published in Atmospheric Environment,
Volume: 39, Number: 18, Page: 3355-3367), June 5, 2005, referred to in
this report as the "Mechanisms of Mercury Removal Study;" and
(2) "Sources of Mercury Wet Deposition in Eastern Ohio, USA," Keeler, G.J.,
et al., referred to in this report as the "Steubenville Study" (a peer review
of the Steubenville Study was completed in December 2005 and the study
was submitted for publication in a scientific journal in February 2006).
Results of both studies were not available until after EPA issued CAMR in March
2005, and thus could not have been considered in EPA's deliberations on CAMR.
We believe the uncertainties associated with its CAMR analysis underscore the
need for EPA to develop and implement a plan for monitoring the impact of
CAMR on mercury deposition and mercury concentrations in fish tissue. Without
implementation of a monitoring plan and/or improvements to current models,
"utility-attributable" hotspots that can pose health risks may occur and go
undetected.
EPA Analyzed Potential for "Utility-Attributable" Hotspots
In its Revision of December 2000 Regulatory Finding, EPA states it "does not
believe that there will be any [utility-attributable] hot spots after implementation
of CAIR [Clean Air Interstate Rule] and CAMR." EPA's analyses of mercury
hotspots considered many factors that influence the way mercury is deposited to
land and waterbodies. For its CAMR analysis, EPA used the Community
-------
Multiscale Air Quality (CMAQ) model as the principal tool to predict patterns of
mercury deposition and as an important part of assessing the potential for "utility-
attributable" mercury hotspots under CAMR.
EPA considers the CMAQ to be the most capable model available for assessing
the impacts of CAMR on mercury deposition within the United States. The
model is designed to estimate pollutant concentrations and depositions over large
areas, such as the continental United States. The model accounts for variations in
mercury emissions, differences in the atmospheric reactions of mercury, and the
impact of those factors on deposition.
However, there are important limitations associated with some of the inputs EPA
used in CMAQ for its CAMR analysis. The July 2005 Final Report: Second Peer
Review of the CMAQ Model, conducted by an independent panel that included
State, academic, and private organizations, notes the following limitations:
CMAQ is a modeling system that simulates a wide range of physical,
chemical and biological processes... Some of these processes are well
understood, some reasonably well understood, and some only poorly
understood. This wide range in the level of knowledge about the
processes being modeled, and the fact that uncertainties in characterizing
some of the processes correspond to areas of active research worldwide,
means that some parts of the model code are sufficiently well established
as to be considered fixed, while other parts of the code are under
continuing development.
Other models also played a role in EPA's analysis of the potential for hotspots
under CAMR by contributing input data to CMAQ (see Appendix B for details on
some of these other models). For example, a separate model was used to estimate
the amount of mercury emissions from utilities based on certain economic
assumptions, and another was used to predict weather patterns. Both the
emissions and weather data were fed into CMAQ, and CMAQ predictions on
deposition were fed into another model to estimate the effects of deposition on
future mercury fish tissue concentrations.
In its hotspots analysis, the Agency discussed instances where conservative
assumptions were used to avoid underestimating the impact of utilities. For
example, in its hotspots analysis EPA did not screen out watersheds in which
sources of mercury other than air deposition were significant. According to EPA,
this may result in higher concentrations of methylmercury in fish being attributed
to power plants than would be the case had EPA been able to account for non-air
sources. In addition, EPA's hotspots analysis discusses the conservative estimates
used in determining the oral reference dose for mercury (i.e., an estimate of the
daily exposure to the human population, including sensitive subgroups, that is
without an appreciable risk of deleterious effects during a lifetime). The
reference dose and human exposure information were used to establish the water
-------
quality criterion for methylmercury in fish tissue, the criterion used by EPA to
represent a mercury hotspot.
Data and Science Gaps Exist for Mercury Emissions Estimates
While EPA has conducted activities to greatly increase its knowledge of mercury
emissions from coal-fired utility plants, the Agency acknowledges that some
uncertainty still exists when estimating total and speciated3 mercury emissions
and in projecting these emissions after implementation of various control
technologies.
CMAQ requires the input of emissions inventory data to predict how emissions
. will transport and deposit. CMAQ was first run with a full emissions inventory to
establish a base case scenario assuming the presence of all emissions. Next,
CMAQ was run with emissions from coal-fired utilities removed, in what is called
a "zero-out" run, to determine the impact of the variable that was zeroed out.
EPA used this zero-out method to determine that no "utility-attributable" hotspots
would occur after accounting for emissions reductions expected to be achieved
from the Clean Air Interstate Rule and CAMR.
The utility emissions input into CMAQ were developed from the Integrated
Planning Model (IPM). The IPM is a model of the U.S. electric power sector that
can be used to evaluate the cost and emissions impacts of proposed policies to
limit emissions of pollutants, including mercury. EPA has used the IPM in
rulemakings since the mid 1990's. As part of that process, EPA takes comments
on the underlying assumptions of the model and makes changes as a result. For
its Clean Air Interstate Rule and CAMR analyses, EPA used IPM to estimate base
case and future year national inventories of unit-specific mercury emissions under
different control scenarios.
IPM uses equations (emission modification factors) to estimate utility emissions
given the chemical composition of the coal being burned as well as various
operating characteristics of the utility unit (e.g., type of control technology
installed). These equations were based on various coal composition and
emissions testing data collected during a 1999 Information Collection Request
and more recent testing conducted by EPA, the Department of Energy, and
industry participants.
While extensive data have been collected on mercury emissions from coal-fired
utilities, some data and science gaps still exist with respect to understanding the
effectiveness of specific controls in reducing mercury emissions from coal. As
noted in the EPA Office of Research and Development's February 18, 2005,
3 Mercury speciates into three basic forms: elemental, ionic, and participate. Estimating the amount of speciated
mercury emissions is important since the type of mercury emitted impacts how effectively it is captured by control
technologies, and how it will react when emitted into the atmosphere. Differences in atmospheric reactions impact
the amount and location of the mercury's deposition.
-------
update of its study on control of mercury emissions, data and science gaps exist
with respect to existing controls that are intended to reduce emissions of other
pollutants with the co-benefit of reducing mercury, as well as emerging
technologies specifically designed to reduce mercury emissions. The impact of
these uncertainties on EPA's estimates of mercury emissions in base case and
future years is qualitatively discussed in Agency documents but has not been
quantified. The uncertainties could impact the accuracy of the estimated utility
emissions input into CMAQ and CMAQ's resulting deposition estimates.
CMAQ Model Uncertainties and Limitations
CMAQ is useful for predicting regional and national patterns of deposition, but it
has limitations that need to be carefully considered when used for modeling small
areas of localized deposition and, thus, identifying hotspots. When emissions data
are fed into CMAQ, the model averages the data over an area known as a "grid
cell." CMAQ can predict deposition results over grid cells of various sizes (or
resolutions) as specified by the modeler.
For CMAQ, EPA used a 36 kilometer (km) grid resolution (36 km x 36 km) for
its Clean Air Interstate Rule and CAMR modeling, which equates to a surface
area approximately 22 miles wide by 22 miles long, or approximately 484 square
miles. The model provides one average concentration for the entire area. For
example, if there is only one power plant in the corner of a grid square, that
plant's emissions are averaged over the entire 36 km x 36 km area. Averaging
over grid cells may result in a smoothing out of areas of high and low deposition.
EPA acknowledges this limitation in its Effectiveness Technical Support
Document:4
CMAQ immediately dilute[s] simulated emissions into the entire grid
volume in which they are released. This causes an artificially fast dilution
and under-represents direct deposition from air to surfaces near emission
sources . . .
When looking for hotspots, the ability to identify areas of localized deposition is
important. Using the CMAQ model at 36 km x 36 km, in the opinion of some
EPA officials we interviewed, was too coarse a resolution to be able to pinpoint
small areas of localized deposition. Some EPA officials stated that use of a finer
resolution, such as 12 km grid size, is possible in CMAQ. However, at very fine
resolutions - for instance, a 4 km grid size - the meteorological components of
the model probably fall apart and may introduce greater uncertainties in model
results. EPA outlines three reasons for using a 36 km grid square size in its
Effectiveness Technical Support Document. First, the larger grid size would
account for mercury deposition that enters a watershed through groundwater
4
Methodology to Generate Deposition, Fish Tissue Methylmercury Concentrations, and Exposures for Determining
Effectiveness of Utility Emission Controls (Effectiveness Technical Support Document). U.S. EPA, March 15, 2005.
-------
inflow and runoff, as opposed to a smaller grid size that may only account for
direct inputs to surface water. Second, in larger waterbodies where there is
substantial fishing activity, the fish species consumed by humans are likely
migratory and the accumulation of mercury in these fish will come from
deposition over a larger area. Third, many anglers may catch fish from a variety
of waterbodies in a watershed, thus a larger grid size would account for this
fishing pattern.
Study Finds Different Rates of Atmospheric Chemical Reactions
The Mechanisms of Mercury Removal Study developed information on the rates
of atmospheric chemical reactions involving mercury that is different than rates
used by EPA in its CAMR hotspots analysis. The study was published in June
2005 after EPA issued CAMR. Rate constants, which quantify the speed or rate
of chemical reactions, are the most important inputs affecting modeling results.
The accuracy of rate constants can affect the accuracy of modeling results.
Oxidation is an atmospheric process that makes mercury more reactive and is the
most important reaction associated with mercury deposition. The mercury
oxidation rate affects how quickly mercury is deposited and influences its
properties and behavior. For example, oxidation makes elemental mercury more
water soluble and more quickly deposited; if mercury emitted from a source
comes out already oxidized, it can be immediately deposited near the source
(depending on meteorological conditions and other factors).
Results of the Mechanisms of Mercury Removal Study regarding mercury
reactions and associated rates suggest that emissions from global sources
potentially account for less mercury deposition in the United States than
previously believed. This means that the contribution of global sources to U.S.
deposition may have been overestimated in EPA's analysis and the impact from
domestic sources underestimated. According to the Agency scientist responsible
for developing mercury capabilities in CMAQ, if the study's results about rate
constants are accurate, then chemical formulations currently used in all other
atmospheric simulation models, including CMAQ, could be incorrect (when
modeling mercury deposition).
Uncertainties Noted with Methylation and Bioaccumulation
Assumptions about methylation and bioaccumulation directly impact the resulting
predictions about mercury fish tissue concentrations after implementation of
CAMR. Mercury methylation is a complex process that occurs in the
environment when oxidized mercury is transformed into highly toxic
methylmercury, which bioaccumulates (builds up) in fish tissue. Some of the
important factors affecting methylation rates and bioaccumulation were not fully
accounted for in EPA's analysis. Also, a lack of knowledge about some factors
used in EPA's analysis is a source of uncertainty in EPA's conclusions about
mercury fish tissue concentrations.
10
-------
Methylation. Transformation of mercury to methylmercury occurs at varying
speeds in different waterbodies, and EPA's analysis did not fully account for this
variation. Methylation occurs when mercury enters waterbodies and bacteria
transform it to methylmercury, a highly toxic and bioaccumulative form of
mercury. Methylation of mercury occurs in waterbodies at highly variable speeds
depending on various ecosystem-specific factors, including: the bacteria in the
waterbody, the type of land surrounding the waterbody, the quantity of certain
substances such as sulfate and carbon in the waterbody, and the pH (chemistry) of
the waterbody. Thus, two adjacent waterbodies with equal mercury deposition
can have different concentrations of mercury in fish.
EPA's analysis did not address individual differences between waterbodies, or the
time it takes for different waterbodies to adjust to changes in atmospheric
deposition. The modeling assumed that the environmental factors affecting the
formation of methylmercury remain constant. EPA acknowledges that a lack of
knowledge about methylation is "a major contributor to overall uncertainty" in its
analysis; however, the effect of this uncertainty on the Agency's ability to inform
mercury control policies is highly variable. An EPA official stated that variance
in methylation rates was taken into account because actual methylmercury fish
tissue measurements, which reflect varying methylation rates among different
waterbodies where measurements were obtained, were used in the "utility-
attributable" hotspot analysis. As explained in the next section, we found that
concerns remain about these fish tissue measurements, which call into question
how well they address methylation uncertainties.
Bioaccumulation. EPA's analysis did not fully account for the highly variable
ways that mercury bioaccumulates in fish. When mercury deposition to a
waterbody changes because of reductions in emissions, it can take time for those
changes to be reflected in fish tissue methylmercury concentrations. Fish absorb
methylmercury from their food and directly from water as it passes over their
gills. To predict levels of methylmercury in fish tissue, CMAQ deposition results
for a given area were input into a model that assumed a proportional relationship
between declines in atmospheric mercury deposition and declines in mercury fish
tissue concentrations. For example, a 50-percent decrease in mercury deposition
rates was projected to lead to a 50-percent decrease in mercury concentrations in
fish. However, drawing conclusions and making comparisons between different
fish types is limited in that mercury bioaccumulates in highly variable ways
among fish, both between species and within individual fish of a species. To
establish a 2001 baseline estimate of methylmercury fish tissue concentrations,
EPA used data from the National Listing of Fish Advisories and the National
Lake Fish Tissue Survey:
For included locations, samples for the same species are averaged across all
available years (post 1998), and then the highest averaged per species
concentration is used to represent the methylmercury concentration for that
11
-------
sample location. For example, if there are two species at a location, walleye
and pike, with three sampling dates for each species, we would first average
over the three sample dates for each species, and then select walleye if the
average for -walleye is highest, or select pike if the average for pike is highest.
. .. Assignment of the maximum average species concentration recognizes the
greater risk to an individual consuming species -with higher accumulation of
mercury while respecting the fact that each sample for an individual species is
only an estimate of the true mean concentration in that species.
According to EPA staff, the adequacy of current fish tissue data is sparse - it is
patchy, non-standardized from State to State, and only identifies potential
problems where data were actually collected. Regarding EPA's fish tissue data,
an Agency official said, "The data does not support the conclusion that CAMR
will not cause hotspots." In its Effectiveness Technical Support Document, EPA
states that, among other limitations, the model it used to estimate changes in
methylmercury fish tissue concentrations does not account for the time lag
between a reduction in mercury deposition and a reduction in methylmercury
concentrations in fish tissue. However, the document stated that EPA is unaware
of any other tool for performing a national-scale assessment of the change in fish
methylmercury concentrations resulting from reductions in atmospheric
deposition of mercury.
Study Shows Significant Deposition from Local Sources
Results from the Steubenville Study,5 a multiyear study in the Ohio River Valley,
found that approximately 70 percent of mercury wet deposition at Steubenville,
Ohio in 2003 and 2004 was attributable to local/regional coal combustion sources,
predominantly from utility boilers.6 The results of the Steubenville Study suggest
that additional monitoring is necessary to ensure that EPA's CAMR analysis has
properly estimated the contribution of local and regional mercury deposition.
For example, while CMAQ results do not provide an estimate of mercury wet
deposition for Steubenville specifically (due to its 36 km x 36 km grid cell area),
it estimated for 2001 that 44 percent of the wet deposition in the grid cell
containing Steubenville was from coal-fired utilities. Spatial and temporal
differences7 between the Steubenville Study and EPA's CAMR analysis do not
allow for their results to be fully comparable; however, data from other
monitoring sites further suggest that monitoring is needed to ensure that CMAQ
5 A peer review of the Steubenville Study was completed in late December 2005 and it was submitted for
publication in a scientific journal in February 2006.
6 The Steubenville Study results have an uncertainty bound of approximately 15 percent. This uncertainty bound
does not follow a normal distribution pattern but is positively skewed, i.e., the upper bound of the 95 percent
confidence interval extends further from the estimate than the lower bound.
7 The Steubenville Study wet deposition results are for 2003 and 2004 and (1) represent the wet deposition for a
specific monitoring location; (2) include wet deposition for all coal-combustion sources; and (3) have quantified
estimates of uncertainty. Conversely, the CMAQ results are for the year 2001 and (1) represent an estimate for a
much larger area (i.e., a 36 km x 36 km grid cell); (2) represent deposition from coal-fired utilities only; and (3) do
not quantify uncertainty.
12
-------
has not underestimated wet deposition in some locations. An Agency scientist
noted that:
.. .CMAQ runs conducted using 2001 emissions data for CAMR modeling
showed that there are areas in the U.S. where domestic sources create large
areas of enhanced deposition (e.g., up to 60% of wet mercury deposition in
some areas originated from domestic coal combustion sources). The
Steubenville measurements are consistent with these projections. As an
example of uncertainties related to CMAQ... the University of Michigan has
run a network of event-based mercury monitoring sites in the Midwest and
Vermont and the 2001 CMAQ model runs systematically underestimate the
deposition observed at these sites (in some cases by over a factor of 2).
Senior Office of Air and Radiation officials told us that the Steubenville area is
known to have higher-than-average deposition from coal-fired utilities, and that
the preliminary monitoring results were not unexpected. OAQPS noted that for
grid cells neighboring the Steubenville grid cell, the CMAQ model predicted that
a higher percentage of mercury deposition was attributable to utility coal
combustion (i.e., 57 to 71 percent). Preliminary results of the Steubenville Study
were made available to Agency officials shortly after EPA's promulgation of
CAMR and the Revision of December 2000 Regulatory Finding in March 2005,
but were not available for consideration by the Agency during its promulgation of
these rules. The Agency noted that they analyzed a number of scientific studies in
developing CMAQ, but our evaluation did not consider all of the scientific
evidence EPA used in developing CMAQ. As noted in Appendix A, we did not
evaluate all the inputs and assumptions associated with EPA's mercury hotspots
analysis. Additional limitations of our evaluation are listed in Appendix A.
Uncertainties Underscore Need for Mercury Monitoring Plan
In the preamble to the Revision of December 2000 Regulatory Finding, EPA
stated that although it believed the likelihood of a "utility-attributable" hotspot
occurring to be "remote," it intended to closely monitor the potential for hotspots,
continue to advance the state of the science of mercury fate and transport, and
take appropriate action if the possibility of a "utility-attributable" hotspot arose
after implementation of CAMR. However, at the time we completed our field
work, EPA had not yet developed a plan for monitoring hotspots. Given the
uncertainties associated with the inputs to the CMAQ model and the results of
recent studies as noted, it is important for EPA to have a plan to monitor mercury
deposition. Mercury monitoring data could assist the Agency in determining
"utility-attributable" hotspots, and in evaluating and improving the accuracy of its
mercury fate and transport models. Without a mercury monitoring plan, "utility-
attributable" hotspots could potentially occur after implementation of CAMR but
be less likely to be identified due to a lack of deposition data or reliable modeling
techniques to identify mercury sources.
13
-------
Field measurement of mercury deposition could improve EPA's ability to conduct
source apportionment studies to help to determine whether a hotspot was "utility-
attributable." To assess whether CAMR results in "utility-attributable" hotspots,
EPA must have mercury deposition data that enable it to identify the mercury
source. Source-apportionment studies, such as that conducted by EPA in
Steubenville, are designed to accomplish this task. Such studies estimate a
source's contribution to mercury deposition and require the collection of
deposition samples and measurements of trace elements in addition to mercury.
Trace elements are elements that are co-emitted with mercury from particular
sources, and help identify from which source(s) the deposited mercury originally
came. For example, sulfur and selenium are trace elements associated with coal
combustion. When these elements are in samples of deposited mercury, they
indicate the mercury came from coal combustion sources. By employing a
monitoring plan that incorporates more studies of this nature, EPA can better
assess the impact that utilities have on mercury deposition and resulting fish tissue
concentrations.
Mercury deposition data would also help EPA improve its current understanding
of mercury fate and transport, and allow the Agency to validate and improve
mercury deposition estimation models and techniques. Model performance can
be assessed by comparing model predictions to actual field data. While mercury
deposition data are available through the Mercury Deposition Network (MDN),
these data have important limitations for model evaluation, particularly modeling
designed to identify mercury hotspots:
The MDN measures only wet deposition because there is no adequate field
methodology currently available for dry deposition.
The MDN does not generally provide deposition monitoring data for areas
expected to be of greatest concern for deposition from local emissions
sources. This is because MDN monitoring sites are generally located in
rural locations that do not have local sources of emissions.
There are large areas of the nation with few or no MDN monitoring sites.
The MDN collects deposition samples on a weekly basis, so it does not
accurately measure the impacts of individual events, such as rain or
snowfall.
The MDN sites do not collect trace element data, such as sulfur and
selenium data for coal combustion, which is needed to conduct source
apportionment modeling.
Due to the limitations associated with available data from the current mercury
deposition monitoring network, EPA is currently unable to fully assess the
accuracy of CMAQ's mercury deposition predictions against actual field
14
-------
measurements. Agency officials told us that the EPA Office of Research and
Development's National Exposure Research Laboratory was already
implementing a research plan for mercury monitoring, but recent budget
reductions have halted the program.
Conclusion
EPA has acknowledged uncertainties and limitations in its analysis of the
potential for "utility-attributable" hotspots. The results from two studies - the
Mechanisms of Mercury Removal Study and the Steubenville Study - illustrate
uncertainties about some of the key assumptions used in CMAQ and the
deposition results projected by the model. Further consideration of uncertainties
could alter EPA's conclusions about the potential for "utility-attributable"
mercury hotspots. EPA indicated it will closely monitor hotspots, continue to
advance mercury science, and take appropriate actions if hotspots arose. To
accomplish this, the Agency needs to establish a monitoring plan to conduct
source-apportionment studies to measure the impact of CAMR and to assist in
evaluating the accuracy of its model predictions against actual field data.
Recommendation
We recommend that the Acting Assistant Administrator for Air and Radiation:
2-1 Work with the Assistant Administrator for the Office of Research and
Development to develop and implement a mercury monitoring plan,
including milestones and responsible program offices for implementing
each component of the plan, to: (1) assess the impact of CAMR, if
adopted, on mercury deposition and fish tissue; and (2) evaluate and
refine, as necessary, mercury estimation tools and models. This effort
should consider the suitability of the Office of Research and
Development's mercury research plan for addressing these objectives.
Agency Comments and OIG Evaluation
The Agency generally agreed with the recommendation in Chapter 2 of the report.
However, the Agency expressed concern with our characterization of some
scientific issues in the report and offered clarification on three specific issues. We
accepted the Agency's technical clarifications and have made changes to the final
report as appropriate. The Agency also provided us with additional concerns not
specifically addressed in its written response to our draft report. We met with the
Agency to discuss these concerns, and made changes to the final report as
appropriate. In response to our recommendation, the Agency stated that the
Office of Air and Radiation and Office of Research and Development will
continue to work together to ensure that they are using the best possible
information to assess the transport, transformation, deposition, and fate of
mercury emissions in the United States. We support the Agency's commitment to
15
-------
using the best possible information to assess the impact of mercury emissions in
the United States, and continue to recommend the Agency develop a monitoring
plan to better ensure that this happens. The Agency's formal written response is
in Appendix C.
16
-------
-': -.. *'.],..,. IghapterS - :
EPA Needs to'Clari^Gonditions Under Which CAMR
, - - ,», ,-;,,» w .',' , . ,iffc«*ff » ;, !,j,,*s*'A!*i'<-i»«'i*i«i'y/ " '
-------
measure for "hazards to public health," EPA adopted the "utility-attributable"
hotspots definition to determine whether such utility hotspots would remain after
implementation of the Clean Air Interstate Rule and CAMR. Based on the
analysis described in Chapter 2 of this report, EPA stated that it did not believe
that "utility-attributable" mercury hotspots would exist after implementing these
rules, therefore supporting the Agency's decision that utilities did not need to be
regulated under Section 112 of the Clean Air Act.
"Utility-Attributable" Definition Could be Interpreted to Limit
EPA's Ability to Mitigate Hotspots
The "utility-attributable" definition could be interpreted to limit EPA's ability to
address waterbodies with elevated levels of mercury unless utility emissions were
the sole cause of the problem. This could in turn limit EPA's ability to reduce the
number of waterbodies with fish consumption advisories where there is a health
risk due to the combined impact of mercury from all sources, including air
emissions.
As discussed in its December 1997 Mercury Study Report to Congress, EPA
stated that there is "clearly" a need to address the combined impacts of mercury
originating from all sources, including air emissions, wherever the combination of
sources have been related to unacceptably high mercury levels in fish. Further, in
its December 2000 Finding, EPA recognized concerns about the potential local
impact of mercury trading programs and acknowledged that:
. . . approaches that involve economic incentives must be constructed in a
way that assures that communities near the sources of emissions are
adequately protected.
Within CAMR and the Revision of December 2000 Regulatory Finding, EPA
specifies several actions it might take to mitigate the effects of a hotspot in the
event one should be identified. However, our analysis of the revision and CAMR
suggests that the Agency may be precluded from taking any of those actions
unless the hotspot first meets the criteria of a "utility-attributable" hotspot. EPA
officials told us that this was not the intent of the rule, but agreed that the rule
could be clearer.
CAMR and Revision Must be Read Together
Based on our reading of CAMR and the Revision of December 2000 Regulatory
Finding, we conclude that the definition of a hotspot presented in the revision is
intended to apply to CAMR. CAMR and the revision were issued on the same
day and address the same subject matter. In addition, the preamble to the CAMR
restates EPA's conclusion from the revision, but refers to it as part of "this
action":
18
-------
As stated elsewhere in this action EPA does not believe that utility-
attributable hot spots will be an issue after implementation ofCAIR
[Clean Air Interstate Rule] and CAMR.
Because "utility-attributable" hotspots are not discussed "elsewhere" within
CAMR, we conclude that "this action" refers to the other, closely related action
published by EPA on the same day. This action, the Revision of December 2000
Regulatory Finding, defines "utility-attributable" hotspots and also explains that
EPA may address hotspots under "other authorities under the CAA [Clean Air
Act]," should they occur. However, the only mechanism to which EPA refers in
order to addresses potential future hotspots - and the only mechanism presently
promulgated - is CAMR. The revision cites the following ways it could address
"utility-attributable" hotspots:
.. . if in the future we determine that utility-attributable hotspots exist and
that those hotspots occur as the result ofHg emissions from coal-fired
Utility Units, we may promulgate a tighter section 111 standard of
performance, provided we determine the technology can achieve the
contemplated reductions. We could revise the standard of performance by
adjusting the cap-and-trade program to limit trading by high-emitting
Utility Units. .. . Thus, although we cannot conclude today which
statutory authority we would implement to address utility attributable
hotspots because that determination necessarily hinges on the facts
associated with the identified hotspots, we do conclude that were such a
situation to occur, we believe that EPA has adequate authority to address
any such situation that may arise in the future.
When read together, these regulatory actions suggest that a finding of a solely
"utility-attributable" mercury hotspot is necessary to initiate Agency action to
mitigate hotspots under CAMR. If this were the case, EPA would be precluded
from requiring additional mercury reductions from the utility industry, even if it
were determined that utilities were significantly contributing to a hotspot, if the
utilities were not the sole cause of the hotspot. For example, if methylmercury
fish tissue concentrations for a waterbody were at 0.32 mg/kg, EPA's water
quality criterion of 0.3 mg/kg would be exceeded. If, in this hypothetical case,
utility mercury emissions were causing 0.3 mg/kg or less of the total
methylmercury, under the requirement as written, utilities would be excluded
from any additional reductions to help mitigate the problem.
We discussed our interpretation with Office of Air and Radiation officials. These
officials confirmed that the "utility-attributable" hotspot definition in the revision
applies to the CAMR. However, these officials told us that this definition does
not establish a criterion for when the Agency can adjust the performance
standards under CAMR. They noted that under Section 111, performance
standards are to be reviewed every 8 years, and can be adjusted for various
reasons.
19
-------
Conclusion
The two rules related to controlling mercury emissions from coal-fired utilities
were issued on the same day and refer to and are consistent with each other.
Thus, it appears that they are intended to be read together. Further, the "utility-
attributable" definition in the Revision of December 2000 Regulatory Finding
applies to the discussion of hotspots in CAMR, and this definition establishes a
criterion for when the CAMR can be adjusted to address a potential health hazard.
If this were the case, tighter performance standards for utilities contributing to a
hotspot could not be promulgated unless it was first determined that the hotspot
was solely "utility-attributable." Although not the intent of the rulemaking, EPA
officials agreed that the rule could be clearer. We believe CAMR, if adopted,
should be clarified to avoid any possible misinterpretation of how the "utility-
attributable" definition affects EPA's ability to modify utility performance
standards.
Recommendation
If EPA decides to adopt CAMR after the rule reconsideration process, to better
ensure protection of public health and the environment, we recommend that the
Acting Assistant Administrator for Air and Radiation:
3-1 Explain in CAMR that the "utility-attributable" hotspot definition found in
the revision does not establish a prerequisite for making future changes to
the performance standards under CAMR.
Agency Comments and OIG Evaluation
The Agency's response did not specifically address our analysis and conclusion
that CAMR could be interpreted to use the "utility-attributable" hotspot definition
as a prerequisite for future changes to CAMR. The Agency commented that,
while information regarding "utility-attributable" hotspots would be relevant to
future possible revisions to CAMR, such hotspots are not a prerequisite to the
Agency making changes to performance standards under CAMR. We believe the
Agency's intent should be made clear in the final rule. Accordingly, we revised
our final report to recommend that EPA, to better ensure protection of public
health and the environment, explain in the CAMR that the "utility-attributable"
hotspot definition set forth in the revision is not a prerequisite for making changes
to the CAMR. After submitting its formal written response to the draft report the
Agency also suggested clarifying language to parts of Chapter 3. We accepted
some of the suggestions and incorporated them into the final report. The
Agency's formal written response is in Appendix C.
20
-------
Appendix A
Details on Scope and Methodology
We conducted interviews with staff from the following EPA offices:
Office of Air and Radiation, including its Office of Air Quality Planning and Standards
and Office of Atmospheric Programs.
Office of Research and Development, including its National Exposure Research
Laboratory and National Center for Environmental Research.
Office of Policy, Economics, and Innovation.
Office of Water.
We also interviewed officials from the following external organizations: the National Oceanic
and Atmospheric Administration, including its Air Research Laboratory; the Northeast States for
Coordinated Air Use Management; and the Clean Air Task Force.
To understand the variables associated with mercury fate and transport modeling and,
specifically, CMAQ, we reviewed and/or discussed with the above officials selected reports and
studies, including:
Mechanisms of Mercury Removal by Os and OH in the Atmosphere. Calvert Jack G.;
Lindberg Steve E. Atmospheric Environment, Volume: 39, Number: 18, Page: 3355-
3367, June 5, 2005.
The National Oceanic and Atmospheric Administration's September 2005 (draft) report
by Cohen, et al, Report to Congress: Mercury Contamination in the Great Lakes.
EPA's Regulatory Impact Analysis of the Final Clean Air Mercury Rule, March 2005.
EPA's Mercury Study Report to Congress, December 1997.
A slide presentation on EPA's Steubenville, Ohio, study, Preliminary Results from
Steubenville Hg Deposition Source Apportionment Study, April 27, 2005.
The most recent peer review of CMAQ, Final Report: Second Review of the CMAQ
Model.
Technical Support Document: Methodology Used to Generate Deposition, Fish Tissue
Methylmercury Concentrations, and Exposure for Determining Effectiveness of Utility
Emission Controls. Analysis of Mercury from Electricity Generating Units, March 17,
2005 (revised).
Technical Support Document for the Final Clean Air Mercury Rule: Air Quality
Modeling, March 2005.
Emissions Inventory and Emissions Processing for the Clean A ir Mercury Rule, March
2005.
21
-------
To gain an understanding of State and environmental groups' concerns related to EPA's analysis of
potential "utility-attributable" hotspots under CAMR, we reviewed the following selected comments:
« The December 19, 2005 comments submitted In Reconsideration of: Revision of
December 2000 Regulatory Finding on the Emissions of Hazardous Air Pollutants From
Electric Utility Steam Generating Units and the Removal of Coal- and Oil-Fired Electric
Utility Steam Generating Units from the Section 112(c) List 70 Fed. Reg. 62200 (Oct.
28, 2005); and Standards of Performance for New and Existing Stationary Sources:
Electric Utility Steam Generating Units 70 Fed. Reg. 62213 (Oct. 28, 2005). Comments
Submitted by: The States of New Jersey, California, Connecticut, Delaware, Illinois,
Maine, Massachusetts, Minnesota, New Hampshire, New Mexico, New York,
Pennsylvania, Rhode Island, Vermont, and Wisconsin, Docket No. OAR-2002-0056.
The December 19, 2005 comments submitted regarding the "Revision of December 2000
Regulatory Finding on the Emissions of Hazardous Air Pollutants From Electric Utility
Steam Generating Units and the Removal of Coal- and Oil-Fired Electric Utility Steam
Generating Units From the Section 112(c) List: Reconsideration, " 70 Fed. Reg. 62,200
(October 28, 2005). Comments of: Clean Air Task Force, Izaak Walton League of
America, Natural Resources Council of Maine, Ohio Environmental Council, U.S. Public
Interest Research Group, Natural Resources Defense Council, Chesapeake Bay
Foundation, Waterkeeper, Aroostook Band ofMicmac Indians, Houlton Band ofMaliseet
Indians, Penobscot Indian Nation, The Passamaquoddy Tribe at Indian Township.
To gain an understanding of EPA's definition of "utility-attributable" hotspots and the basis for
that definition, we reviewed the following regulatory actions:
Regulatory Finding on the Emissions of Hazardous Air Pollutants From Electric Utility
Steam Generating Units, December 20, 2000.
Proposed National Emission Standards for Hazardous Air Pollutants; and, in the
Alternative, Proposed Standards of Performance for New and Existing Stationary
Sources: Electric Utility Stream Generating Units; Proposed Rule, January 30, 2004.
Final Rule - Preamble - Standards of Performance for New and Existing Stationary
Sources: Electric Utility Steam Generating Units, March 15, 2005.
Final Rule - Regulatory Text - Standards of Performance for New and Existing
Stationary Sources: Electric Utility Steam Generating Units, March 15, 2005.
Final Rule - Revision of December 2000 Regulatory Finding on the Emissions of
Hazardous Air Pollutants from Electric Utility Steam Generating Units and the Removal
of Coal- and Oil-fired Electric Utility Steam Generating Units from the Section 112 (c)
List, March 15, 2005.
Reconsideration: Revision of December 2000 Regulatory Finding on the Emissions of
Hazardous Air Pollutants from Electric Utility Steam Generating Units and the Removal
of Coal- and Oil-fired Electric Utility Steam Generating Units from the Section 112 (c)
List, October 21,2005.
Reconsideration: Standards of Performance for New and Existing Stationary Sources:
Electric Utility Steam Generating Units, October 21, 2005.
22
-------
Prior Coverage
In a prior EPA OIG report, Additional Analyses of Mercury Emissions Needed Before EPA
Finalizes Rules for Coal-Fired Electric Utilities (2005-P-00003), dated February 3, 2005, we
cited concerns about EPA's limited assessment of the potential for mercury hotspots resulting
from its (then proposed) cap-and-trade program under CAMR. In that report, we recommended
that EPA further assess the risk of hotspots and, if necessary, identify how the Agency would
reassess the hotspot issue. In response to our recommendation, EPA stated that it did not
believe utility emissions would result in hotspots based on additional analyses it had performed,
particularly after implementation of the Clean Air Interstate Rule and CAMR, but it would
monitor the situation and take action if necessary. For this current review, we evaluated EPA's
analysis of hotspots, its conclusion that there will be no "utility-attributable" hotspots after
implementation of the Clean Air Interstate Rule and CAMR, and plans the Agency may have in
place to continue to monitor the issue. Details on what we found, including recommendations,
are in Chapters 2 and 3 of this current report.
Internal Controls
Government Auditing Standards require that auditors obtain an understanding of internal
control significant to the audit objectives and consider whether specific internal control
procedures have been properly designed and placed in operation. This evaluation was a
limited-scope assessment of certain analyses pertaining to a rulemaking. Thus, we determined
whether the Agency's hotspots analysis and conclusions were peer reviewed, and if the key
model used in this analysis was separately peer reviewed. Peer review is a key internal control
for ensuring the acceptability of scientific data and processes. We found that CMAQ, the main
model used by EPA in its hotspots analysis, was peer reviewed; however, we found no evidence
that the Agency's overall hotspots analysis, described in the document Methodology to
Generate Deposition, Fish Tissue Methylmercury Concentrations, and Exposures for
Determining Effectiveness of Utility Emission Controls, was peer reviewed. The Agency's
Ecosystem Scale Modeling for Mercury Benefits Analysis, part of the Regulatory Impact
Analysis for the CAMR, was peer reviewed. The benefits analysis was similar to the hotspots
analysis, but it assessed the impact of CAMR on a national scale, as opposed to identifying
localized hotspots or local-scale impacts.
Limitations
Our work had several limitations. Specifically, we did not:
Review every model that contributed to EPA's analysis of the potential for "utility-
attributable" hotspots under CAMR.
Evaluate all of the inputs and assumptions'associated with EPA's mercury hotspots
analysis.
Evaluate the adequacy of EPA's water quality criterion to protect human health.
23
-------
Appendix B
Models Used in CAMR Analysis
The following diagram depicts how data from each model were used in EPA's hotspot analysis.
Details on each model follow the diagram.
MM5
4
IPM -»CMAQ -» MMaps - Changes in methylmercury levels in fish tissue
T
GEOS-CHEM
Integrated Planning
Model (IPM)
To analyze future cost and emissions
impacts of proposed environmental
regulations upon utilities.
Estimates mercury emissions
from utilities after implementation
of Clean Air Interstate Rule and
CAMR.
Mesoscale Model (MM5)
To provide meteorological information,
such as wind, temperature, precipitation,
and sea level pressure.
Simulates weather patterns,
which affect where mercury
deposits.
Goddard Earth Observing
System-CHEMistry
(GEOS-CHEM) Global
Model
To provide a global three-dimensional
model of atmospheric chemistry driven by
meteorology.
Uses global chemistry and
transport information to provide
global/background mercury
concentrations.
Community Multiscale Air
Quality (CMAQ) Model
To estimate mercury deposition.
To simulate various chemical and physical
processes thought to be important in the
atmospheric transformation and distribution
of mercury.
Estimates amount of mercury
deposition occuring within 36 km2
grid cells after implementation of
Clean Air Interstate Rule and
CAMR.
Mercury Maps (MMaps)
To relate changes in mercury air deposition
rates to changes in mercury fish tissue
concentrations on a national scale.
Uses CMAQ deposition data to
estimate fish tissue
concentrations of methylmercury
based on the assumption of a
1-to-1 ratio between reductions in
air deposition and reductions in
average methylmercury fish
tissue concentrations.
24
-------
Appendix C
Agency Response to Draft Report
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, D.C. 20460
Bill Roderick, Acting Inspector General
Office of the Inspector General
Office of Program Evaluation
1301 Constitution Ave. NW (2400 T)
EPA West Building
Washington, DC 20004
Dear Mr. Roderick:
Thank you for the opportunity to comment on the draft Office of the Inspector General
(OIG) report entitled "Monitoring Needed to Assess Impact of EPA's Clean Air Mercury Rule
on Potential Hotspots." In reviewing the draft report, we acknowledge your acceptance of the
majority of the issues we identified in our earlier review. We have also recently supplied your
office with additional written comments pertaining to the modeling analyses associated with the
Steubenville project. We believe that the collective scientific and engineering expertise within
EPA's Offices of Air and Radiation (OAR) and Research and Development (ORD) puts our
offices in a unique position to assess the current state-of-the-science with respect to mercury
transport, deposition, and fate, and its impact on the creation of utility-attributable hotspots.
We continue to have concerns about the portrayal of some scientific issues in the report,
and note three areas where we would like to provide clarifying remarks. First, with respect to the
potential changes in the atmospheric reaction rates within the Community Multiscale Air Quality
(CMAQ) model (see pages 9 and 10), such changes would be made uniformly in all mercury
transport/deposition models, not just CMAQ. Thus, the enhancements would create different
results in any assessment using these numerical simulation technologies. Second, regarding our
need to improve ambient monitoring (see page 13), the report should acknowledge that the
Mercury Deposition Network (MDN) currently measures only wet deposition because there is no
adequate field methodology available for dry deposition.
25
-------
Finally, in terms of how EPA addressed the uncertainties in methylation and
bioaccumulation rates between different fresh water bodies, our supporting health benefits
assessment materials describe in great detail our complete understanding of these processes.
You are correct to point out, and we clearly acknowledge in our documents, the uncertainties
associated with mercury transport, deposition, and effects. At the same time, it should be
acknowledged that the magnitude of uncertainties and their effect on our ability to inform
mercury control policies is highly variable. We believe we have clearly explained the science
and the uncertainties and provided a solid foundation for the Clean Air Mercury Rule (CAMR).
In your draft report, you recommend two specific follow-up actions for the Agency.
Below we address each of these recommendations.
Recommendation 2-1: Work with the Assistant Administrator for the Office of Research
and Development to develop and implement a mercury
monitoring plan, including milestones and responsible program
offices for implementing each component of the plan, to: (1)
assess the impact of CAMR, if adopted, on mercury deposition
and fish tissue, and, (2) evaluate and refine, as necessary,
mercury estimation tools and models. This effort should
consider the suitability of the Office of Research and
Development's mercury research plan for addressing these
objectives.
EPA currently operates the MDN, which is located predominantly in the eastern U.S. and
monitors only wet deposition. In the technical support documents supporting CAMR, EPA
has continually highlighted the need for and the willingness to support additional ambient
monitoring, including the development of dry deposition monitoring, to enhance our ability to
assess the numerical accuracy of our sophisticated simulation tools - e.g., the CMAQ model.
As you are aware, ORE) has been heavily involved over the past decade in developing the
CMAQ model, and is actively engaged in utilizing ambient data and the latest scientific
information to update the model to reflect the best possible chemistry and physics. OAR and
ORD will continue to work together to ensure that we are using the best possible information
to assess the transport, transformation, deposition, and fate of mercury emissions in the U.S.
Recommendation 3-1: If EPA decides to adopt CAMR after the rule reconsideration
process, we recommend that the Acting Assistant Administrator
for Air and Radiation: Specifically explain what role the "utility-
attributable" hotspot definition has in determining whether to
make any future changes to the performance standards under
CAMR.
EPA has explained to your staff that while information regarding utility-attributable
hotspots would be relevant to future possible revisions to CAMR, such hotspots are not a
prerequisite. CAMR controls are based on the new source performance standards (NSPS) as
set forth in section 111 of the Clean Air Act. To this end, the Agency is required by law to
review and revise, as necessary, these limits every eight years. In conducting such a review,
26
-------
we will analyze and evaluate the availability of new mercury control technologies installed
since the previous review, and to the extent they provide additional cost-effective control, the
Agency can move to change the existing NSPS limits. Additionally, the Agency continues to
update its understanding of the science associated with mercury emissions, transport,
transformation, and deposition, both from ambient data collection and monitoring and
through continued enhancements to our analytical tool box. Thus, we feel that OAR and
ORD are uniquely positioned to monitor this situation and provide the best possible solution
for the protection of public health and the environment.
In closing, we direct the OIG staff to the numerous technical documents supporting
the final CAMR, particularly the benefits assessment materials in which we outline in detail
the variability associated with methylation and bioaccumulation rates in different water
bodies. In these documents, EPA has demonstrated that the conclusions reached in the
CAMR are based firmly in sound scientific principles, utilizing the best information
available. If your staff have additional questions in researching these documents, our
scientists, engineers, and modelers would be happy to assist them.
Sincerely,
William Wehrum
Acting Assistant Administrator
Office of Air and Radiation
George Gray
Assistant Administrator
Office of Research and Development
27
-------
Appendix D
Distribution
Office of the Administrator
Acting Assistant Administrator for Air and Radiation
Assistant Administrator for Research and Development
Deputy Assistant Administrator for Air and Radiation
Acting Deputy Assistant Administrator for Science, Office of Research and Development
Acting Deputy Assistant Administrator for Management, Office of Research and Development
General Counsel
Agency Followup Official (the CFO)
Agency Followup Coordinator
Audit Followup Coordinator, Office of Air and Radiation
Audit Followup Coordinator, Office of Research and Development
Associate Administrator for Congressional and Intergovernmental Relations
Associate Administrator for Public Affairs
Director, Office of Air Quality Planning and Standards
Deputy Director, Office of Air Quality Planning and Standards
Audit Liaison, Office of Air Quality Planning and Standards
Acting Inspector General
28
------- |